Cardiac Electrophysiology, From Cell To Bedside, 6th Edition

Sixth Edition
Cardiac
Electrophysiology
From Cell to Bedside
DOUGLAS P. ZIPES, MD
Distinguished Professor
Emeritus Professor of Medicine, Pharmacology, and Toxicology
Emeritus Director, Division of Cardiology and the Krannert Institute of Cardiology
Indiana University School of Medicine
Indianapolis, Indiana
Editor-in-Chief, HeartRhythm
Editor-in-Chief, PracticeUpdate/Cardiology
JOSÉ JALIFE, MD
Cyrus and Jane Farrehi Professor of Cardiovascular Research
Professor of Internal Medicine
Professor of Molecular and Integrative Physiology
University of Michigan Medical School
Co-Director, University of Michigan Center for Arrhythmia Research
Ann Arbor, Michigan
With 742 illustrations

1600 John F. Kennedy Blvd.
Ste 1800
Philadelphia, PA 19103-2899
Cardiac Electrophysiology: From Cell to Bedside ISBN: 978-1-4557-2856-5

Copyright © 2014, 2009, 2004, 2000, 1995, 1990 by Saunders, an imprint of Elsevier Inc.
All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any
means, electronic or mechanical, including photocopying, recording, or any information storage and
retrieval system, without permission in writing from the publisher. Details on how to seek permission
and further information about the Publisher’s permissions policies and our arrangements with
organizations such as the Copyright Clearance Center and the Copyright Licensing Agency can be
found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the
Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience
broaden our understanding, changes in research methods, professional practices, or medical
treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described herein. In
using such information or methods they should be mindful of their own safety and the safety of
others, including parties for whom they have a professional responsibility.
With respect to any drug or pharmaceutical products identified, readers are advised to check the
most current information provided (i) on procedures featured or (ii) by the manufacturer of each
product to be administered, to verify the recommended dose or formula, the method and duration of
administration, and contraindications. It is the responsibility of practitioners, relying on their own
experience and knowledge of their patients, to make diagnoses, to determine dosages and the best
treatment for each individual patient, and to take all appropriate safety precautions.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors
assume any liability for any injury and/or damage to persons or property as a matter of products
liability, negligence or otherwise, or from any use or operation of any methods, products,
instructions, or ideas contained in the material herein.
Library of Congress Cataloging-in-Publication Data
Cardiac electrophysiology (2014)

Cardiac electrophysiology : from cell to bedside / [edited by] Douglas P. Zipes, Jose Jalife.—Sixth
edition.
p. ; cm.
Includes bibliographical references and index.
ISBN 978-1-4557-2856-5 (hardcover : alk. paper)
I. Zipes, Douglas P., editor of compilation. II. Jalife, Jos?, editor of compilation. III. Title.
[DNLM: 1. Arrhythmias, Cardiac. 2. Electrophysiologic Techniques, Cardiac.
3. Heart—physiology. WG 330]
RC685.A65
616.1′28—dc23
2013020141
Senior Editor: Dolores Meloni

Senior Content Development Specialist: Janice Gaillard
Publishing Services Manager: Catherine Jackson
Senior Project Manager: Rachel E. McMullen
Design Direction: Louis Forgione
Printed in Canada
Last digit is the print number: 9 8 7 6 5 4 3 2 1

Contributors
Hugues Abriel, MD, PhD Justus M. B. Anumonwo, PhD

Professor of Pathophysiology Department of Internal Medicine
Director, Department of Clinical Research Department of Molecular and Integrative Physiology
University of Bern University of Michigan
Bern, Switzerland Ann Arbor, Michigan
Wayne O. Adkisson, MD Luciana Armaganijan, MD, PhD, MHS

Department of Medicine Electrophysiologist
Cardiac Arrhythmia Center Division of Electrophysiology and Cardiac Arrhythmias
Cardiovascular Division Dante Pazzanese Institute of Cardiology
University of Minnesota Medical School São Paulo, Brazil
Minneapolis, Minnesota
Hiroshi Ashikaga, MD, PhD
Esperanza Agullo-Pascual, PhD Division of Cardiology
Postdoctoral Fellow Department of Biomedical Engineering
Department of Medicine Johns Hopkins University School of Medicine
Leon H Charney Division of Cardiology Baltimore, Maryland
New York University School of Medicine
New York, New York Felipe Atienza, MD, PhD
Assistant Professor
Olujimi A. Ajijola, MD, PhD School of Medicine
Clinical Instructor Universidad Complutense de Madrid
UCLA Cardiac Arrhythmia Center Senior Electrophysiologist
UCLA Health System Cardiology Department
Los Angeles, California Hospital General Universitario Gregorio Marañón
Madrid, Spain
Amin Al-Ahmad, MD
Assistant Professor Uma Mahesh R. Avula, MD
Division of Cardiovascular Medicine Center for Arrhythmia Research
Director Cardiac Electrophysiology Laboratory Internal Medicine/Cardiology
Stanford University University of Michigan
Stanford, California Ann Arbor, Michigan
Oluseun Alli, MD Peter H. Backx, DVM, PhD

Assistant Professor Medicine Professor Physiology and Medicine
Director Structural Heart Program University of Toronto
Division of Cardiology Toronto, Ontario, Canada
University of Alabama at Birmingham, University of Alabama
Hospital Elise Balse, PhD
Birmingham, Alabama Université Pierre et Marie Curie-Paris 6–INSERM
ICAN, Institute of Cardiometabolism and Nutrition
Robert K. Altman, MD Paris, France
Attending Physician
Al-Sabah Arrhythmia Institute Conor D. Barrett, MD
St. Luke’s Roosevelt Hospital Director
New York, New York Al-Sabah Arrhythmia Institute
St. Lukes’ Roosevelt Hospital Center
Elad Anter, MD New York, New York
Cardiac Electrophysiologist
Cardiology David G. Benditt, GBScEE, MD
Beth Israel Deaconess Medical Center Professor Cardiovascular Medicine
Instructor University of Minnesota Medical Center
Harvard Medical School Minneapolis, Minnesota
Boston, Massachusetts
Omer Berenfeld, PhD
Charles Antzelevitch, PhD Associate Professor Internal Medicine and Biomedical
Executive Director and Director of Research Engineering
Masonic Medical Research Laboratory University of Michigan
Utica, New York Ann Arbor, Michigan
iii
iv Contributors
Donald M. Bers, PhD Pedro Brugada, MD, PhD

Senior Research Scientist Professor, Chairman
Department of Medicine Cardiovascular Division
Leon H Charney Division of Cardiology UZ Brussel-VUB
New York University School of Medicine Brussels, Belgium
New York, New York
Ramon Brugada, MD, PhD
Charles I. Berul, MD Dean, School of Medicine
Professor Pediatrics and Integrative Systems Biology University of Girona-Spain
George Washington University Cardiovascular Genetics Centre
Division Chief Cardiology University of Girona-IDIBGI
Children’s National Medical Center Girona, Spain
Washington, D.C.
Victoria Brugada, BA
A. Christian Blank, MD Cardiovascular Division
Department of Pediatric Cardiology UZ Brussel-VUB
Wilhelmina Children’s Hospital Brussels, Belgium
University Medical Center
Utrecht, The Netherlands Eric Buch, MD, MS
Assistant Professor of Medicine
Raffaella Bloise, MD UCLA Cardiac Arrhythmia Center
Molecular Cardiology UCLA Health System
IRCCS Fondazione Salvatore Maugeri Los Angeles, California
Pavia, Italy
Feliksas F. Bukauskas, PhD
Frank Matthias Bogun, MD Dr Habil Professor
Associate Professor Medicine Dominick P Purpura Department of Neuroscience
Internal Medicine Albert Einstein College of Medicine
University of Michigan Bronx, New York
Ann Arbor, Michigan
J. David Burkhardt, MD
Martin Borggrefe, PhD, FESC Director of Research
Professor St David’s Medical Center
Department of Medicine Texas Cardiac Arrhythmia Institute
University Medical Centre Mannheim Austin, Texas
Mannheim, Germany
Nenad Bursac, PhD
Noel G. Boyle, MD, PhD, FHRS Associate Professor Biomedical Engineering
Professor of Medicine Faculty Cardiology
UCLA Health system Duke University
Los Angeles, California Durham, North Carolina
Günter Breithardt, MD Hugh Calkins, MD

Department Kardiologie und Angiologie Professor Medicine
Kompetenznetz Vorhofflimmern (AFNET) Director of Electrophysiology
Universitätsklinikum Münster Nicholas J Fortuin Professor of Cardiology
Münster, Germany Johns Hopkins Medical Institutions
Baltimore, Maryland
Marisa Brini, PhD
Department of Biology David J. Callans, MD
University of Padova Associate Director of Electrophysiology
Padova, Italy UPHS Medicine
University of Pennsylvania
Peter R. Brink, PhD Philadelphia, Pennsylvania
Department of Physiology and Biophysics
Institute for Molecular Cardiology Oscar Campuzano, BSc, PhD
Stony Brook University Cardiovascular Genetics Centre
Stony Brook, New York University of Girona-IDIBGI
Girona, Spain
Josep Brugada, MD, PhD
Medical Director Sean M. Caples, MD
Hospital of the University of Barcelona Division of Pulmonary and Critical Care Medicine
Barcelona, Spain Sleep Disorders Clinic
Department of Internal Medicine
Pau Brugada, MSc Mayo Clinic College of Medicine
Cardiovascular Genetics Centre Rochester, Minnesota
University of Girona-IDIBGI
Girona, Spain
Contributors v
Ernesto Carafoli, MD Stuart J. Connolly, MD, FRCPC

Venetian Institute for Molecular Medicine Director, Division of Cardiology
Padova, Italy Salim Yusuf Chair in Cardiology
Population Health Research Institute
Augustin Castellanos, MD Hamilton Health Sciences
Emeritus Professor of Medicine McMaster University
Division of Cardiology David Braley Cardiovascular and Stroke Research Institute
University of Miami Miller School of Medicine Hamilton, Ontario, Canada
Miami, Florida
Jason Constantino, PhD
William Catterall, BA, PhD Research Assistant
Professor & Chair Pharmacology Cardiovascular Research
University of Washington Department of Biomedical Engineering
Seattle, Washington And Institute of Computational Medicine
Johns Hopkins University
Marina Cerrone, MD Baltimore, Maryland
Senior Research Scientist
Department of Medicine Lia Crotti, MD, PhD
Leon H Charney Division of Cardiology Section of Cardiology
New York University School of Medicine Department of Nuclear Medicine
New York, New York University of Pavia
Department of Cardiology and Molecular Cardiology
Lan S. Chen, MD Laboratory
Professor of Clinical Neurology IRCCS Fondazione Policlinico S. Matteo
Department of Neurology Pavia, Italy
Riley Hospital for Children Institute of Human Genetics
Indiana University Helmholtz Zentrum Muenchen
Indianapolis, Indiana Neuherberg, Germany
Lei Chen, MD Frank A. Cuoco, Jr., MD, MBA, MS

Department of Pharmacology Assistant Professor of Medicine
College of Physicians & Surgeons of Columbia University Division of Cardiology
New York, New York Cardiac Electrophysiology
Medical University of South Carolina
Peng-Sheng Chen, MD Charleston, South Carolina
Medtronic Zipes Chair of Cardiology
Director Krannert Institute of Cardiology Anne B. Curtis, MD
Chief Division of Cardiology Charles and Mary Bauer Professor and Chair
Cardiologist Medicine
Cardiovascular Service Line University at Buffalo
Indiana University Health Buffalo, New York
Ralph J. Damiano, Jr., MD
Ashley Chin, MBChB John M. Schoenberg Professor of Surgery
Department of Medicine Chief of Cardiac Surgery
University of Cape Town Vice Chairman
Director of Electrophysiology and Pacing Department of Surgery
Division of Cardiology Barnes-Jewish Hospital
Groote Schuur Hospital Washington University School of Medicine
Cape Town, South Africa St. Louis, Missouri
Aman Chugh, MD Dawood Darbar, MBChB, MD

Section of Electrophysiology Associate Professor Medicine
University of Michigan Health System Director Vanderbilt Arrhythmia Service
Ann Arbor, Michigan Vanderbilt University
Nashville, Tennessee
Ira S. Cohen, MD, PhD
Department of Physiology and Biophysics Mithilesh K. Das, MD
Institute for Molecular Cardiology Associate Professor Clinical Medicine
Stony Brook University Cardiology/Medicine
Stony Brook, New York Krannert Institute of Cardiology
Mario Delmar, MD, PhD

Professor Medicine/Cardiology
New York University
New York, New York
vi Contributors
Eva Delpón, BPharm, PhD Igor R. Efimov, PhD, FA, HA, FHRS
Associate Professor Pharmacology Lucy & Stanley Lopata Distinguished Professor
School of Medicine Biomedical Engineering
Universidad Complutense Washington University
Madrid, Spain Professor Medicine and Radiology
Professor Cell Biology and Physiology
Luigi Di Biase, MD, PhD Washington University School of Medicine
Associate Professor, Department of Medicine, (Cardiology) St. Louis, Missouri
Albert Einstein College of Medicine at Montefiore Hospital
New York, USA; Kenneth A. Ellenbogen, MD
Senior Researcher Kontos Professor Medicine
Cardiology, Electrophysiology Cardiology
Texas Cardiac Arrhythmia Institute at St David’s Medical VCU School of Medicine
Center Richmond, Virginia
Austin, Texas;
Assistant Professor Cardiology Patrick T. Ellinor, MD, PhD
University of Foggia Associate Professor
Foggia, Italy Harvard Medical School;
Associate Physician
Sanjay Dixit, MD Cardiac Arrhythmia Service
Associate Professor Medicine Cardiovascular Research Center
Cardiovascular Division Center for Human Genetic Research
Hospital of the University of Pennsylvania, Massachusetts General Hospital
Director Cardiac Electrophysiology Cardiology-Medicine Boston, Massachusetts
Philadelphia Veterans Affairs Medical Center
Philadelphia, Pennsylvania Emilia Entcheva, PhD
Associate Professor Biomedical Engineering
Dobromir Dobrev, MD Stony Brook University
Professor of Pharmacology and Toxicology Stony Brook, New York
Director Institute of Pharmacology
Faculty of Medicine N.A. Mark Estes III, MD
University Duisburg-Essen Professor of Medicine
Hufelandstrasse, Essen, Germany Tufts University School of Medicine
Director, New England Cardiac Arrhythmia Center
Derek J. Dosdall, PhD Tufts Medical Center
Assistant Professor Internal Medicine Boston, Massachusetts
Cardiology
Comprehensive Arrhythmia Research and Management Center Rodolphe Fischmeister, PhD
University of Utah Director INSERM UMRS 769-LabEx LERMIT
Salt Lake City, Utah University Paris-Sud
Faculty of Pharmacy
John W. Dyer, PhD Chatenay-Malabry, France
Research Assistant Professor
School of Electrical and Computer Engineering John D. Fisher, MD
The University of Oklahoma Professor of Medicine
Norman, Oklahoma Albert Einstein College of Medicine
Biomedical Engineering Fellow Program Director CCEP
Heart Rhythm Institute Montefiore Medical Center
University of Oklahoma Health Sciences Center Bronx, New York
Oklahoma City, Oklahoma
Glenn I. Fishman, MD
Lars Eckardt, MD Director Leon H Charney Division of Cardiology
Professor of Internal Medicine/Cardiology New York University School of Medicine
Division of Electrophysiology New York, New York
Department of Cardiology and Angiology
University of Müenster David S. Frankel, MD
Münster, Germany Assistant Professor Medicine
Cardiovascular Division, Electrophysiology Section
Andrew G. Edwards, PhD Hospital of the University of Pennsylvania
Department of Bioengineering Philadelphia, Pennsylvania
University of California San Diego
La Jolla, California
Contributors vii
Michael R. Franz, MD, PhD, FHRS William J. Groh, MD, MPH

Professor of Medicine and Pharmacology Associate Professor Medicine
Georgetown University Medical Center; Indiana University School of Medicine
Director Indianapolis, Indiana
Myocardial and Arrhythmia Research
Staff Cardiology and Electrophysiologist Blair P. Grubb, MD
Veterans Affairs Medical Center Professor Medicine and Pediatrics
Washington, DC Cardiovascular Medicine
The University of Toledo
Paul A. Friedman, MD Toledo, Ohio
Director, Cardiac Implantable Device Lab
Division of Cardiovascular Medicine Michel Haissaguerre, MD
Mayo Clinic University of Bordeaux, France
Rochester, Minnesota Hospital Cardiologique du Haut-Leveque
Pessac, France
Victor F. Froelicher, MD
Professor Medicine Johan Hake, PhD
Department of Medicine Department of Bioengineering
Stanford University University of California San Diego
Director ECG and Exercise Testing Laboratories LaJolla, California
Veterans Affairs Palo Alto Health Care System Department of Computational Cardiac Modeling
Palo Alto, California Simula Research Laboratory
Medical Director Human Performance Lab Lysaker, Norway
Sportsmedicine
Stanford, California Henry R. Halperin, MD, MA, FAHA, FHRS
David J. Carver Professor of Medicine
Apoor S. Gami, MD, MSc Professor of Radiology and Biomedical Engineering
Cardiac Electrophysiology and Pacing Johns Hopkins University
Advocate Medical Group – Midwest Heart Specialists Baltimore, Maryland
Elmhurst, Illinois
Louise Harris, MBChB, FACC
Alfred L. George, Jr., MD Professor Medicine
Grant W. Liddle Professor University of Toronto
Chief Division of Genetic Medicine Staff Cardiologist/Electrophysiologist
Departments of Medicine and Pharmacology Medicine/Cardiology
Vanderbilt University Toronto Congenital Cardiac Centre for Adults
Nashville, Tennessee Peter Munk Cardiac Centre
University Health Network
Edward P. Gerstenfeld, MD Toronto, Ontario, Canada
Associate Professor Medicine/Cardiology
University of California Stéphane Hatem, MD, PhD
San Francisco, California Professor Medicine
ICAN Institute of Cardiometabolism & Nutrition
Michael R. Gold, MD, PhD UMRS-956 (INSERM/UPMC)
Michael E Assey Professor of Medicine Faculté de Médecine Pitié-Salpétrière
Chief of Cardiology MUSC Paris, France
Charleston, South Carolina
David L. Hayes, MD
Jeffrey J. Goldberger, MD, MBA Professor Medicine and Cardiovascular Diseases
Professor of Medicine Mayo Clinic
Northwestern University Feinberg School of Medicine Rochester, Minnesota
Chicago, Illinois
Meleze Hocini, MD
Eleonora Grandi, PhD University of Bordeaux, France
Assistant Project Scientist Hôpital Cardiologique du Haut-Lévêque
Pharmacology Pessac, France
University of California Davis
Davis, California Stefan H. Hohnloser, MD
Professor Cardiology
Richard A. Gray, PhD Department of Cardiology
Biomedical Engineer JW Goethe University
Center for Devices and Radiological Health Frankfurt, Germany
Office of Science and Engineering Laboratories
Food and Drug Administration David Richard Holmes, Jr., MD
Silver Spring, Maryland Professor Medicine and Cardiovascular Diseases
Mayo Clinic
Rochester, Minnesota
viii Contributors
Masahiko Hoshijima, MD, PhD José Jalife, MD

Associate Adjunct Professor Cyrus and Jane Farrehi Professor of Cardiovascular Research
Center for Research in Biological Systems and Department of Professor of Internal Medicine
Medicine Professor of Molecular and Integrative Physiology
University of California San Diego University of Michigan Medical School
La Jolla, California Co-Director, University of Michigan Center for Arrhythmia
Research
Yuxuan Hu, MSE Ann Arbor, Michigan
Graduate Student
Department of Biomedical Engineering Bong Sook Jhun, PhD
And Institute of Computational Medicine Research Associate
Johns Hopkins University Center for Translational Medicine
Baltimore, Maryland Department of Medicine
Jefferson Medical College
Thomas J. Hund, PhD Thomas Jefferson University
Assistant Professor Philadelphia, Pennsylvania
Departments of Biomedical Engineering and Internal Medicine
Dorothy M. Davis Heart and Lung Research Institute Roy M. John, MBBS, PhD
The Ohio State University Associate Director Cardiac Electrophysiology Laboratory
Columbus, Ohio Brigham and Women’s Hospital
Assistant Professor Medicine
Mathew D. Hutchinson, MD Harvard Medical School
Assistant Professor Medicine Boston, Massachusetts
Cardiovascular Division
University of Pennsylvania Monique Jongbloed, MD, PhD
Philadelphia, Pennsylvania Departments of Cardiology and Anatomy
Leiden University Medical Center
Hye Jin Hwang, MD, PhD Leiden, The Netherlands
Division of Cardiology Mark E. Josephson, MD, FACC, FACP
Yonsei Cardiovascular Research Institute Professor Medicine
Yonsei University College of Medicine Division of Cardiology
Seoul, Republic of Korea Beth Israel Deaconess Medical Center
Department of Biomedical Engineering Harvard Medical School
Washington University in St. Louis Director Harvard-Thorndike Electrophysiology Institute and
St. Louis, Missouri Arrhythmia Service
Chief Cardiovascular Division
Raymond E. Ideker Beth Israel Deaconess Medical Center
Professor Emeritus of Medicine Boston, Massachusetts
University of Alabama at Birmingham
Birmingham, Alabama Alan H. Kadish, MD
President
Leonard Ilkhanoff, MD, MS Touro College
Assistant Professor Medicine New York Medical College
Northwestern University Feinberg School of Medicine; New York, New York
Director, Acquired and Inherited Arrhythmia Program
Northwestern Memorial Hospital Jérôme Kalifa, MD, PhD
Department of Medicine, Division of Cardiology Assistant Professor Internal Medicine
Electrophysiology Section University of Michigan
Chicago, Illinois Ann Arbor, Michigan
Jodie Ingles, PhD Jonathan M. Kalman, MBBS, PhD

Research Officer Director of Cardiac Electrophysiology
Molecular Cardiology Research Program Cardiology Royal Melbourne Hospital,
Centenary Institute Professor of Medicine
Sydney, New South Wales, Australia Medicine University
University of Melbourne
Warren M. Jackman, MD Melbourne, Victoria, Australia
George Lynn Cross Research Professor
Heart Rhythm Institute Timothy J. Kamp, MD, PhD
University of Oklahoma Health Sciences Center Professor, Director, Stem Cell and Regenerative Medicine
Oklahoma City, Oklahoma Center
University of Wisconsin-Madison
Pierre Jais, MD Madison, Wisconsin
University of Bordeaux, France
Hôpital Cardiologique du Haut-Lévêque
Pessac, France
Contributors ix
Mohamed Hani Kanj, MD Peter Kohl, MD, PhD, FHRS

Associate Director Electrophysiology Labs Chair Cardiac Biophysics and Systems Biology
Robert and Suzanne Tomsich Department of Cardiovascular National Heart and Lung Institute
Medicine Imperial College
Cleveland Clinic London, GBR
Cleveland, Ohio Visiting Professor and Reader in Cardiac Physiology
Department of Computer Science
Beverly Karabin, RN, MSN, PhD University of Oxford
Associate Professor of Nursing Oxford, Great Britain
College of Nursing
University of Toledo Medical Center Aravindan Kolandaivelu, MD
Toledo, Ohio Division of Cardiology
Johns Hopkins University School of Medicine
Robert S. Kass, PhD Baltimore, Maryland
Professor and Chair Pharmacology
Columbia University Medical Center Andrew D. Krahn, MD, FHRS
New York, New York Chief of Cardiology
Sauder Family and Heart and Stroke Foundation Chair in
Demosthenes G. Katritsis, MD, PhD Cardiology
Director Cardiology Paul Brunes UBC Professor in Heart Rhythm Disorders
Athens Euroclinic University of British Columbia
GRC, Hon Consultant Cardiologist Vancouver, British Columbia, Canada
Cardiology
Guy’s and St Thomas’ Hospitals Andrew Krumerman, MD
London, Great Britain Associate Professor Clinical Medicine
Division of Cardiac Electrophysiology
Kuljeet Kaur, PhD Albert Einstein College of Medicine/MMC
Center for Arrhythmia Research Bronx, New York
University of Michigan
Ann Arbor, Michigan Saurabh Kumar, BSc(Med), MBBS
Department of Cardiology
Jong J. Kim, PhD and Department of Medicine
Postdoctoral Associate The Royal Melbourne Hospital
Biomedical Engineering University of Melbourne
Duke University Parkville, Victoria
Durham, North Carolina Australia
Paulus Kirchhof, MD, FESC, FHRS Karl-Heinz Kuck, MD, PhD

Professor Cardiovascular Medicine Chief
Centre for Cardiovascular Sciences Department of Medicine
University of Birmingham, Asklepios Klinik St. Georg
Consultant Cardiologist Hamburg, Germany
SWBH NHS Trust
Birmingham, United Kingdom; Edward G. Lakatta, MD
Associate Professor Medicine Director
Department of Cardiology and Angiology Laboratory of Cardiovascular Science
University Hospital Münster Intramural Research Program
Münster, Germany National Institute on Aging
National Institutes of Health
André G. Kléber, MD Baltimore, Maryland
Visiting Professor Pathology
Beth Israel Deaconess Medical Center Rakesh Latchamsetty, MD
Harvard Medical School Clinical Lecturer Electrophysiology
Boston, Massachusetts University of Michigan Hospital
Ann Arbor, Michigan
George J. Klein, MD, FRCP(C)
Professor of Medicine Dennis H. Lau, MBBS, PhD
Schulich School of Medicine Senior Lecturer and NHMRC Postdoctoral Fellow
Western University Centre for Heart Rhythm Disorders
London, Ontario, Canada University of Adelaide
Royal Adelaide Hospital and SAHMRI
Adelaide, Australia
x Contributors
Bruce B. Lerman, MD John C. Lopshire, MD, PhD

H. Altschul Master Professor of Medicine Assistant Professor Medicine and Cellular & Integrative
Chief Division of Cardiology Physiology
Director Cardiac Electrophysiology Laboratory Department of Medicine/Cardiology
Department of Medicine Krannert Institute of Cardiology
Division of Cardiology Indiana University School of Medicine,
Cornell University Medical Center Attending Physician Cardiac Electrophysiology
New York Presbyterian Hospital Department of Medicine/Cardiology
New York, New York Richard L Roudebush Veterans Affairs Medical Center
Jérôme Leroy, PhD
Assistant Professor of Physiology Steven A. Lubitz, MD, MPH
Faculty of Pharmacy IFR141-University Paris-Sud Instructor
Signalisation et Physiopathologie Cardiaque-Inserm UMR-S Harvard Medical School;
769 – LabEx Lermit Assistant in Medicine
Châtenay-Malabry, France Cardiac Arrhythmia Service
Cardiovascular Research Center
William R. Lewis, MD Massachusetts General Hospital
Chief Clinical Cardiology Boston, Massachusetts
Case Western Reserve University School of Medicine
Cleveland, Ohio Christopher Madias, MD
Clinical Cardiac Electrophysiology Service
Shien-Fong Lin, PhD Department of Medicine
Professor Medicine Rush University Medical Center
Division of Cardiology Chicago, Illinois
Department of Medine
Indianapolis, Indiana Aman Mahajan, MD, PhD
Professor of Anesthesiology
Mark S. Link, MD UCLA Cardiac Arrhythmia Center
Professor Internal Medicine UCLA Health System
Tufts University School of Medicine Los Angeles, California
Jonathan C. Makielski, MD
Christopher F. Liu, MD Fellow Medicine
Assistant Professor Medicine Professor Medicine
Division of Cardiology University of Wisconsin
Weill Cornell Medical College Madison, Wisconsin
Assistant Director Cardiac Electrophysiology Laboratory
New York-Presbyterian Hospital/Cornell Medical Center Marek Malik, PhD, MD, DSc
New York, New York Professor of Cardiac Electrophysiology
St. Paul’s Cardiac Electrophysiology
Deborah J. Lockwood, MBBCh St. George’s, University of London
Associate Professor Medicine London, England
Medicine/Cardiovascular Heart Rhythm Institute
OU Health Sciences Center Victor A. Maltsev, PhD
Oklahoma City, Oklahoma Staff Scientist
Laboratory of Cardiovascular Science
Peter Loh, MD, PhD Intramural Research Program
Department of Cardiology National Institute on Aging
Division of Heart and Lungs National Institutes of Health
University Medical Center Baltimore, Maryland
Utrecht, The Netherlands
Francis E. Marchlinski, MD, FACC, FHRS
Anatoli N. Lopatin, PhD Professor Medicine
Associate Professor Molecular and Integrative Physiology University of Pennsylvania School of Medicine
Molecular and Integrative Physiology Director Electrophysiology
University of Michigan Medical School Hospital of the University of Pennsylvania
Ann Arbor, Michigan Philadelphia, Pennsylvania
Ariane J. Marelli, MD, MPH

Associate Professor Medicine
McGill University
Director McGill Adult Unit for Congenital Heart Disease
Division of Cardiology
McGill University Health Centre
Montreal, Quebec, Canada
Contributors xi
Steven M. Markowitz, MD Fred Morady, MD

Professor Medicine Section of Electrophysiology
Cornell University Medical Center University of Michigan Health System
New York Presbyterian Hospital Ann Arbor, Michigan
New York, New York
Robert J. Myerburg, MD
Barry J. Maron, MD Professor Medicine and Physiology
Director Department of Medicine
Hypertrophic Cardiomyopathy Center University of Miami Miller School of Medicine
Minneapolis Heart Institute Foundation Miami, Florida
Minneapolis, Minnesota;
Professor of Medicine Hiroshi Nakagawa, MD, PhD
Tufts Medical Center and School of Medicine Professor of Medicine
Boston, Massachusetts; Director of Clinical Catheter Ablation Program
Professor of Medicine Director of Translational Electrophysiology
Mayo Clinic Heart Rhythm Institute
Rochester, Minnesota The University of Oklahoma HSC
Jeffrey R. Martens, PhD
Associate Professor Pharmacology Carlo Napolitano, MD, PhD
University of Michigan Vice-Director of Molecular Cardiology
Ann Arbor, Michigan IRCCS Fondazione Salvatore Maugeri
Pavia, Italy;
Steven O. Marx, BS, MD Research Associate Professor
Associate Professor Medicine Cardiovascular Genetics Program
Columbia University College of Physicians and Surgeons Leon H Charney Division of Cardiology
New York, New York New York University School of Medicine
New York, New York
Andrew D. McCulloch, PhD
Professor and Jacobs School Distinguished Scholar Sanjiv M. Narayan, MD, PhD
Director, Cardiac Biomedical Science and Engineering Center Professor Medicine
Departments of Bioengineering and Medicine University of California San Diego
University of California San Diego Director Electrophysiology Section
La Jolla, California Veterans Affairs Medical Center
San Diego, California
Andreas Metzner, MD
Staff Electrophysiologist Andrea Natale, MD
Department of Cardiologist Executive Medical Director
Asklepios Klinik St. Georg Texas Cardiac Arrhythmia Institute
Hamburg, Germany St David’s Medical Center
Austin, Texas;
Anuska P. Michailova, PhD Consulting Professor
Department of Bioengineering Division of Cardiology
University of California San Diego Stanford University
LaJolla, California Stanford, California;
Clinical Professor Medicine
John Michael Miller, MD Case Western Reserve University
Professor Medicine Cleveland, Ohio;
Indiana University School of Medicine, Director, Interventional Electrophysiology
Director Cardiac Electrophysiology Services Scripps Clinic
Indiana University Health San Diego, California
Indianapolis, Indianapolis
Stanley Nattel, MD
Michelle Lynne Milstein, PhD Professor and Paul David Chair in Cardiovascular
Postdoctoral Research Fellow Electrophysiology
University of Michigan Medical School Université de Montréal
Department of Internal Medicine Cardiologist and Director Electrophysiology Research Program
University of Michigan Center for Arrhythmia Research Research Center
Ann Arbor, Michigan Montreal Heart Institute
Montreal, Quebec, Canada
Peter Mohler, PhD
Professor and Director Saman Nazarian, MD
Dorothy M Davis Heart and Lung Research Institute Division of Cardiology
Ohio State University Medical Center Johns Hopkins University School of Medicine
Columbus, Ohio Baltimore, Maryland
xii Contributors
Jeanne M. Nerbonne, PhD Douglas L. Packer, MD

Alumni Endowed Professor of Molecular Biology and Professor of Medicine
Pharmacology Heart Rhythm Services
Developmental Biology Cardiovascular Diseases
Washington University Medical School Department of Medicine
St. Louis, Missouri Mayo Clinic
St. Mary’s Hospital
Fu Siong Ng, BSc (Hons), MBBS, MRCP, PhD Rochester, Minnesota
Clinical Lecturer in Cardiology
National Heart and Lung Institute Olle Pahlm, MD, PhD
Imperial College London Emeritus Professor
London, United Kingdom; Clinical Physiology
Clinical Research Fellow Lund University
Department of Biomedical Engineering Lund, Sweden
Washington University in St. Louis
St. Louis, Missouri Sandeep V. Pandit, PhD
Research Assistant Professor
Akihiko Nogami, MD Internal Medicine-Cardiology
Director, Division of Heart Rhythm Management University of Michigan
Yokohama Rosai Hospital Ann Arbor, Michigan
Yokohama, Japan
David S. Park, MD, PhD
Sami F. Noujaim, PhD Instructor
Assistant Professor Medicine Department of Medicine
Tufts University School of Medicine New York, University School of Medicine
Investigator Leon H Charney Division of Cardiology
Molecular Cardiology Research Center Cardiac Electrophysiology
Tufts Medical Center NYU Langone Medical Center
Boston, Massachusetts New York, New York
Brian Olshansky, MD Geoffrey S. Pitt, MD, PhD

Department of Internal Medicine Director Ion Channel Research Unit
University of Iowa Hospitals and Clinics Duke University
Iowa City, Iowa Durham, North Carolina
Hakan Oral, MD Sunny S. Po, MD, PhD

Professor Internal Medicine Professor Medicine
Frederick GL Huetwell Professor of Cardiovascular Medicine Heart Rhythm Institute
Director Cardiac Arrhythmia Service Department of Medicine
University of Michigan University of Oklahoma Health Sciences Center
Ann Arbor, Michigan Oklahoma City, Oklahoma
Jin O-Uchi, MD, PhD Silvia G. Priori, MD, PhD

Instructor Scientific Director
Center for Translational Medicine Director of Cardiology and Molecular Cardiology
Department of Medicine IRCCS Fondazione Salvatore Maugeri;
Jefferson Medical College Associate Professor
Thomas Jefferson University Department of Molecular Medicine
Philadelphia, Pennsylvania University of Pavia
Pavia, Italy;
Feifan Ouyang, MD Professor of Medicine
Director Electrophysiology Laboratories Leon Charney Division of Cardiology
Department of Cardiology New York University School of Medicine
Asklepios Klinik St Georg New York, New York
Hamburg, Germany
Wouter-Jan Rappel, PhD
Cevher Ozcan, MD Center for Theoretical Biological Physics
Assistant Professor of Medicine and Department of Physics
Department of Medicine University of California San Diego
University at Buffalo School of Medicine San Diego, California
and Biomedical Sciences
Buffalo, New York Vivek Y. Reddy, MD
Helmsley Trust Professor of Medicine
Mount Sinai School of Medicine
Director, Cardiac Arrhythmia Service
Mount Sinai Hospital
New York, New York
Contributors xiii
Jason O. Robertson, MD, MS Frank B. Sachse, PhD

Research Fellow Research Associate Professor of Bioengineering
Cardiothoracic Surgery Nora Eccles Harrison Cardiovascular Research and
Washington University Training Institute
Barnes-Jewish Hospital University of Utah
St. Louis, Missouri Salt Lake City, Utah
Richard B. Robinson, PhD Lindsey L. Saint, MD

Department of Pharmacology Research Fellow
Center for Molecular Therapeutics Cardiothoracic Surgery
Columbia University Medical Center Washington University
New York, New York Barnes-Jewish Hospital
St. Louis, Missouri
Dan M. Roden, MD
Professor Medicine & Pharmacology Javier Saiz, PhD
Assistant Vice-Chancellor for Personalized Medicine Professor Biomedical Engineering
Vanderbilt University School of Medicine Universitat Politecnica de Valencia
Nashville, Tennessee Valencia, Spain
Robert A. Rose, PhD José A. Sánchez-Chapula, MD, PhD

Department of Physiology and Biophysics Professor CUIB
Faculty of Medicine Universidad de Colima
Dalhousie University Colima, Mexico
Halifax, Nova Scotia, Canada
Prashanthan Sanders, MBBS, PhD
Michael R. Rosen, MD Professor of Cardiology and NHMRC Practitioner Fellow
Professor Pharmacology and Pediatrics Director, Centre for Heart Rhythm Disorders
Columbia University University of Adelaide, Royal Adelaide Hospital and SAHMRI
New York, New York Adelaide, Australia
Adjunct Professor Physiology and Biophysics
Stony Brook University Michael C. Sanguinetti, PhD
Stony Brook, New York Professor Physiology and Medicine
Nora Eccles Harrison Cardiovascular Research and
Raphael Rosso, MD Training Institute
Department of Cardiology University of Utah
Sourasky Tel-Aviv Medical Center Salt Lake City, Utah
Sackler School of Medicine
Tel-Aviv University Pasquale Santangeli, MD
Tel-Aviv, Israel Research Fellow
St. David’s Medical Center
Yoram Rudy, PhD Texas Cardiac Arrhythmia Institute
The Fred Saigh Distinguished Professor Austin, Texas
Biomedical Engineering, Cell Biology & Physiology
Medicine, Radiology, Pediatrics Georgia Sarquella-Brugada, MD
Director Cardiac Bioelectricity and Arrhythmia Center (CBAC) Pediatric Arrhythmia Unit
Washington University in St. Louis Hospital Sant Joan de Déu
St. Louis, Missouri University of Barcelona
Barcelona, Spain
Jeremy N. Ruskin, MD
Director Jonathan Satin, PhD
Cardiac Arrhythmia Service Professor Physiology
Massachusetts General Hospital; University of Kentucky College of Medicine
Associate Professor of Medicine Lexington, Kentucky
Harvard Medical School
Boston, Massachusetts Martin Jan Schalij, MD, PhD
Chief of Cardiology
Hani N. Sabbah, PhD Leiden University Medical Center
Professor Medicine Leiden, The Netherlands
Wayne State University
Director Cardiovascular Research Benjamin J. Scherlag, PhD
Henry Ford Health System Regents Professor, George Lynn Cross Research
Detroit, Michigan Professor, Helen Webster
Professor, Cardiac Arrhythmias
Professor, Medicine
Heart Rhythm Institute
College of Medicine
University of Oklahoma Health Sciences Center
xiv Contributors
Rainer Schimpf, MD Allan C. Skanes, MD, FRCPC

Department of Medicine-Cardiology Professor of Medicine
University Medical Center Director of EP Lab
Mannheim, Germany University Hospital
Western University
Georg Schmidt, MD London, Ontario, Canada
Professor of Cardiology
Technische Universität München Virend K. Somers, MD, PhD
Head Center of Nonlinear Dynamics in Cardiology Professor of Medicine
Munich, Germany Division of Cardiovascular Diseases
Mayo Clinic
Peter J. Schwartz, MD Rochester, Minnesota
Professor Cardiology, Molecular Medicine
University of Pavia Bruce S. Stambler, MD
Director Laboratory for Cardiovascular Research and Coronary Professor of Medicine
Care Unit Case Western Reserve University
Fondazione IRCCS; Cleveland, Ohio
Policlinico S. Matteo
Co-Director Cardiovascular Genetics Laboratory Adam B. Stein, MD
Pavia, Italy; Assistant Professor Medicine
Hatter Institute for Cardiovascular Research in Africa University of Michigan
Department of Medicine Ann Arbor, Michigan
University of Cape Town
Cape Town, South Africa; Lynne Warner Stevenson, MD
Chair of Sudden Death Director Cardiomyopathy/Heart Failure Program
Department of Family and Community Medicine Brigham and Women’s Hospital;
College of Medicine, King Saud University Professor of Medicine
Riyadh, Saudi Arabia Harvard Medical School
Christopher Semsarian, MBBS, PhD, FRACP
Professor Medicine William G. Stevenson, MD
University of Sydney Director Clinical Electrophysiology
Program Head Molecular Cardiology Cardiovascular Division
Centenary Institute Brigham and Women’s Hospital;
Sydney, New South Wales, Australia Professor Medicine
Harvard Medical School
Ashok J. Shah, MBBS Boston, Massachusetts
University of Bordeaux II
Bordeaux, France Jian Sun, PhD
Department of Rhythmologie Cardiology
Hôpital Cardiologique du Haut-Lévêque Shanghai Jiao Tong University School of Medicine
Pessac, France Attending Physician Cardiology
Xin Hua Hospital
Robin Shaw, MD, PhD Shanghai, China
Associate Professor
Medicine Richard Sutton, DSc
University of California Los Angeles Emeritus Professor Cardiology
Associate Professor/Staff Physician II Imperial College
Cedars-Sinai Heart Institute London, Great Britain
Cedars-Sinai Medical Center
Los Angeles, California Michael O. Sweeney, MD
Cardiac Pacing and Heart Failure Device Therapy
Shey Shing Sheu, PhD Cardiovascular Division
Professor and Associate Director Brigham and Women’s Hospital
Center for Translational Medicine Boston, Massachusetts
Department of Medicine
Jefferson Medical College Charles Swerdlow, MD
Thomas Jefferson University Cedars Sinai Heart Institute
Philadelphia, Pennsylvania Clinical Professor Medicine
Cardiology UCLA
Kalyanam Shivkumar, MD, PhD, FHRS Los Angeles, California
Professor of Medicine & Radiology
UCLA Cardiac Arrhythmia Center Juan Tamargo, MD, PhD, FESC
UCLA Health System Professor Pharmacology
Los Angeles, California School of Medicine, Universidad Complutense
Madrid, Spain
Contributors xv
Harikrishna Tandri, MD Mintu P. Turakhia, MD, MAS

Division of Cardiology Assistant Professor of Medicine
Johns Hopkins University School of Medicine Department of Medicine (Cardiovascular Medicine)
Baltimore, Maryland Stanford University
Stanford, California;
Rabi Tawil, MD Director of Cardiac Electrophysiology
Professor Neurology Veterans Affairs Palo Alto Health Care System
University of Rochester Medical Center Palo Alto, California
Rochester, New York
Ravi Vaidyanathan, ME, PhD
Usha Tedrow, MD, MSc Research Associate
Director Clinical Cardiac Electrophysiology Program Medicine-Cardiology
Department of Medicine, Cardiovascular Division University of Wisconsin
Brigham and Women’s Hospital Madison, Wisconsin
Harvard Medical School Héctor H. Valdivia, MD, PhD
Boston, Massachusetts Frank N Wilson Professor of Cardiovascular Medicine
Internal Medicine
Cecile Terrenoire, MD University of Michigan
Department of Pharmacology Ann Arbor, Michigan
College of Physicians & Surgeons of Columbia University
New York, New York Virginijus Valiunas, PhD
Research Associate Professor
Catalina Tobón, PhD Physiology and Biophysics
Professor GI2B Stony Brook University
Instituto Tecnológico Metropolitano Stony Brook, New York
Medellín, Colombia;
Research Marcel A.G. van der Heyden, PhD
Universitat Politècnica de València Medical Physiology
Valencia, Spain University Medical Center Utrecht
Utrecht, The Netherlands
Jeffrey A. Towbin, MD
Executive Co-Director Christian van der Werf, MD
The Heart Institute; Department of Cardiology
Professor and Chief Heart Failure Research Center
Pediatric Cardiology Academic Medical Center
Kindervelt-Samuel Kaplan Chair in Pediatric Cardiology and Amsterdam, The Netherlands
Cardiac Research
Cincinnati Children’s Hospital Medical Center George F. Van Hare, MD
University of Cincinnati College of Medicine Professor Pediatrics
Cincinnati, Ohio Washington University School of Medicine
Director Pediatric Cardiology
Natalia A. Trayanova, PhD St. Louis Children’s Hospital
Professor St. Louis, Missouri
Department of Biomedical Engineering
Johns Hopkins University Marmar Vaseghi, MD, MS
Baltimore, Maryland Assistant Professor of Medicine
UCLA Cardiac Arrhythmia Center
Martin Tristani-Firouzi, MD UCLA Health System
Professor Pediatric Cardiology Los Angeles, California
University of Utah School of Medicine
Investigator Christian Veltmann, MD
Nora Eccles Harrison Cardiovascular Research and Department of Medicine-Cardiology
Training Institute University Medical Center
Salt Lake City, Utah Mannheim, Germany
Richard G. Trohman, MD Victoria L. Vetter, MD, MPH, MSHP

Clinical Cardiac Electrophysiology Service Director, Youth Heart Watch
Department of Medicine Professor of Pediatrics
Co-Director Section of Cardiology The Children’s Hospital of Philadelphia
Rush University Medical Center Perelman School of Medicine at the University of Pennsylvania
Chicago, Illinois Philadelphia, Pennsylvania
Zian H. Tseng, MD, MAS Sami Viskin, MD

Associate Professor of Medicine in Residence Director Cardiac Hospitalization
Cardiac Electrophysiology Section, Cardiology Division Tel Aviv Medical Center
University of California San Francisco Tel Aviv, Israel
San Francisco, California
xvi Contributors
Niels Voigt, MD Y. Joseph Woo, MD

Institute of Pharmacology Professor of Surgery
Faculty of Medicine Director, Cardiac Transplantation and Mechanical Circulatory
University Duisburg-Essen Assist Program Director
Hufelandstrasse, Essen, Germany Minimally Invasive and Robotic Cardiac Surgery Program
Division of Cardiovascular Surgery
Marc A. Vos, PhD Department of Surgery
Professor and Chair of Medical Physiology University of Pennsylvania
Manager Research and Education Division Heart and Lungs Philadelphia, Pennsylvania
University Medical Center Utrecht
Utrecht University Masatoshi Yamazaki, MD, PhD
Utrecht, The Netherlands Assistant Professor
Cardiovascular Research
Galen S. Wagner, MD Research Institute of Environmental Medicine
Associate Professor of Medicine, Cardiology Nagoya University
Duke University Medical Center Nagoya, Japan
Durham, North Carolina
Felix Yang, MD
Paul J. Wang, MD Associate Director Cardiac Electrophysiology
Professor Medicine Maimonides Medical Center
Director Arrhythmia Service Brooklyn, New York
Stanford University
Stanford, California Yael Yaniv, PhD
Research Fellow
Rukshen Weerasooriya, BMedSc (Hons), MBBS Laboratory of Cardiovascular Science
Clinical Professor National Institute on Aging
Medicine University of Western Australia National Institutes of Health
Crawley, Western Australia; Baltimore, Maryland
Cardiac Electrophysiologist
Cardiology Sing-Chien Yap, MD, PhD
Hollywood Private Hospital Cardiology
Nedlands, Western Australia; Erasmus Medical Center
AF Ablation Clinic Rotterdam, The Netherlands
Nedlands, Western Australia
Raymond Yee, MD
Arthur A. M. Wilde, MD, PhD Professor of Medicine
Professor Cardiology University of Western Ontario
Department of Cardiology Director
Heart Failure Research Center Arrhythmia Service
Academic Medical Center London Health Sciences Centre
University of Amsterdam, London, Ontario, Canada
Amsterdam, The Netherlands
Manuel Zargoso, PhD
Bruce L. Wilkoff, MD Departmento de Fisiología
Professor Internal Medicine Universidad de Valencia
Cardiovascular Medicine Valencia, Spain
Cleveland Clinic Lerner College of Medicine
Case Western Reserve University Katja Zeppenfeld, MD, PhD
Director Cardiac Pacing and Tachyarrhythmia Devices Departments of Cardiology and Anatomy
Associate Chair Section of Cardiac Pacing and Leiden University Medical Center
Electrophysiology Leiden, The Netherlands
Medical Information Officer Heart and Vascular Institute
Cardiovascular Medicine Douglas P. Zipes, MD
Cleveland Clinic Distinguished Professor
Cleveland, Ohio Emeritus Professor of Medicine, Pharmacology, and Toxicology
Emeritus Director, Division of Cardiology and the Krannert
Erik Wissner, MD Institute of Cardiology
Director, Magnetic Navigation Laboratory Indiana University School of Medicine
Department of Cardiology Indianapolis, Indiana
Asklepios Klinik St. Georg Editor-in-Chief, HeartRhythm
Hamburg, Germany Editor-in-Chief, PracticeUpdate/Cardiology
Preface
We are pleased to publish the Sixth Edition of Cardiac Electro- sections on the cell biology of cardiac impulse initiation and
physiology: From Cell to Bedside, four years after publication of the propagation, models of cardiac excitation and arrhythmias, neural
Fifth Edition. We shortened publication time by a year because control of cardiac electrical activity, arrhythmia mechanisms,
of the breakneck speed of new observations, as well as the molecular genetics, and pharmacogenomics; and a new section
increased importance of this specialty for patients with ventricu- on pharmacologic, genetics, and cell therapy approaches to ion
lar arrhythmias or atrial fibrillation. Cell to Bedside continues to channel dysfunction. All chapters have been revised, and 25 chap-
serve as an important “go-to” resource with the latest informa- ters are new.
tion on cardiac electrophysiology for basic scientists and clini- The second half is divided into sections on diagnostic evalu-
cians. As before, we have designed the chapters to appeal to a ation, mechanisms, features and management of supraventricular
broad audience at all stages of learning, from post docs to gray and ventricular tachycardias, syncope and AV block; arrhythmias
beards, literally spanning cell to bedside. Chapters continue in special populations, and pharmacologic, electrical, and surgical
to be written by the most authoritative leaders in each field. therapies, as well as new therapeutic approaches, including vagal
We thank these busy professionals for being a part of this and spinal cord stimulation. Advances in each area called for
endeavor. thorough revision and updating of each chapter, and the addition
So many advances have been made in the past four years that of 28 totally new chapters to the clinical half of the book.
we expanded the number of chapters from 110 to 134. The full As before, and as ALWAYS, we thank our spouses, Joan Zipes
color display provides easier reading, and the electronic version and Paloma Jalife, for providing us with love and support, not
allows portability, while the Web site serves as a place where the only for this endeavor, but for all of the important things in our
reader can view overflow figures, tables, and videos. As Braun- lives that allow us to be who we are!
wald’s Heart Disease: A Textbook of Cardiovascular Medicine has Janice Gaillard and Dolores Meloni at Elsevier put out many
spawned a related family of books called Companions, so too has fires and facilitated publication of this book, and we thank them.
Cardiac Electrophysiology: From Cell to Bedside. The first offspring As we said in the Preface of the Fifth Edition, it is you, the
is Electrocardiography of Arrhythmias: A Comprehensive Review, by reader, who determines the success or failure of this book, and
Das and Zipes. We hope to add more to this family in the coming we thank you for continuing to make Cardiac Electrophysiology:
years. From Cell to Bedside, a part of your learning. We hope this edition
The first half of this book is divided into nine sections includ- lives up to your standards and requirements. Let us know.
ing a section on structural, molecular, and biophysical bases of
cardiac ion channel function; a new section on intramolecular Douglas P. Zipes, MD
interactions and cardiomyocyte electrical function; updated José Jalife, MD
xvii
Video Contents
Chapter 1 Chapter 87
Voltage-Gated Sodium Channels and Electrical Right Ventricular Cardiomyopathy
Excitability of the Heart 5. Aneurysm Formation as a Dyskinetic Region on Cine MR
1. Nav Gating With Inactivation Autoloop Imaging
6. LV Epicardial Fat Infiltration with Reserved Global Function
William A. Catterall, MD
Harikrishna Tandri, MD, and Hugh Calkins, MD
Chapter 70
Chapter 129
Noninvasive Electrocardiographic Imaging (ECGI)
Ventricular Arrhythmias
of Human Ventricular Arrhythmias and
7. Cryo-Thermy Application on Epicardial Surface
Electrophysiologic Substrate 8. Transaortic Visualization of Lateral LV Endocardial LV
2. Patient (LV1): Epicardial Focal VT Surface
3. Patient (LV1): Scar-related Reentry VT 9. Cardiac Exposure Through Partial Sternotomy
4. ECGI Activation Movie: Ventricular Tachycardia (VT) Ini- Sanjay Dixit, MD, and Y. Joseph Woo, MD
tiation, Continuation, and Termination by Antitachycardial
Pacing (ATP)
Yoram Rudy, MD
xxvi
Structural and Molecular Bases PART I
of Ion Channel Function
Voltage-Gated Sodium Channels and

Electrical Excitability of the Heart 1
William A. Catterall
conduction in artificial lipid membranes and expression of the

CHAPTER OUTLINE α-subunit alone is sufficient for physiologic function in recipient
Subunit Structure of Sodium Channels 1 nonexcitable cells, indicating that this subunit has all the neces-
sary structural elements for voltage-dependent gating and ion
Three Dimensional Structure of Sodium Channels 1 conduction.1-3 The primary sequence predicts that the sodium
Sodium Channel Structure and Function 3 channel α-subunit folds into four internally repeated domains
(I-IV), each of which contains six α-helical transmembrane seg-
Sodium Channel Genes 8 ments (S1-S61,3-6; Figure 1-2). In each domain, the S1-S4 seg-
ments serve as the voltage-sensing module, and the S5 and S6
segments and the reentrant P loop between them serve as the
Voltage-gated sodium channels initiate action potentials in pore-forming module. Extracellular loops connect the S5 and S6
cardiac myocytes and other excitable cells, and they are respon- transmembrane segments to the P loop in each domain, whereas
sible for propagation of action potentials through the atria, con- the other extracellular loops are small. Large intracellular loops
duction system, and ventricles of the heart. As shown in Figure link the four homologous domains, and the large N-terminal and
1-1, action potentials in atrial and ventricular muscle fibers rise C-terminal domains also contribute substantially to the mass of
rapidly from a resting potential near −80 mV and reach their peak the internal face of sodium channels. This view of sodium channel
within 1 ms. During this brief interval, cardiac sodium channels architecture, originally derived from hydrophobicity analysis of
respond to the change in pacemaker potential as it reaches the amino acid sequence,4 has been largely confirmed by bio-
threshold and open to allow rapid Na+ entry. Sodium channels chemical, electrophysiologic, and structural experiments.6
begin to inactivate as soon as they open and inactivate to 98% or The auxiliary β subunits were identified in the initial purifica-
99% completion within a few milliseconds. The plateau phase of tion studies of sodium channels.1 These subunits have a single
the cardiac action potential is generated by opening of voltage- transmembrane segment, a large N-terminal extracellular domain
gated calcium channels (see Chapter 2), and the cell is finally that is homologous in structure to a variable chain (V-type)
repolarized by slower opening of voltage-gated potassium chan- immunoglobulin-like fold, and a short C-terminal intracellular
nels (see Chapter 3). The rate of conduction of the action poten- segment (see Figure 1-2).7-10 The β-subunits interact with
tial through the cardiac tissue depends directly on the rate of rise α-subunits through their extracellular immunoglobulin (Ig)-fold
of the cardiac action potential and therefore on the density of domains, modulate α-subunit function, and enhance their cell
sodium channels and their rate of activation. surface expression.11 Like other proteins with an extracellular
Sodium channels are complexes of a large, pore-forming Ig-fold, they also serve as cell adhesion molecules by interacting
α-subunit and smaller β-subunits. Much is now known about the with extracellular matrix proteins, cell adhesion molecules, and
mechanisms of activation, inactivation, and ion conduction by the cytoskeletal linker proteins.12-17 These interactions are thought to
sodium channel protein, as summarized in this chapter. Multiple localize and stabilize sodium channels in specific subcellular com-
genes encode sodium channel subunits, and the distinct sodium partments and to bring crucial signaling molecules to the sodium
channel subtypes have subtle differences in functional properties channel to regulate it. Deletion of the genes encoding β-subunits
and differential distribution in subcellular compartments of causes alterations in sodium channel function; reduced action
cardiac myocytes. These differences in function and localization potential conduction and abnormal development of myelin folds
may contribute to specialized functional roles of sodium channels in axons; hyperexcitability and epilepsy in the brain; and arrhyth-
in cardiac physiology and pharmacology. mias in the heart.18-20
Subunit Structure of Sodium Channels Three-Dimensional Structure of

Sodium Channels
Sodium channel proteins purified from excitable cells are com-
plexes composed of an approximately 260-kD α-subunit in asso- Sodium channel architecture has been revealed in three dimen-
ciation with one or two auxiliary β-subunits of approximately 33 sions by determination of the crystal structure of the bacterial
to 39 kD.1 Purified sodium channel complexes of α- and sodium channel NavAb at high resolution (2.7 Å (0.27 nm);
β-subunits are sufficient for voltage-dependent gating and ion Figure 1-3). This ancient sodium channel has a simple
1
2 STRUCTURAL AND MOLECULAR BASES OF ION CHANNEL FUNCTION
LEFT VENTRICLE LEFT ATRIUM
40 40
20 20
0 0
–20
V(mV)
–20
–40 –40
–60 –60
–80 –80
100 ms 100 ms
Figure 1-1. Cardiac action potential in sheep heart. Cardiac myocytes in left ventricle or left atrium were impaled with a microelectrode, and the cardiac action potential
was recorded.
(Courtesy J. Jalife.)
Figure 1-2. Transmembrane organization of sodium channel subunits. The primary structures of the subunits of the voltage-gated ion channels are illustrated as trans-
membrane folding diagrams. Cylinders represent probable alpha helical segments: blue, S1-S3; green, S4; yellow, S5; red, S6; shaded orange area, outer pore loop; purple,
intracellular S4-S5 helix. Bold lines represent the polypeptide chains of each subunit with length approximately proportional to the number of amino acid residues in the
brain sodium channel subtypes. The extracellular domains of the β1- and β2-subunits are shown as immunoglobulin-like folds Ψ, sites of probable N-linked glycosylation.
P, sites of demonstrated protein phosphorylation by PKA (circles) and PKC (diamonds); white circles, the outer (EEDD) and inner (DEKA) rings of amino residues that form
the ion selectivity filter and the tetrodotoxin binding site; ++, S4 voltage sensors; h in shaded circle, inactivation particle in the inactivation gate loop; open shaded circles,
sites implicated in forming the inactivation gate receptor. The structure of the extracellular domain of the β-subunits is illustrated as an immunoglobulin-like fold based
on amino acid sequence homology to the myelin P0 protein.8,22 Sites of binding of α- and β-scorpion toxins and a site of interaction between α- and β1-subunits are also
shown.
(Adapted from Catterall WA: From ionic currents to molecular mechanisms: the structure and function of voltage-gated sodium channels. Neuron 26:13–25, 2000.)
structure—four identical subunits that are each similar to one surrounded by four pore-forming modules composed of S5 and
homologous domain of a mammalian sodium channel without S6 segments and the intervening pore loop (see Figure 1-3, A).
the large intracellular and extracellular loops of the mammalian Four voltage-sensing modules composed of S1-S4 segments are
protein.21 This structure has revealed a wealth of new informa- symmetrically positioned around the outer rim of the pore
tion about the structural basis for sodium selectivity and conduc- module (see Figure 1-3, A). The transmembrane architecture of
tance, the mechanism for block of the channel by therapeutically NavAb shows that the adjacent subunits have swapped their func-
important drugs, and the mechanism of voltage-dependent tional domains such that each voltage-sensing module is most
gating. As viewed from the top, NavAb has a central pore closely positional around with the pore-forming module of its
Voltage-Gated Sodium Channels and Electrical Excitability of the Heart 3
S4
S3
P P2
S2
A
S1 S5 S6
S4–S5 linker
C S1N
Figure 1-3. Three-dimensional structure of sodium channels. A, Top view of NavAb channels colored according to crystallographic temperature factors of the main-chain
(blue < 50 Å2 to red > 150 Å2). B, Side view of NavAb. C, Structural elements in NavAb. The structural components of one subunit are highlighted. 1-6, Transmembrane
segments S1-S6.
(Adapted from Payandeh J, Scheuer T, Zheng N, et al: The crystal structure of a voltage-gated sodium channel. Nature 475:353–358, 2011.)
neighbor (see Figure 1-3, B). It is likely that this domain-swapped Outer Pore and Selectivity Filter
arrangement enforces concerted gating of the four subunits or
domains of sodium channels. Voltage clamp studies showed that sodium channels are highly
Comparison of the primary structures of the auxiliary selective for sodium versus potassium and other monovalent
β-subunits to those of other proteins revealed a close structural cations.26,27 Analysis of ion selectivity and block by tetrodotoxin
relationship to the family of proteins that contain Ig-like folds, and saxitoxin led to a model of tetrodotoxin and saxitoxin as plugs
which include many cell-adhesion molecules.8 The extracellular of the selectivity filter in the outer pore of sodium channels.26
domains of these type-I single-membrane-spanning proteins are Mutational analysis identified a key glutamate residue in the
predicted to fold in a similar manner as myelin protein P0, membrane-reentrant loop in domain I as a crucial residue for
whose Ig-like fold is known to be formed by a sandwich of two tetrodotoxin and saxitoxin binding.28 Additional studies revealed
β-sheets held together by hydrophobic interactions (see Figure a pair of important amino acid residues, mostly negatively
1-222). Myelin P0 protein is a cell adhesion molecule involved in charged, in analogous positions in all four domains (see Figure
tight wrapping of myelin sheets, and many related cell adhesion 1-2, B, small white circles).29-31 Mutation of a set of four residues
molecules with extracellular Ig-folds and a single membrane- in analogous positions in each domain (aspartate in domain I,
spanning segment are involved in cell-cell interactions among glutamate in domain II, lysine in domain III, and alanine in
neurons and glia.22 As expected from their structure, NaVβ subdomain IV, DEKA) to glutamates confers calcium selectivity,32
units interact with extracellular matrix molecules, other cell indicating that the side chains of these amino acid residues are
adhesion molecules, and intracellular cytoskeletal proteins and likely to interact with sodium ions as they are conducted through
signaling proteins.12-16,23 the ion selectivity filter of the pore. These results showed that
the narrow selectivity filter in the outer pore is formed by the
reentrant P loops between transmembrane segments S5 and S6
of each domain (see Figure 1-2). Mutations in this postulated ring
Sodium Channel Structure and Function of four amino acid residues have strong effects on selectivity for
organic and inorganic monovalent cations, in agreement with the
Hodgkin and Huxley24 defined the three key functions of sodium idea that they form the selectivity filter, and structure-function
channels: (1) voltage-dependent activation, (2) fast inactivation, studies suggest specific structural interactions and functional
and (3) selective ion conductance. Building on this foundation, roles for the P loops from the four domains.33-36
detailed biophysical studies revealed the ion selectivity of the In the bacterial sodium channel NavAb, the overall pore
channel pore, detected the movement of the voltage sensors as architecture includes a large external vestibule, a narrow ion
gating current, and developed mechanistic models for these selectivity filter containing the amino acid residues shown to
essential channel functions.25,26 Recent structure-function studies determine ion selectivity in vertebrate sodium and calcium chan-
using molecular, biochemical, structural, and electrophysiologic nels, a large central cavity that is lined by the S6 segments and
techniques have provided clear understanding of the molecular is filled with water, and an intracellular activation gate formed at
and structural basis for these sodium channel functions. the crossing of the S6 segments at the intracellular surface of the
membrane (Figure 1-4, A). The activation gate is tightly closed opposite to that of K+: negatively charged residues interact with
in the NavAb structure (see Figure 1-4, B), and there is no space Na+ to remove most (but not all) of its waters of hydration, and
for ions or water to move through it. This general architecture Na+ is conducted as a hydrated ion interacting with the pore
resembles voltage-gated potassium channels (see Chapter 3). through its inner shell of bound waters. Theoretical consider-
Although the overall pore architecture of sodium and potassium ations of sodium selectivity and conductance predicted an outer
channels is similar, the structures of their ion selectivity filters high–field-strength site that would only partially dehydrate the
and their mechanisms of ion selectivity and conductance are permeating ion and two inner sites that would conduct and rehy-
completely different. Potassium channels select K+ by direct drate the permeant Na+ ion because of the high energy of hydra-
interaction with a series of four ion coordination sites formed by tion of Na+ (see Figure 1-4, C, D).37
the backbone carbonyls of the amino acid residues that comprise
the ion selectivity filter (Chapter 3). No charged amino acid resi-
dues are involved, and no water molecules intervene between K+ Voltage-Dependent Activation
and its interacting backbone carbonyls in the ion selectivity filter
of potassium channels. In contrast, the NavAb ion selectivity The voltage dependence of the activation of sodium channels
filter has a high field strength site at its extracellular end (see derives from outward movement of approximately 12 gating
Figure 1-4, C; Glu177 = ε177), which is formed by amino acid charges as a consequence of depolarization of the membrane and
residues that are highly conserved and are key determinants of reduction of the membrane electric field.25,38 The S4 segments of
ion selectivity in vertebrate sodium and calcium channels (see each homologous domain serve as the primary voltage sensors
Figure 1-2). Considering its dimensions of approximately 4.6 Å2, for activation.4,5 They contain repeated motifs of a positively
Na+ with two planar waters of hydration could fit in this high– charged amino acid residue followed by two hydrophobic resi-
field-strength site. This outer site is followed by two ion coordi- dues creating a transmembrane spiral of positive charges. Upon
nation sites formed by backbone carbonyls (see Figure 1-4, D). depolarization, outward movement and rotation of S4 is thought
These two carbonyl sites are perfectly designed to bind Na+ with to initiate a conformational change that opens the sodium channel
four planar waters of hydration, but would be much too large to pore.4,5 This “sliding helix” or “helical screw” model is supported
bind Na+ directly. In fact, the NavAb selectivity filter is large by strong evidence. For example, neutralization of the positively
enough to fit the entire K+ channel ion selectivity filter inside it. charged residues in S4 reduce the voltage-dependence of gating.39
Thus, the chemistry of Na+ selectivity and conductance is The outward and rotational gating movement of S4 segment has
Extracellular M221
funnel
Selectivity
filter
Central
cavity
Side S6
view
Activation
A gate B
Top
view Side IonEX
view
5.4 Å
SiteHFS
4.5 Å
SiteCEN
L176*
3.5 Å
T175* SiteIN
C D
Figure 1-4. NavAb pore and selectivity filter. A, Architecture of the NavAb pore. Purple, Glu177 side-chains; gray, pore volume. B, The closed activation gate at the intracel-
lular end of the pore illustrating the close interaction of Met221 residues in closing the pore. C, Top view of the ion selectivity filter. Symmetry-related molecules are colored
white and yellow; P-helix residues are colored green. Hydrogen bonds between Thr175 and Trp179 are indicated by gray dashes. Electron densities from Fo-Fc omit maps
are contoured at 4.0 σ (blue and gray), and subtle differences can be appreciated (small arrows). D, Side view of the selectivity filter. Glu177 (purple) interactions with Gln172,
Ser178, and the backbone of Ser180 are shown in the far subunit. Fo-Fc omit map, 4.75 σ (blue); putative cations or water molecules (red spheres, IonEX). Electron-density
around Leu176 (gray; Fo-Fc omit map at 1.75 σ) and a putative water molecule are shown (grey sphere). Na+-coordination sites: SiteHFS, SiteCEN and SiteIN.
been detected directly by reaction of substituted cysteine residues change in electrostatic force, the S4 segment moves outward
1
in S4 segments with extracellular sulfhydryl reagents following with each positive gating charge interacting with charged amino
channel activation and by analysis of the movement of fluorescent acid side chains in turn to ease their movement through the
probes incorporated into these substituted cysteine residues.40,41 voltage sensor module. When the R3 and R4 gating charges
Other support for this mechanism derives from a wide range of pass the hydrophobic constriction site in the center of the
structure and function studies.42 voltage sensor module, the force on the S4-S5 linker is sufficient
In the structure of the bacterial sodium channel NavAb, the to pull on the pore-forming module and twist and bend the S6
S4 segment is in a transmembrane position in its activated state segment, resulting in opening of the pore at its intracellular end.
and its positive charges are thought to be neutralized by negative During activation of the voltage sensors of sodium channels, the
charges in the nearby S1 and S2 segments (Figure 1-5, A).21 At S4-S5 intracellular linkers in each domain (see Figure 1-2, B,
the resting membrane potential, the force of the electric field, purple) exert a force on the adjacent S6 segments, and the pore
which is negative inside the cell, would pull the positive charges opens by bending the S6 segment (see Figure 1-2, B, red). In
inward. Depolarization would abolish this force and allow an bacterial voltage-gated potassium channels and sodium channels,
outward movement of the S4 helix and its gating charges, cata- bending of the S6 segment occurs at a critical hinge glycine
lyzed by exchange of ion pair partners (see Figure 1-5, B; Movie residue about one-third down the S6 segment,43-45 and this
1-1). After conformational changes have occurred in all four bending motion allows the opening of the inner mouth of the
domains, the transmembrane pore can open and conduct ions pore (Videos 1 and 2; Figure 1-5) and ion rapid movement across
(see Movie 1-1). This structural model shows that the S4 segment the membrane.
and its gating charges move through a narrow gating pore that
focuses the transmembrane electric field to a distance of approxi-
mately 5 Å normal to the membrane and allows the gating Fast Inactivation
charges to move from an intracellular aqueous vestibule to an
extracellular aqueous vestibule with a short transit through the Fast inactivation of the sodium channel is a critical process that
channel protein (see Video 1-1). occurs within milliseconds of channel opening. The generally
accepted model of this process involves a conserved inactivation
gate formed by the intracellular loop connecting domains III and
Pore Opening IV (see Figure 1-2), which serves as a hinged lid that binds to the
intracellular end of the pore and blocks it (Figure 1-6). Intracel-
From these structural models, one can visualize the steps in the lular perfusion of proteases prevents fast inactivation.25 Site-
gating of a voltage-gated ion channel. In the closed state, the directed antipeptide antibodies against the short, highly conserved
negative internal membrane potential of −70 to −90 mV pulls the intracellular loop connecting domains III and IV of the sodium
S4 gating charges inward by electrostatic force. The inward posi- channel α-subunit (see Figure 1-2), but not antibodies directed
tion of the S4 segment exerts a force on the S4-S5 linker, twists to other intracellular domains, were found to prevent fast sodium
and straightens the S6 segment, and closes the pore at its inner channel inactivation.46,47 Moreover, the accessibility of this site
mouth. When the cell is depolarized, the electrostatic force for antibody binding was reduced when the membrane was depo-
pulling the S4 segment inward is relieved. In response to the larized to induce inactivation, suggesting that the loop
S2-S3 loop
membrane
anchor
S1N
A B
Figure 1-5. The voltage-sensing domain (VSD). A, Side view of the VSD illustrating the extracellular negative charge-cluster (red; ENC), the intracellular negative charge-
cluster (red; INC), hydrophobic constriction site (green; HCS), residues of the S1N helix (cyan), and phenylalanine of the S2-S3 loop (purple). S4 segment and gating charges
(R1-R4) are in yellow. B, Transmembrane view of the lowest-energy Rosetta models of the VSD of NaChBac in Resting State 1 (left) and Activated State 3 (right). Side chains
of the gating-charge-carrying arginines in S4 and key residues in S1, S2, and S3 segments are shown in stick representation and labeled. Gray, blue, and red atoms are C,
N, and O, respectively. The HCS is highlighted by orange bars. The lowest-energy models of Resting State 1 predict that R1 forms hydrogen bonds with the backbone
carbonyl of I96 (in S3) at the extracellular edge of the HCS. On the intracellular side of the HCS, R3 makes ionic interactions with the amino acid residues of the intracellular
negatively charged cluster, including E70 (in S2) and D93 (in S3), and R4 forms an ion pair with D93 (in S3). The lowest-energy models for Activated State 3 predict that R1
forms an ion pair with E43 (in S1), R2 forms an ion pair with E43 (in S1), R3 forms hydrogen bond with Y156 (in S5) and makes ionic interactions with D60 (in S2) and E43
(in S1), and R4 forms an ion pair with D60 (in S2). See Videos 1-1 and 1-2.
(Adapted from Payandeh J, Scheuer T, Zheng N, et al: The crystal structure of a voltage-gated sodium channel. Nature 475:353–358, 2011; and Yarov-Yarovoy V, DeCaen PG,
Westenbroek RE, et al: Structural basis for gating charge movement in the voltage sensor of a sodium channel. Proc Natl Acad Sci U S A 109:E93–102, 2012.)
I IV I IV peptide and analysis by multidimensional NMR methods.53

These experiments reveal a rigid α-helix flanked on its N-
II III II III terminal side by two turns, the second of which contains the IFM
motif (see Figure 1-6, B). The fold of the inactivation gate peptide
projects F1489 into the solvent away from the core of the peptide,
an unusual position for a hydrophobic residue in a short peptide.
In this position, F1489 is poised to serve as a tethered ligand that
F
M I
occludes the pore. The nearby threonine (T1491), which is an
I
H N F H N
important residue for inactivation,50 also is in position to interact
3 M 3 with the inactivation gate receptor in the pore. In contrast, the
A
CO2 CO2 methionine of the IFM motif (M1490) is buried in the core of
the peptide, interacting with two tyrosine residues in the alpha
helix. This hydrophobic interaction stabilizes the fold of the
peptide and forces F1489 into its exposed position. The structure
I1488 of the inactivation gate peptide in solution suggests that the rigid
α-helix serves as a scaffold to present the IFM motif and T1491
M1490 S1506 to a receptor in the mouth of the pore as the gate closes.
Scanning mutagenesis experiments have revealed multiple
amino acid residues that can form the inactivation gate receptor
F1489 within and near the intracellular mouth of the pore (see Figure
1-1, B, blue circles), including hydrophobic residues at the intracel-
T1491 lular end of transmembrane segment IVS654 and amino acid resi-
dues in intracellular loops IIIS4-S555 and IVS4-S5.56-59 Mutations
of residues in each of these positions impair inactivation by desta-
bilizing the inactivated state, as expected for disruption of the
inactivation gate receptor. In addition, mutations in intracellular
loop IVS4-S5 impair closed channel block by IFM-containing
B peptides, consistent with function as the inactivation gate recep-
Figure 1-6. The molecular mechanism of fast sodium channel inactivation. A, The tor,57 and paired insertions of charged residues in the IIIS4-S5
hinged-lid mechanism. The intracellular loop connecting domains III and IV of the loop and in the IFM motif indicate that these peptide segments
sodium channel is depicted as forming a hinged lid with the critical phenylalanine interact during inactivation.55 Evidently, multiple peptide seg-
(Phe1489) within the IFM motif shown occluding the mouth of the pore during the ments form a complex inactivation gate receptor into which the
inactivation process. The circles represent the transmembrane helices. B, Three- inactivation gate closes to occlude the inner pore.
dimensional structure of the central segment of the inactivation gate as deter-
mined by multidimensional NMR. Isoleucine 1488, phenylalanine 1489, and
methionine 1490 (IFM) are illustrated in yellow. Threonine 1491, which is important
for inactivation, and serine 1506, which is a site of phosphorylation and modulation Coupling of Activation to Fast Inactivation
by protein kinase C, are also indicated.
Sodium channel inactivation derives most or all of its voltage
(Adapted from Catterall WA: From ionic currents to molecular mechanisms: the struc- dependence from coupling to the activation process driven by
ture and function of voltage-gated sodium channels. Neuron 26:13–25, 2000.) transmembrane movements of the S4 voltage sensors.25 Increas-
ingly strong evidence implicates the S4 segment in domain IV in
this process. Mutations of charged amino acid residues at the
connecting domains III and IV forms an inactivation gate that extracellular end of the IVS4 segment have strong and selective
folds into the channel structure during inactivation.46,47 Cutting effects on inactivation.60 α-Scorpion toxins and sea anemone
the loop between domains III and IV by expression of the sodium toxins uncouple activation from inactivation by binding to a
channel in two pieces greatly slows inactivation.39 Mutagenesis receptor site at the extracellular end of the IVS4 segment and
studies of this region revealed a hydrophobic triad of isoleucine, preventing its normal gating movement,61,62 evidently trapping it
phenylalanine, and methionine (IFM) that is critical for fast inac- in a position that is permissive for activation but not for fast
tivation (see Figure 1-2, B, blue circle with h),48 and peptides inactivation. The IIIS4 and IVS4 segments, detected by cova-
containing this motif can serve as pore blockers and can restore lently incorporated fluorescent probes, are specifically immobi-
inactivation to sodium channels having a mutated inactivation lized in the outward position by fast inactivation, arguing that
gate.49 The latch of this fast inactivation gate is formed by three their movement is coupled to the inactivation process.63 Together,
key hydrophobic residues, IFM, and an adjacent threonine (T). these results provide strong evidence that outward movement of
These results support a model in which the IFM motif serves as the S4 segment in domain IV is the signal to initiate fast inactiva-
a tethered pore blocker that binds to a receptor in the intracel- tion of the sodium channel by closure of the intracellular inacti-
lular mouth of the pore. Inactivation is impaired in proportion vation gate. The molecular mechanism for coupling of this
to the hydrophilicity of amino acid substitutions for the key movement of IVS4 to inactivation gate closure is an interesting
phenylalanine residue (F1489), suggesting that it forms a hydro- subject for further investigation.
phobic interaction with an inactivation gate receptor during inac-
tivation.50 Voltage-dependent movement of the inactivation gate
has been detected by measuring the accessibility of a cysteine Slow Inactivation
residue substituted for F1489.51 This substituted cysteine residue
becomes inaccessible to reaction with sulfhydryl reagents as the In addition to the fast inactivation process discovered by Hodgkin
inactivation gate closes. Glycine and proline residues that flank and Huxley in their classic work, a separate slow inactivation
the IFM motif may serve as molecular hinges to allow closure of process operating on the time scale of 100 ms to seconds also
the inactivation gate like a hinged lid (see Figure 1-6, A).52 terminates the Na+ influx through Na+ channels. This process is
The three-dimensional structure of the central portion of the engaged during repetitive generation of action potentials in nerve
inactivation gate has been determined by expression as a separate and muscle cells and limits the length of trains of repetitive action
potentials. Bacterial Na+ channels whose structure has been pore and corresponding movement of the two adjacent pairs of
1
determined have a slow inactivation process, although their S6 segments away from the axis (Figure 1-7). This movement is
homotetrameric structure means that they do not have a struc- observed at the selectivity filter at the extracellular end of the
tural component analogous to the intracellular loop connecting pore, in the central cavity, and at the activation gate at the intra-
domains III and IV of vertebrate channels, which mediates fast cellular end of the pore (see Figure 1-7). This asymmetric col-
inactivation. The structure of the slow-inactivated bacterial Na+ lapse of the pore is accompanied by subtle rotation of the
channel64 reveals that the pore has partially collapsed by move- voltage-sensing domain around the cylindrical exterior surface of
ment of two opposing S6 segments toward the central axis of the the pore domain (see Figure 1-7). It is likely that the pore collapse
I217C WT-CD
Top
view
I217C WT-AB
I217C WT-AB
S6 S6
Intracellular
view
A B
Figure 1-7. Structural basis for slow inactivation. A, Structure of the functional elements of the pore in the pre-open state. Top, Selectivity filter; middle, central cavity with
amino acid residues of the drug-binding site shown in color; bottom, activation gate. B, Structure of the functional elements of the pore in the slow-inactivated state shown
as in A.
(Adapted from Payandeh J, Gamal El-Din TM, Scheuer T, et al: Crystal structure of a voltage-gated sodium channel in two potentially inactivated states. Nature 486:135–139, 2012.)
is important for stabilization of the sodium channel in the inac- The amino acid residues that form the receptor sites for Na+
tivated state, which requires strong, long-duration hyperpolariza- channel blockers line the inner surface of the S6 segments and
tion for recovery to the resting state. create a three-dimensional drug receptor site whose occupancy
would block the pore (see Figure 1-8, B). Remarkably, fenestra-
tions lead from the lipid phase of the membrane sideways into
Inner Pore and Receptor Site for Local Anesthetic the drug receptor site, providing a hydrophobic access pathway
and Antiarrhythmic Drugs for drug binding (see Figure 1-8, B). This form of drug binding
from the membrane phase was predicted in early studies of the
Sodium channels are the molecular targets for drugs used in mechanism of block of Na+ channels by different local anesthet-
control of cardiac arrhythmias, local anesthesia, prevention of ics.26 Access to the drug binding site in NavAb channels from the
acute pain, and treatment of epilepsy and bipolar disorder. membrane phospholipid bilayer is limited by the side chain of a
Sodium channel–blocking drugs are also in development for single amino acid residue (see Figure 1-8, B ), which could control
treatment of chronic pain. Sodium channel blocking drugs bind drug access and egress from the drug receptor site and possibly
to a specific receptor site within the pore of sodium channels, entry and egress of physiologic lipid modulators.
formed by the S6 segments in domains I, III, and IV21,65-68 (Figure In addition to these important pharmaceutical agents, sodium
1-8). Their binding blocks ion movement through the pore and channels are the molecular targets for a large number of neuro-
stabilizes the inactivated state of sodium channels. Antiarrhyth- toxins that paralyze prey by preventing neuromuscular func-
mic drugs and antiepileptic drugs share similar, overlapping tion.77,78 These toxins act at five or more distinct receptor sites
receptor sites. Complete block of sodium channels would be and either block the pore of the channel or alter the kinetics or
lethal; however, these drugs selectively block sodium channels in voltage dependence of gating. The pore blockers tetrodotoxin
depolarized or rapidly firing cells, such as axons carrying high- and saxitoxin bind to neurotoxin receptor site 1, which is formed
intensity pain information and rapidly firing nerve and cardiac by the P loops in the four domains (see Figure 1-2).
muscle cells that drive epileptic seizures or cardiac arrhythmias.69-71
This selective block arises because the drugs can reach their
binding site in the pore of the sodium channel more rapidly when
the pore is repetitively opened, and they bind with high affinity Sodium Channel Genes
to inactivated sodium channels that are generated in rapidly firing
or depolarized cells. The conformational change observed in the Sodium channels are the founding members of ion channel
central cavity during slow inactivation may be responsible for the superfamily that includes voltage-gated calcium channels, TRP
increased affinity for drug block of the inactivated state (see channels, voltage-gated, inward rectifying, and two-pore-domain
Figure 1-7). The use-dependent action of the sodium channel– potassium channels, and cyclic nucleotide-regulated CNG and
blocking drugs is essential for their therapeutic efficacy. HCN channels.79 In evolution, the four-domain sodium channel
High-affinity binding of local anesthetics to the inactivated was last among the voltage-gated ion channels to appear, and it
state of sodium channels requires two critical amino acid residues, is only found in multicellular organisms. It is thought that sodium
phe1764 and tyr1771 in brain type IIA channels, which are channels evolved by two rounds of gene duplication from ances-
located on the same side of the IVS6 transmembrane segment tral single-domain bacterial sodium channels. Voltage-gated
two α-helical turns apart (see Figure 1-8, A; phe1764 and tyr1771 sodium channel genes are present in a variety of metazoan species,
in blue, etidocaine in red).65-68 It is likely that the tertiary amino including the fly, leech, squid, and jellyfish. The biophysical
group of local anesthetics interacts with phe1764, which is located properties, pharmacology, gene organization, and even intron-
more deeply in the pore, and that the aromatic moiety of the local splice sites of these invertebrate sodium channels are largely
anesthetics interacts with tyr1771, which is located nearer to the similar to the mammalian sodium channels.
intracellular end of the pore (see Figure 1-8, blue residues). Sub- Ten related sodium channel genes are found in vertebrates,
sequent work has shown that sodium channel–blocking drugs of and nine encode voltage-gated sodium channels (Figure 1-9).79,80
diverse structure that are used as antiarrhythmic drugs and as More than 20 exons comprise each of the sodium channel
anticonvulsants also interact with the same site as local anesthet- α-subunit genes in mammals. Genes encoding sodium channels
ics, but also create additional interactions with other nearby Nav1 1, Nav1 2, Nav1 3, and Nav1 7 are localized on chromosome
amino acid residues.72-76 2 in humans, and these channels share similarities in sequence,
Top
Side view
view
Pore
Pore portal
portal
Figure 1-8. Drug binding in the central cavity. A, Three-dimensional model of proposed orientation of amino acid residues within the Na+ channel pore with respect to
the local anesthetic etidocaine. Only transmembrane segments IS6 (red), IIIS6 (green), and IVS6 (blue) are shown. Residues important for etidocaine binding are shown in
space-filling representation. B, Side-view through the pore module illustrating fenestrations (portals) and hydrophobic access to central cavity. Phe203 side-chains, yellow
sticks. Surface representations of NavAb residues aligning with those implicated in drug binding and block. Blue, Thr206; green, Met209; orange, Val213; gray lines, membrane
boundaries. Electron-density from an Fo-Fc omit map is contoured at 2.0 σ. C, Top-view sectioned below the selectivity filter, colored as in B.
rNaV1.1 vs. β2 and β4) may be able to substitute for each other in interac-
1
rNaV1.2 tion with sodium channel α-subunits.
rNaV1.3 Although NaV1.5 channels are primarily expressed in the heart
rNaV1.7 and are often termed the cardiac sodium channel,81 several of the
rNaV1.4
rNaV1.6
brain sodium channel subtypes are also expressed at lower levels
rNaV1.5 in the heart, where they are differentially localized in subcellular
rNaV1.8 compartments in a cell-specific and species-specific manner. In
rNaV1.9 human atrial myocytes, NaV1.5 channels are localized in a striated
pattern on the cell surface at the z-lines in each sarcomere. In
25 20 15 10 5 0 contrast, NaV1.2 channels are localized primarily at the interca-
Amino acid sequence differences (%)
lated discs, but at a lower concentration, and NaV1.1 channels are
localized at low density in a scattered punctate pattern over the
Figure 1-9. Amino acid sequence similarity of voltage-gated sodium channel cell surface. Because all the “brain” and “skeletal muscle” sodium
α-subunits. Amino acid sequence similarity of voltage-gated sodium channel channel subtypes are inhibited by nanomolar concentrations of
α-subunits. A comparison of amino acid identity for rat sodium channels Nav1.1 to
TTX and NaV1.5 channels require micromolar concentrations of
Nav1.9. The comparison was performed with Megalign in the program DNAStar
(using the Clustal method) for the four domains and the cytoplasmic linker con- TTX for inhibition, the contribution of the TTX-sensitive
necting domains III and IV. sodium channels to cardiac contractility can be tested in careful
does-response experiments. Such experiments indicate that 12%
(Adapted from Catterall WA: From ionic currents to molecular mechanisms: the struc- to 27% of sodium current is conducted by the TTX-sensitive
ture and function of voltage-gated sodium channels. Neuron 26:13–25, 2010.) sodium channels, and block of these channels reduces the ampli-
tude and velocity of contraction of atrial muscle strips approxi-
mately 15%.85 The differential localization of these subtypes
biophysical characteristics, block by nanomolar concentrations of suggests specific functions for the NaV1.2 channels in the initia-
tetrodotoxin, and broad expression in neurons. A second cluster tion of the action potential at intercalated discs and the NaV1.5
of genes encoding Nav1 5, Nav1 8, and Nav1 9 channels is local- channels in depolarizing the atrial membrane near the voltage-
ized to human chromosome 3p21-24. Although they are more gated calcium channels at the z-lines in order to efficiently initi-
than 75% identical in sequence to the group of channels on ate excitation-contraction coupling.
chromosome 2, these sodium channels all contain amino acids In human ventricular myocytes, a similar pattern of localiza-
substitutions that confer varying degrees of resistance to the pore tion of sodium channels is observed, with NaV1.2 channels pri-
blocker tetrodotoxin. In Nav1 5, the principal cardiac isoform,81 marily localized at the intercalated discs and the predominant
a single amino acid change from phenylalanine to cysteine in the NaV1.5 channels localized in a striated pattern at the z-lines. In
pore region of domain I, is responsible for a 200-fold reduction contrast to atrial myocytes, which have no T-tubules, the NaV1.5
in tetrodotoxin (TTX) sensitivity compared with those channels channels are localized at the cell surface and extend into the
on chromosome 2.82 At the identical position in Nav1 8 and Nav1 T-tubules in ventricular myocytes. It is likely that these two
9, the amino acid residue is serine, and this change results in even channel types have similar functions as in atrial myocytes—
greater resistance to TTX.83 These two channels are primarily NaV1.2 initiating the action potential at the intercalated disc and
expressed in peripheral sensory neurons. Compared with the NaV1.5 conducting the action potential along the myocyte surface
sodium channels on chromosomes 2 and 3, Nav1 4 (which is and into the T-tubule to reach the calcium channels and activate
expressed in skeletal muscle) and Nav1 6 (which is highly abun- excitation-contraction coupling.
dant in the central nervous system) have greater than 85% In contrast to this distribution of sodium channel subtypes in
sequence identity and similar functional properties, including human cardiac myocytes, the distribution of sodium channels in
TTX-sensitivity in the nanomolar concentration range. A tenth mouse ventricular myocytes is rather different.86-88 In the mouse
sodium channel, Nax, whose gene is located near the sodium heart, NaV1.5 channels are localized in high density in the inter-
channels of chromosome 2, is evolutionarily more distant.80 Key calated discs, whereas NaV1.1 channels are localized in lower
differences in functionally important regions of voltage sensor density at the z-lines.86 The high concentration of sodium chan-
and inactivation gate and lack of functional expression of voltage- nels at the ends of mouse ventricular myocytes suggests that they
gated sodium currents in heterologous cells suggest that Nax conduct action potentials in a salutatory manner, like a myelin-
might not function as a voltage-dependent sodium channel. Con- ated nerve; that is, the large sodium current entering at the
sistent with this conclusion, targeted deletion of the Nax gene in intercalated disc instantaneously depolarizes the entire myocyte
mice causes functional deficits in sensing plasma salt levels.84 and directly activates the high density of sodium channels at the
The NaVβ subunits are encoded by four distinct genes in intercalated discs at the other end of the cell, rather than being
mammals.7-10 The gene encoding β1 maps to chromosome 19q13, conducted progressively across the cell. The instantaneous depo-
whereas β2 and β4 are located on chromosome 11q22-23 and β3 larization of the entire cell surface would then initiate action
is located nearby on chromosome 11q24. The β1 and β3 subunits potentials in the T-tubules conducted by the TTX-sensitive
associate noncovalently with the α-subunits, whereas β2 and β4 sodium channels there. This altered mode of initiation and con-
are covalently linked by a disulfide bond. These distinct modes duction of the cardiac action potential could be a specialization
of association and the corresponding similarities in amino acid that supports the rapid beating rate of the mouse heart (600 bpm)
sequence suggest that the similar pairs of β-subunits (β1 and β3 without loss of synchrony and force of contraction.
sodium channel in planar lipid bilayers. Proc Natl 5. Catterall WA: Molecular properties of voltage-
References Acad Sci USA 82:240–244, 1985. sensitive sodium channels. Annu Rev Biochem
3. Numa S, Noda M: Molecular structure of sodium 55:953–985, 1986.
1. Catterall WA: The molecular basis of neuronal channels. Ann N Y Acad Sci 479:338–355, 1986. 6. Catterall WA: From ionic currents to molecular
excitability. Science 223:653–661, 1984. 4. Guy HR, Seetharamulu P: Molecular model of the mechanisms: The structure and function of
2. Hartshorne RP, Keller BU, Talvenheimo JA, et al: action potential sodium channel. Proc Natl Acad voltage-gated sodium channels. Neuron 26:13–25,
Functional reconstitution of the purified brain Sci USA 508:508–512, 1986. 2000.
7. Isom LL, De Jongh KS, Patton DE, et al: Primary 28. Noda M, Suzuki H, Numa S, et al: A single point 50. Kellenberger S, West JW, Scheuer T, et al: Molec-
structure and functional expression of the b1 mutation confers tetrodotoxin and saxitoxin insen- ular analysis of the putative inactivation particle in
subunit of the rat brain sodium channel. Science sitivity on the sodium channel II. FEBS Lett the inactivation gate of brain type IIA Na+ chan-
256:839–842, 1992. 259:213–216, 1989. nels. J Gen Physiol 109:589–605, 1997.
8. Isom LL, Ragsdale DS, De Jongh KS, et al: Struc- 29. Terlau H, Heinemann SH, Stühmer W, et al: 51. Kellenberger S, Scheuer T, Catterall WA:
ture and function of the beta-2 subunit of brain Mapping the site of block by tetrodotoxin and saxi- Movement of the Na+ channel inactivation gate
sodium channels, a transmembrane glycoprotein toxin of sodium channel II. FEBS Lett 293:93–96, during inactivation. J Biol Chem 271:30971–
with a CAM-motif. Cell 83:433–442, 1995. 1991. 30979, 1996.
9. Morgan K, Stevens EB, Shah B, et al: Beta-3: An 30. Lipkind GM, Fozzard HA: A structural model of 52. Kellenberger S, West JW, Catterall WA, et al:
additional auxiliary subunit of the voltage-sensitive the tetrodotoxin and saxitoxin binding site of the Molecular analysis of potential hinge residues in
sodium channel that modulates channel gating Na+ channel. Biophys J 66:1–13, 1994. the inactivation gate of brain type IIA Na+ chan-
with distinct kinetics. Proc Natl Acad Sci U S A 31. Penzotti JL, Fozzard HA, Lipkind GM, et al: Dif- nels. J Gen Physiol 19:607–617, 1997.
97:2308–2313, 2000. ferences in saxitoxin and tetrodotoxin binding 53. Rohl CA, Boeckman FA, Baker C, et al: Solution
10. Yu FH, Westenbroek RE, Silos-Santiago I, et al: revealed by mutagenesis of the Na+ channel outer structure of the sodium channel inactivation gate.
Sodium channel beta-4, a new disulfide-linked aux- vestibule. Biophys J 75:2647–2657, 1998. Biochemistry 38:855–861, 1999.
iliary subunit with similarity to beta-2. J Neurosci 32. Heinemann SH, Terlau H, Stühmer W, et al: 54. McPhee JC, Ragsdale DS, Scheuer T, et al: A criti-
23:7577–7585, 2003. Calcium channel characteristics conferred on the cal role for transmembrane segment IVS6 of the
11. McCormick KA, Srinivasan J, White K, et al: The sodium channel by single mutations. Nature sodium channel a subunit in fast inactivation. J Biol
extracellular domain of the b1 subunit is both nec- 356:441–443, 1992. Chem 270:12025–12034, 1995.
essary and sufficient for b1-like modulation of 33. Chiamvimonvat N, O’Rourke B, Kamp TJ, et al: 55. Smith MR, Goldin AL. Interaction between the
sodium channel gating. J Biol Chem 274:32638– Functional consequences of sulfhydryl modifica- sodium channel inactivation linker and domain III
32646, 1999. tion in the pore-forming subunits of cardiovascular S4-S5. Biophys J 73:1885–1895, 1997.
12. Ratcliffe CF, Qu Y, McCormick KA, et al: A Ca2+ and Na+ channels. Circ Res 76:325–334, 56. Lerche H, Peter W, Fleischhauer R, et al: Role in
sodium channel signaling complex: Modulation by 1995. fast inactivation of the IV/S4-S5 loop of the human
associated receptor protein tyrosine phosphatase b. 34. Pérez-García MT, Chiamvimonvat N, Marban E, muscle Na+ channel probed by cysteine mutagen-
Nat Neurosci 3:437–444, 2000. et al: Structure of the sodium channel pore revealed esis. J Physiol (Lond) 505:345–352, 1997.
13. Srinivasan J, Schachner M, Catterall WA: Interac- by serial cysteine mutagenesis. Proc Natl Acad Sci 57. McPhee JC, Ragsdale D, Scheuer T, et al: A critical
tion of voltage-gated sodium channels with the USA 93:300–304, 1996. role for the S4-S5 intracellular loop in domain IV
extracellular matrix molecules tenascin-C and 35. Schlief T, Schönherr R, Imoto K, et al: Pore prop- of the sodium channel a subunit in fast inactivation.
tenascin-R. Proc Natl Acad Sci USA 95:15753– erties of rat brain II sodium channels mutated in J Biol Chem 273:1121–1129, 1998.
15757, 1998. the selectivity filter domain. Eur Biophys J 25:75– 58. Tang LH, Chehab N, Wieland SJ, et al:
14. Ratcliffe CF, Westenbroek RE, Curtis R, et al: 91, 1996. Glutamine substitution at Alanine1649 in the
Sodium channel beta-1and beta-3 subunits associ- 36. Sun YM, Favre I, Schild L, et al: On the S4-S5 cytoplasmic loop of domain 4 removes the
ate with neurofascin through their extracellular structural basis for size-selective permeation of voltage sensitivity of fast inactivation in the human
immunoglobulin-like domain. J Cell Biol 154:427– organic cations through the voltage-gated sodium heart sodium channel. J Gen Physiol 111:639–652,
434, 2001. channel—Effect of alanine mutations at the DEKA 1998.
15. Kazarinova-Noyes K, Malhotra JD, McEwen DP, locus on selectivity, inhibition by Ca2+ and H+, 59. Filatov GN, Nguyen TP, Kraner SD, et al: Inacti-
et al: Contactin associates with sodium channels and molecular sieving. J Gen Physiol 110:693–715, vation and secondary structure in the D4/S4-5
and increases their functional expression. J Neuro- 1997. region of the SkM1 sodium channel. J Gen Physiol
sci 21:7517–7525, 2001. 37. Hille B: Ionic selectivity, saturation, and block in 111:703–715, 1998.
16. Malhotra JD, Koopmann MC, Kazen-Gillespie sodium channels. A four-barrier model. J Gen 60. Chen LQ, Santarelli V, Horn R, et al: A unique
KA, et al: Structural requirements for interaction Physiol 66:535–560, 1975. role for the S4 segment of domain 4 in the inactiva-
of sodium channel beta 1 subunits with ankyrin. 38. Hirschberg B, Rovner A, Lieberman M, et al: tion of sodium channels. J Gen Physiol 108:549–
J Biol Chem 277:26681–26688, 2002. Transfer of twelve charges is needed to open skel- 556, 1996.
17. Xiao ZC, Ragsdale DS, Malhotra JD, et al: etal muscle Na+ channels. J Gen Physiol 106:1053– 61. Rogers JC, Qu Y, Tanada TN, et al: Molecular
Tenascin-R is a functional modulator of sodium 1068, 1995. determinants of high affinity binding of α-scorpion
channel b subunits. J Biol Chem 274:26511–26517, 39. Stuhmer W, Conti F, Suzuki H, et al: Structural toxin and sea anemone toxin in the S3-S4 extracel-
1999. parts involved in activation and inactivation lular loop in domain IV of the Na+ channel α
18. Chen C, Bharucha V, Chen Y, et al: Reduced of the sodium channel. Nature 339:597–603, subunit. J Biol Chem 271:15950–15962, 1996.
sodium channel density, altered voltage depen- 1989. 62. Sheets MF, Kyle JW, Kallen RG, et al: The Na
dence of inactivation, and increased susceptibility 40. Yang N, Horn R: Evidence for voltage-dependent channel voltage sensor associated with inactivation
to seizures in mice lacking sodium channel beta S4 movement in sodium channel. Neuron 15:213– is localized to the external charged residues of
2-subunits. Proc Natl Acad Sci U S A 99:17072– 218, 1995. domain IV, S4. Biophys J 77:747–757, 1999.
17077, 2002. 41. Yang NB, George AL, Jr, Horn R: Molecular basis 63. Cha A, Ruben PC, George AL, et al: Voltage
19. Chen C, Westenbroek RE, Xu X, et al: Mice of charge movement in voltage-gated sodium chan- sensors in domains III and IV, but not I and II, are
lacking sodium channel beta1 subunits display nels. Neuron 16:113–122, 1996. immobilized by Na+ channel fast inactivation.
defects in neuronal excitability, sodium channel 42. Catterall WA: Ion channel voltage sensors: struc- Neuron 22:73–87, 1999.
expression, and nodal architecture. J Neurosci ture, function, and pathophysiology. Neuron 64. Payandeh J, Gamal El-Din TM, Scheuer T, et al:
24:4030–4042, 2004. 67:915–928, 2010. Crystal structure of a voltage-gated sodium channel
20. Lopez-Santiago LF, Meadows LS, Ernst SJ, et al: 43. Jiang Y, Lee A, Chen J, et al: Crystal structure and in two potentially inactivated states. Nature
Sodium channel Scn1b null mice exhibit prolonged mechanism of a calcium-gated potassium channel. 486:135–139, 2012.
QT and RR intervals. J Mol Cell Cardiol 43:636– Nature 417:515–522, 2002. 65. Ragsdale DS, McPhee JC, Scheuer T, et al: Molec-
647, 2007. 44. Jiang Y, Lee A, Chen J, et al: The open pore con- ular determinants of state-dependent block of
21. Payandeh J, Scheuer T, Zheng N, et al: The crystal formation of potassium channels. Nature 417:523– sodium channels by local anesthetics. Science
structure of a voltage-gated sodium channel. 526, 2002. 265:1724–1728, 1994.
Nature 475:353–358, 2011. 45. Zhao Y, Yarov-Yarovoy V, Scheuer T, et al: A gating 66. Liu G, Yarov-Yarovoy V, Qu Y, et al: Differential
22. Shapiro L, Doyle JP, Hensley P, et al: Crystal struc- hinge in Na+ channels; a molecular switch for elec- Interactions of lamotrigine and related drugs with
ture of the extracellular domain from Po, the major trical signaling. Neuron 41:859–865, 2004. transmembrane segment IVS6 of voltage-gated
structural protein of peripheral nerve myelin. 46. Vassilev PM, Scheuer T, Catterall WA: Identifica- sodium channels. Neuropharm 44:413–422, 2003.
Neuron 17:435–449, 1996. tion of an intracellular peptide segment involved in 67. Yarov-Yarovoy V, Brown J, Sharp E, et al: Molecu-
23. Isom LL: Sodium channel b-subunits: Anything sodium channel inactivation. Science 241:1658– lar determinants of voltage-dependent gating and
but auxiliary. Neuroscientist 7:42–54, 2001. 1661, 1988. binding of pore-blocking drugs in transmembrane
24. Hodgkin AL, Huxley AF: A quantitative descrip- 47. Vassilev P, Scheuer T, Catterall WA: Inhibition of segment IIIS6 of the Na+ channel a subunit. J Biol
tion of membrane current and its application to inactivation of single sodium channels by a site- Chem 276:20–27, 2001.
conduction and excitation in nerve. J Physiol directed antibody. Proc Natl Acad Sci USA 68. Yarov-Yarovoy V, McPhee JC, Idsvoog D, et al:
117:500–544, 1952. 86:8147–8151, 1989. Role of amino acid residues in transmembrane seg-
25. Armstrong CM: Sodium channels and gating cur- 48. West JW, Patton DE, Scheuer T, et al: A cluster of ments IS6 and IIS6 of the sodium channel alpha
rents. Physiol Rev 61:644–682, 1981. hydrophobic amino acid residues required for fast subunit in voltage-dependent gating and drug
26. Hille B: Ionic Channels of Excitable Membranes, Na+ channel inactivation. Proc Natl Acad Sci USA block. J Biol Chem 277:35393–35401, 2002.
ed 3, Sunderland, MA, 2001, Sinauer Associates 89:10910–10914, 1992. 69. Hille B: Local anesthetics: hydrophilic and hydro-
Inc. 49. Eaholtz G, Scheuer T, Catterall WA: Restoration phobic pathways for the drug-receptor reaction. J
27. Hille B: The permeability of the sodium channel of inactivation and block of open sodium channels Gen Physiol 69:497–515, 1977.
to metal cations in myelinated nerve. J Gen Physiol by an inactivation gate peptide. Neuron 12:1041– 70. Hondeghem LM, Katzung BG: Antiarrhythmic
59:637–658, 1972. 1048, 1994. agents: The modulated receptor mechanism of
action of sodium and calcium channel blocking sodium channels. Mol Pharmacol 68:1611–1622, 84. Hiyama TY, Watanabe E, Okado H, et al: The
drugs. Annu Rev Pharmacol Toxicol 24:387–423, 2005. subfornical organ is the primary locus of sodium-
71.
1984.
Catterall WA: Common modes of drug action on
Na+ channels: Local anesthetics, antiarrhythmics
77. Cestele S, Catterall WA: Molecular mechanisms of
neurotoxin action on voltage-gated sodium chan-
nels. Biochimie 82:883–892, 2000.
level sensing by Na(x) sodium channels for the
control of salt-intake behavior. J Neurosci
24:9276–9281, 2004.
1
and anticonvulsants. Trends Pharmacol Sci 8:57– 78. Catterall WA, Cestele S, Yarov-Yarovoy V, et al: 85. Westenbroek R, Kaufmann S, Lange V, et al: Local-
65, 1987. Voltage-gated ion channels and gating modifier ization and function of sodium channel subtypes in
72. Ragsdale DR, McPhee JC, Scheuer T, et al: toxins. Toxicon 49:124–141, 2007. human heart. Circulation 116:11–12, 2007.
Common molecular determinants of local anes- 79. Yu FH, Catterall WA: The VGL-chanome: a 86. Maier SK, Westenbroek RE, Schenkman KA, et al:
thetic, antiarrhythmic, and anticonvulsant block of protein superfamily specialized for electrical sig- An unexpected role for brain-type sodium channels
voltage-gated Na+ channels. Proc Natl Acad Sci naling and ionic homeostasis. Sci STKE 2004:re15, in coupling of cell surface depolarization to con-
USA 93:9270–9275, 1996. 2004. traction in the heart. Proc Natl Acad Sci USA
73. Wright SN, Wang SY, Wang GK: Lysine point 80. Goldin AL, Barchi RL, Caldwell JH, et al: Nomen- 99:4073–4078, 2002.
mutations in Na+ channel D4-S6 reduce inacti- clature of voltage-gated sodium channels. Neuron 87. Maier SK, Westenbroek RE, Yamanushi TT, et al:
vated channel block by local anesthetics. Mol Phar- 28:365–368, 2000. An unexpected requirement for brain-type sodium
macol 54:733–739, 1998. 81. Rogart RB, Cribbs LL, Muglia LK, et al: Molecu- channels for control of heart rate in the mouse
74. Nau C, Wang SY, Wang GK: Point mutations at lar cloning of a putative tetrodotoxin-resistant rat sinoatrial node. Proc Natl Acad Sci U S A
L1280 in Nav1.4 channel D3-S6 modulate binding heart Na+ channel isoform. Proc Natl Acad Sci 100:3507–3512, 2003.
affinity and stereoselectivity of bupivacaine enan- USA 86:8170–8174, 1989. 88. Maier SK, Westenbroek RE, McCormick KA, et
tiomers. Mol Pharmacol 63:1398–1406, 2003. 82. Satin J, Kyle JW, Chen M, et al: A mutant of TTX- al: Distinct subcellular localization of different
75. Weiser T, Qu Y, Catterall WA, et al: Differential resistant cardiac sodium channels with TTX- sodium channel alpha and beta subunits in single
interaction of R-mexiletine with the local anes- sensitive properties. Science 256:1202–1205, ventricular myocytes from mouse heart. Circula-
thetic receptor site on brain and heart sodium 1992. tion 109:1421–1427, 2004.
channel alpha-subunits. Mol Pharmacol 56:1238– 83. Sivilotti L, Okuse K, Akopian AN, et al: A single 89. Yarov-Yarovoy V, DeCaen PG, Westenbroek RE,
1244, 1999. serine residue confers tetrodotoxin insensitivity on et al: Structural basis for gating charge movement
76. Lipkind GM, Fozzard HA: Molecular modeling of the rat sensory-neuron-specific sodium channel in the voltage sensor of a sodium channel. Proc
local anesthetic drug binding by voltage-gated SNS. FEBS Lett 409:49–52, 1997. Natl Acad Sci U S A 109:E93–E102, 2012.
Calcium Channels in the Heart 2
Robert A. Rose and Peter H. Backx
of CaV1.3 have also been described.8,9 These variants result in the

CHAPTER OUTLINE
expression of long and short forms of CaV1.3, which are charac-
L-type Ca2+ Channels 13 terized by differences in biophysical properties such as the activa-
2+ tion midpoint; however, not all of the short forms of CaV1.3 are
T-type Ca Channels 16
expressed in the heart.8
Transient Receptor Potential (TRP) Channels 18 Ca2+ channel β-subunits (CaVβ1 through CaVβ4) are encoded by
four genes (CACNB1-4) and, like the α1-subunits, can be alterna-
tively spliced to generate a number of isoforms.10,11 Crystal struc-
Calcium (Ca2+) permeable ion channels are involved in a number tures of LTCC β-subunits demonstrate that these subunits
of fundamental processes in the heart, including automaticity in contain conserved Src homology 3 (SH3) and guanylate kinase
the sinoatrial (SAN) and atrioventricular (AVN) nodes, excitation- domains, as seen in scaffolding proteins in the membrane-
contraction coupling in the working myocardium, and regulation associated guanylate kinase (MAGUK) family.12,13 CaVβ-subunits
of gene expression in hypertrophic signaling.1-4 Several ion chan- are intracellular proteins (see Figure 2-1) that bind to a single site,
nels in the plasma membrane of cardiac myocytes and cardiac called the α interaction domain (AID), on the loop connecting
fibroblasts are involved in transporting Ca2+ into the cytosol from domains I and II on the α1-subunit of the LTCC. CaVβ acts as a
the extracellular space (Table 2-1). Included among these are two chaperone protein that facilitates the trafficking of LTCCs to the
forms of L-type Ca2+ channels (LTCC; CaV1.2 and CaV1.3), plasma membrane and also modulates the biophysical properties
T-type Ca2+ channels, and several transient receptor potential of these channels. Specifically, coexpression of CaVβ2 with CaV1.2
(TRP) channels. These different Ca2+ permeable ion channels increases ICa,L amplitude, accelerates activation and inactivation
play distinct roles in different parts of the heart. For example, kinetics, and shifts the steady inactivation curve to more negative
CaV1.2-mediated L-type Ca2+ current (ICa,L) is present in all car- membrane potentials. CaVβ also profoundly enhances the affinity
diomyocytes, whereas CaV1.3-mediated ICa,L is restricted to SAN, of dihydropyridines (blockers of LTCCs) for the α1-subunit by
AVN, and working atrial myocardium. T-type Ca2+ currents decreasing their dissociation from the channel.14
(ICa,T) are mainly expressed in the SAN and AVN in normal Four genes (CACNA2D1-4), which can be alternatively
conditions, but may be expressed in ventricular myocytes in spliced, encode the α2-δ-subunit of the LTCC.15,16 The α2-δ1 and
cardiac hypertrophy. In most cases, TRP channels conduct non- α2-δ3 isoforms are expressed in the heart. The protein is cleaved
selective cation currents that include Ca2+ influx, which may be post-translationally and then is relinked via disulfide interactions.
important in both cardiac myocytes and cardiac fibroblasts, the This subunit is primarily extracellular, with the δ-subunit portion
latter of which don’t typically express robust voltage-gated Ca2+ anchored in the plasma membrane (see Figure 2-1). CaVα2-δ-
channels. The purpose of this chapter is to review the expression subunits facilitate the targeting of LTCCs to the plasma mem-
patterns, biophysical properties, and structure-function relation- brane and increase ICa,L by shifting the voltage dependence of
ships of Ca2+ permeable channels in the heart. activation and inactivation to hyperpolarized membrane
potentials.
Ca2+ channel γ-subunits are membrane-bound proteins
encoded by eight genes (CACNG1-8) with several isoforms,
L-type Ca2+ Channels including γ4, γ6, γ7, and γ8, which are present in cardiac muscle.
These isoforms associate with CaV1.2 and alter activation and
Molecular Composition inactivation properties of the channel.17
LTCCs are multimeric proteins consisting of an α1-subunit that

constitutes the pore of the channel and several accessory subunits Biophysical Properties of ICa,L
denoted β, α2-δ, and γ2,5,6 (Figure 2-1). Currently, four α1-subunits
for L-type Ca2+ channels are known, and two of these, CaV1.2 As their name suggests, LTCCs are highly selective for Ca2+ over
(α1C encoded by the CACNA1C gene) and CaV1.3 (α1D encoded monovalent cations. The channel pore is approximately 6 Å in
by the CACNA1D gene), are expressed in the heart3,5,7 (see Table diameter at its narrowest point,18 indicating that size is not the
2-1). These α1-subunits form the channel pore and contain the main factor in determining selectivity. Rather, a series of four
voltage sensor that controls channel gating, as well as drug glutamate residues (EEEE motif) is responsible for conferring
binding sites and regulatory sites targeted by second messengers. ion selectivity through the ability of this motif to bind divalent
Ca2+ channel α1-subunits have a similar structure to voltage-gated cations (Ca2+ and Mg2+), which block the passage of monovalent
Na+ and K+ channels, whereby they are organized into four cations.19,20 As additional divalent ions enter the channel, bound
domains (I through IV), each containing six transmembrane seg- Ca2+ ions are displaced from the EEEE motif and are pushed
ments (S1 through S6). The voltage sensor is located in the S4 through the pore, which generates ICa,L.
transmembrane segment of each domain, and the pore loops are At the single-channel level, ICa,L conductance in ventricular
located between S5 and S6. Alternative splice variants of CaV1.2 myocytes (i.e., CaV1.2 dependent) has been reported to be
have been documented and can impact regulation by second approximately 5 pS in 2 mM Ca2+ and approximately 15 pS with
messengers and drug binding. Recently, alternative splice variants Ba2+ as the charge carrier.21 Similarly, single-channel conductance
13
Table 2-1. Nomenclature of the Voltage-Gated Ca2+ Channels Along With Chromosome Location and Biophysical and Pharmacological Properties
Chromosome Biophysical
Channel Tissue Localization Pore Subunit Gene Name Location Property Pharmacological Blockers
CaV1.1 skeletal muscle α1S CACNA1S 1q31-32 L-type nifedipine, verapamil, diltiazem
CaV1.2 heart, brain, smooth muscle, α1C CACNA1C 12p13 L-type nifedipine, verapamil, diltiazem
adrenal gland
CaV1.3 brain, pancreas, kidney, cochlea, α1D CACNA1D 3p14.3 L-type nifedipine, verapamil, diltiazem
heart
CaV1.4 retina α1F CACNA1F Xp11.23 L-type nifedipine, verapamil, diltiazem
CaV2.1 brain, cochlea α1A CACNA1A 19p13.1 P/Q-type ω-agatoxin IVA
CaV2.2 brain α1B CACNA1B 9q34 N-type ω-conotoxin GVIA
CaV2.3 brain, heart, cochlea α1E CACNA1E 1q25-31 R-type ω-agatoxin IIA
CaV3.1 brain, heart α1G CACNA1G 17q22 T-type mibefradil, kurtoxin, Ni2+
CaV3.2 brain, heart, kidney, liver α1H CACNA1H 16p13.3 T-type mibefradil, kurtoxin, Ni2+
CaV3.3 brain α1I CACNA1I 22q13 T-type mibefradil, kurtoxin?, Ni2+
γ α1 α2
+H3N CO–2 +H3N

domain:I II III IV δ
outside
+ + + +
12345 6 12345 6 12345 6 12345 6
+ + + +
inside
+H3N CO–2
+H3N
CO–2
+H3N
CO–2
Figure 2-1. Ca2+ channel subunit structure. L-type Ca2+ channels consist of the pore-forming α1-subunit and a series of accessory subunits (β, γ, α2δ). Predicted α-helices
are depicted as cylinders. The lengths of the lines correlate with the approximate length of the polypeptide segments. T-type Ca2+ channels consist of the pore-forming
α1-subunit; however, whether accessory subunits associate with T-type Ca2+ channels in native cardiomyocytes is less clear.
(From Catterall W: Voltage-gated calcium channels. Cold Spring Harb Perspect Biol 3:a003947, 2011.)
for CaV1.3 has been determined in heterologous expression −15 mV. In contrast, recombinant CaV1.3–dependent ICa,L acti-
systems and has been found to be approximately 15 pS with Ba2+ vates at more negative membrane potentials (i.e., the V1/2(act) is
as the charge carrier.8 hyperpolarized) and displays slower inactivation kinetics.8,23,24
ICa,L is mediated by CaV1.2 and CaV1.3 in the heart, and these CaV1.3 is also less sensitive to dihydropyridines than is CaV1.2.
two channel isoforms are distinguished by their unique biophysi- CaV1.2 is the primary determinant of ICa,L in the ventricular
cal properties7,22 (Figure 2-2). CaV1.2-mediated ICa,L, which is myocardium, where it plays a prominent role in excitation-
expressed throughout the myocardium (atrium, ventricles, and contraction coupling and the Ca2+ induced–Ca2+ release process.1
conduction system), has a bell-shaped current-voltage (I-V) rela- Approximately 75% of these LTCCs are located in dyads, in close
tionship in which the current activates at membrane potentials proximity to the ryanodine receptors located in the junctional
positive to −40 mV and peaks between 0 and +10 mV. The V1/2 sarcoplasmic reticulum, within the t-tubular system (see Figure
of channel activation (V1/2(act) ) is typically between −10 and 2-4). Consistent with the critical role of CaV1.2 in the ventricles,
v (mV)
–80 –60 –40 –20 0 20 40 1.5
2
I/max
0.5
HVA CAV1.2
LVA/CA.T
1.0 0.0
–100 –80 –60 –40 –20 0
A “Intermediate” CAV1.3 B mV
Figure 2-2. Voltage dependence of activation and steady state inactivation of ICa,T, CaV1.3-mediated ICa,L, and CaV1.2-mediated ICa,L from mouse SAN. A, Current-voltage (I-V)
relationships illustrating that ICa,T activates most negatively, and CaV1.2-mediated ICa,L activates most positively. CaV1.3-mediated ICa,L has intermediate activation properties.
B, Steady state inactivation curves for ICa,T, CaV1.3-mediated ICa,L, and CaV1.2-mediated ICa,L. Dashed lines indicate the voltage at which 50% of channels are inactivated (V1/2(inact) )
for each Ca2+ channel. Dotted lines indicate points of complete current inactivation and full channel availability.
(From Mangoni MM, Couette B, Marger L, et al: Voltage-dependent calcium channels and cardiac pacemaker activity: from ionic currents to genes. Prog Biophys Mol Biol 90:38,
2006.)
CaV1.2 knockout mice die before birth in association with cardiac Ca2+-dependent inactivation of ICa,L is a calmodulin (CaM)
failure.25 CaV1.2 is also expressed in the SAN, AVN, and atrial dependent process.35,36 CaM is bound to the C-terminus of the
myocardium, along with CaV1.3.22,26 As a result, total ICa,L in these channel via the IQ domain, and when local Ca2+ is elevated, an
supraventricular tissues is dependent on both α1C and α1D pore- increase in Ca2+ binding to CaM occurs. This enhances the inter-
forming subunits and demonstrates biophysical characteristics action of CaM with the IQ domain, thereby causing inactivation
that are different from those of the ventricles. The generation of of the channel. Ca2+-dependent inactivation may serve as a pro-
CaV1.3 knockout mice has been important in determining the tective negative feedback mechanism to prevent Ca2+ overload in
biophysical properties of these channels and in distinguishing cardiomyocytes.
them from CaV1.2-mediated ICa,L (and ICa,T).23,27 Specifically, these Steady state inactivation (i.e., channel availability), like
CaV1.3 knockouts have been used to show that ICa,L in SAN and voltage-dependent activation, follows a sigmoidal relationship,
atrial myocytes activates between −60 and −50 mV in association with a V1/2(inact) of ≈−35 mV for CaV1.2-dependent ICa,L and
with a left shift in the V1/2(act) and peaks at membrane potentials −45 mV for CaV1.3-dependent ICa,L.22 The activation and inacti-
around −10 mV. The steady state ICa,L inactivation curve is also vation curves for ICa,L can overlap, resulting in a “window current,”
shifted to the left when CaV1.3 contributes to total ICa,L. Together, which has been postulated to play a role in the generation of
available data show that CaV1.3-dependent ICa,L has intermediate arrhythmogenic early afterdepolarizations (EADs; see Figure
biophysical properties compared with CaV1.2-dependent ICa,L 2-3, B).37,38
and ICa,T. LTCCs are also known to undergo a process called facilitation,
The unique biophysical properties of CaV1.3 enable ICa,L to whereby a progressive increase in ICa,L amplitude and the time
play a prominent role in pacemaker activity in the SAN and in constant of inactivation can occur during increases in pacing
the electrical conduction in the atrial myocardium. In the SAN, frequency.34,39 The facilitation process occurs as the result of a
CaV1.3-mediated ICa,L contributes to the diastolic depolarization reduction in Ca2+-dependent inactivation and is mediated by
phase of the action potential and thus is a major determinant of calmodulin-dependent kinase II (CaMKII) phosphorylation.40
heart rate in vivo. CaV1.2, on the other hand, contributes more
prominently to the action potential upstroke in SAN myocytes.
Consistent with this, CaV1.3 knockout mice display sinus brady- Organization and Localization of LTCCs
cardia in association with a reduced diastolic depolarization
slope.26-28 CaV1.3 knockout mice also show increased susceptibil- It is now appreciated that LTCC can be organized into distinct
ity to atrial fibrillation,24,29 a very common cardiac arrhythmia, subcellular compartments through the actions of scaffolding pro-
confirming that these Ca2+ channels play an important role in teins, including A-kinase anchoring proteins (AKAPs),41,42 within
atrial conduction. cardiomyocytes43 (Figure 2-4). Examples of subpopulations of
LTCCs include those found in dyads within the T-tubules and
those in separate plasma membrane domains such as caveolae and
ICa,L Inactivation lipid rafts.2 This pattern of organization is critical for the precise
spatio-temporal modulation and regulation of LTCCs in differ-
ICa,L undergoes decay during sustained depolarization, which is ent physiological functions such as excitation-contraction cou-
termed inactivation (Figure 2-3, A). This inactivation process is pling, transcriptional regulation, and responses to β-adrenergic
time, voltage, and calcium dependent.30-32 Voltage-dependent (β-AR) receptor activation.
activation is relatively slow, as is indicated by the degree of inac- Within dyads, LTCCs from a complex with β-ARs, adenylyl
tivation with Ba2+ as the charge carrier, or when monovalent cyclase (AC), protein kinase A (PKA), and AKAPs.2,44,45 These
cations permeate LTCCs in the absence of divalent cations.33 In proteins area also in close apposition to a complex of proteins
the presence of extracellular Ca2+, inactivation is much faster, and associated with the SR, including ryanodine receptors, PKA,
this Ca2+-dependent inactivation is further accelerated in the protein phosphatase 2A (PP2A), phosphodiesterases, and AKAPs
presence of functional SR Ca2+ release (i.e., during normal (AKAP5 and AKAP15).41,42 It is the AKAPs that are thought to
excitation-contraction coupling).34 This indicates that both the be responsible for the organization of these signaling complexes.
Ca2+ entering the myocytes through the LTCC and the Ca2+ Within caveolae, LTCCs associate with β2-ARs, AC, PKA, PP2A,
released from the SR during the Ca2+ transient contribute to and caveolin 3, all of which enable LTCCs in caveolae to be
LTCC inactivation. locally stimulated by β2-AR activation.46
Surface Membrane T-Tubule Caveolae
Cnv-3
Ca2+
Cnv-3
LTCC Ca2– – AR
AP C
RyR2
C
Ca2+ 2
LT
AKAP
PKA Persistent
AK
AC PP2A Ca2+
Ca2+
RyR2 PKA Sparklets
–AR
LTCC
Sarcoplasmic Ca2+ LTCC Ca2+
Reticulum
LTCC
AKAP
PP2B PKCa
EC Coupling Excitability Gene Regulation Nucleus
Figure 2-3. A, Ca2+ channel inactivation with Ca2+, Ba2+, or monovalent cations (ns) as the charge carrier. Currents were measured at 0 mV except for Ins at −30 mV to obtain
comparable activation, and peak currents were normalized. ICa with SR Ca2+ release (i.e., Ca2+ transients occur) was recorded using the perforated patch-clamp technique
and 2 mM external Ca2+. ICa with no SR Ca2+ release (i.e., no Ca2+ transients occur) was recorded in the whole-cell configuration with 10 mM EGTA in the pipette. IBa was
recorded in the whole-cell configuration with 2 mM external Ba2+ and 10 mM EGTA in the pipette. Ins was measured in divalent cation-free conditions. The t1/2 of current
decline progressively increases from top to bottom, illustrating that ICa inactivation is both Ca2+ and voltage dependent. B, Overlap of steady state activation and inactivation
curves for ICa,L illustrates the presence of a window current (dashed lines), which may contribute to the generation of early afterdepolarizations (EADs). ICa,T can also produce
window currents at more negative membrane potentials (between −80 and −40 mV).
(A, From Bers DM, Perez-Reyes E: Ca channels in cardiac myocytes: structure and function in Ca influx and intracellular Ca release. Cardiovasc Res 42:339-360, 1999. B, From
Benitah JP, Alvarez JL, Gomez AM: L-type Ca2+ current in ventricular cardiomyocytes. J Mol Cell Cardiol 48:26-36, 2010.)
Normalized Ca Channel Current LTCC also demonstrate a property called coupled gating, which
ICa (perforated) occurs when LTCCs are grouped together into clusters in the
plasma membrane, resulting in localized regions with signifi-
ICa (ruptured and EGTA) cantly elevated intracellular Ca2+ concentrations. Coupled gating
is enhanced by the scaffolding protein AKAP5 and by protein
IBa kinase C (PKC), which form the signaling complex with the
clusters of LTCCs.43,47
Ins
T-type Ca2+ Channels
50 ms
A
Molecular Composition and Biophysical Properties
1.0 1.0 Although L-type Ca2+ channels get the lion’s share of the atten-
tion in both cardiac and vascular electrophysiology, low voltage-
Availability
Activation
activated T-type Ca2+ channels also play important roles in

0.5 0.5 normal and diseased hearts. Three T-type Ca2+ channels exist in
mammals: Cav3.1 (α1G), Cav3.2 (α1H), and Cav3.3 (α1I), with
Cav3.1 and Cav3.2 channels being the major channels in heart48,49
0.0 0.0 (see Table 2-1). Although L-type Ca2+ channels are complex
–30 0 30 mV
oligomeric proteins that are heavily regulated by many distinct
20 signal transduction mechanisms, much less is known about the
0 auxiliary subunit structure for T-type Ca2+ channels.50 The elec-
EAD
–20 trophysiological properties of T-type α1-subunits expressed
–40 alone are very similar to those observed in native channels, sug-
–60
gesting that T-type Ca2+ channels do not require auxiliary sub-
200 ms units, as do L-type Ca2+ channels, for proper function. Consistent
–80
B with this, antisense depletion of CaVβ-subunits has no effect on
ICa,T (in neurons),51,52 and coexpression of cloned CaVβ-subunits
Figure 2-4. Subcellular localization of L-type Ca2+ channels in cardiomyocytes. has no major effect on ICa,T in heterologous expression systems.53
LTCCs associate with anchoring proteins (AKAPs) and other signaling proteins to CaVγ-subunits have been found to have little or no effect on
create distinct populations of LTCCs in T-tubules and within caveolae. Also, cluster-
CaV3.3 currents54; however, they have been found to accelerate
ing of LTCCs in discrete regions of the plasma membrane results in coupled gating.
See text for details.
inactivation and negatively shift steady state inactivation of the
CaV3.1 current.55 Coexpression of α2δ-subunits with α1G doubled
(From Best JM, Kamp TJ: Different subcellular populations of L-type Ca2+ channels the size of ICa,T, possibly through increased expression of α1-α2δ
exhibit unique regulation and functional roles in cardiomyocytes. J Mol Cell Cardiol complexes at the plasma membrane.53,56 T-type Ca2+ channels
52:376-387, 2012.) have been shown to interact with, and be modulated by, other
proteins such as Kelch-like 1 protein57 and caveolin-3.58 In progenitors. In this regard, T-type Ca2+ channels, along with
2
summary, available data indicate that CaV3 α1-subunits can inter- hyperpolarization activated cyclic nucleotide gated (HCN) chan-
act with some auxiliary subunits, but additional studies are needed nels, are expressed at high levels in embryonic stem cells and
to determine the physiological role for these interactions, par- undergo a subtype switch during maturation from Cav3.2 to
ticularly in the heart. Cav3.1 α-subunits as they differentiate into cardiac myocytes.86
The biophysical properties of T-type Ca2+ channels in heart Control of the developmental and regional expression patterns
are distinct from those of L-type Ca2+ channels.59 For example, of T-type Ca2+ channels is poorly understood. Tbx3 is reported
T-type Ca2+ channels activate and inactivate at lower (more nega- to be important in SAN development, along with expression of
tive) membrane potentials with thresholds for activation observed Cav3.1 channels and other channel genes associated with SAN.87,88
at about −70 mV while inactivation begins to occur at approxi-
mately −90 mV50 (see Figure 2-2). T-type Ca2+ channels also show
much faster entry and exit from inactivation50,60 that is indepen- T-type Ca2+ Channels in Pacemaking, Cardiac
dent of Ca2+. An analysis of the steady state activation and inac- Conduction, and Heart Disease
tivation curves reveals that T-type Ca2+ channels have relatively
large steady state window currents at membrane voltages between Unlike L-type Ca2+ channels, the functional effects of T-type Ca2+
−80 and −40 mV,61 which are expected to facilitate depolarization channel activity have been more challenging to demonstrate.
in myocytes of the SAN and the conducting system, thereby Pharmacologic and genetic studies support the conclusion that
promoting spontaneous action potential firing. Although the per- T-type Ca2+ channels contribute to pacemaker activity by provid-
meation properties of T-type Ca2+ channels have been challenging inward current during the diastolic depolarization,22,62,89
ing to quantify, the single-channel conductance is reported to be although these effects are relatively modest compared with the
2.5-fold smaller than that of L-type Ca2+ channels (when Ba2+ is contributions of other depolarizing currents in SAN myocytes.70,78
the charge carrier).62 ICa,T also contributes to spontaneous firing and conductance prop-
T-type Ca2+ channels have distinct pharmacologic properties erties of the AVN.22,80,90 Mice with Cav3.1 ablation have reduced
compared with L-type Ca2+ channels, and these differences have heart rates and slowed AVN conduction,79 while Cav3.2-deficient
been exploited for dissecting the physiologic role of ICa,T in the mice are indistinguishable from wild type mice.91 These observa-
myocardium. Relative to L-type channels, T-type channels have tions are consistent with the prevalence of Cav3.1 channels, com-
a very high sensitivity to block by Ni2+, and this sensitivity varies pared with Cav3.2 channels, in the SAN of adult mice, as has been
between different T-type channel isoforms (half maximal inhibi- mentioned. A major complication of dissecting the role of T-type
tory concentration [IC50] for Cav3.2 is nearly 10-fold lower than Ca2+ channels in pacemaker function is the presence of Cav1.3-
that for Cav3.1 channels).50,61,63 Putative T-type channel blockers dependent L-type Ca2+ channels,23 along with Cav1.2 channels, in
include pimozide,64 efonidipine,65 and mibefradil,66 although SAN and AVN. Cav1.3 channels activate at voltages that partially
closer scrutiny has revealed that efonidipine and mibefradil also overlap with T-type Ca2+ channels, and no pharmacologic agent
block L-type Ca2+ channels at higher concentrations. The only is sufficiently selective to unequivocally dissect the functional
known specific blocker of T-type Ca2+ currents (generated by the contributions of these channels. Moreover, it is conceivable that
Cav1.3 or Cav3.2 homologues) is kurtoxin67 from scorpions. On T-type channels are expressed in distinct regions of the SAN and
the other hand, many putative selective L-type Ca2+ channel AVN from Cav1.3 channels, thereby complicating further our
blockers (such as dihydropyridines and verapamil) also inhibit understanding of the relative contributions of T-channels to pace-
T-channels.68 The development of more selective agents maker activity. A recent report in mice established that Cav1.3 and
will clearly help to improve our understanding of the role of Cav3.1 contribute synergistically to pacemaker activity in the
T-type Ca2+ channels in physiological and pathophysiological SAN and conduction in the AVN under baseline conditions and
processes.49 in response to β-adrenergic receptor stimulation.80 These results
also confirmed that Cav1.3-mediated ICa,L plays a relatively greater
role in pacemaker activity and AVN conduction than Cav3.1-
Expression Patterns of T-type Ca2+ Channels mediated ICa,T. Given that depolarizing current from Na+/Ca2+
exchanger secondary to Ca2+ release from the sarcoplasmic reticu-
T-type Ca2+ channels show marked developmental and regional lum contributes to diastolic depolarization92,93 (later called the
differences in expression that have been linked to the unique Ca2+ clock94), it remains to be determined to what extent the
functions of these channels. For example, T-type Ca2+ currents action of the T-type Ca2+ channel currents on pacemaker activity
are present in the fetal heart with the expression of Cav3.1 pre- relies on the activation of RYR2 channels in the SR membrane. In
dominating in mouse and Cav3.2 prevailing in rats.69,70 After this regard, conditional overexpression of Cav3.1 using the
birth, ICa,T levels decline progressively in ventricles and become α-MHC promoter suggests that T-type Ca2+ channels are
undetectable in adults while remaining at measurable levels in expressed on the surface sarcolemma in regions distinct from the
SAN (≈3 pA/pF), conducting system (≈3 pA/pF), and atria.49,71-75 L-type Ca2+ channels and remote from RYR2 channels.95
In adult hearts of humans76 and mice,75 Cav3.1 is the more preva- Many studies have concluded, based on pharmacologic inter-
lent isoform, and Cav3.2 seems to be more abundant in canine ventions, that re-expression of ICa,T plays a deleterious role in
Purkinje fibers,71 although this topic is controversial.77 It is inter- cardiac remodeling and arrhythmogenesis seen in diseased
esting to note that ICa,T appear to vary inversely with body size48,78 hearts.83,96 For example, treatment with nickel and mibefradil
in mammals, consistent with the conclusion that these channels reduced cell proliferation of neonatal cardiomyocytes induced
influence basal heart rates. Indeed, mice lacking Cav3.1 channels by hyperglycemia.97 Mibefradil also reduced infarct sizes and
showed bradycardia combined with slowed conduction through improved ventricular function in rats subjected to myocardial
the AVN, and no electrical changes were seen in Cav3.2-knockout infarction (MI)98 while protecting dog ventricles from partial
mice.79-81 A common finding in diseased ventricles is the elevated coronary artery occlusions99 and pacing-induced arrhyth-
expression of Cav3.1 and Cav3.2, along with the reappearance of mias.100,101 Both efonidipine and mibefradil were also more effec-
ICa,T,82-85 possibly resulting from fetal gene reactivation. These tive than nitrendipine in reducing sudden death in mice with
observations are particularly intriguing because elevations in ICa,T heart failure.102 T-type Ca2+ channels have been shown to initiate
are linked to hypertrophy, disease progression, and arrhyth- spontaneous activity in pulmonary veins, thereby potentially
mias.70,78 In addition, ICaT is reported to be preferentially expressed inducing paroxysmal atrial fibrillation.103 These findings are con-
in small mononucleated myocytes of adolescent feline hearts,63 sistent with the ability of efonidipine to prevent atrial remodeling
possibly representing nascent myocytes originating from cardiac in paced dogs.104 By contrast, mibefradil has been shown to cause
deleterious effects in animal models,105,106 and treatment of car- a tetrameric structure in which each subunit contributes to the
diovascular disease in humans with mibefradil has terminated, selectivity filter and channel pore.110 Most TRP channels, with
presumably as a result of off-target effects on the liver.107 The the exception of TRPA and TRPP, contain a stretch of ≈25 intra-
rationale for targeting T-type Ca2+ channels in treating heart cellular residues on the C-terminal side of the S6 transmembrane
disease has come into focus with several recent studies showing domain (TRP box), from which these channels derive their
that not only do increases in ICaT (via Cav3.1 overexpression in name.120
mice) not induce adverse cardiac remodeling,95 they also abrogate TRP channels do not contain the series of arginine residues
cardiac hypertrophy induced by pressure overload, isoproterenol, in the S4 transmembrane segment that are typical of other
and exercise.108 Moreover, Cav3.1-deficient mice show enhanced voltage-gated channels, and the location and/or structure of spe-
adverse remodeling in these models and also show accelerated cific gate(s) that control the passage of ions through the channel
functional deterioration and enhanced arrhythmias following pore are not well resolved.110 Nevertheless, TRP channel activity
MI, in contrast to mice lacking Cav3.2, which responded simi- is often voltage dependent and characterized by conductance-
larly to wild type mice. This consensus, however, is at odds voltage relationships with relatively shallow slope factors. Fur-
with a study showing that Cav3.2 ablation, not Cav3.1 deletion, thermore, many TRP channels exhibit the voltage-dependent
is protective against pressure overload and angiotensin-induced relaxation seen in tail currents following depolarizing voltage
hypertrophy.60 steps.122 TRP channel voltage sensitivity is modulated by a
number of factors, including ligand binding, temperature, osmo-
larity, and mechanical perturbations.
Transient Receptor Potential (TRP) Channels

TRPC Channels and Cardiac Hypertrophy
The TRP channels are a large group of ion channels that were
first described in Drosophila based on mutations in trp genes in Increases in calcium influx have been implicated in pathological
photoreceptors that resulted in impaired transient responses to hypertrophic signaling in the heart due at least in part to the
light.109 Currently, there are 28 mammalian trp genes, which are activation of calcineurin-nuclear factor of activated T cells
divided into six subfamilies denoted as TRPC (canonical), TRPV (NFAT) signaling.123 Although LTCCs have been shown to
(vanilloid), TRPM (melastatin), TRPA (ankyrin), TRPP (poly- contribute to this increase in Ca2+ influx,4,124,125 a number of
cystin), and TRPML (mucolipin).110 All of these channels are studies have demonstrated that TRPC channel expression and
cationic in nature; however, there is a high level of diversity in activity are also upregulated in cardiac hypertrophy and
terms of cation selectivity and gating mechanisms. Most TRP heart failure. For example, pressure overload in rodents results
channels are nonselective cation channels that are permeable to in upregulation of TRPC1 and TRPC3.126,127 Similarly, TRPC6
Na+ and Ca2+ ions (PCa/PNa = 1-10), but this does not apply in all was upregulated in cardiac hypertrophy and in heart failure.128
cases. For example, TRPM3α2, TRPV5, and TRPV6 are Ca2+ TRPC5 has been shown to be elevated in human heart failure
selective channels (PCa/PNa >100), while TRPM3α 1, TRPM4, patients.126 Consistent with these findings, pro-hypertrophic
and TRPM5 are not Ca2+ permeable (PCa/PNa < 0.05). TRPM6 agents such as endothelin, phenylephrine, and angiotensin II
and TRPM7 are permeable to Mg2+ and Ca2+.110,111 cause the upregulation of TRPC1 and TRPC3 in cultured
A number of TRP channels have been detected in the cardiomyocytes.129,130
heart, at least at the mRNA level, including TRPC1, TRPC3-7, Recent studies suggest that Ca2+ influx through TRPC chan-
TRPV2, TRPV5, TRPV5, TRPM4, TRPM5, and TRPM7.111-113 nels initiates signaling pathways associated with hypertrophy and
Roles for some of these specific TRP channels in the heart have pathological cardiac remodeling.114 For example, activation of
been identified in some instances. For example, TRPC channels TRPC channels leads to activation of fetal genes (including atrial
have been implicated in cardiac hypertrophy,114 and both natriuretic peptide and skeletal α-actin), cell enlargement, and
TRPC3115 and TRPM7116 have been shown to play a role in apoptosis. These responses have been shown to be dependent on
cardiac fibroblasts in atrial fibrillation. TRPM7 channels have enhanced calcineurin-NFAT signaling. Consistent with a role for
also been described in ventricular myocytes,117,118 while TRPC TRPC channels in hypertrophic signaling, the TRPC3 blocker
channels have been suggested to play a role in pacemaker activity Pyr3 (a Bis-trifluoromethyl pyrazole compound) has been shown
in the SAN.119 In most cases, however, there is still much to be to prevent cardiac hypertrophy in pressure-overloaded mice,131
learned about TRP channels in the heart in normal and patho- and TRPC1 knockout mice are less susceptible to cardiac hyper-
physiological conditions. trophy and heart failure.127 Similar protective responses have
A complete discussion of TRP channels in the heart is beyond been demonstrated in TRPC3-, TRPC4-, and TRPC6-dominant/
the scope of this chapter. Thus, we will provide an overview negative mutant mice.132 Together, these findings suggest an
of the general structure of TRP channels and a brief discussion important role for Ca2+ entry through TRP channels in hyper-
of the biophysical properties of key members of the TRP family trophy and heart failure.
that have been implicated in cardiovascular disease. The reader
is directed to several comprehensive reviews of the biology of
TRP channels for additional information.110,111,113,120,121 TRP Channels in Cardiac Fibroblasts
Cardiac fibroblasts account for ≈80% of the nonmyocyte popula-
General Structure and Biophysical Properties tion in the heart and are responsible for the synthesis and secre-
of TRP Channels tion of extracellular matrix proteins, including collagen types I
and II, as well as matrix metalloproteinases.133,134 In the setting of
TRP channels are predicted to have a similar membrane topology pathological conditions such as hypertension, cardiac hypertro-
to other voltage-gated (Na+, Ca2+, K+) ion channels in the heart phy, and heart failure, there is an increase in the proliferation of
in which the channel is formed by six transmembrane-spanning cardiac fibroblasts, which differentiate into myofibroblasts that
domains (S1-S6) with a pore between S5 and S6 (Figure 2-5, A). contribute to the deposition of extracellular matrix and cardiac
The N and C termini of these proteins are intracellular and bind fibrosis.133,134 This results in conduction abnormalities, which
a number of regulatory proteins. Similar to the other six trans- may be due to both structural remodeling and possible fibroblast-
membrane domain channels, TRP channels are thought to form myocyte electrical interactions.135
Ca2+ or Na–
La3– 2–APB
Extracellular
S1 S2 S3 S4 S5 S6
Intracellular
N
A C
pA/pF
10 Low Cr (2) 20 pA/pf 10 pA/pf
Ca+ Mg+ free(2)
5 10 5
Low Na+ (2)
Control (1) Control (1)
–100 –50 0 50 100 mV –100 –50 0 50 100 mV –100 –50 0 50 100 mV
–5 –10 –5
Control (1)
–10 –20 –10
10 20 Low Cr 10
Ca2– /Mg2– free
Minimum outward
8
current (pA/pF)
15 1 2 2
Low Cr
6 1 2 10 5 1
4
2 5
0 0 0
0 100 200 300 400 500 0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700
0 0 0
2
Minimum outward
current (pA/pF)
–2 –5 1
–4 1
1 2
–10 –5 2
–6
–8 –15
B –10 –20 –10
pA
6
+80 mV
150 +40 mV
0 mV
3
pA
50
–100 100 mV SC = 38.4 +

– 1.6 pS
C –50 0 60 120 mV
Figure 2-5. Transient receptor potential (TRP) channels in the heart. A, General membrane topology of TRP channels. S1-S6 are the transmembrane segments, and the pore is
located between S5 and S6. Most TRP channels are nonselective and permeable to monovalent and divalent cations. Many TRP channels are blocked by lanthanum (La3+) and
2-aminodiethylphenhyl borate (2-APB). B, Properties of TRPM7-dependent current in human atrial fibroblasts. The I-V relationship (left) elicited by a voltage ramp protocol illus-
trates that TRPM7 currents strongly outwardly rectify and inward currents are very small in normal conditions. Single-channel conductance (right) of TRPM7-mediated current
was 38.4 ± 1.6 pS. C, Properties of TRPC-dependent nonselective cation currents in rat cardiac fibroblasts. Current-voltage (I-V) relations elicited by a voltage ramp protocol (top)
and time course of maximum inward and outward currents (bottom) are illustrated. The current is not affected by Cl− replacement with CH3SO3−. Replacing extracellular Na+ with
NMDG+ selectively reduces the inward current, while removal of divalent cations linearizes the I-V relationship and increases inward and outward currents.
(A, Clapham DE, Runnels LW, Strubig C: The TRP ion channel family. Nat Rev Neurosci 2:387-396, 2001. B, Rose RA, Hatano N, Ohya S, et al: C-type natriuretic peptide activates
a non-selective cation current in acutely isolated rat cardiac fibroblasts via natriuretic peptide C receptor-mediated signalling. J Physiol 580:255-274, 2007. C, Du J, Xie J, Zhang
Z, et al: TRPM7-mediated Ca2+ signals confer fibrogenesis in human atrial fibrillation. Circ Res 106:992-1003, 2010.)
Several TRP channels are expressed in cardiac fibroblasts. In fibrosis in atrial fibrillation. The pathophysiological role of TRP
rat ventricular fibroblasts, mRNA for TRPC2, TRPC3, TRPV2, channels in cardiac arrhythmias is an emerging area of interest,
TRPV6, TRPM4, and TRPM7 has been detected,112 while and the aforementioned studies strongly indicate the need to
human atrial fibroblasts were shown to express mRNA for better understand the relationships between these ion channels
TRPC1, TRPC6, TRPC2, TRPC4, and TRPM7.116 These TRP and conduction disturbances in the heart.136
channels facilitate the influx of Ca2+ into fibroblasts and may serve In conclusion, several ion channels in the heart are responsible
as a primary source of Ca2+ for these nonexcitable cells, which do for mediating Ca2+ influx into myocytes and fibroblasts. CaV1.2-
not typically express voltage-gated Ca2+ channels.112 mediated ICa,L represents a major route for Ca2+ entry in all car-
Recent data indicate that Ca2+ influx into fibroblasts through diomyocytes throughout the heart, while CaV1.3-mediated ICa,L
TRP channels is a significant contributor to fibroblast prolifera- is restricted to the SAN, the cardiac conduction system, and the
tion and fibrosis in atrial fibrillation. Du et al116 demonstrated atrial myocardium. T-type Ca2+ channels primarily contribute to
that TRPM7 (Figure 2-5, B) is a major pathway for Ca2+ entry Ca2+ influx in the specialized pacemaker cells of the SAN and
in human atrial fibroblasts, and that this Ca2+ activates transform- AVN in normal physiology, but may also be expressed in ven-
ing growth factor (TGF)-β1, a critical signaling molecule in atrial tricular myocytes in the diseased heart. Recently, a number
fibrosis. TRPC channels also constitute a major pathway for Ca2+ of TRP channels have emerged as important contributors to
entry into cardiac fibroblasts via nonselective cation currents112 Ca2+ influx in myocytes and fibroblasts. Each of the ion channels
(Figure 2-5, C), and a recent study115 suggests that Ca2+ entry in contributes importantly to the transport of Ca2+ into cells in the
atrial fibroblasts via TRPC3 channels controls fibroblast prolif- heart in normal physiology and in the setting of cardiovascular
eration and differentiation in an extracellular signal–related disease. These channels are importantly regulated by a variety of
kinase (ERK)-dependent fashion in atrial fibrillation. This study neurohumoral signaling molecules, and interest remains high in
further demonstrates a role for microRNA-26 and NFAT in the how these channels are or may be targeted for therapeutic
TRPC3-dependent enhancement of fibroblast proliferation and interventions.
of the alpha 1-subunit. FEBS Lett 352:141–145, nodes: insight gained using gene-targeted null
References 1994. mutant mice. Circ Res 90:981–987, 2002.
15. Davies A, Hendrich J, Van Minh AT, et al: Func- 28. Platzer J, Engel J, Schrott-Fischer A, et al: Con-
1. Bers DM: Cardiac excitation-contraction cou- tional biology of the alpha(2)delta subunits of genital deafness and sinoatrial node dysfunction
pling. Nature 415:198–205, 2002. voltage-gated calcium channels. Trends Pharma- in mice lacking class D L-type Ca2+ channels. Cell
2. Best JM, Kamp TJ: Different subcellular popula- col Sci 28:220–228, 2007. 102:89–97, 2000.
tions of L-type Ca2+ channels exhibit unique 16. Jay SD, Sharp AH, Kahl SD, et al: Structural 29. Mancarella S, Yue Y, Karnabi E, et al: Impaired
regulation and functional roles in cardiomyocytes. characterization of the dihydropyridine-sensitive Ca2+ homeostasis is associated with atrial fibrilla-
J Mol Cell Cardiol 52:376–387, 2012. calcium channel alpha 2-subunit and the associ- tion in the alpha1D L-type Ca2+ channel KO
3. Benitah JP, Alvarez JL, Gomez AM: L-type ated delta peptides. J Biol Chem 266:3287–3293, mouse. Am J Physiol Heart Circ Physiol
Ca(2+) current in ventricular cardiomyocytes. J 1991. 295:H2017–H2024, 2008.
Mol Cell Cardiol 48:26–36, 2010. 17. Yang L, Katchman A, Morrow JP, et al: Cardiac 30. Campbell DL, Giles WR, Hume JR, et al: Inac-
4. Pitt GS, Dun W, Boyden PA: Remodeled cardiac L-type calcium channel (Cav1.2) associates with tivation of calcium current in bull-frog atrial myo-
calcium channels. J Mol Cell Cardiol 41:373–388, gamma subunits. FASEB J 25:928–936, 2011. cytes. J Physiol 403:287–315, 1988.
2006. 18. McCleskey EW, Almers W: The Ca channel in 31. Kass RS, Sanguinetti MC: Inactivation of calcium
5. Catterall WA: Voltage-gated calcium channels. skeletal muscle is a large pore. Proc Natl Acad Sci channel current in the calf cardiac Purkinje fiber:
Cold Spring Harb Perspect Biol 3:a003947, U S A 82:7149–7153, 1985. Evidence for voltage- and calcium-mediated
2011. 19. Chen XH, Bezprozvanny I, Tsien RW: Molecular mechanisms. J Gen Physiol 84:705–726, 1984.
6. Catterall WA: Structure and regulation of basis of proton block of L-type Ca2+ channels. J 32. Lee KS, Marban E, Tsien RW: Inactivation of
voltage-gated Ca2+ channels. Annu Rev Cell Dev Gen Physiol 108:363–374, 1996. calcium channels in mammalian heart cells: Joint
Biol 16:521–555, 2000. 20. Chen XH, Tsien RW: Aspartate substitutions dependence on membrane potential and intracel-
7. Schram G, Pourrier M, Melnyk P, et al: Differen- establish the concerted action of P-region gluta- lular calcium. J Physiol 364:395–411, 1985.
tial distribution of cardiac ion channel expression mates in repeats I and III in forming the proton- 33. Hadley RW, Hume JR: An intrinsic potential-
as a basis for regional specialization in electrical ation site of L-type Ca2+ channels. J Biol Chem dependent inactivation mechanism associated
function. Circ Res 90:939–950, 2002. 272:30002–30008, 1997. with calcium channels in guinea-pig myocytes. J
8. Bock G, Gebhart M, Scharinger A, et al: Func- 21. Guia A, Stern MD, Lakatta EG, et al: Ion Physiol 389:205–222, 1987.
tional properties of a newly identified C-terminal concentration-dependence of rat cardiac unitary 34. Richard S, Perrier E, Fauconnier J, et al: ‘Ca(2+)-
splice variant of Cav1.3 L-type Ca2+ channels. J L-type calcium channel conductance. Biophys J induced Ca(2+) entry’ or how the L-type Ca(2+)
Biol Chem 286:42736–42748, 2011. 80:2742–2750, 2001. channel remodels its own signalling pathway in
9. Tan BZ, Jiang F, Tan MY, et al: Functional char- 22. Mangoni ME, Couette B, Marger L, et al: cardiac cells. Prog Biophys Mol Biol 90:118–135,
acterization of alternative splicing in the C termi- Voltage-dependent calcium channels and cardiac 2006.
nus of L-type CaV1.3 channels. J Biol Chem pacemaker activity: from ionic currents to genes. 35. Peterson BZ, DeMaria CD, Adelman JP, et al:
286:42725–42735, 2011. Prog Biophys Mol Biol 90:38–63, 2006. Calmodulin is the Ca2+ sensor for Ca2+ -dependent
10. Foell JD, Balijepalli RC, Delisle BP, et al: Molec- 23. Mangoni ME, Couette B, Bourinet E, et al: Func- inactivation of L-type calcium channels. Neuron
ular heterogeneity of calcium channel beta- tional role of L-type Cav1.3 Ca2+ channels in 22:549–558, 1999.
subunits in canine and human heart: evidence for cardiac pacemaker activity. Proc Natl Acad Sci 36. Qin N, Olcese R, Bransby M, et al: Ca2+-induced
differential subcellular localization. Physiol U S A 100:5543–5548, 2003. inhibition of the cardiac Ca2+ channel depends on
Genomics 17:183–200, 2004. 24. Zhang Z, He Y, Tuteja D, et al: Functional roles calmodulin. Proc Natl Acad Sci U S A 96:2435–
11. Hofmann F, Biel M, Flockerzi V: Molecular basis of Cav1.3(alpha1D) calcium channels in atria: 2438, 1999.
for Ca2+ channel diversity. Annu Rev Neurosci insights gained from gene-targeted null mutant 37. Bers DM, Perez-Reyes E: Ca channels in cardiac
17:399–418, 1994. mice. Circulation 112:1936–1944, 2005. myocytes: Structure and function in Ca influx and
12. Chen YH, Li MH, Zhang Y, et al: Structural basis 25. Seisenberger C, Specht V, Welling A, et al: Func- intracellular Ca release. Cardiovasc Res 42:339–
of the alpha1-beta subunit interaction of voltage- tional embryonic cardiomyocytes after disruption 360, 1999.
gated Ca2+ channels. Nature 429:675–680, 2004. of the L-type alpha1C (Cav1.2) calcium channel 38. January CT, Riddle JM: Early afterdepolariza-
13. Van Petegem F, Clark KA, Chatelain FC, et al: gene in the mouse. J Biol Chem 275:39193– tions: mechanism of induction and block: A role for
Structure of a complex between a voltage-gated 39199, 2000. L-type Ca2+ current. Circ Res 64:977–990, 1989.
calcium channel beta-subunit and an alpha- 26. Mangoni ME, Nargeot J: Genesis and regulation 39. Hryshko LV, Bers DM: Ca current facilitation
subunit domain. Nature 429:671–675, 2004. of the heart automaticity. Physiol Rev 88:919– during postrest recovery depends on Ca entry. Am
14. Mitterdorfer J, Froschmayr M, Grabner M, et al: 982, 2008. J Physiol 259:H951–H961, 1990.
Calcium channels: The beta-subunit increases the 27. Zhang Z, Xu Y, Song H, et al: Functional roles of 40. Zuhlke RD, Pitt GS, Deisseroth K, et al: Calmod-
affinity of dihydropyridine and Ca2+ binding sites Ca(v)1.3 (alpha(1D) ) calcium channel in sinoatrial ulin supports both inactivation and facilitation of
L-type calcium channels. Nature 399:159–162, 61. Iftinca MC, Zamponi GW: Regulation of neuro- 82. Huang B, Qin D, Deng L, et al: Reexpression of
1999. nal T-type calcium channels. Trends Pharmacol T-type Ca2+ channel gene and current in post-
41. Scott JD, Santana LF: A-kinase anchoring pro-
teins: Getting to the heart of the matter. Circula-
tion 121:1264–1271, 2010.
Sci 30:32–40, 2009.
62. Hagiwara N, Irisawa H, Kameyama M: Contribu-
tion of two types of calcium currents to the pace-
infarction remodeled rat left ventricle. Cardiovasc
Res 46:442–449, 2000.
83. Nuss HB, Houser SR: T-type Ca2+ current is
2
42. Beca S, Aschars-Sobbi R, Panama BK, et al: Regu- maker potentials of rabbit sino-atrial node cells. J expressed in hypertrophied adult feline left ven-
lation of murine cardiac function by phosphodi- Physiol 395:233–253, 1988. tricular myocytes. Circ Res 73:777–782, 1993.
esterases type 3 and 4. Curr Opin Pharmacol 63. Perez-Reyes E: Molecular characterization of 84. Schmitt R, Clozel JP, Iberg N, et al: Mibefradil
11:714–719, 2011. T-type calcium channels. Cell Calcium 40:89–96, prevents neointima formation after vascular injury
43. Santana LF, Navedo MF: Natural inequalities: 2006. in rats: Possible role of the blockade of the T-type
Why some L-type Ca2+ channels work harder 64. Xie X, Van Deusen AL, Vitko I, et al: Validation voltage-operated calcium channel. Arterioscler
than others. J Gen Physiol 136:143–147, 2010. of high throughput screening assays against three Thromb Vasc Biol 15:1161–1165, 1995.
44. Fuller MD, Emrick MA, Sadilek M, et al: Molec- subtypes of Ca(v)3 T-type channels using molecu- 85. Sen L, Smith TW: T-type Ca2+ channels
ular mechanism of calcium channel regulation in lar and pharmacologic approaches. Assay Drug are abnormal in genetically determined cardio-
the fight-or-flight response. Sci Signal 3:ra70, Dev Technol 5:191–203, 2007. myopathic hamster hearts. Circ Res 75:149–155,
2010. 65. Sawada T, Shinke T, Shite J, et al: Impact of cyto- 1994.
45. Hulme JT, Lin TW, Westenbroek RE, et al: Beta- chrome P450 2C19*2 polymorphism on intra- 86. Yanagi K, Takano M, Narazaki G, et al:
adrenergic regulation requires direct anchoring of stent thrombus after drug-eluting stent Hyperpolarization-activated cyclic nucleotide-
PKA to cardiac CaV1.2 channels via a leucine implantation in Japanese patients receiving clopi- gated channels and T-type calcium channels
zipper interaction with A kinase-anchoring dogrel. Circ J 75:99–105, 2011. confer automaticity of embryonic stem cell-
protein 15. Proc Natl Acad Sci U S A 100:13093– 66. Tanaka H, Shigenobu K: Pathophysiological sig- derived cardiomyocytes. Stem Cells 25:2712–
13098, 2003. nificance of T-type Ca2+ channels: T-type Ca2+ 2719, 2007.
46. Balijepalli RC, Foell JD, Hall DD, et al: Localiza- channels and drug development. J Pharmacol Sci 87. Wiese C, Grieskamp T, Airik R, et al: Formation
tion of cardiac L-type Ca(2+) channels to a caveo- 99:214–220, 2005. of the sinus node head and differentiation of sinus
lar macromolecular signaling complex is required 67. Chuang RS, Jaffe H, Cribbs L, et al: Inhibition of node myocardium are independently regulated by
for beta(2)-adrenergic regulation. Proc Natl Acad T-type voltage-gated calcium channels by a new Tbx18 and Tbx3. Circ Res 104:388–397, 2009.
Sci U S A 103:7500–7505, 2006. scorpion toxin. Nat Neurosci 1:668–674, 1998. 88. Hoogaars WM, Engel A, Brons JF, et al: Tbx3
47. Navedo MF, Cheng EP, Yuan C, et al: Increased 68. Lacinova L: T-type calcium channel blockers— controls the sinoatrial node gene program and
coupled gating of L-type Ca2+ channels during new and notable. Gen Physiol Biophys 30:403– imposes pacemaker function on the atria. Genes
hypertension and Timothy syndrome. Circ Res 409, 2011. Dev 21:1098–1112, 2007.
106:748–756, 2010. 69. Cribbs LL, Martin BL, Schroder EA, et al: Iden- 89. Nilius B, Talavera K, Verkhratsky A: T-type
48. Ono K, Iijima T: Pathophysiological significance tification of the t-type calcium channel (Ca(v)3.1d) calcium channels: The never ending story. Cell
of T-type Ca2+ channels: Properties and func- in developing mouse heart. Circ Res 88:403–407, Calcium 40:81–88, 2006.
tional roles of T-type Ca2+ channels in cardiac 2001. 90. Madle A, Linhartova K, Koza J: Effects of the
pacemaking. J Pharmacol Sci 99:197–204, 70. Cribbs L: T-type calcium channel expression and T-type calcium channel blockade with oral mibe-
2005. function in the diseased heart. Channels (Austin) fradil on the electrophysiologic properties of the
49. Vassort G, Talavera K, Alvarez JL: Role of T-type 4:447–452, 2010. human heart. Med Sci Monit 7:74–77, 2001.
Ca2+ channels in the heart. Cell Calcium 40:205– 71. Rosati B, Dun W, Hirose M, et al: Molecular basis 91. Chen CC, Lamping KG, Nuno DW, et al: Abnor-
220, 2006. of the T- and L-type Ca2+ currents in canine Pur- mal coronary function in mice deficient in
50. Perez-Reyes E: Molecular physiology of low- kinje fibres. J Physiol 579:465–471, 2007. alpha1H T-type Ca2+ channels. Science 302:1416–
voltage-activated t-type calcium channels. Physiol 72. Zhou C, Chen H, Lu F, et al: Cav3.1 (alpha1G) 1418, 2003.
Rev 83:117–161, 2003. controls von Willebrand factor secretion in rat 92. Huser J, Blatter LA, Lipsius SL: Intracellular Ca2+
51. Lambert RC, Maulet Y, Mouton J, et al: T-type pulmonary microvascular endothelial cells. Am J release contributes to automaticity in cat atrial
Ca2+ current properties are not modified by Ca2+ Physiol Lung Cell Mol Physiol 292:L833–L844, pacemaker cells. J Physiol 524(Pt 2):415–422,
channel beta subunit depletion in nodosus gan- 2007. 2000.
glion neurons. J Neurosci 17:6621–6628, 1997. 73. Leuranguer V, Monteil A, Bourinet E, et al: T- 93. Hata T, Noda T, Nishimura M, et al: The role of
52. Leuranguer V, Bourinet E, Lory P, et al: Antisense type calcium currents in rat cardiomyocytes Ca2+ release from sarcoplasmic reticulum in the
depletion of beta-subunits fails to affect T-type during postnatal development: contribution to regulation of sinoatrial node automaticity. Heart
calcium channels properties in a neuroblastoma hormone secretion. Am J Physiol Heart Circ Vessels 11:234–241, 1996.
cell line. Neuropharmacology 37:701–708, Physiol 279:H2540–H2548, 2000. 94. Lakatta EG, Maltsev VA, Vinogradova TM: A
1998. 74. Mizuta E, Miake J, Yano S, et al: Subtype switch- coupled system of intracellular Ca2+ clocks and
53. Dolphin AC, Wyatt CN, Richards J, et al: The ing of T-type Ca2+ channels from Cav3.2 to surface membrane voltage clocks controls the
effect of alpha2-delta and other accessory subunits Cav3.1 during differentiation of embryonic stem timekeeping mechanism of the heart’s pacemaker.
on expression and properties of the calcium cells to cardiac cell lineage. Circ J 69:1284–1289, Circ Res 106:659–673, 2010.
channel alpha1G. J Physiol 519(Pt 1):35–45, 2005. 95. Jaleel N, Nakayama H, Chen X, et al: Ca2+ influx
1999. 75. Niwa N, Yasui K, Opthof T, et al: Cav3.2 subunit through T- and L-type Ca2+ channels have differ-
54. Green PJ, Warre R, Hayes PD, et al: Kinetic underlies the functional T-type Ca2+ channel in ent effects on myocyte contractility and induce
modification of the alpha(1I) subunit-mediated murine hearts during the embryonic period. Am unique cardiac phenotypes. Circ Res 103:1109–
T-type Ca(2+) channel by a human neuronal J Physiol Heart Circ Physiol 286:H2257–H2263, 1119, 2008.
Ca(2+) channel gamma subunit. J Physiol 2004. 96. Ertel SI, Ertel EA, Clozel JP: T-type Ca2+ chan-
533:467–478, 2001. 76. Chandler NJ, Greener ID, Tellez JO, et al: nels and pharmacological blockade: Potential
55. Klugbauer N, Dai S, Specht V, et al: A family of Molecular architecture of the human sinus node: pathophysiological relevance. Cardiovasc Drugs
gamma-like calcium channel subunits. FEBS Lett Insights into the function of the cardiac pace- Ther 111:723–739, 1997.
470:189–197, 2000. maker. Circulation 119:1562–1575, 2009. 97. Li M, Zhang M, Huang L, et al: T-type Ca2+
56. Gao B, Sekido Y, Maximov A, et al: Functional 77. Han W, Bao W, Wang Z, et al: Comparison of channels are involved in high glucose-induced rat
properties of a new voltage-dependent calcium ion-channel subunit expression in canine cardiac neonatal cardiomyocyte proliferation. Pediatr Res
channel alpha(2)delta auxiliary subunit gene Purkinje fibers and ventricular muscle. Circ Res 57:550–556, 2005.
(CACNA2D2). J Biol Chem 275:12237–12242, 91:790–797, 2002. 98. Sandmann S, Min JY, Meissner A, et al: Effects of
2000. 78. Ono K, Iijima T: Cardiac T-type Ca(2+) channels the calcium channel antagonist mibefradil on hae-
57. Aromolaran KA, Benzow KA, Cribbs LL, et al: in the heart. J Mol Cell Cardiol 48:65–70, 2010. modynamic parameters and myocardial Ca(2+)-
T-type current modulation by the actin-binding 79. Mangoni ME, Traboulsie A, Leoni AL, et al: Bra- handling in infarct-induced heart failure in
protein Kelch-like 1. Am J Physiol Cell Physiol dycardia and slowing of the atrioventricular rats. Cardiovasc Res 44:67–80, 1999.
298:C1353–1362, 2010. conduction in mice lacking CaV3.1/alpha1G 99. Parent de Curzon O, Ghaleh B, Hittinger L, et
58. Markandeya YS, Fahey JM, Pluteanu F, et al: T-type calcium channels. Circ Res 98:1422–1430, al: Beneficial effects of the T- and L-type calcium
Caveolin-3 regulates protein kinase A modulation 2006. channel antagonist, mibefradil, against exercise-
of the Ca(V)3.2 (alpha1H) T-type Ca2+ channels. 80. Marger L, Mesirca P, Alig J, et al: Functional roles induced myocardial stunning in dogs. J Cardio-
J Biol Chem 286:2433–2444, 2011. of Ca(v)1.3, Ca(v)3.1 and HCN channels in auto- vasc Pharmacol 35:240–248, 2000.
59. Lee JH, Gomora JC, Cribbs LL, et al: Nickel maticity of mouse atrioventricular cells: Insights 100. Fareh S, Benardeau A, Nattel S: Differential
block of three cloned T-type calcium channels: into the atrioventricular pacemaker mechanism. efficacy of L- and T-type calcium channel block-
Low concentrations selectively block alpha1H. Channels (Austin) 5:251–261, 2011. ers in preventing tachycardia-induced atrial
Biophys J 77:3034–3042, 1999. 81. Le Quang K, Naud P, Qi XY, et al: Role of T-type remodeling in dogs. Cardiovasc Res 49:762–770,
60. Chiang CS, Huang CH, Chieng H, et al: The calcium channel subunits in post-myocardial 2001.
Ca(v)3.2 T-type Ca(2+) channel is required for infarction remodelling probed with genetically 101. Fareh S, Benardeau A, Thibault B, et al: The
pressure overload-induced cardiac hypertrophy in engineered mice. Cardiovasc Res 91:420–428, T-type Ca(2+) channel blocker mibefradil
mice. Circ Res 104:522–530, 2009. 2011. prevents the development of a substrate for
atrial fibrillation by tachycardia-induced atrial via natriuretic peptide C receptor-mediated sig- in K(v)4.2N transgenic mice. Circulation 105:
remodeling in dogs. Circulation 100:2191–2197, nalling. J Physiol 580:255–274, 2007. 1850–1856, 2002.
1999. 113. Watanabe H, Murakami M, Ohba T, et al: TRP 126. Bush EW, Hood DB, Papst PJ, et al: Canonical
102. Kinoshita H, Kuwahara K, Takano M, et al: T- channel and cardiovascular disease. Pharmacol transient receptor potential channels promote
type Ca2+ channel blockade prevents sudden death Ther 118:337–351, 2008. cardiomyocyte hypertrophy through activation of
in mice with heart failure. Circulation 120:743– 114. Eder P, Molkentin JD: TRPC channels as effec- calcineurin signaling. J Biol Chem 281:33487–
752, 2009. tors of cardiac hypertrophy. Circ Res 108:265– 33496, 2006.
103. Chen YC, Chen SA, Chen YJ, et al: T-type 272, 2011. 127. Seth M, Zhang ZS, Mao L, et al: TRPC1 chan-
calcium current in electrical activity of cardio- 115. Harada M, Luo X, Qi XY, et al: Transient recep- nels are critical for hypertrophic signaling in the
myocytes isolated from rabbit pulmonary vein. J tor potential canonical-3 channel-dependent heart. Circ Res 105:1023–1030, 2009.
Cardiovasc Electrophysiol 15:567–571, 2004. fibroblast regulation in atrial fibrillation. Circula- 128. Niizeki T, Takeishi Y, Kitahara T, et al: Diacyl
104. Ohashi N, Mitamura H, Tanimoto K, et al: A tion 126:2051–2064, 2012. glycerol kinase-epsilon restores cardiac dysfunc-
comparison between calcium channel blocking 116. Du J, Xie J, Zhang Z, et al: TRPM7-mediated tion under chronic pressure overload: A new
drugs with different potencies for T- and L-type Ca2+ signals confer fibrogenesis in human atrial specific regulator of Galpha(q) signaling cascade.
channels in preventing atrial electrical remodel- fibrillation. Circ Res 106:992–1003, 2010. Am J Physiol Heart Circ Physiol 295:H245–
ing. J Cardiovasc Pharmacol 44:386–392, 117. Gwanyanya A, Amuzescu B, Zakharov SI, et al: H255, 2008.
2004. Magnesium-inhibited, TRPM6/7-like channel in 129. Ohba T, Watanabe H, Murakami M, et al: Upreg-
105. Xia QG, Reinecke A, Dorenkamp M, et al: Com- cardiac myocytes: Permeation of divalent cations ulation of TRPC1 in the development of cardiac
parison of cardioprotective effects of mibefradil and pH-mediated regulation. J Physiol 559:761– hypertrophy. J Mol Cell Cardiol 42:498–507,
and ramipril in stroke-prone spontaneously 776, 2004. 2007.
hypertensive rats. Acta Pharmacol Sin 25:763– 118. Gwanyanya A, Sipido KR, Vereecke J, et al: ATP 130. Brenner JS, Dolmetsch RE: TrpC3 regulates
768, 2004. and PIP2 dependence of the magnesium- hypertrophy-associated gene expression without
106. Takebayashi S, Li Y, Kaku T, et al: Remodeling inhibited, TRPM7-like cation channel in cardiac affecting myocyte beating or cell size. PLoS One
excitation-contraction coupling of hypertrophied myocytes. Am J Physiol Cell Physiol 291:C627– 2:e802, 2007.
ventricular myocytes is dependent on T-type C635, 2006. 131. Kiyonaka S, Kato K, Nishida M, et al: Selective
calcium channels expression. Biochem Biophys 119. Ju YK, Chu Y, Chaulet H, et al: Store-operated and direct inhibition of TRPC3 channels under-
Res Commun 345:766–773, 2006. Ca2+ influx and expression of TRPC genes in lies biological activities of a pyrazole compound.
107. Po AL, Zhang WY: What lessons can be learnt mouse sinoatrial node. Circ Res 100:1605–1614, Proc Natl Acad Sci U S A 106:5400–5405, 2009.
from withdrawal of mibefradil from the market? 2007. 132. Wu X, Eder P, Chang B, et al: TRPC channels
Lancet 351:1829–1830, 1998. 120. Clapham DE: TRP channels as cellular sensors. are necessary mediators of pathologic cardiac
108. Nakayama H, Bodi I, Correll RN, et al: alpha1G- Nature 426:517–524, 2003. hypertrophy. Proc Natl Acad Sci U S A 107:7000–
dependent T-type Ca2+ current antagonizes 121. Clapham DE, Runnels LW, Strubing C: The 7005, 2010.
cardiac hypertrophy through a NOS3-dependent TRP ion channel family. Nat Rev Neurosci 133. Baudino TA, Carver W, Giles W, et al: Cardiac
mechanism in mice. J Clin Invest 119:3787–3796, 2:387–396, 2001. fibroblasts: Friend or foe? Am J Physiol Heart
2009. 122. Nilius B, Talavera K, Owsianik G, et al: Gating Circ Physiol 291:H1015–H1026, 2006.
109. Minke B: Drosophila mutant with a transducer of TRP channels: A voltage connection? J Physiol 134. Souders CA, Bowers SL, Baudino TA: Cardiac
defect. Biophys Struct Mech 3:59–64, 1977. 567:35–44, 2005. fibroblast: The renaissance cell. Circ Res
110. Ramsey IS, Delling M, Clapham DE: An intro- 123. Heineke J, Molkentin JD: Regulation of cardiac 105:1164–1176, 2009.
duction to TRP channels. Ann Rev Physiol hypertrophy by intracellular signalling pathways. 135. Yue L, Xie J, Nattel S: Molecular determinants of
68:619–647, 2006. Nat Rev Mol Cell Biol 7:589–600, 2006. cardiac fibroblast electrical function and thera-
111. Vennekens R: Emerging concepts for the role of 124. Richard S, Leclercq F, Lemaire S, et al: Ca2+ cur- peutic implications for atrial fibrillation. Cardio-
TRP channels in the cardiovascular system. J rents in compensated hypertrophy and heart vasc Res 89:744–753, 2011.
Physiol 589:1527–1534, 2011. failure. Cardiovasc Res 37:300–311, 1998. 136. Rose RA, Belke DD, Maleckar MM, et al: Ca2+
112. Rose RA, Hatano N, Ohya S, et al: C-type natri- 125. Sah R, Oudit GY, Nguyen TT, et al: Inhibition entry through TRP-C channels regulates fibro-
uretic peptide activates a non-selective cation of calcineurin and sarcolemmal Ca2+ influx pro- blast biology in chronic atrial fibrillation. Circula-
current in acutely isolated rat cardiac fibroblasts tects cardiac morphology and ventricular function tion 126:2039–2041, 2012.
Voltage-Regulated
Potassium Channels 3
Jeanne M. Nerbonne
lack functional Nav channels, the upstroke of the action potential

CHAPTER OUTLINE
is substantially slower and is dominated by Ca2+ influx through
Myocardial Voltage-Gated K+ Channels: voltage-gated Ca2+ (Cav) channels. There are also marked
Transient Outward Voltage-Gated K+ Channels 23 regional differences in action potential heights and durations, as
well as in the time courses of action potential repolarization (see
Myocardial Voltage-Gated K+ Channels: Figure 3-1). These differences, which affect the normal spread of
Delayed Rectifier Voltage-Gated K+ Channels 25 excitation in the myocardium and influence the dispersion of
+
Inwardly Rectifying Myocardial K Channels Also repolarization in the ventricles, primarily reflect regional differ-
Contribute to Repolarization 26 ences in the functional expression and the properties of the
outward K+, as well as the inward (Nav and Cav) currents.1
Pore-Forming (α) Subunits of Myocardial Cellular electrophysiologic studies have detailed the proper-
Voltage-Gated K+ Channels 26 ties of the major inward (Nav and Cav) and outward (K+) currents
Accessory/Auxiliary Subunits of Myocardial that shape the waveforms of atrial and ventricular action poten-
tials (Figure 3-2). In contrast to the cardiac Nav and Cav cur-
Voltage-Gated K+ Channels 26
rents, there are multiple types of myocardial voltage-gated K+
Molecular Determinants of Native Myocardial Transient (Kv) and non-voltage, inwardly rectifying K+ (Kir) channels
Outward Voltage-Gated K+ Channels 28 (Table 3-1), many of which are differentially expressed, contrib-
uting to regional variations in myocardial action potential wave-
Molecular Determinants of Native Myocardial forms (see Figure 3-1) and refractoriness.1-3 In addition, changes
Delayed Rectifier Voltage-Gated K+ Channels 28 in the densities, distributions, and properties of Kv and Kir chan-
Molecular Determinants of Myocardial Kir and K2P nels are evident in a variety of myocardial diseases, and these
Channels 29 changes affect repolarization, influence propagation and decrease
rhythmicity, effects that can produce substrates for the generation
of life-threatening arrhythmias.1 There is, therefore, consider-
able interest in defining the molecular mechanisms controlling
the biophysical properties and the functional cell surface expres-
Summary sion of these channels. A large number of Kv and Kir pore-
forming α and accessory β (Table 3-2) subunits have been
Electrophysiologic studies on isolated mammalian cardiac myo- identified,4,5 and considerable progress has been made in defining
cytes have identified the presence of multiple types of voltage- the relationships between these subunits and functional myocar-
gated K+ (Kv) channel currents with distinct time- and dial Kv and Kir channels.1,6,7 Importantly, the studies completed
voltage-dependent properties that contribute to determining to date have revealed that the molecular correlates of the various
action potential amplitudes, waveforms and durations. A number types of Kv and Kir channels distinguished electrophysiologically
of non–voltage-gated, inwardly rectifying K+ (Kir) channels that (see Table 3-1) are indeed distinct.1 In the long term, defining
contribute to action potential repolarization have also been func- the molecular compositions of myocardial Kv and Kir channels
tionally identified. The various types of Kv and Kir channels are will also facilitate studies aimed at determining the molecular
differentially expressed in different cardiac cell types, contribut- mechanisms controlling the marked regional differences in the
ing to cellular and regional differences in action potential wave- expression of these channels in the normal myocardium, as well
forms and refractoriness. In the diseased myocardium, K+ current as the derangements in the expression or functioning of these
remodeling is evident and correlated with changes in action channels that occur with myocardial disease. In this chapter, the
potential waveforms and durations, increased dispersion of repo- electrophysiologic and molecular diversity of repolarizing myo-
larization and the development of substrates for reentrant cardial Kv and Kir channels and the molecular determinants of
arrhythmias. The cloning of Kv and Kir channel pore-forming native myocardial K+ channels will be reviewed.
(α) and accessory (β) subunits, as well as the subsequently identi-
fied two pore domain K+ (K2P) channel and small conductance
Ca2+-dependent K+ (SK) channel subunits, has provided insights Myocardial Voltage-Gated K+
into the molecular basis of functional myocardial K+ channel
diversity. In addition, a variety of experimental approaches have Channels: Transient Outward
been, and continue to be, used to define the molecular determi- Voltage-Gated K+ Channels
nants of native cardiac K+ channels and facilitated efforts focused
on exploring the molecular mechanisms controlling the proper- Kv currents, activated on membrane depolarization, influence
ties and the functional cell surface expression of these channels myocardial action potential amplitudes and durations and, in
in the normal and in the diseased myocardium. most cells, two broad classes of Kv currents have been distin-
In atrial and ventricular myocytes, the action potential guished: transient outward K+ currents, Ito; and delayed, out-
upstroke, attributed to inward currents through voltage-gated wardly rectifying K+ currents, IK (see Table 3-1). The transient
Na+ (Nav) channels, is rapid (Figure 3-1). In nodal tissues, which currents (Ito) activate rapidly and underlie early (phase 1)
23
repolarization, whereas the delayed rectifiers (IK) determine the

SA node latter phase (phase 3) of action potential repolarization (see
Figure 3-2) back to the resting membrane potential. These clas-
Atrium
sifications are broad, however, and there are actually multiple
types of transient (Ito) and delayed rectifier (IK) Kv currents (see
AV node
Table 3-1) expressed in cardiac cells.1,3 Electrophysiologic and
Purkinje fiber pharmacologic studies, for example, have clearly demonstrated
that there are two types of transient outward K+ currents, now
referred to as Ito,fast (Ito,f) and Ito,slow (Ito,s), and that these currents
Endocardium
are differentially distributed.6,7 The rapidly activating and inacti-
vating transient outward K+ current, Ito,f, is also characterized by
Midmyocardium rapid recovery from steady-state inactivation, whereas Ito,s recov-
ers slowly from inactivation.6,7 In addition, Ito,f is readily distin-
LV Epicardium guished from other Kv currents (including Ito,s) using Heteropoda
RV
toxin-2 or -3 (see Table 3-1).
0.2 sec Although originally identified in Purkinje fibers, Ito,f is a prom-
R inent repolarizing Kv current in atrial and ventricular myocytes,
P T as well as in nodal cells, in most species.1 There are, however,
marked regional differences in Ito,f densities, with the highest
Q densities typically in atrial myocytes. In addition, in mammalian
S
QT ventricles, Ito,f and Ito,s are differentially distributed. In canine left
ventricles (LVs), for example, Ito,f density is fivefold to sixfold
Figure 3-1. Electrical activity in the mammalian myocardium. Schematic of the higher in epicardial and midmyocardial, than in endocardial,
human heart with the different anatomic regions labeled and representative action
potential waveforms recorded in these different regions illustrated. In the lower
cells.2 There are also marked regional differences in Ito,f densities
panel, a schematic of a surface electrocardiogram is presented with four sequential in adult mouse ventricles.8-10 Specifically, Ito,f density is higher in
beats displayed and the P, Q, R, S, and T waves marked on the last beat. right than in left ventricular myocytes, and within the LV, Ito,f
densities are higher in apex than in base myocytes.8-10 In the
mouse, even greater Kv current heterogeneity is seen in cells
isolated from the septum: all ventricular septum cells express Ito,s,
and most (approximately 80%) also express Ito,f.8 When present,
however, Ito,f density is significantly (p < 0.001) lower in septum
Ventricular Atrial
Phase 1
Phase 2
Phase 0
Phase 3
Pore-
Phase 4 100 ms 100 ms forming
subunit
INa SCN5A
ICa,L CACNA1C
I10,L KCND3
I10,L KCNA4
IK,L KCNQ1
IK KCNH2
IKLr KCNA5
Icc KCNK2/3
IKl KCNJ2/4/12
IKl(Ca) KCNN1/2/3
IK∆TP KCNJ8/11
Figure 3-2. Schematics of action potential waveforms and underlying ionic currents in adult human ventricular and atrial myocytes. The major ionic currents shaping action
potentials in human atrial and ventricular myocytes are schematized, and the main pore-forming α-subunits underlying these currents are listed. As discussed in the text,
there are regional differences in the relative expression levels of some of the repolarizing K+ currents and the contributions of the various K+ currents to shaping action
potential waveforms and controlling repolarization.
Voltage-Regulated Potassium Channels 25
Table 3-1. Potassium Conductances Expressed in the Mammalian Myocardium
Current
Ito, f
Activation
Fast
Pharmacology
mM 4-AP
Pore-Forming (α) Subunits
K CND 3
Expression
Atria, ventricles,
Heterogeneous Expression*
Yes
3
HaTX Purkinje
HpTX
Ba2+
Ito, s Fast mM 4-AP K CNA 4 Ventricles Yes
IKr Moderate E-4031 K CNH 2 Ventricles Yes
Dofetilide
IKs Very slow NE-10064 K CNQ 1 Ventricles Yes
NE-10133
IKur Fast µM 4-AP K CNA 5 Atria No
IKp Fast Ba 2+
?? Ventricles ??
IK Slow mM TEA ?? Ventricles ??
IK, slow1 Fast mM 4-AP K CNA 5 Atria, ventricles No
IK, slow2 Fast mM TEA K CNB 1 Atria, ventricles No
Iss Slow mM TEA K CNK 2/3 Atria, ventricles No
A1899
IKI _ Ba2+ K CNJ 2/4/12 Atria, ventricles No
IK(Ca) Ca2+ Apamin K CNN 1/2/3 Atria No
IK(Ach) Acetylcholine Tertiapin-Q K CNJ 3/5 Atria No
IK(ATP) ATP SUR K CNJ 8/11 Atria, ventricles No
depletion
*Heterogeneous expression in ventricles.
4-AP, 4-aminopyridine; HaTX, hanatoxin; HPTx, heteropodatoxin; SUR, sulfonylureas; TEA, tetraethylammonium.
Table 3-2. Kv Channel Accessory β-Subunits than in RV or LV cells.8 Ito,f and Ito,s are also differentially expressed
Family Protein Gene Current
in ferret LV, and Ito,s is detected only in endocardial LV cells.11
Despite heterogeneities in functional expression, the properties
I I
Kvβ Kvβ1* KCNAB1 to, f K, slow1 of Ito,f and Ito,s in different cardiac cell types (and species) are
Kvβ2* KCNAB2 ?? remarkably similar, leading to suggestions that the molecular
compositions of the underlying (Ito,f and Ito,s) channels are also
Kvβ3 KCNAB3 similar.3
Kvβ4
I
KCNE MinK* KCNE1 Ks
MiRP1* KCNE2 I I I
Kr to, f f
Myocardial Voltage-Gated K+ Channels:
MiRP2* KCNE3 I
to, f
Delayed Rectifier Voltage-Gated K+ Channels
MiRP3 KCNE4 Electrophysiologic and pharmacologic studies have also distin-
MiRP4 KCNE5 guished multiple types of cardiac delayed rectifier K+ currents, IK
I I (see Table 3-1). In atrial myocytes, for example, the dominant
KChAP KChAP* PIAS3 to, f k
repolarizing K+ current is a rapidly activating, non-inactivating
KChIP KChIP1 KCNIP1 K+ current, IKur (IK,ultrarapid), which is not detected in ventricular or
KChIP2* KCNIP2 I /I
to, f CaL
nodal cells.1 In most ventricular myocytes, there are two promi-
nent components of delayed rectification, IKr (IK,rapid) and IKs
KChIP3 KCNIP3 (IK,slow), that are different from IKur in terms of time- and
KChIP4.2 CSEN voltage-dependent properties.1 The biophysical properties of IKr
and IKs are distinct: IKr activates and inactivates rapidly,
KChIP4.3 KCNIP4
displays marked inward rectification, and is selectively blocked
NCS NCS-1* FREQ I
to, f by class III anti-arrhythmics, including dofetilide and sotalol.12
DPPX DPP6* DPP6 I
to, f
In contrast, IKs activates slowly and does not display inward
I
rectification.12
DPP10* DPP10 to, f
Similar to the transient outward K+ currents, there are also
*Expressed in the heart. marked regional differences in the functional expression of IKs
Kv, Voltage-gated K+. and IKr in mammalian ventricular myocytes.1,2 The density of IKs
in canine LV, for example, is higher in epicardial and endocardial
than in M cells.2 There are also regional differences in the func- channels,21 suggesting that action potentials will be shortened
tional expression of IKr and IKs channels in guinea pig LV.13 In markedly when only a few IKATP channels are activated.
cells isolated from the LV free wall, for example, the density of
IKr is higher in subepicardial than in midmyocardial or subendo-
cardial myocytes.13 At the base of the LV, in contrast, IKr and IKs
densities are significantly lower in endocardial, than in either Pore-Forming (α) Subunits of Myocardial
midmyocardial or epicardial, LV cells.13 Differences in Kv current Voltage-Gated K+ Channels
densities contribute to the variations in action potential wave-
forms recorded in different regions (i.e., atria and ventricles, right Kv channel pore-forming (α) subunits are six transmembrane-
and left ventricles, apex and base of the ventricles) of the heart, spanning domain proteins (Figure 3-3) with a region between the
as well as in different layers (epicardial, midmyocardial, and fifth and sixth transmembrane domains that contributes to the
endocardial) of the LV and RV walls.1-3,8-13 K+-selective pore.4 The positively charged fourth transmembrane
In rodent ventricles, there are additional components of IK domain in the Kv α-subunits (see Figure 3-3) is homologous to
with properties distinct from IKs and IKr (see Table 3-1). In mouse the corresponding regions in Nav and Cav channel α-subunits,
and rat ventricular myocytes, for example, there are novel delayed placing them in the S4 superfamily of voltage-gated channels.3,4
rectifier Kv currents that have been referred to as IK, IKslow and Iss In contrast to Nav and Cav channels, in which only a single
(see Table 3-1).1,3 Mouse ventricular IK,slow was first identified as α-subunit is required to form a channel, functional Kv channels
a rapidly activating and slowly inactivating K+ current with prop- comprise four α-subunits (see Figure 3-3). Similar to the diversity
erties distinct from Ito,f, Ito,s and Iss expressed in the same cells.14 of functional myocardial Kv channels (see Table 3-1), however,
In addition, IK,slow was shown to be blocked selectively by micro- multiple Kv α-subunits have been identified. These subunits
molar concentrations of 4-aminopyridine (4-AP), which does not comprise several homologous Kv α-subunit subfamilies, Kv1.x,
affect Ito,f or Ito,s.14 Subsequent work, however, revealed the pres- Kv2.x, Kv3.x, Kv4.x, Kv10.x, Kv11.x, and many members of these
ence of two components of mouse ventricular IK,slow: IK,slow1, which Kv α-subunit subfamilies are expressed in the mammalian heart.1
is blocked by µM 4-AP; and, IK,slow2, which is blocked selectively In addition, further functional Kv channel diversity could arise
by TEA.15-18 In addition, it has been demonstrated that IK,slow1 and through alternative splicing of transcripts and through the forma-
IK,slow2 reflect the expression of distinct molecular entities.15-18 In tion of heteromultimeric channels between two or more Kv
contrast to the differential distribution of Ito,f and Ito,s, however, α-subunit proteins in the same Kv α-subunit subfamily.1,3,4
IK,slow1, IK,slow2, and Iss appear to be uniformly expressed in mouse Additional subfamilies of Kv α-subunits were revealed with
atrial and ventricular myocytes.8-10,15-19 the cloning of the human eag-related (HERG) gene, KCNH2,
subsequently identified as the locus of mutations underlying one
form of familial long QT-syndrome, LQT2, and KCNQ1
(KvLQT1), the locus of mutations in another inherited long QT
Inwardly Rectifying Myocardial K+ Channels syndrome, LQT1.23 Heterologous expression of KCNH2 (ERG1)
Also Contribute to Repolarization reveals inwardly rectifying Kv currents23 with properties similar
to cardiac IKr (see Table 3-1). Although there are several ERG
In addition to the Kv currents, Kir currents, specifically IK1 and (KCNH) subfamily members, only KCNH2 (which encodes
the ATP-dependent K+ current, IKATP (Table 3-1), contribute to ERG1) appears to be expressed in the myocardium.1 Heterolo-
shaping myocardial action potential waveforms.20-22 Similar to the gous expression of KCNQ1 (KvLQT1) alone reveals rapidly acti-
Kv currents, the densities of the Kir currents vary in different vating and non-inactivating Kv currents, whereas co-expression
regions of the heart (e.g., atria, ventricles, conducting tissue).1,3 with the Kv accessory subunit, minK (see Table 3-2), produces
In contrast to the Kv currents, however, myocardial Kir current slowly activating Kv currents that resemble the slow component
densities are similar in myocytes in different regions of the ven- of cardiac-delayed rectification, IKs.1,3,23
tricles.1 In mammalian atrial and ventricular myocytes, IK1 plays
a role in establishing resting membrane potentials and plateau
potentials, and contributes to phase 3 repolarization (see Figure
3-2). The fact that the conductances of IK1 channels are high at Accessory/Auxiliary Subunits of
negative membrane potentials underlies the contribution of IK1 Myocardial Voltage-Gated K+ Channels
to atrial and ventricular resting membrane potentials.20 Although
the voltage-dependent properties of IK1 channels are such that In addition to the Kv α-subunits, a number of Kv channel acces-
conductance is low at potentials positive to approximately sory (Kv β) subunits have also been identified (see Table 3-2).
−40 mV, these channels nevertheless contribute outward K+ cur- The first of these subunits was KCNE1, which encodes a small
rents during the plateau phase of the action potential in ventricu- (130-aa) protein (minK) with a single transmembrane spanning
lar cells,20 as well as during phase 3 repolarization (see Figure domain.24 It appears that minK coassembles with KvLQT1 to
3-2), because the driving force on K+ is high at depolarized mem- form functional cardiac IKs channels.23,24 Additional minK homo-
brane potentials. logs, MiRP1 (KCNE2), MiRP2 (KCNE3), and MiRP3 (KCNE4)
Myocardial ATP-dependent K+ channels are weak, inwardly have also been identified (see Table 3-2), and it has been sug-
rectifying channels that are inhibited by (elevated) intracellular gested that MiRP1 (KCNE2) functions as an accessory subunit
ATP and activated by nucleotide diphosphates.21 In ventricular coassembling with ERG1 to generate cardiac IKr.25 It has also
myocytes, activation of IKATP channels during periods of hypoxia been reported that the MiRP subunits interact with multiple Kv
and ischemia results in action potential shortening, suggesting α-subunit subfamilies and modify channel properties.24 MiRP2,
that these channels provide a link between cellular metabolism for example, coassembles with Kv3.4 in mammalian skeletal
and membrane potential.21,22 The opening of IKATP channels muscle,26 and MiRP1 coassembles with Kv4.x α-subunits when
appears to contribute to the cardioprotection resulting from is coexpressed in heterologous cells.27 These observations suggest
chemic preconditioning.22 In contrast with some ventricular Kv that the MiRP (KCNE) accessory subunits can assemble with a
channels, IKATP channels appear to be distributed uniformly in the variety of Kv α-subunits and contribute to the formation of
right and left ventricles and through the thickness of the right or multiple types of myocardial Kv channels. Direct experimental
left ventricular walls. Interestingly, IKATP channels are expressed support for this hypothesis, however, has not been provided, and
at much higher densities than other sarcolemmal myocardial K+ the roles of the various KCNE subunits in the generation of
3
Kv Kir K2P
– +H N
CO2 +H N 2
+H N 2 –
2 CO2 CO2
–
Kv Kir K2P
+H N
2
– +H
+H N
2
CO2 2N
+H N +H
2 2N 2–
–
+H CO CO2
B 2N
MiRP
KChIP
Kirβ Kvβ
C
Figure 3-3. Pore-forming α-subunits and assembly of functional K+ channels. A, Linear sequences and transmembrane topologies of the pore-forming α-subunits encoding
voltage-gated (Kv), inwardly rectifying (Kir), and two-pore domain (K2P) K+ channels are illustrated. B, The tetrameric assembly of Kv and Kir α-subunits and the dimeric
assembly of K2P α-subunits are illustrated below the primary sequences of the α-subunits. C, Assembled, functional K+ channels composed of pore-forming α-subunits
and one or more cytosolic and transmembrane accessory subunits.
functional cardiac Kv26,27 and other non-Kv28 channels remain to domains.31 When coexpressed with Kv4 α-subunits, each of the
be defined. KChIPs increases K+ current densities, slows inactivation, speeds
Another type of Kv channel accessory subunit was revealed recovery from inactivation, and shifts the voltage dependence of
with the isolation of low molecular weight (approximately 45 kD) current activation.31 In contrast, KChIP coexpression does not
cytosolic (Kvβ) subunits from brain.5 Three homologous Kv affect the properties or the densities of Kv1.4- or Kv2.1-encoded
β-subunits, Kvβ1, Kvβ2, and Kvβ3 (see Table 3-2), as well as K+ currents, consistent with the suggestion that the modulatory
alternatively spliced transcripts, have been identified, and both effects of the KChIPs are specific for Kv4 α-subunit–encoded Kv
Kvβ1 and Kvβ2 are expressed in heart.1 Previous studies have channels.31 More recent studies, however, suggest that the
shown that Kvβ subunits interact with the intracellular domains KChIPs modulate the functional cell surface expression of Kv1.5-
of Kv1 α-subunits, and functional studies demonstrated that Kvβ encoded Kv channels,38 as well as myocardial Cav channels.39 In
subunits affect the biophysical properties and increase the cell addition, although KChIP binding to Kv4 α-subunits is not Ca2+
surface expression of heterologously expressed Kv1 α-subunit– dependent, mutations in EF hand domains 2, 3, and 4 eliminate
encoded Kv currents.5 Because Kv α- and β-subunits appear to the modulatory effects of KChIP1 on Kv4- and Kv1.5-encoded
coassemble in the endoplasmic reticulum,29 the increase in func- Kv currents,31,38 suggesting a role for voltage-dependent Ca2+
tional channel expression suggests that the Kv β-subunits affect entry and intracellular Ca2+ levels in the regulation of functional
channel assembly, processing or stability or, alternatively, func- cardiac (Kv4-encoded) Ito,f channels, as has been demonstrated
tion as chaperone proteins. Although Kvβ1 and Kvβ2 have been for neuronal Kv4-encoded channels.40
shown to associate with Kv4 subunits in the mouse myocardium The transmembrane diaminopeptidyl transferase-like protein
and the targeted deletion of Kvβ1 results in the attenuation of 6 (DPP6 or DPPX) has also been suggested as an accessory
Kv4-encoded Ito,f in mouse ventricular myocytes,30 the roles of subunit of cardiac41 and neuronal42 Kv4-encoded channels (see
Kvβ subunits in the generation of myocardial Kv channels in Table 3-2). Lacking enzymatic activity, DPP6 increases the cell
other species have not been explored. surface expression of Kv4 α-subunits, shifts the voltage depen-
A yeast two-hybrid screen, using the N terminus of Kv4.2 as dences of activation and inactivation of heterologously expressed
the bait, led to the identification of accessory Kv channel interact- Kv4 currents to more negative potentials, and accelerates the
ing proteins, (KChIPs) (see Table 3-2) in brain.31 Of the four rates of current activation, inactivation, and recovery.41,42 Inter-
KChIPs identified,31,32 only KChIP2 is expressed in heart,31,33 estingly, heterologous coexpression of DPP6 with Kv4.3 and
although there are several KChIP2 splice variants.34-36 Interest- KChIP2 produces Kv currents that closely resemble native
ingly, the KChIPs belong to the recovering family of neuronal cardiac Ito,f.41 Another member of the DPP-like subfamily of
Ca2+-sensing (NCS) proteins,37 and contain four EF-hand proteins, DPP10, has been demonstrated to associate with Kv4.2
and KChIP3 in the rat brain and to have regulatory effects similar α-subunits.61 The protein expression pattern of this putative Kv4
to DPP6 on heterologously expressed Kv4 (with and without channel accessory subunit has not been examined in canine or
KChIP) channels.43,44 In addition, DPP10 has been shown to be human hearts, and a role for NCS-1 in controlling functional
expressed in normal and failing human ventricles.45 myocardial Ito,f densities has not been determined directly.
Although accumulating evidence suggests that cardiac Ito,f chan-
nels function in macromolecular complexes, comprising Kv4
Molecular Determinants of Native α-subunits and multiple cytosolic and transmembrane accessory
subunits (Figure 3-4), the molecular compositions of native myo-
Myocardial Transient Outward cardial Ito,f channels have not been determined. In addition, the
Voltage-Gated K+ Channels functional roles of additional Kv accessory subunits, including
members of the KCNE2762 and DPPX4145 subfamilies, in the
Considerable experimental evidence has accumulated demon- generation of native myocardial Ito,f channels and in determining
strating a critical role for Kv α-subunits of the Kv4 subfamily in regional differences in myocardial Ito,f densities remain to be
the generation of cardiac Ito,f channels. In rat and mouse ventricu- defined.
lar myocytes exposed to antisense oligodeoxynucleotides targeted The kinetic and pharmacologic properties of the slow tran-
against Kv4.2 or Kv4.3, Ito,f density is reduced by approximately sient outward K+ currents, Ito,s, in ventricular myocytes are dif-
50%.46,47 Rat ventricular Ito,f density is also reduced in cells ferent from Ito,f (see Table 3-1), which is interpreted as suggesting
exposed to an adenoviral construct encoding a truncated Kv4.2 that the molecular correlates of ventricular Ito,s and Ito,f channels
subunit (Kv4.2ST) that functions as a dominant negative.48 In are also distinct. Direct experimental support for this hypothesis
addition, Ito,f is eliminated in ventricular and in atrial myocytes was provided in electrophysiologic experiments on myocytes iso-
isolated from transgenic mice expressing a pore mutant of Kv4.2, lated from mice with a targeted deletion of the Kv1.4 gene,
Kv4.2W362F (Kv4.2DN), which also functions as a dominant Kv1.4–/–,63 demonstrating that Ito,s is undetectable in septum
negative.49,50 Although biochemical and electrophysiologic studies cells.64 The properties and the densities of Ito,f, IK,slow1, IK,slow2, and
suggested that Kv4.2 and Kv4.3 are associated in adult mouse Iss in Kv1.4–/– left and right ventricular (and in atrial) myocytes,
ventricles and that functional mouse ventricular Ito,f channels are however, are indistinguishable from those measured in wild type
heteromeric,47 targeted deletion of Kv4.2 eliminates mouse ven- cells.64,65 Interestingly, upregulation of Ito,s is evident in the ven-
tricular Ito,f,51 whereas elimination of Kv4.3 has no effect.52 In tricles of Kv4.2DN-expressing mice in which Ito,f is eliminated.65
mouse ventricles, therefore, Kv4.2 is the critical α-subunit Given the similarities in the time- and voltage-dependent proper-
required for the generation of functional Ito,f channels.51 Given ties of the slow transient outward K+ currents in other species
the similarities in the properties of Ito,f (see Table 3-1), it seems with mouse Ito,s (1), it seems reasonable to suggest that Kv1.4
reasonable to suggest that Kv4 α-subunits also underlie Ito,f in likely also encodes Ito,s in ferret, rabbit, human, and canine ven-
other species. In canine and in human myocardium, however, the tricular myocytes.
candidate subunit is Kv4.3 because Kv4.2 is barely detectable.53
Although two splice variants of Kv4.3 have been identified,54 the
expression levels of the two predicted Kv4.3 proteins and the
functional roles of these variants in the generation of functional Molecular Determinants of Native Myocardial
cardiac Ito,f channels have been determined. Delayed Rectifier Voltage-Gated K+ Channels
It has also been demonstrated that the Kv channel accessory
subunit KChIP2 coimmunoprecipitates with Kv4 α-subunits As noted earlier, ERG1 is the locus of mutations in LQT2, and
from adult mouse ventricles, which is consistent with a role for heterologous expression of ERG1 reveals voltage-gated, inwardly
this subunit in the generation of Kv4-encoded mouse ventricular rectifying K+-selective channels that are similar to cardiac IKr.23
Ito,f channels.47 In ferret and canine hearts, a gradient in KChIP2 Alternatively processed forms of ERG1, with unique N- and
message expression is observed through the thickness of the ven- C-termini, have also been identified in mouse and human hearts
tricular wall,35,36 leading to suggestions that the differential and suggested to have roles in the generation of native cardiac
expression of KChIP2 underlies the epicardial–endocardial dif- IKr channels.66-68 It has also been suggested that functional cardiac
ferences in Ito,f densities. Subsequent studies revealed that the IKr channels are multimeric, comprising ERG1 and minK, and
patterns of expression of the KChIP2 message, the KChIP2 biochemical studies have demonstrated coimmunoprecipitation
protein, and Ito,f densities in canine ventricles are indeed similar,55 of ERG1 and minK from equine ventricles.69 It is not clear
which is consistent with an important role for KChIP2 in deter- whether ERG1 and minK or other members of the KCNE family
mining functional canine and human ventricular Ito,f densities. In are also found in association in other species.
contrast, in rat and mouse ventricles, there is little or no gradient Although heterologous expression of KCNQ1, the locus of
in KChIP2 message or protein (51,56) expression, and it appears mutations in LQT1, reveals rapidly activating, non-inactivating
that regional differences in Kv4.2 expression underlie the hetero- Kv currents, coexpression with minK produces slowly activating
geneities in Ito,f densities in rodent ventricles.56,57 Molecular Kv currents similar to cardiac IKs.23 These observations, together
insights into the regulation of the observed regional variations in with biochemical data demonstrating that heterologously
the expression of the Kv4.2 transcripts were provided with the expressed KvLQT1 and minK associate, were interpreted as sug-
demonstrations that the expression levels of the transcription gesting that minK coassembles with KvLQT1 to form functional
factors, Irx5 and NFAT, are positively and negatively correlated, cardiac IKs channels.1,23 Direct biochemical evidence for the in
respectively, with the differences in Kv4.2 expression and Ito,f situ coassembly of KvLQT1 and minK was recently provided in
densities.58,59 Interestingly, approximately 25 transcription factors studies of equine ventricles.69 Similar data for human ventricular
were subsequently shown to be differentially expressed in the IKs are yet to be provided. Interestingly, it was reported that
ventricles,60 although the functional import of these findings KvLQT1 modulates the distribution and properties of ERG1-
remains to be determined. encoded channels, an observation interpreted as suggesting
A role for the KChIP2 splice variant, KChIP2b, in determin- that cardiac IKs and IKr channels are regulated directly
ing regional differences in ferret ventricular Ito,f densities has also through Kv α-subunit–Kv α-subunit interactions.70 The molecu-
been proposed.35 The KChIP-related (NCS family) protein, lar mechanisms controlling the cell surface expression of func-
NCS-1, appears to be uniformly expressed in adult mouse ventional IKs channels, the regional differences in functional
tricles and, in addition, coimmunoprecipitates with Kv4 ventricular IKs densities, and the interactions between cardiac IKs
DPP6/10
MinK/MiRP
T1
KChIP2
A Kvβ2 B
Figure 3-4. Schematic of putative Kv4.3-encoded, myocardial Ito,f channel macromolecular complex. A, Cross section of a Kv4.3 channel in a membrane showing two Kv4.3
α-subunits (blue), generated based on the structure of Kv1.2,91 each interacting with a cytosolic KChIP2 (red) and a cytosolic Kvβ (green) accessory subunit (1 : 1 : 1 : stoichi-
ometry) through distinct, non-overlapping N-terminal domains. The transmembrane accessory subunits, DPP6/10 (brown)92 and MinK/MiRPs (yellow), which have also been
proposed to interact with Kv4.3 α-subunits (1 : 2 stoichiometry) and to contribute to the formation of functional cardiac Ito,f channels. B, Structural analyses of Kv4.3N-KChIP1
complexes93 revealed a 1 : 1 stoichiometry with each KChIP (red) bridging two adjacent Kv4.3 N termini (blue), anchoring hydrophobic Kv4.3 N terminal residues in a hydro-
phobic binding pocket in KChIP1. Protein structures illustrated were generated based on published structural data91-93 using PyMOL.
and IKr (and possibly other cardiac channels), however, remain to

be determined. Molecular Determinants of Myocardial Kir
Similar to the transient outward Kv currents, molecular and K2P Channels
genetic methods, in combination with biochemistry and electro-
physiology primarily in mice, have provided molecular insights In cardiac and other cells, inwardly rectifying K+ (IK1) channels
into the basis of functional delayed rectifier Kv channel diversity are encoded by a large and diverse subfamily of inward rectifier
in the murine myocardium. A role for Kv1 α-subunits in the K+ (Kir) channel pore-forming α-subunit genes,4 each of which
generation of mouse ventricular IK,slow, for example, was revealed encodes a protein with two transmembrane domains that assem-
with the demonstration that IK,slow is selectively attenuated in ble as tetramers to form K+ selective pores (see Figure 3-3). Based
ventricular myocytes isolated from transgenic mice expressing a on the properties of the currents produced in heterologous
truncated, dominant negative Kv1 α-subunit, Kv1.1DN.14 It was expression systems, Kir2 α-subunits were long thought to encode
subsequently shown, however, that IK,slow is also reduced in ven- the strongly inwardly rectifying cardiac IK1 channels,20 and several
tricular myocytes expressing a dominant negative Kv 2.1 mutant, members of the Kir2 subfamily are expressed in the myocar-
Kv2.1DN,15 revealing that there are two molecularly distinct dium.71 Direct insights into the roles of Kir2 α-subunits in the
components of mouse ventricular IK,slow : IK,slow1 that is sensitive to generation of cardiac IK1 channels were provided in studies com-
micromolar concentrations of 4-AP and encoded by Kv1 pleted on myocytes isolated from mice lacking KCNJ2 (Kir2.1–/–)
α-subunits, and IK,slow2 that is sensitive to TEA and encoded by or KCNJ12 (Kir2.2–/–).72 Although Kir2.1–/– mice have cleft palate
Kv2 α-subunits.15 Subsequent studies revealed that IK,slow1 is elim- and die shortly after birth, precluding electrophysiologic studies
inated in ventricular myocytes isolated from mice harboring the on myocytes from adult animals, voltage-clamp recordings from
targeted disruption of the KCNA5 (Kv1.5) locus, revealing that isolated newborn Kir2.1–/– ventricular myocytes revealed that IK1
Kv1.5 encodes IK,slow1.16 These findings, together with the previ- is absent, whereas a quantitative reduction in IK1 was observed in
ous studies completed on cells from Kv1.4–/– animals63 in which adult Kir2.2–/– ventricular myocytes,72 suggesting that both Kir2.1
Ito,s is eliminated,64 reveal that, in contrast to the Kv4 α-subunits, and Kir2.2 contribute to mouse ventricular IK1 channels and that
Kv4.2, and Kv4.3,47 the Kv1 α-subunits, Kv1.4 and Kv1.5, do not functional cardiac IK1 channels are heteromeric. The quantitative
associate in adult mouse ventricles in situ. Rather, functional Kv1 differences between the effects of the deletion of KCNJ2 and
α-subunit–encoded Kv channels in mouse ventricular myocytes KCNJ12 further suggest that Kv2.1 (KCNJ2) is the critical
are homomeric, composed of Kv1.4 α-subunits (Ito,s)63 or Kv1.5 subunit underlying (mouse) IK1 channels.72
α-subunits (IK,slow1).16 The roles of Kv accessory subunits in the Mutations in KCNJ2 have been linked to congenital long
generation of functional Ito,s, IK,slow1 and IK,slow2 channels and the QT (Andersen-Tawil syndrome or LQT7) and short QT
molecular mechanisms controlling the differential expression of syndromes,73,74 and increasing expression of Kir2.1 in the
these channels remain to be determined. mouse heart, which results in the upregulation of IK1,
is proarrhythmic.75,76 Previous studies have identified regional

Table 3-3. Two-Pore K+ (K2P) Channel α-Subunits
differences in myocardial IK1 expression and properties in adult
mouse heart,77,78 and it has been suggested that these differences Family Subfamily Protein Gene Cardiac
reflect the variable subunit compositions of the channels, as well Current
as differences in polyamine concentrations.79 Studies focused on
Two-Pore TWIK TWIK-1* KCNK1 ??
testing this hypothesis directly and on defining the molecular
mechanisms controlling regional differences in the expression TWIK-2* KCNK6 ??
and functioning of native IK1 channels in species other than mice TWIK-3 KCNK7
are also clearly warranted.
In the mammalian heart, IKATP channels appear not to have a TWIK-4 KCNK8
prominent role in action potential repolarization, but are thought TREK TREK-1* KCNK2 ??
to be important in myocardial ischemia and preconditioning.21,22
TREK-2 KCNK10
In heterologous systems, IKATP channels can be reconstituted by
coexpression of Kir6.x subunits with one or more ATP-binding TASK TASK-1* KCNK3 IKp??
cassette proteins that encode sulfonylurea receptors, SURx.80 TASK-2 KCNK5
Pharmacologic and mRNA expression studies suggested that
cardiac sarcolemmal IKATP channels are encoded by Kir6.2 and TASK-3 KCNK9
SUR2A, and the essential role of Kir6.2 was demonstrated TASK-4 KCNK14
directly in experiments showing that IKATP channels are undetect-
TASK-5 KCNK15
able in ventricular myocytes isolated from Kir6.2–/– animals.81 In
addition, cardiac IKATP channel activity is reduced in SUR2–/– TRAAK TRAAK-1* KCNK4
myocytes82 and unaffected in SUR1-/- myocytes,83 suggesting an THIK THIK-1 KCNK13
important role for SUR2. Interestingly, the properties of the
residual IKATP channels in SUR2–/– myocytes are similar to those THIK-2* KCNK12 ??
produced in heterologous cells on coexpression of Kir6.2 and TALK TALK-1 KCNK16
SUR1,82 suggesting that SUR1 also contributes to the generation
TALK-2* KCNK17 ??
of functional cardiac IKATP channels by coassembling with Kir6.2
α-subunits alone or with SUR2A. Biochemical and molecular *Expressed in the heart.
genetic strategies will need to be combined to define the molecu-
lar components of native cardiac IKATP channels directly.
Although action potential waveforms in Kir6.2–/– and wild suggest that K2P subunits might contribute to background or
type ventricular myocytes are indistinguishable, the action poten- “leak” K+ channels (i.e., channels with properties similar to the
tial shortening observed in wild type cells during ischemia or steady-state, non-inactivating K+ current [Iss] that is expressed or
metabolic blockade is abolished in the Kir6.2–/– cells,81 consistent characterized in rodent myocytes).8,9,85-87 In recent studies com-
with the hypothesis that cardiac IKATP channels have a role under pleted on rat atrial and ventricular myocytes, roles for TASK188-90
pathophysiologic conditions, particularly those involving meta- and TREK186,87 subunits in the generation of Iss have been sug-
bolic stress. Interestingly, action potential durations are largely gested. Clearly, experiments focused on testing these hypotheses
unaffected in transgenic animals expressing mutant IKATP chan- are needed to provide clear insights into the molecular basis of
nels with markedly (fortyfold) reduced ATP sensitivity, suggest- Iss and to allow further study focused on defining the mechanisms
ing that there are additional inhibitory mechanisms that regulate controlling the physiologic and the pathophysiologic regulation
cardiac IKATP channel activity in vivo.84 Similar to the strong of these (Iss) channels.
inwardly rectifying myocardial Kir currents, IK1, further study is
needed to provide molecular insights into the mechanisms con-
trolling regional differences in the functional expression of myo-
cardial IKATP channels. Conclusions
In addition to Kv and Kir channel α-subunits that assemble
as tetramers, a novel type of K+ α-subunit with four transmem- Cellular electrophysiologic studies have distinguished multiple
brane spanning regions and two pore domains (see Figure 3-3) types of voltage-gated inward and outward currents that contrib-
was identified with the cloning of TWIK-1.85 Both pore domains ute to action potential repolarization in mammalian cardiac cells
contribute to the formation of the K+ selective pore, and TWIK-1 (see Table 3-1). The outward (K+) currents are more numerous
subunits assemble as dimers.86 A large number of two-pore and more diverse than the inward (Na+, Ca2+) currents, and most
domain K+ (K2P) channel α-subunit genes have been identified, cardiac cells express a repertoire of voltage-gated and inwardly
and several are expressed in the mammalian myocardium (Table rectifying K+ channels (see Table 3-1). In addition, several of
3-3). Heterologous expression of K2P subunits reveals currents these K+ channels are expressed differentially in different myo-
with distinct biophysical properties and sensitivities to several cardial cell types and through the thickness of the ventricular
potential intracellular and extracellular modulators, including walls. Molecular cloning led to the identification of multiple
anesthetics, change in pH, and fatty acids.85 voltage-gated (Kv), non–voltage-gated, inwardly rectifying (Kir),
The multiplicity of K2P α-subunits (see Table 3-3), the wide- and weakly rectifying, non-inactivating (K2P) K+ channel pore-
spread distribution of expressed subunits and the findings that forming α-subunits and a number of channel accessory (β) sub-
the properties of the channels encoded by K2P subunits are regu- units (see Table 3-2) thought to contribute to the formation of
lated by a variety of physiologically (and pathophysiologically) the various cardiac K+ currents that have been distinguished elec-
relevant stimuli suggest that K2P channels likely subserve a trophysiologically (see Table 3-1). In recent years, considerable
variety of important functions. As with other cell types, the physi- progress has been made in identifying the pore-forming Kv and
ologic roles of these subunits and channels in the myocardium Kir α-subunits contributing to the formation of most of the K+
are just beginning to be explored. Both TREK-1 and TASK-1 channels expressed in mammalian cardiac myocytes (see Figure
are detected in heart, and heterologous expression of these sub- 3-2). In addition, biochemical studies have provided some insights
units gives rise to instantaneous, non-inactivating K+ currents into the molecular mechanisms underlying the observed hetero-
that display little or no voltage dependence.85 These properties geneities in the expression of myocardial Kv and Kir currents.
For cardiac Ito,f, regional differences in current densities are cor- roles of the various Kvβ and the K2P α-subunits in the genera-
3
related with differences in Kv4.2 protein expression in adult tion of functional myocardial K+ channels and on defining the
mouse ventricles,57 whereas variable expression of the Ito,f channel molecular mechanisms controlling the properties and the cell
accessory protein, KChIP2, has been suggested as underlying the surface expression of myocardial K+ channels encoded by the
transmural gradient in Ito,f densities in canine and human ven- various Kv α and β, Kir α and β, and K2P α-subunits. In addition,
tricles.55 For cardiac IK1 channels, in contrast, recent studies as numerous studies have documented changes in functional K+
suggest that differences in Kir channel α-subunit composition or channel expression in a variety of myocardial disease states,
differences in the concentrations of intracellular polyamines changes that could reflect modifications in channel properties, as
appear to have roles in regulating the functional diversity of these well as alterations in the molecular compositions of the channels
channels.78 or the processing of the underlying channel subunits, it seems
Accumulating evidence suggests that native myocardial Kv clear that a major focus of future research will be on defining
channels likely function in macromolecular protein complexes these mechanisms in detail. There are numerous possible mecha-
(see Figure 3-4), comprising pore-forming α-subunits and mul- nisms, including transcriptional, translational, and posttransla-
tiple cytosolic and transmembrane accessory subunits (see Table tional, that likely are important in regulating the functional
3-2). The molecular compositions of native myocardial Kv chan- expression and the properties of myocardial K+ channels in the
nels, however, have not been determined. In contrast to the normal, as well as in the damaged or diseased, myocardium.
progress made in defining the Kv (and the Kir) α-subunits encod- Defining these mechanisms will have an important effect on the
ing native myocardial Kv (and Kir) channels or currents, however, field, leading to new insights into the mechanisms that regulate
much less is known about the functional roles of the putative myocardial K+ channel expression and functioning and to the
accessory subunits (see Table 3-2) of these channels, as well as development of novel therapeutic strategies to prevent or reverse
about the functioning of the various K2P subunits (see Table 3-3). the remodeling of these channels associated with systemic or
An important focus of future work will likely be on defining the myocardial disease.
rectifier K+ current by dofetilide. Circ Res 72:75– 25. Abbott GW, Sesti F, Splawski I, et al: MiRP1 forms
References 83, 1993. IKr potassium channels with HERG and is associ-
13. Bryant SM, Wan X, Shipsey SJ, et al: Regional ated with cardiac arrhythmia. Cell 97:175–187,
1. Nerbonne JM, Kass RS: Molecular physiology of differences in the delayed rectifier current (IKr and 1999.
cardiac repolarization. Physiol Rev 85:1205–1253, IKs) contribute to the differences in action potential 26. Abbott GW, Butler MH, Bendahhou S, et al:
2005. duration in basal left ventricular myocytes in MiRP2 forms potassium channels in skeletal
2. Antzelevitch C, Dumaine R: Electrical heterogene- guinea-pig. Cardiovasc Res 40:322–331, 1998. muscle with Kv3.4 and is associated with periodic
ity in the heart: physiological, pharmacological and 14. London B, Jeron A, Zhou J, et al: Long QT and paralysis. Cell 104:217–231, 2001.
clinical implications. In Page E, Fozzard HA, ventricular arrhythmias in transgenic mice express- 27. Zhang M, Jiang M, Tseng GN: minK-related
Solaro RJ, editors: Handbook of Physiology, ing the N terminus and first transmembrane peptide 1 associates with Kv4.2 and modulates its
Section 2: The Cardiovascular System, Vol. 1, New segment of a voltage-gated potassium channel. gating function: potential role as beta subunit of
York, 2002, Oxford, pp 654–692. Proc Natl Acad Sci U S A 95:2926–2931, 1998. cardiac transient outward channel? Circ Res
3. Nerbonne JM: Molecular basis of functional 15. Xu H, Barry DM, Li H, et al: Attenuation of the 88:1012–1019, 2001.
voltage-gated K+ channel diversity in the mam- slow component of delayed rectification, action 28. Yu H, Wu J, Potapova I, et al: MinK-related
malian myocardium. J Physiol (Lond) 525:285– potential prolongation, and triggered activity in peptide 1: A beta subunit for the HCN ion channel
298, 2000. mice expressing a dominant-negative Kv2 alpha subunit family enhances expression and speeds
4. Coetzee WA, Amarillo Y, Chiu J, et al: Molecular subunit. Circ Res 85:623–633, 1999. activation. Circ Res 88:E84–E87, 2001.
diversity of K+ channels. Ann NY Acad Sci 16. London B, Guo W, Pan XH, et al: Targeted 29. Nagaya N, Papazian DM: Potassium channel alpha
868:233–285, 1999. replacement of Kv1.5 in the mouse leads to loss of and beta subunits assemble in the endoplasmic
5. Pongs O, Leicher T, Berger M, et al: Functional the 4-aminopyridine-sensitive component of reticulum. J Biol Chem 272:3022–3027, 1997.
and molecular aspects of voltage-gated K+ channel I(K,slow) and resistance to drug-induced QT prolon- 30. Aimond F, Kwak SP, Rhodes KJ, et al: Accessory
β subunits. Ann NY Acad Sci 868:344–355, 1999. gation. Circ Res 88:940–946, 2001. Kvbeta1 subunits differentially modulate the func-
6. Oudit GY, Kassiri Z, Sah R, et al: The molecular 17. Zhou J, Kodirov S, Murata M, et al: Regional tional expression of voltage-gated K+ channels in
physiology of the cardiac transient outward potas- upregulation of Kv2.1-encoded current, IK,slow2, in mouse ventricular myoctyes. Circ Res 96:451–458,
sium current (Ito) in normal and diseased myocar- Kv1DN mice is abolished by crossbreeding with 2005.
dium. J Mol Cell Cardiol 33:851–872, 2001. Kv2DN mice. Am J Physiol Heart Circ Physiol 31. An WF, Bowlby MR, Betty M, et al: Modulation
7. Niwa N, Nerbonne JM: Molecular determinants of 284:H491–H500, 2003. of A-type potassium channels by a family of
cardiac transient outward K+ current (Ito) expres- 18. Li H, Guo W, Yamada KA, et al: Selective elimina- calcium sensors. Nature 403:553–556, 2000.
sion and regulation. J Mol Cell Cardiol 48:12–25, tion of I(K,slow1) in mouse ventricular myocytes 32. Morohashi Y, Hatano N, Ohya S, et al: Molecular
2010. expressing a dominant negative Kv1.5alpha cloning and characterization of CALP/KChIP4, a
8. Xu H, Guo W, Nerbonne JM: Four kinetically subunit. Am J Physiol Heart Circ Physiol novel EF-hand protein interacting with presenilin
distinct depolarization-activated K+ currents in 286:H319–H328, 2004. 2 and voltage-gated potassium channel subunit
adult mouse ventricular myocytes. J Gen Physiol 19. Bou-Abboud E, Li H, Nerbonne JM: Molecular Kv4. J Biol Chem 277:14965–14975, 2002.
113:661–678, 1999. diversity of the repolarizing voltage-gated K+ cur- 33. Rosati B, Pan Z, Lypen S, et al: Regulation of
9. Brunet S, Aimond F, Li H, et al: Heterogeneous rents in mouse atrial cells. J Physiol 529:345–358, KChIP2 potassium channel beta subunit gene
expression of repolarizing voltage-gated K+ cur- 2000. expression underlies the gradient of transient
rents in adult mouse ventricles. J Physiol 559:103– 20. Lopatin AN, Nichols CG: Inward rectifiers in the outward current in canine and human ventricle. J
120, 2004. heart: an update on I(K1). J Mol Cell Cardiol Physiol 533:119–125, 2001.
10. Liu J, Kim K-H, London B, et al: Dissection of the 33:625–638, 2001. 34. Deschenes I, DiSilvestre D, Juang GJ, et al: Regu-
voltage-activated potassium outward currents in 21. Flagg TP, Nichols CG: Sarcolemmal K(ATP) lation of Kv4.3 current by KChIP2 splice variants:
adult mouse ventricular myocytes: Ito,f, Ito,s, IKslow1, channels: what do we really know? J Mol Cell a component of native cardiac I(to)? Circulation
IKslow2 and Iss. Basic Res Cardiol 106:189–204, 2011. Cardiol 39:61–70, 2005. 106:423–429, 2002.
11. Brahmajothi MV, Campbell DL, Rasmussen RL, 22. Grover GJ, Garlid KD: ATP-Sensitive potassium 35. Patel S, Campbell DL, Morales MJ, et al: Hetero-
et al: Distinct transient outward potassium current channels: a review of their cardioprotective phar- geneous expression of KChIP2 isoforms in the
(Ito) phenotypes and distribution of fast- macology. J Mol Cell Cardiol 32:677–695, 2000. ferret heart. J Physiol 539:649–656, 2002.
inactivating potassium channel alpha subunits in 23. Keating MT, Sanguinetti MC: Molecular and cel- 36. Decher N, Barth AS, Gonzalez T, et al: Novel
ferret left ventricular myocytes. J Gen Physiol lular mechanisms of cardiac arrhythmias. Cell KChIP2 isoforms increase functional diversity of
113:581–600, 1999. 104:569–580, 2001. transient outward potassium currents. J Physiol
12. Jurkiewicz NK, Sanguinetti MC: Rate-dependent 24. Abbott GW, Goldstein SA: A superfamily of small 557:761–772, 2004.
prolongation of cardiac action potentials by a potassium channel subunits: form and function of 37. Burgoyne RD, Weiss JL: The neuronal calcium
methanesulfonanilide class III antiarrhythmic the MinK-related peptides (MiRPs). Q Rev sensor family of Ca2+-binding proteins. Biochem J
agent. Specific block of rapidly activating delayed Biophys 31:357–398, 1998. 353:1–12, 2001.
38. Li H, Guo W, Mellor RL, et al: KChIP2 modulates isoforms in the murine heart based on laser capture mutation in the KCNJ2 gene. Circ Res 96:800–
the cell surface expression of Kv1.5-encoded K(+) microdissection. Cardiovasc Res 73:739–749, 807, 2005.
channels. J Mol Cell Cardiol 39:121–132, 2005. 2007. 75. Noujaim SF, Pandit SV, Berenfeld O, et al:
39. Thomsen MB, Wang C, Ozgen N, et al: Accessory 57. Dixon JE, McKinnon D: Quantitative analysis of Up-regulation of the inward rectifier K+ current
subunit KChIP2 modulates the cardiac L-type potassium channel mRNA expression in atrial and (IK1) in the mouse heart accelerates and stabilizes
calcium current. Circ Res 104:1382–1389, 2009. ventricular muscle of rats. Circ Res 75:252–260, rotors. J Physiol 578:315–326, 2007.
40. Anderson D, Mehaffey WH, Iftinca M, et al: Regu- 1994. 76. Piao L, Li J, McLerie M, et al: Transgenic upregu-
lation of neuronal activity by Cav3-Kv4 channel 58. Costantini DL, Arruda EP, Agarwal P, et al: The lation of IK1 in the mouse heart is proarrhythmic.
signaling complexes. Nat Neurosci 13:333–337, homeodomain transcription factor Irx5 establishes Basic Res Cardiol 102:416–428, 2007.
2010. the mouse cardiac ventricular repolarization gradi- 77. Dhamoon AS, Pandit SV, Sarmast F, et al: Unique
41. Radicke S, Cotella D, Graf EM, et al: Expression ent. Cell 123:347–358, 2005. Kir2.x properties determine regional and species
and function of dipeptidyl-aminopeptidase-like 59. Rossow CF, Dilly KW, Santana LF: Differential differences in the cardiac inward rectifier K+
protein 6 as a putative b-subunit of human cardiac calcineurin/NFATc3 activity contributes to the Ito current. Circ Res 94:1332–1339, 2004.
transient outward current encoded by Kv4.3. transmural gradient in the mouse heart. Circ Res 78. Panama BK, McLerie M, Lopatin AN: Heteroge-
J Physiol 565:751–756, 2005. 98:1306–1313, 2006. neity of IK1 in the mouse heart. Am J Physiol Heart
42. Nadal MS, Ozaita A, Amarillo Y, et al: The CD26- 60. Rosati B, Grau F, McKinnon D: Regional variation Circ Physiol 293:H3558–H3567, 2007.
related dipeptidyl aminopeptidase-like protein in mRNA transcript abundance within the ven- 79. Panama BK, Lopatin AN: Differential polyamine
DPPX is a critical component of neuronal A-type tricular wall. J Mol Cell Cardiol 40:295–302, 2006. sensitivity in inwardly rectifying Kir2 potassium
K+ channels. Neuron 37:449–461, 2003. 61. Guo W, Malin SA, Johns DC, et al: Modulation of channels. J Physiol 571:287–302, 2006.
43. Jerng HH, Qian Y, Pfaffinger PJ: Modulation of Kv4-encoded K+ currents in the mammalian myo- 80. Babenko AP, Aguilar-Bryan L, Bryan J: A view of
Kv4.2 channel expression and gating by dipeptidyl cardium by neuronal calcium sensor-1. J Biol sur/KIR6.X, KATP channels. Ann Rev Physiol
peptidase 10 (DPP10). Biophys J 87:2380–2396, Chem 277:26436–26443, 2002. 60:667–687, 1998.
2004. 62. Radicke S, Cotella D, Graf EM, et al: Functional 81. Suzuki M, Li RA, Miki T, et al: Functional roles of
44. Ren X, Hayashi Y, Yoshimura N, et al: Transmem- modulation of the transient outward current Ito by cardiac and vascular ATP-sensitive potassium
brane interaction mediates complex formation KCNE beta-subunits and regional distribution in channels clarified by Kir6.2-knockout mice. Circ
between peptidase homologues and Kv4 channels. human non-failing and failing hearts. Cardiovasc Res 88:570–577, 2001.
Mol Cell Neurosci 29:320–332, 2005 Res 71:695–703, 2006. 82. Pu J, Wada T, Valdivia C, et al: Evidence of KATP
45. Radicke S, Cotella D, Bortoluzzi A, et al: 63. London B, Wang DW, Hill JA, et al: The transient channels in native cardiac cells without SUR.
Dpp10—A new putative regulatory β subunit of outward current in mice lacking the potassium Biophys J 80:625–626, 2001.
Ito in failing and non-failing human heart. channel gene Kv1.4. J Physiol 509:171–182, 1998. 83. Seghers V, Nakazaki M, DeMayo F, et al: SUR1
Circulat. 64. Guo W, Xu H, London B, et al: Molecular basis of knockout mice. A model for K(ATP) channel-
46. Fiset C, Clark RB, Shimoni Y, et al: Shal-type transient outward K+ current diversity in mouse independent regulation of insulin secretion. J Biol
channels contribute to the Ca2+-independent tran- ventricular myocytes. J Physiol 521:587–599, 1999. Chem 275:9270–9277, 2000.
sient outward K+ current in rat ventricle. J Physiol 65. Guo W, Li H, London B, et al: Functional conse- 84. Koster JC, Knopp A, Flagg TP, et al: Tolerance for
500:51–64, 1997. quences of elimination of I(to,f) and I(to,s): early ATP-insensitive K(ATP) channels in transgenic
47. Guo W, Li H, Aimond F, et al: Role of heteromul- afterdepolarizations, atrioventricular block, and mice. Circ Res 89:1022–1029, 2001.
timers in the generation of myocardial transient ventricular arrhythmias in mice lacking Kv1.4 and 85. Lesage F, Lazdunski M: Molecular and functional
outward K+ currents. Circ Res 90:586–593, 2002. expressing a dominant-negative Kv4 alpha subunit. properties of two-pore-domain potassium chan-
48. Johns DC, Nuss HB, Marban E: Suppression of Circ Res 87:73–79, 2000. nels. Am J Physiol Renal Physiol 279:F793–F801,
neuronal and cardiac transient outward currents by 66. Pond AL, Scheve BK, Benedict AT, et al: Expres- 2000.
viral gene transfer of dominant-negative Kv4.2 sion of distinct ERG proteins in rat, mouse, and 86. Goldstein SA, Bockenhauer D, O’Kelly I, et al:
constructs. J Biol Chem 272:31598–31603, 1997. human heart. Relation to functional I(Kr) channels. Potassium leak channels and the KCNK family of
49. Barry DM, Xu H, Schuessler RB, et al: Functional J Biol Chem 275:5997–6006, 2000. two-P-domain subunits. Nat Rev Neurosci 2:175–
knockout of the transient outward current, 67. Jones EM, Roti Roti EC, Wang J, et al: Cardiac IKr 184, 2001.
long-QT syndrome, and cardiac modeling in mice channels minimally comprise hERG 1a and 1b sub- 87. Terrenoire C, Lauritzen I, Lesage F, et al: A
expressing a dominant-negative Kv4 alpha subunit. units. J Biol Chem 279:44690–44694, 2004. TREK-1-like potassium channel in atrial cells
Circ Res 83:560–567, 1998. 68. Phartiyal P, Jones EM, Robertson GA: Hetero- inhibited by beta-adrenergic stimulation and acti-
50. Xu H, Li H, Nerbonne JM: Elimination of the meric assembly of human ether-a-go-go-related vated by volatile anesthetics. Circ Res 89:336–332,
transient outward current and action potential pro- gene (hERG) 1a/1b channels occurs cotranslation- 2001.
longation in mouse atrial myocytes expressing a ally via N-terminal interactions. J Biol Chem 88. Barbuti A, Ishii S, Shimizu T, et al: Block of the
dominant negative Kv4 alpha subunit. J Physiol 282:9874–9882, 2007. background K+ channel TASK-1 contributes to
519:11–21, 1999. 69. Finley MR, Li Y, Hua F, et al: Expression and coas- arrhythmogenic effects of platelet-activating factor.
51. Guo W, Jung WE, Marionneau C, et al: Targeted sociation of ERG1, KCNQ1, and KCNE1 potas- Am J Physiol Heart Circ Physiol 282:H2024–
deletion of Kv4.2 eliminates I(to,f) and results in sium channel proteins in horse heart. Am J Physiol H2030, 2002.
electrical and molecular remodeling, with no evi- 283:H126–H138, 2002. 89. Besana A, Barbuti A, Tateyama MA, et al: Activa-
dence of ventricular hypertrophy or myocardial 70. Ehrlich JR, Pourrier M, Weerapura M, et al: tion of protein kinase C epsilon inhibits the two-
dysfunction. Circ Res 97:1342–1350, 2005. KvLQT1 modulates the distribution and biophysi- pore domain K+ channel, TASK-1, inducing
52. Niwa N, Wang W, Sha Q, et al: Kv4.3 is not cal properties of HERG. A novel alpha-subunit repolarization abnormalities in cardiac ventricular
required for the generation of functional I(to,f) interaction between delayed rectifier currents. J myocytes. J Biol Chem 279:33154–33160, 2004.
channels in adult mouse ventricles. J Mol Cell Biol Chem 279:1233–1241, 2004. 90. Putzke C, Wemhoner K, Sachse FB, et al: The
Cardiol 44:95–104, 2008. 71. Liu GX, Derst C, Schlichthorl G, et al: Compari- acid-sensitive potassium channel TASK-1 in rat
53. Dixon JE, Shi W, Wang HS, et al: Role of the son of cloned Kir2 channels with native inward cardiac muscle. Cardiovasc Res 75:59–68, 2007.
Kv4.3 K+ channel in ventricular muscle. A molecu- rectifier K+ channels from guinea-pig cardiomyo- 91. Long SB, Tao X, Campbell EB, et al: Atomic struc-
lar correlate for the transient outward current. Circ cytes. J Physiol 532:115–126, 2001. ture of a voltage-dependent K+ channel in a lipid
Res 79:659–668, 1996. 72. Zaritsky JJ, Redell JB, Tempel BL, et al: The con- membrane-like environment. Nature 450:376–
54. Kong W, Po S, Yamagishi T, et al: Isolation and sequences of disrupting cardiac inwardly rectifying 382, 2009.
characterization of the human gene encoding Ito: K(+) current (IK1) as revealed by the targeted dele- 92. Strop P, Bankovich AJ, Hansen KC, et al: Structure
further diversity by alternative mRNA splicing. Am tion of the murine Kir2.1 and Kir2.2 genes. of a human A-type potassium channel interacting
J Physiol 275:H1963–H1970, 1998. J Physiol 533:697–710, 2001. protein DPPX, a member of the dipeptidyl
55. Rosati B, Grau F, Rodriguez S, et al: Concordant 73. Tristani-Firouzi M, Jensen JL, Donaldson MR, aminopeptidase family. J Mol Biol 343:1055–1065,
expression of KChIP2 mRNA, protein and tran- et al: Functional and clinical characterization of 2004.
sient outward current throughout the canine ven- KCNJ2 mutations associated with LQT7 (Ander- 93. Pioletti M, Findeisen F, Hura GL, et al: Three-
tricle. J Physiol 548:815–822, 2003. sen Syndrome). J Clin Invest 110:381–388, 2002. dimensional structure of the KChIP1-Kv4.3 T1
56. Teutsch C, Kondo RP, Dederko DA, et al: Spatial 74. Priori SG, Pandit SV, Rivolta I, et al: A novel form complex reveals a cross-shaped octamer. Nat Struct
distributions of Kv4 channels and KChIP2 of short QT syndrome (SQT3) is caused by a Mol Biol 13:987–995, 2006.
Structural and Molecular Bases
of Cardiac Inward Rectifier
Potassium Channel Function 4
Anatoli N. Lopatin and Justus M.B. Anumonwo
the resulting channels depends partially or solely on interactions

CHAPTER OUTLINE
with additional (accessory) proteins.2 From a functional perspec-
Background 33 tive, there are three distinct classes of cardiac Kir channels (IK1,
IKACh, and IKATP). As will be discussed later in this chapter, based
A Family of Genes Encode Inward Rectifier
on the degree of inward rectification the underlying channels
Potassium Channels 33 can be considered as strong (IK1; KACh) or weak (KATP) inward
Classical Cardiac Inward Rectifier Potassium Channels 33 rectifiers.2
There is ample molecular and electrophysiologic evidence for
Acetylcholine-Activated Potassium Channels 37 the expression of Kir2, Kir3, and Kir6 subfamily members in
ATP-Sensitive Potassium Channels 39 myocardial tissue, subunits representing the molecular correlates
of IK1, IKACh, and IKATP, respectively. Given that the other members
Conclusion 40 of Kir family (Kir1, Kir4, Kir5, and Kir7) are thought to be
primarily important for K+ transport in other tissues, they will
not be discussed further in this chapter.
Background
Inwardly rectifying potassium (Kir) channels are important for
Classical Cardiac Inward Rectifier
stabilizing the resting membrane potential, establishing the Potassium Channels
threshold of excitation, and modulating the repolarization phase
of the cardiac action potential.1 Inward rectification is a process Structure and Function
in which the conductance of the Kir channel increases with
membrane hyperpolarization, but decreases with depolarization Kir2 Subfamily Underlies Cardiac IK1
to potentials positive to the potassium equilibrium potential. The Kir2 subfamily consists of five members (Kir2.1-Kir2.5), of
In essence, Kir channels behave as “bio-diodes,” preferentially which only Kir2.1-Kir2.4 are expressed in the mammalian heart.
passing current in one direction. The molecular basis of rectifica- Kir2.5 is expressed in the fish. As shown in Table 4-1, the follow-
tion in Kir channels is a physical occlusion of the ion permeation ing genes encode the mammalian cardiac Kir channels: KCNJ2,
pathway by depolarization-induced movement of intracellular KCNJ12, KCNJ4, and KCNJ14, for Kir2.1 through Kir2.4,
cations, such as magnesium and polyamines.1 Only few Kir chan- respectively. There is evidence that in the mammalian heart
nels, however, display strong rectification.2 Strong rectification Kir2.1-Kir2.3 isoforms are expressed in cardiac myocytes, and
enables the Kir channels to stabilize the resting membrane poten- that Kir2.4 is probably only expressed in neuronal cells. It is well
tial, as well as to protect the cell from an excessive loss of K+ ions established that members of the Kir2 subfamily underlie IK1,
during the plateau phase of an action potential.1 The molecular although the subunit composition varies among species and cell
correlates of these channels are primary subunits encoded by the types, and channel complexes are likely formed as hetero-
KCNJ family of genes.3 Mutations in KCNJ genes have been tetrameric structures.
associated with various channelopathies,2 demonstrating the
importance of Kir channels in normal cardiac excitation. This Crystal Structure of Kir2 Channels
chapter will focus on three well-studied Kir channels in the myo- In recent years, x-ray crystallographic structures of both bacterial
cardial cells: the classical inward rectifier potassium channels and mammalian homologs of several Kir channels have been
(IK1), the acetylcholine-activated potassium channels (KACh), and obtained.5 Figure 4-2 highlights some important and common
the adenosine triphosphate (ATP)-sensitive potassium channels features of Kir channel structure based on the results of work
(KATP). Additional information on the topics covered in this with Kir2.1 and Kir2.2 mammalian channels.6 It is now firmly
chapter can be obtained from recent review articles.2,4 established that Kir channels are tetramers of distinct subunits,
each having two transmembrane domains (M1 and M2), rela-
tively small N-terminal, and large C-terminal cytoplasmic
domains, and a pore-forming structure between M1 and M2
A Family of Genes Encode Inward Rectifier (see Figure 4-2). The pore structure contains pore helix directed
Potassium Channels toward the conduction pathway and the characteristic GYG (or
GFG) motif, also known as K+ channel signature sequence, that
Channels belonging to the Kir family are structurally and func- contributes to the selectivity filter in all potassium channels. The
tionally different from voltage-gated potassium channels.2,4 The M1 and M2 transmembrane domains in each subunit are arranged
genes that encode Kir channels are ascribed the KCNJ nomen- as an antiparallel coiled-coil and make contact with each other.
clature and are categorized into seven subfamilies based on the Kir2 channels have a negatively charged amino acid (D172 in
gene products (Kir1-7; Table 4-1; Figure 4-1). In general, Kir Kir2.1) located in approximately the middle of the pore, which
channels consist of homomeric or heteromeric complexes of the has a critical role in the phenomenon of inward rectification
respective Kir subunits, but as will be discussed, functionality of discussed later.
33
allowing the helixes to bend and change the size of the pore at
Table 4-1. Diversity of α-Subunit Proteins in the Family of Inward
the bundle crossing.
Rectifier Potassium Channels
The so-called slide helix formed by the N-terminus region
Subfamily Protein Gene Human Mouse Current just preceding the M1 helix is another unique and important
regulatory feature of both bacterial and mammalian Kir chan-
Kir1 Kir1.1 KCNJ1 11q25 ?
nels.7 This helix intercalates between the inner leaflet of the
Kir2 Kir2.1 KCNJ2 17q23 11 IK1 plasma membrane and cytosol likely because of its amphiphilic
Kir2.2 KCNJ12 17p11.2 11 IK1 nature. The important functional role of the slide helix is high-
lighted by the fact that many loss-of-function mutations associ-
Kir2.3 KCNJ4 22U IK1 ated with Andersen-Tawil syndrome (LQT7) are located in the
Kir2.4 KCNJ14 19q13.4 ? slide helix.
A large C-terminal domain provided by four Kir2.x subunits
Kir3 Kir3.1 KCNJ3 2 IKACh
consists primarily of β-strands and potentially strongly interacts
Kir3.2 KCNJ6 21q22 with a smaller N-terminal domain. As shown in Figure 4-2, C-
KCNJ7 16 terminal domain forms a large intracellular vestibule, approxi-
mately 30 Å in length, for easy ion passage and likely provides
Kir3.3 KCNJ9 1q21 1 binding sites for various intracellular agents. The cytoplasmic
Kir3.4 KCNJ5 11q25 9 IKACh domain harbors a number of residues (e.g., E224 and E299 in
Figure 4-2) known to contribute to inward rectification.
Kir4 Kir4.1 KCNJ10 1q21 1
A membrane phospholipid phosphatidylinositol 4,5-
Kir4.2 KCNJ15 21q22 16 bisphosphate (PIP2) is an important structural and regulatory
Kir5 Kir5.1 KCNJ16 17q25 component of Kir channels.8 In the membrane, PIP2 likely inter-
acts with hydrophobic amino acids on both M1 and M2 trans-
Kir6 Kir6.1 KCNJ8 12p11.1 6 IKATP membrane helices as well as with a well-conserved (among Kir
Kir6.2 KCNJ11 11p15 IKATP channels) RWR motif located just at the end of the M1 domain
(see Figure 4-2). The interaction of PIP2 with a specific region
Kir7 Kir7.1 KCNJ13 2 1 ?
in the CTD leads to an approximately 6 Å translation of the
(Modified from Nerbonne JM, Nichols CG, Schwarz TL, et al: Genetic entire CTD toward the membrane associated with the movement
manipulation of cardiac K+ channel function in mice: what have we learned, of the M1 helix, which ultimately leads to the opening of the
and where do we go from here? Circ Res 89:944–956, 2001). gate.6
Mechanism of Polyamine-Induced Rectification

The key experiments conducted in mid-1990s with the first
cloned members of the Kir2 subfamily have clearly demonstrated
that almost all essential properties of classical strong inward rec-
Kir1 Kir2 Kir3 Kir4 Kir5 Kir6 Kir7 tification could be explained by a voltage-dependent block of the
channel by ubiquitous intracellular organic cations, the poly-
4.1 5.1 amines.9 Micromolar concentrations of free polyamines (primar-
SU
R
SU
5.1 4.1
ily spermine and spermidine) are sufficient to reproduce the
degree of rectification observed in native cells. The strength of
R
SU
SU
R
rectification varies among the different members of the Kir sub-

family, and although every Kir channel shows some degree of
SUR1 SUR2 Kir6.1 Kir6.2 KATP
inward rectification, they can be broadly grouped as either
strongly rectifying (Kir2 and Kir3) or weakly rectifying (Kir1 and
Kir6) channels. It is well established that the time-dependent
Kir3.1 Kir3.2 Kir3.3 Kir3.4 KACh
activation of strong, inward rectifiers during membrane hyper-
polarization reflects the exit of polyamines (primarily sperimine)
from the pore.
Kir2.1 Kir2.2 Kir2.3 Kir2.4 IK1
Rectification Properties Are Related to Electrostatic
Figure 4-1. The family of inward rectifier potassium channels. All members of the
Kir family share significant structural similarity, but only Kir2 and Kir3 subfamilies
Interactions in the Cytoplasmic Pore of a Kir Channel
represent channels carrying classical strongly rectifying currents. Four members of Early work with cloned Kir channels established that strong
each Kir2 and Kir3 subfamilies were cloned in mammals. Heteromeric assemblies inward rectification by intracellular polyamines depends on three
of Kir2.1, Kir2.2, and Kir2.3 subunits underlie IK1 current, and heteromeric assembly negatively charged residues located in the second transmembrane
of Kir3.1 and Kir3.4 subunits underlies IKACh current. Weakly rectifying KATP channels domain (D172; the rectification controller; see Figure 4-2) and
are composed of pore-forming Kir6.1 and Kir6.2 subunits and auxiliary SUR1and in the C-terminal tail (E224 and E299; Kir2.1 amino acid [aa]
SUR2 subunits. numbering; see Figure 4-2). Positively charged polyamines enter
(Modified from Anumonwo JM, Lopatin AN: Cardiac strong inward rectifier potassium
the Kir channel pore to physically occlude it, a process aided by
channels. J Mol Cell Cardiol 48:45–54, 2010.)
negative charges provided by aspartate and glutamate residues.
Both electrophysiologic and structural data are consistent with
the idea that channel block occurs in two sequential steps. A
There are at least three distinct regions of the intramembrane weakly voltage-dependent (shallow) blocking step involves entry
part of the pore: selectivity filter, a centrally located water filled of polyamine into the Kir2.1 pore at a site provided by a ring of
cavity of approximately 10 Å in diameter, and the narrowing of negative charges (E224 and E299) in C-terminus. A more strongly
the pore at the cytoplasmic side as the pore-lining M2 helixes voltage-dependent blocking step reflects the movement of poly-
come closer to each other (known as bundle-crossing). M2 helixes amine to its deep binding site near D172 residue. It is also
also possess a highly conserved glycine residue (a “hinge”; G168 believed that the strong voltage dependence of the polyamine
in Kir2.1) that likely contributes to the channel gating by block arises not only from the high valency (z) of polyamines (z≈4
Structural and Molecular Bases of Cardiac Inward Rectifier Potassium Channel Function 35
Pore helix
M2 helix
4
M2
Selectivity Inner helix G168
filter
M1 M1 helix
Outer helix
RWR
Water cavity D172
PIP2
Bundle Slide helix
crossing G-loop
N N E299
E224
Spm
Cytoplasmic pore
C C
C C
Figure 4-2. Architecture of a typical Kir channel. X-ray structure of chicken Kir2.2 crystallized in the presence of PIP2 (Accession code 3SPI, available in the Protein Data Bank;
from Hansen et al.6). Only two opposing subunits are shown for clarity. Right, The structure is rotated approximately 90 degrees to better present the intracellular ion
conduction pathway and bound molecules of PIP2. Numbering of the amino acids corresponds to that in Kir2.1. Red circles represent K+ ions in the selectivity filter. Blue
circles show the location of three residues most critical for inward rectification. Spermine molecule (Spm) is shown in the middle at the same scale as the Kir2.2 structure.
Analysis and visual presentation of the structures were performed using DeepView Swiss-PdbViewer and PyMol software.
for spermine) but also from a displacement (push) of K+ ions Kir2.1 and Kir2.3 channels. The voltage dependence of steady-
through the pore.10 Among the polyamines, spermine is the most state rectification is different between Kir2.x channels, as are
voltage-dependent blocker and also has the highest potency for single-channel conductance and kinetics properties of the cur-
blocking the channel. rents. In particular, polyamine unblock (activation) at negative
A characteristic feature of Kir2.1 and Kir3 channels is a flex- membrane potentials in Kir2.3 channels is several-fold slower
ible cytoplasmic pore-facing G-loop (see Figure 4-2) that forms than in Kir2.1 and Kir2.2 channels. Single-channel conductance
a girdle around the central axis of the Kir channel.11 It was is smallest in Kir2.3 (approximately 10 pS), medium in Kir.2.1
estimated that this girdle constricts the ion permeation pathway (approximately 25 pS) and largest in Kir2.2 (approximately 35 pS).
to approximately 3 Å. Mutations in the G-loop were shown
to disrupt inward rectification. In addition to the previously
described E224 and E299 residues, it was shown that A255 and Cellular and Membrane Localization
A259 located farther away from the pore axis are also involved in
channel rectification. Electrophysiologic experiments using cys- IK1 currents, and thus the underlying Kir2.x subunits, display a
teine modifications in the pore region of a mutant Kir6.2 channel distinct regional distribution in the heart. It has been shown that
(N160D; equivalent to D172 in Kir2.1) showed that spermine inward IK1 is generally more prominent in ventricular myocytes
binds at a deep site beyond the rectification controller residue and Purkinje fibers and is significantly smaller in atrial myocytes
D172, a site that is close to the extracellular mouth of the pore. (with one known exception of the mouse heart). The density of
In another study, refinement of the crystal structures of bacterial inward IK1 currents (i.e., currents normalized to membrane
KirBac1.1 and KirBac3.1 allowed identification of the shallow capacitance) is small in pacemaker cells of the sinoatrial node in
polyamine binding sites at the cytoplasmic interface between the mice and rats and essentially undetectable in the rabbit sinoatrial
two subunits.12 These observations notwithstanding, the precise nodes. IK1 cannot be detected in atrioventricular node of rabbit
mechanism involved in rectification is still being worked out, and but relatively large IK1 currents can be recorded in the guinea-pig
the exact location of polyamine binding sites in Kir channels is atrioventricular node. Moreover, the density of inward IK1 varies
still a controversial issue. across the ventricular myocardium. For example, in the mouse
heart inward, IK1 is larger in apical myocytes compared with
Differential Properties of Kir2.x Subfamily epicardial cells, and IK1 is larger in right ventricular than in left
Consistent with their overall significant sequence homology, all ventricular myocytes.
Kir2 channels have the three mentioned residues (D172, E224, Molecular biologic studies are also consistent with location-
and E299) important for rectification at the equivalent positions. dependent expression of specific Kir2 isoforms. In particular,
Recent studies, however, unexpectedly showed that rectification real-time reverse-transcriptase polymerase chain reaction
properties are rather different in the three Kir2 isoforms.13 Spe- analysis of Kir2 transcripts in the human heart showed the follow-
cifically, the data show that Kir2.2 channels display significantly ing relative expression levels: in Purkinje fibers, Kir2.1 > Kir2.3 >
stronger voltage dependence of rectification than that observed in Kir2.2; and in the right ventricle, Kir2.1 > Kir2.2 > Kir2.3; and
the sequence was reversed in right atrium, Kir2.3 > Kir2.2 > intracellular pH (pHi) is dependent on the species and the type
Kir2.1.14 Information on expression patterns of Kir2 subunits can of tissue. For example, rat and guinea pig ventricular IK1 is not
also be gleaned from functional data using the unique properties sensitive to physiologically relevant changes in pHi. In contrast,
of corresponding channels. For example, in cardiac atrial and IK1 in sheep ventricular myocytes is inhibited by intracellular H+
ventricular myocytes, unitary conductance values display a wide with pKa of approximately 7.4. The difference in pH sensitivity
spectrum ranging from 10 to 15 pS (as in Kir2.3 channels) to as is likely due to differences in subunit composition of IK1. For
high as 40 to 45 pS (as in Kir2.2 channels). example, channels composed of Kir2.3 subunits exhibit strong
Evidence shows that IK1 channels are located not only in the sensitivity to pHi within a physiologically relevant range, whereas
non–T-tubular component of sarcolemma of ventricular myo- Kir2.1 and Kir2.2 channels are relatively insensitive to H+.
cytes, but also in the intercalated discs and in the T-tubules. For Pharmacologic tools for modulation of Kir2.x and IK1 channels
example, accumulation and depletion of K+ only in T-tubules lead are limited. The most useful research tool for studying Kir2
to changes in whole-cell IK1.15 IK1 accumulation/depletion phe- channels is Ba2+ (inhibitory concentration of 50% [IC50] for
nomena are not observed in atrial cells, which essentially lack inward currents, 0.5 to 10 µM). Ba2+ action is subunit dependent.
T-tubules and in ventricular myocytes, in which T-tubules are For example, Ba2+ blocks Kir2.2 channels fivefold to sevenfold
removed by osmotic shock.15 more efficiently than Kir2.1 channels. In the early 1990s, a com-
Alternatively, Kir2.1, Kir2.2, and Kir2.3 subunits were pound RP 58866 and its active enantiomer, terikalant, were
localized to the T-tubular membrane using immunolabeling shown to be selective blockers of IK1 (IC50, 5 to 20 µM). However,
with specific antibodies. Intercalated discs were not studied elec- later studies revealed that terikalant also inhibits many other K+
trophysiologically in regard to the presence of IK1 channels, channels, some with even higher potency (IC50 in submicromolar
although labeling with various Kir2 antibodies can clearly be range for IKr channels). Similarly, it was found that LY97119
observed at this location. Moreover, there is evidence that in compound (LY, a tertiary homolog of clofilium) blocks IK1 in the
canine ventricular and atrial myocytes, Kir2.3 subunits are low micromolar range, but it also blocked Ito (transient outward
expressed at higher levels in intercalated disc membranes relative current) at submicromolar concentrations. Perhaps chloroquine
to T-tubules. (an antimalarial drug) is the most potent blocker of IK1 with an
IC50 of approximately 0.5 µM. However, chloroquine does not
discriminate among IK1, IKACh, and IKATP, and it has been shown
Pharmacology and Regulation to affect other currents (e.g., INa) in the low micromolar range.
Activators of Kir2 and IK1 channels were also described. In par-
Kir2.x and IK1 channels can be regulated in a number of ways.2 ticular, flecainide, a widely used antiarrhythmic drug, increases
Most studies on adrenergic stimulation show that inward IK1 cur- Kir2.1 currents by approximately 50% at a concentration of
rents are suppressed by activation of both α and β receptors, 1 µM, but has no effect on current carried through Kir2.2 and
although opposite effects were also described. In addition, adren- Kir2.3 channels.19 Arachidonic acid and the antiinflammatory
ergic regulation is clearly dependent on the type of receptors and agent tenidap were shown to specifically activate Kir2.3 channels,
subunit composition of the channel. with a greater than twofold maximum increase in inward
Both isoproterenol and forskolin inhibit IK1 in human ven- current with IC50 of approximately 0.5 and 1.3 µM, respectively.
tricular myocytes, suggesting involvement of protein kinase A Zacopride, a gastrointestinal prokinetic drug, was recently found
(PKA)-mediated phosphorylation of underlying Kir2.x subunits. to be a selective IK1 channel agonist. The activating effect,
Molecular details of the phenomenon, however, are contradic- however, is modest, with a maximum increase in IK1 of approxi-
tory. For example, it has been shown that the application of a mately 34% at a concentration of 1 µM.
catalytic subunit of cyclic adenosine monophosphate (cAMP)
dependent PKA leads to activation of Kir2.1 channels expressed
in Xenopus oocytes, but to Kir2.1 inhibition when the channels Channelopathies
they are expressed in a mammalian cell line (COS-7). The data
on PKA regulation of native IK1 channels is limited and somewhat There are at least four known channelopathies associated with
controversial; however, most of the studies show that IK1 channels IK1 channels, all originating from mutations in KCNJ2: ATS,
are inhibited by exposure of the cytosolic side of the membrane short QT (SQT) syndrome, familial atrial fibrillation (FAF) and
to purified catalytic subunit of PKA. Studies in exogenous- catecholaminergic polymorphic ventricular tachycardia (CPVT)
expressing systems also showed the involvement of PKC in nega- (Figure 4-3).2
tive regulation of Kir2.1 channels. Consistent with the latter, ATS is characterized by a triad of pathologic clinical pheno-
experiments using human atrial myocytes show that α1-adrenergic types including morphogenesis and functioning of skeletal and
stimulation likely reduces IK1 via a PKC-dependent mechanism. cardiac muscle. One of the prominent features of ATS is cardiac
Kir channels are also targets for phosphorylation by tyrosine electrical abnormalities, including brief episodes of ventricular
kinases. In Kir2.1 channels, the site of downregulation was tar- tachycardia, multi-focal ventricular ectopy induced by adrenergic
geted to a single Y242 residue in the C-terminus.16 stimulation and prolongation of QT interval. ATS1 is used to
PIP2 is an important component in membrane-delimited differentiate ATS patients carrying mutations in KCNJ2. Many
second messenger signaling system and a powerful activator of ATS1 patients display clear QT prolongation and have been
Kir channels.17 There is significant evidence that PIP2 regulates referred to, perhaps questionably in some cases, as LQT7.
the channel gating primarily through specific electrostatic inter- LQT7/ATS1 mutations are numerous (see Figure 4-3) and
actions with the cytosolic part of the channel (see Figure 4-2).6 result in nonfunctional channels when exogenously expressed in
It is also clear that various properties of Kir2 channels are heterologous system.20 Because affected patients are heterozy-
modulated by PIP2. For example, pH sensitivity of Kir2.3 chan- gous for the mutant and wild type (WT) alleles, and the channel
nels is strongly dependent on the strength of channel-PIP2 is composed of four subunits, a dominant-negative effect of
interaction. mutated subunit can lead to reduced IK1 current in cardiac myo-
Various common cations, such as Ca2+ and H+, contribute to cytes (and other cells). The dominant-negative effect of LQT7/
regulation of Kir2 channels as well.2 Intracellular Ca2+ blocks IK1 ATS1 mutations has been demonstrated using cloned channels.
channels in a voltage-dependent manner. There is evidence that For example, coexpression of WT and D71V mutants results in
a transient increase in intracellular Ca2+ during action potential barely measurable inward Kir2 currents compared with those
can lead to significant blockage of IK1.18 Regulation of IK1 by produced by expression of WT subunits alone. The magnitude
LQT7/ATS1 G144A/S G146D/S display a normal QT interval in contrast to patients affected by
4
SQT3 S136F C154F D172N mutation.
CPVT Other V123G Mutations in KCNJ2 were also linked to another type of
FAF excitability disorder—catecholaminergic polymorphic ventricu-
lar tachycardia (CPVT).24 CPVT is a heritable disorder charac-
terized by frequent ventricular arrhythmias and sudden cardiac
death associated with physical activity or adrenergic stimulation.
C101R Four KCNJ2 mutations associated with CPVT have been identi-
∆ 163-164
∆ 95-98 M1 M2 fied (see Figure 4-3). Three of them are located in the N-terminus
D172N and one in the distal C-terminus of Kir2.1. The electrical phe-
V931 notype of these mutations included prominent U waves, ventricu-
lar ectopy, and polymorphic ventricular tachycardia. In contrast
to closely located mutations associated with LQT7/ATS1 syn-
E303K drome, CPVT mutations were not associated with dysmorphic
R82Q R82W P186L V302M features or with skeletal muscle abnormalities. Electrophysio-
D78G/Y R189I M301K logic analysis of exogenously expressed mutants showed that they
T75A/M/R T75M T192A G300D/V T305A did not produce any measurable current. In addition, two of the
M307I mutations (T75M, R82W) displayed strong dominant negative
T74A G211T effects when coexpressed with WT channels. T305A mutation
D71/N/V G215D T309I did not exert a significant dominant-negative effect on inward
Y68D N216H R312C currents, but strongly increased the strength of inward rectifica-
∆ 314-315 tion, leading to a relative decrease in outward current. Finally,
R67W R67Q L217P
some mutations in KCNJ2 could not be easily classified into the
R218Q/W P351S
C54F four categories mentioned previously, and for this reason they are
N C referred to as other in Figure 4-3. For example, C54F mutation
Figure 4-3. Channelopathies of the classical inward rectifier channel, IK1, associ- was found in a patient who showed cardiac abnormalities
ated with mutations in Kir2.1 subunit. Mutant residues are color coded to represent (arrhythmia was not specified) upon corticosteroid intake.
the long QT (LQT7/ATS1; black), catecholaminergic polymorphic ventricular tachy-
cardia (CPVT; red), familial atrial fibrillation (FAF; green), and short QT3 (SQT3; blue).
(Modified from Anumonwo JM, Lopatin AN: Cardiac strong inward rectifier potassium Acetylcholine-Activated Potassium Channels
channels. J Mol Cell Cardiol 48:45–54, 2010.)
Structure and Function

of dominant negative effect, however, is variable in different
mutants. In the heart, Kir3.1/Kir3.4 channels are responsible for the
SQT syndrome is characterized by an abnormally short QT effects of acetylcholine and adenosine, and they act through a
interval (less than 300 ms) and increased risk of having fibrillation coupling mechanism involving a receptor, a G protein (Go/Gi
and sudden death. Currently, three forms of SQT syndrome have family), and the potassium channel.2,25 Because channel gating
been described. SQT1 and SQT2 syndromes result from muta- requires a G protein, Kir3.1/Kir3.4 channels are considered a
tions in genes underlying two voltage-gated potassium channels, type of KG channel.4 For channel activation, the G-protein–
HERG (KCNH2) and IKs (KCNQ1). A third variant of SQT coupled receptor (Figure 4-4) can be a muscarinic (M2) or a
syndrome (SQT3) originating from mutation in KCNJ2 has purinergic (P1) receptor, which are activated by acetylcholine or
been described recently.21 Genetic analysis revealed a charge- adenosine, respectively. The ultimate result of channel activation
neutralizing substitution (D172N) in the critical place of the is the opening of Kir3.1/Kir3.4 channels, which permits K+ efflux
channel responsible for strong inward rectification (see Figure and consequently hyperpolarizes the cell membrane.
4-2). Accordingly, coexpression of WT and D172N mutant As in a typical potassium channel, ion selectivity in Kir3.x
subunits showed decreased rectification of heteromeric channel. channels is conferred by the presence of the signature sequence
Computer simulations showed that reduced rectification of IK1 (T-X-G-Y/F-G),4,26 and the mechanism of rectification involves
can explain some of the characteristic features of an electrocar- an asparagine or an aspartate residue for interactions with the
diograph, such as tall and asymmetric T waves, observed in polyamines (details of rectification have been discussed previ-
affected patients. ously). Thus, the mechanism of rectification is similar to that
Recently, a second mutation in KCNJ2 associated with SQT described for Kir2 channels, and Kir3.x channels belong to the
syndrome has been identified (M301K).22 In contrast to D172N class of strong inward rectifier channels.2 Membrane topology of
mutation, exogenous expression of M301K subunits alone did not Kir3 channels is similar to that described for Kir2 channels (see
produce measurable currents. However, coexpression of both Figure 4-2).
WT and M301K subunits resulted in large K+ currents displaying Exactly how do Kir3.1/Kir3.4 channels and G proteins inter-
significantly reduced inward rectification (relatively larger act to cause channel opening? First, a brief discussion of G pro-
outward currents). Given that the M301K mutation resides on teins is necessary. G proteins are complexes consisting of an α
KCNJ2, a gene already associated with SQT3 syndrome, it is (molecular weight [MW], ~40,000), a β (MW, ~35,000) and a γ
reasonable to classify M301K as another SQT3 mutation. (MW, ~8000) subunit, which transduce signals from membrane
One study described association of single V93I mutation in receptors (e.g., muscarinic [M2]) to an effector, such as a KG
KCNJ2 with familial atrial fibrillation.23 Electrophysiologic channel.4,25 Four subfamily members of the guanosine diphos-
experiments with cloned channels showed, in particular, that phate (GDP) Gα subunit dictate selectivity of signaling (activa-
V93I mutation leads to a relative increase in the magnitude of tion of adenylate cyclase [Gαs], inhibition of adenylate cyclase
the outward current, or a decrease in the strength of inward [Gαi], and activation of phospholipase [Gq]).25 Normally, Gα is
rectification. Regarding the inward rectification, the effect of bound to GDP in the absence of an agonist, and the Gα/GDP
V91I mutation on Kir2.1 current resembles that found in D172N complex is coupled to the receptor and has low GTPase activity.
mutant channels. However, patients carrying the V93I mutation With vagal stimulation and ligand-receptor interaction, GDP is
GPCR K+ tissues.2,4 Following the cloning of genes underlying Kir3 chan-

(M2 or P1) Agonist nels in the early to middle 1990s, investigators have probed myo-
(ACh/Ado) cardial tissues to examine localization of the gene products.
Overall, these studies reported an abundance of Kir3.1 and Kir3.4
mRNA in nodal and atrial, but not in ventricular tissues, which
would be consistent with tissue distribution of IKACh. In one such
α β β study,28 a comprehensive analysis was performed using Western
γ γ blot and immunofluorescence to examine channel distribution in
sinus-nodal, atrial, and ventricular tissues, and showed similar
Giαβγ GDP
A GTP results of expression pattern across species (rat, ferret, and guinea
pig hearts). It was reported that, whereas there was minimal
Cell-attached Inside-out expression of Kir3.1 in the ventricles of all species tested, Kir3.1
and Kir 3.4 were highly expressed in atrial tissue of all the species.
ACh It was noted that although there were relatively high levels of
or
Ado
expression in the atria, significant quantitative differences in
Kir3.1 and Kir3.4 protein levels were found in the different
Cell species. Furthermore, it was demonstrated that Kir3.1 and M2
GTP 100 µM receptor colocalized in the sinoatrial node. In nodal and atrial
3 min tissue, immunofluorescence showed localization of Kir3.x/M2
ACh 1.1 µM receptors more on the outer (lateral) membranes than in T-tubular
–60 mV membranes.28 Similar quantitative differences in expression have
been reported for atrial versus ventricular tissues in the human
myocardium.28,29
GTP 100 µM
2.5 min
Ado 10 µM
–80 mV
Pharmacology and Regulation
B It is well established that Kir3.1/Kir3.4 are generally insensitive

to classical potassium channel blockers such as or 4-AP or TEA.4
Figure 4-4. Activation of Kir3.1/Kir3.4 channels by muscarinic (M2) and purinergic However, experiments in cardiac myocytes isolated from rabbit
(P1) G-protein coupled receptors (GPCRs). A, Membrane-delimited pathway for hearts show that the toxin, tertiapin, is a selective blocker of
activation of acetylcholine (Ach) and adenosine (Ado) gated Kir channels. Ligand-
GPCR interaction enhances GTP association with G-protein α-subunit, and results
IKACh.30 Channel block is highly potent, with an affinity in the
in the release of the β/γ-subunits for the activation of the Kir channel. Potassium nanomolar range. Cardiac Kir3.1/Kir3.4 channels are also inhib-
efflux hyperpolarizes the cell membrane. B, Requirement of GTP for Kir3.1/Kir3.4 ited by quinine, quinidine, and verapamil; however, affinities of
activation by ACh (top single channel traces) and Ado (bottom single channel the these drugs are in the micromolar range.4 Based on the results
traces). Holding potentials for experiments are annotated. Channel activity is of experiments in heterologous (oocyte) expression systems,
present in the cell-attached patch mode, with ACh or Ado present in the pipette current through Kir3.x channels is enhanced by “intoxicating”
(top inset). Patch excision (at arrow; inside-out patch configuration) loss of GTP concentrations of ethanol, and an approximately 43-aa stretch on
results in loss of channel activity. Subsequent application of GTP (100 µM; intracel- the C-terminus has been identified as important for the ethanol
lular side) restored channel activity. effect.31 Similarly, in a heterologous expression system, Kir3.x
(A, Modified from Breitwieser GE: GIRK channels: hierarchy of control. Am J Physiol channels displayed sensitivity to intracellular acidification, an
Cell Physiol 289:C509–C511, 2005. B, Modified from Hibino H, Inanobe A, Furutani K inhibitory effect that was dependent on histidine residues in the
et al: Inwardly rectifying potassium channels: their structure, function, and physiolog- N- and C-terminal regions of the channels.4 A variety of other
ical roles. Physiol Rev 90:291–366, 2010.) agents have also been shown to modulate Kir3.x channels. For
example, similar to all Kir channels, Kir3.1/Kir3.4 require PIP2
for channel activity.4 Curiously, however, the PIP2 effect is
enhanced by other factors such as the G-protein Gβγ complex, as
exchanged for GTP and results in the uncoupling of Gβγ from well as by intracellular cations. In general, Kir3.x channel activity
Gα (see Figure 4-4, A). The released Gβγ subunits in turn activate is also sensitive to mechanical stretch, can be modified by phos-
the Kir3.1/Kir3.4 channels. There is a critical requirement for phorylating agents, and is sensitive (negatively) to regulators of
GTP for channel activation (see Figure 4-4, B). A variety of G-protein signaling.4
experiments have determined the precise molecular interactions
involved in the Gβγ-induced Kir3.1/Kir3.4 channel gating, and
have shown that the cytoplasmic region of the channel is inti- Channelopathies
mately involved with gating. For example, crystal structure analy-
ses of the cytoplasmic region of Kir3.1 suggest that the C-terminus Over four decades ago, a mouse with a striking locomotor defi-
of two neighboring Kir3 channels subunits bind to each other, ciency (weaving) was described, and the defect has subsequently
and that an N-terminus is positioned between the two C-terminal been traced to a naturally occurring gain-in-function mutation
domains.27 The study also showed that for Kir3.1 channels, the in the Kir3.2 channel.32 Additional neurologic defects attendant
cytoplasmic residues L262, L333, and E336 are important in to this mutation earned the weaver mouse the title of the “most
gating, and the equivalent residues are H64 and L262 in Kir3.4 cantankerous rodent.”33 There is relatively little information
channels. available on inherited cardiac channelopathies associated with
Kir3.1/Kir3.4 channels. However, alterations in Kir3.1/Kir3.4
channel activity have been reported in certain cardiac rhythm
Cellular and Membrane Localization abnormalities.34 Electrical remodeling in atrial fibrillation patients
has been shown to increase the constitutively active component
There are tissue-dependent differences in expression of IKACh in of IKACh, which could cause an abbreviation of action potential
the myocardium, with a very high density of expression in nodal duration. More recently, a study was performed in a Long QT
syndrome large Chinese family (49 individuals) with autosomal- Kir2.1; N160D in Kir6.2) converts Kir6.2 into a strongly rectify-
4
dominant Long QT syndrome (LQTS).35 The locus of the ing channel.37 A defining property of KATP channels is their
LQTS-associated gene was mapped to chromosome 11q23.3- characteristic sensitivity to intracellular ATP (hence the name
24.3. A combination of biochemistry and cell electrophysiology ATP-dependent K+ channels). Under normal conditions, both
in heterologous expression systems was used to demonstrate that exogenously expressed Kir6.2/SUR2 channels and native chan-
a heterozygous G387R mutation on the Kir3.4 (KCNJ5) identi- nels in cardiac myocytes are inhibited by ATP in the micromolar
fied in all affected family members was responsible for reduced range (10 to 100 µM) by direct binding to the channel rather
channel expression on the sarcolemma. than through phosphorylation mechanisms. In the absence of
intracellular Mg2+ ions, adenosine diphosphate (ADP) and other
nucleotides also inhibit the channel. Modeling studies suggests
that the ATP binding pocket is located at the interface of the N
ATP-Sensitive Potassium Channels and C terminus of each Kir6.x subunit; therefore, the KATP
channel possesses four ATP binding sites.
Structure and Function The overall structure of auxiliary SUR subunit is presented in
Figure 4-5, B. It is believed that SUR interacts with Kir6.x sub-
Among all inward rectifiers, KATP channels are the most unique units to modulate channel gating through TMD0 domain and
in their molecular architecture.36 As in other Kir channels, the L0 linker region. Experimental data are consistent with the
ion conduction pathway is provided by a tetrameric arrangement idea that intracellular ATP induces dimerization of nucleotide-
of pore-forming subunits, but additional auxiliary regulatory sub- binding domains NBD1 and NBD2, converting them into a cata-
units are necessary for the channel to be fully functional (Figure lytically active site for Mg2+-dependent hydrolysis of ATP (leading
4-5, B). The two known isoforms of the pore-forming subunits to MgATP). Hydrolysis of ATP is followed by a conformational
are encoded by Kir6.1 (KCNJ8) and Kir6.2 (KCNJ11) genes. The change that is then transduced to the Kir6.x subunit; however,
two isoforms of auxiliary subunits are encoded by SUR1 (ABCC8) MgADP is an even more potent activator of the KATP channel.
and SUR2 (ABCC9) genes (the SUR name originates from “sul- The latter puts the hydrolysis hypothesis into question. Accord-
fonylureas,” a class of drugs known to inhibit KATP channels by ingly, it has been suggested that NBDs can be locked in a post-
acting on the auxiliary subunit). SUR2 gene can be alternatively hydrolytic state by MgADP and other nucleotides to sustain the
spliced at the very C-terminus (last 42 aa) leading to SUR2A and active state of the channel.
SUR2B isoforms. Genomic arrangements of SUR and Kir6 genes
are unique as well. Specifically, SUR1 is followed by Kir6.2 in a
close proximity on chromosome 11pp15.1, whereas while SUR2 Cellular and Membrane Localization
is followed by Kir6.1 on chromosome 12p12.1. The consequences
of this arrangement regarding the regulation at a genomic level KATP channels are found in virtually every kind of cardiac tissues,
are not clear at this time. Although heteromeric assembly of but they are most prominent in cardiac myocytes and smooth
Kir6.x subunits produces functional channels in vitro, it remains muscle cells. Detailed experimental analysis reveals significant
unclear whether heteromeric Kir6.x complexes exist in native differences in various properties of native KATP channels, suggest-
tissues. In contrast, both SUR1/Kir6.2 and SUR2/Kir6.2 chan- ing different subunit organization in every case.38 A significant
nels likely exist in native tissues. Membrane topology and general amount of evidence suggest that ventricular sarcolemmal KATP
organization of Kir6.x subunits is highly similar to that in well- channels are likely composed of Kir6.2 and SUR2A subunits. In
characterized Kir2.x channels (see Figure 4-2). particular, functional sarcolemmal KATP channels are absent in
KATP channels display weak rectification; however, it has been Kir6.2 knockout mice, and various channel properties (including
shown that just a single amino acid substitution in the so-called large single channel conductance, high sensitivity to pinacidil and
rectification controller region of inward rectifiers (see Figure 4-2, cromakalim, and low sensitivity to diazoxide) are similar to those
P-helix
TM1 TM2 TMD0 TMD1 TMD2
SURx
88 Å
Slide helix
Kir6
N-terminus L0
NBD1 NBD2
A C-terminus B C
Figure 4-5. Molecular structure of KATP channel. A, A pore-forming subunit of KATP channel is encoded by Kir6.x genes and contains two transmembrane helical domains
TM1 and TM2, a pore-forming region (P-helix), a large C-terminal domain, and characteristic N-terminal domain (slide helix) interfacing inner leaflet of the membrane and
cytoplasmic C-terminus. B, An auxiliary subunit to the channel is encoded by SUR1 and SUR 2A/B genes and consists of the seventeen transmembrane helices organized
in several distinct domains and several important cytoplasmic regions. TMD0 domain and L0 region are responsible for the interaction with Kir6.x subunits and regulation
of gating of the channel. TMD1 and TMD2 domains are followed by nucleotide binding domains, NBD1 and NBD2, which form two nucleotide binding sites at their interface.
C, Arrangement of Kir6.x and SUR.x subunits in an octameric KATP channel complex. Modeling studies predict one adenosine triphosphate binding site at each of the Kir6
interfaces and Mg2+-nucleotide binding sites in the NBD domains of SUR subunit.
(Modified from Flagg TP, Enkvetchakul D, Koster JC, et al: Muscle KATP channels: recent insights to energy sensing and myoprotection. Physiol Rev 90:799–829, 2010.)
obtained from ventricular myocytes. The latter is also supported increased by stretch suggesting the involvement of channel-
by a relatively low level of SUR1 (a pancreatic isoform) in the cytoskeleton interactions. Intracellular pH is another potent
ventricles, and the finding that the activity of KATP channels is regulator of native KATP channels with acidification leading to
essentially unaffected in ventricular myocytes from SUR1 knock- increase in channel activity.
out mice. Recent experiments with mice showed, however, that KATP channels can be inhibited or activated by a variety of
SUR1 subunit of KATP channel may be a dominant isoform in drugs, all acting on the SUR subunit. Sulfonylureas such as ace-
atrial tissue. In particular, activity of KATP channels could not be tohexamide, glipizide, glibenclamide, tolbutamide, and HMR
detected in atrial myocytes isolated from SUR1 knockout mice, 1098 are prominent inhibitors. Clinically, sulfonylureas are used
and the pharmacologic profile of atrial KATP channels is reminis- exclusively for the treatment of type 2 diabetes, but they are also
cent of that conferred by SUR1 (higher sensitivity to diazoxide, a useful research tool in work with cardiac preparations.
lower to pinacidil) rather than SUR2 subunits. It remains unre- KATP channel openers (KCO) include pinacidil, cromakalim,
solved whether this subunit composition exists in atrial tissue of rimakalim, nicorandil, diazoxide, and minoxidil sulfate. In con-
other animals and humans. trast to sulfonylureas, KCOs are useful in treating cardiovascular
KATP channels in smooth muscle display a number of proper- disorders such as myocardial ischemia and congestive heart
ties distinct from that in cardiac myocytes, suggesting unique failure. KCOs display strong selectivity to subunit composition
subunit composition.36 Significantly smaller channel densities (up of KATP channels. In particular, KATP channels in pancreatic β-cells
to 100-fold per cell), generally low single channel conductance (SUR1 based) are strongly activated by diazoxide, but not affected
(~30 pS), and a lack of channel activity upon excision of the by cromakalim or nicorandil while ventricular KATP channels
membrane patch are some of the common features of smooth (SUR2A based) are strongly activated by cromakalim or nicor-
muscle KATP channels. Studies with exogenously expressing andil but not by diazoxide. Smooth muscle KATP channels (SUR2B
systems showed that cloned KATP channels originating from coex- based) are activated by all these drugs. Mitochondrial KATP chan-
pression of Kir6.1 and SUR2B subunits resemble native KATP nels are known to be potently activated by diazoxide and inhib-
channels in smooth muscle, for the most part. A strongest support ited by 5-hydroxydecanoate (5-HD).
for Kir6.1/SUR2B composition of smooth muscle KATP channels
comes from experiments with genetically modified mice. Specifi-
cally, KATP currents cannot be recorded in aortic smooth muscle Channelopathies
cells isolated from either Kir6.1 or SUR2 knockout mice, whereas
the activity of KATP channels is preserved in these cells in Kir6.2 Mutations in genes underlying KATP channels have been associ-
knockout mice. ated with cardiovascular disorders. Specifically, F1524S and
Mitochondrial KATP channels received a lot of attention since A1513T substitutions in the SUR2A gene (missense and frame-
they were first described at this location in early 1990s. In shift, respectively) were linked to dilated cardiomyopathy. Both
contrast to sarcolemmal KATP channels, however, their molecular mutations were mapped to the locus of SUR2A, which is respon-
identity remains highly controversial.36 Although exogenously sible for catalytic activity of NBD2 domains, and likely exert their
coexpressed Kir6.1 and SUR1 subunits produce KATP channels actions through reduced activation of KATP channel. The other
with many properties resembling those found in mitochondria, reported mutation in the SUR2A gene was associated with adren-
the activity of mitochondrial KATP channels in Kir6.1 knockout ergic atrial fibrillation originating in the vein of Marshall, a well-
mice was not affected. Recent promising developments in this known location for this type of fibrillation. As in the previous
area include the identification of mitochondria-specific short- case, the underlying missense mutation in the SUR2A gene
form of SUR2 subunits generated by a nonconventional (T1547I substitution) likely affects the functioning of NBD2
intraexonic splicing, which can underlie mitochondrial KATP leading to a compromised channel regulation by adenine nucleo-
channels.39 There is also strong evidence that the pore-forming tides. Recently, the novel gain-of-function mutation (S422L sub-
subunit of mitochondrial KATP channels is encoded by KCNJ1 stitution) in the pore-forming Kir6.1 subunit (when coexpressed
(Kir1.140). with SUR2A) was associated with ventricular fibrillation and
linked to J-wave syndrome susceptibility.41
Pharmacology and Regulation

Pharmacology of KATP channels is extensive, and regulation is Conclusion
complex relative to other members of Kir family, which is in part
due to the structural complexity of the channel.4 Cardiac KATP The biophysical and regulatory properties of Kir channels are
channels are regulated by a variety of mechanisms, and important crucial for cardiac electrical activity. Significant experimental evi-
quantitative details of their modulation surely depend on the dence clearly implicates several members of the Kir subfamily as
specific subunit composition. The most prominent and charac- molecular determinants underlying the three major inward recti-
teristic mechanism involves regulation by intracellular nucleo- fier potassium currents in native cardiac cells: IK1, IKACh, and IKATP.
tides such as ATP and ADP. Intracellular ATP is a potent blocker The general architecture of Kir channels has been well estab-
of the channel while ADP (in the presence of Mg2+ ions) has an lished, and fine details of their structure and function have been
activating effect. An overwhelming level of ATP under normal revealed with the aid of several available crystal structures of
conditions keeps the channel essentially shut (native channels in cloned channels. Nevertheless, many important questions remain
cardiac myocytes are half blocked by 50 to 100 µM ATP), whereas unanswered. For example, how do the differences in the bio-
metabolic disturbances leading to both a drop in ATP levels and physical and regulatory properties of Kir2 isoforms affect hetero-
a consequent rise in ADP levels result in channel opening (in the meric channel complexes that underlie the native IK1 in different
presence of intracellular Mg2+ ions). Experiments with isolated species and in different parts of the heart? What is the subunit
membrane patches show that phospholipids, especially PIP2, are composition of mitochondrial KATP channel? How are Kir chan-
potent activators of native (and cloned) KATP channels while their nels sorted into microdomains in the sarcolemma, such as
hydrolysis reduces channel activity. Regulation of KATP channels T-tubules or intercalated discs, and how do they interact with
by PKA-dependent phosphorylation is well documented in other proteins within these microdomains? These questions
smooth muscles, but the data are limited in cardiac myocytes. It undoubtedly will be the focus of much investigation in the near
has been shown that the activity of atrial KATP channels can be future.
J Physiol 582(Pt 2):675-693, 2007. PMCID: sinoatrial node of heart. J Histochem Cytochem
References 2075332. 49(10):1221-1234, 2001.
1. Lopatin AN, Nichols CG: Inward rectifiers in the

heart: an update on I(K1). J Mol Cell Cardiol
15. Cheng L, Wang F, Lopatin AN: Metabolic stress
in isolated mouse ventricular myocytes leads to
remodeling of t-tubules. Am J Physiol Heart Circ
29. Koumi S, Wasserstrom JA: Acetylcholine-sensitive
muscarinic K+ channels in mammalian ventricular
myocytes. Am J Physiol 266(5 Pt 2):H1812-H1821,
4
33(4):625-638, 2001. Physiol 301:H1984-H1995, 2011. 1994.
2. Anumonwo JM, Lopatin AN: Cardiac strong 16. Wischmeyer E, Doring F, Karschin A: Acute sup- 30. Kitamura H, Yokoyama M, Akita H, et al: Tertiapin
inward rectifier potassium channels. J Mol Cell pression of inwardly rectifying Kir2.1 channels potently and selectively blocks muscarinic K(+)
Cardiol 48(1):45-54, 2010. by direct tyrosine kinase phosphorylation. J Biol channels in rabbit cardiac myocytes. J Pharmacol
3. Nerbonne JM, Nichols CG, Schwarz TL, et al: Chem 273(51):34063-34068, 1998. Exp Ther 293(1):196-205, 2000.
Genetic manipulation of cardiac K(+) channel 17. Huang CL, Feng S, Hilgemann DW: Direct acti- 31. Aryal P, Dvir H, Choe S, et al: A discrete alcohol
function in mice: what have we learned, and where vation of inward rectifier potassium channels by pocket involved in GIRK channel activation. Nat
do we go from here? Circ Res 89(11):944-956, PIP2 and its stabilization by Gbetagamma. Nature Neurosci 12(8):988-995, 2009. PMCID: 2717173.
2001. 391(6669):803-806, 1998. 32. Patil N, Cox DR, Bhat D, et al: A potassium
4. Hibino H, Inanobe A, Furutani K, et al: Inwardly 18. Zaza A, Rocchetti M, Brioschi A, et al: Dynamic channel mutation in weaver mice implicates mem-
rectifying potassium channels: their structure, Ca2+-induced inward rectification of K+ current brane excitability in granule cell differentiation.
function, and physiological roles. Physiol Rev during the ventricular action potential. Circ Res Nat Genet 11(2):126-129, 1995.
90(1):291-366, 2010. 82(9):947-956, 1998. 33. Herrup K: The weaver mouse: a most cantanker-
5. Doyle DA, Morais Cabral J, Pfuetzner RA, et al: 19. Caballero R, Dolz-Gaiton P, Gomez R, et al: Fle- ous rodent. Proc Natl Acad U S A 93(20):10541-
The structure of the potassium channel: molecular cainide increases Kir2.1 currents by interacting 10542, 1996. PMCID: 38187.
basis of K+ conduction and selectivity. Science with cysteine 311, decreasing the polyamine- 34. Voigt N, Trausch A, Knaut M, et al: Left-to-right
280(5360):69-77, 1998. induced rectification. Proc Natl Acad Sci U S A atrial inward rectifier potassium current gradients
6. Hansen SB, Tao X, MacKinnon R: Structural basis 107(35):15631-15636, 2010. PMCID: 2932566. in patients with paroxysmal versus chronic
of PIP2 activation of the classical inward rectifier 20. Plaster NM, Tawil R, Tristani-Firouzi M, et al: atrial fibrillation. Circ Arrhythm Electrophysiol
K+ channel Kir2.2. Nature 477(7365):495-498, Mutations in Kir2.1 cause the developmental and 3(5):472-480, 2010.
2011. PMCID: 3324908. episodic electrical phenotypes of Andersen’s syn- 35. Yang Y, Liang B, Liu J, et al: Identification of a
7. Enkvetchakul D, Jeliazkova I, Bhattacharyya J, drome. Cell 105(4):511-519, 2001. Kir3.4 mutation in congenital long QT syndrome.
et al: Control of inward rectifier K channel 21. Priori SG, Pandit SV, Rivolta I, et al: A novel form Am J Hum Genet 86(6):872-880, 2010. PMCID:
activity by lipid tethering of cytoplasmic domains. of short QT syndrome (SQT3) is caused by a 3032079.
J Gen Physiol 130(3):329-334, 2007. PMCID: mutation in the KCNJ2 gene. Circ Res 96(7):800- 36. Flagg TP, Enkvetchakul D, Koster JC, et al:
2151642. 807, 2005. Muscle KATP channels: recent insights to energy
8. Logothetis DE, Jin T, Lupyan D, et al: 22. Hattori T, Makiyama T, Akao M, et al: A novel sensing and myoprotection. Physiol Rev 90(3):799-
Phosphoinositide-mediated gating of inwardly rec- gain-of-function KCNJ2 mutation associated with 829, 2010.
tifying K(+) channels. Pflugers Arch 455(1):83-95, short-QT syndrome impairs inward rectification of 37. Shyng S, Ferrigni T, Nichols CG: Control of rec-
2007. Kir2.1 currents. Cardiovasc Res 93:666-673, 2012. tification and gating of cloned KATP channels by
9. Lopatin AN, Makhina EN, Nichols CG: The 23. Xia M, Jin Q, Bendahhou S, et al: A Kir2.1 gain- the Kir6.2 subunit. J Gen Physiol 110(2):141-153,
mechanism of inward rectification of potassium of-function mutation underlies familial atrial 1997. PMCID: 2233786.
channels: “long-pore plugging” by cytoplasmic fibrillation. Biochem Biophys Res Commun 38. Bao L, Kefaloyianni E, Lader J, et al: Unique prop-
polyamines. J Gen Physiol 106(5):923-955, 1995. 332(4):1012-1019, 2005. erties of the ATP-sensitive K(+) channel in the
10. Xu Y, Shin HG, Szep S, et al: Physical determi- 24. Watanabe H, Knollmann BC: Mechanism underly- mouse ventricular cardiac conduction system.
nants of strong voltage sensitivity of K(+) channel ing catecholaminergic polymorphic ventricular Circ Arrhythm Electrophysiol 4(6):926-935, 2011.
block. Nat Struct Mol Biol 16:1252-1258, 2009. tachycardia and approaches to therapy. J Electro- PMCID: 3247655.
11. Pegan S, Arrabit C, Zhou W, et al: Cytoplasmic cardiol 44(6):650-655, 2011. 39. Ye B, Kroboth SL, Pu JL, et al: Molecular
domain structures of Kir2.1 and Kir3.1 show sites 25. Wettschureck N, Offermanns S: Mammalian G identification and functional characterization
for modulating gating and rectification. Nat Neu- proteins and their cell type specific functions. of a mitochondrial sulfonylurea receptor 2 splice
rosci 8(3):279-287, 2005. Physiol Rev 85(4):1159-1204, 2005. variant generated by intraexonic splicing. Circ Res
12. Clarke OB, Caputo AT, Hill AP, et al: Domain 26. Bichet D, Haass FA, Jan LY: Merging functional 105(11):1083-1093, 2009. PMCID: 2988690.
reorientation and rotation of an intracellular studies with structures of inward-rectifier K(+) 40. Foster DB, Ho AS, Rucker J, et al: Mitochondrial
assembly regulate conduction in Kir potassium channels. Nat Rev Neurosci 4(12):957-967, 2003. ROMK channel is a molecular component of
channels. Cell 141(6):1018-1029, 2010. 27. Nishida M, MacKinnon R: Structural basis of MitoKATP. Circ Res 111:446-454, 2012.
13. Panama BK, Lopatin AN: Differential polyamine inward rectification: cytoplasmic pore of the G 41. Medeiros-Domingo A, Tan BH, Crotti L, et al:
sensitivity in inwardly rectifying Kir2 potassium protein-gated inward rectifier GIRK1 at 1.8 A Gain-of-function mutation S422L in the KCNJ8-
channels. J Physiol 571(2):287-302, 2006. resolution. Cell 111(7):957-965, 2002. encoded cardiac K(ATP) channel Kir6.1 as a
14. Gaborit N, Le Bouter S, Szuts V, et al: Regional 28. Dobrzynski H, Marples DD, Musa H, et al: Dis- pathogenic substrate for J-wave syndromes.
and tissue specific transcript signatures of ion tribution of the muscarinic K+ channel proteins Heart Rhythm 7(10):1466-1471, 2010. PMCID:
channel genes in the non-diseased human heart. Kir3.1 and Kir3.4 in the ventricle, atrium, and 3049900.
Mammalian Calcium Pumps
in Health and Disease 5
Marisa Brini and Ernesto Carafoli
much more complex, but the basic E1/E2 nomenclature has been
CHAPTER OUTLINE
retained. Importantly, each step of the reaction cycle is reversible,
2+
Sarco/Endoplasmic Reticulum Ca ATPase 43 so that ATP can be produced by reversing the direction of the ion
2+ transport process. Reversal of the SERCA pump, with production
Secretory Pathway Ca ATPase 46
of ATP, had in fact already been demonstrated in one of the first
2+
Plasma Membrane Ca ATPase 47 experiments on the transport of Ca2+ by vesicular preparations of
2+ sarcoplasmic reticulum.5 A simplified version of the cycle, but
Ca Pumps in the Disease Process 49
adapted to Ca2+ pumps, is shown in Figure 5-1.
Conclusions 51 Several Ca2+ pump isoforms have been described in animal
cells, differing essentially in tissue distribution, regulatory prop-
erties, and some mechanistic peculiarities. The isoform diversity
Ca2+-transporting adenosine triphosphatases (ATPases; Ca2+ reflects the existence of separate basic gene products, but also the
pumps) have been described in animal and plant cells and in cells occurrence of complex patterns of alternative splicing that
of lower eukaryotes. This chapter will focus on the ATPases of increase very significantly the number of variants of each of the
animal cells and on the disease processes linked to their dysfunc- three pumps. The analysis of the differential properties of the
tion. The three animal Ca2+ pumps belong to the large superfam- Ca2+ pump isoforms is a vigorously investigated topic that has
ily of P-type ATPases, which have been so defined because their important linkages to the general process of cellular Ca2+ homeo-
reaction cycle is characterized by the formation of an acid-stable stasis, which in animal cells is regulated by a number of nonmem-
phosphorylated Aspartate (Asp) residue (the P intermediate) in a brane Ca2+-binding proteins and of membrane-intrinsic Ca2+
highly conserved sequence (SDKTGT[L/IV/M][T/I/S]).1 The channels and transporters. The transporters interact with Ca2+
family now contains hundreds of members and eight subfamilies.2 with high or low affinity, and thus function either as fine tuners
The subfamilies have been identified based essentially on trans- of cytosolic Ca2+ or come into play whenever the concentration
ported substrate specificity, the evolutionary appearance of which of Ca2+ increases to levels adequate for their low affinity. The Na/
having been accompanied by abrupt changes in sequence. The Ca-exchanger of the plasma membrane and the mitochondrial
changes, however, do not involve eight conserved structurally and Ca2+ uptake and release systems are the low-affinity regulators of
mechanistically important regions that define the core of the cytosolic Ca2+. The three pumps, by contrast, control Ca2+ effi-
superfamily. Five branches have been identified in the phyloge- ciently even in the low concentrations of the cytosol at rest. Pump
netic tree of the superfamily: two animal Ca2+ pumps belong to activity is fundamental to the correct functioning of the machin-
subgroup II A (the sarco/endoplasmic reticulum Ca2+ [SERCA] ery of animal cells: dysfunctions, genetic or otherwise, of their
and secretory pathway Ca2+ [SPCA] pumps), one to subgroup II B operation might not necessarily induce cell death, but invariably
(the plasma membrane Ca2+ [PMCA] pump). All P-type ATPases, generate disease phenotypes.
including the three that transport Ca2+ in animal cells, are multi-
domain proteins that share the essential properties of the reaction
mechanism, have molecular masses varying between 70 and
150 kDa, and share the presence of 10 hydrophobic trans- Sarco/Endoplasmic Reticulum Ca2+ ATPase
membrane (TM) spanning domains (however, some have only six
or eight). The number of TMs being even and the N- and The SERCA pump is a key mechanism to adjust the Ca2+ homeo-
C-termini of all P-type pumps are on the same membrane side stasis in the endoplasmic reticulum (ER) lumen. Considering that
(i.e., the cytosol); one exception is a splice variant of the SERCA the ER Ca2+ is involved in a multitude of signaling events and in
pump that has 11 TM). The P-type ATPases also share the sensi- housekeeping functions that control cell growth, differentiation
tivity to the transition state analog orthovanadate and, with some and apoptosis, the activity of the SERCA pump is a key element
specific differences (see below), to La3+. Other inhibitors only in cell wellness.
affect selected members of the superfamily. The three-dimensional The SERCA pump is inhibited by La3+ and orthovanadate,
(3D) structures of four P-type ATPases have become available and the discovery of specific inhibitors such as thaspigargin,6
following the landmark solution of the 3D structure of the SERCA cyclopiazonic acid,7 and 2.5-di(t-butyl)hydroquinone8 repre-
pump 12 years ago3: molecular modeling on templates of the sented a significant advantage in the biochemical and structural
SERCA pump structure has indicated that all P-type ATPases characterization of the pump.
share the general principles of 3D structure. The reaction cycle of The SERCA protein is organized in the membrane with 10
P-type ATPases originally envisaged only the E1 and E2 steps, TMs: numerous mutagenesis studies and the solution of its 3D
characterized by distinct conformations and affinities for adenos- structure have clarified essential molecular details of its function,
ine triphosphate (ATP) and the transported ion. For example, which will be summarized in this chapter. Full details are avail-
Ca2+ pumps in the E1 state engage Ca2+ with high affinity at one able in a number of more comprehensive reviews.9-11
side of the membrane, and in the state their E2 lowered affinity Analysis of the 3D structure of the SERCA1 pump isoform
for Ca2+ releases it to the opposite membrane side.4 Later, addi- has revealed that the single polypeptide chain folds in three
tional intermediate states were added that made the reaction cycle cytosolic domains and in one transmembrane sector (M)
43
composed of the 10 formerly predicted TMs (Figure 5-2). The of the luminal gate that releases Ca2+ to the lumen and permits
three cytosolic domains have been named according to their role the E1P-E2P transition. The closure of luminal gate, and thus
in the reaction cycle: the nucleotide binding domain (N) binds the E2P→E2 Pi transition, then occurs because a second rotation
ATP, the phosphorylation domain (P) drives ATP hydrolysis of the A domain locks it to the P domain. A highly conserved
leading the phosphorylation of the catalytic Asp, and the actuator TGES motif in the A domain fills the gap between the N and P
domain (A) catalyzes the dephosphorylation of the P-domain. domains after the second rotation of the A domain, eventually
The A and P domains are connected to the transmembrane M permitting the release of Ca2+ into the lumen. The rearrange-
domain that contains the 2 Ca2+ binding sites: the SERCA pump ments of transmembrane helices M1-M6 induced by the rotation
transports 2 Ca2+ per ATP hydrolyzed. The N domain is instead of the A domain allow protons and water molecules to enter and
connected to the P domain. During the cycle, phosphorylation stabilize the empty Ca2+ binding sites. The rearrangements also
and dephosphorylation events promote conformational changes induce the retraction of the TGES from the phosphorylation site
that control the access of Ca2+ to the two binding sites (site I and and the entrance of one water molecule to the phosphorylation
site II), which exist in high- and low-affinity states (Figure 5-3). site, inducing the release of phosphate (and Mg2+) and the com-
The two sites are located near the cytoplasmic surface of the plete closure of the luminal gate.
membrane, but site I faces the cytoplasmic side and site II is Three SERCA genes generate three isoforms. Their number
closer to the luminal side. Once Ca2+ becomes bound to site I, a is increased by alternative splicing processes. SERCA1 is almost
conformational change increases the affinity of site II and permits exclusively expressed in muscle tissues, specifically in fast-twitch
the phosphorylation of the catalytic Asp by ATP, leading to the skeletal muscles. Interestingly, the generation of truncated and
transition E2→E1→E1•2Ca2+ E1P. The binding of ATP cross- less active SERCA1 variants has been described, contribute to
links the P and N domains, permitting the interaction of P reduce the Ca2+ concentration in the ER lumen and cause apop-
domain with the A domain, which rotates inducing the opening totic cell death. SERCA2b and SERCA2a are the two major
SERCA protein isoforms, with the former having a housekeeping
function and the latter having a more specialized function.
SERCA2a is found in slow skeletal and cardiac muscles; it is also
expressed in low amounts in smooth muscles and in neurons. The
Cytosol splice variants SERCA2c and SERCA2d have also been found in
ATP Ca2+ low amounts in the heart. The expression of the SERCA3 pump,
ADP
which has a limited cell-type distribution, is variable. In several
cell types this expression is induced by differentiation, and it is
E1 E1 ATP E1 P decreased during tumorigenesis and blastic transformation. In
cells of hematopoietic origin and in various epithelial cells, the
Membrane
Ca2+ Ca2+
SERCA2 pump gene is coexpressed with that of the SERCA3
pump. All SERCA3 splice variants (SERCA3a-f) have lower Ca2+
Ca2+ affinity than the SERCA2 pump, which raises doubts on their
role in the presence of higher affinity SERCA pump variants.
E2 E2 P E2 P Differences in their spatial cellular distribution could justify their
copresence. The SERCA3 pump is confined to environments
P Ca2+ with high Ca2+ concentration, such as those close to the plasma
Extracellular membrane of cardiomyocytes, at basal regions in epithelial cells,
space or lumen and in the membrane of acidic Ca2+ stores in platelets. Although
the SERCA1a pump is the best characterized isoform in terms of
Figure 5-1. A simplified reaction cycle of the P-type adenosine triphosphatases
structure-function relationship, the regulatory aspects of SERCA
(pumps) adapted to the Ca2+ pumps. The two original conformational states of the
pumps are envisaged. The E1 pumps bind Ca2+ with high affinity at one membrane pumps have instead been better defined on the SERCA2 isoform.
side (the cytosol), the E2 pumps have much lower Ca2+ affinity, and release Ca2+ to The SERCA2a pump is the major isoform of developing and
the opposite membrane side. Adenosine triphosphate phosphorylates a conserved adult mammalian heart (SERCA2b is also expressed there, its
Asp in the active site. level being unchanged during the development).12 SERCA2a is
Cytoplasm Cytoplasm
Membrane E1-2Ca2+ E2 Membrane
Lumen Lumen
Figure 5-2. The three-dimensional structure of the SERCA pump, showing the open configuration of the three cytoplasmic domains A, N, and P in the presence of Ca2+,
and the more compact configuration of the cytosolic sector in its absence. The two purple spheres in the upper panel represent the two bound Ca2+. The E2 structure shown
contains the inhibitor thapsigargin (TG). The TMs are the transmembrane domains. Several residues of importance not discussed in the text are also shown.
(Modified from Toyoshima C: Structural aspects of ion pumping by Ca2+-ATPase of sarcoplasmic reticulum. Arch Biochem Biophys 476:3–11, 2008.)
Mammalian Calcium Pumps in Health and Disease 45
Figure 5-3. The conformational changes of the cytosolic and mem-

brane domains of the SERCA pump during the transport of Ca2+
5
ATP 30° 90°
across the molecule.
Mg2+
(Modified from Toyoshima C: Structural aspects of ion pumping by
Ca2+-ATPase of sarcoplasmic reticulum. Arch Biochem Biophys 476:3–
11, 2008.)
ADP
H+
20°
110°
30°
Mg2+
Pi
Ca2+
H+ Ca2+
the most abundant protein in the heart membrane; its increased would bind to the pump in transmembrane sites of PLB.14 Studies
sarcoplasmic reticulum (SR) expression during the development on SLN are still limited, but SLN overexpression in ventricle
paralleled the increasing rate of Ca2+ uptake by the SR lumen and cells of animal models, where SLN is essentially absent, caused
the shortening of the relaxation time in adults, with respect to a decrease in the Ca2+ affinity of SERCA2a and slowed relaxation,
neonatal heart. SERCA2a pump levels are higher in atria than in suggesting that SLN is as effective an inhibitor of SERCA pump
ventricles, partially accounting for the shorter duration of con- activity as PLB. Interestingly, the overexpression of SLN in the
traction in atrial than in ventricular tissues. heart of PLB-null mice caused a decrease in the affinity of the
The expression and the activity of the SERCA2a pump have SERCA2a pump for Ca2+ and impaired contractility. The finding
been studied extensively. The primary mechanism of the regula- that isoproterenol, a β-adrenergic agonist, relieved the inhibition
tion of the pump is mediated by phospholamban (PLB).13 PLB suggests that SLN could mediate the β-adrenergic response in
is a 52-residue protein composed of a hydrophobic helical the heart. Ablation of the SLN gene increases the affinity of the
C-terminal portion inserted into the SR membrane and a hydro- SERCA pump for Ca2+, resulting in enhanced rates of SR Ca2+
philic N-terminal region that protrudes into the cytosol and uptake.15
contains phosphorylation sites (Ser16 and Thr 17) for PKA and, The SERCA2b pump is the acknowledged housekeeping
possibly, CAMKII. Dephosphorylated PLB binds to the pump, isoform, and it has a dual role. By transporting Ca2+ from the
decreasing its Ca2+ affinity; phosphorylation by PKA, and possi- cytosol to the ER lumen, the SERCA2b pump contributes to the
bly CaMKII, releases the inhibition and increases the affinity of maintenance of cytosolic Ca2+ at the low resting levels (≈100 nM),
the pump for Ca2+ and thus Ca2+ transport. The hydrophilic at the same time ensuring the high Ca2+ levels (in the mM range)
N-terminal portion of PLB interacts with a domain close to the in the lumen of ER that make the ER the main intracellular Ca2+
active site of the pump and, within the membrane, with trans- store that controls cellular activities (e.g., contraction, prolifera-
membrane helices 2, 4, 6, and 9. It is generally believed that PLN tion, differentiation, cell death). The pump also ensures the
exists both as a pentamer and a monomer. It is not clear how the proper internal Ca2+ ambient for the ER enzymes (e.g., those
conversion between the two forms occurs, but several lines of involved in protein folding and lipid synthesis).
evidence indicate that monomeric PLB is the active form. The SERCA2b pump shares 85% sequence identity with the
However, structural observations indicate that pentamers can also SERCA1a counterpart but differs functionally from it and from
interact with the pump.14 the SERCA2a isoform, which is characterized by a unique
Another small (31 residues) transmembrane protein, sarco- C-terminal extension—the so-called 2b-tail, which forms a
lipin (SLN), originally identified as it copurified with the luminal sequence extension and an additional TM segment
SERCA1a pump, has recently also received attention. SLN is (TM11).16 The extension has regulatory properties: the longer
predominantly expressed in the atrial compartment of the heart, pump has a twofold higher affinity for cytosolic Ca2+ and a lower
and its sequence is similar to that of the transmembrane sector maximal turnover rate. According to a model based on the
of PLN. Modeling studies have revealed that the SLN and PLB SERCA1a structure, and on the NMR structure of TM11, the
interactions with the SERCA pump may be similar—that is, SLN interaction of TM11 with TM7 and TM10 has been proposed to
stabilize the pump in the Ca2+-bound E1 conformation, with the (plasma membrane ATPase-related) and PMA1 (plasma mem-
high-affinity Ca2+ binding sites facing the cytosol. The TM11 has brane H+-ATPase) by two independent groups. Later, it was
also been proposed as a novel regulator of the SERCA pump. The detected in mammals and higher vertebrates.
co-reconstitution of the 18-residue long TM11 with the SERCA1a Two basic SPCA genes have been found in higher vertebrates
protein in vitro reduced the maximum reaction rate (Vmax) and coding for the SPCA1 and SPCA2 pumps. Alternative splicing of
increased the Ca2+ affinity of the latter. The regulation of the the primary transcript has been described for SPCA1, but not for
SERCA pump by PLB and SLN, and by the 2b-tail in the 2b the SPCA2 pump.
variant, resembles the interaction of the β- and γ-subunits of the The SPCA1 pump is the housekeeping Ca2+ and Mn2+ pump,
Na+/K+-ATPase with the catalytic subunit α. The regulation also because it is expressed in all cell types, even if with different levels
resembles the regulatory interaction of the PMCA pump with in various tissues. The alternatively spliced SPCA1 mRNA gen-
calmodulin, suggesting operationally similar molecular mecha- erates SPCA1a-d proteins, which differ in the C-terminal portion.
nisms of P-type pump regulation.17 In addition, the C-terminal The expression of SPCA2 pump in human tissues is more
portion of the ePMCa pump can to some extent replace PLB as restricted than that of the SPCA1 pump, suggesting a more spe-
an inhibitor of the SERCA pump. Apart PLB and SLN, which are cialized physiologic function for it. The SPCA2 pump is particu-
the best-studied regulators of the SERCA pump, other proteins larly abundant in the gastrointestinal tract, trachea, prostate, and
interact with the pump. Those interacting at the luminal side have thyroid, salivary, and mammary glands.
an important role. The two chaperones calreticulin and calnexin The SPCA pump is also predicted to be organized in 10 TMs,
contain a globular N-domain that binds carbohydrates—an with a large headpiece portion protruding into the cytosol. On
extended P-domain that mediates the binding of ERp57 and an the basis of the SERCA pump 3D structure, the cytosolic portion
acidic C-terminal domain, that in the case of calreticulin binds is divided in the three canonical A, P, and N portions (Figure
25 mol of Ca2+ per molecule of protein with low affinity (Kd, 5-4). The SPCA pump is shorter than the SERCA and PMCA
2 mM). Luminal Ca2+ buffering by calnexin is less significant, and pumps, because it does not have the long cytoplasmic C-terminal
the acidic C-terminus of the protein protrudes into the cytosol. tail of the PMCA pump, and its intracellular loops are shorter.
Calreticulin and calnexin have been suggested to interact through The SPCA pump differs from the SERCA pump in having only
their N-domain with a putative glycosylated residue present in the one Ca2+ binding site, corresponding to site II in the SERCA1
C-terminal tail of the SERCA2b pump, but absent in SERCA2a. pump. The suggested TM packing and possibly some distant
Glycosylation-independent interaction with the SERCA2b pump residues would define the structure of the site, making it adequate
has been shown to occur, and molecular modeling of the SERCA2b to bind Mn2+ with high affinity and to transport it. This structural
molecule has suggested that its 2b-tail is located in luminal loops, aspect is a peculiarity of the SPCA pump. Gln783 in TM6 and
thus making it inaccessible to interacting proteins. ERp57, a Val335 in TM4 appear to be essential for Mn2+ transport, because
member of the PDI family of proteins with thio-oxidoreductase their mutation (in the Pmr1 yeast pump) abolished Mn2+ trans-
activity, is recruited by the SERCA2b pump–chaperone complex port. Another distinction of the SPCA pump with respect to the
to catalyze the formation of an inhibitory disulfide bridge between SERCA and PMCA pumps, which function as obligatory H+
Cys875 and Cys887 in the luminal loop L7-8 of the SERCA2b
pump.18 The formation of a disulfide bridge could be a regulatory
mechanism, but the proposal is controversial, because mutations
of either or both Cys residues resulted in the loss of Ca2+ transport
but not of the activity in SERCA1.19
Two additional luminal Ca2+ binding proteins have been
shown to interact with the SERCA2 pump: the ubiquitously
expressed calumenin and the histidine-rich Ca2+-binding protein
(HRC). Both decrease the apparent Ca2+ affinity of the pump.
HRC binds Ca2+ with high capacity and low affinity and could
mediate both SR Ca2+ uptake and release through its interaction
with SERCA, when the SR Ca2+ is low, and with triadin, which
is part of the RyR Ca2+ release complex when it is saturated by
Ca2+.20
Secretory Pathway Ca2+ ATPase

The Golgi Ca2+ ATPase (the SPCA pump) shares with the
SERCA pump the role of loading Ca2+ in the Golgi apparatus.21,22
The relative contribution of the SPCA and SERCA pumps to the
total Ca2+ uptake in the Golgi apparatus is cell-type dependent.
The use of the SERCA pump selective inhibitor thapsigargin,
which does not inhibit the SPCA pump, has confirmed that the
SERCA pump contribution is significant in numerous cell types,
but not in keratinocytes that mainly use the SPCA pump. This
point is interesting because SPCA pump mutations that impair
the function of the pump lead to a specific human disease, Hailey-
Hailey disease (discussed later). In contrast to the SERCA pump,
the SPCA pump also transports Mn2+ and thus serves the dual
function of supplying this metal to the Golgi lumen, where it acts
Figure 5-4. A molecular modeling representation of the structure of the SPCA
as a cofactor for the glycosyl-transferases, and to remove it from pump (white) superimposed on a template of the SERCA pump (red).
the cytosol, where its accumulation could be toxic. The Golgi
Ca2+ ATPase was originally discovered in yeast and named Pmr1 (Courtesy L. Raeymakers, Leuven, Belgium.)
exchangers, is the finding that it does not appear to countertrans- detection of the PMCA in phosphate gels of membrane prepara-
5
port H+. No specific SPCA pump inhibitors have been described, tions containing much larger amounts of other P-type pumps).
and no endogenous activators are known. Mutagenesis studies on The other significant difference regarding the SERCA pump,
the Pmr1 yeast pump have generated phenotypes that are impor- which is shared by the SPCA pump, is the presence of only one
tant tools that could provide information on the role of the SPCA Ca2+ binding site. The single site of the PMCA pump corre-
pump in higher eukaryotic organisms. The expression levels and sponds to site 2 of the SERCA pump, and it coordinates Ca2+ to
activity of SPCAs change in response to altered physiologic six residues (essentially the same of the SERCA pump) in TM4
needs. SPCA1 pump expression and activity changed in the brain and TM6. The site that would correspond to site 1 of the SERCA
after an ischemic event,23 and the SPCA1 pump levels increased pump is not operational in the PMCA pump, most likely because
in mammary gland during lactation. This last finding is shared of the absence of an essential acidic residue in transmembrane
by the PMCA2 pump (discussed later). The role of the SPCA domain 5. Interestingly, the insertion by mutagenesis of the
pump in the maintenance of Golgi Ca2+ homeostasis deserves a missing acidic residue (or even of a Gln) in transmembrane 5
comment, because recent evidence have indicated that the Golgi confers to the PMCA pump properties (high cooperativity of
apparatus is a heterogeneous and highly dynamic Ca2+ store.24 Ca2+ binding, Hill coefficient approaching 2, and negative effect
The apparatus has been shown to be an InsP3-sensitive Ca2+ of La3+ on the phosphorylated Asp) that resemble those of the
store, implying a role for it in the generation of local cytosolic SERCA pump—that is, the second Ca2+ binding appears to have
Ca2+ signals. The specific distribution of the SPCA pump in the been made operational.27 All PMCA pumps are sensitive to the
Golgi membranes appears important; it can vary with the cell classical inhibitor of P-type pumps orthovanadate and are insen-
type (i.e., in some cases it colocalizes with Golgi markers and in sitive to the specific inhibitors of the SERCA pump. Some sup-
others with those of the trans-Golgi). However, a consensus has posedly specific inhibitors have been described, such as eosin or
now been reached that the SPCA1 pump–containing Golgi sub- caloxin, but their specificity, and thus their usefulness, does not
compartment is insensitive (or mildly sensitive) to InsP3, and match that of the SERCA pump inhibitors.
thus appears not to be involved in generating cytosolic Ca2+ In mammals, the plasma membrane Ca2+ ATPase is the
signals. The kinetics of Ca2+ release from the Golgi apparatus product of four separate genes, which encode four basic pumps
differed from those of the ER. In particular, although the latency (PMCA 1 to 4) that differ in tissue distribution, functional prop-
following agonist application and the initial rate of Ca2+ release erties, and interaction with protein partners. PMCA 1 and PMCA
were similar for the two organelles, the release of Ca2+ from the 4 have wide tissue distribution and have traditionally been con-
Golgi apparatus terminates faster than that from the ER. These sidered as the housekeeping enzymes. This concept is still valid
findings would be compatible with distinct Ca2+ subcompart- for the PMCA 1 pump, but has recently been challenged for the
ments in the Golgi apparatus endowed with differentCa2+ regu- PMCA 4 pump, which has been shown to have critical Ca2+ sig-
lating molecular components. naling roles in a number of cell types. For example, in the sper-
matozoon, where it is the predominant PMCA pump form and
is required for hyperactivated sperm motility,28 ablation of its
gene limits greatly sperm motility and generates male infertility.29
Plasma Membrane Ca2+ ATPase In the heart, where it has a role in the contractility process that
is unrelated to the beat-to-beat regulation of bulk cytosolic Ca2+,
The plasma membrane Ca2+ ATPase (the PMCA pump) is a high- it modulates the activity of the Ca2+-dependent neuronal nitric
affinity, low-capacity Ca2+ exporting system. Most of the initial oxide synthase (nNOS), which is important to the control of the
work on the pump dealt with erythrocytes, but it gradually excitation–contraction coupling of the cardiomyocyte.30
became clear that the pump is present and active in all animal At variance with PMCA 1 and PMCA 4, the PMCA 2 and
cells, including those of excitable tissues (e.g., heart),25 where it PMCA 3 have restricted tissue distribution: they are particularly
is generally assumed that the export of Ca2+ was performed solely represented in neurons, and have much higher affinity for
by the Na/Ca-exchanger. The discovery of a PMCA pump in calmodulin, which is the natural regulator of the PMCA pump,
heart sarcolemma came as a surprise, and a role for it in the regu- than the two ubiquitous isoforms (Table 5-1).
lation of the bulk Ca2+ homeostasis in the cardiomyocyte was The PMCA pumps are organized in the membrane with the
never considered realistic. According to solid and abundant evi- canonical ten TMs of Ca2+ pumps. The 3D structure of the
dence, and thus to general consensus, the beat-to-beat export of PMCA pump has not yet been solved, but molecular modeling
bulk Ca2+ from heart cells is indeed performed by the Na/ work on SERCA pump templates has predictably indicated the
Ca-exchanger. However, recent evidence has shown that the same structural organization, with the cytoplasmic A, N, and P
PMCA pump of heart sarcolemma has a specific role in Ca2+ domains (Figure 5-5). Differences are also evident, beginning
signaling that is still related to the regulation of the excitation– with the long, unstructured C-terminal tail that contains the most
relaxation cycle, but that is not linked to the general regulation important regulatory sites of the pump and a phospholipid
of the homeostasis of Ca2+ (discussed later). binding stretch (discussed later) in the cytosolic loop that con-
A number of comprehensive reviews describe the properties nects TM 2 and 3.
of the PMCA pump9,26; for reasons of space limitations, this As mentioned, calmodulin is the (most important) natural
contribution will in most cases refer to them, rather than to regulator of the PMCA pump and has been exploited to purify
specific reports. The general aspects of the reaction mechanism the enzyme using calmodulin columns.31 Calmodulin interacts
of the PMCA pump are those of all P-type pumps (see Figure with a domain in the C-terminal tail of the pump that has the
5-1). However, the reaction cycle of the PMCA pump differs in amphipathic helix configuration of canonical calmodulin binding
two important aspects from that of the SERCA pump, which is domains. The affinity of the domain for calmodulin is high (Kd
normally taken as a reference because it is known in atomic detail. in the nM range), and the interaction with the PMCA pump has
One concerns the effect of La3+ on the phosphorylated Asp inter- characteristics that distinguish it from that of other targets of
mediate; in all P-type pumps, including the SERCA and SPCA calmodulin regulation. The PMCA pump apparently only inter-
pumps (see previously), La3+ is supposed to replace Ca2+ as the acts with the C-terminal portion of calmodulin,32 which does not
metal ion that activates the phosphorylation of the catalytic Asp, collapse hairpinlike around its binding PMCA domain. The
thus inhibiting the formation of the phosphoenzyme. In the matter of the interaction of the PMCA pump with calmodulin
PMCA pump, La3+ greatly increases the steady state level of the has recently seen some unexpected developments. A second
phosphoenzyme (a convenient observation that permits the calmodulin binding domain, with much lower affinity for
Table 5-1. Tissue Distribution, Calmodulin Affinity, and Calpain Sensitivity of the Four Basic PMCA Pump Isoforms
PMCA 1b* PMCA 2b PMCA 3b PMCA 4b

Tissue distribution Ubiquitous Restricted (brain) Restricted (brain) Ubiquitous
Kd calmodulin (nM) 40-50 2-4 8 30-40
Calpain sensitivity High Low Low Low
*The notation 1b to 4b refers to the full-length pump, without splicing inserts. The information given in the table, such as calmodulin affinity, has been extracted from
Enyedi A, Verma AK, Heim R, et al: The Ca2+ affinity of the plasma membrane Ca2+ pump is controlled by alternative splicing. J Biol Chem 269:41–43, 1994; Caride AJ,
Filoteo AG, Penniston JT, et al: The plasma membrane Ca2+ pump isoform 4a differs from isoform 4b in the mechanism of calmodulin binding and activation kinetics:
implications for Ca2+ signaling. J Biol Chem 282:25640–25648, 2007; Brini M, Coletto L, Pierobon N, et al: A comparative functional analysis of plasma membrane Ca2+
pump isoforms in intact cells. J Biol Chem 278:24500–24508, 2003; Strehler EE, Zacharias DA: Role of alternative splicing in generating isoform diversity among plasma
membrane calcium pumps. Physiol Rev 81:21–50, 2001.
site of the pump, locking it in an autoinhibited state.33,34 Calmod-

ulin would swing its binding domain, and presumably the entire
C-terminal tail, away from the sites of auto-inhibition, restoring
full activity to the pump.
One of the distinctive properties of PMCA pumps is the
multiplicity of activating mechanisms. Next to calmodulin, the
most important probably are acidic phospholipids, the most
effective among them being the doubly phosphorylated deriva-
tive of phosphatidyl-inositol (PIP2). Acidic phospholipids bind
to two sites: one is the basic C-terminal calmodulin binding
domain, and the other is a stretch of approximately 40 predomi-
nantly basic amino acids in the cytosolic loop connecting trans-
membrane domains 2 and 3. The activation by acidic phospholipids
could be important in vivo. It has been calculated that the con-
centration of phosphatidyl-serine in the surroundings of the
pump would in principle be adequate for approximately 50%
stimulation of its activity.35 The concentration of phosphatidyl-
serine in the membrane is not known to be modulated, and it
thus appears unrealistic to propose a role for it in the reversible
activation of the pump. PIP2, which is the most effective activa-
tor, instead becomes modulated in the membrane during Ca2+-
mediated signaling processes. A PIP2-mediated reversible process
of pump activation could thus appear be a realistic possibility.
Kinases have also been found to activate the pump by phosphory-
lating residues in its C-terminal tail. Meanwhile, protein kinase
C acts on all pump variants, and protein kinase A acts on only
one of the isoforms. An intriguing mechanism of pump activation
is that generated by a dimerization (oligomerization) process that
occurs through the C-terminal calmodulin binding domain; its
Figure 5-5. A molecular modeling of the structure of the PMCA pump (yellow)
physiologic significance is obscure. The concentration of the
superimposed on that of the SERCA pump (red). In the case of perfect match PMCA pump in the plasma membrane of all animal cells is
between the two structures, one sees the red color of SERCA pump. The long, extremely low (probably less than 0.1% of the total membrane
unstructured C-terminal tail of the PMCA pump is not represented. The binding protein); therefore the possibility of random association of two
region in the cytosolic loop connection TM2 and TM3 of the PMCA pump is the or more pump monomers in situ to generate dimers or oligomers
most evident difference between the two pump structures. The following amino appears unrealistic. All mechanisms of activation act by increas-
acids of the SERCA pump are highlighted with colored sphere: N-terminal metionin ing the Ca2+ affinity of the pump; in their absence, the Km (Ca2+)
(green); threonine 242 (gray); lysin 400 (blue) and the C-terminal glycin (purple). of the pump is as high as 20 µM, but drops to 0.5 µM or less in
(Modified from Krebs J, Helms V, Griesinger C: The regulation of Ca2+ signal by mem- the presence of, for example, calmodulin or acidic phospholipids.
brane pumps. Helvetica Chimica Acta 86:3875-3388, 2003.) The pump can be also activated irreversibly, and that occurs when
its C-terminal tail, which includes the calmodulin binding
domain, is shaved off by the Ca2+-dependent protease calpain. In
calmodulin (Kd in the micromolar range) has been identified this case, the activation is linked to the removal of the auto-
downstream of the canonical one.32a The function of this second inhibitory C-terminal tail of the pump. The irreversible activa-
domain, which appears to be present only in some splice isoforms tion by calpain could become significant in conditions of
of the pump, is still not established. However, regardless of the pathologic Ca2+ overload that would demand the uninterrupted
problem of the existence of one or two calmodulin-binding maximal ability of the pump to extrude Ca2+ from the cytosol.
domains, it is generally accepted that the mechanism of calmodu- In addition to these activating mechanisms, a number of
lin activation depends on its ability to release the PMCA pump protein partners have recently been shown to become reversibly
from the state of autoinhibition that prevails in its absence: the associated with the PMCA pumps, sometimes even specifically
C-terminal tail of the pump, including the canonical calmodulin with some isoforms.36 Interestingly, these interactions could
binding domain, folds over to bind to two sites next to the active either be activating or inhibitory. Among them of special interest
are those with the numerous proteins that contain the PDZ for PMCA2. A large exon of 172 bp can be inserted into its
5
binding domain, which is recognized by the extreme C-terminal mRNA, again piecemeal, generating variants c and d (in frame),
portion of most PMCA pump variants. The protein partners that and, presumably, e (out of frame). The insertion of the full 172-bp
are recognized by the PMCA pump via their PDZ domain exon generates the prematurely truncated form a, the mRNA
include, among others, the MAGUK guanylate kinases and without inserts generates form b (an additional, smaller 55 bp
Ania-3, a member of the Homer family of scaffolding proteins. exon can also be optionally inserted in the case of PMCA 2).
In addition to the C-terminal region, other portions of the Information on the consequences of the slicing operation at site
PMCA molecule also interact with regulatory partners; for C on the activity of the pump is more abundant, and recent find-
example, the main intracellular loop interacts with calcineurin ings can lead to important functional developments. The inser-
and the N-terminal cytosolic region with protein 14.3.3. Of par- tion of the novel sequence roughly in the middle of the calmodulin
ticular interest is the interplay of PMCA 4 with the nitric oxide binding domain leads, as expected, to the lowering of the calmod-
synthase in heart.30 It has been shown that the pump tethers ulin affinity for the pump.39,40 The consequence of this decrease
nNOS to a compartmentalized domain in which its activity would in affinity should lower the ability of calmodulin to remove its
reduce the concentration of Ca2+ and thus the production of binding domain from the autoinhibitory sites in the main body
cGMP. The consequent increase of cAMP would then exert its of the pump, decreasing pump activity. However, predictions on
well-known positive inotropic effects while the interaction of the the effects to be expected from the insertion of the novel sequence
pump with nNOS has a defined physiologic effect. In other cases, in the calmodulin binding domain are complicated by the unex-
the effects are less well understood from a physiologic perspec- pected observation that the insert tends to reconstitute the origi-
tive; they vary from the modulation of activity, to the targeting nal entire calmodulin binding domain; 8 of the first 10 residues
to membrane domains, and to the recruiting of PMCA pump of the insert are indeed either identical or conservative, with
variants to cell components (e.g., the cytoskeleton). respect to those or the original C-terminal half of the calmodulin
As mentioned earlier, alternative splicing processes affect all binding domain they replace.
four basic primary transcripts of the pump, greatly increasing the The discussion of the splicing operation has shown that the
number of isoforms. Most of the splice variants described in PMCA 2 pump differs from the other three basic pump isoforms.
the literature have also been documented at the protein level. Other properties of the PMCA 2 pump underscore the difference
The splicing operation occurs at two transcripts sites. Site A cor- and set the PMCA 2 pump apart from the other three basic
responds to the cytosolic loop of the pump molecule that con- isoforms—for example, the high affinity for calmodulin (see
nects transmembrane domains 2 and 3, site C to the C-terminal Table 5-1) and the finding that the PMCA 2 pump has a pecu-
calmodulin binding domain. The A site inserts are always in liarly high level of activity in its absence41,42 (Table 5-2). The
frame; they affect the properties of the pumps, but do not sub- anomaly is not due to the presence of tightly bound calmodulin
stantially alter their structure. Instead, the C site inserts might in purified PMCA 2 pump preparations. It could thus be reason-
not maintain an open protein reading frame, thus resulting in the ably related to the reduced ability of the C-terminal tail of the
truncation of the pump molecule downstream of its regular PMCA 2 pump to autoinhibit the pump. Regardless of the mech-
C-terminus. The full details of the splicing operations and its anism, however, the anomaly generates a pump variant that
complexities are discussed elsewhere.37 Only the aspects that are would function nearly optimally in the absence of activation by
specially significant to the themes of this contribution will thus calmodulin; this would satisfy the demands that particular cell
be considered here. At site A, one exon of 39 bp is apparently types have for a continuous and vigorous Ca2+ exporting function
invariably inserted in the mature mRNA of pump 1, whereas an not depending on pump activation. One last important point on
exon of 42 or 36 bp can be optionally inserted or excluded in the the PMCA 2 pump is its high concentration in the mammary
transcripts of pumps 3, and 4, respectively. The pump variants gland, where its levels increase during lactation (see earlier for
without the inserts are termed z; those with the extra exon are the similar finding on the SPCA pump).
termed x. The A splice for PMCA2 is more complex: three exons
of 33, 60, and 42 bp can be inserted or excluded; however, only
some of the possible combinations have been detected in the
messenger RNAs (mRNAs) of various tissues. For example, in Ca2+ Pumps in the Disease Process
humans only variant w (all exons included), variant x (only the 42
bp exon included), and variant z (no extra exon included) have One of the distinctive properties of the Ca2+ signal is ambiva-
been detected. Only scarce information is available on the func- lence. As must be expected, the central role of Ca2+ in the regula-
tional consequences of the splicing operation at site A. Impor- tion of the most important cell activities demands its precise
tantly, however, the splicing at site A has been found to alter the
membrane targeting of the PMCA2 pump.38 The insertion of all
three extra exons (45 residues, the w form) determined its apical
targeting in polarized MDCK epithelial cells, whereas the inser- Table 5-2. Calmodulin-Dependent ATPase Activity of PMCA 2 and
tion of only one exon (14 residues, the x form) or the absence of PMCA 4
inserts (the z form) determined the sorting of the pump to the
nM ATP per mg protein per min
basolateral plasma membrane.
The splicing at site C is more complex: a large exon (154 bp – Calmodulin + Calmodulin
in PMCA1 and PMCA3, 175 bp in PMCA4) can be optionally PMCA 2 4.21 5.87
included in a site corresponding to the middle of the calmodulin-
binding domain, leading to the loss of the open reading frame PMCA 4 1.45 6.15
and to the premature truncation of the pump. The truncated ATP, Adenosine triphosphate; ATPase, adenosine triphosphatase.
pump is termed a, and the form without insertions is termed b. From Elwess NL, Filoteo AG, Enyedi A, et al: Plasma membrane Ca2+ pump
An additional smaller 68-bp exon can also be optionally included isoforms 2a and 2b are unusually responsive to calmodulin and Ca2+. J Biol
in PMCA3; however, the large extra exon contains multiple inter- Chem 272:17981–17986, 1997; and Hilfiker H, Guerini D, Carafoli E: Cloning and
nal donor sites and can thus be inserted piecemeal. The inserts expression of isoform 2 of the human plasma membrane Ca2+ ATPase.
of 87 and 114 bp are in frame and generate the c and d forms, Functional properties of the enzyme and its splicing products. J Biol Chem.
that of 152 and 154 bp are not and generates the truncated a 269:26178–26183, 1994.
form. Again, the splicing operation at site C is peculiarly complex
spatial and temporal control. The vast array of Ca2+ binding characterized by the loss of adhesion between epidermal cells and
proteins and Ca2+ transporters expressed in all animal cells under- by abnormal keratinization. More than 130 mutations in the
lies the concept. Defects in the control of Ca2+ unavoidably gen- SERCA2 pump gene have been reported in human patients,
erate states of cell suffering that could culminate, in cases of distributed along all the pump sequence without evidence of
massive and protracted Ca2+ overload, in the death of the cell; clustering in specific regions.49 Impaired ability to transport Ca2+
this is the meaning of the ambivalence concept. Having chosen and increased protein instability are considered the main conse-
Ca2+ as a determinant for function, cells undoubtedly benefit quence of the mutations; however, the fact that the symptoms are
from its unlimited availability in their surroundings; however, the confined to the skin suggests that specific elements of Ca2+
choice forces cells to live in a permanent condition of controlled homeostasis in keratinocytes are important to the pathogenesis
risk. Increases of Ca2+ significantly greater than normal levels can of the disorder. Alternatively, compensatory mechanisms for the
be handled by cells if the increase lasts for a limited time, because SPCA pump defect could act in other tissues. No cardiac pheno-
of the existence of the uptake system of mitochondria, which can types have been described in patients carrying the SERCA2
efficiently sequester for a while large amounts of Ca2+.43 However, pump mutations, even if the large majority of patients carry
if the overload situation persists, the fate of the cell becomes mutations in the region common to SERCA2a and 2b pump
sealed. Apart from these extreme and obvious cases of Ca2+ catas- isoforms. Darier disease shares properties with Hailey-Hailey
trophe, the cellular homeostasis of Ca2+ can also become deregu- disease, a genetic skin disorder caused by SPCA pump mutations
lated in less dramatic ways by defects in the individual participants (discussed later), indicating that the maintaining of a proper Ca2+
in the Ca2+ controlling operation. These defects are compatible pumping activity is of particular relevance to the epidermis.
with the continuation of cell life, but generate cell discomfort Nongenetic disease phenotypes linked to altered SERCA
phenotypes with diverse degrees of severity. Among them, those expression or activity have also been described; those associated
generated by defects of the Ca2+ pumps are currently receiving with heart failure and cancer are the best documented. A consen-
increased attention. Several disease phenotypes caused by genetic sus has been reached on the reduction in SERCA2a pump level
and nongenetic defects of the Ca2+ pumps have been described. or activity in failing hearts of both animal models and patients;
The problem with the nongenetic alterations is the frequent dif- however, the mutations have been ascribed mostly to PLB rather
ficulty of deciding whether the detected defects are really the sole than the SERCA pump. A single report has described mutations
cause of the disease phenotype or whether they are primitive or in the SERCA2 gene that can predispose to heart failure.50 The
secondary. By contrast, the pathologic genetic phenotypes are relationship between cancer and the SERCA2 and SERCA3
certainly causative and are mechanistically well defined. The dis- pumps has attracted increasing interest in the last years. Numer-
cussion will thus focus essentially on these phenotypes, with a ous alterations in the SERCA2 gene have been found in patients
discussion of the nongenetic conditions limited to some special with colon or lung cancer. However, as is frequently the case with
cases. nongenetic alteration of Ca2+ pumps, it is not clear whether they
are primitive and directly involved in the early events of the
carcinogenesis.
SERCA Pump As for SERCA3, its expression becomes progressively lost
during the multistep process of colon tumorigenesis, and its
Two human diseases associated with genetic mutations in the expression levels are inversely proportional to the loss of differ-
SERCA pump have been described, but numerous nongenetic entiation of the lesions. Finally, studies on human liver tumors
pathologic conditions (e.g., heart failure, cancer development, have shown that the hepatitis B virus DNA could be integrated
diabetes) have been associated with its malfunction or decreased in the SERCA1 pump gene causing the expression of a truncated
level of expression. Humans and some large animals develop a mutant protein with consequent decrease in ER Ca2+ content and
similar muscular disease, termed Brody disease in humans and increase in apoptosis.51
congenital pseudomyotonia in Chianina cattle, as a consequence
of mutations in the gene of the SERCA1 pump. Brody disease is
a rare recessive myopathy characterized by impaired relaxation, SPCA Pump
painless muscular cramps, and muscle stiffness following exercise.
The clinical diagnosis is difficult because the symptoms are het- Hailey-Hailey disease is the only known genetic disease associ-
erogeneous, but sometimes muscle biopsies show reduced ated with an SPCA1 pump mutation. The disease has a dominant
SERCA1 pump expression. Six different mutations in the autosomal inheritance and is phenotypically characterized by the
SERCA1 pump have been identified, four of them introducing a increased propensity to the formation of skin lesions, mainly at
premature stop codon that truncates the pump, creating variants the site of sweating and friction. The lesions are due to the loss
with higher instability.44,45 of adhesion of keratinocytes and to the abnormal keratinization
In the deficiency of SERCA1 pump function in skeletal of the skin. At least 87 different mutations have been described,
muscles observed in Chianina cattle affected by pseudomyotonia, located all along the SPCA pump sequence.52 Some mutations
the SERCA1 pump activity was decreased by approximately 70%. are responsible for reduced Ca2+ transport activity, but the large
Linkage analysis has shown that a mutation in the SERCA1 pump majority lead to protein instability. When the mutant SPCA
was associated to the phenotype, and subsequent mutation analy- pump is overexpressed in model cells, its level of protein expres-
sis has revealed its association with a missense Arg164His muta- sion is extremely low, despite normal levels of mRNA, and correct
tion in exon 6 of the gene.46 Arg164 is a strongly conserved localization to the Golgi compartment.53,54 As mentioned earlier,
residue in the SERCA1 pumps, but its role in the function of the Hailey-Hailey disease shares phenotypic properties with Darier
pump has not been defined. Another mutation (Arg559Cys) in disease; however, at variance with the keratinocytes of patients
the SERCA1 pump has been described in Belgian Blue Cattle with Darier disease, the keratinocytes of patients with Hailey-
affected by congenital muscular dystonia 147 and in a Dutch Hailey disease show an abnormal response to extracellular Ca2+,
Improved Red and White cross-breed calf.48 The disease pheno- possibly because the upregulation of the SPCA1 pump in Darier
type is the same, suggesting that the loss of SERCA pump activity disease keratinocytes could compensate for the SERCA dysfunc-
could be a common element in these different myopathies. tion. Hailey-Hailey disease is essentially benign, but squamous
Another disease associated to SERCA pump mutations is cell carcinoma can develop from the skin lesions. Interestingly,
Darier disease, a rare autosomal human dominant skin disorder in mice the loss of SPCA pump function causes skin cancer, but
not the acantholytic skin disorder it causes in humans. This Finally, a disease phenotype has been identified recently for a
5
finding suggests that elements favoring the survival or the apop- defect in the PMCA3 pump gene in humans. The PMCA3 pump
totic death depending on the keratinocyte type could have a role is the least known of the PMCA isoform. As mentioned, it is
in the development of the disease. essentially restricted to brain, although it is also found in rat
Recently, linkage analysis on chromosome 16q has revealed skeletal muscles; it is particularly abundant in the choroid plexus.
an association between the SPCA2 gene and specific language A study using X-exome sequencing has recently identified a mis-
impairment (SLI), a common developmental disorder character- sense mutation in the calmodulin binding domain of the PMCA3
ized by difficulties in language acquisition in individuals pump in a family with X-linked congenital cerebellar ataxia.63 A
with otherwise normal development.55 The linkage suggests a molecular study of the mutant PMCA3 pump expressed in model
possible role of the SPCA pump in the context of language cells has revealed its impaired ability to extrude Ca2+. A missense
impairment. mutation (Tyr543Met) in the PMCA3 pump gene has also
been detected in human pancreatic cancer cells.64 The effects of
the mutation on the activity of the pump have not been
PMCA Pump investigated.
Several genetic pathologies linked to the dysfunction of the

PMCA pumps have been described; they affect animals (mice) and
humans, and are either spontaneous (e.g., those in humans) or Conclusions
induced by the ablation, partial disruption, or mutations of PMCA
genes. The first described disease phenotype related to a PMCA The pumps that transport Ca2+ across membranes control the
pump defect is a form of hereditary deafness that involves the homeostasis of Ca2+ in all animal cells. In those of nonexcitable
PMCA2 isoform of the pump, which is abundantly expressed in tissues, which are not exposed to the periodical increases of cyto-
the stereocilia of the hair cells of the Corti organ in the inner ear. solic Ca2+ to levels that can reach the micromolar range, they are
As mentioned previously, PMCA2 has a peculiar characteristic: the most important or even the sole systems that maintain cyto-
high activity even in the absence of calmodulin. This property solic Ca2+ at the physiologic nanomolar level. They can do so
evidently satisfies the necessity of maintaining a constant flow of because they interact with Ca2+ with the appropriate high affinity.
Ca2+ from the stereocilia to the endolymph that bathes them. The In excitable tissues, the ejection of Ca2+ from the cytosol is instead
tight control of the homeostasis of Ca2+ in the endolymph is essen- mostly performed by a larger, lower-affinity system, the Na/Ca
tial for the functioning of the stereocilia bundle that gates mecha- exchanger. Ca2+ pumps still exist in the internal membranes of
noelectrical channels through which K+ (and Ca2+) flow into the excitable cells, and even in their plasma membrane. Their role in
hair cell to generate (or modulate) the acoustic signals. The the control of Ca2+ homeostasis, at least at the peak of the Ca2+
PMCA2 pump isoform of the stereocilia is the wa variant,56 which transients that are essential for the physiology of these cells, is
has the site A insert that targets it to the stereocilia, and is overshadowed by the activity of the exchanger. The case of the
C-terminally truncated to depress even further the calmodulin heart is special. PMCA pumps have been conclusively docu-
response with respect to the wild type PMCA2 pump. The first mented in heart sarcolemma, but their role in the ejection of bulk
indication of the role of the PMCA2 pump in the inner ear came Ca2+ from the sarcoplasm, especially in phase with the contraction–
from the deafness phenotype caused by the ablation of its gene.57 relaxation cycle, has always been considered unrealistic. Recent
Several spontaneous and induced mutations of the mouse gene evidence, however, indicates that the sarcolemmal Ca2+ pump
were then described that confirmed the role of the PMCA 2 wa (isoform 4) has a Ca2+ signaling role that is quantitatively unim-
variant in the function of the outer hair cells.9 Two human families portant, as expected in the control of bulk sarcoplasmic Ca2+, but
with an hereditary deafness phenotype caused by two different that is still critical to the contraction–relaxation cycle in more
point mutations in the PMCA2 gene were then also described.58,59 subtle ways and is linked to the modulation of the cyclic nucleo-
The defect of the PMCA2 pump was analyzed molecularly and tide balance. Perhaps, a role for the PMCA pumps in Ca2+ signal-
was characterized as a specific impairment of the long-term ability ing not immediately linked to the bulk ejection of Ca2+ will in
of a mutant pump to export Ca2+ from the cell.59,60 the future be extended to other excitable cells as well.
The inhibition of sperm motility and the male infertility phe-
notype caused by the ablation of the PMCA4 gene in mice have
been already mentioned. The ablation of the PMCA4 gene,
however, also has important effects on the function of the heart. Acknowledgments
The contractility and the amplitude of the Ca2+ transient linked
to the L-channel activity were increased in the cardiomyocytes The authors thank Dr. C. Toyoshima (Tokyo, Japan) from pro-
of PMCA 4 null mice,30 supporting the suggestions for a role of viding the image of the SERCA pump structure of Figure 5-2
the PMCA4 in the regulation of heart contractility. In addition, and the illustration in Figure 5-3, Dr. L. Raeymaekers (Leuven,
the ablation of the PMCA4 gene also impaired phasic contraction Belgium) for providing the structural model of the SPCA pump
and caused apoptosis in the smooth muscle of the portal vein, but in Figure 5-4, and Dr. S. Pantano (Montevideo, Uruguay) from
only in mice carrying a null mutation in one copy of the PMCA1 providing the PMCA for Figure 5-5.
pump gene.28 The original work by the authors has been supported over the
A recent Korean genome-wide association study aimed at years by grants from the Italian Ministry of University and
identifying genetic factors that influence blood pressure and Research (FIRB2001 to E.C., PRIN 2003, 2005, 2008 to M.B),
hypertension risk has located the most significant single nucleo- the Telethon Foundation (Project GGP04169), the Italian
tide polymorphism in the gene for the PMCA1 pump.61,62 Based National Research Council (CNR), the University of Padova
on the observation mentioned previously suggesting that the (Progetto di Ateneo 2008, CPDA082825) to M.B. and the FP6
PMCA1 pump gene can function as a modifier locus for program of the European Union (FP6 Network of Excellence
the phasic vein contraction linked to the PMCA4 pump gene, the NeuroNe, LSH-2003-2.1.3-3, Integrated Project Eurohear), the
Korean authors looked for a gene-gene interaction between vari- Human Frontier Science Program Organization, the Fondazione
ants of the PMCA1 and PMCA4 pump genes in the genome- Cariparo [Progetti di Eccellenza 2008-2009, ERA-Net Neuron
wide association dataset. They found only modest evidence. (grant nEUROsyn 2008) to E.C.
21. Vandecaetsbeek I, Vangheluwe P, Raeymaekers L, pump isoforms in intact cells. J Biol Chem
References et al: The Ca2+ pumps of the endoplasmic 278:24500–24508, 2003.
reticulum and Golgi apparatus. Cold Spring Harb 40. Preiano BS, Guerini D, Carafoli E: Expression and
1. Pedersen P, Carafoli E: Ion motive ATPases. I. Perspect Biol 3, 2011. functional characterization of isoforms 4 of the
Ubiquity, properties, and significance to cell func- 22. Missiaen L, Van Acker K, Van Baelen K, et al: plasma membrane calcium pump. Biochemistry
tion. Trends Biochem Sci 12:146–150, 1987. Calcium release from the Golgi apparatus and 35:7946–7953, 1996.
2. Axelsen KB, Palmgren MG: Evolution of substrate the endoplasmic reticulum in HeLa cells stably 41. Elwess NL, Filoteo AG, Enyedi A, et al: Plasma
specificities in the P-type ATPase superfamily. expressing targeted aequorin to these compart- membrane Ca2+ pump isoforms 2a and 2b are
J Mol Evol 46:84–101, 1998. ments. Cell Calcium 36:479–487, 2004. unusually responsive to calmodulin and Ca2+.
3. Toyoshima C, Nakasako M, Nomura H, et al: 23. Pavlikova M, Tatarkova Z, Sivonova M, et al: J Biol Chem 272:17981–17986, 1997.
Crystal structure of the calcium pump of sarcoplas- Alterations induced by ischemic preconditioning 42. Hilfiker H, Guerini D, Carafoli E: Cloning and
mic reticulum at 2.6 A resolution. Nature 405:647– on secretory pathways Ca2+-ATPase (SPCA) gene expression of isoform 2 of the human plasma mem-
655, 2000. expression and oxidative damage after global cere- brane Ca2+ ATPase. Functional properties of the
4. de Meis L, Vianna AL: Energy interconversion by bral ischemia/reperfusion in rats. Cel Mol Neuro- enzyme and its splicing products. J Biol Chem
the Ca2+-dependent ATPase of the sarcoplasmic biol 29:909–916, 2009. 269:26178–26183, 1994.
reticulum. Annu Rev Biochem 48:275–292, 24. Pizzo P, Lissandron V, Capitanio P, et al: Ca(2+) 43. Carafoli E: The interplay of mitochondria with
1979. signalling in the Golgi apparatus. Cell Calcium calcium: An historical appraisal. Cell Calcium
5. Hasselbach W, Makinose M: [The calcium pump 50:184–192, 2011. 52:1–8, 2012.
of the “relaxing granules” of muscle and its depen- 25. Caroni P, Carafoli E: An ATP-dependent Ca2+- 44. Odermatt A, Barton K, Khanna VK, et al: The
dence on ATP-splitting.]. Biochem Z 333:518–528, pumping system in dog heart sarcolemma. Nature mutation of Pro789 to Leu reduces the activity of
1961. 283:765–767, 1980. the fast-twitch skeletal muscle sarco(endo)plasmic
6. Thastrup O, Cullen PJ, Drobak BK, et al: Thapsi- 26. Carafoli E: Biogenesis: plasma membrane calcium reticulum Ca2+ ATPase (SERCA1) and is associ-
gargin, a tumor promoter, discharges intracellular ATPase: 15 years of work on the purified enzyme, ated with Brody disease. Hum Genet 106:482–491,
Ca2+ stores by specific inhibition of the endoplas- FASEB J 8:993–1002, 1994. 2000.
mic reticulum Ca2(+)-ATPase. Proc Natl Acad Sci 27. Guerini D, Zecca-Mazza A, Carafoli E: Single 45. Odermatt A, Taschner PE, Khanna VK, et al:
U S A 87:2466–2470, 1990. amino acid mutations in transmembrane domain 5 Mutations in the gene-encoding SERCA1, the fast-
7. Seidler NW, Jona I, Vegh M, et al: Cyclopiazonic confer to the plasma membrane Ca2+ pump prop- twitch skeletal muscle sarcoplasmic reticulum
acid is a specific inhibitor of the Ca2+-ATPase of erties typical of the Ca2+ pump of endo(sarco) Ca2+ ATPase, are associated with Brody disease.
sarcoplasmic reticulum. J Biol Chem 264:17816– plasmic reticulum. J Biol Chem 275:31361–31368, Nat Genet 14:191–194, 1996.
17823, 1989. 2000. 46. Drogemuller C, Drogemuller M, Leeb T, et al:
8. Oldershaw KA, Taylor CW: 2,5-Di-(tert-butyl)- 28. Okunade GW, Miller ML, Pyne GJ, et al: Targeted Identification of a missense mutation in the bovine
1,4-benzohydroquinone mobilizes inositol 1,4,5- ablation of plasma membrane Ca2+-ATPase ATP2A1 gene in congenital pseudomyotonia of
trisphosphate-sensitive and -insensitive Ca2+ (PMCA) 1 and 4 indicates a major housekeeping Chianina cattle: an animal model of human Brody
stores. FEBS Lett 274:214–216, 1990. function for PMCA1 and a critical role in disease. Genomics 92:474–477, 2008.
9. Brini M, Carafoli E: Calcium pumps in health and hyperactivated sperm motility and male fertility 47. Charlier C, Coppieters W, Rollin F, et al: Highly
disease. Physiol Rev 89:1341–1378, 2009. for PMCA4. J Biol Chem 279:33742–33750, effective SNP-based association mapping and man-
10. Moller JV, Nissen P, Sorensen TL, et al: Transport 2004. agement of recessive defects in livestock. Nature
mechanism of the sarcoplasmic reticulum Ca2+ 29. Schuh K, Cartwright EJ, Jankevics E, et al: Plasma genetics 40:449–454, 2008.
-ATPase pump. Current opinion in structural membrane Ca2+ ATPase 4 is required for sperm 48. Grunberg W, Sacchetto R, Wijnberg I, et al: Pseu-
biology 15:387–393, 2005. motility and male fertility. J Biol Chem 279:28220– domyotonia, a muscle function disorder associated
11. Toyoshima C: How Ca2+-ATPase pumps ions 28226, 2004. with an inherited ATP2A1 (SERCA1) defect in a
across the sarcoplasmic reticulum membrane. 30. Mohamed TM, Oceandy D, Zi M, et al: Plasma Dutch Improved Red and White cross-breed calf.
Biochim Biophys Acta 1793:941–946, 2009. membrane calcium pump (PMCA4)-neuronal Neuromusc Disorder 20:467–470, 2010.
12. Periasamy M, Bhupathy P, Babu GJ: Regulation nitric-oxide synthase complex regulates cardiac 49. Dhitavat J, Macfarlane S, Dode L, et al: Acrokera-
of sarcoplasmic reticulum Ca2+ ATPase pump contractility through modulation of a compart- tosis verruciformis of Hopf is caused by
expression and its relevance to cardiac muscle mentalized cyclic nucleotide microdomain. J Biol mutation in ATP2A2: evidence that it is allelic to
physiology and pathology. Cardiovasc Res 77:265– Chem 286:41520–41529, 2011. Darier’s disease. J Invest Dermatol 120:229–232,
273, 2008. 31. Niggli V, Penniston JT, Carafoli E: Purification of 2003.
13. Kranias EG, Hajjar RJ: Modulation of the (Ca2+-Mg2+)-ATPase from human erythro- 50. Schmidt AG, Haghighi K, Frank B, et al: Polymor-
Cardiac Contractility by the Phopholamban/ cyte membranes using a calmodulin affinity phic SERCA2a variants do not account for inter-
SERCA2a Regulatome. Circ Res 110:1646–1660, column. J Biol Chem 254:9955–9958, 1979. individual differences in phospholamban-SERCA2a
2012. 32. Elshorst B, Hennig M, Forsterling H, et al: NMR interactions in human heart failure. J Mole Cell
14. Traaseth NJ, Ha KN, Verardi R, et al: Structural solution structure of a complex of calmodulin with Cardiol 35:867–870, 2003.
and dynamic basis of phospholamban and sarco- a binding peptide of the Ca2+ pump. Biochemistry 51. Chami M, Gozuacik D, Saigo K, et al: Hepatitis B
lipin inhibition of Ca(2+)-ATPase. Biochemistry 38:12320–12332, 1999. virus-related insertional mutagenesis implicates
47:3–13, 2008. 32a. Tidow H, Poulsen LR, Andreeva A, et al: A SERCA1 gene in the control of apoptosis. Onco-
15. Gramolini AO, Trivieri MG, Oudit GY, et al: bimodular mechanism of calcium control in gene 19:2877–2886, 2000.
Cardiac-specific overexpression of sarcolipin in eukaryotes. Nature 491:468–472, 2012. 52. Ikeda S, Shigihara T, Mayuzumi N, et al: Muta-
phospholamban null mice impairs myocyte func- 33. Falchetto R, Vorherr T, Brunner J, Carafoli E: The tions of ATP2C1 in Japanese patients with Hailey-
tion that is restored by phosphorylation. Proc Natl plasma membrane Ca2+ pump contains a site that Hailey disease: intrafamilial and interfamilial
Acad Sci U S A 103:2446–2451, 2006. interacts with its calmodulin-binding domain. phenotype variations and lack of correlation with
16. Vandecaetsbeek I, Trekels M, De Maeyer M, et al: J Biol Chem 266:2930–2936, 1991. mutation patterns. J Invest Dermatol 117:1654–
Structural basis for the high Ca2+ affinity of the 34. Falchetto R, Vorherr T, Carafoli E: The 1656, 2001.
ubiquitous SERCA2b Ca2+ pump. Proc Natl Acad calmodulin-binding site of the plasma membrane 53. Dode L, Andersen JP, Leslie N, et al: Hovnanian,
Sci U S A 106:18533–18538, 2009. Ca2+ pump interacts with the transduction domain Dissection of the functional differences between
17. Gorski PA, Trieber CA, Lariviere E, et al: Trans- of the enzyme. Protein Sci 1:1613–1621, 1992. sarco(endo)plasmic reticulum Ca2+-ATPase
membrane helix 11 is a genuine regulator of the 35. Niggli V, Adunyah ES, Carafoli E: Acidic phospho- (SERCA) 1 and 2 isoforms and characterization of
endoplasmic reticulum Ca2+ pump and acts as a lipids, unsaturated fatty acids, and limited prote- Darier disease (SERCA2) mutants by steady-state
functional parallel of beta-subunit on alpha- olysis mimic the effect of calmodulin on the and transient kinetic analyses. J Biol Chem
Na+,K+-ATPase. J Biol Chem 287:19876–19885, purified erythrocyte Ca2+ - ATPase. J Biol Chem 278:47877–47889, 2003.
2012. 256:8588–8592, 1981. 54. Fairclough RJ, Lonie L, Van Baelen K, et al:
18. Li Y, Camacho P: Ca2+-dependent redox modula- 36. Ortega C, Ortolano S, Carafoli E: The plasma Hailey-Hailey disease: identification of novel
tion of SERCA 2b by ERp57. J Cell Biol 164:35– membrane calcium pump. In Krebs J, Michalak M, mutations in ATP2C1 and effect of missense muta-
46, 2004. editors: Calcium: A Matter of Life or Death, 41, tion A528P on protein expression levels. J Invest
19. Daiho T, Yamasaki K, Saino T, et al: Mutations of 2007, Springer, pp 179–197. Dermatol 123:67–71, 2004.
either or both Cys876 and Cys888 residues of sar- 37. Strehler EE, Zacharias DA: Role of alternative 55. Newbury DF, Winchester L, Addis L, et al: CMIP
coplasmic reticulum Ca2+-ATPase result in a com- splicing in generating isoform diversity among and ATP2C2 modulate phonological short-term
plete loss of Ca2+ transport activity without a loss plasma membrane calcium pumps. Physiol Rev memory in language impairment. Am J Hum
of Ca2+-dependent ATPase activity. Role of the 81:21–50, 2001. Genet 85:264–272, 2009.
CYS876-CYS888 disulfide bond. J Biol Chem 38. Chicka MC, Strehler EE: Alternative splicing 56. Grati M, Aggarwal N, Strehler EE, et al: Molecular
276:32771–32778, 2001. of the first intracellular loop of plasma membrane determinants for differential membrane trafficking
20. Arvanitis DA, Vafiadaki E, Fan GC, et al: Histidine- Ca2+-ATPase isoform 2 alters its membrane tar- of PMCA1 and PMCA2 in mammalian hair cells.
rich Ca-binding protein interacts with sarcoplas- geting. J Biol Chem 278:18464–18470, 2003. J Cell Sci 119:2995–3007, 2006.
mic reticulum Ca-ATPase. Am Journal Physiol 39. Brini M, Coletto L, Pierobon N, et al: A compara- 57. Street VA, McKee-Johnson JW, Fonseca RC, et al:
Heart Circ Physiol 293:H1581–H1589, 2007. tive functional analysis of plasma membrane Ca2+ Mutations in a plasma membrane Ca2+-ATPase
gene cause deafness in deafwaddler mice. Nat molecular analysis of the pump defect. Cell 63. Zanni G, Calì T, Kalscheuerc V, et al: Mutation of
Genet 19:390–394, 1998. Calcium 50:569–576, 2011. plasma membrane Ca2+ ATPase isoform 3 in a
58. Schultz JM, Yang Y, Caride AJ, et al: Modification
of human hearing loss by plasma-membrane
calcium pump PMCA2. N Engl J Med 352:1557–
61. Cho YS, Go MJ, Kim YJ, et al: A large-scale
genome-wide association study of Asian popula-
tions uncovers genetic factors influencing eight
family with x-linked congenital cerebellar ataxia
impairs Ca2+ homeostasis, Proc Natl Acad Sci U S
A 109:14514–14519, 2012.
5
1564, 2005. quantitative traits. Nat Genet 41:527–534, 64. Jones S, Zhang X, Parsons DW, et al: Core signal-
59. Ficarella R, Di Leva F, Bortolozzi M, et al: A func- 2009. ing pathways in human pancreatic cancers revealed
tional study of plasma-membrane calcium-pump 62. Hong KW, Go MJ, Jin HS, et al: Genetic varia- by global genomic analyses. Science 321:1801–
isoform 2 mutants causing digenic deafness. Proc tions in ATP2B1, CSK, ARSG and CSMD1 loci 1806, 2008.
Natl Acad Sci U S A 104:1516–1521, 2007. are related to blood pressure and/or hypertension 65. Toyoshima C: Structural aspects of ion pumping by
60. Giacomello M, De Mario A, Lopreiato R, et al: in two Korean cohorts. J Hum Hypertens 24:367– Ca2+-ATPase of sarcoplasmic reticulum. Arch
Mutations in PMCA2 and hereditary deafness: a 372, 2010. Biochem Biophys 476:3–11, 2008.
Structural and Molecular Bases
of Sarcoplasmic Reticulum Ion
Channel Function 6
Héctor H. Valdivia
participate in E-C coupling, different processes link membrane

CHAPTER OUTLINE
depolarization to Ca2+ release in cardiac and skeletal muscle.
The Sarcoplasmic Reticulum 55 Some of these differences stem in part from the ultrastructural
arrangement of the muscle fiber’s organelles. It is well accepted
Methods of Recording Sarcoplasmic Reticulum
that E-C coupling in skeletal muscle is greatly dependent on Ca2+
Ion Channels 56 release from the SR, with little or no Ca2+ entry across the sar-
Molecular Structure of the Cardiac Ryanodine colemma during a single twitch. Accordingly, the SR of skeletal
Receptor (RyR2) 57 muscle is robust and contains large saccular enlargements,
whereas T-tubules, which are invaginations of the external mem-
Accessory Proteins of the RyR2 Channel 58 brane that carry the depolarizing stimulus to the interior of the
Ca2+ Regulation of RyR2 Channels 58 cell, are narrow (~20 to 40 nm) and elongated. By contrast, Ca2+
entry across the sarcolemma is an obligatory step for E-C cou-
Modulation of RyR2 Channel Function 61 pling in cardiac muscle.3 T-tubules, which not only carry the
Exogenous Ligands of RyR2 Channels and Drugs electrical stimulus to the interior of the cell but must serve also
That Affect SR Ca2+ Release 62 as Ca2+ reservoirs, are therefore larger in diameter (~200 nm) in
cardiac muscle. The SR appears thinner and less organized than
RyR2 Channels in Disease 64 in skeletal muscle, but exhibits multiple points of contact with
Monovalent Cation Channels 65 the sarcolemma and T-tubules.4 Depending on the animal species
and the type of cardiac cell, the SR volume can be as high as 12%
of cell volume (mouse atrium), to as little as 0.8% (finch ventri-
cle).1 These sharp differences likely reflect the relative impor-
tance of Ca2+ entry versus SR Ca2+ release for contraction among
The Sarcoplasmic Reticulum different cells: whereas mouse cardiomyocytes rely heavily on
Ca2+ release from the SR, finch (and all avian) ventricular cells
Structural Arrangement of the lack T-tubules altogether and thus rely almost entirely on Ca2+
Sarcoplasmic Reticulum fluxes across the sarcolemma.
The sarcoplasmic reticulum (SR) is a membrane-delimited intra-

cellular organelle present in striated muscle of almost all species.1 Molecular Players of Excitation-
Its name literally means “network of structures located in the Contraction Coupling
cytoplasm of a muscle fiber,” referencing the interconnected
network of tubules and vesicles that spans the sarcomere and In both cardiac and skeletal muscle, the L-type Ca2+ channels or
wraps up the contractile myofilaments. The SR is not continuous dihydropyridine receptors (DHPRs) are the voltage sensors of
with the external membrane, but recent data indicate that it is sarcolemma and T-tubules that initiate E-C coupling, and the SR
continuous with the nuclear envelope.2 In cardiac and skeletal Ca2+ release channels that provide the majority of Ca2+ for con-
muscle, the main function of the SR is to provide the majority of traction are also known as ryanodine receptors (RyRs).4 The manner
Ca2+ ions that are needed to activate the contractile proteins of in which DHPRs trigger RyRs to open is particular to each
the myofilaments, and to resequester Ca2+ from the myoplasm muscle. Because skeletal muscle fibers can twitch in the absence
to allow for relaxation. The Ca2+ concentration within the SR of external Ca2+, and Ca2+ release follows sarcolemmal depolariza-
([Ca2+]SR ) is approximately 1 mM, with a larger pool of Ca2+ tion with virtually no delay, a physical coupling between DHPRs
bound to calsequestrin and other Ca2+-binding proteins.1 There- and RyRs is thought to initiate E-C coupling in skeletal muscle.5
fore, when myofilaments are relaxed and bathed in cytoplasmic Indeed, identification of the structural domains of DHPRs and
milieu containing approximately 0.1 µM Ca2+, there is an approx- RyRs involved in their mechanical interaction continues in
imately 10,000-fold [Ca2+] gradient across the SR membrane. earnest,6 although the presence of proteins “sandwiched” between
The compartmentalization of muscle fibers into small (~2 µm) these channels has not been completely ruled out. By contrast,
structural-functional units (sarcomeres) ensures that Ca2+ diffu- RyRs of cardiac muscle are closely apposed but do not appear
sion from release sites is not a limiting step in muscle contraction. mechanically joined to DHPRs4; instead, they are activated by a
Similarly, the enveloping of myofilaments by the SR ensures small influx of external Ca2+ (the inward Ca2+ current provided
rapid reuptake of Ca2+, leaving rate-limit steps of relaxation to by DHPRs, or ICa), releasing greater amount of Ca2+ from the
contractile proteins. SR.7 This process, Ca2+-induced Ca2+ release (CICR),8 is a signa-
It is useful to compare the structural arrangement of the SR ture event of RyRs but is not exclusive to cardiac muscle. In fact,
in cardiac and skeletal muscle to facilitate discussion of the mech- CICR was first described in skeletal muscle,9 where it apparently
anisms that trigger Ca2+ release in these two muscles. The series serves to activate RyRs that are not coupled to DHPRs.10
of events by which depolarization of the sarcolemma generates a Again, the structural arrangement of DHPRs and RyRs is
mechanical contraction is termed excitation-contraction coupling congruent with the type of E-C coupling they exert in skeletal
(E-C coupling). Although common subcellular structures and cardiac muscle. In the former, a great proportion of DHPRs
55
appears organized as tetrads that rest on top of a single RyR is indispensable for the rational design of drugs for various RyR-
channel in the junctional SR, the specialized region of the SR linked diseases, would be extremely difficult without single-
that is proximal to T-tubules.4 Because RyRs are homotetrameric channel recordings in lipid bilayers. Nevertheless, it is important
protein complexes (discussed later), each DHPR in a tetrad is to recognize the limitations of this technique, most of which stem
presumed to interact with one RyR monomer. RyRs, on the other from the extremely difficult recreation of the conditions under
hand, are clustered in a paracrystalline lattice where they touch which RyRs operate in situ. In their intracellular environment,
each other at the corners.11 In this orderly array, every other RyR RyRs are activated by fast and transient Ca2+ stimuli (as opposed
interacts with a tetrad of DHPRs and is directly activated by the to stationary Ca2+ levels as is routinely used in lipid bilayer experi-
physical coupling mechanism; the DHPR-free RyR is presum- ments), which modify RyR activity importantly (discussed further
ably activated by CICR and cooperative interaction between later). Furthermore, RyRs seldom work in isolation, as in single-
RyRs.4,10 Multiple (i.e., 10 to 20) such DHPR-RyR interactions channel experiments, but are clustered in arrays that exhibit
occur in a couplon, the structural domain of the junctional SR cooperativity and synergism.11,12 Lastly, various accessory pro-
where E-C coupling actually takes place.12 In cardiac muscle, teins of the RyR control its response to cytosolic and luminal
DHPRs are irregularly and sparsely distributed (not forming Ca2+, and these proteins might not be present in reconstituted
tetrads) in a couplon, but RyRs keep their lattice array, thus channels.
changing the stoichiometry and the mode of interaction of these
two channels. A typical couplon in cardiac myocytes can contain
approximately 100 RyRs and only approximately 20 DHPRs.1 Confocal Imaging of RyR Ca2+ Fluxes
Thus, in principle, one DHPR injects Ca2+ into a couplon to
activate approximately 5 RyRs, but current estimates indicate that The optical detection of Ca2+ fluxes has provided new insight into
E-C coupling is normally effected with a high safety factor—that the working principles of the supramolecular Ca2+ signaling
is, only a portion of the existing DHPRs is needed to fully acti- system of many cells. The discrete, transient, and presumably
vate RyRs13 (discussed later ). elemental Ca2+ signaling events of cardiac myocytes, also known
In addition to the junctional SR, DHPRs and RyRs interact at as Ca2+ sparks, are indisputable signs of RyR gating in situ.18 The
the peripheral couplings, the points of contact between SR and introduction of fluorescein- and rhodamine-based Ca2+ indicators
sarcolemmal surface.4 Despite the fact that T-tubules contain the of high dynamic range and the arrival of low-cost versatile confo-
highest density of DHPRs, couplons of the peripheral couplings cal microscopes greatly facilitated the discovery of Ca2+ sparks,
apparently have DHPR/RyR stoichiometry that is similar to those first detected in ventricular myocytes19 and later in smooth20 and
of junctional SR couplons. Thus, RyRs of peripheral couplings are skeletal21 muscle cells. The low myoplasmic [Ca2+] of resting cells
expected to release Ca2+ by a model similar to that described keeps the open probability (Po) of RyRs extremely low; still, RyR
earlier for RyRs of junctional SR, although this has not been channels open with a finite rate that depends on several factors
tested rigorously. A different mechanism, however, must operate (most notably, [Ca2+] on the cytoplasmic and luminal side of the
to activate RyRs in the corbular SR, a basketlike form of SR that RyR), giving rise to spontaneous Ca2+ sparks. An estimate of
is connected to the SR network at one point only, and completely spontaneous spark rates in ventricular myocytes was initially 100/
lacking T-tubular or sarcolemmal contacts. Activation of corbular cell/sec,18 which suggested an opening rate for RyRs of 10−4 s−1,
RyR channels probably requires a combination of Ca2+ diffusion assuming that a typical ventricular myocyte contains approxi-
from release sites and a propagating wave of CICR. Corbular SR mately 1 million RyRs.19 However, spark rates vary widely among
can contain up to approximately 35% of the total RyRs and is investigators apparently using the same conditions and can even
particularly prominent in atrial cells and Purkinje fibers.14 be influenced by experimental artifacts such as cell damage during
isolation, making this estimate less reliable. Initially, Ca2+ sparks
were thought to originate from the opening of a single RyR,18
Methods of Recording Sarcoplasmic but the unitary RyR channel conductance in near physiologic
conditions (~0.5 pA)22 appears to be too low to deliver the
Reticulum Ion Channels fluorescence signal mass of a typical Ca2+ spark.23 Additional
studies detected smaller RyR-originated Ca2+ signals (e.g., “Ca2+
Reconstitution of Sarcoplasmic Reticulum quarks,”24 “Ca2+ embers,”25 “Ca2+ syntilla,”26), somewhat demot-
Channels in Lipid Bilayers ing the elemental adjective bestowed on Ca2+ sparks as indivisible
events of E-C coupling in cardiac cells. Because Ca2+ sparks typi-
Advancements in confocal imaging techniques have started to cally give rise to a twofold increase in fluorescence intensity in
shed new light on SR ion channels in their cellular environment, an area of approximately 2 µm, it is likely that they result from
but at present the majority of their functional and biochemical the coordinated opening of a portion of (or all) RyRs that are
properties have been defined using isolated SR vesicles. Homog- clustered in a couplon (~100). Regardless of the number of RyRs
enization of cardiac and skeletal muscle produces fragmented SR that intervene to form a Ca2+ spark, these Ca2+ signaling events
that can be partially segregated by sucrose density centrifugation have brought fresh insights into the mechanisms that modulate
into RyR-enriched “heavy” SR (corresponding to junctional SR the activity of RyRs in their intracellular environment.
and peripheral couplings), and Ca2+ pump–enriched “light” SR
(corresponding to longitudinal SR). Because RyRs are confined
to regions of the cell that are largely inaccessible to patch elec- [3H]Ryanodine Binding As an Index
trodes, reconstitution of heavy SR into artificial lipid bilayers of RyR Activity
remains the most versatile and effective method to characterize
the biophysical and pharmacologic properties of RyRs at the Measurements of RyR density and activity can be readily obtained
single channel level.15,16 A vast body of knowledge has already by performing [3H]ryanodine binding assays in purified SR ves-
been gained with this technique,16,17 which remains the method icles or even in whole-tissue homogenates. This is possible
of choice in situations that require precise control of modulators because of the high affinity and specificity of [3H]ryanodine for
on both faces of the channel (e.g., in cytosolic or luminal Ca2+ its receptor, which yield an excellent signal-to-noise ratio, and
titrations), and in measurements of unitary channel conductance because the alkaloid binds to open RyR channels only.27 The
under quasi-physiologic conditions. In addition, defining the binding of [3H]ryanodine is enhanced by activators of Ca2+ release
potency and mechanism (open- or closed-channel block), which (Ca2+, adenosine triphosphate [ATP], caffeine) and decreased by
Structural and Molecular Bases of Sarcoplasmic Reticulum Ion Channel Function 57
inhibitors of Ca2+ release (Mg2+, H+, calmodulin), suggesting that

Molecular Structure of the Cardiac Ryanodine
6
the alkaloid binds to a conformationally sensitive domain of the
RyR protein.28 Therefore, [3H]ryanodine can be used as a probe Receptor (RyR2)
of the functional state of the RyR channel. This approach has
2+
contributed to the purification of the RyR itself and to the iden- RyRs are the main pathway for Ca release from the SR, but they
tification of novel ligands, endogenous modulators, and accessory are not restricted to striated muscle. To gain the functional flex-
proteins of RyRs17,27,29; however, ryanodine also displays some ibility necessary to respond to different triggering signals, at least
disadvantages as ligand of RyRs. Mainly, ryanodine has slow three isoforms of RyR are expressed in mammals.17,29,31 RyR1 is
association and dissociation rates, which require long incubation expressed predominantly in fast- and slow-twitch skeletal muscle
times to reach equilibrium (>15 h at room temperature) and and in cerebellar Purkinje cells. RyR2 is found in cardiac muscle
predisposes the RyR to protein degradation. Furthermore, its (presumably the only isoform expressed there), but is also robustly
sluggishness to bind to its receptor renders ryanodine incapable expressed in the brain and in visceral and arterial smooth muscle.
of tracking dynamic changes in RyR activity, which are usually RyR3 is the least understood of the RyR isoforms and appears to
transient and swift (discussed later). Despite these drawbacks, play its most important role during development, although in
[3H]ryanodine binding is the assay of choice to measure the aver- mature cells it is found in the diaphragm, epithelial cells, brain,
aged activity of literally billions of RyRs at the same time and and smooth muscle. Several structural and functional character-
under multiple conditions. istics confer to RyRs, a distinctive earmark. RyRs are homotet-
ramers of large molecular size (~2 million Da); they form
Ca2+-gated Ca2+-permeable channels of large conductance,32 and
Electron Probe X-Ray Microanalysis of Ca2+ and they are distinctively affected by the plant alkaloid ryanodine.27
Other Ions Within the Sarcoplasmic Reticulum Elucidating the structure of RyRs has been difficult because of
the channel’s massive size; however, some details of RyR structure
Ca2+ movements across the SR can be measured in situ by elec- have been obtained through cryoelectron microscopy,6,33,34 com-
tron probe analysis of ultrathin cryosections of muscle rapidly parative modeling,35 and recently x-ray crystallography of small
frozen at different times in the contraction and relaxation cycle.30 RyR segments.36,37 In electron micrographs, RyRs are seen as
These pioneering experiments have provided critical data on the quatrefoil or cloverleaf-shaped structures,33,34,38 or in three-
spatial and temporal movement of Ca2+ and its countercharge dimensional renderings as mushroom-shaped structures, with a
ions (mainly K+ and Mg2+) across the distinct regions of SR large (27 × 27 × 12 nm) cytoplasmic assembly and a smaller
during the E-C coupling cycle. Foundational notions, such as the transmembrane “stalk” spanning approximately 6.5 nm from the
SR undergoing no charge deficit during Ca2+ release (and thus base of the cytoplasmic region and extending into the SR lumen38
maintaining no transmembrane potential) and the terminal cis- (Figure 6-1). The carboxyl-terminal segment crosses the SR
ternae being the main Ca2+ release site of the SR, were derived membrane as few as four and as many as 10 times (depending on
largely from electron probe analysis experiments.30 the model, although the consensus is for six transmembrane
Clamp
FKBP12
Handle 180°
Transmembrane
A B domain
90°
Cyt constriction
Inner branch
Cytoplasmic
domain (open) ion gate
Inner helix
Transmembrane Pore helix
domain
C 10 nm
Figure 6-1. Three-dimensional surface representations of the ryanodine receptor (RyR) channel. Surface representations of the RyR1 at 10 Å resolution as seen from the
T-tubule (A), the SR side (B), and from the side (orthogonal to the SR membrane) (C). The inset is the magnified region indicated by the square, at a higher density threshold,
and cut through the fourfold axis to better visualize the ion pathway. The cutting plane is indicated in blue.
(Courtesy M. Samsó, Virginia Commonwealth University, Richmond, Va.).

domains) and forms the Ca2+-permeable pore, whereas the bulk has been found to hold anchoring sites for PKA, PP2A, cyclic
of the protein (~90%) protrudes into the cytosol to bridge a 15- adenosine monophosphate (cAMP)-specific phosphodiesterase
to 20-nm gap between the SR and T-tubule membranes. The (PDE4D3) and CaMKII,50 underscoring the importance of RyR2
quatrefoil structure results from the symmetric arrangement of regulation by phosphorylation. Thus, the RyR2 protein is not
four identical subunits of approximately 5000 amino acids each; only capable of forming ordered multi-channel arrays,4,11 but is
therefore, a single tetrameric channel encompasses approximately also at the center of a massive macromolecular complex that
20,000 amino acids. includes numerous regulatory proteins. Although the intrinsic
function of each of these accessory proteins is known to some
extent, it is still unclear how this complex of proteins interplays
in situ to regulate Ca2+ release.
Accessory Proteins of the RyR2 Channel
The RyR2 channel can be viewed as a molecular switchboard that
integrates a multitude of cytosolic signals, such as dynamic and Ca2+ Regulation of RyR2 Channels
steady Ca2+ fluctuations, β-adrenergic stimulation (phosphoryla-
tion), oxidation, and metabolic states and that transduce these Regulation of Isolated RyR2 channels
cytosolic signals to the transmembrane domain to release appro-
priate amounts of Ca2+. Furthermore, Ca2+ release is critically Ca2+ is the most potent, versatile, and multifaceted regulator of
influenced by luminal (intra-SR) factors such as Ca2+ content and RyR2 channels. Ca2+ ions turn on, turn off, and permeate through
protein interactions, thus conferring RyR2 channels an additional the RyR2. Ca2+ binds to cytosolic and luminal domains of RyR2s
role as integrative switch-valves that offset cytosolic-luminal Ca2+ and regulates the activity of the channel in a concentration- and
imbalances. Most of the signal-decoding structures are integral time-dependent manner. Because RyRs work cooperatively in
domains of the RyR2 protein. In addition to the huge structural their native environment and are activated by ICa, which is an
tetrameric assembly being highly complex, RyR2 channels are extremely fast and transient Ca2+ stimulus, this mode of regula-
also capable of protein-protein interactions that allow them to tion is extremely complex.
bind, in some cases steadily and in other cases in a time- and It is necessary to discuss the effect of Ca2+ on isolated RyRs
Ca2+-dependent manner, to small and independently regulated before considering the intricacies of Ca2+ regulation of RyRs in
accessory proteins that add another layer of versatility (and com- situ. Under stationary (nonfluctuating) [Ca2+], the activity of
plexity) to regulation of Ca2+ release in vivo (Figure 6-2). The single RyR2 channels is a bell-shaped function of cytosolic [Ca2+]
best-known RyR2-interacting proteins are calmodulin, which (Figure 6-3) because of the presence of Ca2+-activating and inac-
tonically inhibits Ca2+ release39,40; FKBP12.6, which presumably tivating sites.17,29,51 Ca2+ in the range of 0.1 to 10 µM binds to at
stabilizes RyR2 closures41,42,43; sorcin, which inhibits Ca2+ release least one Ca2+-binding domain that activates RyR2s. Higher
in a Ca2+-dependent manner44-46; and the ternary complex triadin- [Ca2+] (100 µM to 3 mM) then inactivates the channel.51 The
junctin-calsequestrin, which “senses” luminal Ca2+ content and affinity and cooperativity of the activating and inactivating Ca2+-
modulates RyR2 activity by acting either as a direct channel binding sites vary greatly under the presence of other relevant
ligand or as an immediate source of releasable Ca2+.47,48 S100A1,49 modulators (e.g., ATP, Mg2+, H+) and constitute powerful mecha-
like calmodulin and sorcin, inhibits RyR2 more conspicuously nisms by which posttranslational modifications of the channel
when [Ca2+] is high; therefore these three proteins might play a protein (e.g., oxidation, phosphorylation, nitrosylation) modulate
role in Ca2+-mediated CICR termination. More recently, RyR2 Ca2+ release. Among the RyR isoforms, RyR2 is particularly
recalcitrant to Ca2+-dependent inactivation, requiring supraphys-
iologic [Ca2+] (~10 mM) for complete inactivation.17,29,51 For this
reason, the functional role of this process is questionable,
although its participation in CICR termination has not been
ruled out (discussed later). This overall picture represents the
Ca2+ response of RyR2 channels to stationary levels of Ca2+, but is
Ca2+
Sorcin (dimer) different from that obtained under dynamic (fast and transient)
PKA Ca2+ stimuli. In what are now landmark studies delineating CICR,
FKBP12.6 Ca2+ Fabiato8,52 observed that the magnitude of Ca2+ release was
mAKAP
dependent on the velocity of the Ca2+ pulse (d[Ca2+]/dt) applied
S100A1
Ca2+ to skinned cardiac fibers. Similarly, single-channel studies have
revealed that a fast Ca2+ stimulus elicits a transient increase in the
CaMKII Calmodulin Cytosol Po of RyR2s. If Ca2+ is applied rapidly but remains sustained for
PP2A
long periods, RyR2 Po will relax, or adapt, to a new steady state
that follows the activation curve obtained under stationary [Ca2+].
Thus, Po–[Ca2+] curves of greater magnitude emerge from titra-
Junctin Triadin tion of RyR2 activity with fast, calibrated Ca2+ pulses compared
SR lumen
with those obtained under similar but steady applications of Ca2+
(Figure 6-3).53-55 The implication of these findings is that there
are components of RyR2 activity that remain undetected in most
Calsequestrin Calsequestrin of the assays presently used, and these aspects of RyR regulation
might explain seemingly discrepant results, as in the case of RyR2
Figure 6-2. Three-dimensional model of the ryanodine receptor (RyR) supramo- phosphorylation (discussed later).
lecular complex. The faded blue structure is the cryo-electron microscopy surface
Regulation of RyR2 channels by luminal (intra-SR) Ca2+ has
representation of the RyR1 protein at 10 Å resolution as shown in Figure 1, C. The
multi-colored proteins are represented by their actual crystal structures. Junctin
gained preeminence as a control mechanism of Ca2+ release, but
and triadin were generated by Song et al.136 using homology modeling. Binding this process remains as complex as cytosolic Ca2+ regulation. The
sites for FKBP12.6 and calmodulin are based on cryo-EM data.33,34 The other protein crucial observation in ventricular myocytes is that, beyond a
interacting sites are idealized. Only some of the most relevant proteins that interact certain threshold, small changes in SR Ca2+ load result in far
with the RyR2 channel are shown. greater increases in Ca2+ release, with the relationship described
Figure 6-3. Modulation of RyR2 channel by Ca2+. (A) Activation of

A RyR2 by fast changes of the [Ca2+] surrounding the cytosolic side of
a 0.1 → 1 µM Ca2+
Cytosolic Ca2+ response

the channel. The resting [Ca2+] was 0.1 µM in all traces. Calibrated
steps increases of [Ca2+] were achieved by laser photolysis of caged
Ca2+ (nitrophenyl-ethylene glycol tetraacetic acid [NP-EGTA]), as
6
1.0
Peak described.54 The RyR2 openings were elicited by fast increase of
0.8 [Ca2+] to 1 µM (a and b) or to 10 µM (c) produced by single laser
Steady-state pulses. Mg2+ (1 mM) was present in b and c. Traces in all panels were
b 0.1 → 1 µM Ca2+ + 1 mM Mg2+ 0.6 recorded from the same channel. (d) Ensemble currents were gener-
P0
ated by the sum data of multiple sweeps. (e) The amplitude and time
0.4
course of the change in [Ca2+] in the cytosolic side of the channel.
0.2 (B) Ca2+-Po curves of RyR2 channels. Activity was measured at the
indicated stationary concentrations of Ca2+ (steady-state) and right
20 pA 0.0 after a fast Ca2+ pulse (“peak”), as in A. (C) Activation of RyR2 channel
c 0.1 → 10 µM Ca2+ + 1 mM Mg2+ by luminal [Ca2+].
8 7 6 5 4 3 2 1
B pCa (A, Modified from Valdivia HH, Kaplan JH, Ellis-Davies GC, et al: Rapid
adaptation of cardiac ryanodine receptors: modulation by Mg2+ and
1.0 Luminal Ca2+ response phosphorylation. Science 267:1997–2000, 1995.)
Peak 500 ms
d 0.8
a Steady-state
0.6 0.6
c
P0
P0
b 0.4
0.0
0.2
e
10 5 0.0
Ca2+ (µM)
c
a b 0.1 1 10
1 0.5
C Luminal [Ca2+]
0.1 Flash
as an inverse hyperbole with a high degree of cooperativity56 (see ICa and quickly terminates in intact cells. What counters the
Figure 6-3). This is clearly an adaptive mechanism that promotes inherently positive feedback of CICR? At least three mechanisms
greater Ca2+ release (and stronger contractions) under conditions have been proposed and, given the importance of CICR termina-
of increased Ca2+ uptake (such as occurs under β-adrenergic tion, it is likely that several mechanisms intervene (even redun-
stimulation) and suggests that RyR2 channels increase their Po in dantly) to avoid overflowing the cytosol with Ca2+.
response to elevated luminal [Ca2+]. The latter has been detected Ca2+-dependent inactivation of Ca2+ release was envisioned by
in single channel experiments,57,58 but the mechanism underlying Fabiato in his studies of CICR.3,8,52 The simplicity of this scheme
luminal Ca2+ regulation at the molecular level remains uncertain makes it elegant. It involves two types of Ca2+ sites controlling
and is possibly due to a combination of several of the following the RyR2 channel: a fast-action, low-affinity Ca2+ activation site
processes: (1) direct Ca2+ binding to RyR2 domains available only and a high-affinity but slow-action Ca2+ inactivation site.8 Accord-
from inside the SR58,59; (2) activation of cytosolic sites by the Ca2+ ingly, a fast Ca2+ stimulus evokes a transient burst of RyR2 activity
ions being permeated by the channel (feed-through mecha- because the channel is activated by Ca2+ acting on the activation
nism)60; and (3) binding of Ca2+ to calsequestrin and subsequent site and then shut down as Ca2+ slowly binds to the inactivation
regulation of RyR2 through junctin, triadin, or both.47,61,62 site. On the other hand, a slow Ca2+ stimulus binds to the higher-
affinity inactivation site and prevents channel openings, at least
until the stimulus is removed. At sustained and high Ca2+ levels,
Regulation of RyR2 Channels In Situ estimated by Fabiato to be approximately 100 µM, the inactiva-
tion sites are saturated, leaving the channel in an absorbing inac-
In each of the approximately 10,000 individual couplons that tivated state.52 Thus, this straightforward scheme allows for
occur in a typical ventricular myocyte, RyR2 channels are packed activation of Ca2+ release by the fast ICa and termination of CICR
forming a paracrystalline array that promotes synchronized Ca2+ by the [Ca2+] lingering in the couplon after Ca2+ release. Recovery
release.4,11 Each of these couplons is independently activated by (repriming) of RyR2 channels from this inactivated (refractory)
Ca2+ entering the cell through the juxtaposed DHPRs of the state requires removal of Ca2+ from the couplon, as expected to
T-tubules (junctional SR) or sarcolemma (peripheral couplings).1,4 occur on a beat-to-beat basis. As intelligible as it sounds, this
The rapid spread of sarcolemmal depolarization into the interior mechanism has encountered experimental and theoretical chal-
of the cell by the T-tubules ensures that all individual couplons lenges. First, SR Ca2+ release was shown to be independent of the
are simultaneously activated by ICa, generating a synchronized interval between two consecutive Ca2+ stimuli,64 which was not
wave of Ca2+ release that spreads quickly to virtually all corners expected if RyR2 channels were refractory for a certain time after
of the cardiomyocyte. In adult ventricular cells, ICa is insufficient the first stimulus. In addition, when a sustained Ca2+ stimulus was
to induce full contractions but is amplified threefold to eightfold used to inactivate the RyR2 channel, a subsequent Ca2+ stimulus
(depending on the species) by Ca2+ release from the SR.1 Because reactivated Ca2+ release,65 which again was unexpected of a refrac-
Ca2+ is simultaneously the input and output signal of RyR2s, tory process. Perhaps the most irreconcilable data come from
this amplification process (CICR) should intuitively be self- single-channel experiments in which RyR2 channels inactivate at
regenerating and all-or-none.63 In other words, the Ca2+ released [Ca2+] >1 to 3 mM, which are levels unlikely to be reached during
by an RyR2 channel should activate further the same channel or a normal contraction. Still, it is possible that at least some level
its neighbors in an apparently interminable cycle. However, this of Ca2+-dependent inactivation of Ca2+ release occurs in vivo, as
is not observed experimentally; instead, CICR is finely graded by mathematical models estimate the [Ca2+] in the dyadic cleft at
levels that are compatible with those required for partial inactiva- probability of triggering RyR2 channels to open also increases,
tion of single RyR2 channels.66 In addition, RyR2 channels in situ and this increases SR Ca2+ release and the amplitude of the [Ca2+]
may be more sensible to Ca2+ than those recorded under artificial transient. If [Ca2+]SR increases, the Po of RyR2s increases (see
environments. Figure 6-3) and SR Ca2+ release also increases. In this dynamic
SR Ca2+ depletion has attracted attention as a mechanism to environment, what is the net effect of blockers or agonists of
terminate CICR, but nagging issues remain. Although it is clear RyR2 channels? Combining experimental data and mathematical
that RyRs terminate Ca2+ release once the Ca2+ inside the SR modeling, the Eisner group74,75 has elegantly advanced the notion
drops to a certain threshold (~50% of the total Ca2+ inside the that potentiating or decreasing RyR2 Po alone does not have
store),67,68 the mechanisms underlying luminal Ca2+ regulation of lasting (steady-state) effects on the amplitude of the [Ca2+] tran-
RyRs remains unclear. In his classical studies that characterized sient. This notion appears counterintuitive at first, but it is rea-
calsequestrin (the major Ca2+-binding protein of the SR), Ikemoto sonable when all cellular Ca2+ fluxes that affect RyR2 activity are
et al.69 observed that the amount and the speed of Ca2+ release considered. Figure 6-4, A, shows the effect of caffeine, a RyR2
from SR vesicles depended directly on vesicular calsequestrin agonist, on [Ca2+] transients and the concomitant effect on sar-
content rather than Ca2+ content, thus implying that calseques- colemmal Ca2+ fluxes and [Ca2+]SR of rat ventricular myocytes.
trin “senses” the SR Ca2+ load and regulates the activity of RyRs. The addition of caffeine produces immediately an increase of the
Boding well with this notion, Ca2+ sparks lasted longer or termi- systolic Ca2+ transient. However, in the continued presence of
nated prematurely after overexpression or partial depletion of caffeine, the amplitude of the Ca2+ transient decreases progres-
calsequestrin levels in ventricular myocytes, respectively.70 Fur- sively until it reaches a steady state. Remarkably, the amplitude
thermore, purified RyR2 channels reconstituted in lipid bilayers of the Ca2+ transient in the new steady state is exactly the same
exhibited little activation by luminal [Ca2+], but the addition of as before the addition of caffeine. Thus, despite the maintained
calsequestrin and its “anchors” junctin and triadin restored effect of caffeine on the RyR2 channel, the Ca2+ transient relaxes
luminal Ca2+ sensitivity.57 Thus, the above data portrays calse- toward its control amplitude. The explanation of this apparent
questrin as an indispensable component of the signaling mecha- paradox lies on exquisite feedback mechanisms that operate in
nism that allows RyRs to terminate Ca2+ release upon [Ca2+]SR the cell. Before application of caffeine, net sarcolemmal Ca2+ flux
depletion; however, other data bestow little role on this protein. is equal (Ca2+ influx – Ca2+ efflux = 0; see Figure 6-4, B).
For example, RyR2 channels expressed in heterologous systems The addition of caffeine increases the amplitude of the Ca2+
(and bearing no calsequestrin) exhibit robust luminal Ca2+ regula- transient and as a result, Ca2+ efflux becomes larger than influx
tion and are activated by Ca2+ overload (store-overload induced (see Figure 6-4, B). A larger Ca2+ efflux leads to reduced [Ca2+]SR
Ca2+ release).58,59 In addition, in direct contradiction with some
of these data, recombinant RyR2 channels (calsequestrin-free)
reconstituted in lipid bilayers were activated by luminal [Ca2+],59
implying that luminal Ca2+ regulation is an intrinsic property of
the channel protein. As a result, it is clear that RyRs are capable
of sensing dropping levels of [Ca2+]SR to terminate Ca2+ release, Caffeine
but it is unclear whether they do it directly or through the
calsequestrin-junctin-triadin ternary complex. As derangement (i)(ii)
3
of RyR refractoriness leads to Ca2+-dependent arrhythmias and
[Ca2+]i 2
may be involved in heart failure progression,71,72 luminal Ca2+ (µmol/L)
regulation of RyRs appears to be critical in the overall scheme to 1
terminate CICR and likely depends on multiple factors. A 0
Stochastic attrition, the random closing of individual RyR2
channels, has been deduced from mathematical modeling as a
Net 4
potential mechanism to terminate CICR.63 High local [Ca2+] gra-
dients build up in the couplon immediately right after CICR, but sarcolemmal 0
Ca flux –4
they quickly dissipate.66 If the number of RyR2 channels in a (µmol/L) –8
couplon was only one, it should be obvious that a random closure B –12
of that one RyR2 channel could be responsible for Ca2+ release
termination (and dissipation of the [Ca2+] gradient). Stern63
inferred that stochastic attrition could terminate local Ca2+ release 110
SR Ca
if the number of RyR2 channels in a couplon was fewer than 10, content 55
but higher number of channels would decrease its probability (µmol/L)
precipitously. Because ultrastructural and functional data esti- C 0
mate this number to be close to 100,1,18 stochastic attrition as a
means to terminate local CICR appears unlikely, at least math-
ematically. However, RyR2 channels have the intrinsic ability to
join each other at their corners and form a paracrystalline 1.0
lattice.4,11 It is therefore possible that RyR2 channels in a couplon Fractional 0.5
gate in a coordinated, or coupled, manner73 and that the random release
closure of one channel promotes the closure of all channels in 0.0
the lattice. D
50 sec
Self-Regulation of Sarcoplasmic Figure 6-4. Transient effect of caffeine, a RyR2 channel agonist, on sarcoplasmic
reticulum (SR) Ca2+ release. (A) Intracellular Ca2+ transients before (i) and immedi-
Reticulum Ca2+ Release ately after (ii) application of 500 µM caffeine (indicated by the bar). (B) Net sarco-
lemmal Ca2+ flux per pulse calculated as Ca2+ influx minus Ca2+ efflux. (C) Calculated
As discussed earlier, RyR2 channels in their intracellular environ- time course of the SR Ca2+ content. (D) Fraction of SR Ca2+ released in each pulse.
ment are regulated by ICa and [Ca2+]SR, both of which undergo
dynamic changes during a single beat. If ICa increases, the (Reproduced with permission from Trafford, et al.74)
(see Figure 6-4, C), an effect that continues until the [Ca2+]SR is phosphorylation increases,54,82,83 decreases,54,84 or has no effect85
6
reduced to such a level that systolic Ca2+ amplitude becomes on RyR2 activity. Several factors preclude an easy interpretation
equal to control despite the potentiating effect of caffeine on of phosphorylation results. First, RyR2 channels contain multiple
RyR2 Po. The opposite effect occurs when caffeine is removed. phosphorylation sites that, depending on their phosphorylation
Initially, the systolic Ca2+ amplitude decreases to below the state, can attenuate or synergize the effect of the other sites, or
control level because [Ca2+]SR is lower than in controls, but RyR2 can require prior phosphorylation to activate the whole protein.
channels are no longer potentiated by caffeine. After a few con- To date, three phosphorylation sites have been recognized:
tractions, the amplitude of the Ca2+ transient is the same as the Ser2809 (mouse RyR2 nomenclature) was first identified by
initial value. In this scenario, caffeine produces transient effects Witcher et al.82 as a CaMKII site, and later, Wehrens et al. sug-
on systolic Ca2+ transients by modifying exclusively the fractional gested it as the only PKA site,86 although RyR2 from mice with
release (Ca2+ release/[ Ca2+]SR; see Figure 6-4, D). Thus, SR Ca2+ genetic ablation of this site can still be phosphorylated by PKA.77
release has direct and immediate effect on sarcolemmal Ca2+ Ser2815 and Ser2030 seem to be exclusively phosphorylated by
fluxes; this scheme precludes a persistent effect of RyR2 modula- CaMKII87 and PKA,88 respectively, although this exclusivity has
tors on SR Ca2+ release, because Ca2+ influx and efflux need to be not been rigorously tested in vivo. Second, as mentioned previ-
equal over the long run. This mode of regulation has important ously, RyR2 activity is strongly dependent on the speed of Ca2+
implications for interpreting maneuvers that presumably modu- application, which in turn can greatly influence the overall effect
late RyR2 activity in physiologic (phosphorylation) and patho- of phosphorylation. For example, PKA phosphorylation of iso-
logic states (discussed later). lated RyR2 increases a transient component of activity (peak
activation) but accelerates the rate of adaptation to a lower
steady-state level of activity (see Figure 6-3). In cellular settings,
this effect would translate into greater and faster rates of Ca2+
Modulation of RyR2 Channel Function release in response to a given Ca2+ entry. In agreement with this
notion, experiments in which SR Ca2+ load and ICa were kept
Phosphorylation constant showed that β-adrenergic stimulation of ventricular
myocytes accelerates the rate of Ca2+ release and has little effect
Sympathetic stimulation of the heart during exercise or under on the magnitude of the [Ca2+]i transient.89 A recent study sug-
emotional stress improves cardiac output by accelerating heart gested that differences in the redox state of RyR could account
rate and increasing the force of contractions. Activation of the for some the reported discrepancies.80 Although appealing, this
β-adrenergic receptor pathway by catecholamines triggers a hypothesis still needs to be corroborated. Thus, more work is
cascade of events that increases cAMP, which in turn activates needed to resolve the role of RyR2 phosphorylation in cardiac
PKA. PKA then phosphorylates several target proteins in the performance in general, and SR Ca2+ leak and heart failure in
sarcolemma, the SR, and myofilaments, notably, L-type Ca2+ particular. Further discussion of RyR2 phosphorylation appears
channels/DHPR, phospholamban, RyR2, and troponin I and C, under Role of RyR2 Channels in Heart Failure in this chapter.
leading to increased Ca2+ entry, enhanced CICR, and faster Ca2+
uptake and relaxation rates, all of which contribute to the inotro-
pic and lusitropic effects of β-adrenergic stimulation on the Oxidation and Nitrosylation
heartbeat.7 RyR2 channels are structural scaffolds for important
kinases and phosphatases,41,50 and in the heart, RyR2 is one of the RyR2 channels (as well as RyR1 channels) can be powerfully
first proteins to undergo phosphorylation during β-adrenergic modulated in vitro by oxidative modifications of thiol residues in
stimulation.76-78 Thus, RyR2 channels appear to be integral com- free cysteines, such as S-nitrosylation, S-glutathionylation, and
ponents of the multiprotein response that leads to increased disulfide oxidation.90-92 Both RyR2 and RyR1 channels contain
cardiac performance during β-adrenergic stimulation, but the approximately 100 cysteine residues per monomer, and although
extent of their participation in the fight-or-flight response many are suitable for modification, only a few appear highly
remains highly debated. Ca2+ release by RyR2 is steeply graded reactant to oxidants.93 In addition, whereas RyR2 and RyR1
by ICa and the SR Ca2+ load, both of which increase during activities depend on pO2, RyR2 does not appear to be activated
β-adrenergic stimulation because of PKA phosphorylation of or S-nitrosylated directly by nitric oxide (NO) but requires
L-type Ca2+ channels and phospholamban, respectively7; there- S-nitrosoglutathione.92 The functional response of each RyR
fore, it appears that RyR2 channels need only be responsive to isoform varies depending on the cysteine residue being modified
extrinsic cues (ICa and [Ca2+]SR) to increase Ca2+ release during and the type of oxidative species that targets it.91,92,94 In addition,
sympathetic stimulation, but this passive compliance would not oxidative modifications can also affect the binding of accessory
explain why RyR2 channels undergo direct protein phosphoryla- proteins. For example, exposing RyR1 to NO increases channel
tion, as observed experimentally.76-78 Some outstanding questions activity,95 an effect that is more pronounced in the presence of
include: What are the intrinsic changes in RyR2 function brought calmodulin, suggesting that S-nitrosylation of RyR1 leads to
about by PKA phosphorylation? Is this functional modulation calmodulin detachment and therefore reversal of the inhibitory
necessary and sufficient to modify global Ca2+ transients, cardiac effect of CaM over RyR1. The cellular effects of NO donors are
performance, and be a substrate for development of cardiac clear, but the molecular reactions underlying such effects are hard
arrhythmias and heart failure progression? These important to pinpoint. The NO donor S-nitroso-N-acetyl penicillamine,
questions are difficult to answer at present, partly because all which targets several EC-coupling proteins and increases inot-
potential functional outcomes have been attributed to RyR2 ropy of cardiac myocytes at low concentrations, but decreases it
phosphorylation. On one hand, Li et al.79 found that PKA phos- at high concentrations.96 Reduced glutathione could quickly react
phorylation of RyR2 has little functional relevance for diastolic with and scavenge NO in cardiac cells.92 Under these circum-
Ca2+ release if SR Ca2+ levels remain constant. On the other hand, stances, nitroso-glutathione or other small nitrosylated molecules
other groups have suggested that PKA phosphorylation of RyR2 would be responsible for RyR2 oxidation.92 It is also possible that
is so essential to intracellular Ca2+ homeostasis that derangement the close proximity of NO synthase (NOS)-3, xanthine oxidase,
of this process may be the basis for heart failure41,42,80 and cate- and RyR2 in cardiac caveolae creates a microenvironment capable
cholaminergic polymorphic ventricular tachycardia (CPVT) epi- of directly nitrosylating RyR2,97 although in physiologic condi-
sodes81 (discussed later). Between these two extremes, other tions the main targets of NOS seem to be other EC-coupling
results, mainly from in vitro experiments, imply that PKA proteins.98
Cytosolic Modulators binding106 that renders the onset of activation incompatible with
the timeframe of many cellular experiments. In addition, the
In addition to Ca2+, which is uncontestably the preeminent mod- recovery of SR function after treatment with ryanodine is difficult
ulator of RyR2 activity (discussed earlier), other cytosolic factors to assess because its slow dissociation rate106 makes the pharma-
have important roles in regulating Ca2+ release, especially in cologic effect essentially irreversible. Structurally speaking,
pathologic states. The concentration of several cytosolic modula- ryanodine is intrinsically minimized (see Table 6-1); thus, conju-
tors of RyR2 channels such as Mg2+, H+, and ATP does not gation of any of its major domains decreases substantially its
change substantially during normal beat-to-beat contractions or affinity and specificity toward RyRs, and fluorescent derivatives
in constant healthy conditions, but may do so under altered that could effectively define RyR location and function such as
metabolic states (e.g., diabetes or ischemia). As expected from the BODIPY-ryanodine, have had limited success.
identification of divalent cation- and nucleotide-binding sites in Scorpion toxins have traditionally been excellent sources of
the RyR2 channel protein, most assays of RyR2 channel activity ion channels ligands. Valdivia et al.107 found in the venom of
indicate that free Mg2+ (and H+) inhibit RyR2 channels, whereas selected scorpions a set of peptide toxins displaying high affinity
ATP increases their activity. Mg2+ inhibits RyR2 by at least two and exquisite selectivity against RyRs. Imperatoxin A (IpTxa), the
mechanisms, one being an ineffective occupation of Ca2+-activa- founder of this novel group of toxins, is a small (3.7-kDa), highly
tion sites and the other being an effective occupation of Ca2+- basic (pH 8.9), globular and thermostable peptide that activates
inactivation sites.99 ATP, however, binds to nucleotide-binding RyRs with high affinity (KD ≈ 5 to 10 nM) and specificity (no
sites on the channel protein and promotes Ca2+ release by acting other target proteins known to date).108 The presence of IpTxa in
as a ligand (i.e., without undergoing energy-yielding hydroly- scorpion venom was surprising because as a basic peptide (see
sis).100 The fall in pH and ATP levels (plus the concomitant rise Table 6-1), this ionized molecule was presumed to be incapable
in free Mg2+) that are characteristic of prolonged ischemia is of penetrating cellular membranes to reach its intended target.
therefore expected to depress RyR2 activity and decrease Ca2+ However, IpTxa, like Maurocalcin (another scorpion peptide
release. Intuitively, this could deteriorate cellular contractility homologous to IpTxa)109 permeates cellular membranes of intact
even more; however, the negative effect of these modulators on cardiac myocytes and mobilizes intracellular Ca2+ with remark-
RyR2s may actually be protective for the heart by means of able speed and with several degrees of potency,110 in effect enter-
reducing energy consumption. This potential benefit needs to be ing the field as the first cell-penetrating peptide RyR-specific
assessed in experiments with intact cells and whole hearts. Other Ca2+ mobilizer of high dynamic range. IpTxa spawned the discov-
cytosolic molecules such as protamines, fatty acids, cyclic adenos- ery of calcins, a small but growing group of scorpion peptides
ine diphosphate (cADP) ribose, and nicotinic acid adenine nucle- that selectively mobilize Ca2+ from RyR-gated stores. The defin-
otide phosphate are known to regulate some aspects of RyR2 ing characteristic of calcins is their capacity to stabilize RyR
function, but their role in E-C coupling in particular and cardiac openings in a long-lasting, subconducting state. This effect is
performance in general is less understood. Recent reviews cover nearly analogous to that of ryanodine, but unlike ryanodine,
most aspects of this regulation.31,101 calcins bind rapidly to RyRs (fast association rate), freely dissoci-
ate from their binding site (reversible effect), display a dose- and
sequence-variable effect, and are amenable for derivatization
Exogenous Ligands of RyR2 Channels and without undergoing major loss in receptor affinity.
Caffeine is a classical agonist of RyRs that is widely used to
Drugs That Affect Sarcoplasmic Reticulum assess the presence and size of RyR-gated Ca2+ stores. In cardiac
Ca2+ Release myocytes, caffeine readily permeates the external membrane and
elicits Ca2+ release by increasing the sensitivity of RyR2 channels
Many toxins and drugs bind to the RyR2 channel and modulate to cytosolic Ca2+.111 In addition, caffeine can increase the sensitiv-
its activity, but interestingly, there is not a single drug in the ity of the channels to luminal Ca2+,112 but this is not firmly estab-
market specifically designed for this purpose (although some are lished.113 Caffeine is easily available and when used at high doses
in clinical trials). Despite the rich assortment of chemicals that (5 to 20 mM) it yields a fair estimate of the SR Ca2+ content, but
affect the activity of RyR2 channels,102,103 only two classes of as with other ligands, it has several disadvantages. Like ryano-
toxins bind to RyRs with high (nanomolar) affinity: ryanodine dine, caffeine is also an alkaloid, but unlike ryanodine, it is a
and its derivatives and a group of peptide toxins termed calcins poor activator of RyRs. Its EC50 on [3H]ryanodine binding, with
(Table 6-1). This section discusses the most outstanding attri- “naked” RyRs (directly exposed) and under steady-state condi-
butes of these and other classical ligands of RyRs, as well as drugs tions, is approximately 300 µM.111 Caffeine is, however, an effec-
that hold promise to treat RyR2-associated disorders. tive inhibitor of cAMP phosphodiesterases and accelerates
Ryanodine is a small plant alkaloid (molecular weight, 494 Da) cellular metabolism by delaying cAMP turnover.114 This effect,
that has been an invaluable tool in defining the pharmacologic however, is negligible when caffeine is used for short periods, but
profile of RyRs.27 As described earlier, ryanodine binds to the prominent under prolonged exposures. Caffeine increases the
open state of the channel, which allows experimenters to use [3H] Ca2+ sensitivity of myofilament proteins, inhibits glycogen phos-
ryanodine as a probe of the functional state of the channel. This phorylase and adenosine receptors, stimulates the Na+/K+ pump,
approach has contributed to the isolation of the RyR1 and RyR2 and impairs phosphoinositide metabolism114; it also quenches the
channels themselves, and to the characterization of many phar- fluorescence of common Ca2+ indicators such as Fluo-3. As a
macologic properties of RyR channels. However, ryanodine result, caffeine displays pleiotropic effects of which the cardiac
exhibits some undesirable characteristics. The effects of ryano- electrophysiologist should be aware.
dine on single RyRs are complex and highly dependent on its Tetracaine is a local anesthetic that blocks voltage-dependent
concentration. At low concentrations (5 to 50 nM), ryanodine Na+ channels with high affinity, but at high doses it also acts as
increases the mean open time of the channel without modifying an allosteric inhibitor of RyRs.115 Because tetracaine stabilizes the
its unitary conductance.104 At intermediate concentrations (50 closed state of RyRs, it is used in cardiac myocytes as a versatile
nM to 10 µM), ryanodine locks the channel in an open subcon- indicator of spontaneous Ca2+ sparks and to assess passive SR Ca2+
ductance state that is recalcitrant to dissociation. Higher concen- leak.116 Ruthenium red also inhibits RyR2 channels117 by prolong-
trations (>10 µM) of the alkaloid fully and irreversibly close the ing the time the channel spends in the closed state.118 It inhibits
channel.105 Thus, ryanodine can act as an agonist and a blocker RyRs with an EC50 of approximately 1 µM, but it cannot perme-
of the RyR. Ryanodine also displays a slow association rate of ate membranes and binds to multiple targets. These attributes
Table 6-1. Ligands of RyR channels and Drugs that Affect SR Ca2+ release by modulating RyR2 channel activity. (Proof & electronic file on art CD)
Ligand
Ryanodine
Chemical Structure Mode of Action
Binds to a conformationally sensitive
Main Properties
High affinity and specificity for the RyR but
6
state of the channel and “locks” the slow association and dissociation
channel in a subconductance state28 rates104,106; dual mode of action105;
intrinsically minimized
Imperatoxin A Binds to open RyRs; stabilizes the 33-amino acid basic peptide with high
channel in a subconductance state; affinity (~5-10 nM) and specificity for the
unlike ryanodine, fast and reversible RyR108; fast association rate; cell-
effect penetrating peptide110; reversible effect;
amenable for derivatization without major
loss in RyR affinity108
Caffeine Increases the open probability of RyR May be used to assess the amount of Ca2+
channels by increasing their inside RyR-gated Ca2+ stores; fast
sensitivity to cytosolic Ca2+ (111) association rate but low affinity and
specificity; several off-target effects114
Tetracaine Decreases the open probability of Can be used to assess passive SR Ca2+
RyRs by stabilizing the channel in the leak116; displays low affinity for RyRs but
closed state115 high affinity for voltage-dependent Na2+
channels
Ruthenium red Decreases the open probability of Moderately high affinity for RyRs (~1 µM);
RyRs by producing “flickery” block of membrane impermeable; inhibits several
the channel118 ion channels and transporters or
cardiomyocytes with high affinity118
Doxorubicin Increases the open probability of Similar effect as caffeine, but with higher
RyRs by increasing their sensitivity to affinity for RyRs; anticancer drug with
cytosolic Ca2+.119 potent cardiotoxic effects119
K201 (JTV-519) Stabilizes the RyR2-FKBP12.6 Prevents Ca2+ overload-induced cell

interaction124 death122; in cardiac myocytes it blocks INa,
IK1, ICa, and Ikp SERCA2a and PKC103; effect
on RyR2 not clearly defined123-126
S107 Stabilizes the RyR2-FKBP12.6 Presumably of higher affinity and
interaction128-129 specificity for RyR2 channels compared
with K201129
Flecainide Decreases the open probability of Prevents tachyarrhythmias in CPVT

RyR2 channels by producing open patients and mice130; does not Increase SR
channel blockade131 Ca2+ load; blocks Na+ channels with high
affinity; mechanism of action to prevent
arrhythmias is controversial132
Carvedilol Decreases the open probability of Prevents tachyarrhythmias in CPVT mice;

RyR2 channels by decreasing their suppresses store overload-induced calcium
mean open time134 release134; combined action as β-blocker
and RyR2 channel blocker
RyR, Ryanodine receptor; SR, sarcoplasmic reticulum.

decrease its usefulness in cellular experiments. Doxorubicin and spontaneous Ca2+ waves was also shown in normal rat ventricular
other anthraquinones are potent stimulators of Ca2+ release that, myocytes, indicating that flecainide inhibits the activity of normal
like caffeine, sensitize the RyR channel to cytosolic Ca2+.119,120 and CPVT mutant RyR2 channels, thus extending its potential
Doxorubicin and caffeine likely bind to different sites in the RyR therapeutic range to all RyR2-generated arrhythmias. In single
channel, but both are presumed to modify the redox state of the RyR2 channel experiments, flecainide produced a fast open
RyR protein, thereby inducing similar effects. channel blockade,131 an effect that differed from that of other Na+
A prevalent hypothesis in RyR2-mediated arrhythmogenesis channel blockers that also suppress RyR2 activity, such as the
is an increased in diastolic SR Ca2+ leak,103,121 and several drugs local anesthetic tetracaine. Unlike tetracaine, flecainide does not
are being tested with the aim of stopping this process. K201, also seem to increase SR Ca2+ content despite its RyR2 blocking
known as JTV-519, is a diltiazem derivative first described by properties. These characteristics might partly explain the supe-
Kaneko122 as a drug that prevents Ca2+ overload-induced cardiac rior effect of flecainide on RyR2-triggered arrhythmias over
cell death. The drug has demonstrated pleiotropic effects, affect- other Na+ channel blockers. However, Priori’s group132 has pro-
ing several sarcolemmal ion currents including INa, IK1, ICa, and posed that the beneficial effects of flecainide in preventing
IKr and other targets such as PKC, SERCA2a, and annexin V.103 arrhythmias are due to the drug’s ability to block Na+ channels
Given that K201 was originally advanced as a drug to prevent and increase the threshold for triggered activity, rather than by
calcium overload in cardiac cells, Yano et al.123 used K201 in a its direct effects on RyR2 channels. Therefore, it seems clear that
canine model of heart failure with the rationale that RyR2- flecainide effectively blocks RyR2s in vitro, but it is debatable
mediated SR Ca2+ leak owing to FKBP12.6 dissociation (as whether targets other than RyR2 are required for the antiar-
advanced by Marks’ group)124 was critically involved in the patho- rhythmic effects of flecainide in the heart.
genesis of this syndrome. K201 preserved left ventricular systolic Carvedilol is a classical, nonselective β-adrenergic receptor
and diastolic function and prevented left ventricular remodeling, blocker (β-blocker) that was also found to prevent redox-
effects attributed to restoration of the RyR2-FKBP12.6 stoichi- dependent Ca2+ leak in failing cardiomyocytes by inhibiting
ometry toward control (nonfailing) levels. Later, Wehrens et al.124 the RyR2 channel.133 Chen’s group demonstrated that carvedilol
reported that K201 prevented ventricular tachycardia and effectively suppresses arrhythmogenic spontaneous Ca2+ release
sudden death in mice with reduced expression of FKBP12.6 in isolated ventricular myocytes.134 Because carvedilol also sup-
(FKBP12.6–/+), but had no effect on mice with complete ablation presses store overload–induced Ca2+ release in heterologous cells
of FKBP12.6 (FKBP12.6−/−), which is consistent with the notion expressing RyR2 channels and it blocks the open duration of the
that K201 prevents SR Ca2+ leak by stabilizing the RyR2- isolated RyR2 channel,134 the antiarrhythmic effect of this drug
FKBP12.6 interaction, but at odds with other studies that have is probably due to its combined β-blocking activity and RyR2
found that FKBP12.6 is not required for K201 effects on RyR2 channel inhibition. However, the carvedilol analog VK-II-86,
channels.125 Regardless of the exact mechanism, these studies which exhibits minimal β-blocking activity but retains its capacity
assign to correction of RyR2 dysfunction a critical role for K201 to inhibit RyR2 channels, prevented stress-induced ventricular
therapeutic effect, but this conclusion should be tempered by the tachyarrhythmias in a mouse model of CPVT (RyR2-R4496C+/−),
fact that the drug modulates multiple targets that affect SR func- although it did so more effectively when combined with either
tion. In addition, K201 was found ineffective in preventing of the selective β-blockers metoprolol or bisoprolol.134 Thus,
RyR2-mediated Ca2+ leak in another canine model of heart direct inhibition of RyR2 channels is probably an essential com-
failure126 and in a mouse model of human CPVT ponent of effective antiarrhythmic therapy in CPVT, but effec-
(RyR2-R4496C).127 tiveness may be greatly increased by inhibition of upstream
S107 is a novel derivative of K201128 that reportedly stabilizes effectors of Ca2+ release, such as the β-adrenergic receptor (carve-
the RyR2-FKBKP12.6 interaction at nanomolar concentrations dilol) or the voltage-dependent Na+ channel (flecainide).
but lacks the pleiotropic effects of K201 when used at concentra-
tions as high as 10 µM129; however, data documenting the selec-
tivity of S107 have not been published. S107 prevented
exercise-induced arrhythmias in another mouse model of CPVT RyR2 Channels in Disease
(RyR2-R2474S)129 and preserved ejection fraction (postmyocar-
dial infarction) in a mouse with decreased RyR2-FKBP12.6 inter- RyR channels are involved in several genetic diseases affecting
action owing to constitutive phosphorylation of RyR2-S2808 cardiac and skeletal muscle. This section discusses only CPVT
(RyR2-S2808D+/+).80 Thus, according to these results, S107 acts and heart failure as inherited and acquired, respectively, RyR2-
exclusively by stabilizing the RyR2-FKBP12.6 interaction. linked syndromes.
Because this interaction (and its counterpart reaction in skeletal
muscle, the RyR1-FKBP12 interaction) is reportedly faulty in
several diseases, including heart failure, ventricular arrhythmias, Catecholaminergic Polymorphic
atrial fibrillation, brain seizures, age-related muscle weakness and Ventricular Tachycardia
muscular dystrophy, the potential therapeutic value of this drug
is tremendous. However, a number of studies do not support the More than 160 different mutations (and increasing) in RYR2, the
link between RyR2 PKA-mediated hyperphosphorylation and gene encoding the RyR2 channel protein, have been associated
dissociation of FKBP12.6 from RyR2 and disagree on the relative with CPVT, an autosomal-dominant inherited cardiac syndrome
importance of the phosphorylation site (Ser2030 vs. Ser2808) at characterized by exercise- or stress-induced tachyarrhythmia epi-
which RyR2 is phosphorylated by PKA.75,77-79,85,88 sodes in the absence of apparent structural heart disease or
Flecainide is a class 1c antiarrhythmic drug classically used to prolonged QT interval.121,135 The clinical features of CPVT are
treat tachyarrhythmias (atrial fibrillation, supraventricular tachy- discussed elsewhere in this book. This section reviews the molec-
cardia) that has been found recently to prevent CPVT episodes ular and cellular basis of the cardiac arrhythmias resulting from
by blocking RyR2 channels.130 Experiments in mice and humans the RyR2-associated CPVT mutations. Excessive Ca2+ release
support the therapeutic value of flecainide in CPVT.130 In cardiac from the SR, especially during diastole, is accepted as the under-
myocytes isolated from a calsequestrin knockout mouse model of lying mechanism that gives rise to ventricular tachyarrhythmias
CPVT, flecainide decreased the frequency of isoproterenol- in CPVT.59,121 However, there is no consensus on whether the
induced diastolic Ca2+ waves caused by spontaneous openings excessive Ca2+ release occurs via a reduction in the threshold
of RyR2 channels.130,131 The ability of flecainide to suppress for activation of mutant RyR2 channels by luminal Ca2+,59,121
enhanced dissociation of FKBP12.6,124,129 defective RyR2 inter- some find alterations in RyR2 function.142 Although naturally a
6
domain interaction,123,125 or a combination thereof (see next system to increase efficiency and competence of the heart during
section). Whatever the mechanism, the enhanced diastolic Ca2+ acute periods of stress or altered hemodynamics, chronic
leak can overload the Na+/Ca2+ exchanger, which generates an β-adrenergic signaling may ultimately be damaging to the heart
inward current as it extrudes the released Ca2+. The inward by altering Ca2+ regulatory protein function, leading to impaired
current, in turn, gradually depolarizes the cell to threshold, favor- Ca2+ cycling. There is growing consensus that Ca2+ mishandling
ing delayed after depolarizations (DADs).121 During depolariza- leads to altered gene transcription, resulting in maladaptive struc-
tion, lack of Ca2+-dependent inactivation of ICa due to previous tural transformations (cardiac hypertrophy) that eventually
depletion of the SR leads to higher Ca2+ entry and reloading of hamper basic cardiac functioning (congestive HF). In fact, HF is
the SR, which triggers another DAD in the next beat. Successive an inexorable evolving stage in transgenic mice with constitutively
repetition of this altered Ca2+ cycle could probably result in par- activated PKA143 or CaMKII overexpression.144 Conceivably,
oxysmal tachycardia and arrhythmias even if only a few foci of hyperphosphorylation of Ca2+ cycling proteins in these animals
ventricular cells are involved. eventually results in abnormal Ca2+ homeostasis. Marks’ group
Although the hypothetical scheme logically relates RyR2 dys- has advanced the hypothesis that RyR2 channels are hyperphos-
function with ventricular tachycardia, it is unclear exactly what phorylated in HF, which leads to enhanced diastolic Ca2+
mechanism induces a group of apparently normal RyR2s to leak.41,42,50,86,124 In this scheme, PKA hyperphosphorylation of
behave aberrantly and to generate sudden tachycardia. Because RyR2-S2809 causes FKBP12.6 dissociation from the RyR2 and
infusion of catecholamines also triggers CPVT, it is likely that altered RyR2 gating, analogous to that reported for displacement
activation of the β-adrenergic system plays an important role. In of FKBP12.6 from the RyR2 by FK506 or rapamycin.41,42 In
this regard, it has been suggested that phosphorylation of RyR2 single-channel recordings, overall RyR2 Po was increased in HF,
by PKA, the kinase linking β1-adrenergic receptor activation to and most openings were to subconducting levels. The net effect
cellular effects, dissociates FKBP12.6,* an accessory protein that was greater ion flux, which in cellular terms would translate into
presumably stabilizes RyR2 in the closed state. Phosphorylation increased diastolic SR Ca2+ leak and be a primary cause of reduced
of a mutant RyR2 would therefore remove a stabilizer from a SR Ca2+ content. They attributed these alterations in HF to a
channel on the verge of dysfunction and would cause the patho- hyperadrenergic state and loss of RyR2-associated phosphatases
genic events described here. However, others have found no despite increased global myocyte phosphatase activity. More
evidence of FKBP12.6 dissociation in CPVT mutation-harboring recently, Wehrens’ group reported that preventing phosphoryla-
RyR2.59 Thus, structural alterations of the RyR2 channel complex tion of RyR2-S2814 (CaMKII site) improves cardiac performance
seem to be more important than dysregulation by accessory after pressure overload-induced HF, but has no protective effect
factors in the pathogenesis of CPVT. Consistent with this notion after myocardial ischemia-induced HF.145 Overall, then, this is an
is the fact that CPVT-associated mutations of RYR2 occur in interesting hypothesis, but many of its central tenets have not
domains corresponding exactly to mutation-containing domains been confirmed by others.103,142,146,147 Therefore, more research is
that give rise to malignant hyperthermia, a RyR1-linked syn- needed to determine whether RyR2 channels are central players
drome that affects skeletal muscle.137 Although some of these in the pathogenesis of HF or mere bystanders. Natural gain-of-
mutations are close to the apparent FKBP12.x-binding domain, function RyR2 mutations that produce “leaky” RyR2 channels can
the majority are not. Recent studies indicate a predominant role generate life-threatening arrhythmias and sudden death,121,135 but
of Purkinje cells in the genesis of ventricular arrhythmias.138,139 none have been reported to lead to overt HF.
Whatever the triggering mechanism, SR Ca2+ load and release
seem to be crucial, because mutations in CSQ2, the gene encod-
ing for cardiac calsequestrin, also generate CPVT.140 Accordingly,
a prominent hypothesis, advocated mainly by Chen and collabo- Monovalent Cation Channels
rators, proposes that cardiac myocytes have a threshold SR Ca2+
load for spontaneous Ca2+ release, and that CPVT-associated Based on early electron microprobe analysis in skeletal muscle, it
mutations decrease the sensitivity to luminal Ca2+ (i.e., decrease was postulated that SR Ca2+ release could not possibly generate
the threshold).121,134 Therefore, during β-adrenergic stimulation, large electrical potentials across the SR membrane because the
when SR Ca2+ load increases as a result of enhanced Ca2+ entry equilibrium potential for Ca2+ would be rapidly reached, limiting
and uptake, CPVT mutant channels reach their threshold and further Ca2+ release.148 This hypothesis was incompatible with the
generate spontaneous Ca2+-release waves, creating a substrate sustained Ca2+ release observed during a tetanic stimulation.
favorable for Ca2+-dependent arrhythmias. Alternatively, CPVT- Therefore, a countercharge movement was necessary to maintain
related mutations could increase the SR load, making it easier to electroneutrality across the SR membrane, and extensive move-
reach the threshold upon adrenergic stimulation.141 ment across the SR of K+ (but not Na+ or Cl–) during SR Ca2+
release were observed.30 Indeed, several types of K+ and Cl– chan-
nels from SR vesicles have been functionally characterized after
Role of RyR2 Channels in Heart Failure reconstitution in lipid bilayers, but their most outstanding struc-
tural features (molecular organization, amino acid sequence, and
A broad array of primary insults to the cardiovascular system can structural domains) remain a mystery. A trimeric intracellular
ultimately lead to the syndrome of heart failure (HF). The patho- cation channel (TRIC) of cardiac and skeletal SR has been char-
genesis of contractile dysfunction at the cardiomyocyte level acterized in detail at the molecular level. This channel is the most
remains unclear, although alterations in E-C coupling seem to be viable candidate to mediate at least a portion of the K+ conduc-
a consistent and important feature of all animal models of HF and tance in the SR membrane and will be discussed separately. An
in the limited number of studies examining failing human myo- alternative hypothesis on the identity of the ion channels respon-
cardium.42,102,103,142 Most studies reveal a blunted Ca2+ transient in sible for countercurrent movement posits that RyR channels
HF cells that explains, at least partly, the characteristic contractile themselves conduct most of this countercurrent,149 but this is
dysfunction of HF. Most studies also find alterations in SR Ca2+ mostly based on theoretical models and needs further experimen-
transport, with decreased SERCA2a expression and function, and tal testing.
The SR membrane apparently contains several K+ channels,
but prominent among them is a Cs+-blocked large-conductance
*References 41, 42, 81, 86, 124, 129. K+ channel that is also blocked by decamethonium, gallamine,
neomycin and, more physiologically relevant, Ca2+ in the milli- past their adolescent age, but homozygous ablation of TRIC-B
molar range.150-152 Thus, SR Ca2+ release could possibly relieve is lethal as the TRIC-B-/- mice die at neonatal stage. Aggravated
Ca2+ block of these channels, thereby promoting K+ entry across embryonic lethality is observed with the TRIC-A-/-TRIC-B-/-
the SR membrane and facilitating the countercharge movement mice, suggesting an important role of TRIC channels
necessary for sustained Ca2+ release. One study reported that a in development.156 Cardiomyocytes isolated from TRIC-A–/–
canine cardiac SR K+ channel is blocked by 4-aminopyridine TRIC-B–/– embryos have swollen SR structures and abnormally
(4-AP),153 but another study could not detect 4-AP sensitivity in small Ca2+ oscillations. In addition, skeletal muscle fibers of
an SR K+ channel isolated from human and sheep cardiac TRIC-A knockout mice display abnormal Ca2+ sparks and
atrium.154 It is possible that these are in fact two different types impaired CICR.156 These findings suggest that TRIC-A has an
of channels, but this is an understudied area and more work is important role in providing the countercharge movement that
needed to answer this and many other questions. sustains SR Ca2+ release in cardiac and skeletal muscle.
Trimeric Intracellular Cation Channel Chloride Channels

Using an ingenious approach based on a monoclonal antibody Most of the functional characterization of SR Cl– channels has
library against triad junction proteins, Takeshima et al.155 identi- been conducted using skeletal muscle, perhaps because of its
fied several proteins (collectively termed mitsugumins, Japanese more robust SR network compared with cardiac muscle. Based
for “triad junction”) that reside inside the SR or are closely asso- on their most elemental biophysical properties, two types of Cl–
ciated with the RyR macromolecular complex.155 The role of a channels are prominent in skeletal muscle SR: a large-conductance
33-kD protein (previously termed mitsugumin 33) has started to (250 pS) channel and a small-conductance (70 pS) channel.157 It
emerge. This novel protein forms a TRIC channel that resides is not clear whether exact functional counterparts of these chan-
in the SR and endoplasmic reticulum membranes.156 Two iso- nels occur in cardiac SR, but one study found a cardiac Cl–
forms of this protein have been described. TRIC-A is predomi- channel that conducts larger anions, such as phosphate and even
nantly expressed in the SR of muscle cells and is particularly adenine nucleotides.158 It is important to consider that these
relevant to this discussion, and TRIC-B, which is expressed in studies were conducted using purified SR vesicles reconstituted
the endoplasmic reticulum of many tissues. Single-channel in lipid bilayers. Although markers of SR such as RyRs and
recordings of TRIC-A reconstituted in lipid bilayers show that SERCA are indeed greatly increased in these preparations, it is
this channel has a K+ conductance of approximately 110 pS (in difficult to eliminate contamination with other organelles such as
200 mM K+) but discriminates poorly other physiologically rel- mitochondria or nuclear membranes. Therefore, the exact origin
evant cations (K+/Na+ permeability ratio, 1.5). TRIC-A is not of these channels needs to be verified by independent techniques.
blocked by Ca2+ or decamethonium; therefore, this channel seems In addition, the physiologic role of Cl– channels remains unclear
pharmacologically different from the SR K+ channel described in because the Cl– concentration inside the SR does not change
the preceding paragraph. Mutant mice lacking TRIC-A survive appreciably during Ca2+ release, at least in skeletal muscle.30
11. Yin CC, Lai FA: Intrinsic lattice formation by the 22. Kettlun C, González A, Ríos E, et al: Unitary
References ryanodine receptor calcium-release channel. Nat Ca2+ current through mammalian cardiac and
Cell Biol 2(9):669–671, 2000. amphibian skeletal muscle ryanodine receptor
1. Bers DM: Excitation-contraction coupling and 12. Stern MD, Pizarro G, Ríos E: Local control Channels under near-physiological ionic condi-
cardiac contractile force, ed 2, Dordrecht, model of excitation-contraction coupling in skel- tions. J Gen Physiol 122(4):407–417, 2003.
Germany, 2001, Kluwer Academic. etal muscle. J Gen Physiol 110(4):415–440, 1997. 23. Wang SQ, Song LS, Lakatta EG, et al: Ca2+
2. Wu X, Bers DM: Sarcoplasmic reticulum and 13. Altamirano J, Bers DM: Voltage dependence of signalling between single L-type Ca2+ channels
nuclear envelope are one highly interconnected cardiac excitation-contraction coupling: unitary and ryanodine receptors in heart cells. Nature
Ca2+ store throughout cardiac myocyte. Circ Res Ca2+ current amplitude and open channel prob- 410(6828):592–596, 2001.
99(3):283–291, 2006. ability. Circ Res 101(6):590–597, 2007. 24. Lipp P, Niggli E: Submicroscopic calcium signals
3. Fabiato A: Calcium-induced release of calcium 14. Sommer JR, Johnson EA: Ultraestructure of as fundamental events of excitation–contraction
from the cardiac sarcoplasmic reticulum. Am J cardiac muscle. In Berne RM, editor: Handbook coupling in guinea-pig cardiac myocytes. J Physiol
Physiol 245(1):C1–14, 1983. of Physiology. Section 2. The Cardiovascular 492(Pt 1):31–38, 1996.
4. Franzini-Armstrong C, Protasi F, Ramesh V: System. Bethesda, MD, Am Physiol Soc. Vol. 1. 25. González A, Kirsch WG, Shirokova N, et al: The
Comparative ultrastructure of Ca2+ release units 1979. spark and its ember: separately gated local com-
in skeletal and cardiac muscle. Ann N Y Acad Sci 15. Coronado R, Kawano S, Lee CJ, et al: Planar ponents of Ca(2+) release in skeletal muscle.
853:20–30, 1998. bilayer recording of ryanodine receptors of sarco- J Gen Physiol 115(2):139–158, 2000.
5. Ríos E, Ma JJ, González A: The mechanical plasmic reticulum. Methods Enzymol 207:699– 26. ZhuGe R, DeCrescenzo V, Sorrentino V, et al:
hypothesis of excitation-contraction (EC) cou- 707, 1992. Syntillas release Ca2+ at a site different from the
pling in skeletal muscle. J Muscle Res Cell Motil 16. Viero C, Thomas NL, Euden J, et al: Techniques microdomain where exocytosis occurs in mouse
12(2):127–135, 1991. and methodologies to study the ryanodine recep- chromaffin cells. Biophys J 90(6):2027–2037,
6. Hamilton SL, Serysheva II: Ryanodine receptor tor at the molecular, subcellular and cellular level. 2006.
structure: progress and challenges. J Biol Chem Adv Exp Med Biol 740:183–215, 2012. 27. Sutko JL, Airey JA, Welch W, et al: The pharma-
284(7):4047–4051, 2009. 17. Fill M, Copello JA: Ryanodine receptor calcium cology of ryanodine and related compounds.
7. Bers DM: Cardiac excitation-contraction cou- release channels. Physiol Rev 82(4):893–922, Pharmacol Rev 49(1):53–98, 1997.
pling. Nature 415(6868):198–205, 2002. 2002. 28. Zimányi I, Pessah IN: Comparison of [3H]ryano-
8. Fabiato A: Time and calcium dependence of acti- 18. Cheng H, Lederer WJ: Calcium sparks. Physiol dine receptors and Ca++ release from rat cardiac
vation and inactivation of calcium-induced release Rev 88(4):1491–1545, 2008. and rabbit skeletal muscle sarcoplasmic reticulum.
of calcium from the sarcoplasmic reticulum of a 19. Cheng H, Lederer WJ, Cannell MB: Calcium J Pharmacol Exp Ther 256(3):938–946, 1991.
skinned canine cardiac Purkinje cell. J Gen sparks: elementary events underlying excitation- 29. Capes EM, Loaiza R, Valdivia HH: Ryanodine
Physiol 85:247–289, 1985. contraction coupling in heart muscle. Science receptors. Skelet Muscle 1(1):18, 2011.
9. Endo M, Tanaka M, Ogawa Y: Calcium induced 262(5134):740–744, 1993. 30. Somlyo AV, Gonzalez-Serratos HG, Shuman H,
release of calcium from the sarcoplasmic reticu- 20. Nelson MT, Cheng H, Rubart M, et al: Relax- et al: Calcium release and ionic changes in the
lum of skinned skeletal muscle fibres. Nature ation of arterial smooth muscle by calcium sarcoplasmic reticulum of tetanized muscle: An
228:34–36, 1970. sparks. Science 270(5236):633–637, 1995. electron-probe study. J Cell Biol 90:577–594,
10. Ríos E, Brum G: Ca2+ release flux underlying 21. Klein MG, Cheng H, Santana LF, et al: Two 1981.
Ca2+ transients and Ca2+ sparks in skeletal mechanisms of quantized calcium release in skel- 31. Lanner JT, Georgiou DK, Joshi AD, et al: Ryano-
muscle. Front Biosci 7:d1195–211, 2002. etal muscle. Nature 379(6564):455–458, 1996. dine receptors: structure, expression, molecular
53. Gyorke S, Fill M: Ryanodine receptor adaptation: 74. Trafford AW, Díaz ME, Sibbring GC, et al: Mod-
details, and function in calcium release. Cold control mechanism of Ca(2+)-induced Ca2+ ulation of CICR has no maintained effect on sys-
32.
Spring Harb Perspect Biol 2(11):a003996,
2010.
Smith JS, Coronado R, Meissner G: Single
release in heart. Science 260:807–809, 1993.
54. Valdivia HH, Kaplan JH, Ellis-Davies GC, et al:
Rapid adaptation of cardiac ryanodine receptors:
tolic Ca2+: simultaneous measurements of
sarcoplasmic reticulum and sarcolemmal Ca2+
fluxes in rat ventricular myocytes. J Physiol. 522
6
channel measurements of the calcium release modulation by Mg2+ and phosphorylation. Pt 2:259–270, 2000.
channel from skeletal muscle sarcoplasmic reticu- Science 267:1997–2000, 1995. 75. Eisner DA, Kashimura T, Venetucci LA, et al:
lum. Activation by Ca2+ and ATP and modulation 55. Velez P, Gyorke S, Escobar AL, et al: Adaptation From the ryanodine receptor to cardiac arrhyth-
by Mg2+. J Gen Physiol 88(5):573–588, 1986. of single cardiac ryanodine receptor channels. mias. Circ J 73(9):1561–1567, 2009.
33. Wagenknecht T, Samsó M: Three-dimensional Biophys J 72:691–697, 1997. 76. Takasago T, Imagawa T, Furukawa K, et al: Regu-
reconstruction of ryanodine receptors. Front 56. Shannon TR, Ginsburg KS, Bers DM: Quantita- lation of the cardiac ryanodine receptor by
Biosci 7:d1464–1474, 2002. tive assessment of the SR Ca2+ leak-load relation- protein kinase-dependent phosphorylation.
34. Samsó M, Wagenknecht T, Allen PD: Internal ship. Circ Res 91(7):594–600, 2002. J Biochem 109:163–170, 1991.
structure and visualization of transmembrane 57. Gyorke S, Terentyev D: Modulation of ryanodine 77. Benkusky NA, Weber CS, Scherman JA, et al:
domains of the RyR1 calcium release channel receptor by luminal calcium and accessory pro- Intact β-adrenergic response and unaltered pro-
by cryo-EM. Nat Struct Mol Biol 12:539–544, teins in health and cardiac disease. Cardiovasc Res gression towards heart failure in mice with genetic
2005. 77:245–255, 2008. ablation of a major PKA site in the cardiac ryano-
35. Welch W, Rheault S, West DJ, et al: A model of 58. Tang Y, Tian X, Wang R, et al: Abnormal termi- dine receptor. Circ Res 101:819–829, 2007.
the putative pore region of the cardiac ryanodine nation of Ca2+ release is a common defect of 78. MacDonnell S, García-Rivas G, Kubo H, et al:
receptor channel. Biophys J 87:2335–2351, 2004. RyR2 mutations associated with cardiomyopa- Adrenergic regulation of cardiac contractility
36. Tung CC, Lobo PA, Kimlicka L, et al: The thies. Circ Res 110(7):968–977, 2012. does not involve phosphorylation of the cardiac
amino-terminal disease hotspot of ryanodine 59. MacLennan DH, Chen SR: Store overload- ryanodine receptor at serine 2808. Circ Res
receptors forms a cytoplasmic vestibule. Nature induced Ca2+ release as a triggering mechanism 102(8):e65–72, 2008.
468(7323):585–588, 2010. for CPVT and MH episodes caused by mutations 79. Li Y, Kranias EG, Mignery GA, et al: Protein
37. Van Petegem F: Ryanodine Receptors: Structure in RYR and CASQ genes. J Physiol 587 kinase A phosphorylation of the ryanodine recep-
and Function. J Biol Chem. 2012. July 20. [Epub (Pt 13):3113–3115, 2009. tor does not affect calcium sparks in mouse ven-
ahead of print] 60. Laver DR, O’Neill ER, Lamb GD: Luminal tricular myocytes. Circ Res 90:309–316, 2002.
38. Serysheva II: Structural insights into excitation- Ca2+-regulated Mg2+ inhibition of skeletal RyRs 80. Shan J, Betzenhauser MJ, Kushnir A, et al: Role
contraction coupling by electron cryomicroscopy. reconstituted as isolated channels or coupled clus- of chronic ryanodine receptor phosphorylation in
Biochemistry (Mosc) 69:1226–1232, 2004. ters. J Gen Physiol 124:741–758, 2004. heart failure and β-adrenergic receptor blockade
39. Yamaguchi N, Takahashi N, Xu L, et al: Early 61. Yuan Q, Fan GC, Dong M, et al: Sarcoplasmic in mice. J Clin Invest. 120(12):4375–4387, 2010.
cardiac hypertrophy in mice with impaired reticulum calcium overloading in junctin defi- 81. Lehnart SE, Terrenoire C, Reiken S, et al: Stabi-
calmodulin regulation of cardiac muscle Ca ciency enhances cardiac contractility but increases lization of cardiac ryanodine receptor prevents
release channel. J Clin Invest 117(5):1344–1353, ventricular automaticity. Circulation 115:300– intracellular calcium leak and arrhythmias.
2007. 309, 2007. Proc Natl Acad Sci U S A 103(20):7906–7910,
40. Gangopadhyay JP, Ikemoto N: Aberrant interac- 62. Qin J, Valle G, Nani A, et al: Ryanodine receptor 2006.
tion of calmodulin with the ryanodine receptor luminal Ca2+ regulation: swapping calsequestrin 82. Witcher DR, Kovacs RJ, Schulman H, et al:
develops hypertrophy in the neonatal cardiomyo- and channel isoforms. Biophys J 97:1961–1970, Unique phosphorylation site on the cardiac
cyte. Biochem J 438(2):379–387, 2011. 2009. ryanodine receptor regulates calcium channel
41. Marx SO, Reiken S, Hisamatsu Y, et al: PKA 63. Stern MD: Theory of excitation-contraction cou- activity. J Biol Chem 266:11144–11152, 1991.
phosphorylation dissociates FKBP12.6 from the pling in cardiac muscle. Biophys J 63(2):497–517, 83. Hain J, Onoue H, Mayrleitner M, et al: Phos-
calcium release channel (ryanodine receptor): 1992. phorylation modulates the function of the calcium
defective regulation in failing hearts. Cell 64. Nabauer M, Morad M: Ca2+-induced Ca2+ release channel of sarcoplasmic reticulum from
101(4):365–376, 2000. release as examined by photolysis of caged Ca2+ cardiac muscle. J Biol Chem 270:2074–2081,
42. Marks AR: Ryanodine receptors, FKBP12, and in single ventricular myocytes. Am J Physiol Cell 1995.
heart failure. Front Biosci 7:d970–7, 2002. Physiol 285:C189–C193, 1990. 84. Terentyev D, Viatchenko-Karpinski S, Gyorke I,
43. Xiao J, Tian X, Jones PP, et al: Removal of 65. Yasui K, Palade P, Gyorke S: Negative control et al: Protein phosphatases decrease sarcoplasmic
FKBP12.6 does not alter the conductance and mechanism with features of adaptation controls reticulum calcium content by stimulating calcium
activation of the cardiac ryanodine receptor or Ca2+ release in cardiac myocytes. Biophys J 67: release in cardiac myocytes. J Physiol. 552:109–
the susceptibility to stress-induced ventricular 457–460, 1994. 118, 2003.
arrhythmias. J Biol Chem 282(48):34828–34838, 66. Soeller C, Cannell MB: Analysing cardiac 85. Stange M, Xu L, Balshaw D, et al: Characteriza-
2007. excitation-contraction coupling with mathemati- tion of recombinant skeletal muscle (Ser-2843)
44. Meyers MB, Fischer A, Sun YJ, et al: Sorcin regu- cal models of local control. Prog Biophys Mol and cardiac muscle (Ser-2809) ryanodine receptor
lates excitation-contraction coupling in the heart. Biol 85(2-3):141–162, 2004. phosphorylation mutants. J Biol Chem
J Biol Chem 278(31):28865–28871, 2003. 67. Varro A, Negretti N, Hester SB, et al: An estimate 278:51693–51702, 2003.
45. Farrell EF, Antaramian A, Rueda A, et al: Sorcin of the calcium content of the sarcoplasmic reticu- 86. Wehrens XH, Lehnart SE, Reiken S, et al: Ryano-
inhibits calcium release and modulates excitation- lum in rat ventricular myocytes. Pflugers Arch dine receptor/calcium release channel PKA phos-
contraction coupling in the heart. J Biol Chem 423:158–160, 1993. phorylation: a critical mediator of heart failure
278:34660–34666, 2003. 68. Shannon TR, Guo T, Bers DM: Ca2+ scraps: progression. Proc Natl Acad Sci USA 103:511–
46. Rueda A, Song M, Stefani E, et al: Sorcin modula- Local depletions of free [Ca2+] in cardiac sarco- 518, 2006.
tion of Ca2+ sparks in rat vascular smooth muscle plasmic reticulum during contractions leave 87. Wehrens XH, Lehnart SE, Reiken SR, et al:
cells. J Physiol (Lond) 576:887–901, 2006. substantial Ca2+ reserve. Circ Res 93:40–45, Ca2+/calmodulin-dependent protein kinase II
47. Gyorke I, Hester N, Jones LR, et al: The role of 2003. phosphorylation regulates the cardiac ryanodine
calsequestrin, triadin, and junctin in conferring 69. Ikemoto N, Ronjat M, Meszaros LG, et al: Pos- receptor. Circ Res 94(6):e61–70, 2004.
cardiac ryanodine receptor responsiveness to tulated role of calsequestrin in the regulation of 88. Xiao B, Zhong G, Obayashi M, et al: Ser-2030,
luminal calcium. Biophys J 86(4):2121–2128, calcium release from sarcoplasmic reticulum. Bio- but not Ser-2808, is the major phosphorylation
2004. chemistry 28:6764–6771, 1989. site in cardiac ryanodine receptors responding to
48. Knollmann BC: New roles of calsequestrin 70. Terentyev D, Viatchenko-Karpinski S, Gyorke I, protein kinase A activation upon beta-adrenergic
and triadin in cardiac muscle. J Physiol. 587 et al: Calsequestrin determines the functional size stimulation in normal and failing hearts. Biochem
(Pt 13):3081–3087, 2009. and stability of cardiac intracellular calcium J 396:7–16, 2006.
49. Prosser BL, Hernández-Ochoa EO, Schneider stores: Mechanism for hereditary arrhythmia. 89. Ginsburg KS, Bers DM: Modulation of excitation-
MF: S100A1 and calmodulin regulation of ryano- Proc Natl Acad Sci U S A 100:11759–11764, contraction coupling by isoproterenol in cardio-
dine receptor in striated muscle. Cell Calcium 2003. myocytes with controlled SR Ca2+ load and Ca2+
50(4):323–331, 2011. 71. Kubalova Z, Terentyev D, Viatchenko-Karpinski current trigger. J Physiol 556:463–480, 2004.
50. Lehnart SE, Wehrens XH, Reiken S, et al: Phos- S, et al: Abnormal intrastore calcium signaling in 90. Xu L, Eu JP, Meissner G, et al: Activation of the
phodiesterase 4D deficiency in the ryanodine- chronic heart failure. Proc Natl Acad Sci U S A cardiac calcium release channel (ryanodine recep-
receptor complex promotes heart failure and 102:14104–14109, 2005. tor) by poly-S-nitrosylation. Science 279:234–
arrhythmias. Cell 123:25–35, 2005. 72. Guo T, Ai X, Shannon TR, et al: Intrasarcoplas- 237, 1998.
51. Meissner G, Rios E, Tripathy A, et al: Regulation mic reticulum free [Ca2+] and buffering in 91. Aracena P, Sanchez G, Donoso P, et al:
of skeletal muscle Ca2+ release channel (ryano- arrhythmogenic failing rabbit heart. Circ Res S-glutathionylation decreases Mg2+ inhibition
dine receptor) by Ca2+ and monovalent cations 101:802–810, 2007. and S-nitrosylation enhances Ca2+ activation of
and anions. J Biol Chem 272(3):1628–1638, 1997. 73. Marx SO, Gaburjakova J, Gaburjakova M, et al: RyR1 channels. J Biol Chem 278:42927–42935,
52. Fabiato A: Calcium-induced release of calcium Coupled gating between cardiac calcium release 2003.
from the cardiac sarcoplasmic reticulum. Am J channels (ryanodine receptors). Circ Res 92. Sun J, Yamaguchi N, Xu L, et al: Regulation of
Physiol 245(1):C1–14, 1983. 88:1151–1158, 2001. the cardiac muscle ryanodine receptor by O(2)
tension and S-nitrosoglutathione. Biochemistry luminal Ca2+ activation of the ryanodine recep- catecholaminergic polymorphic ventricular
47:13985–13990, 2008. tor. Biochem J 414(3):441–452, 2008. tachycardia by increasing the threshold for trig-
93. Voss AA, Lango J, Ernst-Russell M, et al: Identi- 113. Porta M, Zima AV, Nani A, et al: Single ryanodine gered activity. Circ Res 109(3):291–295, 2011.
fication of hyperreactive cysteines within ryano- receptor channel basis of caffeine’s action on Ca2+ 133. Mochizuki M, Yano M, Oda T, et al: Scavenging
dine receptor type 1 by mass spectrometry. J Biol sparks. Biophys J 100(4):931–938, 2011. free radicals by low-dose carvedilol prevents
Chem 279(33):34514–34520, 2004. 114. Magkos F, Kavouras SA: Caffeine use in sports, redox-dependent Ca2+ leak via stabilization of
94. Aracena-Parks P, Goonasekera SA, Gilman CP, pharmacokinetics in man, and cellular mecha- ryanodine receptor in heart failure. J Am Coll
et al: Identification of cysteines involved in nisms of action. Crit Rev Food Sci Nutr 45 Cardiol 49(16):1722–1732, 2007.
s-nitrosylation, sglutathionylation, and oxidation (7-8):535–562, 2005. 134. Zhou Q, Xiao J, Jiang D, et al: Carvedilol and its
to disulfides in ryanodine receptor type 1. J Biol 115. Overend CL, O’Neill SC, Eisner DA: The effect new analogs suppress arrhythmogenic store
Chem 281:40354–40368, 2006. of tetracaine on stimulated contractions, sarco- overload-induced Ca2+ release. Nat Med
95. Eu JP, Sun J, Xu L, et al: The skeletal muscle plasmic reticulum Ca2+ content and membrane 17(8):1003–1009, 2011.
calcium release channel: coupled O2 sensor and current in isolated rat ventricular myocytes. 135. Laitinen PJ, Swan H, Piippo K, et al: Genes, exer-
NO signaling functions. Cell 102:499–509, 2000. J Physiol 507:759–769, 1998. cise and sudden death: molecular basis of familial
96. Gonzalez DR, Fernandez IC, Ordenes PP, et al: 116. Curran J, Hinton MJ, Ríos E, et al: Beta- catecholaminergic polymorphic ventricular
Differential role of S-nitrosylation and the NO- adrenergic enhancement of sarcoplasmic reticu- tachycardia. Ann Med 36(Suppl 1):81–86,
cGMP-PKG pathway in cardiac contractility. lum calcium leak in cardiac myocytes is mediated 2004.
Nitric Oxide 18:157–167, 2008. by calcium/calmodulin-dependent protein kinase. 136. Song DW, Lee JG, Youn HS, et al: Ryanodine
97. Ziolo MT: The fork in the nitric oxide road: cyclic Circ Res 100(3):391–398, 2007. receptor assembly: a novel systems biology
GMP or nitrosylation? Nitric Oxide 18:153–156, 117. Lukyanenko V, Gyorke I, Subramanian S, et al: approach to 3D mapping. Prog Biophys Mol Biol.
2008. Inhibition of Ca2+ sparks by ruthenium red in 105(3):145–161, 2011.
98. Wang H, Viatchenko-Karpinski S, Sun J, et al: permeabilized rat ventricular myocytes. Biophys J 137. Yano M, Yamamoto T, Ikeda Y, et al: Mechanisms
Regulation of myocyte contraction via neuronal 79:1273–1284, 2000. of Disease: ryanodine receptor defects in heart
nitric oxide synthase: role of ryanodine receptor 118. Ma J: Block by ruthenium red of the ryanodine- failure and fatal arrhythmia. Nat Clin Pract Car-
S-nitrosylation. J Physiol 588:2905–2917, 2010. activated calcium release channel of skeletal diovasc Med 3:43–52, 2006.
99. Laver DR, Honen BN: Luminal Mg2+, a key muscle. J Gen Physiol 102(6):1031–1056, 1993. 138. Cerrone M, Noujaim SF, Tolkacheva EG, et al:
factor controlling RYR2-mediated Ca2+ release: 119. Abramson JJ, Buck E, Salama G, et al: Mechanism Arrhythmogenic mechanisms in a mouse model
cytoplasmic and luminal regulation modeled in a of anthraquinone-induced calcium release from of catecholaminergic polymorphic ventricular
tetrameric channel. J Gen Physiol 132(4):429– skeletal muscle sarcoplasmic reticulum. J Biol tachycardia. Circ Res 101:1039–1048, 2007.
446, 2008. Chem 263(35):18750–18758, 1988. 139. Herron TJ, Milstein ML, Anumonwo J, et al:
100. Xu L, Mann G, Meissner G: Regulation of cardiac 120. Zuppinger C, Timolati F, Suter TM: Pathophysi- Purkinje cell calcium dysregulation is the cellular
Ca2+ release channel (ryanodine receptor) by ology and diagnosis of cancer drug induced car- mechanism that underlies catecholaminergic
Ca2+, H+, Mg2+, and adenine nucleotides under diomyopathy. Cardiovasc Toxicol 7:61–66, 2007. polymorphic ventricular tachycardia. Heart
normal and simulated ischemic conditions. Circ 121. Priori SG, Chen SR: Inherited dysfunction of Rhythm 7:1122–1128, 2010.
Res 79:1100–1109, 1996. sarcoplasmic reticulum Ca2+ handling and 140. Faggioni M, Knollmann BC: Calsequestrin 2 and
101. Meissner G: Molecular regulation of cardiac arrhythmogenesis. Circ Res 108(7):871–883, arrhythmias. Am J Physiol Heart Circ Physiol
ryanodine receptor ion channel. Cell Calcium 2011. 302(6):H1250–H1260, 2012.
35:621–628, 2004. 122. Kaneko N: New 1,4-benzothiazepine derivative, 141. Kashimura T, Briston SJ, Trafford AW, et al: In
102. Mackrill JJ: Ryanodine receptor calcium channels K201, demonstrates cardioprotective effects the RyR2(R4496C) mouse model of CPVT, beta
and their partners as drug targets. Biochem Phar- against sudden cardiac cell death and intracellular adrenergic stimulation induces Ca waves by
macol 79(11):1535–1543, 2010. calcium blocking action. Drug Dev Res 33:429– increasing SR Ca content and not by decreasing
103. Currie S, Elliott EB, Smith GL, et al: Two can- 438, 1994. the threshold for Ca waves. Circ Res 107:1483–
didates at the heart of dysfunction: The ryanodine 123. Yano M, Kobayashi S, Kohno, et al: FKBP12.6- 1489, 2010.
receptor and calcium/calmodulin protein kinase II mediated stabilization of calcium release channel 142. George CH: Sarcoplasmic reticulum Ca2+ leak
as potential targets for therapeutic intervention-An (ryanodine receptor) as a novel therapeutic strat- in heart failure: mere observation or functional
in vivo perspective. Pharmacol Ther 131(2):204– egy against heart failure. Circulation 107:464– relevance? Cardiovasc Res 77(2):302–314,
220, 2011. 477, 2003. 2008.
104. Buck E, Zimany I, Abramson JA, et al: Ryanodine 124. Wehrens XH, Lehnart SE, Reiken, et al: Protec- 143. Antos CL, Frey N, Marx SO, et al: Dilated car-
stabilizes multiple conductance states of the skel- tion from cardiac arrhythmia through ryanodine diomyopathy and sudden death resulting from
etal muscle calcium release channel. J Biol Chem receptor stabilizing protein calstabin-2. Science constitutive activation of protein kinase a. Circ
267, 23560–23567, 1992 304, 292–296, 2004. Res 89(11):997–1004, 2001.
105. Lindsay ARG, Tinker A, Williams AJ: How does 125. Yamamoto T, Yano M, Xu X, et al: Identification 144. Zhang T, Maier LS, Dalton ND, et al: The
ryanodine modify ion handling in the sheep of target domains of the cardiac ryanodine recep- deltaC isoform of CaMKII is activated in cardiac
cardiac sarcoplasmic reticulum Ca2+-release tor to correct channel disorder in failing hearts. hypertrophy and induces dilated cardiomyopathy
channel? J Gen Physiol 104:425–447, 1994. Circulation 117:762–772, 2008. and heart failure. Circ Res 92(8):912–919,
106. Pessah I, Stambuk RA, Casida JE: Ca2+ activated 126. Tateishi H, Yano M, Mochizuki M, et al: Defec- 2003.
ryanodine binding: mechanism of activation and tive domain–domain interactions within the 145. Respress JL, van Oort RJ, Li N, et al: Role of
intensity modulation by Mg2+, caffeine and ryanodine receptor as a critical cause of diastolic RyR2 phosphorylation at S2814 during heart
adenine nucleotides. Mol Pharmacol 31:232–238, Ca2+ leak in failing hearts. Cardiovasc Res failure progression. Circ Res 110(11):1474–1483,
1987. 81:536–545, 2009. 2012.
107. Valdivia HH, Kirby MS, Lederer WJ, et al: Scor- 127. Liu N, Colombi B, Memmi S, et al: Arrhythmo- 146. Bers DM: Ryanodine receptor S2808 phosphory-
pion toxins targeted against the sarcoplasmic genesis in catecholaminergic polymorphic ven- lation in heart failure: smoking gun or red herring.
reticulum Ca2+ release channel of skeletal and tricular tachycardia: insights from a RyR2 Circ Res 110(6):796–799, 2012.
cardiac muscle. Proc Natl Acad Sci U S A knock-in mouse model. Circ Res 99:292–298, 147. Valdivia HH: Ryanodine receptor phosphoryla-
89:12185–12189, 1992. 2006. tion and heart failure: phasing out S2808 and
108. Zamudio FZ, Gurrola GB, Arévalo C, et al: 128. Bellinger AM, Reiken S, Dura M, et al: Remodel- “criminalizing” S2814. Circ Res 110(11):1398–
Primary structure and synthesis of Imperatoxin A ling of ryanodine receptor complex causes ‘leaky’ 1402, 2012.
(IpTxa), a peptide activator of Ca2+ release channels: a molecular mechanism for decreased 148. Oetliker H: An appraisal of the evidence for a
channels/ryanodine receptors. FEBS Letters exercise capacity. Proc Natl Acad Sci U S A sarcoplasmic reticulum membrane potential and
405:385–389, 1997. 105:2198–2202, 2008. its relation to calcium release in skeletal muscle.
109. Esteve E, Mabrouk K, Dupuis A, et al: Transduc- 129. Lehnart SE, Mongillo M, Bellinger A, et al: Leaky J Muscle Res Cell Motil 3:247–272, 1982.
tion of the scorpion toxin maurocalcine into cells. Ca release channel/ryanodine receptor 2 causes 149. Gillespie D, Fill M: Intracellular calcium release
Evidence that the toxin crosses the plasma mem- seizures and sudden cardiac death in mice. J Clin channels mediate their own countercurrent: the
brane. J Biol Chem 280(13):12833–12839, 2005. Invest 118:2230–2245, 2008. ryanodine receptor case study. Biophys J
110. Gurrola GB, Capes EM, Zamudio FZ, 130. Watanabe H, Chopra N, Laver D, et al: Fle- 2008;95(8):3706–3714.
et al: Imperatoxin A, a cell-penetrating peptide cainide prevents catecholaminergic polymorphic 150. Coronado R, Miller C: Decamethonium and
from scorpion venom. as a probe of Ca-release ventricular tachycardia in mice and humans. Nat hexamethonium block K+ channels of sarcoplas-
channels/ryanodine receptors. Pharmaceuticals Med 15(4):380–383, 2009. mic reticulum. Nature 288:495–497, 1980.
(Basel) 3(4):1093–1107, 2010. 131. Hilliard FA, Steele DS, Laver D, et al: Flecainide 151. Cukierman S, Yellen G, Miller C: The K+ channel
111. Rousseau E, Meissner G: Single cardiac sarcoplas- inhibits arrhythmogenic Ca2+ waves by open of sarcoplasmic reticulum. A new look at Cs+
mic reticulum Ca2+-release channel: activation by state block of ryanodine receptor Ca2+ release block. Biophys J 48:477–484, 1985.
caffeine. Am J Physiol 256(2 Pt 2):H328–H333, channels and reduction of Ca2+ spark mass. J Mol 152. Oosawa Y, Sokabe M: Voltage-dependent amino-
1989. Cell Cardiol 48:293–301, 2010. glycoside blockade of the sarcoplasmic reticulum
112. Kong H, Jones PP, Koop A, et al: Caffeine induces 132. Liu N, Denegri M, Ruan Y, et al: Flecainide exerts K+ channel. Am J Physiol 250:C361–C364,
Ca2+ release by reducing the threshold for an antiarrhythmic effect in a mouse model of 1986.
153. Liu QY, Rasmusson RL, Liu QX, et al: Voltage- 155. Takeshima H, Shimuta M, Komazaki S, et al: plasmic reticulum vesicles from rabbit skeletal
dependent, open channel blockade of the cardiac Mitsugumin29, a novel synaptophysin family muscle. Biophys J 70:202–221, 1996.
sarcoplasmic reticulum potassium channel by
4-aminopyridine. Can J Cardiol 14:275–280, 1998.
154. Picard L, Cote K, Teijeira J, et al: Sarcoplasmic
member from the triad junction in skeletal
muscle. Biochem J 1998;331(1):317–322.
156. Yazawa M, Ferrante C, Feng J, et al: TRIC chan-
158. Kawano S, Kuruma A, Hirayama Y, et al: Anion
permeability and conduction of adenine nucleo-
tides through a chloride channel in cardiac sarco-
6
reticulum K+ channels from human and nels are essential for Ca2+ handling in intracel- plasmic reticulum. J Biol Chem 274:2085–2092,
sheep atrial cells display a specific electro- lular stores. Nature 448:78–82, 2007. 1999.
pharmacological profile. J Mol Cell Cardiol 157. Kourie JI, Laver DR, Junankar PR, et al: Charac-
34:1163–1172, 2002. teristics of two types of chloride channel in sarco-
Structural and Molecular Bases of
Mitochondrial Ion Channel Function 7
Jin O-Uchi, Bong Sook Jhun, and Shey-Shing Sheu
different Ca2+ affinity, uptake kinetics, and pharmacological char-

CHAPTER OUTLINE
acteristics from the original MCU theory (see review3) (Figure
Ion Channels/Transporters at Mitochondrial 7-1, C). These distinct properties are critical for enabling mito-
Inner Membrane 74 chondria to carry out multiple Ca2+-mediated functions with
optimal spatial and temporal effectiveness. Among these studies,
Ion Channels/Transporters at Mitochondrial RyR1 was found as the first mitochondrial Ca2+ influx mechanism
Outer Membrane 78 with a known molecular identity reported from our group (Figure
Mitochondrial Ion Channels/Transporters in Cardiac 7-1, C). IMiCa was recently recorded from mitoplasts (used mito-
Function and Dysfunction 80 chondria without the outer mitochondrial membrane [OMM]),5
providing direct electrophysiological evidence for the existence
Closing Remarks 82 of a Ca2+ selective ion channel, which would possibly fit he origi-
nally predicted channel nature of MCU. Through RNA interfer-
ence studies, several groups have recently proposed novel
candidate proteins that are involved in mitochondrial Ca2+ uptake
Historical Overview of Mitochondrial Ion mechanisms such as leucine zipper-EF-hand containing trans-
Channel/Transporter Research membrane protein 1 (Letm1)6 and mitochondrial calcium update
(MICU1),7 in addition to our RyR1 discovery (Figure 7-1, C).
Mitochondria are the power plants of all kinds of tissues/cells. In Finally, two very recent articles reported that the coiled-coil
heart/cardiomyocytes, mitochondria use glucose and fatty acids domain containing protein 109A (CCDC109A) is the MCU
to produce adenosine triphosphate (ATP) that drives muscle con- molecular identity.8,9
traction and relaxation for pumping blood that circulates through In addition to Ca2+, the movements of various electrolytes and
the entire body during each heart beat unceasingly, up to a cen- metabolites across the inner mitochondrial membrane (IMM) as
tenary human life. Indeed, mitochondria were originally found well as the OMM are important for the regulation of major
and studied mostly as a cellular “powerhouse” in the first half of mitochondrial functions, including ATP synthesis, Ca2+ homeo-
the 20th century. Soon it was also recognized that Ca2+ stimulates stasis, and reactive oxygen species (ROS) and nitrogen species
oxidative phosphorylation (OXPHOS) and electron transport (RNS) generation. Unlike other organelles, mitochondria possess
chain (ETC) activity, which results in the stimulation of ATP unique double-membrane structures with distinctive phospholip-
synthesis1 (Figure 7-1, A, B). Early studies in the 1960s to 1970s ids and protein compositions, which allow mitochondrial mem-
revealed that isolated mitochondria can take up a large quantity branes to maintain a mitochondrial membrane potential (Δψm)
of Ca2+. It is surprising to note that super-physiological high Ca2+ and unique architecture including cristae (see next section). It is
concentrations ([Ca2+]) (10 to 100 µM) were required to activate interesting to note that IMM and OMM have different sets of
Ca2+ uptake into isolated mitochondria (see reviews2-4). However, ion channels/transporters, as summarized in Figure 7-1A. These
in the intact cells, less than a 10-µM [Ca2+] global increase in the include (1) proton (H+) movement related to ETC activity and
cytosol propagated into the mitochondria matrix. This discrep- uncoupling proteins (UCPs) for the maintenance of Δψm at IMM;
ancy between isolated mitochondria and intact cells was partially (2) K+-selective and anion-selective pathways at IMM and across
resolved by the finding of high cytosolic [Ca2+] ([Ca2+]c) at the the two membranes, such as the mitochondrial permeability tran-
microdomains between mitochondria and endoplasmic reticulum sition pore (mPTP), which are important for the maintenance of
(ER)/sarcoplasmic reticulum (SR) during ER/SR Ca2+ release via mitochondrial volume; (3) the movement of metabolites includ-
inositol triphosphate (IP3) receptor and/or ryanodine receptor ing ATP, adenosine diphosphate (ADP), and phosphate (Pi) at
(RyR), as a result of the juxtaposition of these two organelles. IMM and OMM, and (4) release of proapoptotic proteins, which
These seminal discoveries have finally positioned mitochondria potentially leads to cell death mainly at OMM (see III) or mPTP.
as one of the key players in the dynamic regulation of physiologi- In this chapter, we summarize recent progress in mitochondrial
cal Ca2+ signaling and have promoted research in the field of ion channel/transporter research; this is followed by an overview
mitochondrial Ca2+ channels/transporters. Soon it was also dis- of cardiac mitochondrial ion channel/transporter biophysics and
covered that a Ca2+ efflux mechanism exists to dump the accu- cardiac physiology and pathophysiology.
mulated matrix Ca2+ into the cytosol. Although the functional
characteristics of mitochondrial Ca2+ influx/efflux mechanisms
were functionally discovered more than 50 years ago, the molecu- Overview of Mitochondrial Bioenergetics and
lar identities responsible for these mechanisms have remained a Mitochondrial Membrane Potential
mystery until very recently. The mitochondrial Ca2+ influx was
dogmatically considered to result from a single transport mecha- The most prominent contribution of mitochondria to cellular
nism mediated by a mitochondrial Ca2+ uniporter (MCU) prin- metabolism is based on their capacity to generate ATP through
cipally as the result of nearly complete inhibition by ruthenium the tricarboxylic acid (TCA) cycle and OXPHOS through the
red (RuR) and lanthanides, and its channel nature was originally ETC, which is a concerted series of redox reactions catalyzed by
proposed more than 30 years ago.2,4 However, subsequent studies four multi-subunit enzymes embedded in the IMM (complexes
have identified additional Ca2+ uptake pathways, which exhibit I-IV) and two soluble factors, cytochrome c (cyt c) and coenzyme
71
Q10 (Co Q10), which function as electron shuttles within the not only have different phospholipid compositions, but they show
mitochondrial intermembrane space (IMS) (see Figure 7-1, B) different protein-to-lipid ratios (for OMM, about 0.5 : 0.5; and
(see I.A and II.D). In a healthy eukaryotic cell, more than 90% for IMM, about 0.8 : 0.2). This may allow the proteins at the
of the total intracellular ATP is generated by mitochondria. The IMM to possess enzymatic and/or transport functions compared
main driving force of OXPHOS is known as “chemiosmosis,” with those at the OMM, thus the IMM is much less permeable
which is generated by proton (H+) movement across IMM that to ions and small molecules than is the OMM, which also pro-
creates a membrane potential (Δψm, negative in the matrix) and vides the cellular compartmentalization between the mitochon-
a pH gradient (ΔpH, alkaline in the matrix). Chemiosmosis is the drial matrix and cytosol. As is shown in Figure 7-1, B, complexes
movement of ions across a selectively permeable membrane, I, III, and IV are engaging with the translocation of H+ from
down their electrochemical gradient (protonmotive force: Δp), matrix to IMS, which establishes Δψm and ΔpH, as well as Δp.
which is determined by both Δψm and ΔpH components across Therefore, Δψm is usually highly negative (around −180 to
the IMM (Δp =Δψm + ΔpH). The chemiosmotic hypothesis was −190 mV) compared with the resting potential at plasma mem-
first proposed by Peter D. Mitchell in 1961.10 The basic assump- branes. Finally, according to the mentioned mechanisms, this
tion of this chemiosmotic theory is derived from the important large driving force for H+ influx (Δp) is used by complex V to
observation that the IMM is generally impermeable to ions, but produce ATP (see Figure 7-1, B). Other important roles for Δp
it keeps the permeability of H+. The composition of the OMM in addition to ATP synthase at the IMM are that (1) ΔpH drives
is similar to those of sarcolemma (SL) and ER/SR in eukaryotic pyruvate transport through pyruvate carrier (PYRC) into the
cells, whereas the IMM does not possess cholesterol but has a matrix; (2) ΔpH drives Pi transport through Pi carrier (PIC) into
unique dimeric phospholipid, cardiolipin, which is a typical com- the matrix; and (3) Δψm drives ATP/ADP exchange through the
position for bacterial membranes. Cardiolipin has a unique ability adenine nucleotide translocator (ANT) (see Figure 7-1, A, B).
to interact with proteins, including several mitochondrial respira- As mentioned above, Ca2+ uptake into the mitochondrial
tory chain complexes (I, III, IV, and V), and support their activi- matrix stimulates ATP synthesis (see Figure 7-1A). At the resting
ties11 while also contributing to maintaining the structure of state, the electrochemical driving force for Ca2+ uptake is also
cristae, which enhances the efficiency of ETC activity, possibly provided by Δψm across the IMM. For MCU, Ca2+ is taken
through facilitating the formation of super-complexes of respira- into the mitochondrial matrix down its electrochemical gradient
tory chain at the IMM.12 The unique structure of cardiolipin without transport of another ion. Basically, for each Ca2+
serves as an H+ trap at the IMS near the IMM, maintains the pH transported through MCU, there is a net transfer of two
change near the IMM, and efficiently pools H+ or releases H+ to positive charges into matrix, resulting in a drop in Δψm, which is
the mitochondrial ATP-synthase (complex V) at IMM (Figure energetically unfavorable. However, the Ca2+-stimulated respira-
7-1, B).10,11 It is interesting to note that the IMM and the OMM tion not only will compensate the loss of Δψm by the efflux of H+
Figure 7-1. Overview of Mitochondrial Ion Channels/Transporters A, Major ion movements in mitochondria. Ca2+-dependent ion channels/transporters and
enzymes are indicated with stars.
Structural and Molecular Bases of Mitochondrial Ion Channel Function 73
Figure 7-1, cont’d. B, Mitochondrial electron transport chain (ETC) and possible sites of superoxide production. Red “explosion” symbols indicate places where superoxide
production occurs. ETF-QF, electron transferring flavoprotein-quinone oxidoreductase; Q, coenzyme Q10. C, Ca2+-influx/efflux mechanisms: The channels/transporters for
which molecular identities are still unknown are shown as black. Red arrows show Ca2+ movements, and blue arrows show other ion movements. DroCRC, Ca2+-release
channel in Drosophila mitochondria; RaM, the rapid mode of uptake.
(©O-Uchi J et al: Perspectives on: SGP symposium on mitochondrial physiology and medicine: Molecular identities of mitochondrial Ca2+ influx mechanism: Updated passwords
for accessing mitochondrial Ca2+-linked health and disease. J Gen Physiol 139:435–443, 2012. Originally published in Journal of General Physiology, doi:10.1085/jgp.201210795.)
through ETC, it will also produce a net gain of ATP. In addition, form a diverse array of additional ROS and RNS.13 High
multiple Ca2+ efflux mechanisms work in concert to expedite a levels of ROS and RNS are known to promote cell damage
transient and an oscillatory nature rather than a tonic and a steady and death, but the production of low to moderate levels of ROS/
state change of matrix [Ca2+] ([Ca2+]m). RNS is critical for the proper regulation of many essential cel-
lular processes, including gene expression, signal transduction,
and cardiac excitation-contraction (E-C) coupling.14 ROS are
Overview of Mitochondrial ROS Generation generated by several different cellular sources: (1) membrane-
and Mitochondrial Membrane Potential associated nicotinamide adenine dinucleotide phosphate
(NADPH) oxidase, (2) cytosolic xanthine and xanthine oxidase,
Superoxide and nitric oxide are ROS and RNS, respectively, and (3) the mitochondrial ETCs at IMMs.
produced in cells under normal physiological conditions; Superoxide is the primary oxygen free radical produced in
both species react with other molecules and with each other to mitochondria via the slippage of an electron from the ETC to
molecular oxygen during OXPHOS (see Figure 7-1, B). This spilling of proapoptotic proteins into the cytosol. In contrast, the
“constitutive” superoxide generation is central to proper cellular other two pathways occur without Δψm depolarization.
redox regulation. Recent studies from our collaborating groups
detected a “stochastic” and “transient” superoxide burst from
either single or restricted clusters of interconnected mitochondria
across a wide variety of cell types, termed a mitochondrial superoxide Ion Channels/Transporters at Mitochondrial
flash (mSOF).15 A transient depolarization of Δψm is associated
with each mSOF. The proposed mechanism of mSOF in cardiac
Inner Membrane
muscle cells is as follows: A small increase in constitutive ROS
production transiently opens a large channel named the mPTP to Mitochondrial Ca2+ Channels/Transporters
cause Δψm depolarization, which subsequently stimulates the Regulating Mitochondrial Ca2+ Influx
ETC to produce a burst of superoxide production. This idea is
similar to the previous observation that the mPTP opens and Mitochondrial Ca2+ Uniporter (MCU)
closes transiently (“flicker”) at its low conductance state and can At first, an important regulatory protein of the mitochondrial
release Ca2+ from matrix2. The frequency of mSOFs varies widely Ca2+ uptake mechanism was identified using bioinformatics and
across different cell types and experimental conditions (disease siRNA screening, termed MICU17. MICU1 has two Ca2+-bind-
models), suggesting that these flashes act as biomarkers of cellular ing EF-hands, but has only one single membrane-spanning
metabolic activity and oxidative condition in physiological and domain, which seems unlikely to form a channel pore and to be
pathophysiological conditions.16 Future studies will clarify the an MCU itself rather than a modulator of MCU (see Figure
contribution of altered mSOF activity to ROS overproduction 7-1, C). Finally, two recent articles from two different groups
and metabolic dysfunction in a wide range of mitochondrial dis- simultaneously reported CCDC109A as the molecular identity
eases and oxidative stress–related disorders. Finally, it should be of the MCU.8,9 The main characteristics of the MCU are as
noted that there is controversy surrounding whether the mSOF follows: (1) CCDC109A has two transmembrane domains, which
might reflect oscillations in mitochondrial pH.17 Based on chemi- seem likely to make a channel pore; (2) knock-down of
osmotic theory, one can envision that whenever a significant drop CCDC109A dramatically reduces mitochondrial Ca2+ uptake,
in Δψm occurs, there should be an alkalization in ΔpH. Thus, the and this effect was rescued by overexpression of MCU;
interrelationship between mSOF, pH transients, and Δψm fluctua- (3) CCDC109A knock-down itself does not affect mitochondrial
tions will be an important research topic in the future. O2 consumption, ATP synthesis, Δψm, and morphology, even
though Ca2+ influxes are critical for regulating these functions;
(4) reconstituted MCU in lipid bilayers shows RuR-sensitive
Overview of Mitochondria-Induced Apoptosis and Ca2+ current with 6-7 pS single-channel activity9; and (5) site-
Mitochondrial Ion Channels specific mutations at the MCU pore region show loss of function.
Regarding the structure of MCU, the topology of MCU is
Programmed cell death is genetically designed for self-killing of still under debate. CCDC109A seems to possess two transmem-
individual cells and is one of the critical mechanisms for main- brane domains that oligomerize to become a Ca2+ channel. Riz-
taining homeostasis of multicellular organs/tissues, including the zuto’s group proposed that C- and N-terminals face to the IMS,
heart.18 Apoptosis is a well-established mechanism of programmed whereas Mootha’s group proposed that they face the opposite
cell death activated by a variety of cellular stresses and signals.19 direction.22 The discrepancy in the topology of MCU will
In mammalian cells, the activation of caspases (a family of cyste- need to be resolved for an understanding of the modulation of
ine proteases) is the central mechanism for apoptosis. Under the MCU functions by signaling molecule from the cytosol or from
resting condition, caspases are tightly kept inactive as a “proen- the matrix.
zyme” form and/or by binding to inhibitory proteins (named
inhibitors of apoptosis) in the cytosol. During apoptosis, caspases Mitochondrial Ryanodine Receptor (mRyR)
are activated by cleavage that changes the “proenzyme” form to One of the candidates for the mitochondrial Ca2+ uptake mecha-
the enzyme form and by dissociation of the inhibitors of apop- nism with a known molecular identity is the mitochondrial RyR
tosis.20 One of the major apoptotic pathways is derived from in cardiac cells reported from our group.23 There are three dif-
mitochondria, and this contributes to the activation of caspases ferent RyR isoforms (RyR1, 2, and 3) with different physiological
in the cytosol. This mechanism is initiated by the release of and pharmacological properties. In cardiac cells, intracellular
proapoptotic factors from mitochondria, including cyt c, Smac/ Ca2+ release during E-C coupling was mainly controlled by RyR2
Diablo, and HtrA2/Omi. Cyt c released to cytosol activates cas- located in the SR (see Figures 7-1, A-C). Although RyR1 is also
pases through its binding to apoptotic protease activating factor detectable at mRNA and protein levels in cardiac tissue, its func-
1 (Apaf-1). Smac/Diablo and HtrA2/Omi bind to the inhibitory tional and physiological roles in the heart have not been fully
proteins to remove their inhibitory effects from caspases. The understood for a long time. We first showed that a low level of
apoptotic DNases, including apoptosis inducing factor (AIF) and functional RyR is also expressed at the heart IMM, and this has
endonuclease G (Endo G), are also released from mitochondria, a role in the fast Ca2+ uptake pathway24 (see Figures 7-1, B and
as are the aforementioned proapoptotic factors. The release of 7-2). Furthermore, RyR in cardiac mitochondria exhibits remark-
these proapoptotic factors is regulated under the tight control of ably similar biochemical, pharmacologic, and functional proper-
mitochondrial membrane permeability.21 This mitochondrial ties to those of RyR1 in skeletal muscle SR, but not to those of
membrane permeabilization is mediated via at least three distinct RyR2 in cardiac SR. Therefore, we termed this as mRyR1 (mito-
mechanisms: (1) physical rupture of the OMM as a result of chondrial RyR1)(see review3). The molecular identity of mRyR1
mitochondrial swelling (usually linked to mPTP opening) (II.B.2), was carefully analyzed and confirmed by a variety of functional
(2) modification of the structure of the voltage-dependent anion and biochemical experiments using not only native heart, but also
channel (VDAC) through interaction with proapoptotic proteins the RyR1 knockout mouse heart. Recent studies from our lab
such as the Bcl-2 protein family, and (3) formation of a new pore using electrophysiological techniques directly demonstrated the
as a consequence of oligomerization and membrane insertion existence of mRyR1 and clearly showed the predicted channel
of proapoptotic proteins, including the Bcl-2 family of proteins. nature of skeletal RyR1. In a conventional lipid bilayer of mRyRs
The first pathway involves Δψm depolarization and swelling of purified from a heart IMM, the activity of RyR1 but not of RyR2
the matrix space, followed by loss of OMM integrity and rupture was observed.25 The biophysical and pharmacologic properties of
(mol/L) (mol/L) RyR/IP3R
[Ca2+]ER-mito
[Ca2+]m
Ram
MCU
10–7 mRyR1 LETM1
10–8
0 0
A
(mol/L) (mol/L)
RyR/IP3R
10–5
[Ca2+]ER-mito
[Ca2+]m
Ram
MCU H+
10–6 mRyR1
10–7
0 0 LETM1
(mol/L) (mol/L)
RyR/IP3R
10–5
[Ca2+]ER-mito
[Ca2+]m
Ram
10–6 MCU H+
10–7 mRyR1
0 0 LETM1
(mol/L) (mol/L)
RyR/IP3R
10–5 5x10–5
[Ca2+]ER-mito
[Ca2+]m
Ram
10–6 MCU
mRyR1
0 0 LETM1
(mol/L) (mol/L)
10–5 5x10–5
[Ca2+]ER-mito
[Ca2+]m
mRyR1 Ram MCU
H+
0 0 NCX LETM1
Na+
E
Figure 7-2. Activation/Deactivation Patterns of Ca2+-Influx/Efflux Mechanisms A, At first, the rapid mode of uptake (RaM) (black) is activated at the very
initial phase of Ca2+ transient at microdomains between mitochondria and ER/SR ([Ca2+]ER-mito) transient (red line) (<200 nM) with faster Ca2+ uptake kinetics (ms time scale).
B, Letm1 (orange) starts to uptake Ca2+ at ≥200 nM [Ca2+]ER-mito. C, mRyR1 (red) starts open at ≈1 µM [Ca2+]ER-mito with a 5-fold faster Ca2+ transport compared with the MCU,
and deactivates before [Ca2+]ER-mito reaches the peak. D, Finally, MCU (blue) starts to activate at >1 µM [Ca2+]ER-mito, and the activity increases in a [Ca2+]ER-mito,-dependent manner.
At this point, Letm1 (orange) shifts from Ca2+-uptake mode to Ca2+-efflux mode. E, mPTP (black) and NCX (purple) contribute to Ca2+ efflux in mammalian cells and form
the decay phase of [Ca2+]m transient (blue line). Letm1 also works as a Ca2+-efflux pathway at this phase.
(©O-Uchi J et al: Perspectives on: SGP symposium on mitochondrial physiology and medicine: Molecular identities of mitochondrial Ca2+ influx mechanism: Updated passwords
for accessing mitochondrial Ca2+-linked health and disease. J Gen Physiol 139:435–443, 2012. Originally published in Journal of General Physiology. doi:10.1085/jgp.201210795.)
native single mRyR1 channels were further characterized in heart Ca2+-efflux mechanism is Na2+-dependent in cardiac
mitoplasts using patch-clamp techniques.26 We observed a novel mitochondria.28
225-pS cation-selective channel in heart mitoplasts that exhibited
multiple subconductance states, which was blocked by high con- Mitochondrial Permeability Transition Pores (mPTP)
centrations of ryanodine and RuR, the known inhibitors of RyRs. The mPTP is a large, nonspecific channel that opens at the IMM
Ryanodine exhibited a concentration-dependent modulation of and is known to form under mitochondrial stress conditions such
this channel, with low concentrations stabilizing a subconduc- as mitochondrial Ca2+ overload and elevated oxidative stress,
tance state and with high concentrations abolishing activity. The which causes the release of huge amounts of Ca2+ and proapop-
channel properties of Ca2+-dependent [3H]ryanodine binding and totic proteins from mitochondria, subsequently leading to cell
the channel modulation by caffeine in isolated cardiac mitochon- death.29 Pioneering studies showed that solutes up to 1.5 kDa can
dria24 are suggesting that the topology of mRyR1 is the same as pass through this channel, suggesting that the pore diameter is
that of RyR1 at SR because of these agonist-binding sites; C- and ≈3 nm. One of the important consequences of the mPTP opening
N-terminals face to the IMS (corresponding to the cytoplasmic is that the IMM no longer maintains a barrier to H+, which leads
side of RyR1 at SR), and S1-S2, S3-S4, and S5-S6 linkers face to to Δp dissipation and thus inhibits ATP production. Because of
the mitochondrial matrix side (corresponding to the SR luminal its pore size, mPTP opening also results in equilibration of cofac-
side of RyR1 at SR). However, further studies will be needed to tors and ions across the IMM, including the release of accumu-
confirm the topology of mRyR1 using other modulators from lated Ca2+. This leads to disruption of metabolic gradients
both the matrix side and the cytosolic side. between mitochondria and cytosol, and influx of water concomi-
Unlike MCU, RyR is a poorly Ca2+-selective, large cation tantly occurs, resulting in swelling of the mitochondria until the
channel.26 Therefore, opening of mRyR1 might collapse Δψm, OMM eventually ruptures. The OMM rupture releases cyt c and
which is energetically unfavorable. This dichotomy would be other proapoptotic proteins, potentially leading to apoptotic cell
explained as follows: (1) The expression number of RyR1 in a death. The mPTP opening plays a significant role in the genera-
single mitochondrion is very small, and the changes in Δψm might tion not only of apoptotic cell death but also of necrotic cell
be minimized locally, ensuring maintenance of a Ca2+-driving death, both of which are involved in the etiology of myocardial
force; (2) the rapid Ca2+-activation and -inactivation profile of infarction. It is now widely recognized that mPTP opening is a
this channel (see Figure 7-2) would minimize the Δψm change major cause of reperfusion injury and is an effective target for
instantly; and (3) any small decrease in Δψm can be readily com- cardioprotection.
pensated by the Ca2+-dependent activation of dehydrogenase It is interesting to note that the mPTP has been also shown
during the TCA cycle and ATP synthesis. Taken together, the to open and close transiently (“flicker”) at its low conductance
mRyR may be uniquely poised to sequester Ca2+ during a tran- state, and this may play a physiological role in the Ca2+-efflux
sient and rapid excitation-contraction coupling process in cardiac mechanism2. Therefore, the mPTP is also considered one of the
muscle cells. important Ca2+-efflux mechanisms both in physiological and
pathophysiological conditions (see Figures 7-1 and 7-2). The
Letm1 (Ca2+/H+ Antiporter) properties of the mPTP are well defined, as summarized in
Letm1 was first found as a K+/H+ exchanger (KHX). Recently, Figure 7-3, A. However, despite extensive research by many labo-
with the use of siRNA genome-wide screening in Drosophila, it ratories, its exact molecular identity remains uncertain (Figure
was proposed also as an H+/Ca2+ exchanger (HCX) at the IMM, 7-3, B). On the basis of previous biochemical and pharmacologic
importing 1 Ca2+ and extracting 1 H+6 (see Figures 7-1 and 7-2). studies, the mPTP was proposed to consist of (1) the VDAC in
Knockdown of Letm1 abolished only the initial fast mitochon- OMM, (2) the ANT in IMM, and (3) cyclophilin-D (CypD) in
drial Ca2+ uptake but still showed sustained Ca2+ increase, sug- the matrix (Figure 7-3, B). Indeed, pharmacologic inhibitors of
gesting that Letm1 works at low [Ca2+]c for Ca2+ uptake (see ANT or CypD also inhibit mPTP opening (see Figure 7-3, A).
Figure 7-2). Letm1 activity was inhibited by both RuR, an inhibi- However, recent genetic studies using knockout mouse models
tor of MCU, and CGP37157, an inhibitor of the Na+/Ca2+ reveal that mPTP opening can still occur in the absence of
exchanger (NCX). This report seems like the revival of the story VDAC and ANT, but much higher concentrations of Ca2+ are
by Moyle and Mitchell in 1977 and raises several points of discus- required to trigger pore opening.29 Therefore, these proteins
sion: (1) 1 Ca2+ for 1 H+ antiporter does not favor Ca2+ influx have been currently established as regulatory components of the
physiologically according to the electrochemical gradients of mPTP, and only CypD remains potentially an essential compo-
Ca2+ and pH; (2) Ca2+ influx by Letm1 might be mediated in part nent of mPTP. A remaining fundamental question of mPTP
by changes in Δψm through K+ fluxes because of Letm1 being research is how to clarify the molecular identities of this channel.
itself a KCX; and (3) CGP37157 has not been shown to inhibit It might be possible to use CypD as bait to fish out channel
an IP3-mediated [Ca2+]m increase (see review3). proteins among CypD partnering proteins using proteomic tech-
nologies. Halestrap’s group has proposed that the PIC may be
one of the components of the mPTP.29 However, PIC knockdown
in HeLa cells could not abolish mPTP opening, and an essential
Mitochondrial Ca2+ Channels/Transporters role for the PIC in mPTP formation remains unproven.30 Simi-
Regulating Mitochondrial Ca2+ Efflux larly, genetic screens in yeast and Drosophila might be able to add
further information to the molecular identities of mPTP compo-
Mitochondrial Na+/Ca2+ Exchanger (NCLX) nents. Bernardi’s group found that Drosophila mitochondria
The main mitochondrial Ca2+-efflux pathways are Na+/Ca2+ possess another selective Ca2+-release channel with unique fea-
exchanger (NCX) and/or H+/Ca2+ exchanger (HCX) (see Figures tured characteristics, including (1) intermediate pharmacologic
7-1 and 7-2). Na+- or Li+-dependent Ca2+ transport (NCLX) was properties between the mPTP of yeast and mammals, such as
first cloned as a sixth member of the K+-dependent NCXs sug- inhibition by Pi but not by ADP and cyclosporin A (CsA), as in
gested to be located at the ER or plasma membrane. Recently, the mPTP of yeast mitochondria; and (2) the existence of voltage-
NCLX was shown as the candidate for the mitochondrial NCX.27 and redox-sensitive regulatory sites, as in the mPTP of mam-
The NCLX expression is particularly robust in excitable cells, malian cells31 (see Figure 7-1, C). Taken together, these novel
whereas the activity of the HCX is primarily found in nonexcit- technologies will provide us with further information related to
able cells, suggesting the existence of tissue-specific mitochon- the molecular identities of the mPTP, as well as its physiological
drial Ca2+-efflux mechanisms. Indeed, the primary mitochondrial and pathophysiological properties in cardiomyocytes.
Proposed Effects to
7
Effectors mechanisms mPTP
via Cyp-D binding via change in nucleotide via Ca2+ binding other
to mPTP binding to the ANT to the mPTP mechanism
Oxidative stress (e.g., reperfusion,
t–butylhydroperoxide or diamide)
Thiol reagents (e.g., eosin
maleimide, phenylarsine oxide,
arsenite)
Increased matrix volume
Chaotropic agents
Adenine nucleotide depletion
Increase in matrix (Pi) and (PPi)
Alkalosis
Acidosis
Depolarization/uncoupling
Long chain fatty acids
CsA
SfA
Mg2+,Mn2+,Sr2+,Ba2+
Subiquinone analogues (e.g.,
decyl-ubiquinone, ubiquinone 10)
Peroxynitrite and hydroxynonenal
Lonidamine
Trifluoperazine
Phenylglyoxal
Monobromobimane
Bongkrekic acid
Activation Inhibition Activation/Inhibition

A
VDAC
? ? Bax
Bak
? ? ANT
ANT
CypD PiC ??
B
Figure 7-3. Mitochondrial Permeability Transition Pores (mPTP) A, List of known mPTP effectors. Probable sites of action for the effectors are shown as black
boxes, red boxes, yellow boxes, and a blue box, indicating the effects to mPTP (activation, inhibition, and both effects, respectively). B, Scheme illustrating the proposed role
of CypD, ANT, and PIC in formation of the mPTP structure.
Mitochondrial K+ Channels mitochondrial proteomic approach.35 This novel finding will

need to be followed up using in vivo cardiac disease models.
Mitochondrial KATP Channel (mitoKATP) Because human ROMK mutations are well known as one of the
responsible genes for the salt-wasting nephropathies (such as
More than 20 years ago, one of the K+ channel openers, nicor- Bartter’s syndrome), further observations will be required before
andil, was demonstrated to protect hearts against ischemia- we can understand the mechanism for the sudden cardiac arrest
reperfusion (I/R) injury (see review32). Earlier studies had and ventricular arrhythmias occasionally reported in these
indicated that the K+ channel openers including nicorandil were patients.
targeting ATP-sensitive K+ channels (KATP) at SL (sarcKATP) using
pharmacologic and genetic approaches in small animals. Later, it Mitochondrial KCa Channel (mitoKCa)
became clear that in larger species including humans, sarcKATP Mitochondrial Ca+-activated K+ channel (mitoKCa) was function-
has a minor role in protecting the heart during ischemia, sug- ally identified using the direct patch-clamp of mitoplasts obtained
gesting that other targets of these drugs such as mitochondria from cardiac myocytes36 (see Figure 7-1, A). This channel shows
have essential roles in protection of the heart from metabolic ≈300 pS conductance and is inhibited by K+ channel toxin cha-
stress. Indeed, by using the direct mitoplast patch-clamp tech- rybdotoxin, which exhibits channel properties similar to Ca2+-
nique, the presence of KATP activity was reported at the IMM activated K+ channel found in SL (sarcKCa). Similar to mitoKATP,
named the mitochondrial KATP (mitoKATP) (see Figure 7-1, A), mitoKCa is thought to be essential for IPC and cardioprotection,
exhibiting channel properties similar to sarcKATP.33 Currently, but the molecular identity of this channel remains unclear.34 Gen-
mitoKATP is thought to be essential for cardioprotection erally, big conductance Ca2+-sensitive K+ (BK) channels are ubiq-
induced by ischemic preconditioning (IPC).34 O’Rourke’s uitously expressed in various cell types, but these channels are
group recently reported that one of the splice variants of the renal not functionally expressed in cardiac SL. A recent study showed
outer medullary K+ channel (ROMK) (Kir1.1) is a suitable can- that BK channel subunits are expressed in the IMM of brain and
didate for the pore-forming subunit of mitoKATP using the cardiomyocytes, and that activation of this channel protects
Intermembrane Side
S193
S103 S136
S12
T51
Nek 1
Hexokinase 2 GSK3 Cytosolic Side
GSK3 PKC
A PKC GSK3
VDAC MAC Bax mPTP TOM

Peak conductance (ns) 0.5–4 1–5 1.0–5.4 1.1+
–0.1 0.7+
–0.1
Ion selectivity Anion Sl Cation Sl Cation Sl Cation Sl Cation
Voltage dependent Yes No No Yes No
B
Figure 7-4. Voltage-Dependent Anion Channel (VDAC) A, Schematic diagram of predicted phosphorylation sites in VDAC1 structure. B, Comparison of biophysi-
cal properties between VDAC, other OMM channels (MAC and TOM), mPTP, and a reconstituted “Bax channel.” Bax channel function was observed in the reconstituted
channel using purified human Bax protein. TOM, Translocase of the outer membrane of mitochondria.
normoxic infant rabbit hearts.37 Further studies including genetic injury.41 mPTP opening is involved in the mechanism of mito-
approaches will be needed to confirm whether the protein of chondrial swelling. Since the 1970s, the mitochondrial K+ cycle,
mitoKCa is identical or similar to the BK channels at SL. especially counterbalancing MCU activity with the KHX, has
been identified as an important component of mitochondrial
volume homeostasis.40 Not only cation movements but also anion
Proton Fluxes and Uncoupling Proteins movements into the matrix induce mitochondrial swelling. Mito-
chondrial swelling experiments show that an anion-selective
Electron flow through the ETC is carried out by four enzyme channel is present in the IMM. This channel is activated by
complexes at the IMM, the cyt c, and the mobile carrier Co Q10 matrix Mg2+ depletion and alkalization and is named the IMM
(see Figure 7-1, A, B). Electron transfer processes through com- anion channel (IMAC).42 The physiological role of IMAC is still
plexes I, III, and IV produce Δp that in turn are used to drive unknown because IMAC appears to conduct ions only under
ATP synthase (complex V) (see Figure 7-1, B). When Δp increases, alkaline matrix conditions. However, this anion efflux through
electron transport in complex III is partially inhibited, resulting IMAC may be well designed to enable mitochondria to restore
in the increased backup of electrons to Co Q10 for binding to their normal volume following pathological swelling. We reported
molecular oxygen, and leading to the generation of superoxide. that single channel recordings of the heart IMM show a variety
The main route for the proton flow, driven by the electrochemi- of types of anion-selective conductance; the most prominent is
cal gradient, is through complex V and “proton leak.” The proton the centum picosiemen conductance channel, which matches the
leak is attributed by UCPs, which can modulate the ATP/ADP channel properties of IMAC.26 Currently, the molecular identity
ratio.38 Indeed, proton leak is known to contribute significantly of IMAC is still unknown.
to the control of respiration in mitochondria in state 4 and to
some extent in state 3. Therefore, UCPs are crucial to sustaining
proton leak, preventing excessively high H+ gradient, and ulti-
mately avoiding excessive ROS production. There are three iso-
Ion Channels/Transporters at Mitochondrial
forms of UCPs (UCP1, 2, and 3) with ≈60% homologies, but the Outer Membrane
isoform-specific roles of UCPs are still not clear.39 UCP2 is
ubiquitously expressed, including in the heart, whereas UCP3 is Voltage-Dependent Anion Channel (VDAC)
highly expressed in skeletal muscle, adipose tissue, and, to a lesser
extent, the heart. The proton leak pathway via UCPs might be VDAC is the most abundant protein in the OMM43 and serves
one of the mechanisms responsible for the cardioprotective effect as the main pathway for metabolite/ion transport between the
of IPC, but very little is known about the function or the regula- cytosol and the IMS of mitochondria (see Figure 7-1, A, C). Cur-
tion of UCPs in the adult heart. Further investigations with rently, three distinct VDAC isoforms are known to have high
sophisticated approaches such as genetic ablation of these pro- sequence homology (65% to 70% identity) and similar struc-
teins in cardiomyocytes will be required to advance this impor- ture.44 The recombinant VDAC1 and 2 isoforms are able to form
tant field of research. pores in lipid bilayers, but recombinant VDAC3 has no evident
pore-forming ability. Usually, VDAC is seven times more perme-
able to Cl− than to K+. VDAC channels can also exist in a variety
Anion Channels and Mitochondrial Volume Control of functional states that differ in their ability to pass nonelectro-
lytes and to conduct ions.45 VDAC exhibits ≈3 nS in 1 M NaCl
Mitochondrial volume has been proposed to modulate the rate in full conductance open state (Figure 7-4, B). In the open
of substrate oxidation, and mitochondrial swelling is a key issue state, it shows a significant preference for anions and especially
in cellular pathophysiology, such as during mitophagy40 and I/R favors metabolic anions. The closed state favors cations. The
A Fluorescence Microscope B Electronic Microscope
C Z-line M-line Z-line

T-tubules
T-tubules
Z-line M-line Z-line
Sarcolemma SR Mitochondria
MCU
L-type Ca2+ channel RyR2
mPTP
RYR1
NCX SERCa pump
NCLX VDAC
Figure 7-5. Overview of Mitochondrial Structure and Function in Heart A, Cardiac mitochondrial structure observed by fluorescence microscopy. Mouse
ventricular tissue was stained by CypD antibodies. B, Cardiac mitochondrial ultrastructure observed by electron microscopy. Mouse papillary muscle obtained from left
ventricle was observed by transmission electron microscopy. C, Schematic diagram of cardiac excitation-contraction coupling including mitochondrial structure and ion
channels. Left panel shows mitochondrial Ca2+-influx mechanisms at systole, and right panel shows mitochondrial Ca2+-efflux mechanisms at diastole.
Continued
Model l
Cyto
Mito
Model II
D
Figure 7-5, cont’d. D, [Ca2+]m oscillation models in cardiomyocytes.
permeability of VDAC to small anions by free diffusion includes mechanisms. The first pathway is mediated by mPTP opening
K+ and Na+ and the double positive charge ion Ca2+. Ca2+ perme- (see Figure 7-3), and the second pathway via modification of
ates through both open and closed states of VDAC; the double VDAC through the interaction of Bcl-2 families. The third
positive charge does not exclude Ca2+ from the open state because pathway is the mitochondrial apoptosis–induced channel (MAC),
anion selectivity is not very high in VDAC. Small nonelectrolytes which forms in the OMM during an early stage of apoptosis and
can also pass through the open channel, allowing the passage of directly provides a route for cyt c to release from the IMS to the
metabolites (ATP, ADP, and Pi). VDAC gating has a bell-shaped cytosol (see review50). Unlike VDAC, MAC is partially cation-
voltage-dependent profile, with peak currents around −10 mV.43 selective and has voltage-independent gating (see Figure 7-4, B).
A variety of factors are reported to modulate VDAC function, The pore size of MAC is sufficient to be permeable to 10 to
including colloidal osmotic pressure, protein(s) at the IMS, poly- 17 kDa proteins, which allows the passage of 12 kDa cyt c. MAC
anions such as charged proteins and nucleic acids, NADH and is regulated by the Bcl-2 family. Studies using reconstituted pro-
MgNADPH, tBid, actin, tubulin, and phosphorylation by intra- teins and knockout cell lines show that MAC activity is tightly
cellular signaling.45 VDAC was identified as a target of various connected to Bax/Bak expression and to their insertion and oligo-
protein kinases, and specific phosphorylation sites have been merization in the OMM, suggesting that at least these two pro-
identified (Figure 7-4, A). It has been reported that the phos- apoptotic proteins are crucial structural components (pore) of
phorylated state enhances tubulin binding to VDAC, followed by MAC (see Figure 7-5, B). Because the molecular mass of oligo-
blockage of channel activity.46 merized Bax obtained from mitochondrial membranes is of a very
Recent studies have focused on the regulation of OMM per- large size, MAC might associate with other mitochondrial pro-
meability by a physiological or pathophysiological mechanism. teins at the OMM.51
Although VDAC is excluded as a component of the mPTP,
VDAC is indeed one of the major factors involved in the regula-
tion of cell death signaling (see Figure 7-3, B). VDAC interacts Mitochondrial Ion Channels/Transporters
with apoptosis regulators, Bcl2-family members, hexokinases,
and the cytoskeleton system, including tubulin, which changes
in Cardiac Function and Dysfunction
VDAC channel activity.47 For instance, Bax forms hetero-
oligomers with VDAC and activates VDAC; this is followed by Overview of Mitochondrial Structure
cyt c release from IMS. Hexokinase-I has been shown to directly and Function in Heart
interact with VDAC1, which induces antiapoptotic effects. The
contribution of VDAC to cell death can be isoform- and stimulus- In adult cardiomyocytes, mitochondria are the dominant intra-
dependent: VDAC1 serves as a proapoptotic protein, whereas cellular organelle, numbering in the range of ≈7000 per cell with
VDAC2 protects from apoptosis.48 De Stefani and colleagues their mass occupying up to 35% of the cell volume.52 Three dif-
showed that the proapoptotic effect of VDAC1 is due to its physi- ferent subpopulations of mitochondria of unique size and shape
cal interaction with the IP3 receptor and to its formation of the are found in the adult heart: interfibrillar, subsarcolemmal,
molecular route for transferring Ca2+ signals to mitochondria in and perinuclear mitochondria.53 Interfibrillar mitochondria are
apoptosis.49 aligned in longitudinal rows between myofibrils in close proxim-
ity to SR Ca2+ release sites. They often span a single sarcomere
from Z-band to Z-band and are relatively uniform in size and
Apoptosis-Induced Channels shape (see Figure 7-5, A, B). This unique spatial localization of
interfibrillar mitochondria enables a privileged Ca2+-mediated
The release of these proapoptotic factors including cyt c is regu- crosstalk between SR and mitochondria (see Figure 7-5, C). In
lated by OMM permeability and by at least three distinct contrast, subsarcolemmal and perinuclear mitochondria appear
less organized and more variable in shape and size. Generally, microdomains involved in this coupling process, including RyR2,
7
mitochondrial morphology and dynamics including mitochon- VDAC, MCU, and mRyR1.
drial fission, fusion, and movement are well correlated with their A second controversy related to the kinetics of [Ca2+]m is
metabolic activity.53 In cardiomyocytes, the regulatory proteins whether mitochondria can take up and release Ca2+ in each heart-
responsible for mitochondrial dynamics including dynamin-like beat (an oscillator) (see Figure 7-5, D, Model I), or if they act as
protein (DLP1), mitofusin (Mfn)1, an integrator by taking up Ca2+ gradually, resulting in a steady
and Mfn2 are highly expressed. Using cardiac-specific Mfn state [Ca2+]m increase (see Figure 7-5, D, Model II).59 It is inter-
knockout mice, Maack and Dorn’s group showed that Mfn2 is esting to note that rabbit and guinea pig adult cardiomyocytes
important for the maintenance of bioenergetic responses via respond to [Ca2+]c transients on a beat-to-beat basis (Model I),
inter-organelle Ca2+ crosstalk, suggesting that Mfn2 serves as a but rat cardiomyocyte mitochondria respond to [Ca2+]c transients
tether of SR to mitochondria in cardiomyocytes.54 However, no with a gradual [Ca2+]m increase without obvious beat-to-beat
data directly show the subcellular localization of Mfn2 in native [Ca2+]m transients (Model II). This incapability of mitochondria
cardiomyocytes. Walsh’s group, using cardiac-specific Mfn to respond to [Ca2+]c transients might be due in part to the slow
knockout mice, showed that Mfn1 and Mfn2 have important kinetics of the mitochondrial Ca2+-efflux mechanism, but the
roles in the structural and metabolic development of cardiac mechanisms underlying these species differences in the kinetics
mitochondria during the postnatal state.55 Further studies are of [Ca2+]m are still unknown. Although mitochondria occupy 35%
required to clarify the specific role of the regulatory proteins for of the cytosolic space and are well known to perform uptake of
mitochondrial form and dynamics, including Mfn in adult Ca2+, it remains controversial whether mitochondrial Ca2+ uptake
cardiomyocytes. contributes to the kinetics of beat-to-beat CaT formation in
cardiomyocytes.60 Shannon and colleagues estimated that the
mitochondria contribute to only ≈1% of total Ca2+ removal from
the cytosol during CaT.61
Mitochondrial Ion Channels/Transporters in The view that MCU is not the sole mechanism for transduc-
Cardiac Excitation-Contraction/Metabolism ing the changes of [Ca2+]c into changes of [Ca2+]m has been gaining
Coupling and ROS Generation recognition. Moreover, it has been recently reported that cardiac
mitochondria contain two modes of Ru360-sensitive Ca2+ uptake:
Ca2+ plays a central role in E-C coupling of cardiac muscle56 (see a high Ca2+ affinity rapid uptake mode, and a low Ca2+ affinity
Figure 7-4, C). Ca2+ entry via the voltage-gated L-type Ca2+ slow uptake mode, which are responsible for modulating
channel (VLCC) triggers the opening of SR-RyR2 faced to the OXPHOS and Ca2+ buffering, respectively.28 Half-maximal acti-
transverse tubules (T-tubules) and induces a release of Ca2+ from vation (K0.5) of MCU current (IMiCa) observed in the mitoplast
SR (Ca2+-induced Ca2+ release [CICR]). The concomitant rise of patch-clamp is ≈20 mM in cases where single-channel activities
[Ca2+]c activates cardiac contraction by binding to troponin C. of MCU were specifically dissected by recognizing its unique
[Ca2+]c is then removed through the SR Ca2+ pump (SR Ca2+- biophysical and pharmacological characteristics.5 This K0.5 for
ATPase [SERCA]) or is extruded from the cell via the sarcolem- purified MCU current is quite different from that determined in
mal NCX. Usually, when amplitude and/or frequency of the isolated cardiac mitochondria, which ranges from 1 to 189 µM.
[Ca2+]c transient (CaT) becomes higher, cardiac work increases, This discrepancy might be due in part to the differences in
except for the Frank-Starling mechanism. Therefore, increased experimental conditions. However, the combination of multiple
Ca2+ cycling is correlated with increased ATP consumption. Ca2+ influx mechanisms with different K0.5 in isolated mitochon-
Accumulating evidence indicates that cardiac mitochondria take dria could also lead to this variation. It is interesting to note that
up [Ca2+]c during E-C coupling and that [Ca2+]m accumulation IMiCa density in the heart appears to be much smaller than in other
serves as a signal to ensure the balance of energy supply and tissues, and that mRNA for MICU1 is also particularly low in
demand (excitation-metabolism coupling).57 However, the kinet- the heart, raising an important question about the relative con-
ics of [Ca2+]m uptake is still a matter of debate. tribution of MCU and other Ca2+ influx mechanisms, especially
The composition and structure of Ca2+ microdomains located in cardiac mitochondria (see review3). Indeed, recent studies on
between SR and mitochondria in cardiomyocytes are still not molecular identification in MCU, which employed nonexcitable
clear. It is widely established that the ER Ca2+ release channel IP3 cells, show a 100-times slower and longer time course of mito-
receptor is involved in Ca2+-mediated crosstalk between ER and chondrial Ca2+ transients than those recorded from cardiomyo-
mitochondria in noncardiac cells (see Figure 7-1, C). However, cytes,8,9 strongly suggesting that cardiac mitochondria contain
Ca2+ release from SR in ventricular cardiomyocytes is exclusively not only MCU but also other Ca2+-influx mechanisms, which
via RyR2 but not the IP3 receptor. Most RyR2 were found along have greater Ca2+ sensitivity and faster Ca2+ uptake kinetics
the T-tubule side of the SR and beneath LCCs (see Figure during the heartbeats, such as mRyR1 and the rapid mode of
7-5, C). This unique localization tightly controls CICR in car- uptake (RaM). We previously showed that this Ca2+ uptake mode
diomyocytes.56 Therefore, in cardiomyocytes it is a reasonable exhibits a much greater magnitude and rate of [Ca2+]m than those
scenario that high local concentrations of Ca2+ after CICR in the observed in response to slow [Ca2+]c pulses and is more sensitive
subspace between T-tubule membrane and SR can subsequently to ryanodine (see review3). Moreover, the [Ca2+]c-mRyR1 activity
diffuse to nearby mitochondria to create microdomains of high relationship is bimodal (see review3). These unique properties
Ca2+ around the cardiac mitochondria. Another scenario is that a lead us to hypothesize that mRyR1 is an ideal candidate for
smaller number of RyR2 are located at the distal end of SR sequestering Ca2+ quickly and transiently during the heartbeat,
facing mitochondria, and they release Ca2+ at the microdomain whereas MCU would be an ideal candidate for sequestering Ca2+
between SR and mitochondria; thus mitochondrial Ca2+ channels slowly and steadily in the higher and plateau phases of Ca2+
such as MCU and RyR1 can take up Ca2+ into the matrix effi- pulses, such as sustained increases in the resting [Ca2+]c.
ciently (see Figure 7-5, C). Kim’s group recently showed the In cardiomyocytes, mitochondria are the main source of ROS
possibility of functional coupling of RyR2 and VDAC2 at the production and have crucial roles in ROS signaling in physiologi-
SR-mitochondrial contact site in HL-1 cardiac cells,58 but it is cal and pathophysiological conditions. Recent studies discovered
unclear whether adult ventricular myocytes also possess this mSOFs in adult cardiomyocytes, which is mediated through tran-
molecular architecture. Further studies using adult cardiomyo- sient openings of the mPTP triggered by a small increase in
cytes with genetic manipulation will be required to resolve the constitutive ROS production at the ETC. Therefore, in cardio-
detailed localization of the channel transporters at the myocytes, it seems that there is an “ROS-induced ROS release”
mechanism.62 To enhance understanding of this functional cou- of contractile function during reperfusion in human patients.66
pling and crosstalk between mPTP “flicker” activity and ETC- These basic and clinical observations indicate that delaying or
dependent superoxide production in cardiomyocytes, the question preventing irreversible mPTP opening can serve as a potential
moves to which superoxide production sites (complexes I and/or therapeutic strategy in treating ischemic heart disease. It is clear
III) are important for the ignition of transient mPTP opening in that an understanding of the identification of molecular compo-
intact cardiomyocytes (see Figure 7-1, B). O’Rourke’s group, nents, their detailed structure, and regulation of the mPTP is
using intact cardiomyocytes, reported that the majority of mito- crucial for designing novel and potent therapeutic drugs that are
chondrial ROS is obtained from complex III,41 whereas Dirksen’s more specifically targeted to the pore (see Figure 7-3).
group proposed that complex I may trigger mPTP opening fol-
lowed by superoxide flashes.16 It also remains to be determined Role of Mitochondrial Ion Channels/Transporters
which sites of ROS production in cardiac mitochondria are acti- in Cardiac Arrhythmias
vated under pathophysiological conditions, such as during Cardiac arrhythmias can cause sudden cardiac death, especially
hypoxia. The next question involves how the spatial and temporal during heart failure. The principal pathological processes under-
organization of the mitochondrial network in cardiomyocytes can lying arrhythmias involve heterogeneity of the cardiac action
regulate mSOF size and frequency (see Figure 7-5). It is known potential at the cellular level, which is commonly linked to ven-
that the molecular architecture of the T-tubules and the SR tricular arrhythmias.67 Ion channels in SL have received much
network regulates Ca2+ sparks from SR-RyR2, and that alteration attention for their ability to influence action potential duration.
of these structures under heart failure can modulate the spark Sarcolemmal ion channel mutations can cause alterations in the
frequency.60 Because mitochondrial structure and alignments are action potential duration (long and short QT syndrome) and
altered under the pathological heart, the state of the mitochon- appearance of early or delayed afterdepolarizations. Increasing
drial three-dimensional (3D) network and local ROS diffusion evidence suggests that altered cardiac ion homeostasis and struc-
might modulate mSOF size and frequency, as in the case of Ca2+ tural remodeling are highly associated with elevated ROS and
sparks. These questions will be resolved in the future by updating metabolic stress, implying the cardiac mitochondria play a key
3D fluorescence image acquirements/analysis systems and by role in the generation of arrhythmia (see also review68).
developing the computational cardiomyocyte model to include Mitochondrial oxidative stress contributes to a wide range of
mitochondrial network.63 perturbations in cardiac ion channel functions. ROS induces
functional and structural alteration of Nav1.5 by both transcrip-
tional and posttranslational mechanisms. Elevated ROS mediates
Mitochondrial Ion Channels/Transporters in Cardiac SERCA inhibition, enhances SR Ca2+ release from RyR2,
Dysfunction: Novel Therapeutic Targets enhances VLCC current, and increases sarcolemmal NCX activ-
ity to increase the intracellular Ca2+ (c) level. ROS-mediated
Role of Mitochondrial Ion Channels/Transporters in CaMKII activation stimulates hyperphosphorylation of RyR2,
Cardioprotection Against Ischemia-Reperfusion Injury resulting in SR Ca2+ leak. These ROS-mediated network changes
in the activities of ion channels/pumps are likely to contribute to
Transient episodes of nonlethal myocardial I/R provide a protec- the pathogenesis of arrhythmias. Conventionally designed antiar-
tive effect against myocardial injury in response to prolonged rhythmic drugs target ion channels at SL to modulate ion
lethal ischemia. This phenomenon is known as IPC. In cardio- currents. The observations mentioned earlier suggest new
myocytes, both mitoKATP and mitoKCa play key roles in cardio- potential therapeutic strategies that can be used to prevent
protection (II.C). Activation of the mitoKATP is augmented by arrhythmias by targeting ROS. However, general antioxidant
PKC or tyrosine kinases, whereas the mitoKCa is activated by strategies have not always been successful in antiarrhythmic
PKA, suggesting that these two channels are regulated by inde- therapies, in part because the physiological levels of ROS are
pendent mechanisms to protect the heart from I/R injury.34 essential for cell signaling, and the dominant source of ROS is
Opening of mitoKATP and mitoKCa depolarizes Δψm, which the mitochondria. Therefore, further improvements, such as
reduces the driving force for Ca2+ influx, thereby attenuating directly targeting the sites of ROS production in mitochondria
mitochondrial Ca2+ overload (see Figure 7-1). Consequently, pre- and using dosages that are within the therapeutic window, will be
vention of matrix Ca2+ overload inhibits mPTP opening and required.
protects against heart cell death. A recent clinical study demon-
strated a significant improvement in patient outcome resulting in
a reduction in major coronary events by antianginal therapy with
a mitoKATP-channel opener (nicorandil, see II.C.1) in patients Closing Remarks
with stable angina.64 These evidences suggest that the activators
of these K+ and/or the activators of upstream signaling for regula- Mitochondrial ion channels/transporters have historically stayed
tion of these channels may provide new therapeutic strategies for an important topic in cell biology, despite relatively slow progress
ischemic heart disease. in this area, including elucidation of their molecular identities.
OMM permeability also regulates the release of proapoptotic Use of multiple research tools, such as gene screening analysis,
factors from mitochondria. Therefore, it is a reasonable assump- genetic manipulation, and updated biochemical, pharmacologic,
tion that controlling OMM permeability is one of the potential cell-biological, and electrophysiological techniques, has led to
therapeutic strategies for protection from I/R injury. Growing recent ground-breaking discoveries in the molecular identities of
evidence shows that inhibiting VDAC interaction with Bcl-2 MICU1, MCU, and NCLX. Advances in cloning of mitochon-
families and/or enhancing the hexokinases-VDAC interaction drial Ca2+ channels/transporters will provide essential informa-
might provide cardioprotective effects.65 tion for studying (1) the regulatory mechanism underlying
As mentioned above, the mPTP has retained much attention mitochondrial Ca2+ uptake, such as posttranslational modifica-
from researchers as a strong candidate and as a potential thera- tions of these channels/transporters; (2) the design or discovery
peutic target against I/R injury and myocardial infarction. Many of more specific inhibitors/activators to each channel/transporter
studies report that in cells or animal models that use mPTP for the potential development of therapeutic drugs; and, further-
inhibition including CsA treatment, a cardioprotective effect more, (3) the molecular mechanisms underlying mitochondrial
such as reducing the infarct size is noted. Finally, recent clinical Ca2+-mediated human disease. However, we need to keep in mind
trials show that CsA can reduce infarct size and improve recovery that the molecular identities of most mitochondrial ion channels/
transporters, including several K+ channels and the mPTP, ROS generation, and cell survival in the heart. Discovering the
7
remain unknown. molecular identities and biophysical characteristics of cardiac
In conclusion, cardiac mitochondrial ion channels/transporters mitochondrial ion channels/transporters will provide us with new
are crucial in governing energy production, Ca2+ homeostasis, therapeutic strategies for treating human cardiac diseases.
22. Drago I, Pizzo P, Pozzan T: After half a century 42. Aon MA, Cortassa S, Akar FG, et al: From mito-
References mitochondrial calcium in- and efflux machineries chondrial dynamics to arrhythmias. Int J Biochem
reveal themselves. EMBO J 30:4119–4125, 2011. Cell Biol 41:1940–1948, 2009.
1. Glancy B, Balaban RS: Role of mitochondrial Ca2+ 23. Ryu SY, Beutner G, Dirksen RT, et al: Mitochon- 43. Shoshan-Barmatz V, De Pinto V, Zweckstetter M,
in the regulation of cellular energetics. Biochemis- drial ryanodine receptors and other mitochondrial et al: VDAC, a multi-functional mitochondrial
try 51:2959–2973, 2012. Ca2+ permeable channels. FEBS Lett 584:1948– protein regulating cell life and death. Mol Aspects
2. Gunter TE, Sheu SS: Characteristics and possible 1955, 2010. Med 31:227–285, 2010.
functions of mitochondrial Ca(2+) transport mech- 24. Beutner G, Sharma VK, Giovannucci DR, et al: 44. Messina A, Reina S, Guarino F, et al: VDAC iso-
anisms. Biochim Biophys Acta 1787:1291–1308, Identification of a ryanodine receptor in rat heart forms in mammals. Biochim Biophys Acta
2009. mitochondria. J Biol Chem 276:21482–21488, 1818:1466–1476, 2012.
3. O-Uchi J, Pan S, Sheu SS: Perspectives on: SGP 2001. 45. Rostovtseva TK, Bezrukov SM: VDAC regulation:
symposium on mitochondrial physiology and med- 25. Altschafl BA, Beutner G, Sharma VK, et al: The Role of cytosolic proteins and mitochondrial lipids.
icine: Molecular identities of mitochondrial Ca2+ mitochondrial ryanodine receptor in rat heart: A J Bioenerg Biomembr 40:163–170, 2008.
influx mechanism: Updated passwords for access- pharmaco-kinetic profile. Biochim Biophys Acta 46. Rostovtseva TK, Bezrukov SM: VDAC inhibition
ing mitochondrial Ca2+-linked health and disease. 1768:1784–1795, 2007. by tubulin and its physiological implications.
J Gen Physiol 139:435–443, 2012. 26. Ryu SY, Beutner G, Kinnally KW, et al: Single Biochim Biophys Acta 1818:1526–1535, 2012.
4. Pizzo P, Drago I, Filadi R, et al: Mitochondrial channel characterization of the mitochondrial 47. Shoshan-Barmatz V, Ben-Hail D: VDAC, a
Ca(2)(+) homeostasis: Mechanism, role, and tissue ryanodine receptor in heart mitoplasts. J Biol multi-functional mitochondrial protein as a phar-
specificities. Pflugers Arch 464:3–17, 2012. Chem 286:21324–21329, 2011. macological target. Mitochondrion 12:24–34,
5. Kirichok Y, Krapivinsky G, Clapham DE: The 27. Palty R, Silverman WF, Hershfinkel M, et al: 2012.
mitochondrial calcium uniporter is a highly selec- NCLX is an essential component of mitochondrial 48. Rostovtseva TK, Tan W, Colombini M: On the
tive ion channel. Nature 427:360–364, 2004. Na+/Ca2+ exchange. Proc Natl Acad Sci U S A role of VDAC in apoptosis: Fact and fiction. J Bio-
6. Jiang D, Zhao L, Clapham DE: Genome-wide 107:436–441, 2010. energ Biomembr 37:129–142, 2005.
RNAi screen identifies Letm1 as a mitochondrial 28. Wei AC, Liu T, Winslow RL, et al: Dynamics 49. De Stefani D, Bononi A, Romagnoli A, et al:
Ca2+/H+ antiporter. Science 326:144–147, 2009. of matrix-free Ca2+ in cardiac mitochondria: VDAC1 selectively transfers apoptotic Ca2+ signals
7. Perocchi F, Gohil VM, Girgis HS, et al: MICU1 Two components of Ca2+ uptake and role of phos- to mitochondria. Cell Death Differ 19:267–273,
encodes a mitochondrial EF hand protein required phate buffering. J Gen Physiol 139:465–478, 2012.
for Ca(2+) uptake. Nature 467:291–296, 2010. 2012. 50. Dejean LM, Ryu SY, Martinez-Caballero S, et al:
8. Baughman JM, Perocchi F, Girgis HS, et al: Inte- 29. Halestrap AP: What is the mitochondrial perme- MAC and Bcl-2 family proteins conspire in a
grative genomics identifies MCU as an essential ability transition pore? J Mol Cell Cardiol 46:821– deadly plot. Biochim Biophys Acta 1797:1231–
component of the mitochondrial calcium uni- 831, 2009. 1238, 2010.
porter. Nature 476:341–345, 2011. 30. Varanyuwatana P, Halestrap AP: The roles of phos- 51. Antonsson B, Montessuit S, Sanchez B, et al:
9. De Stefani D, Raffaello A, Teardo E, et al: A forty- phate and the phosphate carrier in the mitochon- Bax is present as a high molecular weight oligomer/
kilodalton protein of the inner membrane is the drial permeability transition pore. Mitochondrion complex in the mitochondrial membrane of
mitochondrial calcium uniporter. Nature 476:336– 12:120–125, 2012. apoptotic cells. J Biol Chem 276:11615–11623,
340, 2011. 31. von Stockum S, Basso E, Petronilli V, et al: Proper- 2001.
10. Kocherginsky N: Acidic lipids, H(+)-ATPases, and ties of Ca(2+) transport in mitochondria of Dro- 52. Dedkova EN, Blatter LA: Measuring mitochon-
mechanism of oxidative phosphorylation: Physico- sophila melanogaster. J Biol Chem 286:41163–41170, drial function in intact cardiac myocytes. J Mol
chemical ideas 30 years after P. Mitchell’s Nobel 2011. Cell Cardiol 52:48–61, 2012.
Prize award. Prog Biophys Mol Biol 99:20–41, 32. O’Rourke B: Mitochondrial ion channels. Annu 53. Hom J, Sheu SS: Morphological dynamics of
2009. Rev Physiol 69:19–49, 2007. mitochondria—a special emphasis on cardiac
11. Chicco AJ, Sparagna GC: Role of cardiolipin 33. Flagg TP, Enkvetchakul D, Koster JC, et al: muscle cells. J Mol Cell Cardiol 46:811–820,
alterations in mitochondrial dysfunction and Muscle KATP channels: Recent insights to energy 2009.
disease. Am J Physiol Cell Physiol 292:C33–C44, sensing and myoprotection. Physiol Rev 90:799– 54. Chen Y, Csordas G, Jowdy C, et al: Mitofusin
2007. 829, 2010. 2-containing mitochondrial-reticular microdo-
12. Soubannier V, McBride HM: Positioning mito- 34. Nishida H, Sato T, Ogura T, et al: New aspects for mains direct rapid cardiomyocyte bioenergetic
chondrial plasticity within cellular signaling cas- the treatment of cardiac diseases based on the responses via inter-organelle Ca2+ crosstalk. Circ
cades. Biochim Biophys Acta 1793:154–170, 2009. diversity of functional controls on cardiac muscles: Res 111:863–875, 2012.
13. Trachootham D, Lu W, Ogasawara MA, et al: Mitochondrial ion channels and cardioprotection. 55. Papanicolaou KN, Kikuchi R, Ngoh GA, et al:
Redox regulation of cell survival. Antioxid Redox J Pharmacol Sci 109:341–347, 2009. Mitofusins 1 and 2 are essential for postnatal meta-
Signal 10:1343–1374, 2008. 35. Foster DB, Ho AS, Rucker J, et al: Mitochondrial bolic remodeling in heart. Circ Res 111:1012–
14. Santos CX, Anilkumar N, Zhang M, et al: Redox ROMK channel is a molecular component of 1026, 2012.
signaling in cardiac myocytes. Free Radic Biol Med MitoKATP. Circ Res 111:446–454, 2012. 56. Bers DM: Calcium cycling and signaling in cardiac
50:777–793, 2011. 36. Xu W, Liu Y, Wang S, et al: Cytoprotective role of myocytes. Annu Rev Physiol 70:23–49, 2008.
15. Wang W, Fang H, Groom L, et al: Superoxide Ca2+-activated K+ channels in the cardiac inner 57. Liu T, O’Rourke B: Regulation of mitochondrial
flashes in single mitochondria. Cell 134:279–290, mitochondrial membrane. Science 298:1029–1033, Ca2+ and its effects on energetics and redox balance
2008. 2002. in normal and failing heart. J Bioenerg Biomembr
16. Wei L, Dirksen RT: Perspectives on: SGP sympo- 37. Shi Y, Jiang MT, Su J, et al: Mitochondrial big 41:127–132, 2009.
sium on mitochondrial physiology and medicine: conductance KCa channel and cardioprotection in 58. Min CK, Yeom DR, Lee KE, et al: Coupling of
Mitochondrial superoxide flashes: from discovery infant rabbit heart. J Cardiovasc Pharmacol ryanodine receptor 2 and voltage-dependent anion
to new controversies. J Gen Physiol 139:425–434, 50:497–502, 2007. channel 2 is essential for Ca2+ transfer from the
2012. 38. Sluse FE: Uncoupling proteins: Molecular, func- sarcoplasmic reticulum to the mitochondria in the
17. Schwarzlander M, Murphy MP, Duchen MR, et al: tional, regulatory, physiological and pathological heart. Biochem J 447:371–379, 2012.
Mitochondrial “flashes”: A radical concept aspects. Adv Exp Med Biol 942:137–156, 2012. 59. Dedkova EN, Blatter LA: Mitochondrial Ca2+ and
repHined. Trends Cell Biol 22:503–508, 2012. 39. Mailloux RJ, Harper ME: Uncoupling proteins and the heart. Cell Calcium 44:77–91, 2008.
18. Mughal W, Kirshenbaum LA: Cell death signalling the control of mitochondrial reactive oxygen 60. Prosser BL, Ward CW, Lederer WJ: Subcellular
mechanisms in heart failure. Exp Clin Cardiol species production. Free Radic Biol Med 51:1106– Ca2+ signaling in the heart: The role of ryanodine
16:102–108, 2011. 1115, 2011. receptor sensitivity. J Gen Physiol 136:135–142,
19. Estaquier J, Vallette F, Vayssiere JL, et al: The 40. Nowikovsky K, Schweyen RJ, Bernardi P: 2010.
mitochondrial pathways of apoptosis. Adv Exp Pathophysiology of mitochondrial volume 61. Shannon TR, Wang F, Puglisi J, et al: A mathemat-
Med Biol 942:157–183, 2012. homeostasis: Potassium transport and permeability ical treatment of integrated Ca dynamics within the
20. Wang C, Youle RJ: The role of mitochondria in transition. Biochim Biophys Acta 1787: 345–350, ventricular myocyte. Biophys J 87:3351–3371,
apoptosis. Annu Rev Genet 43:95–118, 2009. 2009. 2004.
21. Kroemer G, Galluzzi L, Brenner C: Mitochondrial 41. O’Rourke B, Cortassa S, Aon MA: Mitochondrial 62. Zorov DB, Juhaszova M, Sollott SJ: Mitochondrial
membrane permeabilization in cell death. Physiol ion channels: Gatekeepers of life and death. Physi- ROS-induced ROS release: An update and review.
Rev 87:99–163, 2007. ology (Bethesda) 20:303–315, 2005. Biochim Biophys Acta 1757:509–517, 2006.
63. Zhou L, Aon MA, Almas T, et al: A reaction- 65. Camara AK, Bienengraeber M, Stowe DF: Mito- 67. Goldenberg I, Zareba W, Moss AJ: Long QT syn-
diffusion model of ROS-induced ROS release in a chondrial approaches to protect against cardiac drome. Curr Probl Cardiol 33:629–694, 2008.
mitochondrial network. PLoS Comput Biol ischemia and reperfusion injury. Front Physiol 68. Jeong EM, Liu M, Sturdy M, et al: Metabolic
6:e1000657, 2010. 2:13, 2011. stress, reactive oxygen species, and arrhythmia.
64. IONA Study Group: Effect of nicorandil on coro- 66. Gomez L, Li B, Mewton N, et al: Inhibition of J Mol Cell Cardiol 52:454–463, 2012.
nary events in patients with stable angina: The mitochondrial permeability transition pore
Impact Of Nicorandil in Angina (IONA) ran- opening: Translation to patients. Cardiovasc Res
domised trial. Lancet 359:1269–1275, 2002. 83:226–233, 2009.
Molecular Organization, Gating, and
Function of Gap Junction Channels 8
Feliksas F. Bukauskas
Most likely, the initial step in this process is the formation of

CHAPTER OUTLINE
hemichannel plaques (hemi-plaques) in the plasma membrane,6,7
Connexins Expression and Oligomerization 85 which should stabilize the hemichannels’ stochastic lateral diffu-
sion in the plasma membrane and create an outward-directed
De novo Formation of Gap Junction Channels 85
curvature that should reduce the distance between hemichannels
Voltage and Chemical Gating of Connexin-Based in apposing cells.8 Theoretical studies show that when five or
Channels 86 more proteins aggregate, the cluster becomes stable despite their
pairwise repulsions.8 Subsequent steps likely include a close appo-
Hemichannels: Function and Physiologic Relevance 87 sition between cells, possibly mediated by adhesion molecules
Connexins Expression Pattern and Propagation of and overlapping of hemi-plaques that could lead to formation of
Excitation in the Heart 87 precursors of GJ channels termed formation plaques (FPs), which
gradually transform into junctional plaques (JPs).9 Alignment and
Cell-Cell Coupling and Electrical Anisotropy docking of apposed hemichannels comes thereafter, followed by
in the Heart. 88 the formation of a high-resistance seal to isolate the nascent
Alterations of Cell-Cell Coupling in Cx-Deficient channel pore from the extracellular space, ending with the
Animals and Disease-Related Mutations 91 channel pore opening. Docking of individual hemichannels fur-
thermore stabilizes FPs and reduces the distance between still
undocked hemichannels, which may explain an acceleration of de
novo formation kinetics after functional appearance of the first
channels.10 De novo formation of GJ channels, their clustering
Connexins Expression and Oligomerization into JPs, as well as GJ channels’ turnover and their gating was
shown to be regulated by different intracellular pathways that
Connexins (Cxs), a large family of homologous membrane pro- involve cAMP, G proteins, and phosphorylation.11
teins, form gap junction (GJ) channels that provide a direct Significant progress has been made in understanding Cx traf-
pathway for electrical and metabolic signaling between cells. Gap ficking, JP formation, and internalization using cells expressing
junctional communication is critically important in the spread of Cxs fused with color variants of green fluorescent protein (GFP)
excitation in the heart, communication between neurons and glia, or containing a tetracysteine tag.12,13 JPs serve as “crystallization
metabolic exchange between cells in the lens and other tissues centers,” attracting laterally moving hemichannels and catalyzing
lacking blood circulation, organ formation during development, their docking due to close apposition of hemichannels in the area
and regulation of cell proliferation.1 Each GJ channel is com- of JPs. By combining fluorescence imaging of Cx-GFPs with dual
posed of two hemichannels (connexins). Oligomerization of six whole-cell voltage clamp recording, it was shown that only cell
Cxs into hemichannels starts in the endoplasmic reticulum and pairs that exhibit a JP are electrically coupled.14 Furthermore, it
is completed in the Golgi, from where vesicles approximately 100 was shown that only a small fraction of channels assembled in a
to 150 nm in diameter containing hemichannels travel to fuse JP are functional at a given time and this ratio varies from approx-
with the plasma membrane. Cxs are predicted to have four alpha- imately 0.01 for Cx45 and Cx57 to 0.1 for Cx43.15,16 This fraction
helical transmembrane domains (M1 to M4), intracellular N- and can be modulated in the time scale of a few minutes by pH, long-
C-termini (NT and CT), two extracellular loops (E1 and E2), chain alkanols, [Mg2+]i, and many other chemical factors.15,17 This
and a cytoplasmic loop (CL).2 The density map of the crystal indicates that this type of modulation does not involve an assem-
structure of human Cx26 at 3.5 Å resolution revealed that M1 bly of new GJs or their turnover.
and EI line the channel pore, which is narrowed by six NT Cell-cell communication can be organized through homo-
domains residing in the channel vestibule. These structural data typic (same Cx isotype in both hemichannels), heterotypic (2 Cx
were supported by functional studies using a cysteine scanning isotypes form GJ channels, but each hemichannel is assembled
approach3 demonstrating that NT, M1, and E1 domains are from 1 isotype) and heteromeric (different Cx isotypes at least in
directly involved in defining GJ channel unitary conductance, one of the hemichannels) channels that vary in conductance,
gating properties and permselectivity.4 Thus far, at least 21 Cx permselectivity, and gating properties. Formation of heterotypic
genes have been identified in humans. and heteromeric GJ channels can occur in the heart, particularly
in the conduction system, where cardiomyocytes co-express
several Cx isoforms. Most heterotypic junctions that can be
formed in the heart exhibit asymmetric voltage gating and
De novo Formation of Gap Junction Channels rectification of the current voltage (I-V) relation.18,19 Asymmetric
junctional conductance to transjunctional voltage (gj-Vj) depen-
GJ channel de novo formation was examined by manipulating dence can result in impeded signal transfer in one direction and
two separate cells into contact while monitoring electrical cell- facilitated transfer in the opposite direction, as it was demon-
cell coupling. This approach has been used to study formation of strated for Cx43/Cx45, Cx31/Cx45, and Cx40/Cx45.5,20 This may
GJs between cells of Xenopus blastulae and myoballs, rat cardio- facilitate an initiation of one-directional block, which is critical
myocytes, insect cells, and HeLa cells transfected with Cx40.5 in reentry formation.
85
(see Figure 8-1, C). Figure 8-1, D, shows an example of fast and
Voltage and Chemical Gating of Connexin- slow gate operation during acidification with carbon dioxide
Based Channels (CO2), which gradually reduced the number of operating chan-
nels. When only one channel was left to operate, the fast gate
The existence of multiple Cxs raises the questions of how they was governing fast transitions between open and residual states,
differ and how they interact. Unitary conductances of GJs formed finally being closed by the slow gate with transitions lasting
of Cx isoforms range from approximately 10 pS (e.g., mCx30.2) approximately 10 ms. The opposite sequence of events was
to 300 pS (e.g., Cx37) and channels vary in permselectivity from observed during washout from CO2. Two distinct gating mecha-
being nonselective to preferentially selective for cations or nisms were also demonstrated in Cx-based unapposed/
anions.1,21 A property that appears to be common to GJs formed nonjunctional hemichannels (uHC).25
by any Cx isoform is sensitivity of gj-Vj. Figure 8-1, A shows a Mutational studies revealed that the gating polarity of the
decay of gj until a steady-state level (gj,ss) reached (as indicated by fast gating mechanism is governed by charged residues in the
the arrow) in response to Vj step voltage. The symmetric reduc- N-terminal domain and that this polarity could be reversed inde-
tion in gj,ss with positive or negative Vj has been explained by pendently from the slow gating mechanism. Modifications of
having identical Vj-sensitive gate(s) in each apposed/junctional Cx43, including deletion of the CT domain26 or attachment of
hemichannel (aHC) of the GJ channel.23 Normalized gj,ss-Vj aequorin or enhanced green fluorescent protein (EGFP) to CT,
dependencies show big differences in sensitivity to Vj among four selectively abolish fast gating to the residual state.5 These data
principal Cxs of cardiomyocytes (see Figure 8-1, B). A common demonstrate that in each hemichannel there are two molecularly
feature of Vj-gating is that gj,ss does not decline to zero with distinct gating mechanisms that require further study in deter-
increasing Vj, but instead reaches a residual conductance termed mining Cx domains functioning as gating and sensorial elements
gmin. Single-channel studies have shown that gmin is due, at least for each gate. This hypothesis has received support from studies
in part, to incomplete closure of the GJ channel by Vj, which performed in channels formed of Cx30, Cx31, Cx32, Cx37, Cx40,
causes channels to close to a subconductance (residual) state with Cx43, Cx45, and Cx57,1,5 suggesting that fast and slow gates are
fast gating transitions (~1 ms or less).24 It was shown that Vj as likely to be common to all Cx isoforms.
well as chemical uncouplers can also induce gating transitions to Gating properties of GJ channels can be described using the
the fully closed state and that these transitions are slow, approxi- Boltzmann function,23 assuming that GJ channels have two states,
mately 10 ms. Gating to different levels via fast and slow gating open and fully closed, like most ionic channels, and that each
transitions led to the suggestion that there are two distinct Vj- hemichannel gates independently. The accumulation of data
sensitive gates, termed fast and slow or loop.5 The fast gate closes demonstrating gating to the substates and two distinct gating
channels to the residual state, and the slow gate closes them fully mechanisms stimulated the development of gating models that
0 Fast gate Slow gate

Vi, mV
–80
20
gi, nS
10
0
0 20 40 60
A Time, s
C
gj,ss, norm. 100% CO2 Washout
1.0 mCx30.2
Open
Cx40
Residual
closed
0.5
γopenx2
Cx43
γopen
Cx45
5 pA
–100 –50 0 50 100 γresidual

Vj, mV 1s
B D Fully closed
Figure 8-1. Gating of GJ channels. A, Typically, junctional conductance (gj) measured between cell pairs decays in response to transjunctional voltage (Vj), reaching some
steady-state level (gj,ss) (arrow). B, Normalized gj,ss dependencies on Vj show big differences in sensitivity to Vj among four principal Cxs of cardiomyocytes. C, A schematic
of the GJ channel containing fast and slow gating mechanisms. The fast gate (orange) closes the channel partially, whereas the slow gate (blue) closes the channel fully.
D, The effect of CO2 on voltage gating at the single-channel level in a cell pair of fibroblasts expressing Cx43. Exposure to CO2 caused full uncoupling. Ij was monitored at
Vj = 55 mV just before full uncoupling and at the beginning of CO2 washout. Channels exhibited two types of Ij transition: (1) between open and residual state (~90 pS),
with a transition time of approximately 1 ms (red arrows); and (2) between open and fully-closed states (~120 pS), with a transition time of approximately 10 ms or more
(blue arrows). The signals in the insets (sampled at 1-ms interval) illustrate that the last channel closes with a transition time of approximately 10 ms and the first channel
opens with a transition time of approximately 19 ms. The slow opening of the first channel during washout was followed by fast flickering between the open and residual
states. When two operating channels were in the residual state, gj equals the sum of two γres (dashed lines).
(Modified from Bukauskas FF, Peracchia C: Two distinct gating mechanisms in gap junction channels: Co2-sensitive and voltage-sensitive. Biophys J 72:2137–2142, 1997.)
Molecular Organization, Gating, and Function of Gap Junction Channels 87
more intimately describe GJ channel gating.27,28 Recently, a sto- preconditioning leads to opening of Cx43 hemichannels in the
8
chastic 16-state model (S16SM) of Vj-gating was developed, in inner mitochondrial membrane, with subsequent loss of ATP
which each aHC contained both fast and slow gating mecha- from the mitochondrial matrix. This activates ATP-dependent K+
nisms.24 The model can be used to simulate Vj-gating in homo- channels that appear to be central to the protective effect.39
typic and heterotypic junctions. A S16SM allows for simulation It has been shown that among cardiac Cxs, mCx30.2,29 Cx43,35
of Ij dynamics as well as the gj,ss-Vj plot of GJs depending and Cx4540 form functional uHCs, whereas Cx40 does not.41
on individual gating parameters of four gates. The model Typically, uHCs tend to open at higher positivity on the cyto-
includes more than 12 parameters characterizing certain gating plasmic side. Presumably, open channel probability of uHCs
properties of GJs, and to evaluate them manually from experi- increases during action potentials (APs) and to a greater extent
mental gj,ss-Vj dependencies would resemble the search for “the during tachycardia when cells are depolarized for a longer time.
needle in the haystack.” To automate this fitting process, we have Opening of hemichannels in nodal cells would depolarize them
adopted the Global Coordinate Optimization (GCO) algo- and locally increase the extracellular concentration of K+ ions.
rithms,24 which are based on the Bayesian approach to filter the Depolarization could inactivate inward currents driving excit-
stochastic component and smooth small local minima while ability and contribute to a reduction in conduction velocity. This
searching for the global minimum. GCO of experimental data hypothesis raises a new mechanism explaining a long atrioven-
can be performed using the online version of the algorithm at tricular (AV) delay as a result of mCx30.2 hemichannels func-
http://connexons.aecom.yu.edu. tion.42 This corroborates with a reduction in PQ-interval in
All GJ channels also display chemical gating, which shares mCx30.2 knockout mice.43
several features with the slow voltage sensitive gate, (i.e., it Another class of membrane proteins, pannexins (Panxs),
reduces gj to zero and closes channels fully.)1,5 We assume that which also form uHCs but not GJ channels, are involved in
asynchrony of conformational changes in each of six Cx “sub- paracrine signaling and mechanosensitivity.44 It was shown
gates” of the slow gate leads to multistep gating transitions lasting recently that Panx1 constitutes 300 pS uHCs in cardiomyocytes,
approximately 10 ms or more. This may be caused by a low level suggesting that their opening can depolarize the plasma mem-
of cooperativity between these “slow subgates.” Similar slow brane and trigger premature APs that may lead to arrhythmo-
transitions were observed at the initial stages of channel pore genic activities.45
opening during de novo channel formation and during gating
induced by transmembrane voltage (Vm) or chemical uncouplers,
as illustrated in Figure 8-1, D. Although fast gating to the residual
state is induced exclusively by Vj, the slow gating mechanism can Connexins Expression Pattern and
be triggered by Vj, Vm, and chemical factors. Interestingly, Propagation of Excitation in the Heart
uncoupling induced by alkanols, Ca2+, and H+ was shown to be
partially reversible by changes in Vm or Vj.5 These data suggest To date, four connexins—mCx30.2, Cx40, Cx43, and Cx45—are
that chemical- and voltage-sensitive gating mechanisms interact known to be expressed by cardiomyocytes and form GJ channels,
and may share the common gating element of the slow gate, mediating the spread of excitation in the heart, as well as direct
which is triggered by sensorial elements specific for voltage and intercellular metabolic communication42; human ortholog of
different chemical agents. mCx30.2 is Cx31.9.19 Expression patterns of all four cardiac Cxs
are shown in Figure 8-2. In the mammalian heart, the cardiac
impulse is initiated in the spontaneously active pacemaker cells
of the sinoatrial (SA) node, in which mCx30.2 and Cx45 GJs
Hemichannels: Function and integrate thousands of pacemaker cells with various intrinsic fre-
Physiologic Relevance quencies of excitation into one functional unit. These Cxs form
Surface expression of uHCs has been demonstrated in several cell

lines and primary cultures by electron microscopic, biochemical,
and electrophysiologic methods.29,30 It has long been thought that
hemichannels remain closed until docking with hemichannels SA node, ~0.05 m/s
from an apposed cell because ions and metabolites can leak the Cx30.2 ≈ Cx45
cell through open hemichannels. Furthermore, Cx43 uHCs have
Atria, ~0.5–1 m/s
been implicated in diverse roles in cell physiology and patho- Cx40 ≈ Cx43
physiology, including volume regulation; efflux of glutamate,
NAD+, cAMP, IP3, and ATP acceleration of astrocytes and car- AV node, ~0.05 m/s
diomyocyte death during metabolic inhibition, and transduction Cx30.2 ≈ Cx45 > Cx40
of extracellular signals regulating apoptosis development.31,32 The
opening of hemichannels can be enhanced by cell polarization to AV bundle, ~2 m/s
positive potentials and reduction of Mg2+ and Ca2+ in the extracel- Cx40 > Cx45 > Cx30.2
lular milieu.33 Conversely, closure of hemichannels can be induced
Bundle branches, ~2 m/s
by low intracellular pH, 18-α-glycyrrhetinic acid, trivalent Cx40 > Cx45 > Cx30.2
cations (La3+ or Gd3+), and some chloride channel blockers.31,34
Typically, at the single-channel level, the I-V relation of uHCs Purkinje fibers, ~2 m/s
rectifies and their unitary conductances are approximately twice Cx40 ≈ Cx43
the conductances of the GJ channel formed of the same Cxs.35
Working myocytes, ~0.5 m/s
Lately, Cx43 hexamers/hemichannels were shown in purified Cx43
mitochondrial preparations of the mouse myocardium, which
may contribute to mitochondrial K+ uptake.36 More detailed
Figure 8-2. Schematic of the heart with indications showing connexin expression
studies revealed that Cx43 is expressed in subsarcolemmal, patterns and conduction velocities in different regions of the heart.
but not interfibrillar, mitochondria, and it was proposed that
Cx43 plays a critical role in mediating the cardioprotective func- (Modified from Netter: Atlas der Anatomie des Menschen, ed 4, New York, 1993, Ciba
tion of ischemic preconditioning.37,38 It is hypothesized that Geigy Corp.)
homotypic GJ channels with conductances of 9 pS and 32 pS, Na+-channels, cytoarchitecture of fibers,49 and expression of
respectively, as well as mCx30.2/Cx45 heterotypic GJs with the mCx30.2 and Cx45, both of which exhibit low single-channel
single-channel conductance of approximately 17 pS,5,19 or even conductance, leading to relatively high resistivity of intracellular
heteromeric GJs, which can exhibit a variety of single-channel milieu.50,51 It was shown that only approximately 1% of Cx45 GJ
conductances in a range of about 9 pS to 32 pS.46 Electrical channels are open at a given time under normal conditions.15
coupling in the transitional region between the SA node and the Furthermore, mCx30.2 exhibits the lowest single-channel con-
atrium should not be too high, thus preserving functional identity ductance among all members of the Cx family and forms hetero-
of pacemaker cells, but at the same time allowing for signal typic as well as heteromeric channels with all other cardiac Cxs.
transfer. Figure 8-3 shows images of cell pairs formed of Resulting channels also exhibit low unitary conductance (~17 to
HeLaCx30.2-EGFP (red) with HeLa cells expressing Cx40 18 pS), which raises the possibility that mCx30.2 demonstrates a
tagged with cyan florescent protein (CFP) (green, part A) and dominant-negative effect upon junctional conductance between
Cx43-CFP (part B). Both heterotypic junctions exhibited exten- nodal cells. This effect may also contribute to low conduction
sive formation of JPs (arrows) and Vj-gating asymmetry typical velocity in wild type mice and enhanced AV conduction in
for most of heterotypic junctions.19 mCx30.2-deficient mice.43
Excitation transferred through the crista terminalis, pectinate After passing the transitional N-H region and entering the
muscles, and Bachmann’s bundle spreads with a velocity of 0.5 to bundle of His, the propagation of excitation quickly accelerates to
1 m/s (in the human heart) to the right and left atria, where Cx40 approximately 2 m/s and remains at this level in the subendocar-
and Cx43 are most abundantly expressed,47 both exhibiting high dial network of Purkinje fibers, resulting in coordinated contrac-
single-channel conductances (180 pS and 115 pS, respectively),5,48 tion of both ventricles. Expression of Cx40 increases in bundle of
all of which is consistent with the fast spread of excitation and His and right and left bundle branches, where it is co-expressed
almost synchronous contraction of both atria. Presumably, cell- with Cx43. There is also an expression of mCx30.2 and Cx45 at
cell coupling in the atria is maintained by homotypic Cx40 and low levels, which becomes undetectable at the terminals of the
Cx43 GJ channels because Cx40 and Cx43 are highly incompat- ventricular conductive system. The bundle of His and right and
ible in forming heterotypic GJs.20 left bundle branches are isolated electrically from surrounding
The excitation wave enters the AV node through the transi- working myocardium, preserving the high speed of propagation.
tional A-N zone, and its velocity of propagation gradually Cx40, because of its inability to form heterotypic GJs with Cx43,
decreases to approximately 0.05 m/s in the central region called the principal Cx of ventricular cardiomyocytes, may provide this
the compact node or N region. In a normal heart, the AV node is isolation during development. From the bundle branches, excita-
the sole connection between the atria and ventricles, and the tion propagates through false tendons and the endocardial layer
delay of AV conduction is necessary for the sequential contraction of Purkinje cells, where Cx40 and Cx43 are primarily expressed.47
of both, which is important for optimal hemodynamics. Both the Excitation to working cardiomyocytes, preferentially expressing
relatively long refractory period of the AV-nodal cells and the Cx43, is transmitted through discrete Purkinje fiber-muscle
dependence of the conduction velocity on the frequency of exci- (P-M) junctions52 distributed discretely on the endocardial surface
tation limit the number of impulses within a given time period— with an approximately 1 mm distance between them.53 This
termed the Wenckebach point—that can be transmitted to the pattern allows for Purkinje fibers to maintain their electrical iden-
ventricles. Thus, the AV node protects the ventricular myocar- tity and, therefore, high conduction velocity. Multiple P-M junc-
dium by reducing the frequency of APs transmitted to the bundle tions preserve a satisfactorily high safety factor for signal transfer
of His during atrial fibrillation. Slow AV-nodal conduction under to working myocardium, and the only price “paid” for this struc-
pathophysiologic conditions can also lead to the generation of tural organization is a P-M delay of approximately 4 to 7 ms,
supraventricular tachycardia involving reentry loops. In the which slightly prolongs the PQ interval.
AV-nodal cells, mCx30.2 and Cx45 are primarily expressed,
both exhibiting relatively low single-channel conductances. Fur-
thermore, Cx40 is expressed in the N region, but to a lesser
extent.42 Cell-Cell Coupling and Electrical
Slow conduction in the AV node may be a result of mCx30.2 Anisotropy in the Heart
uHCs that tend to depolarize the resting membrane potential,42
as well as reduced excitability because of lower numbers of Typically, the input resistance (Rin) and the length constant of
electrotonic potential decay (λ; the distance at which the ampli-
tude of electrotonic potential evoked by intracellularly injected
current is reduced e-fold) are measured in isolated heart tissue
preparations to determine electrical intercellular communication.
HeLaCx43-CFP ⋅
For cable-like systems λ = (Rm b/ρi)1/2, where Rm is a specific
resistance of the plasma membrane, ρi is a specific resistance of
the intracellular medium integrating averaged resistances of cyto-
plasm and gap junctions, and b is a ratio of the volume of the cell
HeLaCx40-CFP to an area of the plasma membrane. This applies also to two-
dimensional (2-D) structures when the source of excitation is a
line and three-dimensional (3-D) structures when the source of
excitation is a 2-D plane. To determine λ, the electrotonic
responses were recorded via microelectrodes at different dis-
HeLa tances from intracellular current sources along the general direc-
mCx30.2-EGFP tions of cardiac fibers (λX) and in their perpendicular direction
HeLa
(λY). Microelectrode or suction-electrode (~200 to 300-µm tip
mCx30.2-EGFP
A 10 µm B 10 µm diameter), perfusable with a depolarizing KCl solution, were
used.54,55 Rin is typically measured using a single microelectrode
Figure 8-3. Formation of heterotypic GJs visible as multiple junctional plaques or a double-barreled electrode, where one barrel is used for
(yellow) indicated by arrows and located in between cells expressing mCx30.2-EGFP current injection (I) and the second records electrotonic potential
(red) and those expressing Cx40-CFP (A) or Cx43-CFP (B) (both green). (V); Rin = V/I.
Measurements of Rin, λX, and λY allow one to evaluate that “the leading (true) pacemaker cells are roughly spindle-
8
the ρi,X, ρi,Y, and Rm by using models describing passive electrical shaped with a maximum length of 20 to 30 µm and a diameter
properties of syncytial structures,56-58 where ρi,X and ρi,Y are spe- of less than 8 µm, with the long axis of the cells roughly parallel
cific resistances of the intracellular medium along and across to the crista terminalis.” The findings that “the gap junctions are
fibers. Depending on the complexity, used models can be dif- less developed in size and number than in atrial cells” correlate
ferentiated based on a dimension (cable, 2-D sheet, or 3-D with significantly higher Rin and smaller λs in the central region
volume) or structure of the syncytium (continuous with averaged of the SA node than in the crista terminalis and trabeculae. Values
electrical parameters, compartmental, or a combination of the of λX and λY increased, approaching the intercellular septum,
two). In general, continuous models are more applicable to struc- whereas anisotropy decreased, which may be assumed to be
tures with well-coupled cells (e.g., heart, liver), whereas discrete caused by the more dispersed orientation of cardiomyocytes. In
models are more applicable to poorly coupled cells exemplified summary, based on λX, λY, and Rin measurements and using a 3-D
by neuronal networks. We formulated a continuous 2-D model56 model it was shown that in the central region of the SA node, ρX
resembling the flat-cell model,59 a 3-D model of infinite size,57
and 3-D model of finite thickness,60 which most closely describes
⋅
≈ 600 and ρY ≈ 3000 Ohm cm.60 Thus, ρX in the SA node is
approximately 3.5-fold bigger than ρX in the trabeculae and crista
isolated tissue preparations and created methods known as terminalis, which is approximately 150 to 200 Ohm cm and close ⋅
“mirror reflection” or “virtual images.”58,60,61
Many organs and organ systems, including the heart, exhibit
to a specific resistance of the cytoplasm (~120 Ohm cm).67 ⋅
electrical anisotropy determined as a ratio, λX/λY, which is >1.
Furthermore, λX/λY = (ρi,Y/ρi,X)1/2, (i.e., an anisotropy of intracel-
0.5 mm
Vena cava
lular resistivity) is higher than that estimated from λs as a square superior
degree. By measuring X-Y distributions of electrotonic potentials λ
on the endocardial surface of the atrium while current was
terminalis
0.5 mm
Crista
injected from the epicardial surface, it was shown that λX > λY =
λZ, where λZ is perpendicular to endocardial/epicardial surfaces
(i.e., electrically myocardium exhibits a cylindrical anisotropy).60
In general, every local myocardium domain in which the direc- Trabecula
tion of cardiomyocytes remains similarly aligned exhibits a cylin-
drical electrical anisotropy. The magnitude of the anisotropy is
determined by the ellipsoid/cylinderlike shape of the cardiomyo-
Interatrial septum
cytes and the ratio between end-to-end and lateral gap junctions. Rin
Most proposed reentry mechanisms are related mainly to hetero-
geneities of repolarization, which break the wave of premature
excitation necessary to initiate reentry. However, electrical
anisotropy can explain reentry initiation by excitation during the
vulnerable period of the heart, which is homogeneous with all
other parameters.62,63 Interestingly, electrical anisotropy is a
property of all myocardial structural domains under normal con-
ditions, suggesting that the heart is intended to be arrhythmo-
genic, whereas realization of reentry through electrical anisotropy
requires premature excitations, which is acquired from
pathology. Vena cava
inferior
A
Passive Electrical Properties of the Sinoatrial Node
The first attempt to measure λ in the rabbit SA node was per-
formed by Bonke,64 assuming that the SA node constitutes an 900
isotropic medium. More systematic studies of intercellular com-
Rin, kOhm
munication of different SA node regions and surrounding areas 600

were performed by measuring λX, λY, and Rin.60,65 Intracellular
recordings of action potentials (APs) from different regions of
the SA node allowed the determination of the region of “true” 300
pacemakers or the central region where excitation starts earliest.
Figure 8-4, B shows averaged values of Rin measured along the 0
blue dashed line crossing the central region of the SA node. Rin 0 3 6
in atrial trabeculae and crista terminalis was approximately
350 kOhm, which increased approaching the central region of B Distance, mm
the SA node, reaching a maximum of 1050 kOhm, and was lower Figure 8-4. Passive electrical properties of the sinoatrial (SA) node. A, Schematic
in the intracellular septum, approximately 450 kOhm. Figure of the right atrium of the rabbit. An averaged data of λs measured along and across
8-4, A shows a summary of averaged data of λX and λY in the form the crista terminalis are shown in the form of crosses. Values of λ were evaluated
of crosses. Thus, all regions exhibit electrical anisotropy, which from experimental measurements (λexp) using a 3-D model of passive electrical
is highest in crista terminalis and atrial trabeculae. Electrical properties with finite thickness. All regions exhibited electrical anisotropy, which
was highest in the crista terminalis and atrial trabeculae. Electrical anisotropy is
anisotropy gradually reduces, approaching the central region of
reduced in the central region of the SA node. The calibration is at the top right,
the SA node, but general orientation of anisotropy remained the where lengths of white lines correspond to λ = 0.5 mm. B, Averaged values of Rin
same as in the crista terminalis. Anisotropy of λX and λY (529 vs. measured along the blue dashed line starting from the trabeculae and crossing the
306 µm) was also reported in another study,55 in which measure- central region of the SA node. Rin reached maxima in the central region of the SA
ments of λs were averaged for the entire SA node. These data node (1050 kOhm) at the value that is approximately twofold higher than Rin mea-
correlate with ultrastructural studies of the SA node,66 showing sured in the trabecular structure, crista terminalis, and interatrial septum.
revealed that in central regions of the AV node, Rm =

Passive Electrical Properties of the
Atrioventricular Node
⋅
1300 Ohm cm2, ρi,X = ~600, and ρi,Y = ~1200 Ohm cm; AV-nodal ⋅
cells were assumed to be 8 × 12-µm ellipsoids.69 Thus, ρi,X in the
central region of the AV node is similar to that of the SA node,
Division of the AV-nodal region into functionally distinct domains whereas ρi,Y is approximately twofold smaller. In both structures,
is based on anatomy, AP shape, and delays in the spread of excita- anisotropy of axial resistance is smaller than that in all other
tion.68 Mapping of the AV node by using intracellular microelec- structural domains of the heart. In another study,51 it was reported
trode recordings from different AV-nodal regions allowed us to that in the AV node, λ = 430 µm (not specifying an anisotropy)
differentiate AN, N, and NH regions of the AV node based on and Rin = 880 kOhm, which are comparable with our data.
the shape of APs and delays from the stimulus applied to the crista
terminalis (Figure 8-5, A).50 Figure 8-5, B shows that, on average,
Rin in all three AV-nodal domains is higher than in the crista Passive Electrical Properties of Conductive and
terminalis and the bundle of His. Working Systems of Ventricles and Signal
Averaged values of λ measured in different domains of the AV Transfer in P-M Contacts
node are shown as crosses, where half of their axes are propor-
tional to values of λ in corresponding directions (see Figure 8-5, After passing through the AV node, excitation spreads at high
C). The maximal anisotropy is recorded in the crista terminalis speed (~2 m/s) through the bundle of His, left and right bundle
and the bundles of His, and it is significantly less expressed in the branches, multiple false tendons, and it finally almost instanta-
AV node, mainly because of the reduction in λX. Data analysis neously excites the thin layer of Purkinje fibers covering the
endocardial surface of the ventricles. Figure 8-6 shows isopoten-
tial lines of electrotonic potential in a false tendon, the subendo-
cardial surface Purkinje network, and the working myocardium
from the epicardial surface measured in response to intracellular
current applied through a suction electrode (dashed circle). Iso-
potential lines are one-quarter from those measured in the posi-
Cr tion indicated by the filled dot.
ist
a
te Coronary Values of experimentally measured λX,exp and λY,exp depend on
rm the size of the current sources and dimensionality of the tissue
in sinus
al
is (e.g., a false tendon can be approximated as a cable, whereas
His bundle the subendocardial conductive system and working myocardial
surface can be considered 2-D and 3-D structures, respectively).
Consequently, different models of passive electrical properties
should be used to evaluate the true values of λX and λY, which are
A typically bigger than λX,exp and λY,exp; at a smaller diameter of
current source electrode values, of λX,exp and λY,exp are smaller.
Input resistance, Rin (kOhm) Furthermore, the experimentally measured anisotropy of electro-
tonic potentials is smaller than the “true” anisotropy (i.e., λX/λY
> λX,exp/λY,exp). Figure 8-6 shows that the false tendon exhibits the
False tendon
1/4
1
B Tricuspidal valve 4 mm
Endocardial surface,
Length constant, λ
0.5 mm
network of Purkinje fibers

Cr
ist λ Suction electrode
a perfusable with KCl
te Coronary 0.5 mm
rm
in sinus 1/4
al
is
His bundle
1
1/4
C Ventricles
Figure 8-5. Passive electrical properties of the atrioventricular (AV) node. A, Sche- Epicardial surface,
matic of the right atrium of the rabbit involving an area of the AV node. Different working cardiomyocytes
regions of the AV node demonstrate differences in APs’ shape and stepwise 0.5 mm
responses to intracellular current steps of 50 nA applied through the recording
electrode. B, Averaged Rin values (in kOhm) attached to different functional regions
in the surrounding area of the AV node are shown. C, Averaged data of λs are Figure 8-6. Isopotential lines of electrotonic potential in the false tendon, endo-
shown in the form of crosses. All regions exhibited electrical anisotropy, which was cardial surface of Purkinje network, and working myocardium recorded from the
highest in the crista terminalis and the bundle of His, but smallest in the central epicardial surface in response to intracellular current applied through the suction
region of the AV node. The inset in the right-top corner shows the calibration cross electrode (dashed circle) perfusable with KCl. Isopotential lines are at the level of
where the lengths of white lines correspond to λ = 0.5 mm. one fourth that measured in the position indicated by the filled dot.
highest electrical anisotropy (on average λX/λY = 20) and is there- The P-M delay is less within the zone of P-M connection and
8
fore adapted for high speed of excitation propagation. The endo- greater outside the zone. This discrete organization of the P-M
cardial conductive system exhibits relatively high λs, but smaller junctions in the ventricles creates conditions of electrical inde-
anisotropy, which is presumably a result of the meshlike cytoar- pendence of the virtually 2-D system of Purkinje fibers from the
chitecture of Purkinje terminals. 3-D syncytium of M fibers. Otherwise, a strong link between P
Values of λX,exp in working myocardium are smaller than those and M fibers would reduce the speed of propagation in the thin
in the false tendon, mainly because of differences in the dimen- layer of the Purkinje network to that of the working myocardium,
sionality of systems. In both systems, ρi,X is in the range of 150 which would compromise synchrony in the activation of the
⋅
to 200 Ohm cm (i.e., close to the resistivity of the cytoplasm ventricles. The size of the P-M junctions where prepotentials are
⋅
[~120 Ohm cm]).67 However, ρY in working myocardium is a
few-fold smaller than that in the false tendon, where ρi,Y ≈ 2 to
recorded allows one to speculate, using a 3-D model, that only
few P and M fibers in the P-M junction form direct gap junc-
4 kOhm cm.⋅
Purkinje-muscle (P-M) junctions, which were first described
tional contact (i.e., much smaller than the number of cells located
in the area with prepotentials).
in Purkinje–papillary muscle preparations from the ventricles of In Purkinje terminals, Cx40 and Cx43 are coexpressed with a
dog hearts, mediate an approximately 5-ms P-M delay52 and prevalence of Cx43. Working ventricular cardiomyocytes prefer-
characteristic prepotentials (notch) on the upstroke of AP of the entially express Cx43. It was demonstrated that Cx40 does not
transitional cell (Tr) located between P and M cells. Furthermore, form, or does so with very low efficiency, heterotypic junctions
stepwise penetrations of the microelectrode from the endocardial with Cx43.20 Therefore, electrical cell-cell connections at P-M
side in ventricular preparations of the dog heart revealed two connections are organized predominantly through homotypic
scenarios: (1) the AP of a P cell was followed by the AP of an M Cx43 GJ channels.
cell approximately 4 to 10 ms thereafter—no signs of prepoten-
tial were evident during this delay period, indicating that there
is no direct electrotonic coupling; and (2) the AP of a P cell was Alterations of Cell-Cell Coupling
followed by the AP of a Tr cell exhibiting the prepotential, and
then the AP of the M cell was recorded (Figure 8-7, A).53 Inter- in Cx-Deficient Animals and
estingly, P-M junctions were distributed discretely on the endo- Disease-Related Mutations
cardium surface with distances between them of approximately
0.8 to 2 mm (see Figure 8-7, A). Consequently, excitation from To understand the role of connexin in cardiac impulse propaga-
Purkinje cells is transmitted to working myocardium only at P-M tion, unrestricted and cardiomyocyte-specific null mutant mice
junctions (black dots). Excitation spreads from these loci in all with different cardiac Cx genes were bred. Cx43 gene deletion
directions; dashed lines show isochrones of the spread of excitation resulted in postnatal death as a result of malformation of the right
in working myocardium. Figure 8-7, B shows that an area in ventricular outflow tract.70 The postnatal lethality of Cx43-
which prepotentials are recorded has an elliptical shape of about deficient mice can be circumvented by cardiomyocyte-directed
300 × 150 µm on average; the longer axis is always parallel with deletion. The resulting mice exhibited a reduced conduction
the direction of the muscle fibers. The total area of P-M junctions velocity in the ventricles, thus demonstrating the importance of
constituted about 5% of the endocardial surface of the dog heart. Cx43 in the maintenance of ventricular conduction.71
The cross-section of the

ventricular wall
*
P Tr M P M
Purkinje f.
Working
cardio-
myocytes
A
A face-view to the endocardial side
C P
M
P-M
junction 50 µm
*
*
100 mV
6 ms
*
M
P
M
* * * *
B
Figure 8-7. Schematics of discrete P-M junctions (black spots) between a subendocardial network of Purkinje fibers (pink) and working myocardium (grey). A, A cross-section
of ventricular wall. The spread of excitation in Purkinje fibers and working myocardium are shown by pink arrows and black arrows, respectively. Stepwise penetration of
the microelectrode into P-M junction revealed APs with prepotentials (Tr), whereas in majority of places only APs typical for P and M fibers were recorded. B, A face-view
of the ventricle from the endocardial side showing the spread of excitation in the network of Purkinje fibers (pink dashed lines) and working myocardium (black dashed
lines). C, The inset from B showing that in the P-M junction, an area in which APs with prepotentials are recorded has an elliptical form.
(Modified from Bukauskas FF, Sakson ME, Kukushkin NI: Discrete zones of electrical connection between Purkinje terminals and muscle fibers in dog ventricles [in Russian].
Biofizika 21:887–892, 1976.)
Cx40-deficient mice exhibited impaired function of automa- (c.932delC), leading to deletion of the last 72 amino acids
ticity,72 slower conduction velocity in the atria and the AV-nodal of the putative Cx43 protein.85 HeLa transfectants expressing
region, and development of first-degree AV block associated with c.932delC-Cx43 mutant showed a marked reduction in the ability
bundle branch block,73,74 which correlates with the Cx40 expres- of cell pairs to couple electrically. Cells that were cotransfected
sion pattern. Impaired AV-nodal conduction is reflected by a with c.932delC-Cx43, and Cx40 or Cx43 demonstrated a
prolongation of the AV-nodal refractory period and shift in the dominant-negative effect of mutation on the function of Cx40
Wenckebach point to longer cycle lengths in these mice.75 More- and Cx43, which explains the arrhythmogenic nature of
over, the replacement of Cx43 by Cx40, or Cx40 by Cx45, showed c.932delC-Cx43 mutant.
that some cardiac Cxs can be substituted for others.76,77 Cx40 is Some patients with oculodentodigital dysplasia (ODDD),
critical for normal morphogenesis because a high incidence of which is caused by mutations in the Cx43 gene, were reported to
cardiac malformations occur in Cx40-deficient mice.78 suffer from cardiac dysfunction.86 ODDD is a dominant, inher-
Cx40 is also expressed in the kidney and can influence blood ited disorder with characteristic anomalies of the fingers, toes,
pressure through the renin-angiotensin system.79 Cx40-null eyes, face, and teeth. Cardiac abnormalities, including sick sinus
mutated mice are hypertensive, possibly because of an increase syndrome, ventricular tachycardia, and sudden cardiac death,
in the number of renin-secreting cells80 and a loss of the Cx40- have been reported in two families with ODDD. Thus far, more
mediated calcium-dependent inhibitory effect of angiotensin II.79 than 50 mutations in Cx43 related to ODDD have been identi-
The increase of blood pressure caused by alterations in Cx40 fied, including point mutations, frame shift mutations, and amino
expression or function of its mutations can increase arrhythmo- acid duplications.87 Mice carrying the Cx43*G138R point muta-
genicity of the heart. tion reproduced typical characteristics of ODDD syndrome
The effects of Cx45 on cardiac impulse propagation remain and, in addition, exhibited alterations in the electrocardiogram
elusive because unrestricted and cardiac-actin promoter- and caused spontaneous arrhythmias.88 HeLa cells expressing
Cre–mediated deletion of Cx45 resulted in embryonic Cx43*G138R revealed a loss of junctional communication while
lethality caused by endocardial cushion defects and impaired they formed functional channels with cells expressing wild type
vascularization.81 Cx43 tagged with EGFP. Cx43G138R/Cx43 heterotypic GJs
Deletion of mCx30.2 leads to the acceleration of AV-nodal were clustered in large and multiple junctional plaques but cell-
conduction without obvious morphologic heart abnormalities.43 cell coupling was low and exhibited asymmetric voltage gating.
In addition, the AV-nodal refractory period is shortened in these Single-channel conductance was similar to wild type Cx43 but
mice, which results in faster ventricular response rates during most of the GJ channels were closed because of modifications in
induced episodes of atrial fibrillation. Thus, mCx30.2 contributes voltage gating. Therefore, we assumed that Cx43*G138R GJ
to slow AV-nodal conduction and the protection of the ventricles channels can oligomerize into hemichannels, which can dock and
from atrial tachyarrhythmias. No alterations of excitation propa- form GJ channels that are preferentially closed because of changes
gation were found in atria, bundle branches, or ventricles, where in their gating properties.88
mCx30.2 expression is absent. Recently, three novel point mutations of Cx43 were reported
Thus far, no congenital heart diseases have been linked to in 418 Chinese patients with congenital heart diseases.89 The first
mutation of mCx30.2 (hCx31.9) or Cx45, although Cx45- mutation, R153Q, was located in the third transmembrane region
deficient mice show severe heart abnormalities and embryonic of Cx43; the other two mutations, G261W and A323V, were
death. Alterations in distribution and expression levels of cardiac found in the C-terminal region. All affected amino acids are
Cxs have been described in several noncongenital and some con- highly conserved in Cx43. Most of these patients had atrial and
gestive heart failure types not related to Cxs mutations.82 ventricular septal defects, tetralogy of Fallot, pulmonary atresia,
Some forms of arrhythmias appear to be genetically deter- or stenosis, as well as other forms of cardiac malformations. It is
mined by associated mutations that lead to decreased expression assumed that those malformations are related to changes in func-
of Cx40 and modification (D1275N) of the sodium channel tion including gating of Cx43 GJ channels, which needs to be
(SCN5A).83 In addition, it was shown that a predisposition to supported by more detailed studies of unitary conductance and
atrial standstill in patients carrying polymorphisms in the pro- permselectivity of those mutants.
moter region decreased Cx40 expression. Furthermore, sporadic
cases of atrial fibrillation not related to a family history have been
identified as being caused by somatic or tissue-specific genetic
mutations of Cx40 and Cx43. Somatic mutation of Cx40 Acknowledgments
(Cx40*A96S) was described in patients with atrial fibrillation.84
The Cx43 somatic mutation was identified in the atrium of The author thanks Dr. Romualdas Veteikis for helpful comments.
patients with a long history of atrial fibrillation. It was caused by This work was supported by National Institutes of Health grants
a frame shift rising from a single cytosine nucleotide deletion R01NS072238 and RO1HL084464.
References determinant of charge selectivity in connexin46 9. Johnson RG, Reynhout JK, TenBroek EM, et al:
channels. Biophys J 79:3036–3051, 2000. Gap junction assembly: Roles for the formation
1. González D, Gómez-Hernández JM, Barrio LC: 5. Bukauskas FF, Verselis VK: Gap junction channel plaque and regulation by the c-terminus of con-
Molecular basis of voltage dependence of connexin gating. Biochim Biophys Acta 1662:42–60, 2004. nexin43. Mol Biol Cell 23:71–86, 2012.
channels: An integrative appraisal. Prog Biophys 6. Lal R, John SA, Laird DW, et al: Heart gap junc- 10. Bukauskas FF, Weingart R: Voltage-dependent
Mol Biol 94:66–106, 2007. tion preparations reveal hemiplaques by atomic gating of single gap junction channels in an insect
2. Sosinsky GE, Nicholson BJ: Structural organiza- force microscopy. Am J Physiol 268:C968–C977, cell line. Biophys J 67:613–625, 1994.
tion of gap junction channels. Biochim Biophys 1995. 11. Márquez-Rosado L, Solan JL, Dunn CA, et al:
Acta 1711:99–125, 2005. 7. Zampighi GA, Loo DD, Kreman M, et al: Func- Connexin43 phosphorylation in brain, cardiac,
3. Kronengold J, Trexler EB, Bukauskas FF, et al: tional and morphological correlates of connexin50 endothelial and epithelial tissues. Biochim Biophys
Pore-lining residues identified by single channel expressed in xenopus laevis oocytes. J Gen Physiol Acta 1818:1985–1992, 2012.
scam studies in cx46 hemichannels. Cell Commun 113:507–524, 1999. 12. Gaietta G, Deerinck TJ, Adams SR, et al:
Adhes 10:193–199, 2003. 8. Kim KS, Neu J, Oster G: Curvature-mediated Multicolor and electron microscopic imaging of
4. Trexler EB, Bukauskas FF, Kronengold J, et al: interactions between membrane proteins. Biophys- connexin trafficking. Science 296:503–507,
The first extracellular loop domain is a major ical J 75:2274–2291, 1998. 2002.
13. Falk MM, Baker SM, Gumpert AM, et al: Gap 34. Eugenin EA, Basilio D, Sáez JC, et al: The role of 55. Bouman LN, Duivenvoorden JJ, Bukauskas FF,
junction turnover is achieved by the internalization gap junction channels during physiologic and et al: Anisotropy of electrotonus in the sinoatrial
of small endocytic double-membrane vesicles. Mol
Biol Cell 20:3342–3352, 2009.
14. Bukauskas FF, Jordan K, Bukauskiene A, et al:
pathologic conditions of the human central nervous
system. J Neuroimmune Pharmacol Mar 23, [Epub
ahead of print] 2012.
node of the rabbit heart. J Mol Cell Cardiol
21:407–418, 1989.
56. Bukauskas FF, Kukushkin NI, Sakson ME: Model
8
Clustering of connexin 43-enhanced green fluores- 35. Contreras JE, Saez JC, Bukauskas FF, et al: Gating of a 2-dimensional anisotropic syncytium. Biofizika
cent protein gap junction channels and functional and regulation of connexin 43 (cx43) hemichan- (Russian; translated in English by Pergamon Press)
coupling in living cells. Proc Natl Acad Sci U S A nels. Proc Natl Acad Sci U S A 100:11388–11393, 19:712–716, 1974.
97:2556–2561, 2000. 2003. 57. Bukauskas FF, Veteikis RP, Gutman AM: Model of
15. Palacios-Prado N, Briggs SW, Skeberdis VA, et al: 36. Miro-Casas E, Ruiz-Meana M, Agullo E, et al: a passive 3-dimensional anisotropic syncytium as a
Ph-dependent modulation of voltage gating in Connexin43 in cardiomyocyte mitochondria con- continuous medium. Biofizika (Russian; translated
connexin45 homotypic and connexin45/ tributes to mitochondrial potassium uptake. Car- in English by Pergamon Press) 20:1083–1086,
connexin43 heterotypic gap junctions. Proc Natl diovasc Res 83:747–756, 2009. 1975.
Acad Sci U S A 107:9897–9902, 2010. 37. Boengler K, Stahlhofen S, van de Sand A, et al: 58. Bukauskas F, Bytautas A, Gutman A, et al: Simula-
16. Palacios-Prado N, Sonntag S, Skeberdis VA, et al: Presence of connexin 43 in subsarcolemmal, but tion of passive electrical properties in two- and
Gating, permselectivity and ph-dependent modu- not in interfibrillar cardiomyocyte mitochondria. three-dimensional anisotropic syncytial media. In
lation of channels formed by connexin57, a major Basic Res Cardiol 104:141–147, 2009. Bukauskas F, editor: Intercellular Communication,
connexin of horizontal cells in the mouse retina. 38. Rodríguez-Sinovas A, Sánchez JA, González- Manchester/New York, 1991, Manchester Univer-
J Physiol 587:3251–3269, 2009. Loyola A, et al: Effects of substitution of cx43 by sity Press, pp 203–217.
17. Skeberdis VA, Rimkute L, Skeberdyte A, et al: cx32 on myocardial energy metabolism, tolerance 59. Woodbury JW, Crill WE: On the problem of
Ph-dependent modulation of connexin-based gap to ischaemia and preconditioning protection. impulse conduction in the atrium. Nervous Inhibi-
junctional uncouplers. J Physiol 589:3495–3506, J Physiol 588:1139–1151, 2010. tion Procedings, New York, Pergamon Press,
2011. 39. Rottlaender D, Boengler K, Wolny M, et al: Gly- 124–135, 1961.
18. Moreno AP: Biophysical properties of homomeric cogen synthase kinase 3β transfers cytoprotective 60. Bukauskas FF, Gutman AM, Kisunas KJ, et al:
and heteromultimeric channels formed by signaling through connexin 43 onto mitochondrial Electrical cell coupling in rabbit sino atrial node
cardiac connexins. Cardiovasc Res 62:276–286, atp-sensitive k+ channels. Proc Natl Acad Sci U S and atrium. Experimental and theoretical evalua-
2004. A 109:E242–E251, 2012. tion. In Bouman LN, Jongsma HJ, editors:
19. Kreuzberg MM, Sohl G, Kim J, et al: Functional 40. Valiunas V: Biophysical properties of connexin-45 Cardiac Rate and Rhythm. Physiological, Morpho-
properties of mouse connexin30.2 expressed in the gap junction hemichannels studied in vertebrate logical and Developmental Aspects, Hague/Boston/
conduction system of the heart. Circ Res 96:1169– cells. J Gen Physiol 119:147–164, 2002. London, 1982, Martinus Nijhoff Publishers, pp
1177, 2005. 41. Beahm DL, Hall JE: Opening hemichannels in 195–216.
20. Rackauskas M, Kreuzberg MM, Pranevicius M, nonjunctional membrane stimulates gap junction 61. Deakin MA, Neild TO, Turner RG: The extension
et al: Gating properties of heterotypic gap junction formation. Biophys J 86:781–796, 2004. of two-dimensional cable theory to arteries and
channels formed of connexins 40, 43 and 45. 42. Kreuzberg MM, Willecke K, Bukauskas F: arterioles. Bull Math Biol 47:409–424, 1985.
Biophys J 92:1952–1965, 2007. Connexin-mediated cardiac impulse propagation: 62. Kukushkin NI, Bukauskas FF, Sakson MY, et al:
21. Harris AL: Connexin channel permeability to cyto- Connexin 30.2 slows atrioventricular conduction in Anisotropy of steady speedes and delays of the
plasmic molecules. Prog Biophys Mol Biol 94:120– mouse heart. Trends Cardiovasc Med 16:266–272, extrasystolic waves in the heart of the dog. Biofizika
143, 2007. 2006. (Russian; translated in English by Pergamon Press)
22. Bukauskas FF, Peracchia C: Two distinct gating 43. Kreuzberg M, Schrickel J, Ghanem A, et al: Con- 20:687–692, 1975.
mechanisms in gap junction channels: CO2- nexin30.2 containing gap junction channels 63. Spach MS, Josephson ME: Initiating reentry: The
sensitive and voltage-sensitive. Biophys J 72:2137– decelerate impulse propagation through the atrio- role of nonuniform anisotropy in small circuits.
2142, 1997. ventricular node. Proc Natl Acad Sci U S A J Cardiovasc Electrophysiol 5:182–209, 1994.
23. Harris AL, Spray DC, Bennett MVL: Kinetic 108:5959–5964, 2006. 64. Bonke FI: Electrotonic spread in the sinoatrial
properties of a voltage-dependent junctional con- 44. Bao L, Locovei S, Dahl G: Pannexin membrane node of the rabbit heart. Pflugers Arch 339:17–23,
ductance. J Gen Physiol 77:95–117, 1981. channels are mechanosensitive conduits for ATP. 1973.
24. Paulauskas N, Pranevicius H, Mockus J, et al: FEBS Lett 572:65–68, 2004. 65. Bukauskas FF, Veteikis RP, Gutman AM, et al:
A stochastic 16-state model of voltage-gating of 45. Kienitz MC, Bender K, Dermietzel R, et al: Pan- Intercellular coupling in the sinus node of the
gap junction channels enclosing fast and slow gates. nexin 1 constitutes the large conductance cation rabbit heart. Biofizika (Russian; translated in
Biophys J 102:2471–2480, 2012. channel of cardiac myocytes. J Biol Chem 286:290– English by Pergamon Press) 22:108–112, 1977.
25. Trexler EB, Bennett MV, Bargiello TA, et al: 298, 2011. 66. Masson-Pévet M, Bleeker WK, Mackaay AJ, et al:
Voltage gating and permeation in a gap junction 46. Gemel J, Lin X, Collins R, et al: Cx30.2 can form Sinus node and atrium cells from the rabbit heart:
hemichannel. Proc Natl Acad Sci U S A 93:5836– heteromeric gap junction channels with other A quantitative electron microscopic description
5841, 1996. cardiac connexins. Biochem Biophys Res Commun after electrophysiological localization. J Mol Cell
26. Elenes S, Martinez AD, Delmar M, et al: Hetero- 369:388–394, 2008. Cardiol 11:555–568, 1979.
typic docking of cx43 and cx45 connexons blocks 47. Pfenniger A, Wohlwend A, Kwak BR: Mutations 67. Schanne OF: Measurement of cytoplasmic resistiv-
fast voltage gating of cx43. Biophys. J. 81:1406– in connexin genes and disease. Eur J Clin Invest ity by means of the glass microelectrode. In
1418, 2001. 41:103–116, 2011. Lavellee M, OFS, Herbert NC, editors: Glass
27. Chen-Izu Y, Moreno AP, Spangler RA: Opposing 48. Bukauskas FF, Elfgang C, Willecke K, et al: Bio- Microelectrodes, New York, 1969, Wiley, pp
gates model for voltage gating of gap junction physical properties of gap junction channels 299–321.
channels. Am J Physiol Cell Physiol 281:C1604– formed by mouse connexin40 in induced pairs of 68. Van Capelle FJ, Janse MJ, Varghese PJ, et al:
C1613, 2001. transfected human hela cells. Biophys J 68:2289– Spread of excitation in the atrioventricular node of
28. Ramanan SV, Brink PR, Varadaraj K, et al: A three- 2298, 1995. isolated rabbit hearts studied by multiple micro-
state model for connexin37 gating kinetics. Biophys 49. Kleber AG, Rudy Y: Basic mechanisms of cardiac electrode recording. Circ Res 31:602–616, 1972.
J 76:2520–2529, 1999. impulse propagation and associated arrhythmias. 69. Torii H: Electron microscope observation of the
29. Bukauskas FF, Kreuzberg M, Rackauskas M, et al: Physiol Rev 84:431–488, 2004. S-A and A-V nodes and Purkinje fibers of the
Properties of mouse connexin 30.2 and human 50. Bukauskas FF, Veteikis RP: Passive electrical prop- rabbit. Jap Circ J 26:39–77, 1962.
connexin 31.9 hemichannels; implications for erties of the atrioventricular region of the rabbit 70. Reaume AG, De Sousa PA, Kulkarni S, et al:
atrioventricular conduction in the heart. Proc Natl heart. Biofizika (Russian; translated in English by Cardiac malformation in neonatal mice lacking
Acad Sci U S A 103:9726–9731, 2006. Pergamon Press) 22:499–504, 1977. connexin43. Science 267:1831–1834, 1995.
30. De Vuyst E, Boengler K, Antoons G, et al: Phar- 51. De Mello WC: Passive electrical properties of the 71. van Rijen HV, Eckardt D, Degen J, et al: Slow
macological modulation of connexin-formed chan- atrio-ventricular node. Pflugers Arch 37:135–139, conduction and enhanced anisotropy increase the
nels in cardiac pathophysiology. Br J Pharmacol 1977. propensity for ventricular tachyarrhythmias in
163:469–483, 2011. 52. Mendez C, Mueller WJ, Merideth J, et al: Interac- adult mice with induced deletion of connexin43.
31. Sáez JC, Schalper KA, Retamal MA, et al: Cell tion of transmembrane potentials in canine Pur- Circulation 109:1048–1055, 2004.
membrane permeabilization via connexin hemi- kinje fibers and at Purkinje fiber-muscle junctions. 72. Bagwe S, Berenfeld O, Vaidya D, et al: Altered
channels in living and dying cells. Exp Cell Res Circ Res 24:361–372, 1969. right atrial excitation and propagation in con-
316:2377–2389, 2010. 53. Bukauskas FF, Sakson ME, Kukushkin NI: nexin40 knockout mice. Circulation 112:2245–
32. Scheckenbach KE, Crespin S, Kwak BR, et al: Discrete zones of electrical connection 2253, 2005.
Connexin channel-dependent signaling pathways between purkinje terminals and muscle fibers in 73. Simon AM, Goodenough DA, Paul DL: Mice
in inflammation. J Vasc Res 48:91–103, 2011. dog ventricles. Biofizika (Russian; translated lacking connexin40 have cardiac conduction
33. Paul DL, Ebihara L, Takemoto LJ, et al: Con- in English by Pergamon Press) 21:887–892, abnormalities characteristic of atrioventricular
nexin46, a novel lens gap junction protein, induces 1976. block and bundle branch block. Curr Biol 8:295–
voltage-gated currents in nonjunctional plasma 54. Adomonis VM, Bredikis II, Bukauskas FF: Suction 298, 1998.
membrane of xenopus oocytes. J Cell Biol electrode with internal perfusion. Fiziol Zh SSSR 74. van Rijen HV, van Veen TA, van Kempen MJ, et
115:1077–1089, 1991. Im I M Sechenova 69:272–275, 1983. al: Impaired conduction in the bundle branches of
mouse hearts lacking the gap junction protein con- 80. Krattinger N, Capponi A, Mazzolai L, et al: Con- Physiological characterization of a connexin
nexin40. Circulation 103:1591–1598, 2001. nexin40 regulates renin production and blood pres- 43-deletion mutant identified from atrial tissue.
75. VanderBrink BA, Sellitto C, Saba S, et al: sure. Kidney Int 72:814–822, 2007. Circ Res 122:236–244, 2010.
Connexin40-deficient mice exhibit atrioventricular 81. Kruger O, Plum A, Kim JS, et al: Defective vascu- 86. Paznekas WA, Boyadjiev SA, Shapiro RE, et al:
nodal and infra-hisian conduction abnormalities. lar development in connexin 45-deficient mice. Connexin 43 (gja1) mutations cause the pleiotropic
J Cardiovasc Electrophysiol 11:1270–1276, 2000. Development 127:4179–4193, 2000. phenotype of oculodentodigital dysplasia. Am J
76. Alcolea S, Jarry-Guichard T, de Bakker J, et al: 82. Dupont E, Matsushita T, Kaba RA, et al: Altered Hum Genet 72:408–418, 2003.
Replacement of connexin40 by connexin45 in the connexin expression in human congestive heart 87. Zoidl G, Dermietzel R: Gap junctions in inherited
mouse: Impact on cardiac electrical conduction. failure. J Mol Cell Cardiol 33:359–371, 2001. human disease. Pflugers Arch Eur J Physiol
Circ Res 94:4–6, 2004. 83. Groenewegen WA, Firouzi M, Bezzina CR, et al: 460:451–466, 2010.
77. Taffet SM, Jalife J: Swapping connexin genes: How A cardiac sodium channel mutation cosegregates 88. Dobrowolski R, Sasse P, Schrickel JW, et al: The
big is the gap? Circ Res 94:4–6, 2004. with a rare connexin40 genotype in familial atrial conditional connexin43g138r mouse mutant repre-
78. Gu H, Smith FC, Taffet SM, et al: High incidence standstill. Circ Res 92:14–22, 2003. sents a new model of hereditary oculodentodigital
of cardiac malformations in connexin40-deficient 84. Gollob MH, Jones DL, Krahn AD, et al: Somatic dysplasia in humans. Hum Mol Genet 17:539–554,
mice. Circ Res 93:201–206, 2003. mutations in the connexin 40 gene (gja5) in atrial 2008.
79. Wagner C, de Wit C, Kurtz L, et al: Connexin40 fibrillation. N Engl J Med 354:2677–2688, 2006. 89. Wang B, Wen Q, Xie X, et al: Mutation analysis of
is essential for the pressure control of renin syn- 85. Thibodeau IL, Xu J, Li Q, et al: Paradigm of connexon43 gene in chinese patients with congeni-
thesis and secretion. Circ Res 100:556–563, 2007. genetic mosaicism and lone atrial fibrillation: tal heart defects. Int J Cardiol 145:487–489, 2010.
Biophysics of Cardiac Ion PART II
Channel Function
Biophysics of Normal and Abnormal

Cardiac Sodium Channel Function 9
Thomas J. Hund and Peter J. Mohler
sensitivity within the cell.12,13 Although the α-subunit is capable of

CHAPTER OUTLINE forming a functional channel by itself, several auxiliary β-subunits
Nav Channel Activity in Health and Disease 95 have been identified (β1-β4) that modulate channel gating and/or
trafficking. Human mutations in both the α and β-subunits have
NaV Channel Trafficking and Stabilization in Health been linked to arrhythmia.14,15 Importantly, Nav1.5 and associated
and Disease 97 β-subunits reside with large macromolecular complexes com-
Nav Channel Posttranslational Regulation in Health and prised of adaptor, accessory, cytoskeletal, and regulatory proteins
Disease 98 (Figure 9-2).4,5,16-22 An exciting area of research going forward, and
the focus of this and other chapters in this volume (Chapters 18
Summary and Future Directions 99 and 21), is the physiologic function of Nav macromolecular com-
plexes and their role in arrhythmias and mechanical dysfunction
in the setting of human disease.
Abstract
Voltage-gated sodium channels (Nav) underlie the activity of Nav Channel Activity in Health and Disease
many excitable cells. In the heart, Nav channels are responsible
for the rapid cardiomyocyte action potential upstroke that pro- Over the course of the action potential, Na+ channels undergo a
motes rapid conduction of the electrical impulse leading to coor- remarkable series of conformational changes that are essential for
dinated mechanical contraction. Central to this function, Nav the rapid and controlled influx of Na+ ions that supports the
channels activate (and then inactivate) rapidly in response to a action potential upstroke (detailed in Chapter 1).23 Recent work
small depolarization of the membrane, resulting in a large influx to crystallize the bacterial Na+ channel NavAb (forms as a tetra-
of Na+ ions and further membrane depolarization. Dysfunction mer of identical subunits akin to most K+ channels) has added to
in Nav channel activity results in human diseases and disorders, decades of important research on the complex Nav structure–
including epilepsy, ataxia, cardiac arrhythmia, and myotonia.1-3 function relationship governing channel activity.24,25 Briefly, in
Variants in SCN5A, the gene encoding the primary cardiac Nav response to membrane depolarization, channels first activate
α-subunit Nav1.5, have been linked to human arrhythmia syn- rapidly as the four charged S4 segments (see Figure 9-1) move
dromes including long QT type 3 (LQT3), Brugada syndrome, outward and open the channel pore.26 Channel activation is fol-
cardiac conduction disease, sinus node disease, and atrial fibrilla- lowed almost immediately by rapid inactivation caused by further
tion.2,4,5 A detailed discussion of fundamental aspects of Nav outward movement of the S4 sensors in DIII and DIV and block-
structure function, gating, and pharmacology can be found in ing of the channel pore by the DIII-DIV linker that likely
Chapter 1. Here we discuss current understanding regarding depends on the interaction between hydrophobic residues
regulation of Nav biophysical activity and cellular function in (isoleucine-phenylalanine-methionine, IFM) in the III-IV linker
health and disease. and multiple sites in the linkers between S4 and S5 in DIII and
The Nav channel pore-forming α-subunit in vertebrates is a DIV and the cytoplasmic end of S6 in DIV.27-29 Rapid inactivation
single polypeptide with four homologous transmembrane domains is followed by transition into several different slow inactivation
(DI-DIV) comprised of six membrane-spanning α-helices (S1-S6; states that are controlled, in part, by P segments linking S5 and
Figure 9-1). Nav channels share structural similarities with S6 segments that form the inner channel pore (see Figure 9-1).
voltage-gated Ca2+ channels from which they may have evolved.6 Importantly, disruption of Na+ channel gating at any step during
Ten different Nav channel α-subunits are present in mammals, this highly coordinated set of movements may result in inappro-
each with specific biophysical, expression, and regulatory signa- priate (elevated, reduced) current and give rise to arrhythmias.
tures. The primary cardiac isoform, Nav1.5, is derived from 28 Dysfunction in Nav1.5 activity has been identified as a poten-
exons and undergoes alternative splicing in mammalian heart tial cause of arrhythmia in several forms of congenital and
(including human).4,7-11 Although Nav1.5 is the primary Nav acquired disease (see Figure 9-1). Classic examples of the link
channel expressed in the heart, limited expression of neuronal between inherited defects in Nav1.5 function and disease may be
α-subunits has also been reported and likely introduces important found in two inherited arrhythmia syndromes closely linked to
heterogeneity in Nav biophysical activity, localization, and drug mutations in SCN5A: Brugada syndrome and LQT3. An
95
96 BIOPHYSICS OF CARDIAC ION CHANNEL FUNCTION
P segment
DI DII DIII DIV
1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 β
1795insD
∆KPQ
NH2 IFM
(1505–1507)
Voltage E1053K
sensor A1924T
Long QT type 3
Brugada syndrome COOH
Overlap disorder
Figure 9-1. Topology of cardiac Nav1.5 channel (SCN5A) and human variants linked to congenital cardiac arrhythmia. Charged S4 segments (orange) in each domain func-
tion as voltage sensors. P segments between S5 and S6 form the channel pore. IFM motif in DIII-DIV linker (light blue circle) acts as inactivation particle to facilitate rapid
voltage-dependent inactivation. The location of several human variants in SCN5A linked to LQT3 (red square), Brugada syndrome (blue square), or overlap disorder (green
square) are indicated.
DI DII DIII DIV
1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 β
S516
14-3-3η S1503 FGF12B
S1505
NH2 S526
S529 T594 Calmodulin
Ankyrin-G MOG1
S571 SAP97
Nedd4-2
Interacting protein CaMKII phosphorylation site COOH
PKA phosphorylation site
Syntrophin
PKC phosphorylation site
Figure 9-2. Nav1.5 resides in macromolecular complexes comprised of β-subunits (red) and cytoskeletal, adapter, trafficking, and regulatory molecules. Known interacting
proteins are shown along with binding regions. Phosphorylation sites identified to date are indicated for CaMKII (red), PKA (blue), and protein kinase C (green).
overview of molecular basis for these diseases is provided here; introducing a secondary gain-of-function change in Nav activity
both are addressed in detail in later chapters. Brugada syndrome (e.g., depolarizing shift in voltage dependence of inactivation)
is an autosomal, dominant cardiac arrhythmia syndrome charac- that partially balances the primary loss-of-function change (e.g.,
terized by ST-segment elevation in the right precordial leads depolarizing shift in voltage dependence of activation).
(V1-V3) and sudden death in the absence of overt structural Although Brugada syndrome is a disease associated with Nav
disease.30,31 The syndrome is caused by loss-of-function variants loss-of-function, LQT3 represents the other extreme and illus-
in SCN5A (although identified in only about 20% of Brugada trates how Nav gain-of-function may also produce arrhythmia. A
cases31) that result in loss of Na+ current early in the action heterogeneous autosomal dominant genetic disease, LQT3 is
potential because of a loss of functional channel expression/ associated with abnormal QT-interval prolongation on the elec-
localization or defects in channel gating.20,32,33 The substrate for trocardiogram, syncope, polymorphic ventricular tachycardia,
Brugada syndrome arrhythmia is thought to arise in the epicar- and sudden death.14 Variants in SCN5A are the cause of LQT3,
dium of the right ventricle where high expression of transient which is associated with increased risk of cardiac events during
outward K+ current Ito may shift the balance of current to favor rest and decreased efficacy of β-blocker therapy.14,37,38 At least 50
exaggerated phase 1 repolarization (notch) and even loss of variants in SCN5A have been identified as causal for LQT3, and
the action potential plateau.34,35 Loss-of-function SCN5A variants although these variants are scattered throughout the cytoplasmic
have also been associated with isolated cardiac conduction disease face of the channel (with a cluster in the S4 segments), they
characterized by slow conduction throughout the heart (even primarily produce an increase in Na+-channel current (gain-of-
requiring pacemaker implantation) without the repolarization Nav function) by disrupting rapid inactivation, by shifting the
abnormalities or ventricular tachyarrhythmia observed in Brugada voltage-dependence of activation or inactivation, or by altering
syndrome.36 In general, gene variants that give rise to cardiac recovery from inactivation. A prototypical case study of the causal
conduction disease produce a less severe effect on channel link between SCN5A mutations and LQT3 comes from the first
function than those associated with Brugada syndrome, by reported LQT3 mutation resulting in three–amino acid deletion
Biophysics of Normal and Abnormal Cardiac Sodium Channel Function 97
in the DIII-DIV linker that is important for rapid channel inac- targeting channel regulatory pathways (e.g., CaMKII) may be a
9
tivation (ΔKPQ, see Figure 9-1).38 The variant ΔKPQ allele dis- second, and conceivably more practical, method to alter channel
rupts rapid inactivation, producing a small (~0.5% of peak) but availability for therapeutic benefit.
sustained Na+ current during the action potential plateau that
shifts the balance of current to prolong action potential and
increase the likelihood of arrhythmogenic afterdepolariza-
tions.38,39 Although LQT3 and Brugada syndrome are often pre- NaV Channel Trafficking and Stabilization
sented as two very different faces of Nav dysfunction, overlap in Health and Disease
syndromes have been reported, with features of both diseases
showing QT prolongation at slow rates coupled with ST eleva- As outlined before, Nav function is critical for normal heart func-
tion during exercise. Notably, the 1795insD mutation (see Figure tion, and alteration of Nav activity is a hallmark of many forms
9-1) interrupts rapid inactivation to enhance persistent Na+ of inherited and acquired diseases. Despite major advances, the
current and prolong action potential duration, especially at slow mechanistic link among specific molecular defects, Nav channel
pacing rates, but it also enhances an intermediate inactivation dysfunction, and arrhythmias associated with many human
state that compromises Nav recovery between stimuli, leading to arrhythmia variants and in common disease remains elusive.
loss of current at fast rates.40,41 Thus, as discussed before, for Mounting evidence, in particular from human arrhythmia vari-
cardiac conduction disease, a single molecular defect may alter ants in genes encoding ion channel accessory proteins (e.g.,
channel activity at more than one phase (e.g., rapid and slow adapter and scaffolding proteins, channel subunits, chaperones),
inactivation processes) in the gating cycle, giving rise to a complex highlight the importance of proper Nav channel localization
disease phenotype. within specific cellular domains for normal heart function.62-69
Studies of rare congenital conditions such as LQT3 and Voltage-gated Nav channels are targeted to specific membrane
Brugada syndrome have generated important insight into Nav domains in cardiomyocytes (Figure 9-3). Residing primarily at
channel function and have increased our understanding of cell ends in the region of the intercalated disc (ID), Nav channels
arrhythmogenesis in the more complex setting of acquired heart are also found at secondary sites at lateral and transverse-tubule
disease, where Nav channel dysfunction is a common finding. In membranes (see Chapter 18). Although the cardiac isoform
heart failure, for example, an increase in persistent (late) Na+ Nav1.5 largely comprises the Nav population at the ID, neuronal
current has been observed both in patients and animal models of isoforms (e.g., Nav1.1, Nav1.3, Nav1.6) are found almost exclu-
human disease.42-46 Similar to LQT3, in heart failure it is likely sively at transverse tubules.12,13,70 Although the functional conse-
that increased late Na+ current delays action potential repolariza- quences for differential targeting of Nav isoforms within the
tion and promotes pro-arrhythmogenic afterdepolarizations.45 cardiomyocyte remain unclear, computational studies demon-
Increased Na+ entry via the late current may also disrupt normal strate that concentration of Nav channels at the ID (Nav1.5),
Ca2+ homeostasis and promote mechanical dysfunction and pro- where cells are electrically and mechanically coupled, supports
gression of disease.47 The underlying mechanism for changes in electrical impulse propagation.71,72 Likewise, a role for the neu-
Nav function in heart failure is unknown but is most likely not a ronal isoforms (largely outside the ID) has been explored using
result of changes in expression of Nav α or β-subunits.42,45 In light low doses of tetrodotoxin (TTX) (insufficient to block TTX-
of known defects in β-adrenergic and calcium-dependent signal- resistant Nav1.5). These experiments show that although neuro-
ing pathways in heart failure, it is likely that defects in channel nal channels are not necessary for conduction under normal
posttranslational modification play a role in Nav dysfunction. conditions, they may play modulatory roles in excitation-
Regardless of Na+ current mechanism, mounting studies support contraction coupling and ventricular function.12,13
the late Na+ current as a viable therapeutic target (e.g., inhibition
with ranolazine) for preventing arrhythmias and progression of
disease in heart failure.48-51
There have also been many studies on the link between defects
in Nav function and arrhythmias after myocardial infarction, Lateral membrane
where reentrant arrhythmias are highly localized to the highly
remodeled tissue surrounding the infarct (border zone). In par-
ticular, in the canine heart, dramatic electrical and structural
remodeling has been identified coupled with anisotropic conduc-
tion and reentrant arrhythmias in the border zone region. Con- Intercalated
duction through this region is highly irregular, characterized by Transverse-tubules disc
slow and discontinuous conduction.52 Myocytes isolated from the
border zone regions display a characteristic time course of
changes in ion channel activity and action potential, including Lateral membrane Nav 1.5
significant decrease in peak Na+ current, altered subcellular local- Intercalated
ization, decreased channel availability, and slow recovery from disc
SR
inactivation.53-55 Computational studies demonstrate that mea- Transverse Nav 1.5
sured defects in Nav channel activity prolong refractoriness, tubules
decrease conduction velocity, and increase susceptibility to initia-
tion of reentrant arrhythmias.53,56-59 Although the precise mecha-
nism underlying abnormal Nav function after myocardial ?
infarction is likely multifactorial and not fully understood, roles
for increased oxidative stress and/or increased CaMKII- Nav 1.1 Nav 1.6
dependent channel phosphorylation have been identified.57,60 Nav 1.3
Interestingly, adenoviral overexpression of skeletal muscle Nav Figure 9-3. Nav channels are differentially localized within cardiomyocyte.
channel, with inactivation shifted to more depolarized potentials A, Three-dimensional rendering of isolated murine ventricular cardiomyocyte
compared with cardiac isoform, has been shown to improve con- stained with markers for the nucleus (yellow), intercalated disc (green), and z-line
duction and suppress arrhythmias in the canine after myocardial (blue). B, Schematic indicating Nav localization and interacting partners at specific
infarction.61 It is intriguing to consider the possibility that membrane domains within the cardiomyocyte.
An important outstanding question is how are Nav channels dystrophin-syntrophin complex through direct binding of Nav1.5
differentially targeted and regulated within the cardiomyocyte? C-terminus with a PDZ domain in syntrophin.88-91 Na+ current
The answer may be found in the growing family of adapter, and Nav1.5 expression are decreased in mice lacking dystrophin
accessory, cytoskeletal and regulatory proteins—including α1- (mdx5cv), and human mutations in SNTA1 (encodes α1-syntrophin)
syntrophin, ankyrin-G, βIV-spectrin, glycerol-3-phosphate have been identified as the cause of increased Na+ current in
dehydrogenase 1–like protein, MOG1, Nedd4-2, SAP97, and congenital long QT syndrome, and sudden infant death
calmodulin—that reside in macromolecular complexes with Nav syndrome.90-92,87 It will be interesting in the future to determine
channels (see Figure 9-2).4,5,16-22 Recent studies demonstrate that the precise role of Nav1.5 interacting partners at the intercalated
interactions between Nav and these related proteins: (1) regulate disc and lateral membrane in targeting, retention, and function
channel function, (2) differ depending on local membrane of the channel.
domain, and (3) are altered in disease.
The Nav1.5 α-subunit associates with one or more β-subunits
(β1-β4). These β-subunits are type I integral membrane proteins
with an extracellular N-terminus containing an immunoglobulin Nav Channel Posttranslational Regulation
domain with homology to domains found in cell adhesion molin Health and Disease
ecules, a single transmembrane domain, and a cytoplasmic
C-terminal sequence.73 Differential localization of β-subunits has Tight spatial and temporal control of local signaling domains is
been reported in mouse heart, with co-localization of β1 and β2 essential for proper regulation and activity of Nav channels in
found at the intercalated disc and T-tubules, β3 at T-tubules, and cardiomyocytes. Importantly, changes in posttranslational modi-
β4 at the intercalated disc.12 Beta-subunits are reported to alter fication of membrane proteins are associated with increased sus-
Nav1.5 membrane expression and gating. Importantly, human ceptibility to congenital and acquired arrhythmia.90,93-95 In
mutations in Nav β-subunits have been linked to diseases and particular, Nav channel activity is heavily regulated by posttrans-
conditions, including Brugada syndrome (SCN1B and SCN3B), lational modification, including phosphorylation by protein
long QT (SCN4B), cardiac conduction disease (SCN1B), and kinase C,96-98 protein kinase A (PKA),99-102 CaMKII,82,103-106 and
atrial fibrillation (SCN1B, SCN2B and SCN3B).4,67,73-76 glycosylation.4,107 Important in the setting of heart disease,
The ankyrin family of multifunctional adapter proteins (e.g., β-adrenergic stimulation is known to increase INa through PKA-
ankyrin-R, ankyrin-B, and ankyrin-G expressed in the heart) direct and PKA-indirect modulation.99-101,108 As discussed before,
coordinates macromolecular complexes at specific membrane increased INa (late) likely contributes to arrhythmogenesis in heart
domains in cardiomyocytes. The most studied ankyrin in the failure. In parallel, β-adrenergic stimulation is enhanced in failing
heart, ankyrin-B, is involved in targeting of the Na+/Ca2+ hearts. Therefore, it is interesting to consider the possibility that
exchanger, Na+/K+ adenosine triphosphatase (ATPase), Kir6.2, defects in Nav1.5 posttranslational modification downstream of
InsP3 receptor, and protein phosphatase 2A to transverse β-adrenergic stimulation may underlie INa dysfunction in heart
tubules.77-80 Ankyrin-G, on the other hand, is concentrated at the failure. Although PKA directly phosphorylates Nav1.5 at specific
cardiomyocyte ID, where it associates directly with Nav1.5 residues in the DI-DII linker (see Figure 9-2),102 studies with
through binding to a conserved motif in the DII-DIII linker (see PKA inhibitors suggest that this accounts for only part of the
Figures 9-2 and 9-3).19,20,81 Interfering with ankyrin-G/Nav1.5 effect of β-adrenergic stimulation on Nav function.101,109 PKA-
interaction, either through knockdown of ankyrin-G or expres- independent effects likely involve regulation of Nav1.5 surface
sion of mutant Nav1.5 lacking ankyrin-binding activity, disrupts expression by the stimulatory G protein α-subunit (Gαs) and
Nav1.5 membrane, targeting resulting in loss of Na+ current.19,20 caveolin-3 (CAV3).101,109,110 Human gene variants in CAV3 have
Recently, ankyrin-G has been identified in a functional macro- been linked to increased late (persistent) INa and arrhythmias in
molecular complex in vivo with Nav1.5, the actin-associated poly- long QT syndrome type 9.65 Although the link between mutant
peptide βIV-spectrin, and calmodulin kinase II (CaMKII) for CAV3 and changes in Nav function remain unknown, computa-
targeting and regulation of Nav function and cell membrane tional studies suggest that disruption of CAV3/Gαs interaction
excitability.82 Ankyrin-G also has been shown to associate and may alter the dynamics of caveolae membrane fusion to give rise
co-localize with desmosome and gap junction proteins to increased late Na+ current.111
plakophilin-2 and connexin43,83 supporting the notion that Nav channels are also regulated by Ca2+/CaM-dependent
ankyrin-G is an important nodal point for coordinating mechani- pathways, which are known to be dysregulated in disease.112,113
cal and electrical signaling at the cardiomyocyte ID. Importantly, CaM interacts directly with multiple Nav isoforms, including
defects in ankyrin-based pathways have been identified as the Nav1.5 via a C-terminal IQ motif and a domain in the DIII-DIV
underlying cause for abnormal channel targeting and arrhythmias loop with variable results.113-119 Although the net effect of CaM
in congenital and acquired forms of cardiac disease.20,62,84-86 In on channel activity likely depends on factors such as specific
the case of Nav1.5, a human SCN5A variant linked to Brugada isoform and experimental conditions, binding of CaM to the
syndrome (E1053K) is located in the ankyrin-binding motif IQ motif of human Nav1.5 has been shown to enhance slow
of Nav1.5 and disrupts ankyrin-binding activity in vitro (see component of inactivation,113 with other studies showing
Figure 9-1).20 CaM-dependent effects on steady-state inactivation.105,114,118
Nav channels display unique biophysical properties based on Interestingly, the human mutation A1924T in the Nav1.5 IQ
their specific membrane localization.70 Consistent with these motif (see Figure 9-1) results in Brugada syndrome and alters
findings, work from several groups has recently shown that channel activity by interrupting CaM binding to Nav1.5.113-115
Nav1.5 complexes with distinct partners at the ID versus lateral More recently, a role for Ca2+/CaM regulation of Nav1.5 via
membranes.87 In addition to ankyrin-G, Nav1.5 has been found CaMKII-dependent phosphorylation has been described.82,104-106,112
to associate with synapse-associated protein SAP97, a member of Specifically, initial studies showed that overexpression of
the membrane-associated guanylate kinase family, at the cardio- CaMKIIδc in rabbit (adenoviral, acute) and mouse (transgenic,
myocyte ID (see Figure 9-3).87 This interaction depends on the chronic) ventricular cardiomyocytes shifted steady-state inactiva-
terminal three amino acids of the Nav1.5 C-terminus (see Figure tion in the hyperpolarizing direction (decreased availability),
9-2), and knockdown of SAP97 leads to a decrease in INa and increased intermediate activation, slowed recovery from inactiva-
reorganization of Nav1.5 in adult cardiomyocytes. Nav1.5 tion, and enhanced late Na+ current.104 At the cellular level,
is also found, albeit to a much lower degree, at the lateral CaMKIIδ overexpression resulted in accumulation of intracel-
membrane and is likely regulated by interaction with the lular Na+ and prolongation of APD, which translates to increased
QRS (marker for slow intraventricular conduction) and QT in disease. Proper localization of Nav within local signaling
9
intervals on the electrocardiogram, as well as increased suscepti- domains is essential for normal membrane excitability and heart
bility to ventricular arrhythmias.104 It is now clear that CaMKII function. Importantly, mounting evidence demonstrates that
phosphorylates the channel at multiple sites in the DI-DII linker defects in local signaling and regulation of Nav channels underlie
(including S516, S571, T594) (see Figure 9-2) to alter steady- abnormal cell excitability and arrhythmia in heart disease,
state channel availability, recovery from inactivation and persis- including human heart failure. As we learn more about the con-
tent current.82,105,106 Furthermore, CaMKII-dependent regulation stituency, localization and function of specific Nav macromolecu-
of Nav1.5 depends on direct interaction with the actin-associated lar complexes within the cardiomyocyte, we anticipate the
polypeptide βIV-spectrin, which targets CaMKII with Nav1.5 to discovery of new therapeutic targets and strategies for preventing
the cardiomyocyte intercalated disc.82 Loss of spectrin/CaMKIIδ arrhythmias and improving heart function in human heart disease
interaction disrupts CaMKII regulation of Nav1.5 and abnormal patients.
cell membrane excitability. Future studies should determine the
broader role for βIV-spectrin in coordinating Nav macromolecu-
lar complexes at the cardiomyocyte intercalated disc and its role
in arrhythmogenesis. Acknowledgments
This work was supported by the National Institutes of Health
grants HL084583 and HL083422 (to P.J.M.) and HL096805 (to
Summary and Future Directions T.J.H.), the Pew Scholars Trust (to P.J.M.), the American Heart
Association (to P.J.M.), the Saving Tiny Hearts Society (to
The vertebrate heart has evolved, highly specialized pathways for P.J.M.), the Gilead Sciences Research Scholars Program (to
targeting and regulation of Nav channels, reflecting their central T.J.H.), and the Fondation Leducq Award to the Alliance for
role in control of cardiac excitation-contraction at baseline and Calmodulin Kinase Signaling in Heart Disease (to P.J.M.).
ventricular myocytes from mouse heart. Circula- 24. Payandeh J, Gamal El-Din TM, Scheuer T, et al:
References tion 109:1421–1427, 2004. Crystal structure of a voltage-gated sodium
13. Maier SK, Westenbroek RE, Schenkman KA, et channel in two potentially inactivated states.
1. Nattel S, Maguy A, Le Bouter S, et al: Arrhyth- al: An unexpected role for brain-type sodium Nature 486:135–139, 2012.
mogenic ion-channel remodeling in the heart: channels in coupling of cell surface depolarization 25. Payandeh J, Scheuer T, Zheng N, et al: The
heart failure, myocardial infarction, and atrial to contraction in the heart. Proc Natl Acad Sci crystal structure of a voltage-gated sodium
fibrillation. Physiol Rev 87:425–456, 2007. U S A 99:4073–4078, 2002. channel. Nature 475:353–358, 2011.
2. Wilde AA, Brugada R: Phenotypical manifesta- 14. Moss AJ, Kass RS: Long QT syndrome: from 26. Catterall WA: Ion channel voltage sensors: struc-
tions of mutations in the genes encoding subunits channels to cardiac arrhythmias. J Clin Invest ture, function, and pathophysiology. Neuron
of the cardiac sodium channel. Circ Res 108:884– 115:2018–2024, 2005. 67:915–928, 2010.
897, 2011. 15. Napolitano C, Bloise R, Monteforte N, et al: 27. Vassilev PM, Scheuer T, Catterall WA: Identifica-
3. Waxman SG: Channel, neuronal and clinical Sudden cardiac death and genetic ion channelopa- tion of an intracellular peptide segment involved
function in sodium channelopathies: from geno- thies: long QT, Brugada, short QT, catecholamin- in sodium channel inactivation. Science 241:1658–
type to phenotype. Nat Neurosci 10:405–409, ergic polymorphic ventricular tachycardia, and 1661, 1988.
2007. idiopathic ventricular fibrillation. Circulation 28. West JW, Patton DE, Scheuer T, et al: A cluster
4. Rook MB, Evers MM, Vos MA, et al: Biology of 125:2027–2034, 2012. of hydrophobic amino acid residues required for
cardiac sodium channel Nav1.5 expression. Car- 16. Abriel H, Kass RS: Regulation of the voltage- fast Na(+)-channel inactivation. Proc Natl Acad
diovasc Res 93:12–23, 2012. gated cardiac sodium channel Nav1.5 by interact- Sci U S A 89:10910–10914, 1992.
5. Abriel H: Cardiac sodium channel Na(v)1.5 and ing proteins. Trends Cardiovasc Med 15:35–40, 29. Cha A, Ruben PC, George AL Jr, et al: Voltage
interacting proteins: Physiology and pathophysi- 2005. sensors in domains III and IV, but not I and II,
ology. J Mol Cell Cardiol 48:2–11, 2010. 17. Wu L, Yong SL, Fan C, et al: Identification of a are immobilized by Na+ channel fast inactivation.
6. Zakon HH: Adaptive evolution of voltage-gated new co-factor, MOG1, required for the full func- Neuron 22:73–87, 1999.
sodium channels: The first 800 million years. tion of cardiac sodium channel Nav 1.5. J Biol 30. Brugada P, Brugada J: Right bundle branch block,
Proc Natl Acad Sci U S A 109(Suppl 1):10619– Chem 283:6968–6978, 2008. persistent ST segment elevation and sudden
10625, 2012. 18. Lemaillet G, Walker B, Lambert S: Identification cardiac death: a distinct clinical and electrocardio-
7. Wang Q, Li Z, Shen J, et al: Genomic organiza- of a conserved ankyrin-binding motif in the graphic syndrome. A multicenter report. J Am
tion of the human SCN5A gene encoding the family of sodium channel alpha subunits. J Biol Coll Cardiol 20:1391–1396, 1992.
cardiac sodium channel. Genomics 34:9–16, Chem 278:27333–27339, 2003. 31. Antzelevitch C, Brugada P, Brugada J, et al:
1996. 19. Lowe JS, Palygin O, Bhasin N, et al: Voltage- Brugada syndrome: a decade of progress. Circ Res
8. Schroeter A, Walzik S, Blechschmidt S, et al: gated Nav channel targeting in the heart requires 91:1114–1118, 2002.
Structure and function of splice variants of the an ankyrin-G dependent cellular pathway. J Cell 32. Chen Q, Kirsch GE, Zhang D, et al: Genetic
cardiac voltage-gated sodium channel Na(v)1.5. Biol 180:173–186, 2008. basis and molecular mechanism for idiopathic
J Mol Cell Cardiol 49:16–24, 2010. 20. Mohler PJ, Rivolta I, Napolitano C, et al: Nav1.5 ventricular fibrillation. Nature 392:293–296,
9. Shang LL, Dudley SC Jr: Tandem promoters E1053K mutation causing Brugada syndrome 1998.
and developmentally regulated 5’- and 3’-mRNA blocks binding to ankyrin-G and expression of 33. Dumaine R, Towbin JA, Brugada P, et al: Ionic
untranslated regions of the mouse Scn5a cardiac Nav1.5 on the surface of cardiomyocytes. Proc mechanisms responsible for the electrocardio-
sodium channel. J Biol Chem 280:933–940, 2005. Natl Acad Sci U S A 101:17533–17538, 2004. graphic phenotype of the Brugada syndrome are
10. Shang LL, Pfahnl AE, Sanyal S, et al: Human 21. Abriel H, Kamynina E, Horisberger JD, et al: temperature dependent. Circ Res 85:803–809,
heart failure is associated with abnormal Regulation of the cardiac voltage-gated Na+ 1999.
C-terminal splicing variants in the cardiac sodium channel (H1) by the ubiquitin-protein ligase 34. Yan GX, Antzelevitch C: Cellular basis for the
channel. Circ Res 101:1146–1154, 2007. Nedd4. FEBS Lett 466:377–380, 2000. Brugada syndrome and other mechanisms of
11. Walzik S, Schroeter A, Benndorf K, et al: Alterna- 22. van Bemmelen MX, Rougier JS, Gavillet B, et al: arrhythmogenesis associated with ST-segment
tive splicing of the cardiac sodium channel creates Cardiac voltage-gated sodium channel Nav1.5 is elevation. Circulation 100:1660–1666, 1999.
multiple variants of mutant T1620K channels. regulated by Nedd4-2 mediated ubiquitination. 35. Gima K, Rudy Y: Ionic current basis of electro-
PLoS One 6:e19188, 2011. Circ Res 95:284–291, 2004. cardiographic waveforms: a model study. Circ Res
12. Maier SK, Westenbroek RE, McCormick KA, 23. Catterall WA: Voltage-gated sodium channels at 90:889–896, 2002.
et al: Distinct subcellular localization of different 60: Structure, function, and pathophysiology. 36. Tan HL, Bink-Boelkens MT, Bezzina CR, et al:
sodium channel alpha and beta subunits in single J Physiol 590:2557–2589, 2012. A sodium-channel mutation causes isolated
cardiac conduction disease. Nature 409:1043– canine heart. Alterations in Vmax and the transient 75. Valdivia CR, Medeiros-Domingo A, Ye B, et al:
1047, 2001. outward current. Circulation 85:1175–1188, Loss-of-function mutation of the SCN3B-
37. Wang Q, Shen J, Li Z, et al: Cardiac sodium 1992. encoded sodium channel β3 subunit associated
channel mutations in patients with long QT syn- 56. Cabo C, Boyden P: Electrical remodeling of the with a case of idiopathic ventricular fibrillation.
drome, an inherited cardiac arrhythmia. Hum epicardial border zone in the canine infarcted Cardiovasc Res 86:392–400, 2010.
Mol Genet 4:1603–1607, 1995. heart: a computational analysis. Am J Physiol 76. Watanabe H, Koopmann TT, Le Scouarnec S,
38. Wang Q, Shen J, Splawski I, et al: SCN5A muta- Heart Circ Physiol 284:H372–H384, 2003. et al: Sodium channel beta1 subunit mutations
tions associated with an inherited cardiac arrhyth- 57. Christensen MD, Dun W, Boyden PA, et al: associated with Brugada syndrome and cardiac
mia, long QT syndrome. Cell 80:805–811, 1995. Oxidized calmodulin kinase II regulates conduc- conduction disease in humans. J Clin Invest
39. Clancy CE, Rudy Y: Linking a genetic defect to tion following myocardial infarction: A computa- 118:2260–2268, 2008.
its cellular phenotype in a cardiac arrhythmia. tional analysis. PLoS Comput Biol 5:e1000583, 77. Bhasin N, Cunha SR, Mduannayake M, et al:
Nature 400:566–569, 1999. 2009. Molecular basis for PP2A regulatory subunit B56
40. Bezzina C, Veldkamp MW, van Den Berg MP, 58. Decker KF, Heijman J, Silva JR, et al: Properties alpha targeting in cardiomyocytes. Am J Physiol
et al: A single Na+ channel mutation causing both and ionic mechanisms of action potential adapta- Heart Circ Physiol 293:H109–H119, 2007.
long-QT and Brugada syndromes. Circ Res tion, restitution and accommodation in canine 78. Hashemi SM, Hund TJ, Mohler PJ: Cardiac
85:1206–1213, 1999. epicardium. Am J Physiol Heart Circ Physiol ankyrins in health and disease. J Mol Cell Cardiol
41. Clancy CE, Rudy Y: Na+ channel mutation that 296:H1017–1026, 2009. 47:203–209, 2009.
causes both Brugada and long-QT syndrome phe- 59. Starmer CF, Colatsky TJ, Grant AO: What 79. Cunha SR, Bhasin N, Mohler PJ: Targeting and
notypes: a simulation study of mechanism. Circu- happens when cardiac Na channels lose their stability of Na/Ca exchanger 1 in cardiomyocytes
lation 105:1208–1213, 2002. function? 1—numerical studies of the vulnerable requires direct interaction with the membrane
42. Valdivia CR, Chu WW, Pu J, et al: Increased late period in tissue expressing mutant channels. Car- adaptor ankyrin-B. J Biol Chem 282:4875–4883,
sodium current in myocytes from a canine heart diovasc Res 57:82–91, 2003. 2007.
failure model and from failing human heart. J Mol 60. Fukuda K, Davies SS, Nakajima T, et al: Oxidative 80. Mohler PJ, Davis JQ, Bennett V: Ankyrin-B coor-
Cell Cardiol 38:475–483, 2005. mediated lipid peroxidation recapitulates proar- dinates the Na/K ATPase, Na/Ca exchanger, and
43. Jacques D, Bkaily G, Jasmin G, et al: Early fetal rhythmic effects on cardiac sodium channels. Circ InsP3 receptor in a cardiac T-tubule/SR microdo-
like slow Na+ current in heart cells of cardiomyo- Res 97:1262–1269, 2005. main. PLoS Biol 3:e423, 2005.
pathic hamster. Mol Cell Biochem 176:249–256, 61. Lau DH, Clausen C, Sosunov EA, et al: Epicar- 81. Garrido JJ, Giraud P, Carlier E, et al: A targeting
1997. dial border zone overexpression of skeletal muscle motif involved in sodium channel clustering at the
44. Maltsev VA, Sabbah HN, Higgins RS, et al: sodium channel SkM1 normalizes activation, pre- axonal initial segment. Science 300:2091–2094,
Novel, ultraslow inactivating sodium current in serves conduction, and suppresses ventricular 2003.
human ventricular cardiomyocytes. Circulation arrhythmia: an in silico, in vivo, in vitro study. 82. Hund TJ, Koval OM, Li J, et al: A betaIV
98:2545–2552, 1998. Circulation 119:19–27, 2009. spectrin/CaMKII signaling complex is essential
45. Undrovinas AI, Maltsev VA, Sabbah HN: Repo- 62. Mohler PJ, Schott JJ, Gramolini AO, et al: for membrane excitability in mice. J Clin Invest
larization abnormalities in cardiomyocytes of Ankyrin-B mutation causes type 4 long-QT 120:3508–3519, 2010.
dogs with chronic heart failure: role of sustained cardiac arrhythmia and sudden cardiac death. 83. Sato PY, Coombs W, Lin X, et al: Interactions
inward current. Cell Mol Life Sci 55:494–505, Nature 421:634–639, 2003. between ankyrin-g, plakophilin-2, and con-
1999. 63. Chen L, Marquardt ML, Tester DJ, et al: Muta- nexin43 at the cardiac intercalated disc. Circ Res
46. Maltsev VA, Silverman N, Sabbah HN, et al: tion of an A-kinase-anchoring protein causes 109:193–201, 2011.
Chronic heart failure slows late sodium current in long-QT syndrome. Proc Natl Acad Sci U S A 84. Hund TJ, Wright PJ, Dun W, et al: Regulation
human and canine ventricular myocytes: implica- 104:20990–20995, 2007. of the ankyrin-B-based targeting pathway follow-
tions for repolarization variability. Eur J Heart 64. London B, Michalec M, Mehdi H, et al: Mutation ing myocardial infarction. Cardiovasc Res
Fail 9:219–227, 2007. in glycerol-3-phosphate dehydrogenase 1 like 81:742–749, 2009.
47. Bers DM, Despa S, Bossuyt J: Regulation of Ca2+ gene (GPD1-L) decreases cardiac Na+ current 85. Le Scouarnec S, Bhasin N, Vieyres C, et al: Dys-
and Na+ in normal and failing cardiac myocytes. and causes inherited arrhythmias. Circulation function in ankyrin-B-dependent ion channel and
Ann N Y Acad Sci 1080:165–177, 2006. 116:2260–2268, 2007. transporter targeting causes human sinus node
48. Undrovinas NA, Maltsev VA, Belardinelli L, et al: 65. Vatta M, Ackerman MJ, Ye B, et al: Mutant disease. Proc Natl Acad Sci U S A 105:15617–
Late sodium current contributes to diastolic caveolin-3 induces persistent late sodium current 15622, 2008.
cell Ca2+ accumulation in chronic heart failure. and is associated with long-QT syndrome. Circu- 86. Cunha SR, Hund TJ, Hashemi S, et al: Defects
J Physiol Sci 60:245–257, 2010. lation 114:2104–2112, 2006. in ankyrin-based membrane protein targeting
49. Jacobshagen C, Belardinelli L, Hasenfuss G, et al: 66. Cronk LB, Ye B, Kaku T, et al: Novel mechanism pathways underlie atrial fibrillation. Circulation
Ranolazine for the treatment of heart failure with for sudden infant death syndrome: persistent 124:1212–1222, 2011.
preserved ejection fraction: background, aims, late sodium current secondary to mutations in 87. Petitprez S, Zmoos AF, Ogrodnik J, et al: SAP97
and design of the RALI-DHF study. Clin Cardiol caveolin-3. Heart Rhythm 4:161–166, 2007. and dystrophin macromolecular complexes deter-
34:426–432, 2011. 67. Medeiros-Domingo A, Kaku T, Tester DJ, et al: mine two pools of cardiac sodium channels
50. Scirica BM, Morrow DA, Hod H, et al: Effect of SCN4B-encoded sodium channel beta4 subunit Nav1.5 in cardiomyocytes. Circ Res 108:294–304,
ranolazine, an antianginal agent with novel elec- in congenital long-QT syndrome. Circulation 2011.
trophysiological properties, on the incidence of 116:134–142, 2007. 88. Gavillet B, Rougier JS, Domenighetti AA, et al:
arrhythmias in patients with non ST-segment 68. Abbott GW, Sesti F, Splawski I, et al: MiRP1 Cardiac sodium channel Nav1.5 is regulated
elevation acute coronary syndrome: results from forms IKr potassium channels with HERG and is by a multiprotein complex composed of syntro-
the Metabolic Efficiency With Ranolazine for associated with cardiac arrhythmia. Cell 97:175– phins and dystrophin. Circ Res 99:407–414,
Less Ischemia in Non ST-Elevation Acute Coro- 187, 1999. 2006.
nary Syndrome Thrombolysis in Myocardial 69. Splawski I, Tristani-Firouzi M, Lehmann M, et al: 89. Gee SH, Madhavan R, Levinson SR, et al: Inter-
Infarction 36 (MERLIN-TIMI 36) randomized Mutations in the hminK gene cause long QT syn- action of muscle and brain sodium channels with
controlled trial. Circulation 116:1647–1652, drome and suppress IKs function. Nat Genet multiple members of the syntrophin family of
2007. 17:338–340, 1997. dystrophin-associated proteins. J Neurosci
51. Song Y, Shryock JC, Wu L, et al: Antagonism by 70. Lin X, Liu N, Lu J, et al: Subcellular heterogene- 18:128–137, 1998.
ranolazine of the pro-arrhythmic effects of ity of sodium current properties in adult cardiac 90. Ueda K, Valdivia C, Medeiros-Domingo A, et al:
increasing late INa in guinea pig ventricular myo- ventricular myocytes. Heart Rhythm 8:1923– Syntrophin mutation associated with long QT
cytes. J Cardiovasc Pharmacol 44:192–199, 2004. 1930, 2011. syndrome through activation of the nNOS-
52. El-Sherif N, Scherlag BJ, Lazzara R, et al: 71. Kucera JP, Rohr S, Rudy Y: Localization of SCN5A macromolecular complex. Proc Natl
Re-entrant ventricular arrhythmias in the late sodium channels in intercalated disks modulates Acad Sci U S A 105:9355–9360, 2008.
myocardial infarction period. 1. Conduction char- cardiac conduction. Circ Res 91:1176–1182, 91. Wu G, Ai T, Kim JJ, et al: alpha-1-syntrophin
acteristics in the infarction zone. Circulation 2002. mutation and the long-QT syndrome: a disease of
55:686–702, 1977. 72. Mori Y, Fishman GI, Peskin CS: Ephaptic con- sodium channel disruption. Circ Arrhythm Elec-
53. Baba S, Dun W, Cabo C, et al: Remodeling in duction in a cardiac strand model with 3D electrophysiol 1:193–201, 2008.
cells from different regions of the reentrant trodiffusion. Proc Natl Acad Sci U S A 92. Cheng J, Van Norstrand DW, Medeiros-
circuit during ventricular tachycardia. Circulation 105:6463–6468, 2008. Domingo A, et al: Alpha1-syntrophin mutations
112:2386–2396, 2005. 73. Meadows LS, Isom LL: Sodium channels as mac- identified in sudden infant death syndrome cause
54. Pu J, Boyden PA: Alterations of Na+ currents in romolecular complexes: implications for inherited an increase in late cardiac sodium current. Circ
myocytes from epicardial border zone of the arrhythmia syndromes. Cardiovasc Res 67:448– Arrhythm Electrophysiol 2:667–676, 2009.
infarcted heart. A possible ionic mechanism for 458, 2005. 93. Liu N, Ruan Y, Denegri M, et al: Calmodulin
reduced excitability and postrepolarization refrac- 74. Watanabe H, Darbar D, Kaiser DW, et al: Muta- kinase II inhibition prevents arrhythmias in
toriness. Circ Res 81:110–119, 1997. tions in sodium channel beta1- and beta2-subunits RyR2(R4496C+/-) mice with catecholaminergic
55. Lue WM, Boyden PA: Abnormal electrical prop- associated with atrial fibrillation. Circ Arrhythm polymorphic ventricular tachycardia. J Mol Cell
erties of myocytes from chronically infarcted Electrophysiol 2:268–275, 2009. Cardiol 50:214–222, 2011.
94. Ai X, Curran JW, Shannon TR, et al: Ca2+/ the alpha subunit of the cardiac sodium channel. 111. Besse IM, Mitchell CC, Hund TJ, et al: A com-
calmodulin-dependent protein kinase modulates J Biol Chem 271:28837–28843, 1996. putational investigation of cardiac caveolae as a
cardiac ryanodine receptor phosphorylation and
sarcoplasmic reticulum Ca2+ leak in heart failure.
Circ Res 97:1314–1322, 2005.
103. Yoon JY, Ho WK, Kim ST, et al: Constitutive
CaMKII activity regulates Na+ channel in rat
ventricular myocytes. J Mol Cell Cardiol 47:475–
source of persistent sodium current. Front Physiol
2:87, 2011.
112. Deschenes I, Neyroud N, DiSilvestre D, et al:
9
95. Bers DM: Altered cardiac myocyte Ca regulation 484, 2009. Isoform-specific modulation of voltage-gated Na+
in heart failure. Physiology (Bethesda) 21:380– 104. Wagner S, Dybkova N, Rasenack EC, et al: Ca/ channels by calmodulin. Circ Res 90:E49–57,
387, 2006. calmodulin-dependent protein kinase II regulates 2002.
96. Qu Y, Rogers JC, Tanada TN, et al: Phosphoryla- cardiac Na channels. J Clin Invest 116:3127– 113. Tan HL, Kupershmidt S, Zhang R, et al: A
tion of S1505 in the cardiac Na+ channel 3138, 2006. calcium sensor in the sodium channel modulates
inactivation gate is required for modulation by 105. Aiba T, Hesketh GG, Liu T, et al: Na+ channel cardiac excitability. Nature 415:442–447, 2002.
protein kinase C. J Gen Physiol 108:375–379, regulation by Ca2+/calmodulin and Ca2+/ 114. Potet F, Chagot B, Anghelescu M, et al: Func-
1996. calmodulin-dependent protein kinase II in tional interactions between distinct sodium
97. Qu Y, Rogers J, Tanada T, et al: Modulation of guinea-pig ventricular myocytes. Cardiovasc Res channel cytoplasmic domains through the action
cardiac Na+ channels expressed in a mammalian 85:454–463, 2010. of calmodulin. J Biol Chem 284:8846–8854,
cell line and in ventricular myocytes by protein 106. Ashpole NM, Herren AW, Ginsburg KS, et al: 2009.
kinase C. Proc Natl Acad Sci U S A 91:3289– Ca2+/calmodulin-dependent protein kinase II 115. Shah VN, Wingo TL, Weiss KL, et al: Calcium-
3293, 1994. (CaMKII) regulates cardiac sodium channel dependent regulation of the voltage-gated sodium
98. Murray KT, Hu NN, Daw JR, et al: Functional NaV1.5 gating by multiple phosphorylation sites. channel hH1: intrinsic and extrinsic sensors use a
effects of protein kinase C activation on the J Biol Chem 287:19856–19869, 2012. common molecular switch. Proc Natl Acad Sci
human cardiac Na+ channel. Circ Res 80:370–376, 107. Ufret-Vincenty CA, Baro DJ, Lederer WJ, et al: U S A 103:3592–3597, 2006.
1997. Role of sodium channel deglycosylation in the 116. Kim J, Ghosh S, Liu H, et al: Calmodulin medi-
99. Frohnwieser B, Chen LQ, Schreibmayer W, et al: genesis of cardiac arrhythmias in heart failure. ates Ca2+ sensitivity of sodium channels. J Biol
Modulation of the human cardiac sodium channel J Biol Chem 276:28197–28203, 2001. Chem 279:45004–45012, 2004.
alpha-subunit by cAMP-dependent protein kinase 108. Schreibmayer W, Frohnwieser B, Dascal N, et al: 117. Young KA, Caldwell JH: Modulation of skeletal
and the responsible sequence domain. J Physiol Beta-adrenergic modulation of currents produced and cardiac voltage-gated sodium channels by
498(Pt 2):309–318, 1997. by rat cardiac Na+ channels expressed in Xenopus calmodulin. J Physiol 565:349–370, 2005.
100. Zhou J, Yi J, Hu N, et al: Activation of protein laevis oocytes. Recept Channels 2:339–350, 118. Sarhan MF, Van Petegem F, Ahern CA: A double
kinase A modulates trafficking of the human 1994. tyrosine motif in the cardiac sodium channel
cardiac sodium channel in Xenopus oocytes. Circ 109. Yarbrough TL, Lu T, Lee HC, et al: Localization domain III-IV linker couples calcium-dependent
Res 87:33–38, 2000. of cardiac sodium channels in caveolin-rich mem- calmodulin binding to inactivation gating. J Biol
101. Lu T, Lee HC, Kabat JA, et al: Modulation of rat brane domains: regulation of sodium current Chem 284:33265–33274, 2009.
cardiac sodium channel by the stimulatory G amplitude. Circ Res 90:443–449, 2002. 119. Wang C, Chung BC, Yan H, et al: Crystal struc-
protein alpha subunit. J Physiol 518(Pt 2):371– 110. Palygin OA, Pettus JM, Shibata EF: Regulation ture of the ternary complex of a NaV C-terminal
384, 1999. of caveolar cardiac sodium current by a single domain, a fibroblast growth factor homologous
102. Murphy BJ, Rogers J, Perdichizzi AP, et al: Gsalpha histidine residue. Am J Physiol Heart factor, and calmodulin. Structure 20:1167–1176,
cAMP-dependent phosphorylation of two sites in Circ Physiol 294:H1693–H1699, 2008. 2012.
Regulation of Cardiac
Calcium Channels 10
Jonathan Satin
superimposed effects of permeating cations, as well as modulation

CHAPTER OUTLINE
of gating primarily by channel phosphorylation status, must be
Overview 103 considered. Finally, the cardiac-calcium channel is a heteromul-
timeric protein complex. Accessory proteins and interacting pro-
The Cardiac L-Type Calcium Channel Is a
teins either directly or via scaffolding proteins modify cardiac
Multiprotein Complex 104 calcium–channel gating. Additional details on L-type Ca2+
Voltage Effects on Activation and Inactivation 107 channel in the heart can be found in Chapter 2, focus on
excitation-contraction coupling is covered in Chapter 16,
Calcium Regulates Activation and Inactivation Gating 108 β-adrenergic regulation of cardiac function is reviewed in Chapter
Posttranslational Modifications Effects on 19, and a more in-depth focus on Timothy syndrome is presented
Ca-Channel Gating 108 in Chapter 94.
Other Protein Interactions Affecting L-Type
Ca2+ Channel Gating 110 Calcium Channel Expression in the Myocardium
L-type versus T-type Channel Expression and Gating
The myocardium expresses L-type and T-type voltage-gated
Overview calcium channels. Myocardial L-type calcium channels include
CaV1.2 and CaV1.3. Mature ventricular myocardium almost
The super-family of voltage-gated ion channels is defined by the exclusively expresses the CaV1.2 channel. CaV1.2 and CaV1.3 are
common ability of the protein to sense transmembrane potential. expressed in atrial cardiomyocytes, and CaV1.3 is also expressed
Voltage-gated ion channels, specifically voltage-gated cation by sinoatrial6 and atrioventricular nodal cells.7 Commensurate
channels, share the general structural plan consisting of six with these tissue localizations, CaV1.2 is critically important for
α-helical transmembrane segments and a region of amino acids providing trigger calcium for excitation-contraction coupling,
between transmembrane segments 5 and 6 (S5 and S6) that fold and CaV1.3 contributes to heart rate and cardiac conduction.8,9
in from the extracellular space toward the cytosol to form the In the ventricular myocardium, there is an age- and gender-
outer permeation pathway. The fundamental common property dependent gradient of ICa,L. Prepubertal boys have elevated ICa,L
of voltage sensing is conferred by a transmembrane α-helical at the base, but developing girls do not. In adult rabbits, the
stretch of amino acids in the fourth transmembrane segment (S4). gradients are somewhat reversed, with females displaying higher
The purpose of this chapter is to describe the regulation of ICa,L at baseline compared with males,10 and this may be related
cardiac calcium–channel gating. Gating is not simply movement to estrogen regulation of L-type Ca2+ gene transcription.11
of S4 segments. Additional detail on calcium channels and cardiac T-type Ca2+ channels in the heart are encoded by CaV3.1 and
myocyte physiology can be gleaned from several excellent CaV3.2 pore-forming subunits.12,13 T-type Ca2+ current, ICa,T, is
reviews,1-3 as well as two outstanding books on ion channel not normally observed in the mature mammalian ventricular
biophysics.4,5 myocardium. ICa,T is present in pacemaker cells,14 atrial cells, and
Voltage-gated Ca2+ channels tend to open in response to depo- Purkinje fibers.15 ICa,T is also expressed in developing cardiomyo-
larization and tend to close after repolarization. Channel transit- cytes.13,16,17 Consistent with the adage that pathologic cardiac
ing from an open to a closed state is termed activation gating. hypertrophy is accompanied by reexpression of the fetal gene
Channel gating can be envisioned as a mechanical contraption program, ICa,T is reexpressed in ventricular hypertrophy in the
that allows ionic flow across a barrier (Figure 10-1). It requires cat18 and rat.19
some imagination to picture a complex molecule undergoing The gating properties of ICa,T versus ICa,L are fundamentally
gating. Consider one of the starting gate slots at the Kentucky distinct. T-type versus L-type channels are also classified as low
Derby. In this analogy, the stationary horse represents the poten- voltage–activated versus high voltage–activated, respectively. The
tial energy of a calcium ion affected by its electrochemical gradi- activation range of ICa,T is positive to approximately −60 mV,
ent. Once the gate opens, the calcium ion flows down its whereas ICa,L activates positive to approximately −20 mV. Closed-
electrochemical gradient. Although the ion may interact with the state inactivation also follows a similar general pattern. T-type
gate, and influence the gate, the gate is operated independently. Ca2+ channel steady-state availability is maximal for voltages less
Moreover, there are a series of sequential complex steps between than −90 mV, whereas substantial closed-state L-type channel
the starter pushing the button for the gates to open and the actual inactivation is not observed for potentials as positive as approxi-
gating event. Similarly, initial movement of the S4 in response to mately −40 mV. The positive shift of ICa,L availability (also known
voltage is transmitted in a complex, incompletely understood as steady-state inactivation) has a practical benefit for measuring
fashion to other domains of the channel that alter the permeation ionic current in ventricular cardiomyocytes as well. Voltage-gated
pathway, allowing ionic flux. Na+ channels are largely inactivated at −40 mV, thus a prepulse
Cardiac L-type Ca2+-channel gating is influenced by voltage, to −40 mV is frequently used to isolate ICa,L from prominent
Ca2+-ion posttranslational modifications and protein-protein overlapping INa. An alternative method used to measure ICa,L in
interactions. Voltage is a major determinant of gating, and ventricular cardiomyocytes is to replace external Na+ with an
103
Open gates/coupled gating

Ionic current
Voltage sensor Pore
Voltage + + + + – – – –
sensors
– – – – + + + +
Gate
Gating current
CLOSED OPEN
Ca2+
A Gates 1.0
0.8
Normalized G or Q
0.6
Q G
0.4
0.2
0.0
–100 –50 0 50
B V (mV)
Figure 10-1. The relationship among voltage sensing, ion channel gating, and conductance. A, Schematic shows the principal features of a generalized voltage-gated ion
channel and the metaphorical electric-gated Kentucky Derby starting gate. The α-helical transmembrane segments (S4) sense the electrical gradient. Motion of the S4
segments is transmitted allosterically through the channel protein to open the gate. Ions flux based on their electrochemical gradient. B, Conductance (G) and gating
charge movement (Q) depend on voltage for a voltage-dependent ion channel.
(B, Adapted from Bezanilla F: The voltage sensor in voltage-dependent ion channels. Physiol Rev 80:555–592, 2000.)
impermeant cation. ICa,L can then be measured using more nega- partners have been identified and studied (Figure 10-3). Early
tive holding potentials. In juvenile ventricular cardiomyocytes, electrophysiological studies in nonexcitable cells heterologously
hyperpolarizing the holding potential from −40 to −80 mV has expressing CaV1.2 alone showed little, if any, discernible ionic
only minor effects on peak ICa,L (Figure 10-2, A). By contrast, current. Coexpression of CaVβ subunits is a requirement to study
embryonic ventricular cardiomyocytes exhibit ICa,T and ICa,L. ICa,T L-type Ca current from channels generated by plasmids intro-
manifests itself as a current activating positive to −60 mV when duced into nonexcitable cells.20 CaVβ–CaV1.2 interactions occur
a −80 mV holding potential is used, and this low voltage–activated via a deep hydrophobic pocket of CaVβ complexed with the LI-II
Ca2+ current (ICa,T) is closed state–inactivated by holding the domain of CaV1.2.21 CaVβ–CaV1.2 interactions have gating and
potential at −40 mV (see Figure 10-2, B). Heterologous expres- nongating consequences. Early studies showed that CaVβ masks
sion systems transfected with pore-forming CaV3.x subunits an endoplasmic reticulum retention signal on CaV1.2, thus allow-
alone reconstitute most of the native ICa,T properties.13 In sharp ing CaV1.2 to traffic to the surface membrane.22 More recent
contrast, native ICa,L gating requires auxiliary proteins and in studies show that CaVβ–CaV1.2 binding influences a CaV1.2
some cases is rather difficult to completely recreate in heterolo- carboxyl-terminal rearrangement to promote surface expres-
gous expression systems. The remainder of this chapter will focus sion.23 Despite the relatively high-affinity CaV1.2–CaVβ interac-
on CaV1.2, the predominant L-type Ca2+ channel. tion (2-54 nM), dynamic CaVβ–CaV1.2 interaction may also occur
with respect to L-type Ca2+-channel gating,24 consistent with the
idea that CaVβ modifies CaV1.2 gating in a regulated fashion.
The CaV1.2 carboxyl-terminus spans approximately 300
The Cardiac L-type Calcium Channel amino acids (size differing among splice variants), from the cyto-
Is a Multiprotein Complex solic border of homologous repeat intravenous transmembrane
S6 until the termination of the protein. The distal carboxyl-
The pore-forming CaV1.2 Ca2+ channel does not gate in isolation terminus is proteolytically cleaved, yielding an approximate
in cardiac myocytes, and studies of ionic current in heterologous 37 kDa protein that covalently reassociates with the proximal
expression systems show that a functional ion channel complex carboxyl-terminus to regulate function25 (also discussed later).
requires auxiliary proteins. Although the total number of CaV1.2- The distal carboxyl-terminus also can localize to the nucleus,26
interacting proteins is unknown, several important interacting where it regulates gene transcription, including that for CaV1.2.27
Regulation of Cardiac Calcium Channels 105
1 Table 10-1. CaV1.2 Interacting Proteins

Voltage (mV)
–80 –60 –40 –20

0
0 20 40 60
Interacting
Protein Interaction site on CaV1.2 References
10
–1
CaV1.2 DCT, N-terminus 37-39
–2 CaVβ2 LI-II, PCT 40
Current density (pA/pF)

–3 CaV1.2–DCT PCT 25
α2δ Uncertain
–4
Calmodulin (CaM) IQ motif of PCT See text below
HP = –80 mV
–5
HP = –40 mV Calmodulin kinase PCT 35
HP = –80 mV –6 II (CaMKII)
(50 µM 2-APB) Rem Via CaVβ2, DCT, N-terminus 34; 41
HP = –40 mV –7
Rad Via CaVβ2, N-terminus 41
(50 µM 2-APB)
–8 Calcineurin (CaN = DCT 42; 43
PP2B)
A Juvenile cardiac ventricular myocytes
PP2A DCT 43
PDE4B CaV1.2 by ip, no subdomain 44
determined
1 Akap150/79 PCT–DCT scaffolds with 45
other proteins
Voltage (mV) PKA PCT–DCT via akap 46
0 α-Actinin PCT/DCT 47
–80 –60 –20 0 20 40 60
Sorcin PCT 48
LI-II, Cytosolic linker between homologous repeats I and II; DCT, CaV1.2 distal
–1
Current density (pA/pF)
carboxyl-terminus; PCT, CaV1.2 proximal carboxyl-terminus (defined as the

cytosolic region from termination of α-helical transmembrane segment IVS6
until the proteolytic cleavage site demarcating PCT and DCT); CaV1.2 by ip,
HP = –80 mV –2 association demonstrated by coimmunoprecipitation with CaV1.2.
HP = –40 mV
HP = –80 mV
(50 µM 2-APB)
–3 binding,30,31 and mutation of isoleucine of the IQ motif induces
HP = –40 mV dilated cardiomyopathy and premature death.32 Two CaM mol-
(50 µM 2-APB)
ecules interact with the proximal carboxyl-terminus of CaV1.2,
and the CaM molecules on CaV1.2 are arranged in an antiparallel
–4 fashion.31,33 The proximal carboxyl-terminus of CaV1.2 also inter-
B Embryonic cardiac ventricular myocytes acts with several other proteins that modify channel gating. The
monomeric G-protein Rem functionally competes with CaM for
Figure 10-2. L-type current (ICa,L) is the only discernible ICa in mature cardiac myo- channel regulation at this domain as well.34 CaMKII tethers to
cytes. ICa,L and T-type current are simultaneously functional in developing ventricu- the proximal carboxyl-terminus and is an important modulator
lar myocardium. Current-voltage curve for juvenile (1 to 2-month-old mice) (A) and
embryonic (B) ventricular myocytes. A, ICa,L without detectable ICa,T manifested as
of CaV1.2 activity.35 Currently, it is unclear how many proteins
equivalent peak current for holding potentials (Vhold) of −80 and −40 mV. B, ICa,T is combine to form the native L-type calcium channel complex. An
elicited from Vhold −80 mV and is steady-state inactivated at Vhold −40, resulting in unbiased proteomics screen of the closely related N-type calcium
no low-voltage–activated current detected upon depolarization. The shaded area channel, CaV2.2, revealed channel interactions with 207 pro-
indicates the ICa,T component; ICa,T also known as low-voltage-activated Ca2+ current teins.36 This suggests that multiple protein-protein interactions
is observed between −60 and −20 mV (blue shaded region in panel B). 2-APB does sum to yield native ICa,L properties. In fact, the number of inter-
not block CaV1.2 but does inhibit non–CaV1.2-Ca2+ current. The 2-APB sensitivity acting proteins for the CaV1.2 carboxyl-terminal domain exceeds
thus illustrates the complex mixture of T- and multiple L-type currents in the devel- the restricted space, suggesting that weak protein-protein inter-
oping myocardium, in contrast to ICa,L dominated by CaV1.2 in mature ventricular actions among multiple proteins create the possibility of diverse
cardiomyocytes.
mixtures of proteins for any given L-type Ca2+-channel complex.
(Adapted from Schroder E, Wei Y, Satin J: The developing cardiac myocyte: maturation Some of the known CaV1.2 interacting proteins are summarized
of excitability and excitation-contraction coupling. Ann N Y Acad Sci 1080:63–75, in Table 10-1.
2006.)
CaV1.2 Structure and Gating

The proximal carboxyl-terminus remains contiguous with the
pore-forming CaV1.2. Calmodulin (CaM) is prebound to CaV1.2 The defining functional feature of the super-family of voltage-
on the proximal carboxyl-terminus in the IQ motif28,29 and is a gated ion channels is the conserved property that depolarization
critical determinant of Ca2+-dependent inactivation gating (dis- tends to open channels. The process of opening can be summa-
cussed later). IQ is the single letter abbreviation for isoleucine rized by the term activation. According to classical ion channel
and glutamine. The “I” of the IQ motif is essential for CaM biophysics, CaV activation is a smooth curvilinear function of
α2 Figure 10-3. Subunit structure of L-type Ca2+–channel complex. The

s COOH upper panel depicts four homologous repeats of pore-forming CaV1.2
s subunit, CaVβ bound to LI-II (linker joining repeats I and II), and the
NH2 proximal carboxyl-terminus. The IQ motif is required for CaM binding.
D3 The distal carboxyl-terminus is shown as a proteolytically cleaved
D2 D4 protein. Antibodies raised against this segment show it migrating as
D1 an independent protein from the pore-forming region. The lower
δ
panel summarizes critical interacting proteins along with approxi-
Cav1.2
mate CaV1.2 interaction domains. Green boxes indicate regulators that
tend to increase ICa,L, yellow boxes denote more complex effects, and
β
Proximal carboxyl terminus red boxes inhibit ICa,L. Calcineurin (CaN) is denoted with a blue box
NH2 because of the controversial effects of CaN on ICa,L.
EF hand IQ
P C-term
COOH antibody Distal carboxyl terminus
Ser 1928
Schematic localization of protein binding to the Cavα subunit
Repeat I Repeat II Repeat III Repeat IV
N C
I-II linker II-III linker C-terminus
Cavβ subunit
Cavα2δ subunit
Calmodulin
AKAP-15
Calmodulin-kinase II
PP2a-cα
RGK ?
CCt
CaN (= PP2B)
voltage described by one or more Boltzmann distributions. Con- gating currents, channel blockade is applied or voltage is clamped
versely, repolarization closes (deactivates) channels. Deactivation to the reversal potential of the channel, resulting in no net ionic
gating is distinguished from inactivation gating. Deactivation flux. Gating current yields information of the number of active
relates a rapid, reversible transition between an ion-conducting channels, and the movement of the S4 segments, but gives
channel conformation and a nonconducting conformation. Inac- incomplete information for activation gating. To reiterate, activa-
tivation is a longer-lasting, nonconducting conformation that tion gating begins with voltage sensing (measured as gating
may be influenced by the position of the voltage sensors. As with current). Subsequent allosteric channel rearrangements result in
the closely related NaV channel family, CaV channels contain four activation gating. Gating current normalized to ionic current in
S4 segments that presumably displace toward the extracellular a given cell is a measure of coupling between voltage sensing and
space upon depolarization. This S4 displacement then drives allosteric rearrangements, resulting in channel opening.
allosteric rearrangements, resulting in an increase of channel Inactivation gating is not simply the reverse of activation
conductance. Collectively, the depolarization-dependent increase gating. There is no complete molecular structure data available
of channel conductance is referred to as activation gating. for voltage-gated Ca channels; however, voltage-gated Na chan-
CaV1.2 S4 segments each contain four to eight positive- nels have recently been crystallized in two potentially inactivated
charged amino acid residues (lysine or arginine) in register. states.49 These crystal structure studies support biophysical
Thus, when the transmembrane potential (Vm) is negative, the studies that suggest that inactivation gating consists of a series of
S4-positive charges are electrostatically drawn toward the cytosol. complex molecular motions, whereby the voltage sensing domains
Conversely, depolarization results in relative motion of S4 (S4) shift around the pore, two of the S6 segments are transposed
charges toward the extracellular space. The movement of charge extracellularly, and the other two S6 segments collapse on the
across an electric field creates a current. If no ionic flux occurs pore. Consequentially, the permeation pathway is reshaped.
and a depolarization is applied, S4 segments will move, generat- Drawing on the conservation of such broad structure-function
ing what is commonly called a gating current. To measure CaV1.2 models, it can be inferred that similarly complex motions impart
voltage-dependent inactivation to mammalian voltage-gated experimentally by replacing Ca2+ with an alternative charge
10
calcium channels. carrier—usually Ba2+.55 Resulting IBa,L has a significantly slower
The generalized structure of L-type calcium channels (CaV1.x) time course of decay. Overexpression of mutant calmodulin that
shares the voltage-gated ion channel super-family plan of six cannot bind Ca2+ results in Ca2+-current decay kinetics that match
α-helical transmembrane domains, arranged as four contiguous that of Ba2+ current,28,56 thus supporting the key contribution of
homologous repeats. There are obvious critically important dis- calmodulin to Ca2+-dependent inactivation (CDI). When one
tinctions in structure-function detail between CaV1.x and other evaluates the time course of Ca2+-channel current decay, there is
voltage-gated ion channels. The CaV1.x processed transcript an obvious faster decay for ICa,L than for IBa,L (Figure 10-4). This
encodes on the order of more than 2100 amino acids; the precise widely observed channel behavior leads to the common-sense
number depends on splice variant expression. By contrast to conclusion that voltage-dependent inactivation (VDI) is perhaps
T-type Ca2+ channels, and closely related NaV channels, CaV less important than CDI. However, such a relative unimportance
channels require various subunits to generate basal function. of VDI has been challenged. At first glance, current records such
Perhaps the single most critical class of subunits are the CaVβ— as those displayed in Figure 10-4 suggest that VDI is significantly
mainly CaVβ2 in the myocardium.50 CaV1.x and CaVβ form tight slower than CDI. However, in the absence of β-adrenergic recep-
interactions. Cytosolic CaVβ increases cell surface expression and tor stimulation, VDI is relatively fast.57-59
increases channel gating.51 L-type Ca2+-channel inactivation limits Ca2+ entry during the
Early crystallographic studies identified a hydrophobic groove cardiac action potential. Persistent activation, such as in Timothy
on CaVβ that confers CaV1 interactions.21,33,52 More recent work syndrome, has been linked to inappropriate VDI. Timothy syn-
sheds light on α–β interactions with respect to CaV1 structures drome is a monogenic, autosomal, dominant disease likely caused
and, in doing so, yields insight into L-type Ca2+-channel gating. by a missense mutation in CaV1.2.60 A glycine-to-arginine substi-
More than a decade ago, the cytoplasmic I-II linker (LI-II) of tution in the proximal LI-II domain results in defective VDI.
CaV1.2 was identified as essential for CaVβ interaction.53 Subdo- Patients with Timothy syndrome have a broad spectrum of
mains of the cytoplasmic LI-II are highly conserved among CaV1.x disorders, including cardiac arrhythmias, and the myocardial
channels and across species consistent with conservation of func- phenotypic changes are captured in induced pluripotent
tion. CaVβ–LI-II interaction may also contribute to gating. The
S6 segment of CaV1.2 is thought to form the inner permeation
pathway and lies adjacent to the proximal LI-II of CaV1.2. Thus,
crystallography data support the revised model that CaVβ binding
Cav1.2
transmits changes to inactivation gating of CaV1.2 via a partial
α-helical proximal LI-II segment.54 +10 mV
Voltage Effects on Activation and Inactivation

–90 1000 ms
Activation
Classical ion-channel biophysical descriptions use a Boltzmann
distribution to describe ion-channel activation and inactivation
gating.5 CaV1.2 is voltage gated. In other words, voltage deter-
Ca
mines gating; therefore, to study CaV1.2, gating voltage is con-
trolled. To measure steady-state activation gating, cells are
voltage-clamped at a relatively negative potential often approxi-
mating the diastolic potential of cardiomyocytes. Depolarizing
pulses are then typically used to determine the activation range
for macroscopic current, that is, whole-cell ionic current. Result-
ing current-voltage curves can then be transformed, considering Ba
the driving force as the difference between channel reversal
potential and applied potential, to yield a steady-state conductance-
voltage curve. Typically, cardiac ICa,L activates positive to about
−40 mV under conditions of physiologic concentrations of exter-
nal Ca2+. The steady-state activation-voltage range will vary with
species of permeant cation, for example, Ca2+ versus Ba2+, perme-
ant cation concentration, phosphorylation status of the channel 1 nA
complex, and perhaps even dynamic protein-protein interactions
with the heteromultimeric channel complex.
Inactivation Figure 10-4. Voltage-dependent and calcium-dependent inactivation. HEK 293

cells expressing CaV1.2 + CaVβ2a + α2δ recorded in the whole-cell configuration of
L-type Ca2+-channel inactivation gating is regulated by Ca2+ and the patch-clamp technique. Voltage stepped from Vhold −90 to +10 mV elicits mac-
voltage. Macroscopic (whole-cell) ICa,L evoked by a step depolar- roscopic currents that peak within 20 ms and then decay. Currents with external
ization to a constant depolarized potential evokes an inward Ca2+ solution containing 5 mM Ba2+ or 5 mM Ca2+ are superimposed and the peaks are
normalized. Note that conductance for IBa,L > ICa,L. IBa,L decay is a readout of VDI.
current that peaks relatively quickly and then decays with a time
CaVβ2a slows VDI, emphasizing the difference in decay kinetics to that for CDI. Rela-
course that is influenced by calcium, and to some extent by tively rapid ICa,L decay is dominated by CDI, albeit with VDI superimposed.
voltage. The calcium-dependent component of inactivation is
dominant during a step depolarization (discussed in the next (Data from Liang H, DeMaria CD, Erickson MG, et al: Unified mechanisms of Ca2+ regu-
subsection). The voltage-dependent component is uncovered lation across the Ca2+ channel family. Neuron 39:951–960, 2003.)
stem cell–derived cardiomyocytes.61 The slowed VDI gating in

Timothy syndrome Ca2+ channels is manifested as a slower decay
of IBa,L. In addition to explaining fundamental disease mechanism, 0
discovery of L-type Ca2+-channel involvement in Timothy NV 1234
V (mV)
WT 34
syndrome revealed the importance of VDI to cardiac APD
electrophysiology. 12
Inactivation can also be measured by evaluating channel con- –80
formation at steady state. Regardless, it is imperative to eliminate
Ca2+ flux to experimentally separate CDI from VDI. Several 0 1 2
A t (sec)
manipulations have been performed, and each presents con-
founding factors for data interpretation. Equimolar replacement NV (8)
of Ca2+ with Ba2+ has been widely used. Some concern that Ba2+ WT (5)
weakly interacts with CaM motivated the use of monovalent
1234 (5)
cation flux to measure channel availability. Monovalent flux mea-
sured by removal and chelation of divalent cations yields nonse-
lective current with inactivation that is independent of current 0 1 2
B APD (sec)
flux amplitude.62 However, the inactivation of L-type Ca2+ No virus CaM1234 + GFP
channel in the absence of divalent cations is significantly different
–80 +10 mV
from that measured in the presence of divalent cations.55 Divalent
cations interacting with the permeation pathway likely alter Ca
gating behavior,63 even before barium ions go through the Ba 1 nA
channel.64,65 Thus, VDI is not purely voltage dependent in the
100 msec t = 0.371 ± 0.008 t = 0.057 ± 0.037
sense that permeating Ca2+ interacting with the selectivity filters (n = 3) (n = 5)
may influence gating. The distinction between VDI and CDI is C
that in VDI, ionic flux does not contribute to inactivation,
whereas in CDI, Ca2+ flux necessarily interacts with prebound
cytosolic calmodulin and thus influences inactivation in concert NV +DHP
with voltage. 0
NV (11) –DHP NV
mV
1234 (8) +DHP
1234
Calcium Regulates Activation and +DHP –80
.2 s .2 s
Inactivation Gating 0 1 2
2+
Calcium is an important modifier of cardiac L-type Ca -channel D APD (sec)
gating. Calmodulin bound to the proximal carboxyl-terminus of Figure 10-5. Ventricular cardiomyocyte action potential duration (APD) is regu-
the calcium channel30,66 senses calcium ion fluxed through the lated by CDI via ICa,L-CaM modulation. A, Action potentials superimposed for rep-
channel56,67 and Ca2+ ion in the cytosol. In turn, Ca2+-CaM- resentative cardiomyocytes expressing exogenous CaM (wt), no virus, or CaM
CaV1.2-complex imparts a relatively rapid CDI (Figure 10-5).68 mutants. 12 and 34 denote CaM mutants that are unable to bind Ca2+ at the N- and
CaM modulation of ICa,L targets the plateau phase of the action C-termini of CaM. 1234 represents apoCaM. CaM1234 and CaM34 cause ultralong APD
consistent with the Ca2+ fluxing through the channel-regulating APD via Ca-CaM–
potential. This was shown in an elegant series of experiments dependent inactivation. B, Summary data show APD prolonged by apoCaM. C, ICa,L
exemplified by the results shown in Figure 10-5.69 Engineered and IBa,L superimposed in the absence of virus and in the presence of adenoviral-
CaM with its divalent cation sites mutated was introduced into mediated CaM1234 expression. CaM1234 completely eliminates CDI. D, Demonstration
cardiomyocytes. The resulting ICa,L showed significantly slowed that CaM1234 ultralong APD are dependent on ICa,L. Dihydropyridine, L-type Ca2+
decay, and action potential duration became ultralong (see Figure channel blockade, shortens APD in no virus (NV) control and reduces ultralong APD
10-5). Moreover, prolongation of the APD by divalent-cation– in CaM1234-infected cardiomyocytes.
free CaM is reversed by L-type calcium channel blockade with
(Adapted from Alseikhan BA, DeMaria CD, Colecraft HM, Yue DT: Engineered calmodu-
dihydropyridine.69
lins reveal the unexpected eminence of Ca2+ channel inactivation in controlling heart
At the single ion–channel level, Ca2+ CDI is caused by a
excitation. Proc Nat Acad Sci U S A 99:17185–17190, 2002.)
decreased frequency of channel reopenings and a decrease in
mean open time.70 Reopening rate is governed by an absorbing
state from which the channel cannot open, whereas the decrease
of mean open time reflects a faster Ca2+-driven closure of open gating is a sensitive caliper for subdomain cardiomyocyte cyto-
channels. These single-channel studies showed that CDI contrib- solic Ca2+.
utes to slow macroscopic decay, despite earlier findings separating
VDI from CDI based on slow versus fast macroscopic decay.
L-type Ca2+ channels are organized in the junctional mem-
brane in close opposition to ryanodine receptors. Colocalized Posttranslational Modification Effects
ryanodine receptors are present with a four- to tenfold excess to on Ca-Channel Gating
the number of L-type Ca2+ channels.71 Therefore, Ca2+-induced
Ca2+ release results in an amplified local elevation of subdomain β-Adrenergic–Modulated Ca-Channel Gating
Ca2+, with physiologically relevant gating consequences for
L-type Ca2+ channels. L-type Ca2+ channels initiate sarcoplasmic
reticulum (SR) Ca2+ release, and in turn this SR Ca2+ release β-Adrenergic receptor (β-AR)–stimulated modulation of ICa,L is a
promotes L-type Ca2+-channel inactivation, thus creating a clas- major contributor to the cardiac response to increased sympa-
sical physiological negative feedback loop. Experimentally, the thetic tone. β-AR agonists increase macroscopic Ca2+-channel
ability of SR Ca2+ release to inactivate ICa,L allows the use of ICa,L conductance and shift the current voltage relationship to more
as a reporter for SR Ca2+ release.72 In this way, L-type–channel negative potentials. The activation gating shift has a large effect
1.0 Basal – no phosphorylation β - adrenergic signaling –
10
DCT rearrangement
Normalized conductance 0.8
0.6
PCT
0.4
PCT
akap
0.2
DCT DCT
0.0
–40 –20 0 20
Membrane potential (mV) Figure 10-7. L-type Ca2+–channel gating at basal phosphorylation status and after
Figure 10-6. Steady-state activation gating shift of voltage dependence is a potent β-adrenergic stimulation. Left, Basal channel state assumes dephosphorylation of
mechanism for increase of ICa,L. Single Boltzmann distributions are drawn depicting proximal and distal carboxyl-terminal substrates for PKA. The key feature is that at
basal (red) and β-adrenergic–stimulated (blue) L-type Ca2+ channels. For the basal the basal state, the distal carboxyl-terminus (DCT) interacts with the proximal
state, the midpoint of activation is 0, and a −10 mV shift simulates β-adrenergic carboxyl terminus (PCT), perhaps aided by the akap scaffold. In the basal state, DCT
stimulation. All else being equal, the shift of the steady-state activation curve is autoinhibitory for ICa,L. Right, β-adrenergic–stimulated channels. The large P on
increases channel conductance at 0 mV from 50% to more than 90% (vertical the DCT is not required for modulation but is a conserved PKA substrate. The large
arrow). Boltzmann distribution of the form: Gmax / (1 + exp(V½ − V) / k), where Gmax P on the PCT represents two or more identified PKA substrate sites of this domain.
is maximal conductance, Erev is reversal potential, V½ is activation midpoint poten- In this model, phosphorylation of the PCT rearranges the DCT-channel interaction
tial, and k is the slope factor. to alleviate DCT inhibition.
on Ca2+ flux. Consider a steady-state ICa,L activation curve with a identification of CaV1.2 truncation at a consensus calpain sub-
50% maximal conductance at 0 mV. All else being equal, a strate site,80-82 coupled with the presence of approximately 37 kD
−10 mV shift of the steady-state activation curve will result in protein recognized by a DCT antibody,27 suggests that distal
more than 90% of maximal probability of channel activation at carboxyl-terminus is generated by proteolytic cleavage of the
the same potential (Figure 10-6). Thus, the shifting of activation full-length CaV1.2 protein. A requirement for Ca2+ and calpain
gating is a powerful mechanism for increasing ICa,L in response activity for carboxyl terminal cleavage in cardiomyocytes is
to β-AR stimulation. The detail structure-function underpin- inferred from studies on the skeletal muscle homolog (CaV1.1)80
nings of β-AR modulation are intensively studied, yet gaps in and from sequence data conserved between CaV1.1 and CaV1.2.
knowledge still remain. β-AR stimulation via Gs signaling acti- As yet, there is no direct evidence for either a Ca2+ or calpain
vates protein kinase A (PKA). Several unambiguous substrates for requirement for DCT liberation in cardiomyocytes. Neverthe-
PKA have been identified on CaV1.2. These include sites on the less, the majority of studies showing Western blots probed with
proximal- and distal-carboxyl–termini.73-75 In addition. CaVβ2 is anti-CaV1.2 antibodies reveal a protein migrating to approxi-
phosphorylated by PKA.76,77 However, genetically modified mice mately 190 kD rather than the approximately 240 kD expected
carrying a knock-in of CaV1.2 containing a serine-to-alanine for the full-length, that is, unproteolyzed CaV1.2. It should be
mutation in the distal carboxyl-terminus site at position 1928, noted that heterologously expressed CaV1.2 is not processed, and
rendering the channel phospho-deficient,78 or phospo-deficient adenoviral-CaV1.2 constructs introduced into cardiomyocytes
CaV1.2-Ser1928Ala in combination with a truncated phospho- also express full-length channel. This suggests that in cardiomyo-
deficient CaVβ2, retained β-AR modulation of ICa,L in cardiomyocytes, native CaV1.2 mRNA and/or protein is processed relatively
cytes.79 Sites on the CaV1.2 proximal carboxyl-terminus, S1700 early during synthesis. Regardless, reassociation of DCT with
and T1704, are probable substrates for PKA that mediate current CaV1.2 results in gating modifications. Coexpression of DCT
modulation. In this vein, heterologous expression studies showed with CaV1.2 truncated at the position corresponding to proteo-
that a complex series of events can result in recapturing L-type– lytic cleavage results in modification of activation gating. Trun-
channel modulation in a reconstituted system. First, CaV1.2 cated CaV1.2 steady-state activation requires a two-additive
channels truncated at the predicted proteolytic cleavage site were Boltzmann distribution to describe the data. DCT shifts steady-
expressed along with A-kinase anchoring protein 15 (akap15) and state activation by +15 mV.25 It was also noted that the shift was
distal carboxyl-terminus.75 The idea is that distal carboxyl- greater when DCT and truncated CaV1.2 was coexpressed, as
terminus is autoinhibitory for channel gating.25 Akap15 scaffolds opposed to conditions whereby DCT was retained as part of the
distal carboxyl-terminus to proximal carboxyl-terminus25 and, full-length channel construct.25 Mechanistically, DCT might
when expressed in optimal ratios, confers PKA modulation on restrict coupling between gating current (movement of S4 seg-
channels via phosphorylation of the CaV1.2 proximal carboxyl- ments) and allosteric rearrangements that result in ionic flux.25
terminus at positions S1700 and T1704.75 In cardiomyocytes, The broader implication that requires further testing in cardio-
akap150/79 is required for β-AR modulation of ICa,L.45 This is a myocytes is that DCT-proximal carboxyl-terminus interactions
provocative model—specifically, that PKA modulates CaV1.2 restrict activation, and elevated PKA via akap-scaffolded proteins
gating via disruption of an autoinhibitory function (Figure 10-7). disrupts the proximal-carboxyl-terminus-DCT interactions,
Some uncertainty to this model includes the paucity of data resulting in the well-known negative shift of Ca2+-channel activa-
showing distal carboxyl-terminus autoinhibition of Ca2+ current tion in response to β-AR signaling (see Figure 10-7).
in cardiomyocytes and incomplete understanding of the mecha-
nisms of gating modification of L-type Ca2+ channels by the
proteolytically separated distal carboxyl-terminus domain. Facilitation
Pioneering work from the Catterall laboratory underscores
the importance of the distal carboxyl-terminus domain (DCT) in Excessive activation of calmodulin-kinase II (CaMKII) signaling
the precise regulation of L-type Ca2+-channel function. The causes arrhythmias and heart failure.83 CaVβ2 is a target for
CaMKII-induced phosphorylation and CaVβ2 mediates CaMKII- profoundly block ICa,L in heterologous expression systems or car-
triggered cardiomyocyte afterdepolarization and death.84 The diomyocytes,97,98 the RGK abrogation of ICa,L is modified by coex-
implication is that altered L-type Ca2+-channel gating links pression of CaM.34 Rem and CaM co-overexpression not only
CaMKII activity to arrhythmias and cardiomyocyte survival. A blunted the ability of Rem to completely block ICa,L, but slowed
1-Hz train of depolarizations elicits a progressive increase in ICa,L Ca2+-dependent inactivation (Figure 10-8). These findings are
that accumulates over about 10 pulses.85-87 This positive ICa,L consistent with a mechanism that includes Rem interference with
staircase is mediated by CaMKII-dependent phosphorylation.88-90 CaM modulation. Initial biochemical studies showed CaVβ–RGK
In addition to the preeminent CaMKII phosphorylation targets interactions.93,94 However, RGK overexpression does not neces-
on CaVβ2, sites on CaV1.2 have been also identified.35,91 Activated sarily inhibit surface expression of CaV1.2.95,99 Moreover, Rem
CaMKII promotes “Mode 2” L-type Ca2+-channel gating.90 and Rad engineered to prevent RGK-CaVβ interaction retained
Mode 2 is the classical description of L-type channel gating with the ability to block L-type current.41 Recent studies demonstrate
frequent, relatively long-duration openings.92 Interestingly, direct RGK–CaV1.2 interactions. Rem interacts with the
CaMKII tethers to multiple sites on the CaV1.2 proximal CaV1.2N-terminus,41 and the CaV1.2–CaV1.2 proximal carboxyl-
carboxyl-terminus adjacent to the CaM-binding IQ motif.35 terminus on a domain overlapping with the CaM interaction
site.34 RGK modulation of L-type Ca2+–channel function under-
scores two key themes: (1) the carboxyl-terminus of CaV1.2 is a
hotspot for multiple protein interactions and (2) gating may be
Other Protein Interactions Affecting L-Type modified depending on the precise composition of proteins in the
Ca2+–Channel Gating heteromultimeric L-type Ca2+–channel complex.
RGK Proteins
Calcineurin
L-type calcium–channel ionic current is potently inhibited by
RGK GTPases including Rem, Rad, and Kir/Gem.93,94 RGK Calcineurin controls VDI of L-type Ca2+ channels.100 In Timothy
blockade of L-type Ca2+-channel current does not necessarily Syndrome patients L-type Ca2+ channels display significantly
require interference with trafficking of CaV1.2 to the surface slower VDI, and calcineurin restores gating towards a normal
membrane.95 Rather, RGK proteins interfere with L-type Ca2+– state.100 Calcineurin binds to both the N-terminus and the
channel gating. In mice carrying deletion of Rem expression, CaV1.2-DCT.42 In neurons, calcineurin–CaV1.2 interaction
cardiomyocyte ICa,L density is elevated at voltages corresponding requires an akap79/150-scaffold protein.101 In distinction, myo-
to maximal activation gating, and steady state activation is cardial calcineurin–CaV1.2 interaction occurs via direct protein-
positive-shifted on the voltage axis.96 The positive shift of steady- protein association.42 Interest in calcineurin extends far beyond
state activation might be a homeostatic mechanism to counteract L-type Ca2+–channel gating. Calcineurin links cytosolic Ca2+ to
increased current density. Moreover, Rem knockout sheds light transcription signaling responsible for cardiac hypertrophy.102
on the contribution of Rem to L-type Ca2+-channel activation The parallels of bifunctionality for calcineurin and for CaV1.2-
gating. Although all RGK family proteins, when overexpressed, DCT are striking. Both proteins regulate L-type Ca2+–channel
Cav1.2 + β2a + GFP + PKH3-CaM (20)

Cav1.2 + β2a + GFP-Rem + PKH3 (8)
Cav1.2 + β2a + GFP-Rem + PKH3-CaM (16)
Cav1.2 + β2a + GFP + PKH3 (11) 30mM Ca2+
0 300 ms
2.5 30mM Ca2+
pA/pF
–40 –20 20 40
mV
.5
–5
Cav1.2 + β2a + GFP + PKH3
Cav1.2 + β2a + GFP + PKH3-CaM
–7.5 Cav1.2 + β2a + GFP-Rem + PKH3-CaM
–1
A –10 B
Figure 10-8. Ca2+-CaM–dependent inactivation is regulated by L-type Ca2+–channel complex interacting proteins. A, Current voltage curves obtained from whole-cell ICa,L
from HEK 293 cells heterologously expressing CaV1.2 + CaVβ2a and empty vector (PKH3), CaM, Rem, or Rem + CaM. Rem expressed alone results in no detectable ICa,L (open
circles). Rem coexpressed with CaM results in an approximate 50% reduction of peak ICa,L (closed triangles). CaM (closed squares) had no detectable effect on ICa,L compared
with empty vector (closed squares). Thus, CaM in more closely matched stoichiometries to overexpressed Rem can abrogate profound Rem blockade of current. B, Rem
may slow Ca2+-CaM–dependent inactivation. ICa,L elicited by a test potential step to +20 mV for empty vector (green), CaM (red), or Rem + CaM (black). Currents are normal-
ized to peak value. Rem + CaM ICa,L decay kinetics are significantly slower than empty-vector or CaM-alone expression. These data suggest the general notion that the native
stoichiometries of the various channel-complex proteins are important determinants for channel gating.
(Adapted from Pang C, Crump SM, Jin L, et al: Rem GTPase interacts with the proximal CaV1.2 C-terminus and modulates calcium-dependent channel inactivation. Channels
[Austin] 4:192–202, 2010.)
gating, and both proteins are involved in transcriptional N-Terminal–C-Terminal Interaction
10
regulation.
The cardiac CaV1.2 isoform contains a relatively long N-terminus
that inhibits IBa,L, and this long cardiac N-terminal domain is
crucial for the CaVβ2 increase of open probability.39 The
CaV1.2-Coupled Gating CaV1.2 N-terminal also interacts with its proximal carboxyl-
terminus via Ca2+-calmodulin.38
Early studies of single L-type Ca2+ channels showed that ensem- In conclusion, L-type Ca2+ channels provide the main route
ble averages of single-channel events do not necessarily scale to of entry into cardiomyocytes and are essential for providing
macroscopic ICa,L.103 This suggested the notion that CaV1.2 may trigger Ca2+ for excitation-contraction coupling. The mature car-
cluster, and that clustering behavior may alter channel gating. diomyocyte L-type Ca2+ channel is a complex of many proteins.
Oligomerization of CaV1.2 via carboxyl-terminal interactions The sum of interactions yields distinct channel-gating properties.
amplify Ca2+ signaling.37 In Timothy syndrome mutant CaV1.2 A large number of proteins can interact with the CaV1.2 carboxyl-
channels, the slower VDI kinetics require akap150.104 In the terminus, raising the notion that heterogeneous channel com-
absence of akap150, Timothy syndrome slow VDI reverts to plexes exist, even within a given cell. Finally, modulation of ICa,L
normal kinetics. Thus, CaV1.2–CaV1.2 interactions bridged by is dependent on multiple protein interactions that dictate channel
akap contribute to L-type Ca2+–channel inactivation gating. gating, and in turn, cardiomyocyte function.
15. Rosati B, Dun W, Hirose M, et al: Molecular basis Ca(2+) channels revealed by FRET in single
References of the T- and L-type Ca2+ currents in canine living cells. Neuron 31(6):973–985, 2001.
Purkinje fibres. J Physiol 579(Pt 2):465–471, 29. Pitt GS, Zuhlke RD, Hudmon A, et al: Molecular
1. Bers DM: Calcium cycling and signaling in 2007. basis of CaM tethering and Ca2+-dependent inac-
cardiac myocytes. Annu Rev Physiol 70:23–49, 16. Gaughan JP, Hefner CA, Houser SR: Electro- tivation of L-type Ca2+ channels. J Biol Chem
2008. physiological properties of neonatal rat ventricu- 2001:M104959200.
2. Catterall WA: Ion channel voltage sensors: Struc- lar myocytes with alpha1-adrenergic-induced 30. Zuhlke RD, Pitt GS, Deisseroth K, et al: Calmod-
ture, function, and pathophysiology. Neuron hypertrophy. Am J Physiol 275(2 Pt 2):H577– ulin supports both inactivation and facilitation of
67(6):915–928, 2010. H590, 1998. L-type calcium channels. Nature 399(6732):159–
3. Benitah JP, Alvarez JL, Gomez AM: L-type 17. Wetzel GT, Chen F, Friedman WF, et al: Calcium 162, 1999.
Ca(2+) current in ventricular cardiomyocytes. current measurements in acutely isolated neonatal 31. Van Petegem F, Chatelain FC, Minor DL Jr:
J Mol Cell Cardiol 48(1):26–36, 2010. cardiac myocytes. Pediatr Res 30(1):83–88, Insights into voltage-gated calcium channel regu-
4. Hille B: Ion Channels of Excitable Membranes, 1991. lation from the structure of the CaV1.2 IQ
3rd ed. Sunderland, MA, 2001, Sinauer Associ- 18. Nuss HB, Houser SR: T-type Ca2+ current is domain-Ca2+/calmodulin complex. Nat Struct
ates, Inc. expressed in hypertrophied adult feline left ven- Mol Biol 12(12):1108–1115, 2005.
5. DeFelice LJ: Electrical Properties of Cells: Patch tricular myocytes. Circ Res 73(4):777–782, 1993. 32. Blaich A, Pahlavan S, Tian Q, et al: Mutation of
Clamp for Biologists. New York and London, 19. Martinez ML, Heredia MP, Delgado C: Expres- the CaM-binding motif IQ of the L-type Cav1.2
1997, Plenum Press. sion of T-type Ca(2+) channels in ventricular cells Ca2+ channel to EQ induces dilated cardiomy-
6. Zhang Q, Timofeyev V, Qiu H, et al: Expression from hypertrophied rat hearts. J Mol Cell Cardiol opathy and death. J Biol Chem 287(27):22616–
and roles of Cav1.3 (alpha1D) L-type Ca(2)+ 31(9):1617–1625, 1999. 22625, 2012.
channel in atrioventricular node automaticity. 20. Gao T, Chien AJ, Hosey MM: Complexes of the 33. Van Petegem F, Clark KA, Chatelain FC, et al:
J Mol Cell Cardiol 50(1):194–202, 2011. alpha 1C and beta subunits generate the necessary Structure of a complex between a voltage-gated
7. Zhang Z, He Y, Tuteja D, et al: Functional roles signal for membrane targeting of class C L-type calcium channel beta-subunit and an alpha-
of Cav1.3(alpha1D) calcium channels in atria: calcium channels. J Biol Chem 274(4):2137–2144, subunit domain. Nature 429(6992):671–675,
Insights gained from gene-targeted null mutant 1999. 2004.
mice. Circulation 112(13):1936–1944, 2005. 21. Opatowsky Y, Chen CC, Campbell KP, et al: 34. Pang C, Crump SM, Jin L, et al: Rem GTPase
8. Mangoni ME, Couette B, Bourinet E, et al: Func- Structural analysis of the voltage-dependent interacts with the proximal Ca(V)1.2 C-terminus
tional role of L-type Cav1.3 Ca2+ channels in calcium channel beta subunit functional core and and modulates calcium-dependent channel inacti-
cardiac pacemaker activity. Proc Natl Acad Sci its complex with the alpha 1 interaction domain. vation. Channels (Austin) 4(3):192–202, 2010.
U S A 100(9):5543–5548, 2003. Neuron 42(3):387–399, 2004. 35. Hudmon A, Schulman H, Kim J, et al: CaMKII
9. Platzer J, Engel J, Schrott-Fischer A, et al: Con- 22. Bichet D, Cornet V, Geib S, et al: The I-II loop tethers to L-type Ca2+ channels, establishing
genital deafness and sinoatrial node dysfunction of the Ca2+ channel alpha1 subunit contains an a local and dedicated integrator of Ca2+ signals
in mice lacking class D L-type Ca2+ channels. endoplasmic reticulum retention signal antago- for facilitation. J Cell Biol 171(3):537–547,
Cell 102(1):89–97, 2000. nized by the beta subunit. Neuron 25(1):177–190, 2005.
10. Sims C, Reisenweber S, Viswanathan PC, et al: 2000. 36. Muller CS, Haupt A, Bildl W, et al: Quantita-
Sex, age, and regional differences in L-type 23. Fang K, Colecraft HM: Mechanism of auxiliary tive proteomics of the Cav2 channel nano-
calcium current are important determinants of beta-subunit-mediated membrane targeting of environments in the mammalian brain. Proc Natl
arrhythmia phenotype in rabbit hearts with drug- L-type (Ca(V)1.2) channels. J Physiol 589(Pt Acad Sci U S A 107(34):14950–14957, 2010.
induced long QT type 2. Circ Res 102(9):e86– 18):4437–4455, 2011. 37. Dixon RE, Yuan C, Cheng EP, et al: Ca2+ signal-
100, 2008. 24. Jangsangthong W, Kuzmenkina E, Bohnke AK, ing amplification by oligomerization of L-type
11. Yang X, Chen G, Papp R, et al: Oestrogen upreg- et al: Single-channel monitoring of reversible Cav1.2 channels. Proc Natl Acad Sci U S A
ulates L-type Ca(2)(+) channels via oestrogen- L-type Ca(2+) channel Ca(V)alpha(1)-Ca(V)beta 109(5):1749–1754, 2012.
receptor by a regional genomic mechanism in subunit interaction. Biophys J 101(11):2661– 38. Dick IE, Tadross MR, Liang H, et al: A modular
female rabbit hearts. J Physiol 590(Pt 3):493–508, 2670, 2012. switch for spatial Ca2+ selectivity in the cal
2012. 25. Hulme JT, Yarov-Yarovoy V, Lin TW, et al: Auto- modulin regulation of CaV channels. Nature
12. Cribbs LL, Lee J-H, Yang J, et al: Cloning and inhibitory control of the CaV1.2 channel by its 451(7180):830–834, 2008.
characterization of a1H from human heart, a proteolytically processed distal C-terminal 39. Kanevsky N, Dascal N: Regulation of maximal
member of the T-type calcium channel gene domain. J Physiol 576(Pt 1):87–102, 2006. open probability is a separable function of Ca-v
family. Circ Res 83:103–109, 1998. 26. Gomez-Ospina N, Tsuruta F, Barreto-Chang O, beta subunit in L-type Ca2+ channel, dependent
13. Cribbs LL, Martin BL, Schroder EA, et al: Iden- et al: The C terminus of the L-type voltage-gated on NH2 terminus of alpha(1C) (Ca(v)1.2 alpha).
tification of the T-type calcium channel (CaV3.1d) calcium channel ca(v)1.2 encodes a transcription J Gen Physiol 128(1):15–36, 2006.
in developing mouse heart. Circ Res 88(4):403– factor. Cell 127(3):591–606, 2006. 40. Minor DL Jr, Findeisen F: Progress in the struc-
407, 2001. 27. Schroder E, Byse M, Satin J: L-type calcium tural understanding of voltage-gated calcium
14. Hagiwara N, Irisawa H, Kameyama M: Contribu- channel C terminus autoregulates transcription. channel (CaV) function and modulation. Chan-
tion of two types of calcium currents to the pace- Circ Res 104(12):1373–1381, 2009. nels (Austin) 4(6):459–474, 2010.
maker potentials of rabbit sino-atrial node cells. 28. Erickson MG, Alseikhan BA, Peterson BZ, et al: 41. Yang T, Puckerin A, Colecraft HM: Distinct
J Physiol 395:233–253, 1988. Preassociation of calmodulin with voltage-gated RGK GTPases differentially use alpha1- and
auxiliary beta-binding-dependent mechanisms to phenotypes in Timothy syndrome. Nature channels in skeletal muscle. Proc Natl Acad Sci U
inhibit CaV1.2/CaV2.2 channels. PLoS One 471(7337):230–U120, 2011. S A 102(14):5274–5279, 2005.
7(5):e37079, 2012. 62. Brunet S, Scheuer T, Catterall WA: Cooperative 81. De Jongh KS, Colvin AA, Wang KK, et al:
42. Tandan S, Wang Y, Wang TT, et al: Physical and regulation of Ca(v)1.2 channels by intracellular Differential proteolysis of the full-length form
functional interaction between calcineurin and Mg(2+), the proximal C-terminal EF-hand, and of the L-type calcium channel alpha 1 subunit
the cardiac L-type Ca2+ channel. Circ Res the distal C-terminal domain. J Gen Physiol by calpain. J Neurochem 63(4):1558–1564,
105(1):51–60, 2009. 134(2):81–94, 2009. 1994.
43. Xu H, Ginsburg KS, Hall DD, et al: Targeting 63. Josephson IR, Guia A, Lakatta EG, et al: Modula- 82. Gerhardstein BL, Gao T, Bunemann M, et al:
of protein phosphatases PP2A and PP2B to tion of the gating of unitary cardiac L-type Ca(2+) Proteolytic processing of the C terminus of the
the C-terminus of the L-type calcium channel channels by conditioning voltage and divalent alpha 1C subunit of L-type calcium channels and
Ca v1.2. Biochemistry 49(48):10298–10307, ions. Biophys J 83(5):2575–2586, 2002. the role of a proline-rich domain in membrane
2010. 64. Josephson IR, Guia A, Lakatta EG, et al: Modula- tethering of proteolytic fragments. J Biol Chem
44. Leroy J, Richter W, Mika D, et al: Phosphodies- tion of the gating of unitary cardiac L-type Ca2+ 275(12):8556–8563, 2000.
terase 4B in the cardiac L-type Ca(2)(+) channel channels by conditioning voltage and divalent 83. Swaminathan PD, Purohit A, Hund TJ, et al:
complex regulates Ca(2)(+) current and protects ions. Biophys J 83(5):2575–2586, 2002. Calmodulin-dependent protein kinase II: Linking
against ventricular arrhythmias in mice. J Clin 65. Ferreira G, Yi J, Rios E, et al: Ion-dependent heart failure and arrhythmias. Circ Res
Invest 121(7):2651–2661, 2011. inactivation of barium current through L-type 110(12):1661–1677, 2012.
45. Nichols CB, Rossow CF, Navedo MF, et al: Sym- calcium channels. J Gen Physiol 109(4):449–461, 84. Koval OM, Guan X, Wu Y, et al: CaV1.2
pathetic stimulation of adult cardiomyocytes 1997. beta-subunit coordinates CaMKII-triggered
requires association of AKAP5 with a subpopula- 66. Erickson MG, Alseikhan BA, Peterson BZ, et al: cardiomyocyte death and afterdepolarizations.
tion of L-type calcium channels. Circ Res Preassociation of calmodulin with voltage-gated Proc Natl Acad Sci U S A 107(11):4996–5000,
107(6):747–756, 2010. Ca(2+) channels revealed by FRET in single 2010.
46. Dai S, Hall DD, Hell JW: Supramolecular assem- living cells.PG - 973-85. Neuron 31(6):973–985, 85. Hryshko LV, Bers DM: Ca current facilitation
blies and localized regulation of voltage-gated ion 2001. during postrest recovery depends on Ca entry. Am
channels. Physiol Rev 89(2):411–452, 2009. 67. Tadross MR, Dick IE, Yue DT: Mechanism of J Physiol 259(3 Pt 2):H951–H961, 1990.
47. Lu L, Zhang Q, Timofeyev V, et al: Molecular local and global Ca2+ sensing by calmodulin in 86. Picht E, DeSantiago J, Huke S, et al: CaMKII
coupling of a Ca2+-activated K+ channel to complex with a Ca2+ channel. Cell 133(7):1228– inhibition targeted to the sarcoplasmic reticulum
L-type Ca2+ channels via α-actinin2. Circ Res 1240, 2008. inhibits frequency-dependent acceleration of
100:112–120, 2007. 68. Liang H, DeMaria CD, Erickson MG, et al: relaxation and Ca2+ current facilitation. J Mol
48. Meyers MB, Puri TS, Chien AJ, et al: Sorcin Unified mechanisms of Ca2+ regulation across Cell Cardiol 42(1):196–205, 2007.
associates with the pore-forming subunit of the Ca2+ channel family. Neuron 39(6):951–960, 87. Zygmunt AC, Maylie J: Stimulation-dependent
voltage-dependent L-type Ca2+ channels. J Biol 2003. facilitation of the high threshold calcium current
Chem 273(30):18930–18935, 1998. 69. Alseikhan BA, DeMaria CD, Colecraft HM, et al: in guinea-pig ventricular myocytes. J Physiol
49. Payandeh J, Gamal El-Din TM, Scheuer T, et al: Engineered calmodulins reveal the unexpected 428:653–671, 1990.
Crystal structure of a voltage-gated sodium eminence of Ca2+ channel inactivation in 88. Yuan W, Bers DM: Ca-dependent facilitation of
channel in two potentially inactivated states. controlling heart excitation. Proc Natl Acad Sci cardiac Ca current is due to Ca-calmodulin-
Nature 486(7401):135–139, 2012. U S A 99(26):17185–17190, 2002. dependent protein kinase. Am J Physiol 267(3 Pt
50. Foell JD, Balijepalli RC, Delisle BP, et al: Molec- 70. Josephson IR, Guia A, Lakatta EG, et al: Ca(2+)- 2):H982–H993, 1994.
ular heterogeneity of calcium channel beta- dependent components of inactivation of unitary 89. Anderson ME, Braun AP, Schulman H, et al:
subunits in canine and human heart: evidence for cardiac L-type Ca(2+) channels. J Physiol 588(Pt Multifunctional Ca2+/calmodulin-dependent
differential subcellular localization. Physiol 1):213–223, 2010. protein kinase mediates Ca(2+)-induced enhance-
Genomics 17(2):183–200, 2004. 71. Bers DM, Stiffel VM: Ratio of ryanodine to dihy- ment of the L-type Ca2+ current in rabbit
51. Buraei Z, Yang JA: The beta subunit of voltage- dropyridine receptors in cardiac and skeletal ventricular myocytes. Circ Res 75(5):854–861,
gated Ca2+ channels. Physiol Rev 90(4):1461– muscle and implications for E-C coupling. Am J 1994.
1506, 2010. Physiol 264(6 Pt 1):C1587–1593, 1993. 90. Dzhura I, Wu Y, Colbran RJ, et al: Calmodulin
52. Yu H, Chen JK, Feng S, et al: Structural basis for 72. Acsai K, Antoons G, Livshitz L, et al: Microdo- kinase determines calcium-dependent facilitation
the binding of proline-rich peptides to SH3 main [Ca(2)(+)] near ryanodine receptors as of L-type calcium channels. Nat Cell Biol
domains. Cell 76(5):933–945, 1994. reported by L-type Ca(2)(+) and Na+/Ca(2)(+) 2(3):173–177, 2000.
53. Pragnell M, De Waard M, Mori Y, et al: exchange currents. J Physiol 589(Pt 10):2569– 91. Lee TS, Karl R, Moosmang S, et al: Calmodulin
Calcium channel beta-subunit binds to a con- 2583, 2011. kinase II is involved in voltage-dependent facilita-
served motif in the I-II cytoplasmic linker of the 73. Kamp TJ, Hell JW: Regulation of cardiac L-type tion of the L-type Cav1.2 calcium channel: Iden-
alpha 1-subunit. Nature 368(6466):67–70, calcium channels by protein kinase A and protein tification of the phosphorylation sites. J Biol
1994. kinase C. Circ Res 87(12):1095–1102, 2000. Chem 281(35):25560–25567, 2006.
54. Almagor L, Chomsky-Hecht O, Ben-Mocha A, 74. Hulme JT, Westenbroek RE, Scheuer T, et al: 92. Hess P, Lansman JB, Tsien RW: Different modes
et al: The role of a voltage-dependent Ca2+ Phosphorylation of serine 1928 in the distal of Ca channel gating behaviour favoured by dihy-
channel intracellular linker: A structure-function C-terminal domain of cardiac CaV1.2 channels dropyridine Ca agonists and antagonists. Nature
analysis. J Neurosci 32(22):7602–7613, 2012. during β1-adrenergic regulation. Proc Natl Acad 311(5986):538–544, 1984.
55. Grandi E, Morotti S, Ginsburg KS, et al: Inter- Sci U S A 103(44):16574–16579, 2006. 93. Finlin BS, Crump SM, Satin J, et al: Regulation
play of voltage and Ca-dependent inactivation of 75. Fuller MD, Emrick MA, Sadilek M, et al: Molec- of voltage-gated calcium channel activity by
L-type Ca current. Progr Biophys Mol Biol ular mechanism of calcium channel regulation in the Rem and Rad GTPases. Proc Natl Acad Sci
103(1):44–50, 2010. the fight-or-flight response. Sci Signal 3(141):ra70, U S A 100(24):14469–14474, 2003.
56. Peterson BZ, DeMaria CD, Adelman JP, et al: 2010. 94. Beguin P, Nagashima K, Gonoi T, et al: Regula-
Calmodulin is the Ca2+ sensor for Ca2+- 76. Perez-Reyes E, Castellano A, Kim HS, et al: tion of Ca2+ channel expression at the cell surface
dependent inactivation of L-type calcium chan- Cloning and expression of a cardiac/brain beta by the small G-protein kir/Gem. Nature
nels. Neuron 22(3):549–558, 1999. subunit of the L-type calcium channel. J Biol 100(24):701–706, 2001.
57. Mitarai S, Kaibara M, Yano K, et al: Two Chem 267(3):1792–1797, 1992. 95. Correll RN, Pang C, Finlin BS, et al: Plasma
distinct inactivation processes related to phos- 77. Bunemann M, Gerhardstein BL, Gao T, et al: membrane targeting is essential for Rem-
phorylation in cardiac L-type Ca(2+) channel cur- Functional regulation of L-type calcium channels mediated Ca2+ channel inhibition. J Biol Chem
rents. Am J Physiol Cell Physiol 279(3):C603–610, via protein kinase A—mediated phosphorylation 282(39):28431–28440, 2007.
2000. of the beta 2 subunit. J Biol Chem 274(48):33851– 96. Magyar J, Kiper CE, Sievert G, et al: Rem-
58. Findlay I: Beta-adrenergic and muscarinic ago- 33854, 1999. GTPase regulates cardiac myocyte L-type
nists modulate inactivation of L-type Ca2+ 78. Lemke T, Welling A, Christel CJ, et al: Unchanged calcium current. Channels (Austin) 6(3):1–8,
channel currents in guinea-pig ventricular myo- beta-adrenergic stimulation of cardiac L-type 2012.
cytes. J Physiol 545(Pt 2):375–388, 2002. calcium channels in Ca v 1.2 phosphorylation site 97. Crump SM, Correll RN, Schroder EA, et al: L-
59. Findlay I: beta-Adrenergic stimulation modulates S1928A mutant mice. J Biol Chem 283(50):34738– type calcium channel alpha-subunit and protein
Ca2+- and voltage-dependent inactivation of 34744, 2008. kinase inhibitors modulate Rem-mediated regula-
L-type Ca2+ channel currents in guinea-pig ven- 79. Brandmayr J, Poomvanicha M, Domes K, et al: tion of current. Am J Physiol Heart Circ Physiol
tricular myocytes. J Physiol 541(Pt 3):741–751, Deletion of the C-terminal phosphorylation 291(4):H1959–H1971, 2006.
2002. sites in the cardiac beta subunit does not affect 98. Wang G, Zhu X, Xie W, et al: Rad as a novel
60. Splawski I, Timothy KW, Sharpe LM, et al: the basic beta-adrenergic response of the heart regulator of excitation-contraction coupling and
Ca(v)1.2 calcium channel dysfunction causes a and the Cav1.2 channel. J Biol Chem beta-adrenergic signaling in heart. Circ Res
multisystem disorder including arrhythmia and 287(27):22584–22592, 2012. 106(2):317–327, 2010.
autism. Cell 119(1):19–31, 2004. 80. Hulme JT, Konoki K, Lin TW, et al: Sites 99. Yang T, Xu X, Kernan T, et al: Rem, a member
61. Yazawa M, Hsueh B, Jia XL, et al: Using induced of proteolytic processing and noncovalent asso- of the RGK GTPases, inhibits recombinant
pluripotent stem cells to investigate cardiac ciation of the distal C-terminal domain of CaV1.1 CaV1.2 channels using multiple mechanisms that
require distinct conformations of the GTPase. neuronal L-type Ca2+ channel activity and ventricle cells: Dependence on voltage, current
J Physiol 588(Pt 10):1665–1681, 2010. nuclear signaling. Neuron 55(2):261–275, and channel density. J Physiol 443:307–334,
100. Cohen-Kutner M, Yahalom Y, Trus M, et al: Cal-
cineurin controls voltage-dependent-inactivation
(VDI) of the normal and Timothy cardiac chan-
2007.
102. Molkentin JD: Dichotomy of Ca2+ in the heart:
Contraction versus intracellular signaling. J Clin
1991.
104. Cheng EP, Yuan C, Navedo MF, et al: Restoration
of normal L-type Ca2+ channel function during
10
nels. Sci Rep 2:366, 2012. Invest 116(3):623–626, 2006. Timothy syndrome by ablation of an anchoring
101. Oliveria SF, Dell’Acqua ML, Sather WA: 103. Mazzanti M, DeFelice LJ, Liu YM: Gating protein. Circ Res 109(3):255–261, 2011.
AKAP79/150 anchoring of calcineurin controls of L-type Ca2+ channels in embryonic chick
KCNQ1/KCNE1 Macromolecular
Signaling Complex: Channel
Microdomains and Human Disease 11
Lei Chen, Cecile Terrenoire, and Robert S. Kass
experiencing fatal cardiac arrhythmias in the face of elevated SNS

CHAPTER OUTLINE activity.8 Unraveling the molecular links between the SNS and
β-AR Signaling: Coordination of Localized Regulation regulation of the KCNQ1/KCNE1 channel has direct implica-
of Channel Proteins by A-Kinase Anchoring Proteins 115 tions for the mechanistic basis of triggers of arrhythmias
in LQTS.
Iks Channel Regulation and Human Diseases 117
Using Stem Cells as a New Tool to Study Role of Iks
Channels in Diseases 118 β-AR Signaling: Coordination of Localized
Summary 118 Regulation of Channel Proteins by A-Kinase
Anchoring Proteins
Autonomic nervous system control of heart rate and cardiac con- Key to the complex regulation of ion channels by β-AR stimula-
tractility, through sympathetic and parasympathetic activity, is a tion is the spatiotemporal control of local cAMP concentration.
fundamental property of the cardiovascular system. Exercise or This is mediated by the A-kinase anchoring proteins (AKAPs).
emotional stress stimulates the sympathetic nervous system AKAPs are a group of structurally diverse proteins with the
(SNS), resulting in a rapid and dramatic increase in heart rate. common function of binding to, but not being limited to, PKA
To ensure adequate diastolic filling time, the increase in heart regulatory subunits.9-16 AKAPs provide structural scaffolding to
rate is accompanied by a concomitant reduction in the ventricular integrate various enzymes and their substrates to form a com-
action potential duration (APD) and the corresponding QT partmentalized environment. This enables spatiotemporal
interval on the electrocardiogram (ECG). Defective regulation control of the regulatory enzymes (i.e., to present the enzymes
of cardiac electrical activity in the face of sympathetic nervous at high concentrations at the site of their substrates when
system activity can lead to arrhythmias.1 needed).17 Disruption of local complexes can unbalance the
SNS control of cardiac electrical activity is mediated by the response and may have pathophysiological consequences.
activation of β-adrenergic receptors (β-ARs) that regulate the The IKs channel forms a macromolecular signaling complex
function of select ion channels via phosphorylation by cyclic that is coordinated by the binding of AKAP9, also known as
adenosine monophosphate (cAMP)-dependent protein kinase A yotiao,18, 19 via a leucine zipper (LZ) motif in the KCNQ1 carboxy
(PKA). PKA-dependent phosphorylation up-regulates the activ- (C)-terminus domain. Recent studies suggest that AKAP9 associ-
ity of L-type calcium channels, leading to enhanced calcium entry ates not only with the PKA regulatory subunit (RII), but also with
that contributes to action potential prolongation, as well as an protein phosphatase 1 (PP1),18 phosphodiesterase (PDE),20 and
increase in intracellular calcium available for subsequent uptake adenylyl cyclase (AC)21-23 (Figure 11-2). Together these enzymes
by the cardiac sarcoplasmic reticulum (SR). PKA phosphoryla- control the phosphorylation state of the channel via the cAMP/
tion also activates the major intracellular calcium release channel PKA pathway.
on the SR, the type 2 ryanodine receptor (RyR2),2 which is
responsible for releasing calcium to trigger muscle contraction.
Sympathetic stimulation also leads to a PKA-dependent Role of Leucine Zipper in
increase in a slowly activating potassium channel current, IKs Protein-Protein Interaction
(Figure 11-1). IKs channels consist of the pore-forming α-subunit
KCNQ1 and the auxiliary β-subunit KCNE13,4 and contribute Marks and colleagues were the first to show that the cardiac
to cardiac repolarization. PKA-dependent modulation increases calcium release channel/RyR2 is regulated by a macromolecular
the repolarization current to counter the stimulatory effects of signaling complex in which kinases and phosphatases are targeted
PKA on L-type calcium channels5 to achieve a balance of inward to the channel via AKAP and leucine/isoleucine zipper (LZ)
and outward membrane currents. This balance of modulated motifs.2,24,25 They suggested that this may be a common motif for
currents is thought of as a necessary mechanism to regulate coordination of ion channel signaling complexes.25 Subsequent
calcium homeostasis in the face of sympathetic activity. investigations have shown that this is, in fact, the case for at least
Inherited mutations in ion channels have been associated with two other ion channels that are regulated by PKA: L-type calcium
disorders that are exacerbated by SNS activity. For example, the channels26, 27 and KCNQ1/KCNE1 potassium channels.18 The
genes that encode KCNQ1 and KCNE13,4 have been linked to LZ domain is an α-helical structure that forms coiled coils and
the congenital long QT syndrome (LQTS). LQTS, a rare disease was originally identified as highly conserved motifs mediating the
in which the QT interval of the ECG is prolonged as the result binding of transcription factors to DNA.28 Coiled coils are com-
of dysfunctional ventricular repolarization, can precipitate lethal posed of heptad repeats (abcdefg)n in which hydrophobic residues
polymorphic ventricular tachycardia and is associated with occur at positions “a” and “d” and form the hydrophobic face of
syncope, seizures, and sudden death.6 Mutations in KCNQ1 cause the helix, while “b, c, e, f, and g” are hydrophilic residues that
LQT-1, and mutations in KCNE1 cause LQT-5.7 In affected form the solvent-exposed part of the coiled coil.29 LZs classically
patients, triggers of arrhythmias are gene-specific, and those with contain a leucine in position “d” because of its flexible side chain,
mutations in KCNQ1 or KCNE1 are at greatest risk of although the canonical leucine residue can be replaced by an
115
isoleucine or valine. Electrostatic interactions between side de-phosphorylates the channel, respectively. The second pair
chains in the “e” and “g” sites from neighboring helices are involves AC and PDE, which increase and decrease the local
believed to help specify binding partners.30 In vitro site-directed cAMP gradient, respectively (see Figure 11-2). Working in
mutagenesis studies have successfully revealed the sequence spec- concert, these enzymes fine-tune the IKs channel function.34
ificity of interacting helices in proteins such as the GCN4 DNA • PKA. AKAP9-bound PKA regulates IKs channel function by
binding domain,31 phospholamban,32 the myosin binding subunit/ phosphorylation of a single amino acid residue (Ser27) on the
cGKIα,33 and ryanodine receptor types 1 and 2.25 In the case of N-terminus of KCNQ1.18 The biophysical consequence of
the IKs channel (KCNQ1/KCNE1), Marx et al. identified an LZ PKA phosphorylation of the IKs channel is a profound increase
motif in the C-terminus of the KCNQ1 subunit critical to in current amplitude, an accelerated onset of activation, a
AKAP9 interaction.18 Substitution of an alanine for one or more hyperpolarizing shift in the voltage dependence of channel
of the “d” position leucines or isoleucines in the LZ motif disin- activation, and a slowing of deactivation (the return of acti-
tegrated the IKs/AKAP9 complex and rendered the channel vated [open] to resting [closed] channels during diastole).35
unable to be regulated by PKA. The combined effect ensures that during voltage depolariza-
tion, KCNQ1/KCNE1 channel activity is increased in the
presence of SNS stimulation. Consequently, the substantial
Molecular Components of the Iks/AKAP9 Complex repolarization reserve is activated in the face of SNS-mediated
activity, and this reserve potassium channel current can offset
Associated with the IKs/AKAP9 complex are two pairs of enzymes SNS-mediated increases in calcium channel currents, which
with opposing effects on channel phosphorylation state. The would prolong APD.5 It is interesting to note that AKAP9 itself
first pair, which involves PKA and PP1, phosphorylates or was shown to be a substrate of PKA. Indeed, the phosphoryla-
tion of AKAP9 seemed to participate in the regulation of IKs
channels.36 Early works identified a region on AKAP9 (residues
K+ 1440-1457: LEEEVAKVIVSMSIAFAQ) as the primary
KCNE1
binding site for PKA RII subunits.37,38
KCNQ1 IKs • PP1. PP1 is a nonspecific serine/threonine phosphatase that
P dephosphorylates its substrate. AKAP9 provides a platform
cAMP that allows PP1 to specifically target the IKs channel.18 Thus,
PKA
PP1 can reverse the effect of PKA on the channel and attenuate
NE AKAP9 the channel activity. This was evidenced by an experiment in
PKA which it was found that addition of okadaic acid, a PP1 inhibi-
-AR PKA tor, enhanced the effect of PKA-dependent IKs channel regula-
tion.18 The PP1 binding site on AKAP9 was shown to be
Ca2+ located in a region that comprises residues 1171-1229.37
AKAP
P • AC. Upstream in the PKA pathway, ACs are activated by
ICaL Gs-coupled receptors, such as the β-ARs. ACs are responsible
for cAMP synthesis, which then activates PKA. Evidence now
Figure 11-1. SNS regulation of IKs currents. Activation of the β-adrenergic receptor
suggests that ACs are associated with various AKAPs, such as
(β-AR) by norepinephrine (NE) results in an increase in the intracellular cAMP con-
centration, which, in turn, turns on PKA. IKs channels are phosphorylated by PKA.
AKAP79/150, mAKAP, and AKAP9.21,39-41 AKAP9 associates
This regulation is mediated by a macromolecular complex coordinated by AKAP9. with AC1, 2, 3, and 9 but not 4, 5, and 6. Residues 808-957 of
SNS, Sympathetic nervous system; NE, norepinephrine; β-AR, β-adrenergic receptor; AKAP9 bind directly to the AC2 N-terminus. Expression of
PKA, protein kinase A; AKAP, A-kinase anchoring protein; ICaL, L-type Calcium AKAP9 inhibited the activity of AC2 and 3, but not AC1 or 9.22
channel. It has been demonstrated that AC9 was a member of the
KCNE1 KCNQ1
cAMP
S27 P + –
+ +
–
PKA AC9 PDE4D3
PP1
AKAP9
Figure 11-2. The KCNQ1/KCNE1/AKAP9 macromolecular complex. AKAP9 recruits PKA, phosphatase 1 (PP1), adenylyl cyclase 9 (AC9), and phosphodiesterase 4D3 (PDE4D3)
to regulate the phosphorylation state of IKs channels.
KCNQ1/KCNE1 Macromolecular Signaling Complex: Channel Microdomains and Human Disease 117
IKs/AKAP9 complex in the cardiac myocytes of both IKs trans- KCNQ1 G589D and LQT-1
11
genic mice and guinea pigs. AC9 association with the complex
sensitizes PKA phosphorylation of KCNQ1 to SNS The inherited G589D mutation has been linked to the long QT
stimulation.23 syndrome variant 1 (LQT-1) in Finnish families.52 The naturally
• PDE. Phosphodiesterases (PDEs) constitute the sole route for occurring G589D mutation occupies the “e” position in the
degrading cyclic nucleotides in cells. In the mammalian heart, KCNQ1 LZ motif that targets AKAP9.18 As has been noted,
the temporal and spatial dynamics of cAMP gradients are con- residues occupying the “e” and “g” positions in the LZ motif
trolled mainly by PDE3s and PDE4s with a prevailing role of exhibit a limited range of substitutions based on the volume
PDE4s, which are considered the cAMP-specific PDEs.42-44 In occluded by adjacent structures. Indeed, the G589D mutation
transgenic murine cardiac myocytes expressing IKs channels, disrupts the interaction between the IKs channel and AKAP9, and
PDE4s were shown to regulate the basal phosphorylation level dislocates PKA and PP1 from the macromolecular complex.
of IKs channels. Two PDE4D isoforms, likely PDE4D3 and Functionally, the G589D mutation renders the IKs channel unre-
PDE4D5, were found to interact with the channels.20 However, sponsive to cAMP stimulation.18 Affected LQTS patients with
in the heterologous expression system, only PDE4D3, which this mutation suffer from dysfunctional regulation of QT dura-
is known to interact with AKAPs (AKAP-18,45 AKAP-250,46,47 tion during mental and physical stress53 and are at risk of arrhyth-
and AKAP-45047,48), was shown to be specifically recruited to mia and sudden cardiac death during exercise.52 Thus an inherited
KCNQ1 by AKAP9 and to regulate the amplitude of channels mutation of a single residue on the KCNQ1 channel disrupts the
in response to cAMP stimulation.20 The binding site for IKs/AKAP9 signaling complex and raises the risk of arrhythmia
PDE4D3 on AKAP9 currently is not known. in affected patients.
Iks Channel Regulation and Human Diseases KCNE1 D76N Mutation and LQT-5
Phosphorylation of the IKs channel, stimulated by SNS and medi- Kurokawa et al demonstrated that cAMP-mediated functional
ated by AKAP9, causes an increase in current density during regulation of KCNQ1/KCNE1 channels requires the expression
depolarization, as well as a slowing of channel deactivation, which of KCNQ1 with its auxiliary subunit KCNE1.54 Compared with
results in an accumulation of open channels on a beat-by-beat the twofold increase in KCNQ1/KCNE1 current amplitude
basis.18,35,49 The net result is an increased outward current reserve caused by cAMP and okadaic acid, KCNQ1 alone did not respond
to counterbalance the increased activities of L-type calcium chan- to stimulation.54 This suggests that the auxiliary subunit KCNE1
nels and RYR in the face of β-AR stimulation. Mutations that might play a role in transducing protein phosphorylation into
occur within the IKs/AKAP9 complex have been reported to channel functional regulation and represents a new paradigm that
disrupt the physical and functional integrity of the macromolecu- disruption of channel regulation by mutations on the auxiliary
lar complex, leading to an imbalance in intracellular calcium subunit may result in disease. This proved to be true in the case
homeostasis and predisposing cells to two types of calcium- of a naturally occurring LQT-5 mutation KCNE1 D76N. Kuro-
mediated rhythm disturbances: early afterdepolarizations kawa et al. demonstrated that the D76N mutation ablated func-
(EADs)50,51 and delayed after depolarizations (DADs).5 It is now tional regulation of the channels by cAMP,54 resulting in an
well established that disturbance of IKs channel regulation causes expected delay in the onset of repolarization that is more pro-
various cardiac arrhythmic disorders (Figure 11-3). nounced in the face of SNS stimulation.
KCNE1 KCNQ1
D76N
AKAP9 G589D
S1570L
Figure 11-3. Mutations compromising the physical and functional integrity of the IKs/AKAP9 complex are associated with long QT syndrome (LQTS). Illustrated are the
KCNQ1 mutation G589D, the KCNE1 mutation D76N, and the AKAP9 mutation S1570L. All are associated with LQTS.
AKAP9 S1570L Mutation and LQT-11 IKs channels accumulate significantly at a fast rate as a result of
slow deactivation and fast activation caused by phosphorylation,
Given its critical role in the regulation of IKs channel function, it suggesting a large role of this slow channel in APD shortening
is conceivable to postulate that AKAP9 might be a candidate gene and arrhythmia susceptibility in the face of β-adrenergic stimula-
for LQTS. Chen et al.55 characterized the molecular mechanism tion.64 These findings raise the possibility that drugs that target
of KCNQ1/AKAP9 interaction. The interaction between the IKs channel regulation may be used to treat AF.
two molecules is a three-way binding that involves both AKAP9
N- and C-termini as well as the LZ on the KCNQ1 C-terminus.
Of particular interest, the AKAP9 C-terminal binding site (resi-
dues 1574-1643) contains an LZ motif that potentially matches Using Stem Cells as a New Tool to Study Role
the reciprocal counterpart in KCNQ1.55 Identification of AKAP9 of Iks Channels in Diseases
binding motifs enabled quick screening in specific exons of the
AKAP9 gene that encode for these important binding modules. Until recently, little was known regarding the biophysical proper-
In 50 LQTS patients with a strong clinical diagnosis (QTc ≥480 ties of native IKs channels in isolated human cardiomyocytes, as
or Schwartz score ≥3.0) but negative for mutations in all LQTS recordings of IKs in human cardiac cells had proved very difficult
genes, including KCNQ1 and KCNE1, a missense mutation to obtain.65-75 As a result, most of our current knowledge of the
(S1570L) was identified in one patient but was absent in 1320 cardiac IKs channel complex and its pathologic role had been
ethnically matched reference alleles. The patient had experienced obtained from studies using heterologous expression systems,76,77
episodes of syncope and a prolonged QTc of 485 milliseconds, as genetically altered mice,18 and rabbits.78 Over the past few years,
well as a positive family history. AKAP9 S1570L mutation is a new cellular model has emerged with the development of
located close to its C-terminal binding site for KCNQ1 and cardiac myocytes derived from human embryonic stem cells
has a negative impact on the formation of the IKs/AKAP9 (hESC-CMs)79 and human induced pluripotent stem cells
complex. The mutation also reduces PKA-induced phosphoryla- (hiPSC-CMs).80,81 The biophysical properties of IKs channels in
tion of KCNQ1, rendering the IKs channels insensitive to these cells were found to be remarkably similar to those recorded
PKA regulation. Moreover a computational modeling study in myocytes isolated from adult human heart but significantly
revealed that the AKAP9 S1570L mutation significantly pro- different from KCNQ1/KCNE1 channels in established heter-
longed the cardiac action potential duration, especially under the ologous systems, including mammalian cell lines18,61,82,83: (1)
impact of SNS innervations.55 These findings not only estab- Endogenous IKs currents are small in human cardiac myocytes74,84
lished AKAP9 as a novel causal gene for LQTS but also rein- but are much larger in heterologous systems as the result of
forced the idea that AKAPs play a critical role in many overexpression of recombinant channels; and (2) the midpoint of
physiological processes. activation of IKs channels in human cardiac myocytes is signifi-
cantly less depolarized than in heterologous systems.73,84 Quanti-
tative polymerase chain reaction (qPCR) measurements showed
IKs Channel and Atrial Fibrillation that KCNE1 is the major KCNE isoform expressed in hESC-
CMs. Functional experiments suggest that KCNE1 is expressed
The role of IKs channels in the pathophysiology of human cardiac at moderate levels in these cells such that IKs channels have vari-
arrhythmia is not limited to the LQTS. Recent findings suggest able α- and β-subunit stoichiometry that may be modulated
that mutations of the channel may also lead to short QT syn- further with changes in KCNE1 expression that occur in the
drome56 and familial atrial fibrillation (AF).57-60 Of particular developing heart or with disease.84 hiPSC-CMs are a potential
interest are two adjacent KCNQ1 mutations (S140G57 and source of functional human cardiac tissue that can be used as a
V141M58) that are associated with AF. Both mutations are located model system; however, they also offer the possibility of investi-
in the first transmembrane helix (S1) that interfaces with gating the mechanistic basis of heritable cardiac rhythm
KCNE1.61 In intact IKs channels, these two mutations disrupt disorders—channelopathies—in genetic backgrounds specific for
channel deactivation.62 With channels unable to close properly, individual patients. Recently, hiPSC-CMs derived from LQT-1
currents accumulate with each heartbeat, giving rise to a gain-of- patients were characterized and showed a phenotype consistent
function phenotype and shortened APD. Chan et al63 observed with the clinical manifestations.85
that the phenotype of S140G and V141M mutations showed dif-
ferential dependence on KCNE1. Although V141M requires the
presence of KCNE1 to confer its deleterious effect on channel
closing, S140G does not. This is thought to reflect the different Summary
physical distances between the two KCNQ1 residues and
KCNE1.63 The functional consequences of the AF mutations Tightly regulated by SNS, IKs channels play an important role in
resemble the regulatory response of the IKs channel to SNS stim- cardiac repolarization. The channels associate with AKAP9 to
ulation. Both require the co-assembly of KCNQ1 and KCNE1, form a macromolecular complex that includes PKA, PP1, AC9,
and both result in a beat-dependent increase in outward current. and PDE4D3. These regulatory enzymes work in concert to
Is it possible that adrenergic regulation of the IKs channel may regulate the phosphorylation state and biophysical function of
play a role in AF? This was illustrated in transgenic mice express- the channel. Disruptions in the physical and functional integrity
ing IKs channels.64 Compared with wild type mice, which do not of the complex seen in mutations of its various members have
express functional IKs channels, these transgenic mice showed been shown to cause cardiac rhythm disorders, particularly
increased susceptibility to atrial arrhythmia upon β-AR stimula- LQTS. The IKs/AKAP9 macromolecular complex is of significant
tion. Computational simulation demonstrated that the stimulated importance as a potential therapeutic target.
adrenergic receptro stimulation and blockade. Part 3. Sanguinetti MC, Curran ME, Zou A, et al: Coas-
References C. Am Heart J 90:795–803, 1975. sembly of K(V)LQT1 and minK (IsK) proteins to
2. Marx SO, Reiken S, Hisamatsu Y, et al: PKA phos- form cardiac I(Ks) potassium channel. Nature
1. Wit AL, Hoffman BF, Rosen MR: Electrophysiol- phorylation dissociates FKBP12.6 from the calcium 384:80–83, 1996.
ogy and pharmacology of cardiac arrhythmias. release channel (ryanodine receptor): Defective 4. Barhanin J, Lesage F, Guillemare E, et al: K(V)
IX. Cardiac electrophysiologic effects of beta regulation in failing hearts. Cell 101:365–376, 2000. LQT1 and lsK (minK) proteins associate to form
KCNQ1/KCNE1 Macromolecular Signaling Complex: Channel Microdomains and Human Disease 119
the I(Ks) cardiac potassium current. Nature zipper interaction with A kinase-anchoring protein 46. Willoughby D, Wong W, Schaack J, et al: An
384:78–80, 1996. 15. Proc Natl Acad Sci U S A 100:13093–13098, anchored PKA and PDE4 complex regulates sub-
5. Kass RS, Wiegers SE: The ionic basis of
concentration-related effects of noradrenaline on
the action potential of calf cardiac purkinje fibres.
27.
2003.
Hulme JT, Ahn M, Hauschka SD, et al: A novel
leucine zipper targets AKAP15 and cyclic AMP-
plasmalemmal cAMP dynamics. EMBO J 25:2051–
2061, 2006.
47. McCahill A, McSorley T, Huston E, et al: In
11
J Physiol 322:541–558, 1982. dependent protein kinase to the C terminus of the resting COS1 cells a dominant negative approach
6. Keating MT, Sanguinetti MC: Molecular and cel- skeletal muscle Ca2+ channel and modulates its shows that specific, anchored PDE4 cAMP phos-
lular mechanisms of cardiac arrhythmias. Cell function. J Biol Chem 277:4079–4087, 2002. phodiesterase isoforms gate the activation, by basal
104:569–580, 2001. 28. Landschulz WH, Johnson PF, McKnight SL: The cyclic AMP production, of AKAP-tethered protein
7. Splawski I, Shen J, Timothy KW, et al: Spectrum leucine zipper: A hypothetical structure common kinase A type II located in the centrosomal region.
of mutations in long-QT syndrome genes: to a new class of DNA binding proteins. Science Cell Signal 17:1158–1173, 2005.
KVLQT1, HERG, SCN5A, KCNE1, and 240:1759–1764, 1988. 48. Tasken KA, Collas P, Kemmner WA, et al: Phos-
KCNE2. Circulation 102:1178–1185, 2000. 29. Lupas A: Coiled coils: New structures and new phodiesterase 4D and protein kinase A type II con-
8. Schwartz PJ, Priori SG, Spazzolini C, et al: functions. Trends Biochem Sci 21:375–382, 1996. stitute a signaling unit in the centrosomal area.
Genotype-phenotype correlation in the long-QT 30. Walshaw J, Shipway JM, Woolfson DN: Guide- J Biol Chem 276:21999–22002, 2001.
syndrome: Gene-specific triggers for life- lines for the assembly of novel coiled-coil struc- 49. Walsh KB, Kass RS: Regulation of a heart potas-
threatening arrhythmias. Circulation 103:89–95, tures: Alpha-sheets and alpha-cylinders. Biochem sium channel by protein kinase A and C. Science
2001. Soc Symp 111–123, 2001. 242:67–69, 1988.
9. Carnegie GK, Means CK, Scott JD: A-kinase 31. Harbury PB, Zhang T, Kim PS, Alber T: A switch 50. January CT, Moscucci A: Cellular mechanisms of
anchoring proteins: From protein complexes to between two-, three-, and four-stranded coiled early afterdepolarizations. Ann N Y Acad Sci
physiology and disease. IUBMB Life 61:394–406, coils in GCN4 leucine zipper mutants. Science 644:23–32, 1992.
2009. 262:1401–1407, 1993. 51. January CT, Riddle JM: Early afterdepolarizations:
10. Beene DL, Scott JD: A-kinase anchoring proteins 32. Simmerman HK, Kobayashi YM, Autry JM, et al: Mechanism of induction and block. A role for
take shape. Curr Opin Cell Biol 19:192–198, 2007. A leucine zipper stabilizes the pentameric mem- L-type Ca2+ current. Circ Res 64:977–990, 1989.
11. McConnachie G, Langeberg LK, Scott JD: AKAP brane domain of phospholamban and forms a 52. Piippo K, Swan H, Pasternack M, et al: A founder
signaling complexes: getting to the heart of the coiled-coil pore structure. J Biol Chem 271:5941– mutation of the potassium channel KCNQ1 in
matter. Trends Mol Med 12:317–323, 2006. 5946, 1996. long QT syndrome: implications for estimation of
12. Smith FD, Langeberg LK, Scott JD: The where’s 33. Surks HK, Mochizuki N, Kasai Y, et al: Regulation disease prevalence and molecular diagnostics. J Am
and when’s of kinase anchoring. Trends Biochem of myosin phosphatase by a specific interaction Coll Cardiol 37:562–568, 2001.
Sci 31:316–323, 2006. with cGMP- dependent protein kinase Ialpha. 53. Paavonen KJ, Swan H, Piippo K, et al: Response
13. Dodge-Kafka KL, Langeberg L, Scott JD: Com- Science 286:1583–1587, 1999. of the QT interval to mental and physical stress in
partmentation of cyclic nucleotide signaling in the 34. Chen L, Kass RS: A-kinase anchoring protein 9 types LQT1 and LQT2 of the long QT syndrome.
heart: The role of A-kinase anchoring proteins. and IKs channel regulation. J Cardiovasc Pharma- Heart 86:39–44, 2001.
Circ Res 98:993–1001, 2006. col 58:459–413, 2011. 54. Kurokawa J, Chen L, Kass RS: Requirement of
14. Wong W, Scott JD: AKAP signalling complexes: 35. Terrenoire C, Clancy CE, Cormier JW, et al: Auto- subunit expression for cAMP-mediated regulation
Focal points in space and time. Nat Rev Mol Cell nomic control of cardiac action potentials: Role of of a heart potassium channel. Proc Natl Acad Sci
Biol 5:959–970, 2004. potassium channel kinetics in response to sympa- U S A 100:2122–2127, 2003.
15. Michel JJ, Scott JD: AKAP mediated signal trans- thetic stimulation. Circ Res 96:e25–e34, 2005. 55. Chen L, Marquardt ML, Tester DJ, et al: Mutation
duction. Annu Rev Pharmacol Toxicol 42:235–257, 36. Chen L, Kurokawa J, Kass RS: Phosphorylation of of an A-kinase-anchoring protein causes long-QT
2002. the A-kinase-anchoring protein Yotiao contributes syndrome. Proc Natl Acad Sci U S A 104:20990–
16. Colledge M, Scott JD: AKAPs: From structure to to protein kinase A regulation of a heart potassium 20995, 2007.
function. Trends Cell Biol 9:216–221, 1999. channel. J Biol Chem 280:31347–31352, 2005. 56. Bellocq C, van Ginneken AC, Bezzina CR, et al:
17. Gray PC, Scott JD, Catterall WA: Regulation of 37. Westphal RS, Tavalin SJ, Lin JW, et al: Regulation Mutation in the KCNQ1 gene leading to the short
ion channels by cAMP-dependent protein kinase of NMDA receptors by an associated phosphatase- QT-interval syndrome. Circulation 109:2394–
and A-kinase anchoring proteins. Curr Opin Neu- kinase signaling complex. Science 285:93–96, 2397, 2004.
robiol 8:330–334, 1998. 1999. 57. Chen YH, Xu SJ, Bendahhou S, et al: KCNQ1
18. Marx SO, Kurokawa J, Reiken S, et al: Require- 38. Feliciello A, Cardone L, Garbi C, et al: Yotiao gain-of-function mutation in familial atrial fibrilla-
ment of a macromolecular signaling complex for protein, a ligand for the NMDA receptor, binds tion. Science 299:251–254, 2003.
beta adrenergic receptor modulation of the and targets cAMP-dependent protein kinase II(1). 58. Hong K, Piper DR, Diaz-Valdecantos A, et al: De
KCNQ1-KCNE1 potassium channel. Science FEBS Lett 464:174–178, 1999. novo KCNQ1 mutation responsible for atrial fibril-
295:496–499, 2002. 39. Bauman AL, Soughayer J, Nguyen BT, et al: lation and short QT syndrome in utero. Cardiovasc
19. Potet F, Scott JD, Mohammad-Panah R, et al: Dynamic regulation of cAMP synthesis through Res 68:433–440, 2005.
AKAP proteins anchor cAMP-dependent protein anchored PKA-adenylyl cyclase V/VI complexes. 59. Lundby A, Ravn LS, Svendsen JH, et al: KCNQ1
kinase to KvLQT1/IsK channel complex. Am J Mol Cell 23:925–931, 2006. mutation Q147R is associated with atrial fibrilla-
Physiol Heart Circ Physiol 280:H2038–H2045, 40. Kapiloff MS, Piggott LA, Sadana R, et al: An tion and prolonged QT interval. Heart Rhythm
2001. adenylyl cyclase-mAKAPbeta signaling complex 4:1532–1541, 2007.
20. Terrenoire C, Houslay MD, Baillie GS, et al: The regulates cAMP levels in cardiac myocytes. J Biol 60. Das S, Makino S, Melman YF, et al: Mutation in
cardiac IKs potassium channel macromolecular Chem 284:23540–23546, 2009. the S3 segment of KCNQ1 results in familial lone
complex includes the phosphodiesterase PDE4D3. 41. Efendiev R, Samelson BK, Nguyen BT, et al: atrial fibrillation. Heart Rhythm 6:1146–1153,
J Biol Chem 284:9140–9146, 2009. AKAP79 interacts with multiple adenylyl cyclase 2009.
21. Dessauer CW: Adenylyl cyclase—A-kinase anchor- (AC) isoforms and scaffolds AC5 and -6 to alpha- 61. Chung DY, Chan PJ, Bankston JR, et al: Location
ing protein complexes: The next dimension in amino-3-hydroxyl-5-methyl-4-isoxazole-propio- of KCNE1 relative to KCNQ1 in the I(KS) potas-
cAMP signaling. Mol Pharmacol 76:935–941, nate (AMPA) receptors. J Biol Chem 285: sium channel by disulfide cross-linking of substi-
2009. 14450–14458, 2010. tuted cysteines. Proc Natl Acad Sci U S A
22. Piggott LA, Bauman AL, Scott JD, et al: The 42. Mongillo M, McSorley T, Evellin S, et al: Fluores- 106:743–748, 2009.
A-kinase anchoring protein Yotiao binds and regu- cence resonance energy transfer-based analysis of 62. Restier L, Cheng L, Sanguinetti MC: Mechanisms
lates adenylyl cyclase in brain. Proc Natl Acad Sci cAMP dynamics in live neonatal rat cardiac myo- by which atrial fibrillation-associated mutations in
U S A 105:13835–13840, 2008. cytes reveals distinct functions of compartmental- the S1 domain of KCNQ1 slow deactivation of IKs
23. Li Y, Chen L, Kass RS, Dessauer CW: The ized phosphodiesterases. Circ Res 95:67–75, 2004. channels. J Physiol 586:4179–4191, 2008.
A-kinase anchoring protein Yotiao facilitates 43. Rochais F, Abi-Gerges A, Horner K, et al: A spe- 63. Chan PJ, Osteen JD, Xiong D, et al: Characteriza-
complex formation between type 9 adenylyl cyclase cific pattern of phosphodiesterases controls the tion of KCNQ1 atrial fibrillation mutations reveals
and the IKs potassium channel in heart. J Biol cAMP signals generated by different Gs-coupled distinct dependence on KCNE1. J Gen Physiol
Chem 287:29815–29824, 2012. receptors in adult rat ventricular myocytes. Circ 139:135–144, 2012.
24. Kass RS, Kurokawa J, Marx SO, et al: Leucine/ Res 98:1081–1088, 2006. 64. Sampson KJ, Terrenoire C, Cervantes DO, et al:
isoleucine zipper coordination of ion channel mac- 44. Houslay MD, Baillie GS, Maurice DH: cAMP- Adrenergic regulation of a key cardiac potassium
romolecular signaling complexes in the heart: Specific phosphodiesterase-4 enzymes in the car- channel can contribute to atrial fibrillation: Evi-
Roles in inherited arrhythmias. Trends Cardiovasc diovascular system: a molecular toolbox for dence from an I Ks transgenic mouse. J Physiol
Med 13:52–56, 2003. generating compartmentalized cAMP signaling. 586:627–637, 2008.
25. Marx SO, Reiken S, Hisamatsu Y, et al: Circ Res 100:950–966, 2007. 65. Beuckelmann DJ, Nabauer M, Erdmann E: Altera-
Phosphorylation-dependent regulation of ryano- 45. Stefan E, Wiesner B, Baillie GS, et al: Compart- tions of K+ currents in isolated human ventricular
dine receptors: A novel role for leucine/isoleucine mentalization of cAMP-dependent signaling by myocytes from patients with terminal heart failure.
zippers. J Cell Biol 153:699–708, 2001. phosphodiesterase-4D is involved in the regulation Circ Res 73:379–385, 1993.
26. Hulme JT, Lin TW, Westenbroek RE, et al: Beta- of vasopressin-mediated water reabsorption in 66. Varro A, Nanasi PP, Lathrop DA: Potassium cur-
adrenergic regulation requires direct anchoring of renal principal cells. J Am Soc Nephrol 18:199– rents in isolated human atrial and ventricular car-
PKA to cardiac CaV1.2 channels via a leucine 212, 2007. diocytes. Acta Physiol Scand 149:133–142, 1993.
67. Konarzewska H, Peeters GA, Sanguinetti MC: 73. Li GR, Feng J, Yue L, et al: Evidence for two com- 80. Zhang J, Wilson GF, Soerens AG, et al: Functional
Repolarizing K+ currents in nonfailing human ponents of delayed rectifier K+ current in human cardiomyocytes derived from human induced plu-
hearts: Similarities between right septal subendo- ventricular myocytes. Circ Res 78:689–696, 1996. ripotent stem cells. Circ Res 104:e30–e41, 2009.
cardial and left subepicardial ventricular myocytes. 74. Jost N, Virag L, Bitay M, et al: Restricting exces- 81. Yoshida Y, Yamanaka S: Recent stem cell advances:
Circulation 92:1179–1187, 1995. sive cardiac action potential and QT prolongation: Induced pluripotent stem cells for disease model-
68. Veldkamp MW, van Ginneken AC, Opthof T, et A vital role for IKs in human ventricular muscle. ing and stem cell-based regeneration. Circulation
al: Delayed rectifier channels in human ventricular Circulation 112:1392–1399, 2005. 122:80–87, 2010.
myocytes. Circulation 92:3497–3504, 1995. 75. Virag L, Iost N, Opincariu M, et al: The slow 82. Imredy JP, Penniman JR, Dech SJ, et al: Modeling
69. Amos GJ, Wettwer E, Metzger F, et al: Differences component of the delayed rectifier potassium of the adrenergic response of the human IKs
between outward currents of human atrial and sub- current in undiseased human ventricular myocytes. current (hKCNQ1/hKCNE1) stably expressed in
epicardial ventricular myocytes. J Physiol 491(Pt Cardiovasc Res 49:790–797, 2001. HEK-293 cells. Am J Physiol Heart Circ Physiol
1):31–50, 1996. 76. Charpentier F, Merot J, Loussouarn G, et al: 295:H1867–H1881, 2008.
70. Iost N, Virag L, Opincariu M, et al: Delayed recti- Delayed rectifier K(+) currents and cardiac repo- 83. Wang W, Xia J, Kass RS: MinK-KvLQT1 fusion
fier potassium current in undiseased human ven- larization. J Mol Cell Cardiol 48:37–44, 2010. proteins, evidence for multiple stoichiometries of
tricular myocytes. Cardiovasc Res 40:508–515, 77. Lu JT, Kass RS: Recent progress in congenital long the assembled IsK channel. J Biol Chem
1998. QT syndrome. Curr Opin Cardiol 25:216–221, 273:34069–34074, 1998.
71. Bosch RF, Gaspo R, Busch AE, et al: Effects of the 2010. 84. Wang K, Terrenoire C, Sampson KJ, et al: Bio-
chromanol 293B, a selective blocker of the slow, 78. Brunner M, Peng X, Liu GX, et al: Mechanisms of physical properties of slow potassium channels in
component of the delayed rectifier K+ current, on cardiac arrhythmias and sudden death in transgenic human embryonic stem cell derived cardiomyo-
repolarization in human and guinea pig ventricular rabbits with long QT syndrome. J Clin Invest cytes implicate subunit stoichiometry. J Physiol
myocytes. Cardiovasc Res 38:441–450, 1998. 118:2246–2259, 2008. 589:6093–6104, 2011.
72. Li GR, Yang B, Feng J, et al: Transmembrane ICa 79. Vidarsson H, Hyllner J, Sartipy P: Differentiation 85. Moretti A, Bellin M, Welling A, et al: Patient-
contributes to rate-dependent changes of action of human embryonic stem cells to cardiomyocytes specific induced pluripotent stem-cell models for
potentials in human ventricular myocytes. Am J for in vitro and in vivo applications. Stem Cell Rev long-QT syndrome. N Engl J Med 363:1397–
Physiol 276:H98–H106, 1999. 6:108–120, 2010. 1409, 2010.
Structural Determinants and
Biophysical Properties of hERG1
Channel Gating 12
Michael C. Sanguinetti, Martin Tristani-Firouzi, and Frank B. Sachse
structure of the highly conserved potassium selectivity filter and

CHAPTER OUTLINE
prevents K+ ion conductance. A few congenital mutations in
Background 121 HERG1 (N588K, T618I) induce a gain in channel function and
cause short QT syndrome6,7; these two mutations shift the voltage
Biophysical Properties of hERG1 Channels 121
dependence of channel inactivation to more positive potentials,
Markov Models of hERG1 Gating 124 greatly increase outward IKr, and thereby shorten the duration of
ventricular action potentials.
The Structural Basis of hERG1 Channel Gating 125
Biophysical Properties of hERG1 Channels

Background
The kinetics and voltage dependence of hERG1 currents are
In the mammalian heart, hERG1 K+ channels are largely respon- commonly studied in heterologous expression systems, such as
sible for terminal repolarization of action potentials, and con- mammalian cells transfected with cDNA or Xenopus oocytes
genital mutations of this channel are a major cause of long QT injected with cRNA. Currents are usually activated by pulsing to
syndrome, a life-threatening cardiac arrhythmia. The discovery test potentials from a negative holding potential. In the example
of the HERG1 (KCNH2) gene began with genetic and neuro- shown in Figure 12-1, A, an oocyte expressing hERG1 channels
physiologic studies of mutant flies. The eag (ether-a-go-go) locus was voltage-clamped to a holding potential of −80 mV and two-
of a mutant strain of the Mediterranean fruit fly (Drosophila mela- second test pulses were applied in 10-mV increments to voltages
nogaster) was associated with repetitive firing of motor neurons, between −70 and +50 mV. Channels were activated at potentials
an ether-induced leg-shaking phenotype, and altered K+ cur- greater than −60 mV, and the resulting whole-cell currents acti-
rents.1 Subsequent cloning and heterologous expression of eag vated slowly throughout the two-second test pulses in response
revealed that the fly gene encoded the first member of a novel to test potentials from −50 to −10 mV. Activation of hERG1
class of voltage-gated K+ channels. A low stringency screen and occurs with a time constant of approximately 150 ms at +10 mV.8
degenerate polymerase chain reaction was later used to identify At test potentials of 0 mV and greater, outward currents progres-
additional channel genes, including erg (eag-related gene) and elk sively activate faster and reach a smaller peak magnitude. These
(eag-like).2 The HERG1 gene (human ether-a-go-go–related currents are typically analyzed by plotting the peak outward
gene) encodes the hERG1 (KV11.1) protein, an α-subunit that current as a function of test potential. The resulting bell-shaped
coassembles to form functional hERG1 K+ channels that conduct current-voltage (I-V) relationship (see Figure 12-1, B) is the
the rapid delayed rectifier K+ current (IKr) in the heart.3 Two hallmark characteristic of IKr measured in the native cardiomyo-
highly related channels, hERG2 and hERG3, form K+ channels cytes of all species. The decrease in current magnitude associated
with similar biophysical properties, but these channels are pri- with more depolarized test potentials is caused by progressive
marily expressed in the central nervous system and not in the channel inactivation.
heart.4 After each test pulse, the cell was clamped to −70 mV to elicit
IKr is a major determinant of plateau phase duration and the channel deactivation, observed as a slowly decaying outward
rate of phases 2 and 3 repolarization of action potentials in human “tail” current (see Figure 12-1, A). A plot of the peak amplitude
and most nonrodent mammalian cardiomyocytes. hERG1 chan- of tail currents as a function of test potential (see Figure 12-1, C)
nels rapidly activate (open) and inactivate when the cardiomyo- defines the voltage dependence of current activation. The rela-
cyte is depolarized (e.g., during an action potential). Most channels tionship is fitted with a Boltzmann function to determine the
remain inactivated during the plateau phase, greatly reducing the half-point (V0.5, in this case −20 mV) for activation. The magni-
magnitude of outward IKr, an adaptation required for prolonged tude of outward tail current measured at −70 mV is larger than
action potential duration that typifies cardiomyocytes. During the outward current elicited by its preceding test pulse (see
phase 3 repolarization, hERG1 channels rapidly recover from Figure 12-1, A) despite the fact that the electrical driving force
inactivation and reenter their open state. A rapid rate of recovery for outward flux of K+ is smaller at −70 mV than it is for the more
from inactivation combined with a slow rate of deactivation depolarized test pulses. This paradox can be explained by the
(channel closure elicited by membrane repolarization) facilitates kinetics of channel inactivation versus deactivation. Immediately
rapid repolarization during phase 3 of the action potential. after the cell is repolarized to −70 mV, channels first rapidly
In the human heart, loss of function mutations in HERG1 recover from inactivation (i.e., reopen) before they slowly close
delay ventricular repolarization and cause type 2 long QT syn- (deactivate). The time constant for recovery from inactivation at
drome. Delayed ventricular repolarization increases the inci- −70 mV is approximately 10 ms (at room temperature), approxi-
dence of Torsades de pointes arrhythmia that can lead to syncope mately 30 to 100 times faster than deactivation at this potential.
and sudden death. The majority of HERG1 mutations disrupt Most importantly, because channels are far less inactivated at
folding or trafficking of channels to the plasma membrane.5 −70 mV compared with more depolarized test potentials, tail
Inherited mutations that alter channel gating are very rare. currents are actually larger than test currents despite the consid-
Examples include a point mutation (G638S) that disrupts the erably smaller electrical driving force.
121
1.0 I-V
Activation/ Deactivation
inactivation 0.8
Relative current
0.6
0.5
µA 0.4
0.2
1 sec
+50 mV
0.0
–70 mV –80 –60 –40 –20 0 20 40
–80
A B mV
1.0 1.0
Channel availability (inactivation)

0.8 0.8
Relative activation
V1/2: –40 mV
0.6 V1/2: –20 mV 0.6
0.4 0.4
0.2 0.2
0.0 0.0
–80 –60 –40 –20 0 20 40 –120 –80 –40 0 40
C mV D mV
Figure 12-1. Biophysical properties of hERG1 channel currents. A, Whole-cell currents (top panel) elicited in response to step changes in membrane voltage as indicated
(bottom panel). Currents were measured in Xenopus oocytes heterologously expressing hERG1 channels. B, Normalized I-V relationship for peak activating currents shown
in A. C, Voltage dependence of hERG1 channel activation based on plot of normalized peak tail currents illustrated in A. D, Voltage dependence of hERG1 channel
inactivation.
Two different voltage clamp protocols (two-step or three-step) pulse is varied and used to observe the onset of current inactiva-
have been used to characterize the rate of onset and voltage tion. Using this method, the time constants for inactivation vary
dependence of hERG1 inactivation. The two-step protocol between 16 ms at −20 mV and 2 ms at +50 mV.3,10
includes a prepulse to +40 mV, followed by a test pulse applied Ionic currents conducted by hERG1 channels can also be
to a variable potential.3 During the prepulse, channels reach a studied by recording single-channel currents, usually in the pres-
“fully-activated” condition (most channels are inactivated and ence of high [K+]e (e.g., 120 mM) to increase the signal-to-noise
only a minority remain open). The second (test) pulse is applied ratio. In cell-attached patches, single hERG1 channel activity is
to a voltage that is varied between −140 and +30 mV. The peak mostly absent at positive voltages (e.g., +40 mV in Figure 12-2,
initial current measured for test pulse is divided by the product A) because the probability of channel inactivation is very high.
of the maximum slope conductance and the driving force for K+ Upon repolarization to a negative potential, the channel reopens
(test potential − reversal potential), is normalized to a maximum after a variable delay (the period between the dotted line and
value of 1, and is plotted as a function of test voltage, Vt. The initial channel openings shown in Figure 12-2, A). The delay
resulting relationship is shown in Figure 12-1, D, and represents reflects the time required for the channel to recover from inac-
the voltage dependence of channel availability that varies from tivation and is shorter at more negative potentials. Once opened,
1.0 (no inactivation) to nearly 0 (all channels inactivated). The channels briefly close and reopen repetitively until they finally
second method to measure the voltage dependence of hERG1 enter a stable closed state. Analysis of these brief open and closed
inactivation uses a three-step voltage clamp protocol in which times indicates that single channels have at least two open and
channels are first activated/inactivated by a prepulse to a positive two closed states. At −90 mV, the mean open times are approxi-
potential (e.g., +40 mV). The prepulse is followed by a short mately 3 and 12 ms, and the mean closed times are approximately
(e.g., 10 ms) interpulse to a variable potential to allow channels 0.5 and 15 ms.11 The single-channel current amplitude (i) at each
to recover from inactivation but not appreciably deactivate. return potential is estimated by fitting data to a Gaussian distri-
Finally, a third pulse is applied to a fixed positive potential to bution of an all-points histogram of digitized current (see Figure
measure the relative proportion of channels that were in the open 12-2, B). The resulting single-channel current-voltage relation-
state at the end of the second pulse.9,10 The half-point for inac- ship (i-V) is then used to calculate slope conductance, which
tivation of hERG1 estimated with this protocol is approximately varies as a function of voltage (see Figure 12-2, C). The single-
−90 mV. The time course of hERG1 inactivation can be deter- channel conductance of hERG1 is approximately 12 pS when
mined by modification of the triple-pulse protocol. Here, the measured at negative potentials and decreases to 5 pS at positive
interpulse voltage is kept constant and the voltage of the final potentials when measured in the presence of high [K+]e.11
Structural Determinants and Biophysical Properties of hERG1 Channel Gating 123
+40 mV pA
12
3
–50 mV 0.4
–90 mV –120 –80 –40
Events (x104)
–110 mV 2 mV
40 80
–0.4
1
–0.8
–1.2
–1 0 1 2
1 pA
Amplitude (pA)
A 200 ms B C
Figure 12-2. Single hERG1 channel currents. A, Currents recorded from a cell-attached patch of a Xenopus oocyte overexpressing hERG1 channels. The arrow points to a
short opening of the channel at +40 mV. The dotted line indicates onset on repolarization to the test potential indicated in the upper panel. B, All-events histogram used
to determine single hERG1-channel current amplitude (I, ~0.9 pA) at −90 mV. C, Average i-V relationship for hERG1 channels for cell-attached patches ([K+]e = 120 mM).
Solid line indicates a slope conductance of 12 pS.
(Adapted from Zou A, Curran ME, Keating MT, et al: Single HERG delayed rectifier K+ channels in Xenopus oocytes. Am J Physiol 272:H1309–H1314, 1997.)
Vt
Vt –110 mV
–110 mV
1.0
+10 mV
0.8
Relative Q or G
0.6
0 mV
Q G
0.4
–10 mV 0.2
0.0
0.5 µA –20 mV
100 ms 100 nA –120 –80 –40 0 40
A B 100 ms C Vt (mV)
Figure 12-3. hERG1 channel gating currents. A, Outward hERG ionic currents recorded from Xenopus oocytes, elicited by depolarizing pulses from −50 to +30 mV (in 20-mV
increments); inward tail currents were measured after repolarization to −110 mV. Boxed area represents fast component of gating current that precedes onset of outward
ionic currents. B, Gating currents (red traces) in response to membrane depolarization to the indicated test potential (Vt) and upon return to −110 mV. Gating currents were
measured using the cut-open oocyte voltage clamp method in the absence of ionic currents. C, Plot of normalized charge (Q) and conductance (G) as a function of Vt for
hERG1 channels.
Unique insights into the mechanisms of hERG1 channel gating current represents intramembrane displacement of the
gating have been facilitated by measurement of gating currents. highly-charged S4 domain. Intramembrane charge displacement
As opposed to ionic currents, which result from movement of ions invoked by membrane depolarization is estimated by integration
through the channel pore from one side of the membrane to the of either the ON or OFF gating current elicited by depolarization
other, gating currents represent the intramembrane displacement or repolarization, respectively. In the example illustrated in Figure
of charged residues in the voltage sensors of the channel protein. 12-3, B, after a pulse to a variable Vt, OFF gating currents are
The unusual kinetics of hERG ionic current, with its slow activa- elicited by repolarization to −110 mV. The integral of the OFF
tion and rapid inactivation, imply that the underlying movement gating current is the charge (Q), and when plotted as a function
of the voltage sensor is different from most voltage-gated K (Kv) of Vt defines the Q-V relationship that occurs over a more nega-
channels, in which activation is fast and inactivation is relatively tive voltage range than ionic currents defined by the conductance-
slow. Gating currents can be observed as very small outward cur- voltage (G-V) relationship (see Figure 12-3, C). In simple terms,
rents that occur immediately after membrane depolarization and this means that voltage sensor movement occurs at more negative
before the relatively slow onset of ionic currents (Figure 12-3, A). potentials than that required for channel opening or closing.
Voltage clamp fluorimetry12 and direct measurement of gating KCNE2 encodes a 123 amino acid protein (MiRP1) with a
currents13 of hERG reveal two distinct components of charge single transmembrane domain that can serve as ancillary β-
displacement (see Figure 12-3, B). The slow component is approx- subunit for hERG1 to modulate channel expression and may14 or
imately 100-fold slower than Kv1 channel gating currents, carries may not15 alter hERG1 gating kinetics and response to certain
approximately 90% of the gating charge, and is associated with drugs. Like many Kv channels, hERG1 channels are also
the slow rate of hERG activation. The fast-gating component modulated by phosphatidylinositol-4,5-biphosphate (PIP2). This
likely represents rapid transitions between channel states during phosphoinositide increases current magnitude and slows
the early steps of the activation pathway. The majority of the deactivation.16
D
A
C
Figure 12-4. Schematics of Markov models for hERG1 channel gating. A, Five-state model with inactivation through open state. B, Five-state model with inactivation
through open and closed states. C, Two-compartment model comprising models for normal gating (green) and gating in the presence of an hERG1 channel activator (red).
D, Model for reconstruction of gating currents.
(C, Modified from Perry M, Sachse FB, Sanguinetti MC: Structural basis of action for a human ether-a-go-go-related gene 1 potassium channel activator, Proc Natl Acad Sci
U S A 104:13827–13832, 2007. D, From Abbruzzese J, Sachse FB, Tristani-Firouzi M, et al: Modification of hERG1 channel gating by Cd2+. J Gen Physiol 136:203–224, 2010.)
exception of α2 and β2. Voltage-dependent forward rates α and

Markov Models of hERG1 Gating backward rates β were defined as:
zα V m F
A variety of mathematical models have been developed to gain
α = α oe RT
insights into the function of hERG1 channels and their modula-
tion in disease or by drugs. Gating of hERG1 channels has been zβ V m F
−
described using Markov models or Hodgkin-Huxley–type for- β = β oe RT
mulations, with the former approach more prevalent than the

latter. Markov models describe systems such as hERG1 channels with the rates α0 and β0 at 0 mV, the charges zα and zβ, the mem-
by a finite number of states and transitions between them. Gener- brane patential Vm, the temperature T, Faraday constant F, and
ally, this modeling leads to a system of ordinary first-order dif- the gas constant R.
ferential equations, such as: Comparison of measured versus simulated current traces in a
subsequent study reinforced that a model with two closed states
dP ( t ) is inferior to models with three close states.19 Furthermore, dif-
= P (t ) Q
dt ferent variants of integration of the inactivated state into a model
with three closed states (C↔C↔C↔O) were investigated. In this
with the probability of states, P, and the matrix of transition rates, study, some single-channel recordings exhibited absence of
Q. Each matrix element Qij specifies the transition rate from the channel openings during depolarization, which were nevertheless
i-th to the j-th state. The major task of model development is the followed by openings during repolarization. Models with inacti-
definition of these states and parameterization of transition rates. vation from the final closed state, for example, the model pre-
Typical states of the hERG1 channels models are the closed (C), sented in Figure 12-4, B, appeared more adequate to reconstruct
open (O), and inactivated states (I). Transitions are commonly these recordings. However, a recent study suggested an alterna-
described with voltage-dependent or constant rate coefficients. tive interpretation of these recordings and introduced a model
Markov models of hERG1 channels originate from descrip- without direct inactivation from the final closed states, which can
tions of voltage-gated K+ channels, in particular, the Markov explain the experimental data.20 This study compared models of
models of IKr of ferret atrial myocytes17 and rabbit ventricular different topologies. The only model that effectively reproduced
myocytes.18 Studies by Wang et al8 and Kiehn et al19 aimed at experimental data was based on the topology presented in Figure
reconstruction of ion currents through hERG1 channels with 12-4, A.
hidden Markov models. In these studies, voltage-clamp protocols Several studies applied the model topologies presented in
were applied to characterize hERG1 channels expressed in Figure 12-4, A-B, to derive hERG1 channel models by parame-
Xenopus oocytes. Elicited current traces were described with terization of transition rates.21-23 Transition rates were commonly
parameters from fitting to exponential functions. Measured defined as voltage dependent, with the exception of α2 and β2.
current traces were compared with current traces simulated with Models of the topology presented in Figure 12-4, B, have been
Markov models of different topologies, that is, sets of states and implemented for several studies of hERG1 mutations. Clancy and
transitions. On the basis of this comparison, appropriate model Rudy derived hERG1 channel models of mutations linked to
topologies were identified and transitions rates were defined. long-QT syndrome (T474I and R56Q) and a mutation (N629D)
Analyses of the degree of sigmoidicity of current traces after causing reduced C-type inactivation and rendering the channel
activation of the channels suggested that models with three nonselective for cations.24 These models of hERG1 channels with
(C↔C↔C↔O) or four closed states (C↔C↔C↔C↔O) are mutations were integrated into computational models of guinea
more appropriate than models with one (C↔O) or two closed pig ventricular myocytes. Simulations with the cell models
states (C↔C↔O) to reconstruct measured data.8 Further analy- suggested prolonged action potential duration for T474I and
ses led to the development of a five-state model, accounting for R56Q, as well as increased probability of early afterdepolariza-
inactivation from the open state only (Figure 12-4, A). In this tions for R56Q and N629D. Rhodes et al used the same modeling
model, all transition rates were voltage dependent, with the approach to study several mutations linked to sudden infant death
syndrome.25 In particular, the heterozygous mutant hERG1-

The Structural Basis of hERG1 Channel Gating
12
R273Q exhibiting reduced current amplitude was investigated. A
model for hERG1-R273Q was created and integrated in a myocyte
model. Simulations suggested that hERG1-R273Q prolongs General Structure of hERG1 Channels
action potential duration and can cause early afterdepolarizations.
Adeniran et al studied effects of a mutation (N588K) linked to Similar to other Kv channels, functional hERG1 channels are
short QT syndrome.26 N588K is a gain-of-function mutation, and formed by coassembly of four α-subunits. Each hERG1 α-subunit
channels with this mutation exhibit increased current amplitudes. (1159 amino acids, mass of 127 kDa) has six transmembrane
N588K models based on the topologies presented in Figure 12-4, α-helical segments (S1 to S6) and very large cytoplasmic N- and
A-B, were derived and integrated in models of human ventricular C-termini (Figure 12-5, A). The S1 to S4 segments together form
myocytes. Simulations with models of subepicardial, midwall, and an independent voltage-sensing domain. The ion-conducting
subendocardial myocytes indicated that N588K reduces action pore domain includes the S5 and S6 segments, turret and pore
potential duration in all of these cell types. helices contributed by all four subunits. The S6 segments from
Markov models of hERG1 channels play an important role in each of the four subunits line the water-filled central cavity of the
studies of drug effects. The first Markov model of hERG1 chan- pore domain. Figure 12-5, B, presents a structural model, as
nels was developed for studies of dofetilide,27 an hERG1 open- viewed from the extracellular side, of an open Kv channel (based
channel blocker. The model comprised four closed states, an on Kv1.2) that illustrates the arrangement of the S5 and S6 seg-
open state, an inactivated state, and a nonconducting state, with ments to form the central pore domain and how the voltage
drug concentration−dependent transition from the open state sensors are situated on the periphery of the ion-conducting struc-
accounting for the reconstruction of drug effects. The model ture. The extracellular region that links each S5 to its pore helix
suggested that the apparent rate of block by the drug is modu- is called the turret, and this structure is especially large in hERG1
lated by the rate of inactivation. A recent study applied a two- channels. The C-terminal ends of the pore helices are tilted
compartment Markov model to investigate effects of RPR-260243, inward and each is connected to a short and highly conserved
an activator of hERG1 channels (see Figure 12-4, C).28 The peptide sequence (TSVGFG) that forms the selectivity filter. The
model is composed of two five-state models, presented in backbone carbonyl oxygen atoms of the selectivity filter residues
Figure 12-4, B. One of the models described normal hERG1 are oriented toward each other and coordinate the short-lived
channel gating; the other described gating in the presence of the binding of dehydrated K+ ions as they move inward or outward
activator. Drug-dependent transitions were defined between the through the open channel in response to their transmembrane
open and inactivated states of the two models. Three rate electrochemical gradient.32 On the basis of sequence similarity, it
constants were reduced versus the model of normal gating, is assumed that the basic topology of hERG1 is very similar to
which suggested that the drug affects both activation and inacti- the structure of other Kv channels, including the mammalian
vation. A similar two-compartment model was applied in a study KV1.2 channel, whose detailed structure has been determined by
on the antiarrhythmic drug sotalol, which is an inhibitor of x-ray crystallography.33-35 The Kv1.2 channel has been used as a
hERG1 current.29 Normal hERG1 channel gating and gating in template to develop useful homology models that facilitate design
the presence of sotalol were described by models of the topology and interpretation of structure-function studies of hERG1
presented in Figure 12-4, A. Both models were identically param- channels.
eterized.21 The open state in the drug model was replaced by a
nonconductive state. Drug-dependent transitions were defined
from the open state of the model for normal gating to the non- Activation and Deactivation of hERG1 Channels
conductive state. The two-compartment model was integrated in
models of human subepicardial, midwall, and subendocardial In Kv channels, including hERG1, the activation gate is formed
ventricular myocytes. The simulations suggested that sotalol by the C-terminal ends of the S6 helices (the so-called bundle
increases action potential duration in all cell types. Two Markov crossing). In the closed state, crisscrossing of the bundle crossing
models have been developed for reconstruction of gating currents creates a narrow aperture and a barrier to the flux of K+ between
of hERG1 channels.13,30 Previous studies established Markov the cytosol and the central cavity. Based on electron paramagnetic
models of gating currents of tetrameric voltage-gated Shaker resonance spectroscopy of bacterial KcsA channels36 and com-
potassium channels.31 In these models, gating current Igij associ- parison of the crystal structures of K+ channels in the open and
ated with transitions between the i-th state, Si, and the j-th state, closed states,35,37,38 channel opening in response to membrane
Sj, is calculated by: depolarization likely involves outward splaying of the S6 helices,
thereby widening the bundle-crossing aperture to allow ions to
Igij = ( zij ,α + zij ,β ) ( Siα ij − S j βij ) flow into or out of the central cavity. In most Kv channels, a
highly conserved Gly in the S6 segment acts as the hinge point
with the forward transition rate αij, backward transition rate βij, of the activation gate38; however, substitution of this residue in
and the charges zij,α and zij,β. hERG1 (Gly657) did not prevent channel opening, nor did it
In both hERG1 models, gating current of each channel alter the voltage dependence of activation.39 In hERG1, Ala653
subunit was described as a process having three states and two in the S6 segment faces the central cavity. When Ala653 is sub-
transitions (S0↔S1↔S2), with differing time constants for each stituted with an acidic or basic residue, charge repulsion between
transition. A schematic of the resulting Markov model is pre- these residues prevents collapse of the bundle crossing and keeps
sented in Figure 12-4, D. The model of Piper et al.13 extended the channel in the open state.40
this description with two closed, two open, and three inactivated The first four transmembrane segments of each Kv channel
states. This extension allowed reconstruction of potassium cur- subunit form a voltage-sensing structure. The S4 segment
rents conducted by hERG channels. The model of Abbruzzese sequence (GLLKTARLLRLVRVARK) contains four basic Arg
et al.30 split the description of gating currents (see Figure 12-4, (R) and two Lys (K) residues, most separated by two noncharged
D) from the description of potassium currents through the residues in the α-helix structure. The most critical S4 residue for
channel. The two models were coupled by definition of a transi- hERG gating appears to be Arg531 because neutralization of this
tion rate of the potassium current model that is dependent on residue most markedly increases the energy required for channel
states of the gating current model. Interestingly, the gating opening.41,42 Other mutagenesis studies have suggested more spe-
current model alone was able to reproduce major features of cific roles for several basic residues in S4 of hERG1. One study
measured gating currents. suggested that Lys525 stabilizes the closed state, Arg531
Pore
Voltage sensor (S5-S6)
(S1-S4)
T PH
+
+
S1 S2 S3 S4 S5 S6
+
+
S45L
PAS
A cNBD
B
Figure 12-5. Components of a single hERG1 subunit and structural model of a Kv channel. A, Diagram showing major components of a single hERg1 subunit, including
six α-helical transmembrane segments (S1-S6), the extracellular turret (T) and pore helix (PH), and the intracellular N-terminal PAS domain, S4-S5 linker (S45L) and C-terminal
cNBD. B, Model of Kv1.2 channel as viewed from the extracellular side of the membrane. Each subunit of the tetramer is individually color-coded.
(From Horn R: Uncooperative voltage sensors. J Gen Physiol 133:463–466, 2009.)
stabilizes the open state, and Arg534 participates in interactions

that stabilize pre-open closed states.43 Substitution of basic S4
residues with Trp led to the conclusion that Lys525, Arg528, and
Lys538 are especially important in mediating slow activation of
hERG1.44 As first proposed for Shaker K+ channels,45 movement
of the S4 in response to changes in transmembrane voltage is
facilitated, and intermediate closed states defined, by transient
formation of salt bridges between basic residues in S4 (especially
Arg53146) and acidic residues in adjacent S2 (Asp456, Asp460)
and S3 (Asp509) segments of the voltage-sensing domain in
hERG1.47,48
The coupling of voltage-sensor movement to channel
opening in hERG1 appears to involve interactions between the A S4-S5 B S4-S5
S4-S5 linker (S45L) and the S6 segment. The S45L, an amphipa- Figure 12-6. Model of Kv1.2 channel gating, highlighting the proposed role of the
thic α-helix is located near the C-terminal portion of S6.35 S4-S5 linker in channel activation. A, Pore domains of two subunits as viewed from
MacKinnon and colleagues34 proposed that the S45L functions the side. B, Pore domains of all four subunits as viewed from the cytoplasmic side
as a lever driven by voltage-induced changes in S4 that of the membrane. In both panels, the closed state of the channel is shown in red
nudges the S5/S6 helices to close the channel (Figure 12-6). and the open state is shown in blue.
Mutational analysis of the S45L49 and cross-linking of cysteine (From Pathak MM, Yarov-Yarovoy V, Agarwal G, et al: Closing in on the resting state of
residues introduced into the S45L and the C-terminal portion of the Shaker K+ channel. Neuron 56:124–140, 2007.)
the S6 segment50 further suggest that the S45L is the structural
link that couples voltage sensing to hERG1 channel opening and
closing.
The cytoplasmic N-terminal structure of each hERG1 subunit amphipathic α–helix. Similar to disruption of the PAS domain,
has 376 residues and residues 1 to 135 (the “eag” domain) have deletion of a portion of these initial regions or point mutations
been determined and shown to contain a Per-Arnt-Sim (PAS) of Arg5 or Gly6 accelerates hERG1 deactivation.56 These find-
domain.51 Each hERG1 channel subunit also contains a cyclic ings suggest that multiple regions of the N-terminus are required
nucleotide-binding domain (cNBD) within its cytoplasmic for the relatively slow deactivation of hERG1 channels.
C-terminus. The primary function of the PAS domain in hERG1 Alternative splicing of hERG1 RNA produces hERG1b, a
is unknown, but deletion or point mutations of the domain accel- protein with a much shorter and unique N-terminus compared
erates the rate of channel deactivation. It was initially proposed to the full-length hERG1 (known as hERG1a).57,58 The
that mutations in the N-terminus might disrupt its putative inter- N-terminus of hERG1b lacks a PAS domain and is only 36 amino
action with the S45L segment, an intracellular linker between the acids in length. Channels formed from hERG1b alone deactivate
S4 and S5 segments.52,53 More recently, it has been proposed that very rapidly but can be retained in the endoplasmic reticulum
the N-terminal PAS domain may instead interact with the cNBD (ER) because of an “RXR” ER retention signal present on the
to stabilize the open state (i.e., slow deactivation) of the N-terminus.59 Heteromultimeric hERG1a/1b channels are
channel.54,55 In addition to the PAS domain (residues 26-135), the expressed in the human ventricle60 and, unlike hERG1b channels,
initial region of the N-terminal domain has an unstructured traffic efficiently to the plasma membrane and deactivate with
region of nine amino acids followed by an 11-amino-acid kinetics that resemble native IKr measured in isolated myocytes.
As noted before, each hERG1 channel subunit contains a and elevated [K+]e,9 similar to C-type inactivation in Shaker.64
12
cNBD in its cytoplasmic C-terminus. Cyclic-nucleotide–gated Elevated [K+]e is likely to increase ion occupancy of the outermost
(CNG) channels and hyperpolarization-activated, cyclic nucleo- binding site, thereby restricting collapse of the selectivity filter.65
tide-gated (HCN) pacemaker channels also have a cNBD domain. Mutagenesis of hERG1 identified specific residues located in the
The binding of cyclic nucleotides (cAMP, cGMP) is required for turret, selectivity filter or pore helix, that greatly attenuate or
CNG channels to open and facilitates activation of HCN chan- eliminate inactivation, similar to other channels that inactivate by
nels by shifting the voltage dependence of activation to more a (usually slow) “C-type” mechanism caused by subtle rearrange-
negative potentials. In contrast, cAMP has only minor effects on ment of the selectivity filter. Most notably, combined mutation of
the gating of hERG1 channels, shifting the voltage dependence two residues within and just outside of the selectivity filter (G628C
of channel activation by a few millivolts.61 and S631C)9 or a single-point mutation in the pore helix (S620T)66
abolishes inactivation and makes the hERG1 I-V relationship
linear. The S631A mutation causes a +100-mV shift in the voltage
Inactivation of hERG1 Channels dependence of inactivation, with no significant change in the
voltage dependence of activation.67 Moreover, transfer of the
The properties of hERG1 channel inactivation are still being S3-S4 linker of a voltage-independent olfactory CNG K+ channel
explored, but characterization of mutant channels and molecular into hERG1 completely disrupts activation without altering the
dynamics simulation studies have provided remarkable insights voltage dependence of inactivation.68 Together, these findings
into the mechanisms and structural basis of this gating process. suggest that inactivation is not directly coupled to activation in
Cysteine accessibility experiments suggested that conformational hERG1. However, Markov modeling and the finding that S631A
rearrangements in the outer mouth of the pore mediate C-type– does not alter the kinetics of gating currents13 instead argue that
channel inactivation in Shaker channels.62,63 hERG1-channel hERG1 inactivation is coupled to inactivation, similar to that
inactivation is slowed by external tetraethylammonium (TEA) proposed for other Kv channels such as Shaker.69
associated with cardiac arrhythmia. Cell 97:175– 28. Perry M, Sachse FB, Sanguinetti MC: Structural
References 187, 1999. basis of action for a human ether-a-go-go-related
15. Weerapura M, Nattel S, Chartier D, et al: A com- gene 1 potassium channel activator. Proc Natl Acad
1. Ganetzky B, Wu CF: Neurogenetic analysis of parison of currents carried by HERG, with and Sci USA 104:13827–13832, 2007.
potassium currents in Drosophila: Synergistic without coexpression of MiRP1, and the native 29. Brennan T, Fink M, Rodriguez B: Multiscale mod-
effects on neuromuscular transmission in double rapid delayed rectifier current. Is MiRP1 the elling of drug-induced effects on cardiac electro-
mutants. J Neurogenet 1:17–28, 1983. missing link? J Physiol 540:15–27, 2002. physiological activity. Eur J Pharm Sci 36:62–77,
2. Warmke JW, Ganetzky B: A family of potassium 16. Rodriguez N, Amarouch MY, Montnach J, et al: 2009.
channel genes related to eag in Drosophila and Phosphatidylinositol-4,5-bisphosphate (PIP2) sta- 30. Abbruzzese J, Sachse FB, Tristani-Firouzi M, et al:
mammals. Proc Natl Acad Sci U S A 91:3438– bilizes the open pore conformation of the Kv11.1 Modification of hERG1 channel gating by Cd2+. J
3442, 1994. (hERG) channel. Biophys J 99:1110–1118, 2009. Gen Physiol 136:203–224, 2010.
3. Sanguinetti MC, Jiang C, Curran ME, et al: A 17. Liu S, Rasmusson RL, Campbell DL, et al: Activa- 31. Zagotta WN, Hoshi T, Aldrich RW: Shaker potas-
mechanistic link between an inherited and an tion and inactivation kinetics of an E-4031- sium channel gating. III: Evaluation of kinetic
acquired cardiac arrhythmia: HERG encodes the IKr sensitive current from single ferret atrial myocytes. models for activation. J Gen Physiol 103:321–362,
potassium channel. Cell 81:299–307, 1995. Biophys J 70:2704–2715, 1996. 1994.
4. Gutman GA, Chandy KG, Grissmer S, et al: Inter- 18. Clay JR, Ogbaghebriel A, Paquette T, et al: A 32. Zhou Y, Morais-Cabral JH, Kaufman A, et al:
national Union of Pharmacology. LIII. Nomencla- quantitative description of the E-4031-sensitive Chemistry of ion coordination and hydration
ture and molecular relationships of voltage-gated repolarization current in rabbit ventricular myo- revealed by a K+ channel-Fab complex at 2.0 A
potassium channels. Pharmacol Rev 57:473–508, cytes. Biophys J 69:1830–1837, 1995. resolution. Nature 414:43–48, 2001.
2005. 19. Kiehn J, Lacerda AE, Brown AM: Pathways of 33. Jiang Y, Lee A, Chen J, et al: X-ray structure of a
5. Delisle BP, Anson BD, Rajamani S, et al: Biology HERG inactivation. Am J Physiol 277:H199– voltage-dependent K+ channel. Nature 423:33–41,
of cardiac arrhythmias: Ion channel protein traf- H210, 1999. 2003.
ficking. Circ Res 94:1418–1428, 2004. 20. Bett GC, Zhou Q, Rasmusson RL: Models of 34. Long SB, Campbell EB, Mackinnon R: Voltage
6. Brugada R, Hong K, Dumaine R, et al: Sudden HERG gating. Biophys J 101:631–642, 2011. sensor of Kv1.2: Structural basis of electromechan-
death associated with short-QT syndrome linked 21. Fink M, Noble D, Virag L, et al: Contributions of ical coupling. Science 309:903–908, 2005.
to mutations in HERG. Circulation 109:30–35, HERG K+ current to repolarization of the human 35. Long SB, Campbell EB, Mackinnon R: Crystal
2004. ventricular action potential. Prog Biophys Mol Biol structure of a mammalian voltage-dependent
7. Sun Y, Quan XQ, Fromme S, et al: A novel muta- 96:357–376, 2008. Shaker family K+ channel. Science 309:897–903,
tion in the KCNH2 gene associated with short QT 22. Lu Y, Mahaut-Smith MP, Varghese A, et al: Effects 2005.
syndrome. J Mol Cell Cardiol 50:433–441, 2011. of premature stimulation on HERG K+ channels. 36. Perozo E, Cortes DM, Cuello LG: Structural rear-
8. Wang S, Liu S, Morales MJ, et al: A quantitative J Physiol 537:843–851, 2001. rangements underlying K+-channel activation
analysis of the activation and inactivation kinetics 23. Mazhari R, Greenstein JL, Winslow RL, et al: gating. Science 285:73–78, 1999.
of HERG expressed in Xenopus oocytes. J Physiol Molecular interactions between two long-QT syn- 37. Doyle DA: Structural changes during ion channel
502:45–60, 1997. drome gene products, HERG and KCNE2, ratio- gating. Trends Neurosci 27:298–302, 2004.
9. Smith PL, Baukrowitz T, Yellen G: The inward nalized by in vitro and in silico analysis. Circ Res 38. Doyle DA, Morais Cabral J, Pfuetzner RA, et al:
rectification mechanism of the HERG cardiac 89:33–38, 2001. The structure of the potassium channel: molecular
potassium channel. Nature 379:833–836, 1996. 24. Clancy CE, Rudy Y: Cellular consequences of basis of K+ conduction and selectivity. Science
10. Spector PS, Curran ME, Zou A, et al: Fast inactiva- HERG mutations in the long QT syndrome: Pre- 280:69–77, 1998.
tion causes rectification of the IKr channel. J Gen cursors to sudden cardiac death. Cardiovasc Res 39. Hardman RM, Stansfeld PJ, Dalibalta S, et al: Acti-
Physiol 107:611–619, 1996. 50:301–313, 2001. vation gating of hERG potassium channels: S6
11. Zou A, Curran ME, Keating MT, et al: Single 25. Rhodes TE, Abraham RL, Welch RC, et al: glycines are not required as gating hinges. J Biol
HERG delayed rectifier K+ channels in Xenopus Cardiac potassium channel dysfunction in sudden Chem 282:31972–31981, 2007.
oocytes. Am J Physiol 272:H1309–H1314, 1997. infant death syndrome. J Mol Cell Cardiol 44:571– 40. Brown S, Sonntag DP, Sanguinetti MC: A highly
12. Smith PL, Yellen G: Fast and slow voltage sensor 581, 2008. conserved alanine in the S6 domain of the hERG1
movements in HERG potassium channels. J Gen 26. Adeniran I, McPate MJ, Witchel HJ, et al: K+ channel is required for normal gating. Cell
Physiol 119:275–293, 2002. Increased vulnerability of human ventricle to Physiol Biochem 22:601–610, 2008.
13. Piper DR, Varghese A, Sanguinetti MC, et al: re-entrant excitation in hERG-linked variant 41. Piper DR, Hinz WA, Tallurri CK, et al: Regional
Gating currents associated with intramembrane 1 short QT syndrome. PLoS Comput Biol specificity of human ether-a’-go-go-related gene
charge displacement in HERG potassium chan- 7:e1002313, 2011. channel activation and inactivation gating. J Biol
nels. Proc Natl Acad Sci U S A 100:10534–10539, 27. Snyders DJ, Chaudhary A: High affinity open Chem 280:7206–7217, 2005.
2003. channel block by dofetilide of HERG expressed in 42. Zhang M, Liu J, Tseng GN: Gating charges in the
14. Abbott GW, Sesti F, Splawski I, et al: MiRP1 forms a human cell line. Mol Pharmacol 49:949–955, activation and inactivation processes of the HERG
IKr potassium channels with HERG and is 1996. channel. J Gen Physiol 124:703–718, 2004.
51. Morais Cabral JH, Lee A, Cohen SL, et al: Crystal 59. Phartiyal P, Sale H, Jones EM, et al: Endoplasmic
43. Subbiah RN, Clarke CE, Smith DJ, et al: Molecu-
structure and functional analysis of the HERG reticulum retention and rescue by heteromeric
lar basis of slow activation of the human ether-a-
potassium channel N terminus: A eukaryotic PAS assembly regulate human ERG 1a/1b surface
go-go related gene potassium channel. J Physiol
domain. Cell 95:649–655, 1998. channel composition. J Biol Chem 283:3702–3707,
558:417–431, 2004.
52. Wang J, Trudeau MC, Zappia AM, et al: The 2008.
44. Subbiah RN, Kondo M, Campbell TJ, et al: Tryp-
mechanism of N-terminal regulation of deactiva- 60. Jones EM, Roti Roti EC, Wang J, et al: Cardiac IKr
tophan scanning mutagenesis of the HERG K+
tion in HERG potassium channels. Biophys J channels minimally comprise hERG 1a and 1b sub-
channel: The S4 domain is loosely packed and likely
74:A254, 1998. units. J Biol Chem 279:44690–44694, 2004.
to be lipid exposed. J Physiol 569:367–379, 2005.
53. Fernandez-Trillo J, Barros F, Machin A, et al: 61. Cui J, Melman Y, Palma E, et al: Cyclic AMP regu-
45. Tiwari-Woodruff SK, Lin MA, Schulteis CT, et al:
Molecular determinants of interactions between lates the HERG K+ channel by dual pathways. Curr
Voltage-dependent structural interactions in the
the N-terminal domain and the transmembrane Biol 10:671–674, 2000.
shaker K+ channel. J Gen Physiol 115:123–138,
core that modulate hERG K+ channel gating. PLoS 62. Yellen G: The moving parts of voltage-gated ion
2000.
One 6:e24674, 2011. channels. Quart Rev Biophys 31:239–295, 1998.
46. Piper DR, Rupp J, Sachse FB, et al: Cooperative
54. Gustina AS, Trudeau MC: hERG potassium 63. Liu Y, Jurman ME, Yellen G: Dynamic rearrange-
interactions between R531 and acidic residues in
channel gating is mediated by N- and C-terminal ment of the outer mouth of a K+ channel during
the voltage sensing module of hERG1 channels.
region interactions. J Gen Physiol 137:315–325, gating. Neuron 16:859–867, 1996.
Cell Physiol Biochem 21:37–46, 2008.
2011. 64. Rasmusson RL, Morales MJ, Wang S, et al: Inac-
47. Liu J, Zhang M, Jiang M, et al: Negative charges
55. Muskett FW, Thouta S, Thomson SJ, et al: Mecha- tivation of voltage-gated cardiac K+ channels. Circ
in the transmembrane domains of the HERG K
nistic insight into human ether-a-go-go-related gene Res 82:739–750, 1998.
channel are involved in the activation- and
(hERG) K+ channel deactivation gating from the 65. Kiss L, Korn SJ: Modulation of C-type inactivation
deactivation-gating processes. J Gen Physiol
solution structure of the EAG domain. J Biol by K+ at the potassium channel selectivity filter.
121:599–614, 2003.
Chem 286:6184–6191, 2011. Biophys J 74:1840–1849, 1998.
48. Lin MC, Papazian DM: Differences between ion
56. Ng CA, Hunter MJ, Perry MD, et al: The 66. Ficker E, Jarolimek W, Kiehn J, et al: Molecular
binding to eag and HERG voltage sensors contrib-
N-terminal tail of hERG contains an amphipathic determinants of dofetilide block of HERG K+
ute to differential regulation of activation and deac-
alpha-helix that regulates channel deactivation. channels. Circ Res 82:386–395, 1998.
tivation gating. Channels (Austin) 1:429–437,
PLoS One 6:e16191, 2011. 67. Zou A, Xu QP, Sanguinetti MC: A mutation in the
2007.
57. London B, Trudeau MC, Newton KP, et al: Two pore region of HERG K+ channels reduces rectifi-
49. Ng CA, Perry MD, Tan PS, et al: The S4-S5 linker
isoforms of the mouse ether-a-go-go-related gene cation by shifting the voltage dependence of inac-
acts as a signal integrator for HERG K+ channel
coassemble to form channels with properties tivation. J Physiol 509:129–138, 1998.
activation and deactivation gating. PLoS One
similar to the rapidly activating component of the 68. Choveau FS, El Harchi A, Rodriguez N, et al:
7:e31640, 2012.
cardiac delayed rectifier K+ current. Circ Res Transfer of rolf S3-S4 linker to HERG eliminates
50. Ferrer T, Rupp J, Piper DR, et al: The S4-S5 linker
81:870–878, 1997. activation gating but spares inactivation. Biophys J
directly couples voltage sensor movement to the
58. Lees-Miller JP, Kondo C, Wang L, et al: Electro- 97:1323–1334, 2009.
activation gate in the human ether-a’-go-go-related
physiological characterization of an alternatively 69. Panyi G, Deutsch C: Cross talk between activation
gene (hERG) K+ channel. J Biol Chem 281:12858–
processed ERG K+ channel in mouse and human and slow inactivation gates of Shaker potassium
64, 2006.
hearts. Circ Res 81:719–726, 1997. channels. J Gen Physiol 128:547–559, 2006.
Molecular Regulation of Cardiac
Inward Rectifier Potassium Channels
by Pharmacologic Agents 13
José A. Sánchez-Chapula and Marcel A.G. van der Heyden
CHAPTER OUTLINE Inward Rectifier Channel Proteins are

Background 129
Encoded by KCNJ Genes
Inward Rectifier Channel Proteins Are Encoded KIR2.x Proteins Constitute the Classical Inward
by KCNJ Genes 129
Rectifier Current
Cardiac Role of Inward Rectifier Channels 131
The classical IK1 channels are formed by KIR2.x proteins, mainly
Pharmacology of KIR Channels 131 as homotetramers but probably also as heterotetramers. This
Conclusions and Outlook 136 subfamily consists of five members of which KIR2.1, 2.2, and 2.3,
encoded by KCNJ2, KCNJ12, and KCNJ4, respectively, are most
prominently expressed in the heart. The main subunits in the
working myocardium are KIR2.1 and to a lesser extent KIR2.2.
Expression levels of KIR2.3 are much lower. KIR2.x proteins are
Background almost absent in nodal tissues.
Each monomer comprises a transmembrane (TM) and a cyto-
The description in 1949 by Katz1 of the inward rectifier current plasmic part. The transmembrane domain (TMD) consists of the
that he considered “an interesting phenomenon, but difficult to two TM helices, the pore helix positioned between these, and a
explain” in muscle can be regarded as the scientific birth of slide helix located in front of TM1. The cytoplasmic pore domain
inward rectifier currents. Inward rectifier potassium currents (CPD) is compiled by interaction of the amino- and carboxyl-
(KIR) are characterized by their key feature of allowing more termini. In the resulting channel, the four TM2 (inner) helices
inward than outward potassium flow at equivalent membrane of each monomer face the central water-filled pore in the TMD,
potentials negative and positive of the potassium equilibrium with the potassium selectivity filter (SF) positioned in close prox-
potential. Inward rectifier channels are formed by tetrameriza- imity to the cellular exterior (see Figure 13-2). It is believed that
tion of the underlying pore-forming KIR channel proteins. Ion ion conduction may be regulated by two gates in series: one
channel regulation takes place mainly via the intracellularly formed by the inner helices (TM2) of the TMD at the helix-
located amino- and carboxyl-terminal regions of the proteins. bundle crossing and the other by the G loop at the apex of
Functionally, inward rectifiers define, for a large part, the mem- the CPD.2
brane potential of nonexcitable cells and the resting membrane Tetramerization of CPDs creates an extended pore region
potential of excitable cells. into the cytoplasm, separated from the TM pore region by the
The eukaryotic inward rectifier family is encoded by the G-loops. Polyamine block of inwardly rectifying potassium (KIR)
KCNJ genes and includes seven subfamilies (KIR1-7) (Figure channels underlies their key functional property of preferential
13-1). Each protein consists of two transmembrane domains, a conduction of inward K+ currents (inward rectification). Kinetic
short pore loop that harbors the potassium selectivity filter models describe polyamine block as a multistep process, incor-
(GYG) and intracellularly located N- and C-termini (Figure porating sequentially linked “shallow” and “deep” binding steps
13-2). Between subfamilies, sequence homology is as high as of polyamines in the KIR pore. Structurally, these shallow and
40%, whereas sequence identity rises to approximately 60% deep binding steps are conceptualized as an initial weakly voltage-
within subfamilies. Each of the subfamily gene products display dependent binding in the cytoplasmic domain of the channel,
their own characteristics, some having strong and others weak followed by a steeply voltage-dependent step in which spermine
rectifying potential and some responding to metabolic stimuli or migrates to a stable binding site in the TM pore (inner cavity).
neurotransmitters directly. Individual family members are Five negatively charged residues within the cytoplasmic (E224,
involved in many physiologic processes such as cardiac and neural D259, E299, F254, and D255) and one in the TM pore regions
excitability, insulin release, vascular tone, and potassium homeo- (D172) are involved in interacting with the positively charged
stasis.2 Members of three different subfamilies are main contribu- polyamines.5
tors to cardiac electrophysiology (Table 13-1). The classical
inward rectifier current (IK1) channel is formed by KIR2.x proteins Cardiac Acetylcholine Responsive Inward Rectifiers are Formed
and stands at the basis of the resting membrane potential in by KIR3.1 and KIR3.4 Heteromers
working cardiomyocytes; the acetylcholine (ACh) responsive rec- Functional cardiac ACh regulated inward rectifier current (IKACh)
tifiers are built of KIR3.x proteins and play an eminent role in channels are formed by heterotetramers of KIR3.1 and KIR3.4
heart rate modulation, whereas the adenosine triphosphate (also known as GIRK1 and GIRK4), encoded by the KCNJ3 and
(ATP)-sensitive channels are formed by octamers of KIR6.x and KCNJ5 genes, respectively. Expression is confined mainly to the
the regulatory sulfonylurea receptor (SUR) proteins and are atrium, and functional expression in the sinus and atrioventricu-
responsible for coupling the metabolic state of the vulnerable lar nodes is an important feature for heart rate and atrioventricu-
cardiomyocyte to its electrical activity.3,4 lar conduction regulation. Homotetramers of KIR3.1 or KIR3.4
129
produce less or no current when overexpressed in Xenopus IKATP Channels Are Octaheteromers of Four KIR6.x and Four
oocytes.6 It was established that a stochiometry of 1 : 1 and an SURX Channel Proteins
alternating position of KIR3.1 and KIR3.4 subunits within the Compared with IK1 and IKAch channels, an additional level of
channel produced most robust IKAch channel activity.2 complexity is found in the ATP-sensitive rectifier current (IKATP)
The structural organization of the IKAch channel is similar to channels, because a tetramer of SUR proteins embraces the
the classical IK1 channel. The TM2 domains face the water-filled
pore and, confined by the G-gate, an extended pore region pro-
trudes into the cytoplasm. A leucine residue within the C-terminus
of KIR3.1 (L333) and KIR3.4 (L339) appears crucial for Gβγ-
dependent activation. Interestingly, several domains involved in
Gβγ-dependent activation are formed at the cytoplasmic-facing Extracellular
interfaces of the KIR3.1 and KIR3.4 interacting domains.2
SF
TM D172
PD
Kir2.2
Kir2.6 Intracellular
Kir2.3 IK1
Kir2.1 GG E299
E224
Kir2.4
Kir5.1
Kir3.1
Kir3.3 D259
IAch
Kir3.4
Kir3.2
Kir6.1 D255 F254
IKATP
Kir6.2
Kir1.1a CPD
Kir1.1b
Kir4.1
Kir4.2
Kir7.1 Figure 13-2. Structure model of KIR2.1 ion channel. Negatively charged amino acid
residues (D172, E224, F254, D255, and D259) involved in polyamide-mediated rec-
KirBac1.4
tification face the transmembrane and cytoplasmic pore regions, respectively.
Figure 13-1. Cladogram of the human KCNJ gene family. Amino acid sequences CPD, cytoplasmic pore domain; GG, G-gate; SF, selectivity filter; TMPD, transmem-
were aligned by a Clustal-W algorithm, phylogeny was determined by the neighbor- brane pore domain.
joining method. Bacterial KIRBac1.4 was used as outgroup. Subfamilies relevant for
cardiac electrophysiology are color-coded: classical inward rectifier genes (green), (Used with permission from Van der Heyden MA, Sánchez-Chapula JA: Toward specific
acetylcholine-regulated channel genes (orange), and ATP-sensitive inward rectifier cardiac IK1 modulators for in vivo application: old drugs point the way. Heart Rhythm
genes (blue). 8:1076–1080, 2011.)
Table 13-1. Human Cardiac Inward Rectifier Genes, Expression, and Associated Cardiac Disease
Gene Location Protein Residues Main Expression Disease with Cardiac Involvement
KCNJ2 17q24.3 KIR2.1 427 Ventricle, atrium Andersen-Tawil syndrome, atrial fibrillation, short QT syndrome, CPVT
KCNJ12 17p11.2 KIR2.2 433 Ventricle, atrium —
KCNJ4 22q13.1 KIR2.3 445 Atrium, ventricle —
KCNJ3 2q24.1 KIR3.1 501* Atrium —
KCNJ5 11q24 KIR3.4 419 Atrium Long QT syndrome
KCNJ8 12p11.23 KIR6.1 424 Conduction system Sudden infant death syndrome, early repolarization syndrome, Brugada
syndrome, atrial fibrillation
KCNJ11 11p15.1 KIR6.2 390† Ventricle, atrium —
ABCC8 11p15.1 SUR1 1581 Atrium —
ABCC9 12p12.1 SUR2a 1549 Ventricle Atrial fibrillation, dilated cardiomyopathy, Cantú syndrome
SUR2b 1549 Conduction system
*Longest isoform. Shorter isoforms of 235, 253, and 308 amino acids have been reported.
†
Longest isoform. Shorter isoform of 303 amino acids has been reported.
CPVT, Catecholaminergic polymorphic ventricular tachycardia.
Molecular Regulation of Cardiac Inward Rectifier Potassium Channels by Pharmacologic Agents 131
KIR-formed channel.7 Cardiac SUR protein–expressing genes

IKATP Channels Confer Metabolic Status to
13
ABCC8 and ABCC9 are members of the large ATP-binding cas-
sette (ABC) superfamily that consists of 48 genes divided into Electrical Properties
seven subfamilies.8 SUR1 (ABCC8) and SUR2 (ABCC9) contain
the typical ABC “core” that consists of two membrane domains The ATP-sensitive inward rectifier channels are metabolic
(TMD1 and TMD2) of six TM helices, each separated by two sensors that couple the cardiac metabolic state to electrophysi-
cytoplasmic nucleotide–binding domains (NBD1 and NBD2). ologic properties. Under physiologic conditions, ATP favors
Furthermore, the proteins contain an additional N-terminal– closure, whereas ADP levels determine opening of the channels.
located membrane domain consisting of five TM helices (TMD0) Upon metabolic challenges, for example as a result of cardiac
that are linked to the “core” domain by the L0 linker. Interaction ischemia, decreased ATP/ADP ratios result in activation of
with the KIR6.x-formed pore is established mainly by the TMD0 the channel, thereby preserving normal resting potential and
and L0 domains. Nucleotide binding drives dimerization of shortening the cardiac action potential and the period of
NBD1 and NBD2, which results in subsequent rearrangements cell contraction. These processes aid in ATP preservation
of TMD1 and TMD2 and alterations in IKATP channel gating. during circumstances that challenge the vulnerable cardiac
The pore-forming KIR6.x complex resembles the structures of myocyte.
KIR2.x and KIR3.x channels. Cardiac IKATP channels, depending on KIR6.1 knockout mice die suddenly after the occurrence of ST
their location in the heart, consist of octaheteromers of four elevation and subsequent third-degree atrioventricular block and
KIR6.2 (KCNJ11) and four SUR2a subunits (ventricle), KIR6.2 cessation of heart rhythm, which may result from myocardial
and SUR1 (atrium), or KIR6.1 (KCNJ8)/KIR6.2 and SUR2b (con- ischemia as a result of blunted coronary vasoregulation.17 The
duction system). presence of KIR6.1 in the cardiac conduction system may provide
another or additional mechanism of cardiac pathology and con-
duction disturbances in these animals. Furthermore, KIR6.1
knockout mice are more susceptible to endotoxin-mediated
Cardiac Role of Inward Rectifier Channels septic shock, resulting in ST elevation and cardiac death. In the
hearts of KIR6.2 knockout mice, a number of stressors induce
KIR2.x Channels Set the Resting Membrane calcium overload, followed by short-term abnormalities in cardiac
Potential and Contribute to Repolarization rhythm (early afterdepolarization, premature ventricular contrac-
tions, and sinus node regulation) and hemodynamics (increased
In the working myocardium, KIR2.x-mediated IK1 sets the resting left ventricular end-diastolic pressure).18,19 In the long term,
membrane potential close to the potassium reversal potential and cardiac maladaptive remodeling (exaggerated fibrosis and hyper-
contributes outward potassium current during repolarization. trophy) finally results in impaired ejection fraction and conges-
With respect to the heart, neonatal KIR2.1 knockout mice were tive heart failure.
bradycardic and displayed electrocardiogram (ECG) abnormali- SUR1 knockout mice are fertile and phenotypically normal
ties such as lengthening of the PR (~50%) and QT (~50%) but lack atrial IKATP.20 Furthermore, ABCC8 null mutation
intervals. Isolated cardiomyocytes from KIR2.1 knockout animals increases resistance to cardiac, but not neuronal, ischemia-
displayed (1) no IK1; (2) action potential prolongation; and (3) reperfusion injury.21 Knock out of SUR2 resulted in loss of IKATP
spontaneous beating activity.9 In contrast, KIR2.2 knockout mice from all muscles, coronary vasospasms, increased blood pressure,
were viable and lived through adulthood without ECG abnor- and premature death from week 5 onward. Animals displayed
malities. However, their cardiomyocytes displayed a 50% reduc- slight cardiac hypertrophy and increased resistance to cardiac
tion in IK1 densities. In humans, dominant negative KCNJ2 ischemia-reperfusion damage.22
mutations are associated with Andersen-Tawil syndrome 1, char- In humans, a gain-of-function mutation in KCNJ8 has been
acterized by periodic muscle paralysis, biventricular tachycardia associated with early repolarization and Brugada syndrome, both
and sometimes long QT times, and developmental bone abnor- characterized by specific (J-wave) ECG abnormalities and atrial
malities (see Table 13-1).10 Gain-of-function mutations in KCNJ2 fibrillation.23-26 KCNJ8 loss of function has been associated with
have been associated with short QT syndrome, catecholaminer- sudden infant death syndrome.27 ABCC9 mutations have been
gic polymorphic ventricular tachycardia, and congenital atrial linked to atrial fibrillation, dilated cardiomyopathy, and Cantú
fibrillation.11-13 Currently, no disease-causing mutations in syndrome, which causes cardiomegaly and pericardial effusion,
KCNJ12 and KCNJ4 have been described. among other effects (see Table 13-1).28-30 Thus far, KCNJ11 and
ABCC8 mutations have not been associated with cardiac
phenotypes.
IKACh Channels and Heart Rate Regulation
IKACh slows heart rate and atrioventricular conduction in response
to muscarinic M2 receptor activation by ACh released from the Pharmacology of KIR Channels
vagus nerve. In this process, Gβγ directly interacts with and
activates the channel that thereby antagonizes diastolic depolar- Effects of Drugs on KIR Channels
ization of the nodal cardiomyocytes. KCNJ3–/– mice were viable
but showed a lack of carbachol-induced IKACh in atrial cardiomyo- The most commonly used blockers to examine the physiologic
cytes. Moreover, animals showed blunted responses to pharma- roles of KIR channels in native cells and tissues are Ba2+ and Cs+.
cologic interventions directed at vagal-induced bradycardia.14 At micromolar concentration, Ba2+ blocks KIR channels with rela-
KCNJ5–/– mice15 displayed loss of cardiac IKAch. Mice were viable; tive specificity. Externally applied Ba2+ suppresses KIR currents in
in most studies their resting heart rate increased and animals a voltage-dependent manner and Ba2+ inhibits KIR channels more
showed a reduced negative chronotropic response after vagal strongly as the membrane is hyperpolarized.2 However, experi-
stimulation. Furthermore, mice were unable to undergo rapid ments have shown that BaCl2 infusion in vivo produces strong
changes in heart frequency. In human heart function, KCNJ5 deleterious effects—mainly skeletal, cardiac, and smooth muscle
mutations have been associated with long QT syndrome, syncope, dysfunction.5
atrial fibrillation, and sudden cardiac death (see Table 13-1).16 No In the intact heart, IK1 blockers produce membrane depolar-
diseases related to mutations in KCNJ3 have been reported. ization (an effect that slows conduction velocity because of a
voltage-dependent inactivation of Na+ channels), cause prolonga- action on the SUR subunit of the channel. KCOs that activate
tion of the QT and QRS intervals on the surface ECG, and at cardiac IKATP channels include cromakalim, pinacidil, nicorandil,
higher concentrations cause ventricular ectopy and lethal ven- and diazoxide. KCOs act by opening IKATP channels by decreasing
tricular arrhythmias. Conversely, it has been demonstrated that the threshold concentrations of intracellular nucleotides for
increasing inward rectifier currents stabilize reentrant arrhyth- opening of KIR6.2/SUR2a channels, and they also increase the
mias. Recently, it has been shown that pharmacologic inhibition maximum amplitude.2
of inward rectifier currents by chloroquine induces antifibrilla- The rest of this chapter focuses on pharmacologic agents
tory effects.31 Different antiarrhythmic (e.g., quinidine, amioda- that inhibit or activate KIR channels acting directly on the
rone, propafenone, disopyramide) and noncardiac (e.g., channel pore–forming KIR subunits and their underlying blocking
terfenadine, astemizole) drugs at micromolar concentrations have mechanisms.
been found to inhibit the classical cardiac IK1. However, little is
known about their mechanisms of inhibition.32
Gambogic acid (GA) is a potent inducer of apoptosis that has Pore-Blocking Drugs
been considered for anticancer therapy. At 10 µM, GA inhibited
KIR2.1 current differentially during acute applications that pro- Several compounds that partially mimic polyamine binding to IK1
duced 30% of inhibition or during prolonged exposures (up to 3 channels have been identified over the last several years. Block of
hours), which produced 70% of inhibition. The IC50 of KIR2.1 small molecules such as chloroquine, chloroethylclonidine, pent-
inhibition by GA during prolonged exposures was 27 nM. GA– amidine, and quinacrine have been shown to inhibit IK1 and
induced inhibition is not caused by changes in biogenesis or heterologously expressed KIR2.x channels in a voltage-dependent
trafficking to the plasma membrane. It is a lipophilic molecule fashion.37
(LogP ~6.77) that can insert itself into the lipid bilayer and In cardiac myocytes, in contrast to the voltage-independent
thereby might perturb association of KIR2.1 with regulatory lipids mode of IK1 inhibition induced by most of the tested drugs, the
and proteins that specifically facilitate the function of KIR2.1 antimalarial drug chloroquine was found to inhibit IK1 and IKACh
channels. It has been found that pretreatment of HEK293 cells in a voltage-dependent manner, with less pronounced blockade
with GA shifted KIR2.1 and Kv2.1 channels from the at negative test potentials. In addition, unblock was achieved by
Triton X-100–insoluble to the Triton X-100–soluble fraction. hyperpolarizing pulses to potentials negative to the current rever-
GA decreasing KIR2.1 activity is consistent with KIR2.1 sal potential (Figure 13-3). This profile of voltage dependence is
being physiologically active in the context of Triton X-100– consistent with a positively charged molecule blocking the
insoluble microdomains, but not of Triton X-100–soluble channel from the intracellular side and entering the pore to such
microdomains.33 A different drug, the cell-permeable dienone an extent as to be subjected to the transmembrane electrical field.
phenol triterpene celastrol, a neuroprotective compound that is The blocking effects of chloroquine on IK1 caused an increase
also a slow-acting compound, decreased the rate of ion channel in the duration of the ventricular action potential, depolarization
transport and caused a reduction of channel density on the cell of the resting membrane potential, and increase in Purkinje
surface. fibers’ automaticity.38
Muscarinic ACh receptors in the peripheral nervous system Recent studies with small molecules such as chloroquine and
are found primarily on autonomic effector cells innervated by pentamidine revealed the molecular basis by which these drugs
postganglionic parasympathetic nerves. In the heart, the musca- block KIR2.1 in a voltage-dependent manner. Through site-
rinic receptor M2 is the predominant subtype, which is coupled directed mutagenesis and molecular modeling studies, chloro-
to the Gi/o protein family. IKACh channels can be activated by quine and pentamidine were found to block the KIR2.1 channel
muscarinic receptor agonists like cholinomimetic choline esters by plugging the center of the cytoplasmic conduction pathway
(ACh, methacholine, carbachol, and bethanechol). The musca- via interference with negatively charged residues (E224, D259,
rinic receptor antagonist alkaloids such as atropine and synthetic and E299) and for chloroquine with an additional aromatic
and semisynthetic compounds prevented the effects of ACh by residue (F254), which have previously been shown to be involved
blocking its binding to muscarinic receptors.34 in polyamine block (see Figure 13-2).31,38-40 Comparative molecu-
Some antiarrhythmic drugs are known to antagonize at micro- lar modeling and ligand docking of the three-dimensional struc-
molar concentration actions of ACh on IKACh. Evidence from tures of quinidine and chloroquine in the intracellular domain of
different authors suggests that the mechanisms underlying the KIR2.1 predicted that chloroquine effectively blocks potassium
anti-ACh effect of antiarrhythmic drugs at micromolar levels are flow by binding at the center of the ion permeation vestibule of
different among such drugs—that is, quinidine, pilsicainide, diso- KIR2.1. In contrast, quinidine binds the vestibular side, only par-
pyramide, d,l-sotalol, propafenone, cibenzoline, and amiodarone tially blocking ion movement.41 An interesting finding was that
mainly block the muscarinic ACh receptors—whereas flecainide, neutralization of the D172 residue, which faces the transmem-
verapamil, and propafenone inhibit the K+ channel itself as open brane pore and is key in the polyamine block of KIR2.1 (termed
channel blockers. Some relatively novel antiarrhythmic drugs rectifying controller), does not affect chloroquine KIR2.1-blocking
have been found to inhibit IKACh at nanomolar concentrations. potency. A gain-of-function mutation in KCNJ2 was recently
Dronedarone inhibited IKACh in single cells isolated from the reported to cause one form of short QT syndrome (SQT3). A
sinoatrial node with an IC50 of 63 nM.35 The benzopyrene deriva- missense mutation (D172N) located within the transmembrane
tive NIP-151 potently inhibited the heterologously expressed (in ion conduction pathway of KIR2.1 was found in a father and
HEK293 cells) channel KIR3.1/3.4, with an IC50 of 1.6 nM. In daughter with short QT interval.11 The biophysical consequences
both cases, it was suggested that dronedarone and NIP-151 of D172N mutation include a reduction in the degree of current
inhibited IKACh channels by a direct inhibitory interaction with rectification at depolarized potentials. Interestingly, chloroquine
the channel protein or by disrupting the G-protein–mediated caused a dose- and voltage-dependent reduction in wild type
activation.36 (WT), D172N, and WT-D172N heteromeric KIR2.1 current. As
There are two major classes of therapeutic compounds that expected, the potency and kinetics of chloroquine block of
target the IKATP channels: sulfonylureas and IKATP channel openers D172N and WTD172N KIR2.1 current were similar to WT. In
(KCOs).2 Sulfonylureas on pancreatic β cells bind to SUR1 silico modeling of the heterozygous WT-D172N KIR2.1 condi-
channel subunits and block IKATP channels, causing membrane tion predicted that 3 µM chloroquine would normalize inward
depolarization and an influx of Ca2+ channels, stimulating insulin rectifier K+ current magnitude, action potential duration, and
release and hypoglycemia.2 KCOs activate IKATP channels via their effective refractory period. This suggested that therapeutic
20
1.0
13
IK1 IK1
10
IK1
1-(IChloroquine-IControl)
0.8
Current (pA/pF)
Control 3 µM Chloroquine 0
–120 –100 –80 –60 –40 –20 0 0.6
–10
Voltage (mV)
–20 0.4
0.5 nA
–30 Control 0.2

–40 3 µM Chloroquine
IC50 = 0.69 µM ± 0.09
200 ms 0.0
–50
0.01 0.1 1 10 100
A B C Chloroquine (µM)
20
1.0
IKACh IKACh IKACh
1 µM Carbachol
10 0.8
Current (pA/pF)
Control 3 µM Chloroquine
0 0.6
–120–100 –80 –60 –40 –20 0 20 40
0.2 nA
–10 Voltage (mV) 0.4

Control
200 ms 0.2
–20 1 µM Carbachol
IC50 = 0.38 µM ± 0.04
3 µM Chloroquine
0.0
–30 0.01 0.1 1 10 100
D E F Chloroquine (µM)
30
IKATP 1.0
100 µM Pinacidil IKATP IKATP
20
0.8
Current (pA/pF)
Control 3 µM Chloroquine 10
0.6
0
–120–100 –80 –60 –40 –20 0 20 0.4
–10 Voltage (mV)
0.2
1 nA
–20 Control
100 µM Pinacidil IC50 = 0.51 µM ± 0.08
200 ms 0.0
–30 3 µM Chloroquine 0.01 0.1 1 10 100
G H I Chloroquine (µM)
Figure 13-3. Blocking efficacy of chloroquine in different inward rectifier channels. Voltage protocol: steps from −120 mV to +20 mV. A, IK1 current traces in ventricular
myocytes. B, I/V relationship from 5 cells. C, Dose-response curve of blocked IK1 at −60 mV. D, Current traces in atrial myocytes in the presence of 1 µM carbachol to activate
IKACh. E, I/V relationship from five cells. F, Dose-response curve of blocked IKACh at −60 mV. G, Current traces in ventricular myocytes in the presence of 100 µM pinacidil to
activate IKATP. H, I/V relationship from five cells. I, Dose-response curve of blocked IKATP at −60 mV.
(Used with permission from Noujaim SF, Stuckey JA, Ponce-Balbuena D, et al: Specific residues of the cytoplasmic domains of cardiac inward rectifier potassium channels are
effective antifibrillatory targets. FASEB J 24:4302–4312, 2012.)
concentrations of chloroquine might lengthen cardiac repolariza- cavity near the membrane side of the protein, leaving a pathway
tion in SQT3.42,43 of 10 Å in diameter, which most likely will not impede the passage
Chloroquine also blocked other cardiac inward rectifier chan- of potassium ions.
nels. In patch-clamp experiments, chloroquine blocked IK1 (IC50 The honey bee venom–isolated toxin tertiapin and the
= 0.69 µM), IKACh (IC50 = 0.38 µM), and IKATP (IC50 = 0.51 µM) oxidation-resistant tertiapin-Q derivative are potent inhibitors of
(see Figure 13-3). Comparative molecular modeling and ligand KIR3.1/3.4 and KIR1.1 channels with nanomolar affinity. In iso-
docking of chloroquine in the intracellular domains (cytoplasmic lated cardiac myocytes, tertiapin Q blocks IKACh without affecting
pore) of KIR2.1, KIR3.1, and KIR6.2 suggested that chloroquine IKATP, IK1, and cardiac potassium voltage–dependent currents like
blocks or reduces potassium flow by interacting with negatively Ito, IKr, or IKs and has shown to have antiarrhythmic efficacy in
charged amino acids facing the ion permeation vestibule of the dogs. Tertiapin binds to the outer vestibule of the conduction
channel in question. However, these detailed analyses revealed pore formed by the linker between the first and second trans-
important differences. The molecular modeling indicated that membrane (M1–M2) segments. An α-helical structure within the
chloroquine binds more effectively to the cytoplasmic domains toxin interacts with a short sequence of aromatic residues located
of KIR2.1 and KIR3.1 than to that of KIR6.2.31 Chloroquine also in the N-terminal part of the linker that confers high affinity for
seemed to interact with acidic residues in the cytoplasmic domain tertiapin.35
of KIR3.1 and KIR3.4; however, the negative electrostatic potential The detailed information provided here provides the struc-
in the channel pore was lower than in KIR2.1, resulting in partial tural framework for the design of safer, alternative compounds
channel block. In KIR6.2, the molecular modeling and ligand that are devoid of KIR channel–blocking properties. Conversely,
docking of chloroquine in the intracellular domains (cytoplasmic these basic science insights may be useful in the development of
pore) of 6.2 suggested that the drug binds to the side of a large novel and potentially useful antifibrillatory pharmacophores. The
chloroquine-binding models for KIR channels provide a starting applications, these drugs share several common physicochemical
point for feasible approaches in the generation of new therapies. properties. CADs typically contain a hydrophilic domain consist-
ing of one or more primary or substituted nitrogen groups posi-
tively charged at physiologic pH and a hydrophobic domain
Drugs That Interfere in PIP2–Channel Interaction consisting of an aromatic and/or aliphatic ring structure, which
may be substituted with one or more halogen moieties. This
Several polyvalent cations, such as polyamines, trivalent metals, results in polar hydrophilic cationic side-chain and apolar ring
neomycin, and polylysine, inhibited KIR channels by screening systems within one molecule. CADs bind to the negatively
the negatively charged head groups of phosphatidylinositol charged phosphoinositides. The cationic group of CADs is nor-
4,5-bisphosphate (PIP2). PIP2 comprises approximately 1% of mally placed between the polar head groups of phospholipids,
plasma membrane phospholipids and is required in many plasma and the hydrophobic portion is directed toward the hydrophobic
membrane processes, including the function of ion channels and interior of the membrane; thus, the drug molecule intercalates
transporters.44,45 Activation by PIP2 is a common feature of all between lipid molecules.
KIR channels; after excision of an inside-out membrane patch Several widely used CADs have been associated with inward
from a cell expressing KIR channels, current amplitude diminishes rectifier current disturbances, and recent evidence points to
over time. Such a rundown of the current results from PIP2 interference of the channel–PIP2 interaction as the underlying
depletion from the cell membrane. The K+ currents can be mechanism of action.50-55 The inhibitory potency of CADs on the
restored partially by exposing the cytoplasmic face of the patch different subfamilies of KIR channels correlates with the sensitiv-
to PIP2. Each PIP2 molecule consists of an inositol head group ity of the channel to rundown arising from PIP2 depletion. Con-
and fatty-acid side chains inserted into the membrane. The nega- sistent with this, channels relatively resistant to rundown mediated
tively charged PIP2 head groups are anchored just below the by PIP2 depletion, such as KIR1.1, 2.1, and 2.2 (“high PIP2 affin-
plane of the membrane by the lipid tails of the molecule. It has ity”), are less sensitive to CADs, in contrast to KIR2.3, 3.x, and
been proposed that the channel opens when the negative-head 6.2, which are more sensitive.
charges of PIP2 interact with positively charged residues in the Experimentally, interference of the drug with the interaction
cytoplasmic domain, near the plasma membrane.46,47 of KIR2.x, KIR3.1/3.4, and KIR6.2/SUR2a channels and PIP2 is
The x-ray crystal structure of a chicken KIR2.2 channel in a suggested from four sources of evidence: (1) slow onset of current
complex with a short-chain (dioctanoyl) derivative of PIP2 has block when CADs are applied from either the inside or outside
been resolved recently.48 A simple allosteric mechanism of gating of the channel; (2) mutation of KIR2.3 (I213L) and KIR6.2 (C166S)
control by PIP2 appears to be responsible for opening the gate at increases their affinity for PIP2 and lowers their sensitivity for
the TM2 helix-bundle crossing. One PIP2 molecule binds to each CADs; (3) mutations of KIR2.1 (L222I and K182Q), which
of the four channel subunits at an interface between the TMD decrease its affinity for PIP2, increase its sensitivity for CADs
and the CPD. On PIP2 binding, the entire CTD translates 6 Å (Figure 13-4); and (4) coapplication of CADs with PIP2 lowers
toward the TMD and becomes tethered to the TMD, and the CAD-mediated current inhibition.50-55
inner helix gate starts to open. The PIP2-binding site comprises Bupivacaine was the first CAD in which the mechanism of
amino acids from the TMD and the CPD, the acyl chains, glyc- inhibition of the inward rectifier channel KIR3.2 by interference
erol backbone, and 1′ (phosphodiester) phosphate of PIP2 inter- with the PIP2–channel interaction was proposed. Bupivacaine is
act with the TMD, whereas the inositol head group makes a local anesthetic drug with a long duration of action that pro-
interactions with the CPD. The acyl chains insert into the mem- duces excellent sensory anesthesia. Bupivacaine produced inhibi-
brane layer and interact with hydrophobic amino acids on both tory effects on G protein–gated KIR3.x channels expressed in
the inner and outer helices, whereas the 1′ phosphate makes Xenopus laevis oocytes but did not affect KIR1.1 or KIR2.1 chan-
interactions with amino acids forming the sequence arginine- nels. The effects of bupivacaine on KIR3.x channels was indepen-
tryptophan-arginine (amino acids 78-80 in KIR2.2). The inositol dent of the method of channel activation, by activation of the
ring of the head group is oriented toward the CPD, where the 4′ muscarinic receptor or directly via coexpressed G protein Gβγ
and 5′ phosphates are positioned to interact directly with K183, subunits or ethanol, which suggested that bupivacaine’s site of
R186, K188, and K189 of the CPD. Other amino acids on the action was at the channel protein itself.50 On the basis that KIR1.1
tether helix, including R190, participate in the formation of a and 2.1 are high-affinity PIP2 binders and the KIR3x subfamily
hydrogen-bonding network that seems to strengthen the interac- channel shows a high sensitivity to PIP2 depletion, it was pro-
tion between the tether helix and other regions of the CPD. posed that bupivacaine antagonized the interaction of PIP2 with
Comparing the KIR2.2 amino acids involved in binding to PIP2 KIR3.x channels, thereby decreasing the current. Consistent with
by sequence alignment with the other members of the KIR channel this hypothesis, it was found that KIR3.2(I234L) and 3.1(M223L)/
family shows that the amino acids are highly conserved among KIR3.2(I234L) mutant channels, which have increased affinity for
the large family. PIP2, were less sensitive to bupivacaine.50
A variety of signaling partners influence KIR channel activity The first-generation antihistamines, mepyramine and diphen-
by modulating the interaction of KIR channels with PIP2. Some hydramine, are H1 competitive receptor antagonists. Both drugs
of the results strongly suggest that the apparent affinity of reduced the KIR2.3 current amplitude by 25% and 17% at
channel-PIP2 interactions is a key element in determining modu- 100-µM concentrations, whereas KIR2.1 current was insensitive
lation of KIR channels, whereby the strength of channel-PIP2 to both antihistamines.56 Furthermore, KIR3.4 current was inhib-
interactions controls the sensitivity of KIR channels to regulation ited by mepyramine similarly to KIR2.3. The inhibitory effects of
by modulating factors. Strong interaction between PIP2 and both of agents were voltage independent. Moreover, mepyramine-
KIR2.1 prevented interference from other modulatory factors like induced KIR2.3 current inhibition was significantly reduced by a
pH, protein kinase C, and membrane receptors such as M1 single-point mutation of KIR2.3 (I213L), which enhanced its
(type-1 muscarinic) and epidermal growth factor receptors, interaction with membrane PIP2. These results suggested that
whereas the weaker interaction between PIP2 and KIR2.3 was membrane PIP2 may likely be involved in the KIR current inhibi-
interfered further by those modulatory factors such that the function by the first-generation antihistamines.
tion of KIR2.3 or KIR3.4* was reduced.46,47 Quinacrine is an antimalarial agent that has been used for a
The cationic amphiphilic drugs (CADs) represent one of the number of additional indications, such as other parasitic infec-
most abundant chemical groups used in pharmacotherapy.49 In tions, as an antifibrillatory agent and for treatment of auto
contrast to the diverse pharmacologic actions and therapeutic immune disorders. Like other CADs, quinacrine has a high
Kir2.1(WT) Kir2.1(K182Q) 1.0

0.2 0.2
Q 0.8 13
Normalized current
Normalized current
1-(IQuinacrine-IControl)
0.0 0.0
–120 –100 –80 –60 –40 –20 –120 –100 –80 –60 –40 –20
–0.2 mV –0.2 mV 0.6
–0.4 Q –0.4
0.4
–0.6 –0.6
0.2 Kir2.1(WT)
–0.8 –0.8 Kir2.1(L222I)
C C Kir2.1(K182Q)
–1.0 –1.0 –20 mV 0.0
-3
01
10
0
00
10 00
00
1
0.
10
3s
1E
0.
10
0
00
10
–80 mV
Quinacrine (µM)
A B –120 mV C
Kir2.3(WT) 0.4 Kir2.3(I213L) 1.0 Kir2.3(WT)

0.2 Kir2.3(I213L)
Q 0.2
0.8
1-(IQuinacrine-IControl)
Normalized current
Normalized current
0.0
–120 –100 –80 –60 –40 –20 0.0
–0.2 –120 –100 –80 –60 –40 –20 0.6
mV –0.2
Q mV
–0.4 –0.4 0.4
–0.6 –0.6
0.2
–0.8 –0.8
–1.0 C –1.0 C 0.0
01
10
00
1
0.
10
0.
10
D E F Quinacrine (µM)
Figure 13-4. Inhibitory effects of quinacrine on KIR2.1 (wild type) (A), KIR2.1 (K182Q) (B), KIR2.3 (wild type) (D), and KIR2.3 (I213L) (E) channels expressed in HEK 293 cells.
Each panel shows typical currents recorded under control conditions (trace C) and in the presence of 30 µM quinacrine (trace Q). Experiments were recorded using the
whole-cell configuration, elicited by the protocol shown in the inset. Current was normalized to the amplitude recorded at −120 mV under control conditions. Panels C
and F show concentration-response curves constructed from current measured at −120 mV. The IC50 of quinacrine on KIR2.1 (wild type) was 65 ± 7 µM, on KIR2.1 (K182Q)
was 0.80 ± 0.07 µM, on KIR2.1 (L222I) was 24 ± 4 µM, on KIR2.3 (wild type) was 0.74 ± 0.15 µM, and on KIR2.3 (I213L) was 6.5 ± 0.9 µM (n = 5). Data shown represent mean ±
SEM.
(Modified from López-Izquierdo A, Aréchiga-Figueroa IA, Moreno-Galindo EG, et al: Mechanisms for Kir channel inhibition by quinacrine: acute pore block of Kir2.x channels and
interference in PIP2 interaction with Kir2.x and Kir6.2 channels. Pflugers Arch 462:505–517, 2011.)
lipophilicity (log P = 5.5) and interacts directly with membrane probably inserts into the lipid membrane and might interfere
phospholipids to induce lipidosis. The side chain of the amino- with PIP2–channel interactions. Evidence from three other
acridine group of quinacrine (pKa = 10.3) is identical to the side sources—KIR2.3(1213 L), KIR6.2(C166S), and exogenous PIP2—
chain of the aminoquinoline drug, chloroquine. Quinacrine dif- further suggested the interference of tamoxifen with the KIR-PIP2
ferentially inhibited KIR channels in the following order of interaction.
potency: KIR6.2 > KIR2.3 > KIR2.1. In addition, it was found in Mefloquine is an antimalarial drug that is used mainly as a
cardiac myocytes that quinacrine inhibited IKATP > IK1 prophylactic for malaria and in cases of chloroquine-resistant
(see Figure 13-4). It was presented evidence that quinacrine dis- malaria. Mefloquine produced inhibition of IKATP channels of
played a double action toward KIR2.1 and 2.3 channels, that is, pancreatic β cells. The effect is mediated by interaction of the
interfering with PIP2 interaction and direct pore block (see drug with the KIR6.2 subunit, rather than the regulatory SUR
Figure 13-4). On KIR6.2/SUR2a channels, the inhibitory effects of subunit. Mefloquine inhibited three different types of cardiac KIR
quinacrine seemed to be only by interference in PIP2-channel channels in the following order of potency, KIR6.2/SUR2A >
interaction.55 KIR2.3 > KIR2.1, as demonstrated by determination of the IC50
Tamoxifen is a drug used in the treatment of hormone- for mefloquine inhibition of the respective KIR channels.53 Meflo-
responsive breast cancer and prevention of breast cancer in quine also inhibited IKATP and IK1 in cardiac myocytes. Interest-
women at high risk. At clinically relevant concentrations, tamoxi- ingly, the characteristics of the mefloquine inhibition of the KIR
fen inhibited the inward rectifier potassium currents IK1, IKATP and channels were similar to those found with tamoxifen.51,52 The
IKACh in feline atrial and ventricular myocytes in the following effect on KIR channels was concentration dependent but voltage
order of sensitivity: IKACh > IKATP > atrial IK1 > ventricular IK1. independent and showed a slow time course inhibition. In addi-
Tamoxifen also inhibited in that same order KIR6.2/SUR2A and tion, the effects were similar whether mefloquine was applied
KIR3.1/3.4 > KIR2.3 > KIR2.1 = KIR2.2 in transfected HEK-293 externally or internally, suggesting that the inhibitory effect was
cells.51,52 The order of sensitivity of the different KIR channels, membrane delimited. In accordance, the KIR2.3 (I213L) mutant
the slow inhibition time course, the independence of internal or was significantly less sensitive and in inside-out patches, continu-
external drug administration, and the voltage independence of ous application of exogenous PIP2 strikingly prevented the meflo-
the inhibition induced by tamoxifen suggested that tamoxifen quine inhibition.
Carvedilol, a β- and α-adrenoceptor blocker, is used to treat mutations in KCNJ2 associated with Andersen-Tawil syndrome.
congestive heart failure, mild to moderate hypertension, and Recently, it was found that flecainide acutely increased human
myocardial infarction. Carvedilol is a CAD with a high lipophilic- IKIR2.1, mainly by reducing the polyamine blockade of the channel.
ity (log P = 3.8) and an α-hydroxyl secondary amine functional In addition, incubation with flecainide increased functional
group (pKa 7.9). Ferrer et al. demonstrated that carvedilol inhib- KIR2.1 channel density. It was proposed that flecainide reduces
its KIR2.3-carried current with at least 100-fold higher potency the inward rectification of the current at potentials positive to the
(IC50 = 0.49 µM) than KIR2.1 (IC50 = 50 µM).54 KIR2.3 channels potassium reversal potential. Flecainide interacts with C311 at
showed similar inhibition characteristics like tamoxifen and the G loop of the CPD of the channel. It was also shown that
mefloquine, results that suggested that carvedilol, as other cat- incubation with flecainide increased expression of functional
ionic amphiphilic drugs, inhibited KIR2.3 channels by interfering KIR2.1 channels on the plasma membrane, an effect also deter-
with the PIP2–channel interaction. Again, the effect was concen- mined by C311. Interestingly, flecainide pharmacologically
tration dependent and voltage-independent, whereas the KIR2.3 rescues R67W, but not R218W, channel mutations found in
(I213L) mutation decreased the potency of block more than patients with Andersen-Tawil syndrome.58
20-fold (IC50 = 11.1 µM), and in the presence of exogenous PIP2, The neurosteroid pregnenolone sulfate has been found to
inhibition by carvedilol was strongly reduced (80 vs. 2% current preferentially activate KIR2.3 with a half-maximal concentration
inhibition).54 of 16 µM and a maximal current potentiation of approximately
The acidic compound thiopental is a short-acting barbiturate 80%. Pregnenolone appeared to be selective in that it had no
used intravenously for anesthesia induction. Recently, it has been effects on KIR1.1, KIR2.1, KIR2.2, or KIR3.1 currents. Preliminary
shown that thiopental inhibited KIR2.1, KIR2.2, KIR2.3, KIR1.1, data suggested that pregnenolone sulfate acts directly on KIR2.3
and KIR6.2/SUR2A currents with similar potency; in whole-cell at an extracellular site, although the precise binding site location
experiments, 30 µM thiopental decreased KIR2.1, KIR2.2, KIR2.3, and mechanism of action have not yet been defined.59
and KIR1.1 currents to 55% ± 6%, 39% ± 8%, 42% ± 5%, and
49% ± 5%, respectively.57 Evidence was presented showing that
the thiopental effect on KIR channels likely involves disruption of
interactions between KIR6.2/SUR2a and PIP2. Thiopental inhib- Conclusions and Outlook
ited all KIR channels in a concentration-dependent and voltage-
independent manner. In addition, the time course of thiopental Three different classes of inward rectifier currents—IK1, IKACh,
inhibition was slow (T1/2-4 minutes) and independent of external and IKATP—have a prominent role in cardiac electrophysiology.
or internal drug application, suggesting that the inhibitory effect Null mutations of KIR channels in experimental animals often
was membrane delimited. In addition, in the presence of exoge- have detrimental effects on normal cardiac electrophysiology.
nous PIP2, inhibition by thiopental on KIR channels was signifi- Loss- and gain-of-function mutations found in several human
cantly decreased, suggesting an interference of PIP2-KIR channel diseases and syndromes underscore their important roles in
interaction. It was proposed that the anionic drug thiopental may proper heart function. Many marketed drugs affect inward recti-
have a different mechanism of action on KIR channels compared fier function by directly modulating the channel pore. These
with polycations like neomycin or CADs. It was suggested that drugs can be divided into two major classes: those that enter the
thiopental interacts with the various KIR channels directly at a channel pore and directly obstruct the conduction pathway and
common site and it interfered with the KIR channel–PIP2 interac- those that interfere with PIP2-dependent channel activation.
tion. A candidate residue for thiopental binding may be R218 in Conversely, their blocking effects may cause serious side effects
KIR2.1, which is conserved among all KIR channels.57 However, and even arrhythmias, and their action potential–prolonging
that hypothesis requires validation, and more work needs to be capacity may be beneficial under other circumstances such as
done to determine whether the inhibition of thiopental depends atrial fibrillation or short QT syndrome. Continuation of the
specifically on interference with KIR channel–PIP2 interactions. already impressive studies on structure-function relationships of
these channels, their interaction with various drugs at the molec-
ular level, and functional studies of inward rectifier–modifying
Drugs That Induce IK1 Activation drugs in dedicated disease models will eventually provide oppor-
tunities to develop very specific and effective new drugs that
The multichannel-blocker antiarrhythmic drug flecainide exhib- target this channel class for treating a number of human cardiac
its effective control of ventricular arrhythmias associated with diseases.
is a heteromultimer of two inwardly rectifying mutation in the KCNJ2 gene. Circ Res 96:800–
References K+-channel proteins. Nature 374:135–141, 807, 2005.
1995. 12. Vega AL, Tester DJ, Ackerman MJ, et al:
1. Katz B: Les constants electriques de la membrane 7. Inagaki N, Gonoi T, Clement JP 4th, et al: Recon- Protein kinase A-dependent biophysical pheno-
du muscle. Arch Sci Physiol 3:285–299, 1949. stitution of IKATP: An inward rectifier subunit pulus type for V227F-KCNJ2 mutation in catecholamin-
2. Hibino H, Inanobe A, Furutani K, et al: Inwardly the sulfonylurea receptor. Science 270:1166–1170, ergic polymorphic ventricular tachycardia. Circ
rectifying potassium channels: Their structure, 1995. Arrhythm Electrophysiol 2:540–547, 2009.
function, and physiological roles. Physiol Rev 8. Dean M, Annilo T: Evolution of the ATP-Binding 13. Hattori T, Makiyama T, Akao M, et al: A novel
90:291–366, 2010. Cassette (ABC) transporter superfamily in verte- gain-of-function KCNJ2 mutation associated with
3. Anumonwo JM, Lopatin AN: Cardiac strong brates. Annu Rev Genom Hum Genet 6:123–142, short-QT syndrome impairs inward rectification
inward rectifier potassium channels. J Mol Cell 2005. of KIR2.1 currents. Cardiovasc Res 93:666–673,
Cardiol 48:45–54, 2010. 9. Zaritsky JJ, Redell JB, Tempel BL, et al: The con- 2012.
4. Flagg TP, Enkvetchakul D, Koster JC, et al: sequences of disrupting cardiac inwardly rectifying 14. Bettahi I, Marker CL, Roman MI, et al: Contribu-
Muscle KATP channels: recent insights to energy K(+) current (IK1) as revealed by the targeted deletion of the KIR3.1 subunit to the muscarinic-gated
sensing and myoprotection. Physiol Rev 90:799– tion of the murine KIR2.1 and KIR2.2 genes. atrial potassium channel IKACh. J Biol Chem
829, 2010. J Physiol 533:697–710, 2001. 277:48282–48288, 2002.
5. De Boer TP, Houtman MJC, Compier M, et al: 10. Plaster NM, Tawil R, Tristani-Firouzi M, et al: 15. Wickman K, Nemec J, Gendler SJ, et al: Abnormal
The mammalian KIR2.x inward rectifier ion Mutations in KIR2.1 cause the developmental and heart rate regulation in GIRK4 knockout mice.
channel family: expression patterns and pathophys- episodic electrical phenotypes of Andersen’s syn- Neuron 20:103–114, 1998.
iology. Acta Physiol 199:243–255, 2010. drome. Cell 105:511–519, 2001. 16. Yang Y, Yang Y, Liang B, et al: Identification of a
6. Krapivinsky G, Gordon EA, Wickman K, et al: 11. Priori SG, Pandit SV, Rivolta I, et al: A novel form KIR3.4 mutation in congenital long QT syndrome.
The G-protein-gated atrial K+ channel IKAch of short QT syndrome (SQT3) is caused by a Am J Hum Genet 86:872–880, 2010.
17. Miki T, Suzuki M, Shibasaki T, et al: Mouse model 31. Noujaim SF, Stuckey JA, Ponce-Balbuena D, et al: ion channels and transporters. Sci STKE RE19,
of Prinzmetal angina by disruption of the inward Specific residues of the cytoplasmic domains of 2001.
rectifier KIR6.1. Nat Med 8:466–472, 2002.
18. Saito T, Sato T, Miki T, et al: Role of ATP-sensitive
K+ channels in electrophysiological alterations
cardiac inward rectifier potassium channels are
effective antifibrillatory targets. FASEB J 4:4302–
4312, 2012.
46. Du X, Zhang H, Lopes C, et al: Characteristic
interactions with phosphatidylinositol 4,5–
bisphosphate determine regulation of KIR channels
13
during myocardial ischemia: A study using KIR6.2- 32. Tamargo J, Caballero R, Gómez R, et al: Pharma- by diverse modulators. J Biol Chem 279:37271–
null mice. Am J Physiol Heart Circ Physiol cology of cardiac potassium channels. Cardiovasc 37281, 2004.
288:352–357, 2005. Res 62:9–33, 2004. 47. Logothetis DE, Jin T, Lupyan D, et al:
19. Liu XK, Yamada S, Kane GC, et al: Genetic 33. Zaks-Makhina E, Li H, Grishin A, et al: Specific Phosphoinositide-mediated gating of inwardly
disruption of KIR6.2, the pore-forming subunit and slow inhibition of the KIR2.1 K+ channel by rectifying K+ channels. Pflugers Arch 455:83–95,
of ATP-sensitive K+ channel, predisposes to gambogic acid. J Biol Chem 284:15432–15438, 2007.
catecholamine-induced ventricular dysrhythmia. 2009. 48. Hansen SB, Tao X, MacKinnon R: Structural basis
Diabetes 53:S165–S168, 2004. 34. Brown JH, Laiken N: Muscarinic receptors ago- of PIP2 activation of the classical inward rectifier
20. Flagg TP, Kurata HT, Masia R, et al: Differential nists and antagonists. In Brunton L, Chabner B, K+ channel KIR2.2. Nature 477:495–498, 2011.
structure of atrial and ventricular KATP: Atrial KATP Knollman B, editors: Goodman and Gilman’s The 49. Lundbæk JA: Lipid bilayer–mediated regulation of
channels require SUR1. Circ Res 103:1458–1465, Pharmacological Basis of Therapeutics, ed 12, ion channel function by amphiphilic drugs. J Gen
2008. New York, 2011, McGraw-Hill. Physiol 131:421–429, 2008.
21. Elrod JW, Harrell M, Flagg TP, et al: Role of 35. Dobrev D, Carlsson L, Nattel S: Novel molecular 50. Zhou W, Arrabit C, Choe S, et al: Mechanism
sulfonylurea receptor type 1 subunits of ATP- targets for atrial fibrillation therapy Nat Rev Drug underlying bupivacaine inhibition of G protein-
sensitive potassium channels in myocardial Discov 11:275–291, 2012. gated inwardly rectifying K+ channels. Proc Natl
ischemia/reperfusion injury. Circulation 117:1405– 36. McBride BF: The emerging role of antiarrhythmic Acad Sci U S A 98:6482–6487, 2001.
1413, 2008. compounds with atrial selectivity in the manage- 51. Ponce-Balbuena D, López-Izquierdo A, Ferrer T,
22. Stoller D, Kakkar R, Smelley M, et al: Mice lacking ment of atrial fibrillation. J Clin Pharmacol et al: Tamoxifen inhibits KIR2.x family of inward
sulfonylurea receptor 2 (SUR2) ATP-sensitive 49:258–267, 2009. rectifier channels by interfering with PIP2-channel
potassium channels are resistant to acute cardiovas- 37. van der Heyden MAF, Sánchez-Chapula JA: intaractions. J Pharm Exp Ther 331:563–573,
cular stress. J Mol Cell Cardiol 43:445–454, 2007. Towards specific cardiac IK1 modulators for in vivo 2009.
23. Haïssaguerre M, Chatel S, Sacher F, et al: Ven- application; old drugs point the way. Hearth 52. Ponce-Balbuena D, Moreno-Galindo EG, López-
tricular fibrillation with prominent early repolar- Rhythm 8:1076–1080, 2011. Izquierdo A, et al: Tamoxifen inhibits cardiac KATP
ization associated with a rare variant of KCNJ8/ 38. Sanchez-Chapula JA, Salinas-Stefanon E, Torres- and KACh currents in part by interfering with PIP2–
KATP channel. J Cardiovasc Electrophysiol 20:93– Jacome J, et al: Blockade of currents by the anti- channel interaction. J Pharm Sci 113:66–75, 2010.
98, 2009. malarial drug chloroquine in feline ventricular 53. López–Izquierdo A, Ponce-Balbuena D. Moreno-
24. Medeiros-Domingo A, Tan BH, Crotti L, et al: myocytes. J Pharm Exp Ther 297:437–445, 2001. Galindo EG, et al: The antimalarial drug meflo-
Gain-of-function mutation S422L in the KCNJ8- 39. Rodríguez-Menchaca AA, Navarro-Polanco RA, quine inhibits cardiac inward rectifier K+ channels:
encoded cardiac KATP channel KIR6.1 as a patho- Ferrer-Villada T, et al: The structural molecular Evidence for inference in PIP2 –channel interac-
genic substrate for J-wave syndromes. Heart basis of chloroquine block of the inward rectifier tion. J Cardiovasc Pharmacol 57:407–415, 2011.
Rhythm 7:1466–1471, 2010. KIR2.1 channel. Proc Natl Acad Sci U S A 54. Ferrer T, Ponce-Balbuena D, López-Izquierdo A,
25. Barajas-Martínez H, Hu D, Ferrer T, et al: Molec- 105:1364–1368, 2008. et al: Carvedilol inhibits KIR2.3 channels by inter-
ular genetic and functional association of Brugada 40. de Boer TP, Nalos L, Stary A et al: The anti- ference with PIP2-channel interaction. Eur J Phar-
and early repolarization syndromes with S422L protozoal drug pentamidine blocks KIR2.x- macol 668:72–77, 2011.
missense mutation in KCNJ8. Heart Rhythm mediated inward rectifier current by entering the 55. Lopez-Izquierdo A, Aréchiga-Figueroa IA,
9:548–555, 2012. cytoplasmic pore region of the channel. Br J Phar- Moreno-Galindo EG, et al: Mechanisms for KIR
26. Delaney JT, Muhammad R, Blair MA, et al: macol 159:1532–1541, 2010. channel inhibition by quinacrine: Acute pore block
A KCNJ8 mutation associated with early repolar- 41. Noujaim SF, Stuckey JA, Ponce-Balbuena D et al: of KIR2.x channels and interference in PIP2 interac-
ization and atrial fibrillation. Europace 2012 in Structural bases for the different antifibrillatory tion with KIR2.x and KIR6.2 channels. Pflügers Arch
press. effects of chloroquine and quinidine. Cardiovasc 462:505–517, 2011.
27. Tester DJ, Tan BH, Medeiros-Domingo A, et al: Res 89:1–8, 2011. 56. Liu B, Jia Z, Geng X, et al: Selective inhibition of
Loss-of-function mutations in the KCNJ8-encoded 42. El Harchi A, McPate MJ, Zhang YH, et al: Action KIR currents by antihistamines. Eur J Pharmacol
KIR6.1 KATP channel and sudden infant death potential clamp and chloroquine sensitivity of 558:21–26, 2007.
syndrome. Circ Cardiovasc Genet 4:510–515, mutant KIR2.1 channels responsible for variant 3 57. López-Izquierdo A, Ponce-Balbuena D, Ferrer T,
2011. short qt syndrome. J Mol Cel Cardiol 47:743–747, et al: Thiopental inhibits function of different
28. Bienengraeber M, Olson TM, Selivanov VA, et al: 2009. inward rectifying potassium channel isoforms by a
ABCC9 mutations identified in human dilated car- 43. Lopez-Izquierdo A, Ponce-Balbuena D, Ferrer T, similar mechanism. Eur J Pharmacol 638:33–41,
diomyopathy disrupt catalytic KATP channel gating. et al: Chloroquine blocks a mutant KIR2.1 channel 2010.
Nat Genet 36:382–387, 2004. responsible for short QT syndrome and normalizes 58. Caballero R, Dolz-Gaitón P, Gómez R, et al: Fle-
29. Olson TM, Alekseev AE, Moreau C, et al: KATP repolarization properties in silico. Cell Physiol cainide Increases KIR2.1 currents by interacting
channel mutation confers risk for vein of Marshall Biochem 24:153–160, 2009. with Cysteine 311 decreasing the polyamine-
adrenergic atrial fibrillation. Nat Clin Pract Car- 44. Fan Z, Makielski JC: Anionic phospholipids induced rectification. Proc Nat Acad Sci U S A
diovasc Med 4:110–116, 2007. activate ATP-sensitive potassium channels. J Biol 107:15631–15636, 2010.
30. Harakalova M, van Harssel JJ, Terhal PA, et al: Chem 272:5388–5395, 1997. 59. Kobayashi T, Washiyama K, Ikeda K: Pregneno-
Dominant missense mutations in ABCC9 cause 45. Hilgemann DW, Feng S, Nasuhoglu C: lone sulfate potentiates the inwardly rectifying K
Cantú syndrome. Nat Genet 44:793–796, 2012. The complex and intriguing lives of PIP2 with channel KIR2.3. PLoS One 4:e6311, 2009.
Cardiac Stretch–Activated Channels
and Mechano-Electric Coupling 14
Peter Kohl
insight into mechanisms underlying MEC, but also for the

CHAPTER OUTLINE
advancement of the topic beyond the perception of a scientifically
Functional Relevance of Cardiac Mechano-Electric ill-founded clinical curiosity. This chapter will discuss acute elec-
Coupling 139 trophysiologic responses of the heart to mechanical stimulation
and the involvement of SAC.
Cardiac Stretch Activated–Ion Channels 143
Pharmacologic Probes 145
Manifestations of Cardiac Stretch–Activated Ion Functional Relevance of Cardiac
Channel Activation 145 Mechano-Electric Coupling
Summary and Outlook 147
Effects of cardiac mechanical stimulation on heart rate and
rhythm have been reported in the medical literature for more
than a century. To name a few key contributions: pioneering work
It is remarkable that basic scientists and clinical practitioners by Felice Meola2a and Ferdinand Riedinger2b in the late nine-
often are inclined to reduce the heart, and what may be wrong teenth century identified Commotio cordis (or Commotio thoracica)
with it, to its electrical function. Of course, it is not a lack of as an independent pathologic entity where cardiac rhythm dis-
action potential (AP) generation (e.g., asystole) or a surplus of turbances of varying severity are initiated by nonpenetrating
uncoordinated AP-cycles (e.g., ventricular fibrillation [VF]) that mechanical stimulation of the precordium in the absence of
is lethal, but the cessation of competent cardiac pump function. visible structural damage to the heart. In the early twentieth
A case in point is pulseless electrical activity, a cause of cardiac century, Eduard Schott29 reported that precordial fist thumps can
arrest whose prevalence has been rising in recent decades. In be used to pace otherwise asystolic ventricles, such as in Adams-
pulseless electrical activity, the electrocardiogram (ECG) can Stokes syndrome. At the same time, Francis Bainbridge2c famously
show near-normal cycles of electrical excitation and repolariza- identified the positive chronotropic response of the heart to
tion, yet these occur in the absence of contractile activity, pres- increased venous return.
sure generation, and, hence, cardiac output. Thus, since the beginning of published reports in modern
The processes that underlie excitation-contraction coupling medical literature, mechanical stimulation of the heart has been
(ECC; see chapter 16), as well as contractile activity of cardiac found to have the potential of inducing and terminating heart
muscle, form major foci of basic and applied research. In contrast, rhythm disturbances, as well as to modulate cardiac pacemaker
the relevance of the mechanical environment for cardiac electri- rate.
cal activity—mechano-electric coupling (MEC)—is often over-
looked or ignored. Thus, mechano-electric dissociation is often
introduced in experimental research on purpose, by applying Mechanical Induction of Nonphysiologic Rhythms
pharmacologic uncouplers, to reduce or abolish motion artifacts
that interfere with the fidelity of electrical signals, even though The fact that mechanical stimuli of sufficient amplitude can be
this uncoupling alters observed electrical behavior.1 In conceptual used to pace otherwise quiescent hearts has been illustrated by
terms, cardiac MEC complements ECC, integrating cardiac elec- Franz et al., using Langendorff-perfused rabbit heart prepara-
trical and mechanical activity into an intracardiac regulatory loop tions where the ventricles (rendered asystolic by ablation of the
(Figure 14-1). atrioventricular node) were stimulated by periodic inflation of an
Myocardium is an exquisite mechano-electric transducer. A intraventricular balloon (Figure 14-2, A).
tangible example is the classic coronary-perfused heart prepara- Similar behavior is believed to underlie “fist pacing” in asys-
tion, established by Oskar Langendorff in the nineteenth century, tolic patients. Energy levels required for mechanical PVB induc-
which can be stopped or restarted at the flick of a finger. This tion by precordial impact have been established by defibrillation
phenomenon is not restricted to the isolated heart. Catheter pioneer Zoll et al.3 in human volunteers as 0.04 to 1.5 J. For
approach to the ventricles can be judged by the appearance of comparison, the lower end of this energy range is equivalent to
premature ventricular beats (PVB). Mechanically induced ectopic dropping a golf ball (46 g) from a height of 9 cm (3.5 in).
excitation by finger-tapping of the exposed ventricular wall is The study by Zoll et al.3 found that, in anaesthetized dogs,
used for reinstatement of sinoatrial node (SAN) rhythm in impacts with energies tenfold greater than the PVB-induction
patients that are being weaned from cardiac bypass, in particular threshold do not induce repetitive responses, ventricular tachy-
if prior electrical defibrillation has caused asystole. cardia (VT), or VF, even if applied in the relative refractory
In fact, manifestations of cardiac MEC are present at all levels period.3 Overall, this is in keeping with the notion that arrhyth-
of structural and functional integration, from in situ and ex vivo mogenesis requires the combination of trigger and sustaining
whole heart, over in vitro tissue and cells, to subcellular domains mechanisms, so that consequences of isolated ectopic beats,
such as membrane patches. Indeed, patch-clamp identification of whether mechanically induced or not, are usually benign.
cardiac stretch-activated ion channels (SACs) by Guharay and Although these examples illustrate the effects of the external
Sachs2 was pivotal not only for the development of quantitative mechanical environment on cardiac MEC (see Figure 14-1,
139
bottom box), mechanical PVB induction can also arise because of acutely by pharmacologic block of the slowly activating delayed
the heart’s own contractile activity (see Figure 14-1, middle box). rectifier potassium current. On this background, additional
This is illustrated in Figure 14-2, B, which shows MAP record- β-adrenergic stimulation by bolus injection of isoproterenol gives
ings from a patient undergoing pulmonary balloon valvuloplasty. rise to ventricular after contractions of increasing amplitude
In this procedure, the stenosed RV outflow valve is widened by (up to 25 mm Hg). Originating from near-endocardial locations,
insertion and inflation of a balloon. During balloon inflation, their onset precedes (by tens of milliseconds) EAD-like epicardial
RV contractions are isovolumic (no ejection) and give rise to potentials and alterations in descending T-wave morphology.
significantly increased RV peak-pressures. This increase is With increasing amplitude, these apparently contraction-
generally associated with AP-shortening, diastolic depolarization, induced afterdepolarizations eventually reach threshold for PVB-
and occurrence of early afterdepolarization (EAD)-like events. If induction, followed by torsades de pointes arrhythmic activity.5
supra-threshold, these depolarization can give trigger PVB.4 As Under certain conditions, acute mechanical stimulation can
before, against an otherwise inconspicuous myocardial back- be sufficient to give rise to both trigger and sustaining mecha-
ground, self-sustained repetitive activity is not normally induced. nisms for maintained arrhythmias, even in otherwise healthy
In contrast, in the presence of preexisting pathologies, PVB myocardium. A prominent example of this is Commotio cordis. A
can give rise to sustained arrhythmias, as shown in whole-animal number of risk factors for the mechanical induction of such
studies of pathologically prolonged QT intervals, conducted by rhythm disturbances have been identified, based on experimental
Volders et al. In their model, QT-prolongation was induced observations from the pioneering work of Schlomka6 to modern
studies by Link.7 Key risk factors include (1) type of impact
(impulse-like stimulation, whose arrhythmogenic risk is inversely
related to projectile compliance and contact area); (2) impact
Extracardiac Control Mechanisms energy (large subcontusional forces, reaching more than 100 J in
(e.g., autonomic nervous system, hormones, drugs) competitive sports); (3) impact location (precordial, or in human
also spinal, areas that offer efficient energy transmission from
body surface to myocardium); and (4) impact timing. Factors 1
Cardiac Mechano-Electric Coupling (MEC) to 3 can be regarded as permissive: only if they are all present
does timing become decisive.8 This presumably explains why the
Cardiac electrophysiology Cardiac mechanics vast majority of precordial (or spinal) impacts result in relatively
ECC benign heart rhythm changes, if any.
Ion currents Vm/AP Calcium Stress/strain
The critical time-window for the worst possible outcome of
Commotio cordis, induction of VF, overlaps with the T wave. The
Biophysical Environment T wave, during which myocardial electrophysiologic heterogene-
(e.g., ion concentrations, electrical fields, stress/strain, temperature) ity is maximal, has long been associated with a period of increased
susceptibility to arrhythmogenesis by electrical stimulation, the
Figure 14-1. Schematic view of cardiac electromechanic integration. Within the so-called vulnerable window. Compared with electrical stimula-
heart (center box), electrical behavior steers mechanical activity via excitation- tion, the vulnerable window for mechanical VF-induction is sur-
contraction coupling (ECC) but, in turn, it is affected by mechano-electric coupling
prisingly narrow (∼15 ms, just before the peak of the T wave in
(MEC). This involves a multitude of interdependent feed-forward and feed-back
pathways that alter transsarcolemmal and intracellular ion transport, membrane
anaesthetized pig).9
and action potential configuration (ΔVm/AP), calcium handling, and stress–strain Quantitative computational modeling suggests that mechani-
dynamics. Intracardiac electromechanic interactions are modulated by the bio- cally induced VF is favored during a short period of time only,
physical environment of the heart (bottom box), and form a target of extracardiac when the mechanically stimulated tissue volume overlaps with
control (top box). the trailing wave of excitation.10,11 In this setting, sustained
MAP 5 mV
LVVol 500 µL
A 100 ms
0
10 mV
B 1s
Figure 14-2. Mechanically induced premature ventricular beats (PVB). A, Mechanical pacing of the Langendorff-perfused asystolic rabbit heart. Pulsatile left-ventricular
(LV) distention (bottom trace, showing increasing LV balloon inflation) causes diastolic membrane depolarization (top trace, blue highlights). The amplitude of depolarization
increases with the mechanical stimulus and, upon reaching threshold levels, gives rise to the generation of action potentials (APs; top trace, red) with each volume pulse.
Note that the first two APs are spontaneous escape beats, unrelated to the mechanical stimuli. B, Cardiac contraction–induced tissue depolarization (blue) and PVB induc-
tion (red) in patients. MAP recording before (left) and during (middle and right) inflation of a balloon in the right-ventricular outflow tract, which gives rise to early
afterdepolarization-like events (blue) and PVB-induction (red). LVVol, volume pulses; MAP, monophasic AP.
(A, From Franz MR, Cima R, Wang D, et al: Electrophysiological effects of myocardial stretch and mechanical determinants of stretch-activated arrhythmias. Circulation 86:968–978,
1992. B, From Levine JH, Guarnieri T, Kadish AH, et al: Changes in myocardial repolarization in patients undergoing balloon valvuloplasty for congenital pulmonary stenosis:
evidence for contraction-excitation feedback in humans. Circulation 77:70–77, 1988.)
Cardiac Stretch–Activated Channels and Mechano-Electric Coupling 141
arrhythmogenesis occurs because, in addition to mechanical PVB In contrast to acute effects, MEC-contributions to arrhyth-
14
induction in tissue that has regained excitability (trigger), the mogenesis in chronic cardiac overload are more difficult to
intersection of mechanically affected myocardium and the trail- uncover. Usually, pathologies that involve cardiac pressure or
ing repolarization wave gives rise to a functional block zone volume overload develop slowly, and they are associated with
around which reentry develops (sustaining mechanism; Figure pronounced structural and functional remodeling. In addition,
14-3). In three-dimensional tissue, the ectopic activation front the causes of overload and tissue remodeling could be proar-
forms around a transmurally directed conelike volume of tissue rhythmogenic in their own right. Nonetheless, mechanical
in which mechanical activation reaches the AP threshold.12 The factors have been implicated in the domestication of atrial fibril-
more forceful the impact, the closer to fully transmural is the lation (AF),13 whose dominant frequency in patients correlates
affected tissue cone. In addition, the more focal the stimulus, with atrial pressure.14 Similarly, ventricular arrhythmogenesis has
the closer to perpendicular is this ectopic wavefront, relative to been linked to pressure and volume overload.15,16
the cardiac surface and, by implication, to the trailing wave-end. A conceptually interesting approach to probing the relevance
This relationship enhances the arrhythmogenic potential, as seen of mechanical factors for arrhythmogenesis in the chronically
in classic S1/S2 cross-stimulus experiments, and could hold a key overloaded heart is the temporary removal of tissue strain. This
to understanding risk factors 1 and 2. can be achieved with the Valsalva maneuver, an attempt to force-
Risk factors 3 and 4 are related to the understanding that the fully exhale against the closed glottis. Intrathoracic pressure
critical window for mechanical VF-induction will differ, depend- increases during the strain phase of the maneuver, reducing
ing on where on the surface of the heart a stimulus is applied; venous return and favoring arterial drainage from the chest. This
therefore, this critical window exists in space and time. However, condition leads to measurable reductions in cardiac dimensions.
only a limited part of myocardial tissue is in close proximity to On this background, VT has been found to convert to sinus
the precordium and, hence, accessible to extracorporeally applied rhythm.17 In this study, 7 of 15 patients showed sustained and 2
local mechanical stimulation. Therefore, only a subset of the showed transient cardioversion. Relief of ventricular wall stress,
(location-specific) critical windows present throughout the heart rather than autonomic nervous system–mediated responses,
form a mechanically accessible target in vivo. This is qualitatively appears to be a causal contributor to this antiarrhythmic effect,
different from electrical stimulation, which is less focused and because successful cardioversion can also be observed in the pres-
less dependent on proximity, and therefore potentially arrhyth- ence of pharmacologic17 or surgical denervation of the heart, such
mogenic over a longer part of the T-wave duration. as in transplant recipients.18
Thus, acute stretch can cause diastolic depolarization, which
can trigger ectopic excitation. Systolic or sustained stretch can
0 ms 3 ms contribute to arrhythmia sustenance by enhancing heterogene-
ities in excitability, refractoriness, and electrical load. These
MEC effects have implications for preventive measures (e.g.,
chest protector design for sports involving fast-moving projec-
60 mV tiles) and for interventions such as hemodynamic unloading,
active and passive cardiac assist, or biventricular pacing. In addi-
tion, defibrillation threshold increases with ventricular preload,19
apparently not only because of geometric factors,20 but also
because of regionally differing MEC-mediated strain effects that
raise background electrophysiologic heterogeneity.21 Cardiac
mechanosensitivity therefore adds an interesting dimension to
5 ms 13 ms the present discussion about whether a period of chest compres-
sions should precede defibrillation attempts in cardiac arrest
victims with prolonged delays to intervention.
–100 mV Mechanical Termination of

Nonphysiologic Rhythms
Region of Acute mechanical stimulation, usually by precordial thump (PT),
functional block can be used as a means of cardiopulmonary resuscitation. PT has
been reported to terminate arrhythmias, including asystole, VT,
Figure 14-3. Effect of local mechanical stimulation during the critical window of and rarely VF (Figure 14-4).
the T wave in a two-dimensional model of ventricular free wall consisting of 256 × In the emergency resuscitation setting, PT is often the fastest
256 electrotonically coupled guinea pig ventricular cell models. A local impact to
resuscitative procedure available. PT is delivered with the ulnar
the epicardium (right-hand side of the grid), timed to occur at 40% tissue repolar-
ization (see trailing wave-end of the previous activation, which travels from right
side of the clinched fist, from a height of approximately 20 cm (8
to left; for color-coding, see scale bar), stimulates cation nonselective stretch- in), followed by active retraction to emphasize the impulse like
activated channels in the mechanically affected tissue (white outline, top left panel). nature of the stimulus. The energy levels required for PT-based
This gives rise to depolarization in tissue that has regained excitability and early termination of VT and VF are one to two orders of magnitude
repolarization of tissue with positive membrane potentials (top right panel). If supra- higher than those involved in mechanical pacing of the acutely
threshold, stretch-induced depolarization causes ectopic excitation (bottom left asystolic ventricle (4 to 10 J vs. 0.04 to 1.5 J in adults).22 The
panel). The intersection of trailing wave and mechanically induced new excitation more powerful thumps are applied preferentially to the lower
forms a functional block zone. This causes the formation of two wave fronts that sternum, rather than the left sternal edge, which is targeted for
progress across the tissue near perpendicularly to the normal direction of electrical
fist-pacing or precordial percussion.23
wave propagation (bottom-right, see white arrows) and give rise to the formation of
two stable reentrant rotors of excitation.
Efficacy of PT in real-life settings is difficult to calibrate and
compare, because patient backgrounds vary, mechanical impact
(Data from Garny A, Kohl P: Mechanical induction of arrhythmias during ventricular properties are not normally monitored, controlled study designs
repolarisation: modelling cellular mechanisms and their interaction in 2D. Ann N Y are not usually possible, and single case reports in particular
Acad Sci 1015:133–143, 2004.) suffer from positive data publication bias. The available
AS NSR
VT NSR
NSR VF NSR
C
Figure 14-4. Electrocardiograph recordings of precordial thump (PT)-induced cardioversion in patients. A, Conversion of ventricular asystole (AS) to normal sinus rhythm
(NSR) by PT. B, PT-conversion of ventricular tachycardia (VT). C, PT-induced termination of early ventricular fibrillation (VF). Arrows indicate PT application.
(A, From Pellis T, Kette F, Lovisa D, et al: Utility of precordial thump for treatment of out of hospital cardiac arrest: a prospective study. Resuscitation 80:17–23, 2009. B, From
Pennington JE, Taylor J, Lown B: Chest thump for reverting ventricular tachycardia. N Engl J Med 283:1192–1195, 1970. C, From Barrett JS: Chest thumps and the heartbeat.
N Engl J Med 284:393, 1971.)
information from case series involving tens24 to 100 patients25 levels of potassium-selective SAC (SACK) such as TREK-134 and
suggest that fist-pacing of the bradycardic heart has higher cation nonselective SAC (SACNS) such as TRPC1.35
success rates (up to 90%)25 than PT-version of tachyarrhythmias Overall, the range of cardiac electrophysiologic responses to
(1% to 60%).24,26 Only three small prospective studies have inves- mechanical stimulation is not principally different from those
tigated the clinical utility of PT.26-28 They showed that PT is caused by electrical energy delivery. Both can be used to pace,
ineffective in most tachyarrhythmias, that acutely-asystolic arrest cardiovert, or arrest the heart. Concepts such as the vulnerable
(for which PT was initially described in 1920)29 may be the most window for VF-induction apply to both electrical and mechanical
amenable target for mechanical cardioversion, and that negative stimuli. Even the effective energy ranges required for pacing,
side effects are rare (none in >700 cases).26-28 cardioversion and defibrillation are not entirely dissimilar. For
In cases where PT causes successful termination of tachyar- example, external electrical defibrillation is usually achieved by
rhythmias, it is believed to act primarily via depolarizing excitable extracorporeal energy application of 150 to 250 J; of this, only 4%
tissue in the excitable gaps.30 However, as these gaps may be in traverses the heart,36 yielding a cardiac energy delivery of 6 to 10 J,
myocardium at sites distant from the tissue underneath the pre- similar to that for the PT-version of tachyarrhythmia. These simi-
cordium, high PT energy levels are necessary and, if there is a larities should not be surprising, because mechanical energy can
multitude of them (e.g., in VF), positive outcomes are rare. be converted into a transmembrane current by ion fluxes through
In the asystolic heart, fist-pacing can trigger cardiac excitation SAC at the intervention’s target site: the cardiac cell.
and active contraction. The hemodynamic efficacy of such
mechanically induced cardiac contractions is not different from
electrically stimulated beats,31 and both are about twice as pro- Mechanical Modulation of Pacemaking
ductive (in terms of volume output) as chest compressions, even
if performed optimally.32 This efficacy confers resuscitatory value Diastolic stretch of cardiac tissue causes membrane depolariza-
to fist-pacing even when normal sinus rhythm is not immediately tion, if strong enough to induce any change. In working myocar-
reinstated; of course, PT should not delay implementation of dium, such depolarization may trigger PVB (see Figure 14-2),
other established resuscitation measures. while in conduction33 and pacemaker tissue37 of many species an
The most common manifestation of PT effects in ECG increase in beating rate (BR) is observed (Figure 14-5). This
recordings, in particular in the asystolic heart, is an impact- positive chronotropic response to stretch is seen predominantly
induced electrical artifact with an electrical axis that tends to in mammals with low resting heart rates, such as guinea pigs,
resemble the direction of normal QRS-complexes. This manifes- rabbits, cats, dogs, and humans. In fast-beating murine hearts, in
tation suggests that mechanically induced excitation in the qui- contrast, sustained stretch reduces BR, although apparently via
escent heart can proceed along a pathway that has an overall the same underlying mechanism: SACNS activation.38
trajectory similar to that of normal activation. It is possible, From a systematic point of view, early induction of the next
therefore, that earliest excitation is triggered preferentially either heartbeat in response to increased venous return is advantageous,
in cells of the secondary/tertiary pacemaker/conduction tissue of unless volume throughput is limited by inflow, such as in species
the heart (e.g., Purkinje fibers that have long been known to be with already very high heart rates. In species with an upright body
mechanosensitive),33 or in subendocardial locations, which appear posture, however, a chief and evidently overriding requirement
to be more mechanosensitive than subepicardial tissue.5 This for survival is the control of cardiac output pressure to ensure
possibility is supported by transmural differences in expression brain perfusion. Therefore, heart rate responses to hemodynamic
14
40
Membrane potential (mV)

20
–20
–40
–60
0 250 500 750 1000
Time (ms)
Figure 14-5. Positive chronotropic response to stretch in whole animal, isolated tissue, and single pacemaker cell. Left, Classic observation by Francis Bainbridge, showing
that intravenous injection of saline (bottom) raises venous pressure (top) and pulse rate (second from top) without coinciding change in arterial blood pressure (BP) in the
anesthetized dog. Top right, Sharp electrode recording of sinoatrial node (SAN) pacemaker cell potential (top) and tissue tension (bottom, contraction pointing upward)
during application of a 30% area increase to isolated right-atrial tissue containing the SAN (stretch indicated by the increase in resting tension). Telltale signs of MEC effects
are reduction of maximal diastolic and maximal systolic potentials, and increased beating rate. Bottom right, Axial stretch by 7% (black curve) of rabbit single SAN cell gives
rise to electrophysiologic changes that match those seen at tissue level.
(A, From Bainbridge FA: The influence of venous filling upon the rate of the heart. J Physiol 50:65–84, 1915. B, From Deck KA: Dehnungseffekte am spontanschlagenden, isolierten
Sinusknoten. Pflug Arch 280:120–130,1964. C, From Cooper PJ, Lei M, Cheng LX, Kohl P. Axial stretch increases spontaneous pacemaker activity in rabbit isolated sinoatrial node
cells. J Appl Physiol 89:2099–2104, 2000.)
stimuli that affect both venous return and arterial pressure, such selection bias. This question is somewhat misleading, of course,
as most changes in body posture (including standard tilt-table because we do not know the evolutionary cost of removing any
experiments), will be determined primarily by arterial pressure trait (in particular if that trait per se does not conflict with repro-
control patterns. This obscured the identification of the positive ductive probability). It is possible, however, that the deleterious
chronotropic response to stretch in human until Donald and electrophysiologic manifestations of cardiac MEC are side effects
Shepherd39 dissociated the increase in venous return from arterial rather than main targets of underlying mechanisms, akin to the
pressure changes, by passively elevating the legs of healthy vol- arrhythmogenic potential of the sodium-calcium exchanger in
unteers in supine position, confirming the positive chronotropic calcium-overloaded cells, as a consequence of its electrogenicity,
response in humans.39 whereas its physiologic function is to maintain ion concentration
Similar mechanosensitive behavior is believed to underlie the gradients. Therefore, it is conceivable that MEC is a side effect
nonnervous component of respiratory sinus arrhythmia (RSA). of mechano-mechanic coupling, perhaps required for the adjust-
Dynamic changes in thoracoabdominal pressure gradients favor ment of individual cells’ calcium concentration to external
venous blood return to the heart during inspiration, causing a mechanical demand (see Summary and Outlook).
relative increase in right atrial filling and an associated rise in
heart rate. Although this mechanical component contributes little
to RSA at rest (when modulation of vagal innervation is the
dominant driver), it is a major cause of RSA during peak exercise Cardiac Stretch–Activated Ion Channels
in healthy subjects (when vagal tone is reduced, while respiratory
effort and associated pressure gradients are enhanced),40,41 and Transsarcolemmal Channels
explains the presence of RSA in heart transplant recipients.41
If mechanical modulation of the heart rate was the only physi- Mechanosensitive ion channels can be found in the sarcolemma
ologically relevant effect, one might be tempted to wonder why of most prokaryotic and eukaryotic cell types. The open probabil-
cardiac MEC has been preserved during evolution, against its ity of these channels is primarily modulated by mechanical
pro-arrhythmic potential that could have exerted a negative stimuli, such as stretch (e.g., SAC) or cell volume changes
(volume-activated channels; VACs). VACs in particular are sarcolemma, combined with a reduction in transsarcolemmal
believed to be a phylogenetically ancient theme, required to allow dimensions.50 These configurational changes would be favored
single-cell organisms to adapt to potentially drastic changes in by increased lipid bilayer tension and by membrane thinning,
osmotic pressure in their external environment. both of which could underlie the mechanosensitive gating of the
The mammalian heart contains both SACs and VACs. SACs channel. In this context, any channel whose area projection in the
respond instantaneously to mechanical stimuli, whereas VACs plane of the cell membrane increases during opening should be
tend to show a significant lag-time (up to minutes) between the sensitive to lipid bilayer tension. The question “why do SAC
onset of cell volume changes and ion channel response. VACs, channels exist?” could be replaced, therefore, by asking “should
possibly encoded by CIC-3,42 are understood to be important not more channels be at least modulated by their mechanical
contributors to cardiac electrophysiology in chronic settings, environment?” Recent reports indicate that the range of mechan-
such as ischemic or postischemic cell swelling and hypertrophy ically modulated ion channels is indeed significantly larger than
(where VAC are constitutionally active).43 In the context of elec- generally assumed.45,51
trophysiologic responses to beat-by-beat changes in the mechani- At the same time, no sequence or structure homologues of
cal environment, however, SACs are implied as a key mechanism MscL have been identified in mammalian myocardium. However,
underlying MEC. recent evidence suggests that TRPC channels (such as TRPC148
SACs were discovered in the 1980s, initially in cultured avian or TRPC652) underlie cardiac sarcolemmal SACNS, whereas
skeletal muscle by Guhara and Sachs,2 and confirmed in mam- cardiac SACK appear to include two pore-domain channels (e.g.,
malian cardiomyocytes by Craelius et al.44 SACs show either little TREK-1),53 inwardly rectifying channels (e.g., Kir),54 and ligand-
selectivity for (predominantly monovalent) cations (SACNS, for activated channels (e.g., KATP, BKCa).46,55 Another promising can-
nonselective), or they preferentially conduct potassium ions didate is formed by piezo proteins that have been shown recently
(SACK). These selectivity profiles determine their transmem- to assemble into large (protein-tetramers containing more than
brane current reversal potentials, which are half-way between 100 transmembrane domains) mechanosensitive channels in a
AP-plateau and resting potentials for SACNS (usually between 0 range of species, from flies to mammals.56
and –25 mV), and close to the potassium equilibrium-potential Detailed functional description of SAC has been complicated
for SACK (approximately –95 mV in cardiomyocytes). by the fact that some of them (in particular SACNS) appear to be
Like other phenomenological classifications, mechanosensi- located in membrane areas that are not easily accessible to patch-
tive ion channel categories are not absolute, and there is overlap clamp investigations in adult mammalian ventricular cells, such
with other types of ion channels. Several mechanosensitive chan- as T-tubules57 and caveolae.58 Single-ion channel data have there-
nels are also voltage or ligand sensitive, and vice versa, voltage fore been obtained mainly on neonatal or atrial cells.
or ligand activated channels can be modulated by their mechani-
cal environment. Thus, the hyperpolarization-activated cyclic
nucleotide gated (HCN) channel is mechanically modulated,45 as Nontranssarcolemmal Channels
is the adenosine triphosphate (ATP)-inactivated potassium
channel (KATP), whose open probability is increased by stretch in In addition to transsarcolemmal SACs that allow ion flux between
atrial46 and ventricular myocytes.47 Mechanosensitivity could cell interior and extracellular spaces, ion channels in certain sub-
explain why these two ion channel populations appear to be less cellular compartments may be mechanosensitive. These com-
active when studied in vitro, compared with predictions based on partments include the sarcoplasmic reticulum (SR), mitochondria,
in situ observations. The contribution of the funny current, if and the nuclear envelope.
(HCN equivalent in cardiac pacemaker cells) to pacemaking, for Calcium release from and reuptake into the SR have been
example, could be underestimated under conditions of reduced shown to be modulated in cardiac cells by the mechanical envi-
external load, such as in isolated cell and tissue preparations. ronment, chiefly invoking length-dependent changes in calcium
Likewise, in vitro activation of KATP channels can occur only after buffer capacity of contractile filaments and direct or secondary
reaching abnormally low ATP levels in mechanically unloaded effects of transsarcolemmal ion fluxes (e.g., Ca2+-fluxes, or Na+-
cells. This would be in keeping with the in situ observation that fluxes that affect the Ca2+-balance via sodium-calcium exchange).
prevention of systolic stretch (or paradoxical segment lengthen- However, SR calcium release events (“sparks”) become more
ing) of ischemic myocardium, achieved using a tripodlike frequent during acute axial stretch, even in resting cardiomyo-
mechanical clamp, reduces or delays extracellular potassium cytes and in the absence of extracellular sodium and calcium
accumulation in the anaesthetized pig.16 (Figure 14-6).59 Similarly, stimulation of cultured atrial and ven-
For practicality, SACs are therefore regarded as the channels tricular cardiomyocytes by the application of small fluid jets can
whose open probability is primarily affected by the mechanical increase calcium spark rate in a way that appears to draw on
environment, and for whom mechanical stimulation in the mitochondrial calcium.60
absence of cell volume changes is sufficient to promote opening. In other cell types, the nuclear envelope has been found to
SACs typically respond to a range of mechanical stimuli, contain SACs that contribute to calcium signaling in this
including local membrane deformation, changes in cell curvature, domain,61 and it would not be surprising if the same applied to
lateral compression, and axial stretch. Whether transfer of the heart.
mechanical energy to the ion channel protein occurs mainly via Finally, connexin channels can be mechanosensitive. Con-
the cytoskeleton, or via the lipid bilayer, is a matter of debate. It nexins are best known for linking the cytosols of two contacting
is likely that both are involved to individually varying degrees. cells; although “sarcolemmal,” they do not connect a cell’s
Some SACs, such as the transient receptor potential channel inside to the outside in this configuration. Noncardiac con-
(TRPC1), can be activated in pure lipid bilayers.48 Other SACs nexin46 has been shown to be mechanosensitive.62 It will be
are sensitive to cytoskeletal integrity, which could either promote interesting to see whether the same applies to cardiac isoforms,
or prevent their opening.49 such as connexin43, whether located in the sarcolemmal or in the
The actual mechanisms of cardiac SAC activation are not well internal membranes of cardiac mitochondria.63 In fact, mitochon-
established. Open and closed state data from large conductance dria could be more important players in cardiac MEC than cus-
prokaryotic mechanosensitive channels (MscL) have identified tomarily assumed, because they are able to generate significant
one mode of action. This action involves an irislike increase in intracellular forces during mitochondrial swelling,64 and hence
pore dimensions during channel opening, involving an increase possibly link metabolic disturbances to mechanisms involved
in the outer circumference of the protein in the plane of the in MEC.
14
20 µm
–2 s
T SL = 1.80 µm
0s
∆SL ∆SL
1
F[Ca2+]REL
SL = 1.94 µm
2s
Figure 14-6. Time course of relative Fluo-4 signal intensity in a rat ventricular resting cardiomyocyte, illustrating dynamic changes in spatially resolved Ca2+ concentration
before and during axial cell stretch. Upright panels show cell images, averaged from 10 confocal XY scans, before (back) and after (front) stretch application. Low signal-
intensity areas, overlapping the cell, reveal carbon fiber (CF) positions (scale bars, 20 µm). In between, fluorescence intensity in each confocal XY scan was added along
the y-axis and plotted as a pseudo-3D XT-sequence of relative Ca2+ fluorescence. Note the corrugated appearance of background intensity, which is indicative of sarcomere
length (SL). Axial stretch was applied by lateral movement of both CF (shading across XT plot indicates period of CF movement), increasing SL in the affected area by
approximately 8%. This is associated with an increase in spark rate.
(Data from Iribe G, Ward CW, Camelliti P et al: Axial stretch of rat single ventricular cardiomyocytes causes an acute and transient increase in Ca2+ spark rate. Circ Res 104:787–795,
2009. Figure design courtesy Dr. Alan Garny, University of Oxford.)
contain streptomycin. In addition, efficacy in vitro cannot simply

Pharmacologic Probes be extrapolated to the in situ setting, where it would appear that
cardiac SACs, accessible to GsMTx-4, can be protected from
In the absence of direct access to several of the relevant ion acute block by streptomycin.66
channel populations by patch-clamp approaches, SAC contribu- The mechanisms of action of SAC block have not been
tions to cardiac electrophysiology have been studied by pharma- resolved, but they might include screening of negative charges
cologic block. There are three main types of SAC blockers (popular blockers are cations with net charges ranging from 2+
commonly used in experiments: the ionic form of gadolinium for streptomycin to 5+ for GsMTx-4), interactions with the lipid
(Gd3+; typically 10 to 100 µM), aminoglycosidic antibiotics (e.g., bilayer, and open channel block (for detail on SACs as pharma-
streptomycin, 30 to 50 µM), and a tarantula venom peptide cologic targets, see the review by White67).
(Grammostola spatulata mechano-toxin 4, GsMTx-4; effective at
concentrations of 100 nM using the native peptide, but rising to
1 to 5 µM with at least some of the commercially available syn-
thetic forms).
Manifestations of Cardiac Stretch–Activated
All these substances have drawbacks. Gadolinium suffers from Ion Channel Activation
a lack in specificity (overlapping concentration range for block of
L-type calcium, sodium, and rapid delayed rectifier potassium Cell-Level Responses
channels, as well as of the sodium-calcium exchanger) and from
precipitation in bicarbonate/phosphate-buffered solutions Whole-cell SAC currents have been recorded from most cardio-
(although both Gd3+ and gadolinium salts may affect SACs). Ami- myocyte types, including SAN pacemaker cells, Purkinje fibers,
noglycosidic antibiotics are neither strictly selective (e.g., L-type and atrial and ventricular myocytes. Their effects on cellular
calcium channel block with a half-maximal inhibition at 1 to electrophysiology depend on the effective stretch target (SACNS
2 mM), nor are they necessarily reliable tools for acute SAC or SACK), stretch timing (relative to the cardiac cycle), and
block in situ (if they were, prescribing them as antibiotics would stretch characteristics (e.g., rate of rise, amplitude).
be unlikely). Finally, high-purity GsMTx-4,65 although selective As a rule, diastolic stretch gives rise to depolarization if it is
and effective in both D- and L-configurations, still suffers from of sufficient amplitude to cause any changes in membrane poten-
limited availability and high cost. tial (Vm; Figure 14-8). This can best be explained by the activa-
Given the above limitations, streptomycin has emerged as a tion of SACs whose reversal potential is positive to the resting
popular pharmacologic probe to study SACs in vitro, where it cell membrane (i.e., SACNS).
efficiently and reasonably specifically (at micromolar concentra- The effects of stretch during the AP are less clear, because AP
tions) blocks whole-cell currents activated by axial stretch (Figure shortening, crossover of repolarization, and delayed repolariza-
14-7). In this context, caution is advised when using cultured cells tion have all been reported. Some of this discrepancy could be
to study mechanosensitive behavior, because many culture media explained by differences in recording techniques, which can affect
80 stretch targets. For instance, sharp electrodes and a perforated or

60 ruptured patch can give rise to opposite AP duration changes,
40
20 hinging on differential effects of these techniques on the interac-
tion between SACs and calcium handling.68 Stretch configuration
mV
0
20 matters as well, because moderate stretch initially affects only AP
40
60 duration (the AP plateau is electrophysiologically more labile
80 than the resting potential), whereas more severe distension alters
A diastolic behavior, potentially causing crossover of late AP repo-
larization and EAD-like behavior (see Figure 14-8). Another
400 confounding aspect is related to the repolarization level at which
Control AP duration is measured. Because early AP shortening may coin-
200
cide with late AP-prolongation, this can give rise to apparently
pA
0
contradictory changes in reported AP duration.
200 Whereas stretch of normal myocardium appears to preferen-
B tially target SACNS (whose block fully eradicates whole-cell
current responses in isolated cells; see Figure 14-7), this could be
400 different under conditions of metabolic impairment. Thus, KATP
ST channels that are normally quiescent in atrial and ventricular
200
pA
0 cardiomyocytes respond to pipette suction46 and axial stretch47

when preactivated by a reduction in ATP concentration. Coacti-
200
vation of a potassium-selective channel by ATP reduction and
C stretch will give rise to a less-positive reversal potential of the
“net whole-cell current” induced by stretch in the ATP-starved
400 heart. This might explain the reduced efficacy of PT (less effi-
SM
200 cient eradication of excitable gaps by reduced net inward current)
pA
0 in myocardium with severe metabolic impairment.47

Stretch effects may vary not only with disease, but also with
200
species, age, cell type, and location. For example, subendocardial
D ventricular cardiomyocytes appear to be more likely than subepi-
cardial cells to respond to mechanical stimulation with EAD-like
400 behavior.5 Differences in mechanosensitivity can be caused by (1)
SM/ST
200 regionally or transmurally varying levels of relevant mechanical
pA
0 stimuli; (2) different mechanical properties of cell and tissue

200 components involved in the transmission of external mechanical
stimuli to the actual mechanotransducers; (3) variable expression
E or responsiveness of mechanotransducers; and (4) distinct elec-
400 trophysiologic background properties of affected cells and tissues.
SM/ST − SM The latter appears to underlie species differences in SAN pace-
200
maker responses to mechanical stimulation.
pA
0 The response of the SAN pacemaker cells to stretch seems to

200 chiefly involve SACNS activation, as shown in rabbit SAN isolated
0.0 0.2 0.4 0.6 0.8 1.0 cells during axial stretch.69 The principal effects of stretch on
F Sec SAN cell AP-morphology are qualitatively equivalent to those
observed in SAN tissue (see Figure 14-5). They include (in addi-
pA/pF 0.8 tion to truncation of diastolic and systolic Vm-maxima) accelera-
0.6 tion of spontaneous diastolic depolarization and of early AP
repolarization, which can both act to increase pacemaker rate,
0.4
combined with a slowing of late AP repolarization when Vm is
0.2 below the SACNS reversal potential, which would prolong the
pacemaker cycle and reduce beating rate if it dominated the
60 40 20 20 40 60 response. In other words, SAN pacemaker potential changes that
0.2 mV
move toward the reversal potential of SACNS (–11 mV in rabbit
0.4 SAN cells)69 are accelerated by stretch, whereas those that move
0.6 away from the reversal potential are slowed.
If one compares the SAN pacemaker potential waveforms of
G 0.8
slow- and fast-beating mammalian heart (e.g., guinea pig or
Figure 14-7. Action potential (AP) clamp recording of whole-cell current, induced rabbit versus mouse or rat), it is apparent that murine SAN pace-
by axial stretch of guinea pig ventricular myocyte, in the absence and presence of 40 maker potentials have a very different AP morphology: both
µM of streptomycin. A, AP recorded in control conditions and reapplied to the same upstroke and initial repolarization are extremely fast, followed by
cell as a voltage-command (AP clamp). Compensation currents (B-E) illustrate the
an extended late repolarization phase. The percentage of the
cell’s response to interventions. Note that compensation currents have the opposite
polarity to native transmembrane currents. B, Control. C, After application of 5% pacemaker cycle during which Vm moves away from the SACNS
stretch (ST). D, After return to control length and application of streptomycin (SM). reversal potential dominates murine SAN pacemaking (71% in
E, During 5% stretch in the presence of streptomycin (SM/ST). F, Difference current mouse, compared to 46% in rabbit)66; this might underlie the
(E minus D), illustrating that any SM-resistant stretch-induced currents are negligible. negative chronotropic response to sustained stretch, observed in
G, Current-voltage relation of the stretch-induced whole-cell current, measured murine hearts. Importantly, both negative and positive chrono-
during AP-clamp repolarization from +40 to –40 mV. This current appears to be com- tropic responses to sustained stretch can be abolished by the
mensurate with SACNS and is completely abolished with 40 µM of streptomycin. application of GsMTx-4,66 confirming the pivotal contribution
(Data from Lei M, Cooper PJ, Kohl P. Unpublished.)
of SACNS. Thus, identical mechanisms can give rise to opposite
100 100
50 50 14
[mV]
[mV]
0 SACNS 0
–50 –50
SACK
–100 –100
0 0.3 0.6 0 0.3 0.6 0.9 1.2
[s] [s]
Figure 14-8. Effects of sustained moderate (≤5%) and severe (≥10%) axial stretch on guinea pig isolated ventricular myocyte action potential (AP) morphology. Perforated
patch-clamp recordings show AP shortening in the absence of diastolic membrane potential changes during moderate stretch, applied using a pair of carbon fibers (left),
whereas more severe distension gives rise to diastolic depolarization, early AP shortening, and crossover of AP repolarization, yielding early afterdepolarization-like behavior
(right). Black, Control; red, during stretch; SACNS and SACK indicate differences in reversal potential of SAC populations with different ion selectivity. Note different time scales.
(Data from Kohl P, Nesbitt AD, Cooper PJ et al: Sudden cardiac death by Commotio cordis: role of mechano-electric feedback. Cardiovasc Res 50:280-289, 2001.)
responses, depending on the electrophysiologic background of velocity (reports in the literature are divided between increase,
affected cells. This reemphasizes the need for caution when reduction, and no change, whereas reported effects depend on
extrapolating observations between species, such as from mouse stretch amplitudes and could differ in conduction system versus
to human. working muscle),70 which would be important for the interpreta-
Of course, SAN tissue in situ will not normally be stretched tion of electrophysiologic ensemble data, and for their patho-
throughout the entire cycle, and more refined experimental study physiologic relevance.
designs are needed to apply mechanical stimulation in a cycle- Given this complexity, the identification of SAC contributions
dependent manner. In addition, it will be interesting to explore to cardiac MEC in multicellular preparations has relied largely
whether individual components of the “voltage- and calcium- on pharmacologic probes. For most SAC types, specific and effi-
clocks” that drive pacemaking are directly mechanosensitive in cient openers are not available, whereas available blockers suffer
native SAN cells (e.g., HCN or SR-release channels) and whether from a lack of cardiomyocyte specificity, as well as the limitations
effects secondary to other MEC-actions, such as changes in intra- discussed earlier. In addition, it is not sufficient for a blocker to
cellular calcium concentration, determine pacemaker responses be SAC specific at the single cell level, but it must also be effective
to stretch.70 upon acute application to native tissue. This is not necessarily the
case for streptomycin,66 which calls for careful interpretation, in
particular of apparently negative observations at the tissue level.
Tissue- and Organ-Level Effects Despite these notes of caution, the available evidence over-
whelmingly demonstrates that block of SACNS abolishes stretch-
Cardiac arrhythmias are inherently multicellular phenomena, induced changes in SAN pacemaker rate. It can also prevent the
and it is therefore necessary to understand mechanisms, modula- mechanical induction of arrhythmogenic triggers, such as single
tors, and outcomes of cardiac MEC at tissue and organ levels. PVB,71 and the mechanical promotion of sustained arrhythmias,
However, linking macroscopic to microscopic events is not such as the preload-dependent amplification of burst-pacing–
without challenges in multicellular biologic systems. induced AF.72 In contrast, block of SACK can increase PVB induc-
On the “input side,” quantification of mechanical interven- ibility.55 Interestingly, however, SACK block may still prevent
tions is even more difficult in tissue than it is in cells, where mechanically induced development of VF.73 Whether this is
sarcomere length can be used as an indicator of strain, or in caused by removal of an arrhythmia-sustaining effect (ectopic
membrane patches where deformation can be optically moni- excitation is still observed), a shift in the space-time sensitive
tored, at least in principle. In most tissue preparations—except narrow vulnerable window for mechanical VF-induction, or
for trabeculae, thin papillary muscles, and live tissue slices— another mechanism, remains to be confirmed.
mechanical deformation usually cannot be quantified or graded
with respect to subcellular or cellular strains. In the absence of
cell deformation data, the characterization of externally applied
mechanical stimuli is helpful. However, because of complex vis- Summary and Outlook
coelastic tissue properties that confer a strong time-varying com-
ponent to the translation of external interventions to effective The currently available data suggest that mechanical stimulation,
stimuli for MEC, this is a restricted surrogate measure only. acting via sarcolemmal SACs, gives rise to changes in cardiomyo-
On the “output side,” many of the standard techniques used cyte Vm and AP morphology. SACNS-mediated effects in particu-
to record electrophysiologic consequences of mechanical stimu- lar appear to be sufficient to explain quantitatively the majority
lation in tissue and organ preparations report ensemble proper- of acute functional consequences of cardiac MEC quantitatively
ties (including ECG, MAP, and optical mapping). In this context, in physiologic conditions (Figure 14-9). Sufficiency should not,
it is important to recall that the heart contains a large number of however, be confused with validity, necessity, or exclusivity.
different cell types, the majority of which are not cardiomyocytes. There are complex interactions of SAC effects with other media-
These types include endothelial cells, fibroblasts, smooth muscle, tors of mechano-electric integration, in particular those affecting
and intracardiac neurons, all of which are mechanosensitive and calcium.74 SACs can contribute to altered calcium handling either
can affect cardiac electrophysiologic responses to mechanical directly (e.g., via changes in SR calcium release or via transsar-
stimulation. In addition, stretch can influence conduction colemmal Ca2+-flux through SACNS), or indirectly (via changes in
Cardiac Mechano-electric Coupling Effects linked cardiac muscle preparations,75 and it could act as an equal-
izer of the inotropic state of individual cardiomyocytes (see
Figure 14-9). This mechanism could underlie efficient function
Pathology Therapy Physiology of cells across regionally and temporally varying myocardial
e.g., Commotio e.g., Pre-cordial e.g., Positive Inotropy stress–strain gradients in the healthy heart—in particular as the
cordis thump chronotropy equalizer activity of individual cells is not controlled by neuromuscular
junctions, in contrast to skeletal muscle. Disturbances of this
balance can be arrhythmogenic if they are sustained, because even
∆ Vm/AP ∆ Ca2+ small wall-motion abnormalities in patients are associated with
increased dispersion of repolarization.76
Non-sarco-
Sarcolemmal SAC
lemmal SAC
So—What Next?
∆ Stress/strain In addition to populating and interconnecting the pockets of
current insight into MEC effects and mechanisms, there are a
number of key challenges. These include:
Figure 14-9. Schematic representation of functional relevance (blue) and key
• Development of improved tools to assess strain and stress in
mechanisms (red) of acute cardiac MEC. Effects of MEC are particularly striking in
the context of induction and termination of arrhythmias (top left and middle), and cells and tissue, such as those heralded by the advent of fluo-
can be explained by effects of stretch-activated ion channels (SACs) on cardiomyo- rescent force-sensors for live-cell studies,77 and by stretchable
cyte electrophysiology. The physiologic relevance of cardiac SAC (top right) is not electronics for whole-heart research78
documented as well, which in addition to preload-dependent modulation of pace- • Identification of the mechanisms underlying SAC activation,
making rate is likely to include an additional role: the matching of individual cell including open and closed state SAC structures, differential
contractility to external demand. The latter is linked to mechanical modulation of contributions of stress and strain, and characterization of
cellular calcium handling (ΔCa2+). Dashed lines indicate some of the possible links links between externally applied and internally sensed
between classic MEC components and cell calcium, such as from SAC-effects on mechanical clues, including transmission to nonsarcolemmal
calcium handling, to modulation of SAN pacemaking (see text for detail). Vm, Mem-
ion channels59
brane potential; AP, action potential.
• Integration of MEC research with mechano-mechanic and
acute mechanochemical transduction studies, such as involving
Vm/AP morphology that alter voltage-sensitive Ca2+ fluxes, or reactive oxygen species79,80
through SACNS Na+-influx with knock-on effects on the balance • Exploration of cross-talk between MEC and pharmacologic
of transsarcolemmal ion exchange). interventions, including the identification of MEC compo-
This interplay of SAC effects and Ca2+ handling could, in fact, nents as drug targets and modulators81
hold the key to a better understanding of the physiologic rele-
vance of cardiac SACs. If, for example, an individual myocyte in
situ was “less contractile” than its neighbors, then it would be Acknowledgments
stretched (or prevented from shortening) during systole. If this
contributed to a gain of additional (or preservation of available) The author thanks Dr. T. Alexander Quinn and Dr. Christian
intracellular calcium, then it could enable affected cells to adapt Bollensdorff for helpful comments on the manuscript. This work
their contractility to external demand on a beat-by-beat basis. was supported by the British Heart Foundation, the U.K. Bio-
Such matching of local contractility to dynamically varying technology and Biological Sciences Research Council, the Euro-
external loads has been shown experimentally in mechanically pean Commission, and the Magdi Yacoub Institute.
5. Gallacher DJ, Van de Water A, van der Linde H, simulation study in 3D. J Mol Histol 35:679–686,
References et al: In vivo mechanisms precipitating torsades de 2004.
pointes in a canine model of drug-induced long- 12. Garny A, Noble D, Kohl P: Dimensionality in
1. Brines L, Such-Miquel L, Gallego D, et al: Modi- QT1 syndrome. Cardiovasc Res 76:247–256, cardiac modelling. Prog Biophys Mol Biol 87:47–
fications of mechanoelectric feedback induce1d by 2007. 66, 2005.
2,3-butanedione monoxime and blebbistatin in 6. Schlomka G: Commotio cordis und ihre Folgen. 13. Allessie MA, Ausma J, Schotten U: Electrical,
Langendorff-perfused rabbit hearts. Acta Physiol Die Einwirkung stumpfer Brustwandtraumen auf contractile and structural remodeling during
206:29–41, 2012. das Herz. Ergeb inn Med Kinderhkd 47:1–91, atrial fibrillation. Cardiovasc Res 54:230–246,
2. Guharay F, Sachs F: Stretch-activated single ion 1934. 2002.
channel currents in tissue-cultured embryonic 7. Link MS: Commotio cordis: sudden death from 14. Yoshida K, Ulfarsson M, Oral H, et al: Left atrial
chick skeletal muscle. J Physiol 352:685–701, 1984. blows to the chest wall. In Kohl P, Sachs F, Franz pressure and dominant frequency of atrial fibrilla-
2a. Meola F: La commozione toracica. G Int Sci Med MR, editors: Cardiac Mechano-Electric Coupling tion in humans. Heart Rhythm 8:181–187, 2011.
1:923–937, 1879. and Arrhythmias, Oxford, 2011, Oxford University 15. Moreno J, Zaitsev AV, Warren M, et al: Effect of
2b. Riedinger F: Über Brusterschütterung. In Fest- Press, pp 325–239. remodelling, stretch and ischaemia on ventricular
chrift zur dritten Saecularfeier der Alma Julia 8. Kohl P, Nesbitt AD, Cooper PJ, et al: Sudden fibrillation frequency and dynamics in a heart
Maximiliana Leipzig. Leipzig, 1882, Verlag von cardiac death by Commotio cordis: role of failure model. Cardiovasc Res 65:158, 2005.
F.C.W. Vogel, pp 221–234. mechano-electric feedback. Cardiovasc Res 16. Bollensdorff C, Lab MJ: Stretch effects on potas-
2c. Bainbridge FA: The influence of venous filling 50:280–289, 2001. sium accumulation and alternans in pathological
upon the rate of the heart. J Physiol 50:65–84, 9. Link MS, Wang PJ, Pandian NG, et al: An experi- myocardium. In Kohl P, Sachs F, Franz MR,
1915. mental model of sudden cardiac death due to low- editors: Cardiac Mechano-Electric Coupling and
3. Zoll PM, Belgard AH, Weintraub MJ, et al: Exter- energy chest-wall impact (Commotio cordis). Arrhythmias, Oxford, 2011, Oxford University
nal mechanical cardiac stimulation. N Engl J Med NEJM 338:1805–1811, 1998. Press, pp 173–179.
294:1274–1275, 1976. 10. Garny A, Kohl P: Mechanical induction of arrhyth- 17. Waxman MB, Wald RW, Finley JP, et al: Valsalva
4. Levine JH, Guarnieri T, Kadish AH, et al: Changes mias during ventricular repolarisation: modelling termination of ventricular tachycardia. Circulation
in myocardial repolarization in patients undergo- cellular mechanisms and their interaction in 2D. 62:843–851, 1980.
ing balloon valvuloplasty for congenital pulmonary Ann NY Acad Sci 1015:133–143, 2004. 18. Ambrosi P, Habib G, Kreitmann B, et al: Valsalva
stenosis: evidence for contraction-excitation feed- 11. Li W, Kohl P, Trayanova N: Induction of ventricu- manoeuvre for supraventricular tachycardia in
back in humans. Circulation 77:70–77, 1988. lar arrhythmias following a mechanical impact: a transplanted heart recipient. Lancet 346:713, 1995.
19. Dosdall DJ, Doppalapudi H, Ideker RE: Mechani- regulatory mechanisms. Cardiovasc Res 12:449– 61. Itano N, Okamoto S, Zhang D, et al: Cell spread-
cal modulation of defibrillation and resuscitation 469, 1978. ing controls endoplasmic and nuclear calcium:
efficacy. In Kohl P, Sachs F, Franz MR, editors:
Cardiac Mechano-Electric Coupling and Arrhyth-
mias, Oxford, 2011, Oxford University Press,
40. Casadei B, Moon J, Johnston J, et al: Is respiratory
sinus arrhythmia a good index of cardiac vagal tone
in exercise? J Appl Physiol 81:556–564, 1996.
a physical gene regulation pathway from the cell
surface to the nucleus. Proc Natl Acad Sci U S A
100:5181–5186, 2003.
14
pp 374–380. 41. Bernardi L, Salvucci F, Suardi R, et al: Evidence 62. Bao L, Sachs F, Dahl G: Connexins are mechano-
20. Trayanova N, Li WH, Eason J, et al: Effect of for an intrinsic mechanism regulating heart rate sensitive. Am J Physiol 278:C1389–C1395, 2004.
stretch-activated channels on defibrillation effi- variability in the transplanted and the intact heart 63. Miro-Casas E, Ruiz-Meana M, Agullo E, et al:
cacy. Heart Rhythm 1:67–77, 2004. during submaximal dynamic exercise? Cardiovasc Connexin43 in cardiomyocyte mitochondria con-
21. Li W, Gurev V, McCulloch AD, et al: The role of Res 24:969–981, 1990. tributes to mitochondrial potassium uptake. Car-
mechanoelectric feedback in vulnerability to elec- 42. Duan DD: The ClC-3 chloride channels in cardio- diovasc Res 83:747–756, 2009.
tric shock. Prog Biophys Mol Biol 97:461–478, vascular disease. Acta Pharmacol Sin 32:675–684, 64. Kaasik A, Kuum M, Joubert F, et al: Mitochondria
2008. 2011. as a source of mechanical signals in cardiomyo-
22. Pellis T, Kohl P: Anti-arrhythmic effects of acute 43. Baumgarten CM, Clemo HF: Swelling-activated cytes. Cardiovas Res 87:83–91, 2010.
mechanical stimulation. In Kohl P, Sachs F, Franz chloride channels in cardiac physiology and patho- 65. Bowman CL, Gottlieb PA, Suchyna TM, et al:
M, editors: Cardiac Mechano-Electric Coupling physiology. Prog Biophys Mol Biol 82:25–42, Mechanosensitive ion channels and the peptide
and Arrhythmias, Oxford, 2011, Oxford University 2003. inhibitor GsMTx-4: history, properties, mechanisms
Press, pp 361–368. 44. Craelius W, Chen V, El-Sherif N: Stretch activated and pharmacology. Toxicon 49:249–270, 2007.
23. Eich C, Bleckmann A, Schwarz SKW: Percussion ion channels in ventricular myocytes. Bosci Rep 66. Cooper PJ, Kohl P: Species- and preparation-
pacing – an almost forgotten procedure for haemo- 8:407–414, 1988. dependence of stretch effects on sino-atrial node
dynamically unstable bradycardias? A report of 45. Lin W, Laitko U, Juranka PF, et al: Dual stretch pacemaking. Ann N Y Acad Sci 1047:324–335, 2005.
three case studies and review of the literature. Brit responses of mHCN2 pacemaker channels: accel- 67. White E: Mechanosensitive channels: therapeutic
J Anaesth 98:429–433, 2007. erated activation, accelerated deactivation. Biophys targets in the myocardium? Current Pharmaceuti-
24. Befeler B: Mechanical stimulation of the heart: its J 92:1559–1572, 2007. cal Design 12:3645–3663, 2006.
therapeutic value in tachyarrhythmias. Chest 46. Van Wagoner DR, Lamorgese M: Ischemia poten- 68. Calaghan SC, Belus A, White E: Do stretch-
73:832–838, 1978. tiates the mechanosensitive modulation of atrial induced changes in intracellular calcium modify
25. Klumbies A, Paliege R, Volkmann H: Mechanische ATP-sensitive potassium channels. Ann N Y Acad the electrical activity of cardiac muscle? Prog
Notfallstimulation bei Asystolie und extremer Bra- Sci 723:392–395, 1994. Biophys Mol Biol 82:81–95, 2003.
dykardie. Z Ges 43:348–352, 1988. 47. Kohl P, Bollensdorff C, Garny A: Effects of 69. Cooper PJ, Lei M, Cheng L-X, et al: Axial stretch
26. Amir O, Schliamser JE, Nemer S, et al: Ineffective- mechano-sensitive ion channels on ventricular increases spontaneous pacemaker activity in rabbit
ness of precordial thump for cardioversion of electrophysiology: experimental and theoretical isolated sinoatrial node cells. J Appl Physiol
malignant ventricular tachyarrhythmias. Pacing models. Exp Physiol 91:307–321, 2006. 89:2099–2104, 2000.
Clin Electro Physiol 30:153–156, 2007. 48. Maroto R, Raso A, Wood TG, et al: TRPC1 forms 70. Quinn TA, Kohl P: Mechano-sensitivity of cardiac
27. Haman L, Parizek P, Vojacek J: Precordial thump the stretch-activated cation channel in vertebrate pacemaker function: pathophysiological relevance,
efficacy in termination of induced ventricular cells. Nature Cell Biol 7:179–185, 2005. experimental implications, and conceptual integra-
arrhythmias. Resuscitation 80:14–16, 2009. 49. Janmey PA: The cytoskeleton and cell signalling: tion with other mechanisms of rhythmicity. Prog
28. Pellis T, Kette F, Lovisa D, et al: Utility of pre- component localization and mechanical coupling. Biophys Mol Biol 110:257–268, 2012.
cordial thump for treatment of out of hospital Physiol Rev 78:763–781, 1998. 71. Hansen DE, Borganelli M, Stacy GPJ, et al: Dose-
cardiac arrest: a prospective study. Resuscitation 50. Perozo E, Kloda A, Cortes DM, et al: Physical dependent inhibition of stretch-induced arrhyth-
80:17–23, 2009. principles underlying the transduction of bilayer mias by gadolinium in isolated canine ventricles.
29. Schott E: Über Ventrikelstillstand (Adam- deformation forces during mechanosensitive Evidence for a unique mode of antiarrhythmic
Stokes’sche Anfälle) nebst Bemerkungen über channel gating. Nat Struct Biol 9:696–703, action. Circ Res 69:820–831, 1991.
andersartige Arrhythmien passagerer Natur. (“On 2002. 72. Bode F, Sachs F, Franz MR: Tarantula peptide
Ventricular Standstill (Adam-Stokes Attacks) 51. Morris CE, Juranka PF: Nav channel mechanosen- inhibits atrial fibrillation. Nature 409:35–36, 2001.
together with other Arrhythmias of Temporary sitivity: activation and inactivation accelerate 73. Link MS, Wang PJ, VanderBrink BA, et al: Selec-
Nature.”). Dtsch Arch Klin Med 131:211–229, 1920. reversibly with stretch. Biophys J 93:822–833, tive activation of the K+ATP channel is a mechanism
30. Pennington JE, Taylor J, Lown B: Chest thump for 2007. by which sudden death is produced by low-energy
reverting ventricular tachycardia. N Eng J Med 52. Dyachenko V, Christ A, Gubanov R, et al: Mis- chest-wall impact (commotio cordis). Circulation
283:1192–1195, 1970. alignment of sarcomeres by mechanical stimuli: an 100:413–418, 1999.
31. Chan L, Reid C, Taylor B: Effect of three emer- input signal for integrin dependent modulation of 74. ter Keurs HE, Wakayama Y, Sugai Y, et al: Role of
gency pacing modalities on cardiac output in ion channels? Prog Biophys Mol Biol 97:196–216, sarcomere mechanics and Ca2+ overload in Ca2+
cardiac arrest due to ventricular asystole. Resusci- 2008. waves and arrhythmias in rat cardiac muscle. Ann
tation 52:117–119, 2002. 53. Terrenoire C, Lauritzen I, Lesage F, et al: TREK- NY Acad Sci 1080:248–267, 2006.
32. Iseri LT, Allen BJ, Baron K, et al: Fist pacing, a 1-like potassium channel in atrial cells inhibited by 75. Solovyova O, Katsnelson L, Konovalov P, et al:
forgotten procedure in bradysystolic cardiac arrest. β-adrenergic stimulation and activated by volatile Activation sequence as a key factor in spatio-
Am Heart J 113:1545–1550, 1987. anesthetics. Circ Res 89:336–342, 2001. temporal optimization of myocardial function. Phil
33. Kaufmann R, Theophile U: Automatie-fördernde 54. Tamargo J, Caballero R, Gomez R, et al: Pharma- Trans Roy Soc 364:1367–1383, 2006.
Dehnungseffekte an Purkinje-Fäden, Papillarmus- cology of cardiac potassium channels. Cardiovasc 76. Opthof T, Sutton P, Coronel R, et al: The associa-
keln und Vorhoftrabekeln von Rhesus-Affen. Pflüg Res 62:9–33, 2004. tion of abnormal ventricular wall motion and
Arch 297:174–189, 1967. 55. Iribe G, Jin H, Naruse K: Role of sarcolemmal increased dispersion of repolarization in humans is
34. Tan JH, Liu W, Saint DA: Differential expression BKCa channels in stretch-induced extrasystoles independent of the presence of myocardial infarc-
of the mechanosensitive potassium channel in isolated chick hearts. Circ J 75:2552–2558, tion. Front Physiol 3:235, 2012.
TREK-1 in epicardial and endocardial myocytes in 2011. 77. Meng F, Suchyna TM, Sachs F: A fluorescence
rat ventricle. Exp Physiol 89:237–242, 2004. 56. Coste B, Xiao B, Santos JS, et al: Piezo proteins energy transfer-based mechanical stress sensor for
35. Stones R, Calaghan SC, Billeter R, et al: Transmu- are pore-forming subunits of mechanically acti- specific proteins in situ. FEBS J 275:3072–3087,
ral variations in gene expression of stretch- vated channels. Nature 483:176–181, 2012. 2008.
modulated proteins in the rat left ventricle. Pflug 57. Zeng T, Bett GCL, Sachs F: Stretch-activated 78. Kim DH, Ghaffari R, Lu N, et al: Flexible and
Arch 454:545–549, 2007. whole cell currents in adult rat cardiac myocytes. stretchable electronics for biointegrated devices.
36. Lerman BB, Deale OC: Relation between transcar- Am J Physiol 278:H548–H557, 2000. Annu Rev Biomed Eng. 14:113–128, 2012.
diac and transthoracic current during defibrillation 58. Kohl P, Cooper PJ, Holloway H: Effects of acute 79. Dyachenko V, Rueckschloss U, Isenberg G: Modu-
in humans. Circ Res 67:1420–1426, 1990. ventricular volume manipulation on in situ cardio- lation of cardiac mechanosensitive ion channels
37. Deck KA: Dehnungseffekte am spontanschlagen- myocyte cell membrane configuration. Prog involves superoxide, nitric oxide and peroxynitrite.
den, isolierten Sinusknoten. Pflüg Arch 280:120– Biophys Mol Biol 82:221–227, 2003. Cell Calcium 45:55–64, 2009.
130, 1964. 59. Iribe G, Ward CW, Camelliti P, et al: Axial stretch 80. Prosser BL, Ward CW, Lederer WJ: X-ROS sig-
38. Cooper PJ, Ravens U: Mechanical modulation of of rat single ventricular cardiomyocytes causes an naling: rapid mechano-chemo transduction in
pacemaker electrophysiology. In Kohl P, Sachs F, acute and transient increase in Ca2+ spark rate. Circ heart. Science 333:1440–1445, 2011.
Franz MR, editors: Cardiac Mechano-Electric Res 104:787–795, 2009. 81. Beyder A, Strege PR, Reyes S, et al: Ranolazine
Coupling and Arrhythmias, Oxford, 2011, Oxford 60. Belmonte S, Morad M: ‘Pressure-flow’-triggered decreases mechanosensitivity of the voltage-gated
University Press, pp 95–102. intracellular Ca2+ transients in rat cardiac myo- sodium ion channel NaV1.5: A novel mechanism
39. Donald DE, Shepherd JT: Reflexes from the heart cytes: possible mechanisms and role of mitochon- of drug action. Circulation 125:2698–2706,
and lungs: physiological curiosities or important dria. J Physiol 586:1379–1397, 2008. 2012.
Biophysical Properties
of Gap Junctions 15
Virginijus Valiunas and Peter R. Brink
CHAPTER OUTLINE Structure of Gap Junction Channels

Background 151 Structural analysis has revealed that gap junction channels are
Structure of Gap Junction Channels 151 composed of two hemichannels or connexons linked together
that provide an intercellular pathway between two adjacent cell
Cardiac Gap Junctions: Homomeric, Heterotypic, interiors (Figure 15-1, A). Each hemichannel is composed of six
and Heteromeric Forms 152 connexins. Each connexin contains four membrane-spanning
Biophysical Properties of Cardiac Gap domains (M1-M4; see Figure 15-1, B). The N-terminus pro-
Junction Channels 153 trudes from M1 into the cytosol. M1 and M2 are connected by
an extracellular loop E1, whereas M2 and M3 are connected by
Action Potential Propagation in the Myocardium: a cytoplasmic loop. M3 and M4 are connected by another extra-
The Role of Connexins 156 cellular loop (E2), and the cytoplasmic extension of M4 is the
Non–Voltage-Dependent Regulators of C-terminus (see Figure 15-1, B). This general model is true for
all the cardiac connexins and the other human connexins. Indi-
Channel Patency 157
vidual gap junction channels within the atrial and ventricular
Extrinsic Uncoupling Agents of Cardiac Gap myocardium are found in abundance at the intercalated discs in
Junction Channels 158 the form of membrane complexes or plaques containing a large
number of channels, ranging from hundreds to thousands. A
Ischemia, Mutations, Arrhythmia, and Gap Junctions 159 small portion of such a plaque is depicted in Figure 15-1, A.
These plaques can also form along the lateral surfaces of myo-
cytes in normal myocardium and can become even larger struc-
tures in stressed or diseased myocardium.2
Background Only a few studies have been successful in elucidating the
molecular structure of gap junction channels using electron crys-
The propagation of the cardiac action potential throughout the tallography. The first analysis was performed on a noncardiac
working myocardium is made possible by voltage-dependent Na+, connexin, Cx26, and more recently has been revisited, revealing
Ca2+, and K+ currents and by gap junction channels. One of the structural detail to a resolution of approximately 0.35 nm.3 Cx43
roles of gap junction channels is to permit the passage of currents has also been analyzed using crystallographic methods with a
from cell to cell that are essential for action potential propagation resolution of approximately 1.0 nm.4 Figure 15-1, C, depicts a
throughout the working myocardium. When considering action representation of the helical membrane domains and the extracel-
potential propagation gap junctions can be best understood as lular loops of a hemichannel as viewed from within the plasma
components of the longitudinal resistance within the functional membrane of a cell. Figure 15-1, D, illustrates the structure of
syncytium of the myocardium. Gap junction channels are in Cx26 from the perspective of looking down its long axis directly
series with the cytoplasmic resistance. The intercellular resis- into the channel from the cytoplasm where the four transmem-
tance of gap junctions is in series with the intracellular resistance brane domains of each connexin can be visualized. The structural
of the cytoplasm, and together they represent longitudinal resis- analysis has not been sufficiently detailed to demonstrate clearly
tance. Both resistances can affect conduction velocity within which membrane-spanning domains are forming the channel
the heart, but it is the gap junction channels composed of con- wall. However, the structure shown in Figure 15-1, D, again
nexins that dominates longitudinal resistance. Dual whole-cell depicting the four transmembrane helices, suggests that M1, M2,
patch clamp studies of cardiac myocytes have been used to quan- M3, and even M4 are potential contributors to the lining of the
tify gap junction membrane resistance or junctional conductance pore or the cytoplasmic vestibule.
in vitro. The estimates of junctional resistance for ventricular One approach that has been used to define the pore lining
myocyte cell pairs reveal that it is often an order of magnitude regions of connexins has been the substituted cysteine accessibil-
less than the input resistance of an isolated myocyte, somewhere ity method (SCAM). Substituting cysteine for other amino acids
in the range of 2 to 10 MΩ or 1000 to 100 nS.1 These values thought to be contributing to the pore wall is the first step to
must be considered under estimates because of the series resis- successful use of SCAM. This step is followed by a demonstration
tance of the pipettes that have similar resistance values to the that the substitution does not affect normal channel activity. To
junctions themselves.1 In addition to creating an electrical syncy- establish whether the substituted group is part of the pore wall,
tium for the heart gap junctions also allow the passage of small a thiol reactive agent such as maleimide or a derivative is then
solutes such as cAMP that are able to affect the function of mul- perfused into the preparation while monitoring channel activity.
tiple systems within cardiac myocytes. To better understand how One possible outcome is altered unitary conductance consistent,
gap junction channels contribute to the normal functions of the with the substituted group being a component of the pore wall.
heart and how they participate in cardiac arrhythmias and isch- In fact, the application of SCAM to connexins has provided
emia, it is necessary to first describe their structure and biophysi- varied results when assessing pore lining regions of the four
cal properties. membrane-spanning domains and extracellular loops and
151
Cell 1
E1 E2
Plasma
Connexon M1 M2 M3 M4
membrane
CL
Connexin Cytoplasm
N-term
Channel pore
C-term
Cell 2
A B
C D
Figure 15-1. A, A portion of a junctional plaque. Yellow bilayers are the plasma membranes of two adjacent cells. Each gap junction channel is composed of two hemi-
channels or connexons that are themselves composed of connexins. Modified after reference.55 B, Schematic representation of a connexin with 4 membrane domains,
M1-M4, 2 extracellular loops, E1 and E2, and 3 cytoplasmic domains, N-terminus, cytoplasmic loop, and the C-terminus. C, En face view the hemichannel from the cytoplasm
of a cell. D, Molecular rendition of connexin26 (Cx26) viewed from within the plasma membrane of a cell. Each connexin is colored differently.
(A, Modified from Makowski L, Caspar DL, Phillips WC, et al: Gap junction structures. VI. Variation and conservation in connexon conformation and packing. Biophys J 45:208–218,
1984.)
intracellular regions. The use of SCAM to elucidate the pore wall Table 15-1. Connexins of the Human Heart
structure on Cx46 has generated data most consistent with pore
lining domains or pore access regions in M1 and the E1 loop,5 Ventricles Cx43 > Cx45 > > Cx40 > > Cx46~Cx37
but Cx32 SCAM data are more consistent with M3 as a major Atrium Cx40 > Cx43 ~ Cx45
pore lining helix.5 No SCAM analysis has been performed on
Cx43, Cx40, Cx45, or Cx37. Site-directed mutagenesis has also BB/CS Cx40 ~ Cx43 ~ Cx45
been used for Cx43 and Cx37 with various substitution strategies. AV node Cx45 > Cx40 ~ Cx43
In Cx43, mutations were introduced in M3 that resulted in silent
channels.5 This is not conclusive evidence that M3 participates SA node Cx45 > Cx40 > > > > Cx43
as a functional member of the pore wall, but it does not exclude BB/CS, Bundle branch/conducting system; AV, ventricular; SA, sinoatrial.
the possibility. For Cx37, site-directed mutations in M3 resulted From Severs NJ, Bruce AF, Dupont E, et al: Remodelling of gap junctions and
in altered conductive states, which is consistent with M3 partici- connexin expression in diseased myocardium. Cardiovasc Res 80:9–19, 2008;
pation in forming the pore wall, but does not exclusively demon- and Jansen JA, van Veen TA, de Bakker JM, et al: Cardiac connexins and impulse
strate that either.6 The crystallographic and mutagenic data lead propagation. J Mol Cell Cardiol 48:76–82, 2010.
to one possible explanation: different connexins do not use the
identical structural motifs in the formation of an intercellular
pore despite functional similarities. The crystallographic struc-
ture shown in Figure 15-1, C, indicates that there are potential Their distribution within the heart is not uniform. For example,
surfaces for all four domains contributing to the channel or pore Cx43 is abundantly expressed within the ventricles but is only
and the cytoplasmic vestibule. sparingly expressed within the AV and SA nodes.8,9 Table 15-1
illustrates the relative connexin expression levels within the ven-
tricles, atrium, SA node, AV node, and bundle branch/conducting
system (BB/CS). Connexin 31.9, the ortholog of mouse Cx30.2
Cardiac Gap Junctions: Homomeric, found in the mouse AV node, has thus far not been detected in
Heterotypic, and Heteromeric Forms the human heart. Cx30.2 has been shown to be in part responsible
for conduction delay at the AV node in mice.9 The absence of
Hemichannels composed of six connexins from two closed Cx31.9 in the human AV node suggests that the delay in humans
aligned cells form a linkage via the extracellular domains E1 and might be the result of reduced channel numbers or the presence
E2 to create a complete gap junction channel. The heart does of heteromeric or heterotypic forms of Cx45, Cx40, and Cx43 or
not express all 21 identified connexins7 that are able to assemble possibly the existence of another unidentified connexin within
into functional gap junction channels (see Figure 15-1, A, B). the AV node whose properties mimic those of mouse Cx30.2.
Instead, a select number of connexins are expressed within the As implied in Table 15-1, myocytes within the different
human heart; they are Cx43, Cx40, Cx45, Cx46, and Cx37.8 regions of the heart are able to coexpress connexins. For example,
Biophysical Properties of Gap Junctions 153
individual atrial myocytes express Cx40, Cx43, and Cx45 simul- in Figure 15-2, B, for Cx45 and illustrates the decline of junc-
15
taneously but in differing amounts with Cx40 being the most tional current with sustained Vj. Ij,inst is the junctional current
abundantly expressed connexin. The expression of a single con- recorded at the onset of a voltage step, and Ij,ss is the steady state
nexin within myocytes has the potential to generate functional current. In some cases, individual channel activity can be observed
gap junction channels composed of two identical hemichannels, as shown in Figure 15-2, C. Single-channel recordings for Cx43,
both composed of the same connexin referred to as homomeric Cx40, and Cx45 are shown. Multichannel and single-channel
(i.e., homotypic) channel. Another type of gap junction channel data have allowed the determination of unitary conductance
is also possible, where each hemichannel of two opposing cells is (γj,main) for the cardiac connexins, which are listed in Table 15-2.
homomeric but each cell expresses a different connexin. This The ability to monitor unitary events has also allowed a better
type of channel is heterotypic.7 Finally, because most myocytes understanding of voltage-dependent gating in connexins, which
express at least two and often three connexins, a hemichannel can has been shown to have at least two distinct mechanisms: fast
potentially contain two or possible three different connexins.7,10 gating and slow gating. Fast gating is characterized by a rapid
This type of hemichannel is heteromeric. Two heteromeric hemi- transition from an open state to a residual state (γj,residual; see
channels will form a heteromeric gap junction channel. dashed lines in Figure 15-2, C) or closed state, whereas slow
gating is manifest as a series of subconductive states transitioning
from an open or closed state.10 A systematic study on Cx30 gap
junctions reported five substrate conductances unevenly spaced
Biophysical Properties of Cardiac Gap between γj,main and γj,residual, suggesting that each of the six con-
Junction Channels nexins of a hemichannel act as a subgate.11 Macroscopic record-
ings of junctional currents (see Figure 15-2, B) have also been
The biophysical properties of gap junction channels are best useful in dissecting the molecular mechanisms of voltage-
illustrated using a dual whole-cell patch clamp on isolated cell dependent gating. Plotting the steady state currents generated in
pairs. Figure 15-2, A, depicts a cell pair coupled by gap junctions response to different Vj steps (see Figure 15-2, B) reveals the
with the equivalent circuit for the cell pair. All the cardiac gap relationship between junctional conductance (gj) and transjunc-
junction channels have been studied in connexin-deficient cells tional potential (Vj). gj is derived from the ratio of gj,ss/gj,inst. A
that are then transfected with specific cardiac connexins to better Boltzmann fit of the normalized steady state junctional conduc-
understand how homotypic gap channel forms of Cx43, Cx40, tance for various amplitude voltage steps is shown in Figure 15-2,
Cx45, Cx46, and Cx37 behave. In all cases, each can be shown to D, for Cx43, Cx40, and Cx45.
gate closed with the application of increased transjunctional An important parameter derived from the Boltzmann fit is Vj,o.
voltage (Vj). A macroscopic record of junctional currents is shown This parameter represents the half inactivation voltage or that
V1 V2 Ij,inst
I1 I2
Ij,ss
Rj Ij
Rm,1
Rm,2
100 pA
2s
A B
1.0
Cx40 125 pS
Ij 26 pS
2 pA 0 Cx43
gj (norm)
2s
Cx43 54 pS 0.5
Ij 11 pS
1 pA 0
2s Cx40
Cx45 Cx45
25 pS
Ij 0 0.0
1 pA –100 –50 0 50 100
C 2s D Vj (mV)
Figure 15-2. A, Dual whole-cell patch clamp equivalent circuit and superimposed cell pair. The patch electrodes ideally are significantly lower resistance than the junctional
(Rj) and nonjunctional resistances (Rm). The circuit allows for the computation of junctional conductance (gj) with the following relationships: Vj = V2 − V1; I1 = Im,1 + Ij; I2 = –Ij;
gj = Ij / Vj. B, Macroscopic record of homotypic Cx45 using Vj steps of ±10, ±30, ±50, ±70, ±90, and ±110 mV of 10 seconds in duration, which illustrates the voltage depen-
dence of Cx45 and the varied time constants of inactivation with Vj. Ij,inst is the instantaneous junctional current and Ij,ss is the steady state current. C, Single-channel properties
of Cx43, Cx40, and Cx45 channels. Records of single-channel currents observed in pairs of HeLa cells transfected with Cx40, Cx43, and Cx45. Solid line, zero current level;
dashed line, residual current level. The current histograms (right) revealed γj,main = 125 pS and γj,residual = 26 pS for Cx40, γj,main = 54 pS and γj,residual = 11 pS for Cx43, and γj,main
=24 pS for Cx45 cell pairs; pipette was filled with 120 mM K+aspartate–. D, Normalized gj plotted against Vj for Cx43, Cx40, and Cx45. The red dots represent the record
shown in (B). Normalized junctional conductance (gj) is steady state conductance gj,ss normalized against the instantaneous conductance gj,inst.
(Data modified from Valiunas V, Beyer EC, Brink PR: Cardiac gap junction channels show quantitative differences in selectivity. Circ Res 91:104–111, 2002; and Valiunas V, Gemel
J, Brink PR, et al: Gap junction channels formed by coexpressed connexin40 and connexin43. Am J Physiol Heart Circ Physiol 281:H1675–H1689, 2001.)
transjunctional voltage steps. The time course of voltage-

Table 15-2. Multichannel and Single-Channel Data of Different Types
dependent inactivation in junctional currents as shown in Figure
of Gap junction Channels
15-2, B, for Cx45 are similar to the behavior of the other cardiac
Channel Type Vj,o (mV) γj,main (pS) References connexins, although it should be noted that Cx45 has the largest
inactivation time constants of the cardiac connexins, approxi-
Homotypic mately 0.1 seconds for large transjunctional potentials and 3
seconds for smaller transjunctional steps.15
Cx43 ±60 55 14, 19, 46, 47 Heterotypic forms that have been studied often generate
Cx40 ±48 125 19, 46, 48, 49 asymmetric voltage dependent I-V curves, an example of hetero-
typic Cx43-Cx45 is shown in Figure 15-3, B, along with hetero-
Cx45 ±20 25 50, 51
typic Cx40-Cx45 (see Figure 15-3, C) and Cx40-Cx43 (see Figure
Cx46 ±42 128 52 15-3, D). Note that for one voltage polarity, a voltage-dependent
Cx37 ±25 350 53 deactivation or decline in junctional current is present, much like
the homotypic forms. For the other polarity there is, in effect,
Heterotypic little or no voltage-dependent closure. In the case of Cx40-Cx45
and to a lesser degree Cx43-Cx45, there is an increase in junc-
Cx40-Cx43 −80/>100 60/100* 14, 46, 47
tional current, which is best illustrated by the plots of gj versus
Cx40-Cx45 n.d./14 40 14, 54 Vj. This observation suggests that heterotypic gap junction chan-
Cx43-Cx45 −100/12 40 14, 47, 51 nels have altered voltage sensing and gating relative to their
homotypic parents and that Po for these forms might be signifi-
Cx37-Cx43 −70/>100 50/175* 53 cantly less than unity, or the asymmetric unitary conductance
Co-expressed
observed in heterotypic channels is itself voltage dependent.
Cx40/Cx43 ±70 31-130† 46, 54

Cx37/Cx43 ±30/>100 35-280† 53 The Number of Functional Channels within
the Intercalated Disc
Values represent averages derived from representative studies. Experimental
conditions: room temperature; pipette solution: K+ aspartate– if available.
*Conductance polarity dependent. The formation of large plaques or aggregates of channels that are
†
Conductances from homotypic, heterotypic, and heteromeric channels are closely packed together is an interesting synapselike feature of
possible. gap junctions (see Figure 15-1), and it raises a number of ques-
tions, such as are all the channels simultaneously active or are
there active and inactive or silent populations? Furthermore, is
each channel functioning as an identical but independent channel,
or is there evidence for nonindependent, interdependent, or
transjunctional potential where half the channels can be consid- cooperative behavior? Or can channels shift between states that
ered closed. The instantaneous junctional conductance remains in the case of gap junctions would be moving from mostly patent
relatively constant regardless of the voltage, implying that many to mostly closed? Recall that active gap junction channels possess
if not all the gap junctions are patent when Vj is zero. Analysis high open probabilities approaching unity. Thus, accurate mea-
of single and multichannel records of Cx43 (see Figure 15-2, C) surement of total junctional conductance and knowing the unitary
have been used to determine open probability under chronic conductance for a particular connexin allows an estimate of the
application of Vj between 20 and 40 mV yielding values ranging total number of functioning channels operating between a cell
from 0.5 to 0.95, consistent with the idea that gap junction chan- pair. To determine whether all of channels within a plaque are
nels are in the open state when there is no applied transjunctional functional first requires a determination of the number of chan-
potential. The application of Vj greater than 50 mV reduces the nels within any one plaque; second, it requires the determination
mean open time, whereas mean closed time remains relatively of the number of functional channels within that particular
constant, which translates into reduced open probability (Po) plaque. By tagging Cx43 with a fluorescent reporter such as GFP,
with increased Vj amplitude.12 This is consistent with the behav- it is possible to image plaques within cell pairs and further directly
ior manifest in macroscopic recordings for Cx43. Nonstationary assess the junctional conductance using a dual whole-cell patch
analyses of Cx43 and Cx37 have also revealed similar results with clamp. An analysis of experiments using this dual approach of
open probabilities less than 0.95 for Cx43 and 0.7 for Cx37.13 imaging and electrophysiologic assessment of junctional conduc-
Table 15-2 lists Vj,o and unitary conductance for the cardiac tance has revealed that approximately 10% of the channels are
connexins, which includes homotypic, heterotypic, and hetero- functioning within junctional membrane plaques16; furthermore,
meric forms. The values of Vj,o for homotypic channels vary from it appears that Cx43 channels displayed non-independent behav-
20 mV for Cx45 to 60 mV for Cx43. Also listed are all the Vj,o iors associated with phenomena such as transitioning between an
values for heterotypic forms that have thus far been determined. active patent state and a silent state on the order of many seconds
In addition, values for cells coexpressing Cx43/Cx40 and Cx43/ to minutes, which represents an example of mode shifting.12
Cx37 are given. Unitary conductance for the cardiac connexins These surprising results beg the question: why are there so many
varies greatly, as seen in Table 15-2. For heterotypic forms, the apparently silent channels (Po = 0) in presumed dynamic equi-
observed unitary conductance can be polarity dependent.14 librium with a lesser population of almost continually open (Po
Observation of cell pairs coexpressing connexins reveals a range ≈ 1) channels? Are there conditions or circumstances in which
of conductances consistent with channels of different connexin the silent channels can be activated rapidly via phosphorylation,
content. for example? Or are the silent channels already designated or
The macroscopic junctional currents shown in Figure 15-2, identified for internalization as connexosomes (internalized gap
B, illustrate that the currents decline (deactivation) with a junction membranes) to be trafficked to lysosomes?17 The signifi-
prescribed time course on the order of 0.1 to 1 seconds cance of these observations might have little bearing on normal
for Cx43.15 The time constant for voltage-dependent inactiva- action potential propagation, but it remains unclear in regard to
tion is itself voltage dependent and becomes shorter with larger proarrhythmic and antiarrhythmic processes.
Heterotypic channel
0 mV 0 mV
15
V1
V2 +90 mV
A –90 mV
Cx43–Cx45 Cx40–Cx45 Cx40–Cx43
Ij Ij Ij
100 pA 100 pA 100 pA
2s 2s 2s
1.5 1.5 1.5
1.0 1.0 1.0

gj (norm)
gj (norm)
gj (norm)
0.5 0.5 0.5
Cx45 negative Cx40 negative Cx43 negative

0.0 0.0 0.0
–100 –50 0 50 100 –100 –50 0 50 100 –100 –50 0 50 100
B Vj (mV) C Vj (mV) D Vj (mV)
Figure 15-3. A, Rendition of the heterotypic channel where each hemichannel is homomeric, but the connexins for each are different. A voltage step profile for the records
shown in B–D are illustrated to the right. B, Macroscopic current record of Cx43-Cx45 heterotypic gap junction channels. The lower panel shows the normalized gj versus
Vj showing both steady state gj,ss (blue circles) and instantaneous gj,inst (green circles). Cx45 gates close when the potential in the opposing cell of a pair is more positive;
therefore, it gates closed with a negative potential relative to its neighbor (Cx45 negative). C, The same paradigm for a Cx40-Cx45 heterotypic junction. D, The same para-
digm for a Cx43-Cx45 heterotypic junction.
to a control versus junctional conductance measured simultane-

Cardiac Gap Junction Permeability to Ions ously.20 CAMP permeability to Cx43 is much more permissive
and Other Solutes than Cx40.
For a number of the exogenous probes and select messengers,
The major intracellular ion involved in intercellular current flow it has been possible to determine their flux relative to K+ for
via gap junctions is K+. Estimates of the number of K+ flowing specific homotypic connexins, namely Cx43 and Cx40. Figure
through a single gap junction channel per second in response to 15-4, D, is a semilog plot of flux of either an ion, probe, or solute
a voltage step of approximately 23.4 mV or the equivalent of a relative to K+. The monovalent cations are Li+, Na+, TMA, and
10× concentration gradient range from approximately 18.3 × 106 TEA. Lucifer yellow, cAMP, and two oligonucleotides (short
ions per channel per second for Cx40 to 8.1 × 106 ions per interforing RNA [siRNA] mimics) represent solutes of varying
channel per second for Cx43 and 3.7 × 106 ions per channel per size and charge. The graph illustrates what is considered a general
second for Cx45.18,19 A number of studies have also determined rule for connexin channels; monovalent cations have permeabili-
the permeability of the major cardiac connexins to a variety of ties relative to K+ that are similar to an Eisenmann series I or II,
exogenous fluorescent probes and endogenous second messen- which is much like their mobility in free solution. As solute
gers such as cAMP.20 diameter increases, differences in permeability for the same
Permeability of gap junction channels has been defined quali- solute begin to appear between Cx43 and Cx40. The permeability
tative and quantitative. A common qualitative approach for such of two synthetic oligonucleotides are also plotted; they are long,
studies is the use of an exogenous fluorescent probe that is intro- rod-shaped molecules (morpholinos) whose minor diameters are
duced in one cell and detected in another cell of a pair (Figure 1.0 to 1.1 nm.21
15-4, A). For quantification of any specific probe’s permeability, The differences in permeability of LY and cAMP for Cx43
it is necessary to simultaneously measure junctional conductance and Cx40 are shown in Figure 15-4, D, and it should be noted
and fluorescence distribution and intensity of the probe via that TMA, a larger probe but of lesser charge density than cAMP,
imaging.19 is more permeable than cAMP with a larger smaller minor diam-
Figure 15-4, A, shows cell to cell the transfer of fluorescent eter for both Cx43 and Cx40. This strongly suggests selectivity
dye Lucifer yellow (LY). Comparison of LY permeability for based not only on size but on charge as well.
Cx43, Cx40, and Cx45 is a notable example of selectivity, An important aspect of gap junction mediated transfer of
as illustrated in Figure 15-4, B. Lucifer yellow with a minor second messengers is best illustrated when considering cAMP
diameter of 0.95 nm and a net negative charge is greater than permeability. It has the potential to affect a multitude of cellular
tenfold more permeable to Cx43 than Cx40 and threefold more functions. One function is the ability to positively shift the
permeable to Cx45.18,19 The relative fluorescence intensity in a voltage dependence of human HCN4, resulting in an increase in
target cell of a pair is plotted at the 12-minute mark versus the Po of HCN4 channels at any particular membrane potential that
measured junctional conductance measured simultaneously in in turn can affect pacemaker rate within the heart.22 The cell-to-
Figure 15-4, B.19 cell diffusion of cAMP might also participate in the generation
Figure 15-4, C, illustrates the differences in cyclic adenosine of a positive inotropic effect. This scenario implies a sparse inner-
monophosphate (cAMP) permeability for Cx43 and Cx40, where vation density in the ventricular myocardium where only one
the y-axis is the ratio of a reporter channel current density relative myocyte of a number of gap junction–coupled myocytes receive
A 1 min 12 min
Relative intensity 1.0 LY Cx43

0.8 Cx45
0.6
0.4
0.2
Cx40
0.0
0 10 20 30 40 50 60 70
B Conductance (nS)
1 Li Na Cs TMA
TEA
0.08 cAMP
Flux (Fx/FK+)
Cx43 cAMP
0.1
LY
Arbitrary units
0.06
0.01 12-mer
0.04 Cx40
0.02 0.001
24-mer
0.00 0.0001
0 10 20 30 40 50 60 0 5 10
C Conductance (nS) D Solute minor diameter (Å)
Cx40 Cx43
Figure 15-4. A, Three images of a cell pair where the electrode on the left contains a known concentration of fluorescent dye Lucifer yellow (LY). Once in whole-cell patch
mode, fluorescence emission is monitored over time (1- and 12- minute records shown). Scale bar, 20 µm. B, Relative fluorescence intensity of LY in target cell, compared
with source cell plotted versus junctional conductance for Cx43, Cx40, and Cx45. All fluorescence measurements are made at the same time interval. C, Cyclic AMP perme-
ability to Cx43 and Cx40 versus conductance. cAMP is approximately 10-fold more permeable to Cx43 than Cx40. Arbitrary units are the ratio of current density for a cAMP
activated channel before and during exposure in elevated intracellular cAMP levels. (D) Semilog plot of Li+, Na+ Cs+, cAMP, TMA, TEA and LY flux relative to K+ ion for Cx43
and Cx40. Permeability of 2 oligonucleotides (12 and 24 nucleotides long) to Cx43 is also plotted.
(Data from Valiunas V, Beyer EC, Brink PR: Cardiac gap junction channels show quantitative differences in selectivity. Circ Res 91:104–111, 2002; Kanaporis G, Mese G, Valiuniene
L et al: Gap junction channels exhibit connexin-specific permeability to cyclic nucleotides. J Gen Physiol 131:293–305, 2008; Valiunas V, Polosina YY, Miller H, et al: Connexin-
specific cell-to-cell transfer of short interfering RNA by gap junctions, J Physiol 568(Pt 2):459–468, 2005; and Brink PR: Gap junctions in vascular smooth muscle. Acta Physiol
Scand 164:349–356, 1998.)
autonomic input. The autonomic input to the innervated myocyte

results in the rapid generation of elevated cAMP within its cyto- Action Potential Propagation in the
plasm. The cAMP then diffuses to surrounding cells that are not Myocardium: The Role of Connexins
innervated. The time to reach half concentration in adjacent cells
coupled along their lateral borders would be on the order of 1 to Two questions arise when considering the role of gap junction
5 seconds, assuming a junctional permeability illustrated in channels in the propagation of the cardiac action potential. First,
Figure 15-4 for cAMP and a cytoplasmic diffusion coefficient of what is the relationship between gap junction number and con-
approximately 1 × 10−6 cm2/s.23 Such a time course is consistent duction velocity? Second, does gap junction channel voltage
with the time course for sympathetic induced inotropy of the dependence matter? The first question has been best addressed
heart. The notion of sympathetic and parasympathetic innerva- using a combination of experimental data and computational
tion density being less than one-to-one for nerve to myocyte, analysis. In vitro studies of cell pairs and isolated tissues have
while never having been quantitatively assessed is consistent shown that a pharmacologically induced reduction in gap junc-
with observations describing low innervation density in the ventional conductance slows conduction and ultimately can
tricular myocardium.24 block conduction, whereas increased expression enhances con-
Another possible role for connexins in the heart and other duction.26 Similarly, many arrhythmias have been associated with
syncytial tissues is the passage of microRNAs and siRNAs able reduced gap junction channel numbers, redistribution of junc-
to affect gene expression and ultimately cellular phenotype.25 As tional plaques on the lateral surfaces, or mutations within con-
a general rule, cardiac connexins can be considered to be poorly nexins.2 Using both experimental and computationally derived
selective toward monovalent cations as an Eisenmann series 1 data, it has been shown that there is a nonlinear relationship
implies and are only moderately more selective toward the mon- between conduction and longitudinal resistance dominated by
ovalent halide anions. With increased size and charge, Cx40 and gap junctions. Simulation of the experimental data predicts that
Cx45 appear to be less permissive or more selective than Cx43. conduction velocity is roughly proportional to the log10 of
80 have an effect on the propagation or conduction of the cardiac
15
action potential. In fact, the normal AV19 node delay of 0.12
seconds is thought to be in part a result of sparse gap junction
density.27 Conditions such as AV block might be manifest by
60
reduced gap junction channel number and voltage-dependent
Conduction velocity
closure, effectively elevating longitudinal resistance to a point

where insufficient current can spread from cell to cell to trigger
40 continued conduction. Results from animal model systems where
connexin knockouts of Cx40 have been constructed are consistent
with this notion.9
20
Pacemaker Activity and Connexins
0 Besides the essential role of connexins in allowing the propaga-
10 100 1000 10000 100000 tion of the cardiac action potential throughout the myocardium,
they are also of paramount importance in the generation of
Number of channels pacing activity of the sinoatrial node (SA node). The myocytes
Figure 15-5. Conduction velocity is plotted against the log of the number of gap of the SA node vary in morphology and form gap junction com-
junction channels. Both simulation data and experimental data have been used to posed of Cx45 and to a lesser degree Cx40 (see Table 15-1).
generate the depicted relationship. . There are no easily demonstrable intercalated discs, rather the
(From Brink PR, Cronin K, Ramanan SV: Gap junctions in excitable cells J. Bioenerg
myocytes form smaller junctional plaques.9 SA node cells are
Biomembr 28:351–358, 1996; and Cole WC, Picone JB, Sperelakis N: Gap junction
characterized by phase 4 depolarization resulting from the activ-
uncoupling and discontinous propagation in the heart. A comparison of experimen-
ity of HCN4 and to a lesser degree HCN1,28 which is sufficient
tal data with computer simulations. Biophys J 53:809–818, 1988.)
to elicit an action potential in the nodal cells and subsequent
propagation to the surrounding atrial myocardium.
Artificial pacemaker units consisting of two coupled cells have
gap junction channel number over a range of 10 to 1000 nS been shown to pace at 1 Hz where one cell of a pair is a non-
(Figure 15-5). myocyte cell expressing HCN, and the other is a ventricular
The second question as to whether gap junction voltage myocyte by itself is not capable of pacemaker activity.29 The data
dependence can have a role in conduction velocity requires defin- clearly demonstrate the role of gap junctions in the initiation of
ing current flow longitudinally within myocytes in response to a pacemaker activity and further illustrate that pacing rate is unaf-
propagating action potential. This definition then allows the fected over an approximately fourfold range of junctional con-
determination of the transjunctional voltage experienced at the ductance (Figure 15-6).
intercalated disc. It is assumed that a myocyte is approximately
100 µm long (L) and has a diameter of approximately 15 µm and
that myoplasmic resistance is approximately 400 Ω-cm. Thus, the
longitudinal resistance of a myocyte is approximately 2 MΩ. Non–Voltage-Dependent Regulators
Assuming that conduction velocity (θ) is 50 cm/s and that the of Channel Patency
maximum rate of rise for the action potential is 100 V/s, the
longitudinal voltage drop along the long axis of the cell can be There are two intrinsic intracellular elements that are able to
determined by Vcell = ([V / s] / θ) × L, or 20 mV. The longitudinal affect junctional conductance: intracellular pH and intracellular
current flow is then 20 mV / 2 MΩ =10 nA. For an intercalated calcium. Lowered intracellular pH, as occurs in ischemia,30 is
disc with 1.8 × 105 functioning channels or a junctional conduc- known to affect many cardiac membrane channels and transport-
tance of approximately 1000 nS, the amount of current flowing ers and can effectively reduce gap junction conductance.31 Figure
through each patent channel is 0.05 pA. The former assumes a 15-7 shows the effect of perfusion with 100% CO2 on junctional
channel population of homotypic Cx43 channels each with a conductance measured between a pair of cells expressing Cx43.
unitary conductance of approximately 55 pS (19). Therefore, the A pH-sensitive fluorescent probe was used to determine intracel-
transjunctional voltage is 1 mV per channel. For 20,000 channels lular pH (see Figure 15-7, B). Many investigators have observed
the value is approximately 10 mV per channel, and for 2000 that short exposures to elevated H+ results in reduced coupling
channels it is approximately 100 mV per channel. Homotypic with subsequent recovery, as illustrated in Figure 15-7. What is
Cx43 unitary conductances of 55 pS are observed when using not shown in Figure 15-7 is the result of prolonged exposure of
K+aspartate– pipette solutions (see Table 15-2) that best mimic many minutes, which results in an irreversible reaction culminat-
the myoplasmic electrolytes. The other factor to consider is the ing in permanent channel closure, implying both rapid and slow
time course of the voltage dependence. It is possible for a trans- processes triggered by H+. The majority of the connexins are
junctional voltage of 10 to 20 mV or larger, as might occur with similarly affected by acidification, responding to elevated intra-
only a 1000 channels or fewer to result in voltage-dependent cellular H+ with reduced junctional conductance that is presumed
channel closure. To assess this possibility, it is necessary to deter- to be the result of increased closed times and reduced open times.
mine how long a transjunctional voltage will persist during the However, the mechanisms of H+-induced closure are apparently
passage of an action potential. What is the duration of the trans not universal. The mechanism of pH-induced alteration of Cx43
junctional voltage for an action potential conducting at 50 cm/s? gap junction channel open probability has been shown to be
For θ = 50 cm/s, the cardiac action potential will traverse a manifest by a ball-and-chain configuration between the
100-µm length in 0.2 ms or 2 ms for θ = 5 cm/s. The time course C-terminus and the cytoplasmic loop between membrane-
of voltage-dependent closure varies from connexin to connexin, spanning domains M2 and M3. In contrast, Cx40 is also pH
but for the cardiac connexins a 2-ms duration would result in a sensitive, but the mechanism of channel closure is not mediated
small reduction in junctional conductance. by a ball-and-chain–like mechanism. The mechanism by which
It is easy to see that only with reduced numbers of functional H+ affects Cx40 channel patency has not been elucidated, but the
channels would their number or voltage dependence begin to pKa for Cx40 is essentially the same as that for Cx43 (6.7).32 The
Carbenoxolone
2.0 2.0
1.5 1.5
Frequency (Hz)
Frequency (Hz)
1.0 1.0
0.5 0.5
0.0 0.0
0 2 4 6 8 10 12 14 16 0.0 0.5 1.0 1.5 2.0 2.5 3.0
A Conductance (nS) B Time (min)
Figure 15-6. A, Pacing frequency generated by a cell transfected with HCN2 coupled to a ventricular myocyte. Frequency is plotted against the measured junctional
conductance. The dashed line corresponds to an average frequency of 1 Hz. B, Carbenoxolone exposure results in a cessation of pacing. Pacing remains constant for more
than 2 minutes and then abruptly stops. This is consistent with the notion that junctional conductance only affects propagation once a critically low level of coupling is
attained.
(Modified from Valiunas V, Kanaporis G, Valiuniene L, et al: Coupling an HCN2-expressing cell to a myocyte creates a two-cell pacing unit. J Physiol 587(Pt 21):5211–5226, 2009.)
CO2 accomplished using an inside-out patch clamp in combination

with two Cx46 constructs, a wild type with an intact C-terminus,
and a deletion mutant of Cx46 missing a significant portion of
the C-terminus. Both forms had the same pH sensitivity with a
Ij pKa of 6.5, clearly suggesting a site or sites for Cx46 other than
100 pA
the ball-and-chain mechanism of Cx43.10
1 min
A Another intracellular ion able to affect Cx43 junctional con-
ductance is calcium. Elevated intracellular calcium (500 nM to
1 µM) reduces the number of functioning Cx43 gap junction
channels and eventually results in complete uncoupling. The
mechanism of calcium-mediated channel closure has been the
center of some controversy, but recently it has been demonstrated
pHI = 7.4 pHI = 6.6 pHI = 6.1 pHI = 7.2 that calcium acts to reduce Cx43 gap junctional conductance via
B
calmodulin.33 It is possible that calcium–pH synergy is mediated
through calmodulin, but there are currently no data to support
such a hypothesis. Interestingly, gap junction channels composed
1.0 of Cx40 are not affected by elevated intracellular calcium and do
not possess the putative calmodulin binding sites found on Cx43.
In addition to affecting the gating of connexins, calcium is also
permeable to connexins and has been associated with cell death
within the myocardium and other syncytial tissues.34 These find-
pHI (norm)
gj (norm)
ings suggest that it is Cx40 that allows the spread of elevated

0.5 intracellular calcium in infarcts, whereas Cx43 is the connexin
working to preserve cellular integrity by isolating healthy myo-
cytes from those that are damaged in an ischemic episode.
Calcium and calmodulin sensitivity of Cx46, Cx45, and Cx37 has
not been studied as completely as it has in Cx43 and Cx40.
0.0
0 2 4 6 8 10 Extrinsic Uncoupling Agents of Cardiac Gap

C Time (min) Junction Channels
Figure 15-7. A, Junctional current is reduced when 100% CO2 is bubbled into the
perfusate. The bar represents the exposure time, which is approximately 2 minutes. Anesthetics such as halothane have shown to reduce mean open
The time course to reduction in junctional current and subsequent recovery are time and increase mean closed time for Cx43 and Cx40 homo-
not the same, but both represent examples of diffusion time within the bath and typic and heteromeric channels in a dose-dependent manner.
buffering capacity of the cells. B, Intracellular pH can be monitored via pH-sensitive
The mechanism by which halothane reduces channel open time
fluorescent dyes. A HeLa cell pair expressing Cx43 is shown. C, Intracellular pH (pHi)
normalized to the initial pHi (7.4) and junctional conductance (gj) normalized to remains unknown, but it has been suggested that interactions at
the initial gj are plotted against time during 100% CO2. the protein-lipid interface are the most likely site of action.
Another class of agents, long chained alcohols such as octanol
and heptanol, are also effective in reducing junctional conduc-
formation of heteromeric or heterotypic forms of Cx43 and Cx40 tance and are thought to act via protein-lipid interactions. Other
shifts the pKa to a more alkaline value of 7.0.32 The other cardiac agents that are able to reduce functional gap junction numbers
connexins have not been studied as extensively as Cx43 and are carbenoxolone, glycyrrhetinic acid, quinine derivatives, reti-
Cx40, with the exception of Cx46, for which the data are most noic acid, arachidonic acid, and spermine.35 All the aforemen-
consistent with a direct effect on the cytoplasmic surface. Rapid tioned agents lack specificity and inhibit other systems within
delivery of H+ to the cytoplasmic side of Cx46 hemichannels was cells.35 Another approach has been the use of mimetic peptides.
A recent study has shown that GAP26, a polypeptide, is able to one dominates in response to ischemia is most likely disease
15
block hemichannels of Cx43 on a time scale of seconds to minutes, dependent.
and longer exposure is able to effectively reduce Cx43 gap Atrial arrhythmias are also associated with electrical remodel-
junction–mediated coupling.36 The mimetic peptides appear to ing and changes in connexin distribution. Of particular interest
function by binding to the extracellular loops, and when the is Cx40, in which abnormal expression results in an increased
hemichannel is subsequently incorporated into a plaque or tendency toward atrial fibrillation.42
attempts to link via E1 or E2 with its counterpart in the adjacent A number of studies have found strong correlates between
cell, the peptide prevents formation of a functioning channel. mutations in Cx43 and disease state, with the most clearly under-
The apparent turnover rate of approximately 30 minutes is faster stood being oculodentodigital dysplasia (ODDD), in which a
than the turnover rate determined via Western blot analysis,37 but number of mutations in Cx43 have been identified.43 Interest-
might be related connexon docking and gap junction channel ingly, patients with ODDD do not display any cardiac anomalies.
assembly or disassembly associated with the scaffolding complex Mutation in adhesion molecules is another way to affect connexin
protein zonula occludens, ZO-1.38 Despite a lack of mechanistic distribution within the intercalated disc. Naxos disease arises
understanding, these data demonstrate a potentially useful clini- because of a mutation within the adhesion molecule plakoglobin.
cal tool if GAP26 is clearly shown to be gap junction specific. This mutant form does not traffic to the intercalated disc prop-
Direct proof of binding to the extracellular loops is a first neces- erly.44 As a result, Cx43 remodeling occurs with lesser junctional
sary step, and tagged peptides must be used to assess whether the complexes being formed, culminating in an arrhythmogenic
peptide might affect function intracellularly possibly via endo- cardiomyopathy.
somal entry.
A potentially clinically relevant feature of mimetic peptides is
manifest in Gap26, a mimetic peptide for Cx43 that has been
shown to protect against induced myocardial ischemia in vivo.39 Summary
Its potential to slow conduction and create proarrhythmic activity
has not been tested. Overall, extrinsic uncoupling agents have Gap junction channels are poorly selective intercellular channels
proved useful in attempting to understand how gap junction that allow the movement of ions, solutes, second messengers, and
channels affect cardiac action potential propagation but, as might even microRNAs and siRNAs from cell to cell exclusive of the
be expected, a reduction in the number of functioning gap junc- extracellular milieu. Gap junctions participate in the propagation
tion channels results in slowed conduction and consequently the of the cardiac action potential and are expressed differentially to
possibility of generating arrhythmogenic activity.26 affect pacing rate, AV node delay, and rapid conduction in Pur-
kinje fibers, ventricular, and atrial myocardium. The evidence is
overwhelmingly clear that the cardiac connexins are essential to
normal cardiac rhythm and are involved in the response to disease
Ischemia, Mutations, Arrhythmia, processes, such as the phenomenon of lateralization. Mutations
and Gap Junctions in cardiac connexins resulting in or causing cardiac dysfunction
must still be considered correlative rather than causative; they
Ischemia reduces or completely occludes blood flow to the myo- have not yet been unequivocally demonstrated to cause heart
cardium, resulting in hypoxia that subsequently triggers the disease,45 unlike Cx26, where the mutations have been shown to
release of intracellular calcium and acidosis, which can result cause deafness.43 For ischemic disease states, it appears that
in cellular remodeling or cell death. A number of studies have cardiac connexins are best considered as part of the effect rather
demonstrated that ischemia triggers an anatomic remodeling of than the cause. Again, the best example is remodeling associated
gap junctions within myocytes, such that fewer junctions are with ischemia. In one sense, remodeling in the form of increased
found in the intercalated disc and more appear on the lateral abundance of connexin along the lateral surfaces is most logically
surfaces of the myocytes.40 Such redistribution or remodeling is considered as an attempt to circumvent a damaged intercalated
predicted to directly affect longitudinal resistance and reduce disc rather than a precipitating causal event. There is a risk of
conduction velocity along the long axis of the myocyte. To further creating a reentrant arrhythmia with lateralization, but it is a
complicate matters, the insertion of more gap junctions laterally better alternative than significant or complete loss of longitudinal
has the potential to create arrhythmias.40 Studies using a canine conductivity because of a dysfunctional gap junction–mediated
heart failure model found a strong correlation between reduced communication at the intercalated disc.
Cx43 expression in vivo and reentrant ventricular arrhythmias.40
A number of studies have also found the phosphorylation
state of Cx43 to be altered in ischemia and in nonischemic
heart failure.41 Phosphorylation state might also trigger anatomic Acknowledgments
remodeling in response to ischemia or other pathophysiologic
challenges to the heart. Clearly, gap junction channels along The authors thank Dr. Robert Weingart for his guidance and
with many other membrane channels participate in electrical and mentorship. This work was supported by National Institutes of
anatomic remodeling in response to ischemic conditions. Which Health grants RO1 GM 088181 and RO1 GM 088180.
3. Maeda S, Nakagawa S, Suga M, et al: Structure of 6. Prochnow N, Hoffmann S, Dermietzel R, et al:

References the connexin 26 gap junction channel at 3.5 A reso- Replacement of a single cysteine in the fourth
lution. Nature 458(7238):597–602, 2009. transmembrane region of zebrafish pannexin 1
1. Yao JA, Gutstein DE, Liu F, et al: Cell 4. Cheng A, Yeager M: Bootstrap resampling for voxel- alters hemichannel gating behavior. Exp Brain Res
coupling between ventricular myocyte pairs from wise variance analysis of three-dimensional density 199(3-4):255–264, 2009.
connexin43-deficient murine hearts. Circulation maps derived by image analysis of two-dimensional 7. Rackauskas M, Neverauskas V, Skeberdis VA:
Research 93(8):736–743, 2003. crystals. J Struct Biol 158(1):19–32, 2007. Diversity and properties of connexin gap
2. Severs NJ, Bruce AF, Dupont E, et al: Remodelling 5. Kovacs JA, Baker KA, Altenberg GA, et al: Molec- junction channels. Medicina (Kaunas) 46(1):1–12,
of gap junctions and connexin expression in dis- ular modeling and mutagenesis of gap junction 2010.
eased myocardium. Cardiovasc Res 80(1):9–19, channels. Prog Biophys Mol Biol 94(1-2):15–28, 8. Duffy HS, Fort AG, Spray DC: Cardiac connexins:
2008. 2007. genes to nexus. Adv Cardiol 42(1-17), 2006.
9. Jansen JA, van Veen TA, de Bakker JM, et al: thesis following myocardial infarction in rats. Brain 40. Wit AL, Peters NS: The role of gap junctions in
Cardiac connexins and impulse propagation. J Mol Res 1124(1):142–154, 2006. the arrhythmias of ischemia and infarction. Heart
Cell Cardiol 48(1):76–82, 2010. 25. Gommans WM, Berezikov E: Controlling miRNA Rhythm 9(2):308–311, 2012.
10. Bukauskas FF, Verselis VK: Gap junction channel regulation in disease. Methods Mol Biol 822:1-18, 41. Schulz R, Heusch G: Connexin43 and ischemic
gating. Biochim Biophys Acta 1662(1-2):42–60, 2012. preconditioning. Adv Cardiol 42, 213–227, 2006.
2004. 26. Jia Z, Bien H, Shiferaw Y, et al: Cardiac cellular 42. Chaldoupi SM, Loh P, Hauer RN, et al: The role
11. Vogel R, Valiunas V, Weingart R: Subconductance coupling and the spread of early instabilities in of connexin40 in atrial fibrillation. Cardiovasc Res
states of Cx30 gap junction channels: data from intracellular Ca2+. Biophys J 102(6):1294–1302, 84(1):15–23, 2009.
transfected HeLa cells versus data from a mathe- 2012. 43. Pfenniger A, Wohlwend A, Kwak BR: Mutations
matical model. Biophys J 91(6):2337–2348, 2006. 27. Greener ID, Monfredi O, Inada S, et al: Molecular in connexin genes and disease. Eur J Clin Invest
12. Brink PR, Ramanan SV, Christ GJ: Human con- architecture of the human specialised atrioven- 41(1):103–116, 2011.
nexin 43 gap junction channel gating: evidence for tricular conduction axis. J Mol Cell Cardiol 44. Saffitz JE: Arrhythmogenic cardiomyopathy:
mode shifts and/or heterogeneity. Am J Physiol 50(4):642–651, 2011. advances in diagnosis and disease pathogenesis.
271(1 Pt 1):C321–31, 1996. 28. Nof E, Antzelevitch C, Glikson M: The Contribu- Circulation 124(15):e390–2, 2011.
13. Ramanan SV, Valiunas V, Brink PR: Non-stationary tion of HCN4 to normal sinus node function in 45. Makita N, Seki A, Sumitomo N, et al: A con-
fluctuation analysis of macroscopic gap junction humans and animal models. Pacing Clin Electro- nexin40 mutation associated with a malignant
channel records. J Membr Biol 205(2):81–88, physiol 33(1):100–106, 2010. variant of progressive familial heart block type I.
2005. 29. Valiunas V, Kanaporis G, Valiuniene L, et al: Cou- Circ Arrhythm Electrophysiol 5(1):163–172, 2012.
14. Valiunas V, Weingart R, Brink PR: Formation pling an HCN2-expressing cell to a myocyte 46. Valiunas V, Gemel J, Brink PR, et al: Gap junction
of heterotypic gap junction channels by connexins creates a two-cell pacing unit. J Physiol 587(Pt channels formed by coexpressed connexin40 and
40 and 43. Circulation Research 86(2):E42–9, 21):5211–5226, 2009. connexin43. Am J Physiol Heart Circ Physiol
2000. 30. Monastyrskaya K, Tschumi F, Babiychuk EB, et al: 281(4):H1675–H1689, 2001.
15. Desplantez T, Halliday D, Dupont E, et al: Cardiac Annexins sense changes in intracellular pH during 47. Desplantez T, Dupont E, Severs NJ, et al: Gap
connexins Cx43 and Cx45: formation of diverse hypoxia. Biochem J 409(1):65–75, 2008. junction channels and cardiac impulse propagation.
gap junction channels with diverse electrical prop- 31. Vaughan-Jones RD, Spitzer KW, Swietach P: J Membr Biol 218(1-3):13–28, 2007.
erties. Pflugers Arch 448(4):363–375, 2004. Intracellular pH regulation in heart. J Mol Cell 48. Bukauskas FF, Elfgang C, Willecke K, et al: Bio-
16. Bukauskas FF, Jordan K, Bukauskiene A, et al: Cardiol 46(3):318–331, 2009. physical properties of gap junction channels
Clustering of connexin 43-enhanced green fluores- 32. Gu H, Ek-Vitorin JF, Taffet SM, et al: Coexpres- formed by mouse connexin40 in induced pairs
cent protein gap junction channels and functional sion of connexins 40 and 43 enhances the pH sen- of transfected human HeLa cells. Biophys J
coupling in living cells. Pro Natl Acad Sci U S A sitivity of gap junctions: a model for synergistic 68(6):2289–2298, 1995.
97(6):2556–2561, 2000. interactions among connexins. Circ Res 86(10): 49. Beblo DA, Wang HZ, Beyer EC, et al: Unique
17. Kjenseth A, Fykerud T, Rivedal E, et al: Regulation E98–E103, 2000. conductance, gating, and selective permeability
of gap junction intercellular communication by the 33. Xu Q, Kopp RF, Chen Y, et al: Gating of connexin properties of gap junction channels formed by con-
ubiquitin system. Cell Signal 22(9):1267–1273, 43 gap junctions by a cytoplasmic loop calmodulin nexin40. Circ Res 77(4):813–822, 1995.
2010. binding domain. Am J Physiol Cell Physiol 50. Valiunas V: Biophysical properties of connexin-45
18. Kanaporis G, Brink PR, Valiunas V: Gap junction 302(10):C1548–56, 2012. gap junction hemichannels studied in vertebrate
permeability: selectivity for anionic and cationic 34. Decrock E, Vinken M, Bol M, et al: Calcium and cells. J Gen Physiol 119(2):147–164, 2002.
probes. Am J Physiol Cell Physiol 300(3):C600–9, connexin-based intercellular communication, a 51. Elenes S, Martinez AD, Delmar M, et al: Hetero-
2011. deadly catch? Cell Calcium 50(3):310–321, 2011. typic docking of Cx43 and Cx45 connexons blocks
19. Valiunas V, Beyer EC, Brink PR: Cardiac gap junc- 35. Bodendiek SB, Raman G: Connexin modulators fast voltage gating of Cx43. Biophys J 81(3):1406–
tion channels show quantitative differences in and their potential targets under the magnifying 1418, 2001.
selectivity. Circ Res 91(2):104–111, 2002. glass. Curr Med Chem 17(34):4191–4230, 2010. 52. Sakai R, Elfgang C, Vogel R, et al: The electrical
20. Kanaporis G, Mese, G., et al: Gap junction chan- 36. Desplantez T, Verma V, Leybaert L, et al: Gap26, behaviour of rat connexin46 gap junction channels
nels exhibit connexin-specific permeability to a connexin mimetic peptide, inhibits currents expressed in transfected HeLa cells. Pflugers Arch
cyclic nucleotides. J Gen Physiol 2600, 293–305, carried by connexin43 hemichannels and gap 446(6):714–727, 2003.
2008. junction channels. Pharmacol Res 65(5):546–552, 53. Brink PR, Cronin K, Banach K, et al: Evidence for
21. Valiunas V, Polosina YY, Miller H, et al: Connexin- 2012. heteromeric gap junction channels formed from rat
specific cell-to-cell transfer of short interfering 37. Falk MM, Baker SM, Gumpert AM, et al: Gap connexin43 and human connexin37. Am J Physiol
RNA by gap junctions. J Physiol 568(Pt 2):459– junction turnover is achieved by the internalization 273(4 Pt 1):C1386–96, 1997.
468, 2005. of small endocytic double-membrane vesicles. Mol 54. Rackauskas M, Kreuzberg MM, Pranevicius M,
22. Alig J, Marger L, Mesirca P, et al: Control of Biol Cell 20(14):3342–3352, 2009. et al: Gating properties of heterotypic gap junction
heart rate by cAMP sensitivity of HCN channels. 38. Rhett JM, Jourdan J, Gourdie RG: Connexin 43 channels formed of connexins 40, 43, and 45.
Proc Natl Acad Sci U S A 106(29):12189–12194, connexon to gap junction transition is regulated by Biophys J 92(6):1952–1965, 2007.
2009. zonula occludens-1. Mol Biol Cell 22(9):1516– 55. Makowski L, Caspar DL, Phillips WC, et al: Gap
23. Chen C, Nakamura T, Koutalos Y: Cyclic AMP 1528, 2011. junction structures. VI. Variation and conservation
diffusion coefficient in frog olfactory cilia. Biophys 39. Hawat G, Benderdour M, Rousseau G, et al: Con- in connexon conformation and packing. Biophys J
J 76(5):2861–2867, 1999. nexin 43 mimetic peptide Gap26 confers protec- 45(1):208–218, 1984.
24. Hasan W, Jama A, Donohue T, et al: Sympathetic tion to intact heart against myocardial ischemia
hyperinnervation and inflammatory cell NGF syn- injury. Pflugers Arch 460(3):583–592, 2010.
Excitation-Contraction Coupling 16
Donald M. Bers
Ca-induced Ca release to the center of the myocyte (via a chain

CHAPTER OUTLINE
of RyR clusters) or fail to propagate such that the surface release
Excitation-Contraction Coupling and Relationship produces only a small and slow [Ca] elevation near the center of
to Action Potentials 161 the cell.
It is this rise in intracellular [Ca] ([Ca]i) that activates the
Sources and Sinks of Ca in Myocytes: Sarcolemma, myofilaments to contract. Ca binds to troponin C on the thin
Sarcoplasmic Reticulum, Mitochondria 161 filaments, which induces a conformational change that allows the
Balance of Fluxes, E-C Coupling Gain, and Fractional heads of myosin molecules that stick out along the thick filament
Ca Release 164 to bind to actin molecules, which form the body of the thin fila-
ment (see Figure 16-1). The myosin head uses energy stored in
Structure of the Couplon and Submembrane Spaces 164 adenosine triphosphate (ATP) to tilt the head (pulling on the
E-C Coupling: Ca-Induced Ca Release, Ca Sparks, and actin filament), thus creating force or sarcomere shortening,
Ca Waves 165 which is responsible for isovolumic contraction and ejection of
blood from the heart. The synchronization of local Ca transients
Cardiac Alternans 168 throughout the heart is therefore essential for synchronous ven-
Ca Fluxes Can Influence the Cardiac Action Potential 168 tricular contraction. The strength of contraction is directly
related to the [Ca] surrounding the myofilaments. Thus a myocyte
(or region thereof) that has a small Ca transient will be weaker
than an adjacent area and can damp the strength of the stronger
region (i.e., the strong cell can expend its strength, stretching a
Excitation-Contraction Coupling and weak neighbor, and thus not contribute to cardiac output).
Relationship to Action Potentials Readily appreciable consequences of this mechanical dyssyn-
chrony include ischemia ventricular fibrillation and spatially dis-
Cardiac excitation-contraction coupling (ECC) refers to the cordant cardiac alternans.
process by which the electrical activation of cardiac myocytes Another point worth mentioning is that Ca signaling is highly
leads to the activation of contraction.1-3 In its broadest use, ECC local, and in that sense it contrasts with electrical signals. As a
refers to everything from the initial membrane depolarization result, the distribution of Ca entry and SR Ca release must be
through the action potential (AP) and activation of the Ca tran- uniform, even at each region within a single myocyte, to ensure
sient, including how the myofilaments respond to the Ca tran- proper function. In contrast, because of the long electrical space
sient to produce contraction. This chapter will focus on the constant in the heart, there could be potassium channels in only
initiation of normal and abnormal Ca transients in myocytes, Ca every other myocyte, and electrical signaling could be normal.
transport and buffering mechanisms, and the interaction of Ca For cardiac relaxation and ventricular refilling to occur, [Ca]i
signaling with the AP, arrhythmias, and gene regulation. In recent must decline, allowing Ca to dissociate from troponin C and
years, it has become increasingly clear that Ca signaling and terminate myofilament cross-bridge cycling. This [Ca]i decline is
cardiac electrophysiology are inextricably interrelated, making it driven by Ca transport, mainly via the SR Ca–adenosine triphos-
essential to understand myocyte Ca regulation to understand phatase (ATPase) and sarcolemmal Na/Ca exchanger (NCX; dis-
arrhythmogenesis. cussed later). Because Ca movement across the sarcolemma via
Figure 16-1 shows the structure and key mediators of ECC Ca channels and the 3Na:1Ca electrogenic NCX carries a charge,
in myocytes. Ca influx via Ca current (ICa) and Ca release from it affects the AP and excitability. There are other Ca-sensitive
the ryanodine receptor (RyR) in the sarcoplasmic reticulum (SR) channels whose electrophysiologic impact is regulated by the
are central to myofilament activation. Ventricular myocytes have local [Ca]i near those channels. Thus, it is important to integrate
a network of transverse or T-tubules that dive into the cell center, our understanding of Ca handling into our electrophysiologic
perpendicular to the long axis of the myocyte. This T-tubular framework.
system also exhibits longitudinal extension in some myocytes.
The role of the T-tubular system is to synchronize the sarcolem-
mal electrical signal (AP) at junctions throughout the cell, where Sources and Sinks of Ca in
the plasma membrane and SR are close together and mediate
Ca-induced Ca release. In heart failure (HF), there is evidence Myocytes: Sarcolemma, Sarcoplasmic
that the T-tubule network is less extensive and less well orga- Reticulum, Mitochondria
nized,4,5 which can reduce efficacy of the overall ECC process.
Atrial myocytes have fewer T-tubules,5 and specialized con- In reference to the cytosolic compartment in which the myofila-
duction fibers (sinoatrial and atrioventricular nodes and Purkinje ments reside and with which the sarcolemma, SR, mitochondria,
fibers) have almost no T-tubules. In cells or regions thereof that and nucleus interface, the sources of Ca that cause cytosolic
lack T-tubules, ICa initiates SR Ca release only at the surface [Ca] to rise and sink, which remove Ca from the cytosol, can
membrane junctions. Then depending on the conditions dis- be considered. Ca channels (L-type and T-type) and the NCX
cussed later, activation can more slowly propagate as a wave of are the main pathways of Ca entry. However, with the huge
161
Cardiac Myocyte 3Na 2K

Na K Cl
Sarcolemma ATP NCX ATP
3Na
Ca Ca
RyR
Na
Ca
ICa Cyt
SR ATP
PLB
Ca Na H
Mitochondria
Ca
Transverse-Tubule
Transverse-Tubule
Myofilaments Thick Filament Thin Filament
Figure 16-1. Schematic of Ca handling systems in cardiac myocytes. Ca2+, Na+, K+, and Cl− channels are shown, as are sarcolemmal and SR ATPase pumps (ATP). The SR
Ca-ATPase is inhibited by phospholamban (PLB), whose inhibitory effect is reversed by phosphorylation by protein kinase A or Ca-calmodulin–dependent protein kinase.
The mitochondrial cytochrome system (Cyt) responsible for pumping protons (H) out of mitochondria is indicated.
electrochemical driving force for Ca influx caused by a 20,000- Although Ca entry via the channel itself can contribute to CDI,
fold concentration gradient (2 mM outside: diastolic [Ca]i = most LTCCs in cardiac myocytes are localized at junctions with
100 nM) and −80 mV membrane potential (Em), even a slight Ca the SR and RyR, and the Ca-induced Ca release is even more
permeability other than these known transporters would consti- powerful in causing CDI. The integrated amount of Ca influx via
tute an additional non-negligible leak pathway. During diastole, ICa in ventricular myocytes during a normal AP is 5 to 10 µmol/L
the Ca channels are tightly closed and NCX (which is reversible) cytosol but is about twice as large if SR Ca release is blocked. In
functions predominantly in the Ca extrusion mode (inward context, in this Ca flux the amount of SR Ca release in mam-
current). malian ventricular myocytes is normally three to ten times larger
During the AP, voltage-dependent Ca channels open and than this, depending on species and conditions. The regulation
allow Ca entry down its electrochemical gradient, causing an of SR Ca release will be discussed in more detail.
inward depolarizing current (see Chapter 2 for Ca-channel gating NCX can also contribute to the rise in [Ca]i, but Ca entry via
details). In ventricular myocytes, the ICa is virtually all mediated NCX is normally small (<1 µmol/L cytosol) and constrained
by L-type Ca channels (LTCC), in particular by the Cav1.2 largely to the very early part of the AP. NCX can bring Ca in or
isoform. Atrial, and especially pacemaker, cells also express the out (outward or inward current) during the AP, depending on the
Cav1.3 LTCC isoform, which activates at a more negative Em, trans-sarcolemmal [Na] and [Ca] gradients and Em (Figure 16-2).
and thus can better contribute to phase 4 depolarization in pace- During diastole, the Em is negative to the electrophysiologic
makers and recruit additional Cav1.2 channels. reverse potential for NCX (ENCX = 3ENa − 2ECa), so Ca efflux is
T-type Ca channels are also seen in atrial, pacemaker, and thermodynamically favored (see Figure 16-2, A). However, the
conduction fibers, but not in ventricular myocytes (except in low diastolic [Ca]i kinetically limits the amount of transport (low
neonatal myocytes and in some pathophysiologic settings). These substrate concentration). The peak of the AP exceeds ENCX and
channels activate at even more negative Em but also inactivate briefly favors Ca influx (an outward current). However, as soon
much more rapidly than LTCCs. They can contribute to trig- as ICa and SR Ca release begin, the local [Ca]i in the cleft and
gered or pacemaker activity, but even when they are present they submembrane space ([Ca]sm) is much higher than [Ca]i and drives
provide less peak ICa and integrated Ca influx than LTCC and Ca extrusion via NCX (see Figure 16-2, B and C). The declining
cannot effectively substitute for LTCC in triggering SR Ca Em during repolarization also more strongly favors Ca efflux, and
release. The latter is probably because they do not target to the that persists throughout [Ca]i decline and diastole (see Figure
sarcolemmal-SR junctions, where Ca-induced Ca release occurs 16-2, A). Thus, NCX is mainly considered a Ca efflux mecha-
during ECC. nism. An exception to this is certain pathophysiologic conditions
ICa is activated by depolarization and exhibits Em- and like HF, in which [Na]i is elevated, NCX expression may be
Ca-dependent inactivation (VDI and CDI), and CDI is by far the elevated, Ca transients are small, and APD is prolonged (see
dominant physiologic mode of inactivation.6 When CDI is abro- Figure 16-2, C). All these factors shift NCX more in favor of Ca
gated experimentally, AP duration (APD) becomes extremely influx, and thus in HF, Ca entry via NCX can persist for most of
long and Ca tends to overload in cells. There appears to be the AP plateau and significantly contribute to the Ca transient.
calmodulin (CaM) constitutively bound to the Cav1.2 in myo- In a sense, this limits the extent of cardiac dysfunction in HF by
cytes and it senses local [Ca]i elevation and induces inactivation.7 bringing Ca in and by indirectly helping load the SR with Ca.
40 Ca influx 50
16
Em (mV)
Em or ENCX (mV) 0 HF
ENCX
0
–50 Ctl ↑APD
[Ca]sm Em
2.0
–40 Ctl
[Ca] (µM)
Ca efflux 1.5
1.0
[Ca]i
–80 0.5 [Ca]sm
HF
500 nM
0 100 200 300 400
time (ms)
A
2
Inward INa/Ca Ca Outward INa/Ca
HF
3Na Ca Na
INCX (A/F)
1 [Na]i = 11.3 mM
↑[Na]i & NCX
NCX NCX expression
0
Cacleft Ctl
–1 [Na]i = 8.5 mM
Casm 3Nasm
Ca flux (µM) ∫ INCX

4
HF
0
–4
Ca
–8 Ctl
–12
–16
0 200 400 600
Time (ms)
B Cai Nai C
Figure 16-2. Na/Ca exchange (NCX) function in cardiac myocytes. A, Normal rabbit ventricular myocyte AP (black), global and submembrane Ca transients ([Ca]i and [Ca]sm,
blue), and the reversal potential for NCX (red) based on [Ca]sm and with [Na]i −8.5 mM. Ca influx is favored only early in the AP (where Em is higher than ENCX). B, During SR
Ca release, cleft and submembrane [Ca] (Cacleft and Casm) are higher than bulk cytosolic [Ca]i. This influences whether INCX is inward or outward. C, INCX during the rabbit
ventricular myocyte during the AP in control and HF conditions. [Ca]sm and the integrated Ca flux via NCX are also shown.
(Redrawn from Despa S, Islam M, Weber CR, et al. Intracellular Na+ concentration is elevated in heart failure, but Na/K pump function is unchanged. Circulation 105:2543–2548,
2002.)
This raises the question: How much Ca enters the cytosol rabbit (and similarly in human, canine, feline, ferret, and guinea
during the AP and where does it go? The sum of Ca coming from pig), the percent contribution to [Ca]i decline is roughly 70%,
ICa, NCX, and SR Ca release is between 50 and 100 µmol/L 28%, 1%, and 1%, respectively. Thus, the SR Ca-ATPase is the
cytosol. Notably, this is the total amount of Ca entering the dominant transporter by a factor of 3 over the NCX. In rat and
myoplasm that contains many Ca buffers (e.g., troponin C, the mouse ventricular myocytes, these numbers are 92%, 7%, 0.5%,
SR Ca-ATPase, CaM, membrane lipids, ATP). Quantitatively, and 0.5%, so the system is much more SR-dominated (14-fold
the most important Ca buffers are troponin C and the SR more than NCX). In HF, there is generally a downregulation of
Ca-ATPase, which are present at approximately 70 and 50 µmol/L SR Ca-ATPase and upregulation of NCX, such that in rabbit and
cytosol, respectively, and have affinity in the 0.5-µM range. human HF, the SR and sarcolemmal fluxes become more equally
Overall, the cytosolic buffering is approximately 100 : 1, such that balanced. There are fewer data available in atrial myocytes, but
50 to 100 µmol/L cytosol of added Ca only raises the free [Ca]i for human atrial myocytes, we estimate 66% and 25% of Ca
from 100 nM to approximately 500 to 1000 nM. However, that removal occur via SR Ca-ATPase and NCX, respectively. This
is sufficient to partially saturate troponin C and activate contrac- predominance of SR Ca-ATPase over NCX in sinus rhythm is
tion. Notably, Ca buffering does not directly remove Ca from the converted to an equal contribution of these Ca-removal processes
cytosol, and energy-dependent Ca transport is required to return (46% vs. 44%) in chronic atrial fibrillation.8,9
[Ca]i and cytosolic buffers back to their diastolic levels. In mitochondria, there is uptake of Ca during the normal
There are four transporters that work in parallel to bring [Ca]i heartbeat via the mitochondrial Ca uniporter,10 but it can be
down and drive relaxation: (1) the SR Ca-ATPase, (2) NCX, (3) inferred that this is only about 0.5 µmol/L cytosol (~1 µmol/L
the plasma membrane Ca-ATPase, and (4) mitochondrial Ca mitochondria). Significant Ca buffering is expected in this com-
uptake.1,2 We have analyzed quantitatively how these processes partment, which is packed with proteins (as in the cytosol), so the
compete in myocytes from different mammalian species. In rise in free [Ca] in the mitochondria at each beat is rather small.11
The extrusion of Ca from mitochondria is mediated mainly by a flux balance principle and analysis of Ca removal, the expected
mitochondrial Na/Ca exchanger (NCLX),12 which is different ECC gain should also be species-dependent (~3 in rabbit or
from the sarcolemmal NCX. This Ca extrusion is very slow, such human and ~15 in rat or mouse). The precise way that gain is
that at higher heart rates, the diastolic mitochondrial [Ca] gradu- calculated varies, because sometimes the numerator is peak [Ca]i
ally accumulates to a higher steady-state level. or Δ[Ca]i and sometimes the denominator is peak ICa or integral
One important effect of increased mitochondrial [Ca] is that over the first 20 ms. ECC gain is highest at negative Em and
it binds to and stimulates several dehydrogenase enzymes in the becomes progressively smaller at positive Em. The high gain at
mitochondria to increase the production of NADH and conse- negative Em is the result of only a few LTCCs being opened
quently increases ATP production.3 This is thought to be an (meaning that they are mostly at different individual junctions),
important feedback pathway, where the mitochondria increase and those that do open have very high flux (caused by high driving
their ATP production under conditions in which the myocyte is force).14 These openings at negative Em are highly efficacious in
using more ATP (i.e., when Ca transients are more frequent and/ triggering SR Ca release. At somewhat more positive Em, more
or of higher amplitude). When heart rate is reduced, the diastolic LTCCs are recruited, so some will be at the same junctions as
Ca efflux has a better opportunity to keep up with the less fre- other open LTCCs, and these would increase the denominator,
quent influx pulses, and mitochondrial [Ca] gradually falls. thereby reducing the calculated gain (redundant ICa at some junc-
Under severe cellular stress, mitochondria can store large tions). At positive Em, the single-channel conductance is progres-
amounts of Ca, in part by progressively increasing their buffering sively smaller, and more than one LTCC opening at a junction
power.13 This is thought to provide temporary protection in the may be required for successful RyR activation. Thus, gain plateaus
myocyte from the dangers of Ca overload. However, too much at positive Em and does not decline as rapidly as at negative Em.
Ca in mitochondria can lead to opening of the permeability Fractional SR Ca release is also a powerful measure of ECC
transition pore and loss of proteins, which can lead to mitochon- efficacy that can be assessed independent of ICa or SR Ca load,
drial demise. both of which are major influences on ECC efficacy. It is the
amount of SR Ca release divided by the SR Ca content (typically
assessed by rapid application of caffeine). This approach demon-
strated that at moderate intra-SR [Ca] ([Ca]SR, approximately
Balance of Fluxes, Excitation-Contraction 40% of the normal load), a normal ICa trigger could not induce
Coupling Gain, and Fractional Ca Release SR Ca release and that as [Ca]SR increases, the fractional SR
Ca release increases nonlinearly (Figure 16-3, A).15 This
During a steady state beat, by definition, the amount of Ca that steep relationship is thought to reflect the ability of high [Ca]SR
enters the cell must equal the exact amount that leaves the cell; to sensitize the RyR to local [Ca]i triggers and agrees with single-
otherwise, the cell would progressively gain or lose Ca. The same channel RyR recordings in lipid bilayers. The failure of ECC
is true for the Ca fluxes across the SR and across mitochondria at moderate [Ca]SR levels is also consistent with this and empha-
at each beat. This means that the sum of the sarcolemmal Ca sizes the critical importance of [Ca]SR in regulating the ECC
entry via ICa, NCX, and leak must equal that removed by the process.
NCX and plasma membrane Ca-ATPase. Likewise, the amount
of Ca released by the SR must be reaccumulated by the SR
Ca-ATPase. This fundamental principle aids in quantitative esti-
mates of overall Ca fluxes because different measurements have Structure of the Couplon and
different limitations. Submembrane Spaces
ECC gain is often used as an index of ECC efficacy and is
ideally defined as the ratio of the amount of SR Ca release to the The initiation of SR Ca release during ECC is a local phenom-
amount of Ca influx that triggers this release. Therefore, using the enon. The close physical proximity of the L-type Ca channel to
80 Diastolic
EC Coupling 40
Fractional SR Ca Release (%)
SR Ca Leak
∆SR Ca (µmol/l cytosol)
60 Ctrl
30
SR Ca Leak
40 20 HF
Constant Fractional
Release
10
20
0
0
0 20 40 60 80 100 120 0 30 60 90 120 150
A SR Ca load (µmol/L cytosol) B SR Ca load (µmol/L cytosol)
Figure 16-3. Influence of SR Ca load on E-C coupling efficacy and diastolic SR Ca leak. A, Voltage-clamp in rabbit ventricular myocyte, where SR Ca load was varied by
prepulse protocols and caffeine-induced SR Ca depletion. ICa amplitude at each test pulse shown was the same for each cell included, regardless of SR Ca load. The dashed
line shows the expectation if fractional SR Ca release was not regulated by [Ca]SR. B, SR Ca load was similarly manipulated and 1 mM tetracaine was applied to block RyR-
mediated leak (in Na-free, Ca-free solution). At a given SR Ca load the leak is higher in HF myocytes.
(A, Data from Shannon TR, Ginsburg KS, Bers DM: Potentiation of fractional sarcoplasmic reticulum calcium release by total and free intra-sarcoplasmic reticulum calcium con-
centration. Biophys J 78:334–343, 2000. B, Data from Shannon TR, Pogwizd SM, Bers DM: Elevated sarcoplasmic reticulum Ca2+ leak in intact ventricular myocytes from rabbits
in heart failure. Circ Res 93:592–594, 2003.)
the RyR at the junction or couplon is critical for the efficacy of surface membrane, but most of the NCX molecules are not found
16
Ca-induced Ca release (Figure 16-4). Skeletal muscle expresses in the junctions.17 It is unclear what protein or signal dictates the
different isoforms of these two channels (Cav1.1 and RyR1) localization of RyR2 and Cav1.2 to the junction, but junctophilin
versus those in cardiac myocytes (Cav1.2 and RyR2), and the is a candidate protein that interacts with both membranes and
ECC process is more highly evolved. That is, in skeletal muscle, RyR2.18 Notably, there are clusters of RyR2 organized in trans-
Cav1.1 physically communicates with RyR1 such that the verse arrays in atrial myocytes, even in the absence of T-tubules.
depolarization-induced change in Ca-channel structure is suffi- Compelling electrophysiologic data show that the local sub-
cient to activate RyR1, even without any Ca influx through the membrane [Ca] sensed by sarcolemmal ion channels and trans-
channel. The structure of these skeletal muscle junctions is rela- porters (even those outside the junctional cleft) differs significantly
tively rigid, where every other RyR1 tetrameric channel (in a corner-to- from that in the general cytoplasm and that sensed by myofila-
corner array) has four L-type Ca channels directly above and ments and cytosolic fluorescent indicators.19 This means that
communicating with it.16 In the heart, there are similar structured Ca-sensitive ionic currents will sense larger and faster rising Ca
arrays of RyR2 tetramers (~100 RyR per junction), and Cav1.2 transients than those sensed by globally situated Ca indicators
channels are clustered in these junctions as well but appear to be (see Figure 16-2; discussed later).
distributed randomly within the cluster with one Cav1.2 per 4 to
10 RyR tetramers (Figure 16-4). The majority of these couplons
(diameter ~200 nm; depth 15 nm between SR and sarcolemma)
are in the T-tubule, but about 25% are found along the external E-C Coupling: Ca-Induced Ca Release, Ca
sarcolemma, and both function similarly with respect to ECC.1 Sparks, and Ca Waves
In rabbit ventricular myocyte, these junctions occupy approxi-
mately 21% of the T-tubule membrane and 5% of the surface ECC requires Ca influx via the Cav1.2 channel (or a surrogate
sarcolemma. Both numbers are slightly higher in rat, but the pathway), which raises cleft [Ca]i, to activate the RyR opening
couplons occupy 11% to 20% of the total sarcolemma. NCX and and SR Ca release. If extracellular Ca is removed or ICa is inhib-
the Na/K-ATPase are more concentrated in T-tubule than in ited, there is an immediate (<1 second) abolition of cardiac Ca
transients (an effect not seen in skeletal muscle). Under physio-
logic conditions, there are probably enough L-type Ca channels
in each junction (10-15) to ensure that at least one will open
during AP, which is sufficient to raise local [Ca]i enough to acti-
vate at least one of the approximately 100 RyR in the cleft. Once
RyR in SR a single RyR channel opens, it can recruit more RyR within the
(1 channel per tetramer) cleft, and it is believed that about six to 15 individual RyR open-
SR
ings may constitute the normal local release event. Within the
tiny cleft space, local [Ca] probably rises to a peak higher than
TT 100 µM. However, as Ca diffuses away, the local [Ca]i at sites
outside the cleft is much lower and declines as one moves farther
away, because of three-dimensional diffusion (peak [Ca]i ~500-
Nonjunctional 1000 nM). Throughout myocytes and the heart, these local
release events are synchronized by the AP and nearly simultane-
ous activation of ICa at each junction. This synchronization is
critical for functionally effective ECC.
Spontaneous SR Ca release events (Ca sparks) can also occur
SR
LTCC in TT
in the absence of Ca current (Figure 16-5, A and B). This is
because RyR opening is stochastic and influenced by [Ca]i, [Ca]SR,
and RyR modulation (e.g., by phosphorylation, oxidation, or
Non-junctional disease-related mutations, as seen in RyR2 or calsequestrin with
A catecholaminergic polymorphic ventricular tachycardia). Even if
the normal open probability is 10−4 s−1 and there are 1 million
Ca RyR per cell, it could be expected that there are approximately
LTCC 100 RyR openings per second in each cell. Each one of these can
+
participate in the same local positive feedback of Ca-induced Ca
Sarcolemma release and cause the firing of a single junction, similar to that
+
initiated by the ICa trigger. These stochastic Ca sparks are thought
CICR of as the fundamental unit of SR Ca release, and the normal ECC
Ca transient is the spatiotemporal summation of approximately
RyR 15,000 of these local Ca-release events, with one happening at
each junction.
Under normal conditions, the SR Ca release at one junction
does not trigger the release of Ca from a neighboring junction
SR that is either 2 µm away longitudinally (at the next Z-line) or
about 0.5 µm transversely. That is, we normally see spatially
discrete Ca sparks. This is because the [Ca]i at that distance away
B Ca Ca Ca is too low to activate the next junction because it requires the
higher local [Ca]i near the mouth of a Ca channel. However,
Figure 16-4. Junctional structure in which T-tubular membrane apposes the SR
membrane. A, Side view of a junction along the T-tubule (top) and view through
there are conditions in which this tight local control is lost and
the T-tubule membrane looking at the LTCCs (red) and array of RyR tetramers (yellow Ca waves can propagate and become arrhythmogenic.
and orange). Some nonjunctional SR regions also have RyR clusters and can con- What conditions favor Ca waves? The same conditions that
tribute to Ca release. B, A LTCC when opened allows Ca influx, which can activate increase RyR excitability and Ca spark frequency tend to increase
SR Ca release via RyR, and release from one RyR can activate neighboring RyRs. the ability of Ca sparks to propagate as Ca waves (see Figure 16-5,
Ca sparks Ca wave
10 µm
A 1s
Ca sparks
Ca waves
430 nM 1000 nM
[Ca]i
µm
20
100 nM
20 µm
200 ms 200 ms
5µ
m
B s C D
100 m
Figure 16-5. Ca sparks and waves in cardiac myocytes. A, A linescan confocal image showing numerous Ca sparks in a mouse ventricular myocyte and a propagating Ca
wave. B-D, Representations of Ca sparks and waves as a tilted 2-D and 3-D plot color-coded for local [Ca]i.
(Images obtained from Y. Yang and Y. Li.)
A, C, and D). The simplest and best known case is that of cellular Instead, the normal mode of ECC in these cells includes propa-
and SR Ca overload. Because RyR opening is favored by both gated Ca-induced Ca release (Ca waves) to activate the center of
high local cleft [Ca] and elevated [Ca]SR, increased Ca spark the cell. It is not known whether this is because atrial myocytes
frequency when Ca loading is favored (e.g., Na/K-ATPase inhibi- tend to be more highly Ca-loaded under baseline conditions (vs.
tion, AP prolongation, β-adrenergic activation, or increased heart ventricular myocytes) or whether there is a structural or regula-
rate). In these cases, there is also greater SR Ca release during tory difference in RyRs. However, this might predispose atrial
each Ca spark (owing to higher [Ca]SR and driving force), and this myocytes to be more on the verge of producing potentially
can superimpose on an elevated [Ca]i background. These factors arrhythmogenic Ca waves.
will increase the rise in [Ca]i that a Ca spark causes at the neigh- How does SR Ca release terminate? Because Ca-induced Ca
boring junction. Furthermore, the next junction has a higher release is an inherently positive feedback system, it is important
sensitivity to [Ca]i because of the higher [Ca]SR, and these effects to consider how SR Ca release terminates. Because release nor-
synergize to increase the likelihood that a Ca spark becomes a mally terminates when [Ca]SR levels are only about 50% depleted
Ca wave. As Ca loading progresses, there can often be small, (during either global ECC events or Ca sparks), there must be
abortive miniwaves before observing full-blown Ca waves that something that breaks the positive feedback. There is compelling
traverse the whole myocyte. It is these cell-wide Ca waves that evidence that cardiac RyR gating is influenced strongly by luminal
can activate sufficient inward NCX current that can cause trig- [Ca]SR such that RyR opening is favored at high [Ca]SR (enhancing
gered arrhythmias. Ca spark initiation) and RyR closing is favored at lower [Ca]SR.
SR Ca leak during diastole is composed of leak that is detect- Experiments show that release terminates at about 400 µM [Ca]SR
able as Ca sparks plus Ca leak that is not detectable as Ca sparks.20 during both Ca sparks and ECC, and when there is little spatial
Thus, although Ca sparks can provide a convenient readout con- gradient of [Ca]SR.23 Thus, there seems to be an internal brake
cerning SR Ca leak, the Shannon-Bers tetracaine block protocol that prevent this positive feedback from completion. This occurs
assesses all RyR-dependent leak and was the first to demonstrate at a similar level of [Ca]SR to that at which ECC and Ca sparks
that SR Ca leak was increased in HF (Figure 16-6, B).21,22 There fail as [Ca]SR decreases (see Figure 16-3), and these are probably
appears to be a small but significant component of SR Ca leak functionally related effects.
that is not mediated by either the RyR or the related InsP3 recep- Why does a specific junction fail to activate? Several factors
tor (InsP3R).20 could be responsible. First, the depolarization could fail to acti-
Atrial myocytes that lack T-tubules cannot use ICa to synchro- vate sufficient entry via ICa. This could happen if the Ca channels
nize release at internal RyR clusters that exist along the Z-lines. are refractory or blocked, but it can also happen at very positive
2.2 Hz
16
40
(mV)
0
Em
-40
-80
6
Big
Shortening
4 Small
(%) 2
0
0.5 s 0.25 s
[Ca]SR
(A.U.)
60
20
A
2.0 Hz Unaltered diastolic [Ca]SR
2300
2200
[Ca]SR (A.U.)
2100
2000
B 1900 1s
2.2 Hz
1.5 Hz
3500
[Ca]SR (A.U.)
3000
2500 1s
C
Figure 16-6. Cardiac alternans in rabbit ventricular myocytes. A, Em, myocyte shortening and intra SR free [Ca] ([Ca]SR) assessed using fluo-5N trapped inside the SR. The
large and small beats are superimposed at right. B, Frequently, alternans are observed without any change in diastolic [Ca]SR between large and small beats. C, At 1.5-Hz
stimulus frequency, this cell did not demonstrate alternans, but as [Ca]SR rose further at 2.2 Hz, the cell exhibited stable alternans.
(Data redrawn from Picht E, DeSantiago J, Blatter LA, et al: Cardiac alternans do not rely on diastolic sarcoplasmic reticulum calcium content fluctuations. Circ Res 99:740–748,
2006.)
potentials relevant to the AP overshoot. At positive potentials, both a lower cleft [Ca] that lasts a shorter time and neighboring
the unitary current through Ca channels is reduced because of RyR that are also less sensitive (i.e., more than the normal trigger
driving force, and thus more Ca channels in the junction may is required to attain amplification). Third, the RyR can also be
need to open to raise cleft [Ca] sufficiently to ensure RyR activa- refractory. Although RyR refractoriness is not exactly like Na
tion. Notably, the early repolarization often seen (phase 1) of the channel refractoriness, it is clear that after a SR Ca release event,
cardiac AP serves to rapidly increase the driving force for Ca some time is required (0.5-10 sec) for the RyR to regain its
entry via activated LTCCs.24 Second, if the [Ca]SR is low, the cleft maximal sensitivity to cleft [Ca]. Note that this is different from
[Ca] required to activate a first RyR is higher, so the same ICa and additional to the time it takes for the SR Ca-ATPase to refill
might fail to initiate RyR activation.25 Indeed, the dependence of the SR between beats. This RyR refractoriness may be an impor-
Ca spark frequency and diastolic Ca leak on [Ca]SR is similar to tant intrinsic protection from arrhythmogenic Ca waves, although
that for fractional SR Ca release during ECC (see Figure 16-3). it may be involved in the development of cardiac alternans.
Even if a first RyR opening occurs, several factors at low [Ca]SR Notably, this RyR refractoriness is not absolute, and recovery can
limit the likelihood that it will recruit other RyR in the couplon. be hastened at increased Ca loads, allowing an abrupt appearance
That is, if the stochastic closure of both the ICa and first open of Ca waves at a certain threshold of SR Ca load. Fourth, physical
RyR occurs before another RyR is recruited, a full release or disruption of the couplon can reduce the efficacy of ICa to activate
spark will be aborted. Several factors contribute to this. At low the RyR. This has been suggested to occur when junctophilin
[Ca]SR, the reduced driving force and shorter open time create 2 is knocked down26 and also in HF, in which T-tubular
organization is reduced27; in these cases, the efficacy of local ECC (ICl(Ca)), Ca-activated K+ current (IK(Ca)), and the delayed rectifier
could be compromised. In the most extreme locations, it would (IKs) and these can influence AP configuration as well. ICl(Ca) is
become rather like the situation in non–T-tubular regions of more prominent at positive Em and contributes to the early repo-
atrial myocytes, where the local ICa cannot trigger release, and larization phase of the AP, where submembrane [Ca] is especially
release would be triggered mainly by propagated Ca-induced Ca high. IK(Ca) is very small or nonexistent in ventricular myocytes
release. but is more prominent in atrial myocytes and again most likely
participates in early repolarization. IKs is known to increase with
higher [Ca]i, but the kinetics of this effect and how it interfaces
with the β-adrenergic effects on IKs are not well known. All of
Cardiac Alternans these currents will be sensitive to changes in Ca transients,
including the large increases in Ca transient associated with
Cardiac alternans was first observed in pulse pressure, but physiologic β-adrenergic activation.
Ca-transient amplitude and APD alternans also occur and have
been recognized as a precursor for ventricular tachycardia and
ventricular fibrillation. Increasing heart rate induces Ca, contrac- Myocyte Ca Handling in Rhythmicity and
tile, and APD alternans hand in hand (see Figure 16-6, A). Arrhythmogenesis
Because Ca alternans persists under a voltage clamp where no Em
change occurs, and changes in Ca-transient amplitude influence The brief focus here is on the role of Ca in triggered activity,
currents in ways that explain APD alternans, it is generally especially early and delayed afterdepolarizations (EADs and
accepted that Ca alternans causes APD alternans.28 In alternans, DADs; Figure 16-7). For DADs, the explanation is fairly simple
the larger Ca transient is usually associated with the longer APD and an extension of properties described before. If after AP repo-
(concordant Ca-Em alternans; see Figure 16-6, A), but it can larization the conditions favor Ca sparks and waves (i.e., Ca
occur in opposite fashion (discordant Ca-Em alternans). Although overload and/or enhanced RyR sensitivity), more of these events
there are several Ca-dependent ionic currents (discussed later), will occur and generate a Ca-activated transient inward current
AP alternans can be adequately explained by effects on ICa and and DAD (see Figure 16-7, A), which is carried almost exclusively
NCX. During the large beat, there will be faster ICa inactivation via NCX current (not ICl(Ca) or nonselective cation current).
because of more robust CDI, and that should shorten APD. Notably, the negative diastolic Em favors inward NCX. If the
However, the larger Ca transient will also drive greater inward inward NCX current is sufficient to bring the myocyte to the
NCX current, which would prolong APD. Thus, in concordant threshold, then an AP can be triggered.
Ca-Em alternans, it seems as if the NCX effect is somewhat more EADs that occur before repolarization is complete are slightly
dominant. Conversely, the ICa effect might predominate for the more complicated because there are two possible mechanisms.
more rare discordant alternans. Ca alternans can also exist out of One mechanism is fundamentally the same as that for DADs (i.e.,
phase within a single myocyte, reemphasizing the idea that Ca spontaneous SR Ca release and inward NCX current; see Figure
signals are very local compared with Em signals. At the whole- 16-7, B). EADs are typically observed under conditions in which
heart level, alternans can also be spatially concordant (all regions APD is long (e.g., IKr block or long-QT syndromes) and this
in phase with each other) or discordant, and the latter is more tends to load cells with Ca because of the prolonged Ca entry
likely to degenerate into ventricular tachycardia or ventricular and reduced diastolic interval for Ca efflux. However, the depo-
fibrillation. larized Em reduces inward INCX at any given [Ca]i, making this
Why do Ca transients alternate during alternans? Possibilities more impactful as repolarization proceeds than at positive plateau
include refractoriness of ICa, changes in SR Ca load, and refrac- potentials. The second, and historically precedent, mechanism is
toriness of SR Ca release. As heart rate increases and alternans that some LTCCs can become re-available during long APD and
occurs, it has been shown that ICa is not different between the reactivate, creating an inward current surge and net depolariza-
large and small beats.29 Alternans also occurs even when diastolic tion (see Figure, 16-7, C). There is good experimental and theo-
[Ca]SR (and SR Ca load) is the same at the beginning of the large retical evidence for both of these mechanisms,6 and they are not
and small beats (see Fig 16-6, B), and it appears that the initial mutually exclusive. For example, an SR Ca release event may
event is caused by refractoriness of the RyR at the small beat.29 produce only a very small plateau INCX and depolarization, which
This creates a small beat during which some RyR junctions fail may be amplified by recruiting ICa to cause a larger EAD.
to fire and thus can recover by the next and larger beat. However, Sinoatrial and atrioventricular node myocytes show spontane-
the large-small Ca transients can also influence SR Ca load. ous Ca transients and APs and are discussed extensively in
Because during the small beat, there is less ICa inactivation and Chapter 25. The SR Ca release and INCX mechanism described
less Ca extrusion via NCX, there will tend to be a net gain of cell before for triggered DAD activity in atrial and ventricular myo-
and SR Ca, and the opposite will occur during the large beat. cytes is an important part of normal pacemaker activity in the
This may be the reason alternations in SR Ca load often accom- sinoatrial node. These cells likely have a relatively Ca-loaded
pany alternans (see Figure, 16-6, C). The direction is such that physiologic state, ancillary ICa that activates at more negative Em,
such changes will tend to amplify the extent of Ca alternans. lower stabilizing IK1 current, and a more favorable source-sink
relationship compared with ventricular myocytes.
In conclusion, cardiac electrophysiology and Ca handling are
inextricably linked functionally, and although Em influences [Ca]i
Ca Fluxes Can Influence the Cardiac and contraction, changes in [Ca]i and Ca transporters function
Action Potential feedback and alter Em in fundamentally important ways. These
can alter AP morphology and excitability and be important in
We have already discussed how Ca carries ionic current (ICa, pacemaking activity and arrhythmias. The Ca regulation dis-
NCX) and that has predictable consequences on the AP configu- cussed in this chapter is also under intense regulatory control by
ration (reduced ICa inactivation such as in Timothy syndrome can numerous signaling pathways in normal and diseased hearts. This
cause APD prolongation). However, there are several other chan- includes, but is not limited to, sympathetic α- and β-adrenergic
nels whose gating is influenced by [Ca]i or, more specifically, pathways and also regulation by calmodulin and Ca-calmodulin–
submembrane [Ca]. These include Ca-activated Cl− current dependent protein kinase II.
DAD EAD EAD
16
Em
ICaL
ICaL react
SR Ca release
INaCa
Figure 16-7. Delayed (DAD) and early afterdepolarizations (EAD) and related ICa and INCX. During DAD, SR Ca release activates INCX that in one case (solid trace) is sufficient
to trigger AP and the other (dashed trace) is not. The purple EAD shows a situation in which the EAD is initiated by SR Ca release and INCX, whereas the red EAD is initiated
by reactivation of L-type Ca current.
is the mitochondrial calcium uniporter. Nature reticulum Ca2+ leak in normal and failing rabbit
References 476(7360):336–340, 2011. ventricular myocytes. J Physiol 588(Pt 23):4743–
11. Andrienko TN, Picht E, Bers DM: Mitochondrial 4757, 2010.
1. Bers DM: Excitation-contraction coupling and free calcium regulation during sarcoplasmic reticu- 21. Shannon TR, Ginsburg KS, Bers DM: Quantita-
cardiac contractile force, ed 2, Dordrecht, the lum calcium release in rat cardiac myocytes. J Mol tive assessment of the SR Ca2+ leak-load relation-
Netherlands, 2001, Kluwer Academic Press, pp Cell Cardiol 46(6):1027–1036, 2009. ship. Circ Res 91(7):594–600, 2002.
1427. 12. Palty R, Sekler I: The mitochondrial Na+/Ca2+ 22. Shannon TR, Pogwizd SM, Bers DM: Elevated
2. Bers DM: Cardiac excitation-contraction coupling. exchanger. Cell Calcium 52(1):9–15, 2012. sarcoplasmic reticulum Ca2+ leak in intact ventricu-
Nature 415(6868):198–205, 2002. 13. Wei AC, Liu T, Winslow RL, et al: Dynamics of lar myocytes from rabbits in heart failure. Circ Res
3. Bers DM: Calcium cycling and signaling in cardiac matrix-free Ca2+ in cardiac mitochondria: two com- 93(7):592–594, 2003.
myocytes. Annu Rev Physiol 70:23–49, 2008. ponents of Ca2+ uptake and role of phosphate buff- 23. Picht E, Zima AV, Shannon TR, et al: Dynamic
4. Lyon AR, MacLeod KT, Zhang Y, et al: Loss of ering. J Gen Physiol 139(6):465–478, 2012. calcium movement inside cardiac sarcoplasmic
T-tubules and other changes to surface topography 14. Altamirano J, Bers DM: Voltage dependence of reticulum during release. Circ Res 108(7):847–856,
in ventricular myocytes from failing human and rat cardiac excitation-contraction coupling: unitary 2011.
heart. Proc Natl Acad Sci U S A 106(16):6854– Ca2+ current amplitude and open channel probabil- 24. Sah R, Ramirez RJ, Oudit GY, et al: Regulation of
6859, 2009. ity. Circ Res 101(6):590–597, 2007. cardiac excitation-contraction coupling by action
5. Dibb KM, Clarke JD, Horn MA, et al: Character- 15. Shannon TR, Ginsburg KS, Bers DM: Potentia- potential repolarization: role of the transient
ization of an extensive transverse tubular network tion of fractional sarcoplasmic reticulum calcium outward potassium current (Ito). J Physiol 546(Pt
in sheep atrial myocytes and its depletion in heart release by total and free intra-sarcoplasmic reticu- 1):5–18, 2003.
failure. Circ Heart Fail 2(5):482–489, 2009. lum calcium concentration. Biophys J 78(1):334– 25. Sato D, Bers DM: How does stochastic ryanodine
6. Morotti S, Grandi E, Summa A, et al: Theoretical 343, 2000. receptor-mediated Ca leak fail to initiate a Ca
study of L-type Ca2+ current inactivation kinetics 16. Block BA, Imagawa T, Campbell KP, et al: Struc- spark? Biophys J 101(10):2370–2379, 2011.
during action potential repolarization and early tural evidence for direct interaction between the 26. van Oort RJ, Garbino A, Wang W, et al: Disrupted
afterdepolarizations. J Physiol 590:4465–4481, molecular components of the transverse tubule/ junctional membrane complexes and hyperactive
2012. sarcoplasmic reticulum junction in skeletal muscle. ryanodine receptors after acute junctophilin knock-
7. Zuhlke RD, Pitt GS, Deisseroth K, et al: Calmodu- J Cell Biol 107(6 Pt 2):2587–2600, 1988. down in mice. Circulation 123(9):979–988, 2011.
lin supports both inactivation and facilitation of 17. Scriven DR, Dan P, Moore ED: Distribution of 27. Song LS, Sobie EA, McCulle S, et al: Orphaned
L-type calcium channels. Nature 399(6732):159– proteins implicated in excitation-contraction cou- ryanodine receptors in the failing heart. Proc Natl
162, 1999. pling in rat ventricular myocytes. Biophys J Acad Sci U S A 103(11):4305–4310, 2006.
8. Grandi E, Pandit SV, Voigt N, et al: Human atrial 79(5):2682–2691, 2000. 28. Weiss JN, Nivala M, Garfinkel A, et al: Alternans
action potential and Ca2+ model: sinus rhythm and 18. Ziman AP, Gomez-Viquez NL, Bloch RJ, et al: and arrhythmias: from cell to heart. Circ Res
chronic atrial fibrillation. Circ Res 109(9):1055– Excitation-contraction coupling changes during 108(1):98–112, 2011.
1066, 2011. postnatal cardiac development. J Mol Cell Cardiol 29. Picht E, DeSantiago J, Blatter LA, et al: Cardiac
9. Voigt N, Li N, Wang Q, et al: Enhanced sarcoplas- 48(2):379–386, 2010. alternans do not rely on diastolic sarcoplasmic
mic reticulum Ca2+ leak and increased Na+-Ca2+ 19. Weber CR, Piacentino V, 3rd, Ginsburg KS, et al: reticulum calcium content fluctuations. Circ Res
exchanger function underlie delayed afterdepolar- Na+-Ca2+ exchange current and submembrane 99(7):740–748, 2006.
izations in patients with chronic atrial fibrillation. [Ca(2+)] during the cardiac action potential. Circ 30. Despa S, Islam M, Weber CR, et al: Intracellular
Circulation 125(17):2059–2070, 2012. Res 90(2):182–189, 2002. Na+ concentration is elevated in heart failure, but
10. De Stefani D, Raffaello A, Teardo E, et al: A 20. Zima AV, Bovo E, Bers DM, et al: Ca2+ spark- Na/K pump function is unchanged. Circulation
forty-kilodalton protein of the inner membrane dependent and -independent sarcoplasmic 105:2543–2548, 2002.
Intermolecular Interactions and PART III
Cardiomyocyte Electrical Function
Ion Channel Trafficking in the Heart

Jeffrey R. Martens and Robin Shaw
17
retrograde trafficking pathways. Anterograde trafficking ensues
CHAPTER OUTLINE
only after proper protein synthesis and processing in the endo-
Ion Channel Trafficking versus Quality Control 171 plasmic reticulum and Golgi apparatus, including quality control
mechanisms, glycosylation, and posttranslation modification
Anterograde Trafficking of Channels in the Heart 171
(Figure 17-1).1 Often, channel trafficking, synthesis, and quality
At Their Destination: Channels in the control are grouped together in common discussion. However,
Myocyte Membrane 173 these processes can be distinct, using unique cellular machinery
and subject to differential regulation. Retrograde movement ini-
Retrograde Trafficking of Channels in the Heart 174 tiates with endocytosis, after which internalized proteins can
Regulation of Channel Trafficking 174 follow multiple routes to different intracellular fates (see Figure
17-1).2 One well-recognized fate is the targeting of internalized
Alterations of Channel Trafficking in proteins to lysosomes, or proteasomes, followed by degradation
Pathophysiological States 174 (see Figure 17-1). Alternatively, trafficking through recycling
Pharmacologic Manipulation of Channel Trafficking endosomes allows proteins to return to the plasma membrane and
and Implications for Cardiac/Antiarrhythmic Therapy 176 protects them from degradation (see Figure 17-1).3 Sorting at
early endosomes to Rab-GTPase–specific compartments is now
established as an important event in determining the intracellular
fate of internalized proteins.4-6 Another important component of
the endocytic machinery regulating protein surface levels is the
Ion Channel Trafficking versus coordinated movement of molecular motors. In general, protein
Quality Control trafficking is highly coordinated between long-range events
involving the microtubule-based kinesin and dynein motors, and
Research into the mechanisms of protein trafficking in the heart short-range events using unconventional myosin motors.7-10
has lagged behind that of other tissues, in particular neuronal There is a significant and growing body of literature concerning
systems. Most attention has been focused on polarized cells, such ion channel trafficking from synthesis to sorting to degradation
as neurons, that have clear compartmentalization. Interest in the in multiple tissues and cells systems that has been reviewed
mechanisms of trafficking in nonpolarized cells has been sparked previously.11-15 Our discussion will center mainly on recent work
in part by newly recognized subcellular compartmentalization on focused on the control of ion channel density at the plasma
the nanometer scale. Recent research related to cardiovascular membrane, in the heart, where relatively little is known about
ion channel trafficking is beginning to fill a gap in our knowledge protein trafficking. We have chosen to organize this chapter
of the molecular machinery controlling ion channel trafficking around the major transport events in ion channel movement
and localization in cardiac myocytes. An understanding of channel within myocytes. Given that numerous and different ion channels
trafficking will undoubtedly contribute to our knowledge of the play an essential role in all chambers of the heart and throughout
events underlying the pathophysiological conditions character- the cardiovascular system, it is impossible to include a compre-
ized by altered ion channel surface expression and will likely hensive review of all the literature. Rather, it is our intention to
provide novel insight into the general mechanism of protein traf- use select examples to highlight the major themes that have been
ficking in the cardiovascular system. Several known ion channel recently uncovered in ion channel trafficking.
mutations result in trafficking defects that contribute to the
development of disease. Importantly, as new genes and gene
mutations are discovered in our quest to understand disease and
arrhythmogenic mechanisms, these findings must be accompa-
Anterograde Trafficking of Channels
nied by studies to identify the molecular machinery and cellular in the Heart
mechanism of protein dysfunction. In the future, therapeutic
strategies designed to manipulate specific ion channel trafficking The Directed Targeting Paradigm
pathways may prove useful in the treatment of cardiovascular
channelopathies and may provide a novel approach to dynami- Posttranscription and posttranslation, the proteins that form
cally tuning electrical excitability in the heart. cardiac ion channels are modified and usually oligomerize in the
The steady-state cell surface density of ion channel proteins Golgi apparatus, where they are inserted into membrane vesicles
is determined by the balance between the anterograde and for delivery to the plasma membrane. These channel-laden
171
172 INTERMOLECULAR INTERACTIONS AND CARDIOMYOCYTE ELECTRICAL FUNCTION
Intercalated Cell surface

T-tubule
disc
Recycling
endosome
Intermediate
endosome
Golgi
apparatus
MTOC
Endoplasmic
reticulum Nucleus
EB1 Cx43 Kinesin Myosin Dynein Myosin
Cardiomyocytes
BIN1 Cav channel Kv channel Rab11a Microtubule Actin
Figure 17-1. Ion channel trafficking pathways in the heart. Anterograde and retrograde transport using kinesin motors on microtubules and myosin motors on actin fila-
ments is illustrated. Channel trafficking to subcellular compartments including intercalated disc, T-tubule, and lateral membrane is highlighted.
vesicles are transported using the cytoskeleton as a path to the junction consists of a pair of abutting hexameric hemichannels or
plasma membrane, with most studies focusing on the role of connexons in adjacent cell membranes. Each connexon consists
microtubule-based forward transport.16-21 A critical aspect of of six connexin proteins. More than 20 Cxs have been identified
forward transport is localization to membrane subdomains. It is and each is usually named according to its respective molecular
entirely possible that channels may be delivered to random mass.28 All connexins contain four transmembrane domains.
regions of the plasma membrane, only to then laterally diffuse Connexin43 is the most abundant connexin in the ventricular
within the membrane to their appropriate subdomain.17 However, myocardium. In individual ventricular cardiomyocytes, Cx43 is
the temporal and stochastic inefficiency of random channel inser- localized at the intercalated disc at longitudinal ends of the cell,
tion, together with unexplained mechanisms of subsequent lateral where Cx43 gap junctions provide rapid action potential trans-
localization other than chance interaction with a subdomain- mission and synchronized cardiac excitation.29,30
specific anchor protein, suggests that specificity of delivery from Cardiac-specific conditional deletion models with greater
the Golgi apparatus to the surface submembrane may also occur. than 50% Cx43 loss develop increased arrhythmia susceptibility
Still to be fully explored, this directed targeting paradigm of ion and sudden cardiac death.31-33 In general, these studies reveal that
channel delivery may be generalizable to all cardiac ion channels severe arrhythmia phenotypes can manifest when Cx43 coupling
and explored in terms of other cytoskeletal elements and anchor drops to less than 20% in the heart, which correlates with theo-
proteins (discussed next). retic predictions.34 Thus, maintenance of proper Cx43 expression
This section of the chapter focuses on mechanisms that govern and localization has a critical role in appropriate electrical cou-
the forward trafficking of two major cardiac ion channels that are pling and ventricular function.
indispensable for cellular excitability, cell-cell coupling, and Data exist for multiple, but not incompatible, models of Cx43
excitation-contraction coupling: (1) the L-type calcium channel trafficking to cell-cell borders of cardiac intercalated discs.17,18,20,25
Cav1.2 and (2) the dominant cardiac ventricular gap junction, There is almost universal agreement that microtubules help
connexin43 (Cx43).22-26 We also discuss in detail a protein with a deliver Cx43 to the plasma membrane. A landmark series of two
newly described role in the heart that is essential for Cav1.2 traf- papers in 2002 found evidence that newly-formed Cx43 appear
ficking and delivery, bridging integrator 1 (BIN1). BIN1 is at the perimeter of the Cx43 plaques and then diffuse into central
rapidly emerging as a multifunctional cardiac player beyond its plaque regions.17,20 Taken together with microtubule delivery, the
previously understood function as a membrane scaffold. model developed that Cx43 hemichannels are inserted into the
general plasma membrane, rapidly diffuse to the edge of dense
plaques, and then more slowly diffuse into the plaque center.
Forward Trafficking of Cx43 Subsequent to these studies, it has been observed that Cx43 can
be inserted directly at the edge and into plaques, and membrane
Gap junctions are intercellular channels that form low-resistance fluidity exits within the plaque region.18,35 It has also been found
pathways, allowing ions and metabolites to flow from cell to cell. that the plaques are internalized not necessarily from the center,
In the heart, gap junctions electrically couple cardiomyocytes but that different segments of plaque can be internalized at any
to orchestrate spatial propagation of action potentials.27 A gap time35 and full plaque internalization can occur in one step.36
Ion Channel Trafficking in the Heart 173
More recently, it has been found that Cx43 occurs in regions combination of the large size of the α-subunit, Cav1.2, the mul-
17
surrounding Cx43 plaques, the “perinexus,” at higher density tiple β-subunits, the difficulty of manipulating cardiomyocytes
than in general membrane where Cx43 proteins may interact with T-tubules in culture, and less dense enrichment of the
with scaffolding proteins and other ion channels.37 These studies channel on the plasma membrane, presenting difficulty to clearly
are providing evidence that the gap junction plaque and sur- characterize by cytochemistry a distinct phenotype. Subunits are
rounding regions are highly dynamic with complex behavior of important for LTCC trafficking. For instance, the β-subunit is
targeted insertion and internalization. necessary for surface expression.42 Moreover, the α2δ-subunit
Given the low density of Cx43 hemichannels in membrane synergizes with the β-subunit to promote surface channel expres-
well away from Cx43 plaque regions, and the technical difficulty sion.43 These auxiliary subunits enhance surface expression by
of distinguishing Cx43 inserted in membrane from submembra- promoting channel export from the endoplasmic reticulum and
nous and still cytoplasmic collections of protein, it is difficult to overall channel stability.44
find studies that can quantify the lateral diffusion coefficient of Despite the necessary localization of LTCC to T-tubules, it
membrane-bound Cx43. It may be that free hemichannels rapidly was not understood until recently how the localization occurs. In
diffuse within the plasma membrane before stopping at plaque 2010, it was found that Cav1.2-based channels adhere to the
regions, but direct evidence for this phenomenon is lacking. The directed targeted paradigm.23 In particular, the membrane-
directed targeting paradigm is based on the observation that de scaffolding protein BIN1 was found to provide a membrane
novo intracellular Cx43 hemichannels can arrive directly in the anchor that attaches dynamic microtubules, allowing delivery of
gap junction plaque region. The hemichannels are targeted to Cav1.2 channels directly to T-tubule membranes. Using HL-1
plaque regions with specificity obtained from the channel protein cells that express Cav1.2 but do not form T-tubules, and non-
(Cx43), microtubule plus-end tracking proteins (EB1 and p150 muscle cell lines that neither express Cav1.2 nor naturally form
[glued], and a membrane anchor [adherens junction structure]).18 T-tubules, it was found that exogenous BIN1 caused the forma-
By this model, the dynamic microtubule highways are anchored tion of deep invaginations in the cell membrane enriched with
and terminate at adherens junction structures, allowing directed endogenous or overexpressed Cav1.2, suggesting that BIN1-
delivery of Cx43 hemichannels to adherens junction–containing containing membrane is sufficient to recruit Cav1.2 channels. To
membrane to occur. It is probable that once inserted into plaque test the possibility that BIN1 serves as an anchoring site for
regions, local hemichannel diffusion occurs within the plaque and microtubules on which Cav1.2 channels are trafficked, the study
between the plaque and the plaque perinexus. tracked growing microtubules extending toward BIN1 clusters
Although most studies of Cx43 forward trafficking focus on and found that the microtubule plus end-tracking proteins pause
the microtubule cytoskeleton, the actin cytoskeleton is also and associate with BIN1 clusters at the cell periphery. Moreover,
involved in delivery of membrane proteins and ion channels. Dye it was determined that the non-BAR cytoskeleton–anchoring
transfer studies revealed the dependence of Cx43 insertion into domain of BIN1 is required for this activity because truncation
plaque on actin, as tested by pharmacologic actin disruption and mutants lacking this domain failed to cluster Cav1.2 at cell surface
anti-actin antibodies.38,39 Actin is also important for Cx43 plaque invaginations. The microtubule plus end-tracking protein that
internalization, although we recently found that when internal- may aid in microtubule anchoring to BIN1, analogous to EB1
ization is blocked, forward delivery is still actin dependent.26 binding to adherens junctions,18,25 has not yet been identified.
Furthermore, a remarkable greater than 80% of cytoplasmic
post-Golgi Cx43 is slow-moving or stationary and actin-
associated. These findings indicate that microtubules work in At Their Destination: Channels
concert with actin to deliver Cx43 to the plasma membrane. It
may be that post-Golgi Cx43 vesicles exist in actin-associated
in the Myocyte Membrane
reservoirs within the cytoplasm, waiting for either the right
microtubule to permit membrane delivery or mass acute delivery Subcellular Localization of Kv Channels
in the case of a metabolic stress. Within Myocytes
Most tissues, and even single cells, express multiple Kv channel
Forward Trafficking of Cav1.2 types belonging to one or more subfamilies. Importantly, the
subcellular localization of these different channel isoforms is nec-
Although connexins electrically couple cardiomyocytes, the essary for proper function and signaling. For example, the
membrane ion channel most responsible for calcium entry and regional and cell-specific distribution of Kv channels contributes
excitation-contraction coupling is the α1C pore-forming subunit to local variations in the shape and duration of the cardiac action
(Cav1.2) of the L-type voltage-gated (Cav) Ca2+ channel (LTCC). potential. At the subcellular level, Kv channels can be expressed
In response to membrane depolarization by sodium currents, within T-tubules, at the intercalated disc, or on the lateral mem-
LTCCs open to allow inward Ca2+ entry during the plateau phase brane, in an isoform specific pattern. Although it is known that
of the cardiac action potential. The L-type currents (ICa,L) that Kv channel expression differs in different regions of the heart—
are generated initiate calcium-induced Ca2+ release from the sar- for example the atria versus ventricles or endocardium versus
coplasmic reticulum (SR) via ryanodine receptors, which leads to epicardium—it is unclear whether the subcellular localization
muscle contraction. The Cav1.2 protein is large and consists of varies within the heart itself. Recently it has been shown that
24 transmembrane segments organized into four homologous ectopically expressed Kv2.1 shows a different subcellular localiza-
domains.40 Auxiliary subunits of the LTCC include the β-subunit, tion between atria and ventricles. It is not known whether this
the α2δ-subunit, and the γ subunit, which help regulate trafficking chamber-specific localization applies to endogenous channel
of Cav1.2 to the cell membrane and regulate the voltage depen- localization. It is also not known how different Kv channel iso-
dence of channel gating.41 forms affect localization.
T-tubule invaginations of ventricular cardiomyocyte plasma Historically, Kv localization is believed to primarily
membrane are enriched with LTCCs. This enrichment is neces- involve protein-protein interactions among channel proteins
sary for calcium-induced calcium release with nearby ryanodine and PDZ-domain–containing scaffolding proteins or the actin
receptors, which is important for beat-to-beat excitation- cytoskeleton. Recent work shows that Kv2.1 channels are immo-
contraction coupling in the heart. Trafficking of the LTCCs is bilized on the lateral membrane of atrial myocytes, where their
not as extensively studied as that of Cx43, probably owing to a diffusion is likely limited by actin cytoskeletal corrals as in
hippocampal neurons.45 In addition, Kv1.5 has been shown to mechanisms. This includes posttranslational modification,
directly couple to α-actinin-2 in HEK 293 cells via a specific channel assembly, and external and internal influences such as
sequence in the amino terminus of the channel.46 In addition, oxidative stress, neurohumoral control, and disease. Perhaps
Kv1.5-cytoskeleton interactions appear to play a role in modula- most studied are posttranslational modifications to channel pro-
tion of the channel by protein kinase A in oocytes.47 Further, the teins, which affect channel stability, function, and, in some cases,
actin binding protein, cortactin, regulates Kv1.5./N-cadherin trafficking. It is well known that glycosylation marks channel
interactions at the intercalated disc. Kvβ-subunits are also likely proteins that are expressed on the cell surface of cardiac myo-
to interact with the cytoskeleton and thus immobilize channels on cytes. In addition, ubiquitin modification occurs with multiple Kv
the cell surface.48 In addition, some Kvβ-subunits behave as chap- channel isoforms, implicating the proteasome in channel degra-
erone proteins, promoting the cell surface expression of a subset dation. However, few papers directly address posttranslational
of co-expressed α-subunits and perhaps influencing localiza- modification of channels and its exact role in Kv channel traffick-
tion.49,50 A role for PDZ domain–containing scaffolding proteins, ing. Two recent exceptions include reports on the phosphoryla-
such as SAP97, in the regulation of Kv1.5 surface localization has tion of KCNQ1/KCNE1 channels and the sulfenic acid
also been described.51,52 Furthermore, it has been shown that the modification of Kv1.5. In the first example, RAB-dependent
K+ channel interacting protein, KChIP2, contributes to the for- KCNQ1/KCNE1 anterograde trafficking was shown to require
mation of functional mouse Kv1.5-encoded ventricular chan- phosphorylation. This process includes the activation of phos-
nels.53 It is also believed that membrane microdomains, perhaps phoinositide 3-phosphate 5-kinase and the generation of PI(3,5)
formed by protein-lipid interactions, are an important and novel P(2) by the serum- and glucocorticoid-inducible kinase 1. Valida-
mechanism of Kv channel localization54,55 in the heart. This is tion of this mechanism in cardiac myocytes will be an important
furthered by the finding that Kv1.5, caveolin, and SAP97 exist as step forward. In the second example, it was shown that under
part of a tripartite complex in the heart56; however, other reports peroxide-induced stress a single cysteine residue (C581) in the
indicate little co-localization of Kv1.5 with caveolin in myocytes. C-terminal domain of Kv1.5 undergoes oxidation to sulfenic acid,
Other work shows that KCNQ1-KCNE1 channels, which encode a novel redox-sensitive posttranslational modification to this
for the K+ current in the heart, are localized by interactions with channel. In turn, oxidation of C581 triggers internalization of
the A-kinase anchoring protein, Yotiao.57 Although progress has Kv1.5, decreases current density, and diverts the channel from
been made in identifying elements involved in Kv channel local- recycling endosomes toward degradation. The sulfenic acid mod-
ization, clustering, and anchoring, less is known regarding the ification was demonstrated on Kv1.5 in the heart, whereas a
molecular machinery controlling their plasma membrane target- global increase in sulfenic acid–modified proteins in human
ing or the regulation of surface expression. patients with atrial fibrillation linked the trafficking regulation to
disease pathophysiology.
Retrograde Trafficking of Channels Regulation of Cx43 Trafficking

in the Heart
Histone deacetylases, which modify histone tails to repress gene
At the plasma membrane, localization to specific membrane transcription, were shown to regulate Cx43 expression in mouse
microdomains, and association with scaffolding proteins into embryonic stem cells.70 In mdx hearts, global elevation of histone
macromolecular signaling complexes, likely contribute to the sta- acetylase activity is associated with lateralized Cx43 and increased
bility and biological function of Kv channels.58-64 Despite associa- physical interaction between the acetylase P300/CBP-associated
tion with scaffolding proteins, Kv channels have been shown to factor with Nε-lysine acetylated Cx43. Addition of the histone
undergo dynamic trafficking at the plasma membrane through acetylase inhibitor anacardic acid reduced Cx43 Nε-lysine acety-
constitutive internalization and recycling.65 Internalization of lation and restored Cx43 localization to the intercalated disc.
Kv1.5 occurs via a dynein-mediated, microtubule-dependent Correspondingly, lateralization of Cx43 was achieved by a short
pathway.65,66 After internalization, sorting of Kv1.5 into specific drug treatment that increased total protein acetylation. These
Rab-dependent endocytic compartments determines the intracel- studies and others have revealed multiple new functions for pro-
lular fate of the channel. Specifically, association of Kv1.5 with teins normally associated with gene expression regulation in
Rab4- or Rab11-containing endocytic vesicles is associated with directly affecting connexin trafficking.
recycling of the channel back to the plasma membrane, whereas
association with rab7-containing vesicles denotes channel
degradation.65-67 In addition, KCNQ1/KCNE1 potassium chan- Alterations of Channel Trafficking in
nels also undergo constitutive recycling involving Rab-specific
compartments.68 Specifically, Rab11 is implicated in exocytic traf- Pathophysiological States
ficking, whereas endocytosis of KCNQ1/KCNE1 is dependent
on Rab5. Finally, data indicate that hERG channels, in heterolo- Kv Channel Trafficking in the Diseased Heart
gous expression systems, undergo endoplasmic reticulum export
in COPII vesicles and endosomal recycling before being pro- Alterations in the cell surface expression of functional Kv chan-
cessed in the Golgi. This process involves the GTPases Sar1 and nels occur in numerous cardiovascular disease states and undoubt-
Rab11B, with a potential minor contribution from ARF1.69 edly contribute to their pathophysiology. One example is Kv1.5
and its role in paroxysmal and persistent atrial fibrillation,71,72 as
well as chronic hypoxic pulmonary hypertension.73 Overexpres-
sion of Kv1.5 in rat cardiomyocytes dramatically shortens action
Regulation of Channel Trafficking potential duration, producing a phenotype similar to that
observed in short QT syndrome.74 Conversely, intracellular
Posttranslational Regulation of Kv sequestration of Kv1.5 in mouse cardiac myocytes, by overexpres-
Channel Trafficking sion of a dominant negative channel fragment, results in reduced
surface levels of Kv1.5 and long QT arrhythmias.75 In patients
The regulation of Kv channel trafficking can occur at multiple with persistent or paroxysmal atrial fibrillation, reduced outward
stages in the transport of the channel and with multiple potassium current is observed. This reduction in current is
caused, in part, by diminished Kv1.5 protein expression, whereas Internalization of Cx43 is clathrin-dependent36 and may also
17
other potassium channels, such as Kv2.1, remain unaffected.71 involve Zonula Occludens-1 (ZO-1), which is a 220-kDa scaf-
Interestingly, Kv1.5 mRNA levels are unchanged, indicating that folding protein that tethers transmembrane proteins such as con-
alterations in steady-state protein levels may be responsible for nexins directly, or via associated adaptor proteins, to the
the aberrant phenotype; however, these changes have not been cytoskeleton.81 ZO-1 associates with Cx43 via its second PDZ
likened directly to channel trafficking. Nevertheless, these vicis- domain at the perimeter of the gap junctional plaque.82-85 In
situdes are intriguing, given that several other K+ channel traf- patients with end-stage congestive heart failure caused by idio-
ficking defects lead to the development of disease. Of significance pathic dilated cardiomyopathy DCM and ischemic cardiomyopa-
are changes that occur with long QT syndrome. thy, decreased Cx43 expression at the cell surface is accompanied
Pioneering work from January and colleagues first reported by increased association with ZO-1.86 These studies suggest that
that hERG channel mutations altered protein trafficking, causing ZO-1 may regulate cell surface gap junction availability by pro-
long QT syndrome type 2.44 These trafficking defects reduce the moting Cx43 endocytosis.
number of functional channels expressed on the myocyte surface,
contributing to a reduction in repolarizing current. These obser-
vations have been confirmed by multiple groups and expanded to Functional Significance of Cav1.2 and BIN1
include numerous long QT syndrome type 2 mutations. Recently, in the Human Heart
it was shown that trafficking-deficient hERG K+ channels linked
to long QT syndrome are regulated by a microtubule-dependent Calcium influx in the working myocardium, which critically
quality control event. This highlights the possibility that many depends on the α1C pore-forming subunit, plays a central role
of these reported alterations in surface density reflect changes in in converting electrical impulses into mechanical activation of the
protein folding/stability events versus the surface transport of contractile machinery. Gain-of-function mutations in Cav1.2,
channels. which cause near complete loss of voltage-dependent channel
One recent example that directly links long QT syndrome inactivation and calcium overload in multiple tissues, are linked
mutants to channel trafficking involves KCNQ1 and KCNE1 to Timothy syndrome.87,88 Prolonged sodium currents resulting
subunits. In this report, it was shown that specific disease- from disrupted channel inactivation delay and repolarization sig-
associated mutations in these channel subunits disrupt normal nificantly increases QT interval duration and the risk of arrhyth-
endosomal recycling of potassium current channels.68 This work mias and sudden death. Interestingly, the disease-causing
provides novel mechanistic insight into potentially fatal cardiac mutations G406R and G402S were found within exons 8 and 8a,
arrhythmias and highlights that trafficking pathways may repre- which are alternatively spliced in a mutually exclusive fashion
sent important therapeutic intervention points for the treatment and are present in different relative amounts in various tissues. It
of cardiac arrhythmias. was suggested that the multitude of severities of symptoms
across multiple organs reflects tissue-specific expression of splice
variants. In patients with hypertrophic heart failure, aberrant
Cx43 Regulation in the Diseased Heart splicing of the mutually exclusive exons 31 and 32 was detected
such that re-expression of the fetal exon contributed to disease
The density and composition of gap junctions determine cell-cell progression.89
coupling efficiency and ensure orchestrated current flow. Changes Loss-of-function mutations in Cav1.2 in patients with Brugada
in Cx43 expression and trafficking can alter conduction and syndrome are characterized by a short QT interval and sudden
impair heart function. Many types of ventricular remodeling that cardiac death.90 When missense mutations at G490R and A39V
occur in humans as a result of cardiac overload are characterized were co-expressed with other LTCC subunits in CHO cells, a
by changes in the expression and distribution of Cx43. Early clear reduction of ICa,L was observed. Confocal microscopy studies
myocardial infarction studies revealed decreased Cx43 at the revealed that channel trafficking was unaffected. Thus, it was
intercalated disc and lateralization of remaining channels in the hypothesized that the G490R mutation, which is located in a
infarct boarder zone.76-78 Altered gap junction plaque size exists linker region, interferes with β-subunit binding to inhibit current
in most forms of heart failure, where changes in the cellular density. Another study revealed that a V2041I mutation in the
distribution of Cx43 affects the spread of excitation in the heart, C-terminus also reduces ICa,L amplitude by decreasing channel
which can be arrhythmogenic. More recent studies have explored conductance and altering current inactivation.90
the mechanisms of altered Cx43 distribution in disease. Abnormal excitation-contraction coupling, resulting from
Reduced cardiac cell-cell coupling in ischemic and nonische deregulation at the onset of Ca2+ influx through the T-tubule
mic hearts is strongly associated with Cx43 dephosphorylation, network, is increasingly implicated in heart failure progression
which generally has been inferred from Western blot band shifts and sudden cardiac death. Heart content of Cav1.2, typically
and phosphospecific antibodies. For instance, Cx43 dephosphor- analyzed by biochemical assay of heart muscle lysates, is not
ylation leads to cell-cell uncoupling in the setting of ischemia, altered in failing hearts.24 However, although it was recently
which can be rescued by direct suppression of Cx43 dephos- confirmed that total cellular Cav1.2 content is unchanged in
phorylation.79,80 The specific Cx43 residues that undergo post- heart failure, we also found that the channels are internalized.24
translational modification during disease are still being explored In testing the forward trafficking machinery of Cav1.2 in failing
and there is likely a rich balance of disease-related phosphoryla- hearts, we found that human heart failure involves a reduction in
tion and dephosphorylation on the Cx43 C-terminus. An elegant BIN1 at both the protein and mRNA message level, implying
study recently found that inhibiting three specific casein kinase that the reduction is transcriptional. Subsequent studies in adult
sites (S325, S328, S330) from dephosphorylation protected car- cardiomyocytes confirmed that decreased BIN1 reduces forward
diomyocytes from Cx43 remodeling and arrhythmias during trafficking of Cav1.2, also diminishing intracellular calcium tran-
ischemia. It is not known whether diminished cell-cell coupling sients (in isolated cells and zebrafish hearts) and contractility (of
is a result of increased rate of plaque internalization or dimin- zebrafish hearts). Recent rat studies confirmed that BIN1 is
ished rate of Cx43 delivery, or both. We recently found that decreased in failing hearts and recovers with successful treat-
hearts with end-stage ischemic cardiomyopathy were character- ment.91 It is therefore possible that reduced contractility of pro-
ized by specific disruption of the cytoskeleton-based Cx43 gressively failing heart can be traced, in part, to decreased
forward trafficking machinery without changes in total transcription of BIN1. Understanding the regulation of BIN1
expression.25 transcription is currently an active investigation.
In keeping with the cell to bedside theme of this book, it is worth trafficking has focused primarily on hERG K+ channels. Altera-
mentioning that the potential for BIN1 as a blood-available bio- tions in hERG-mediated IKr current, whether drug-induced or a
marker for arrhythmogenic right ventricular cardiomyopathy result of the more than 200 naturally occurring mutations of this
(ARVC) was also recently identified.92 ARVC is a primary myo- channel, may induce or contribute to the development of long
cardial disorder with a high incidence of ventricular arrhyth- QT syndrome. Nearly 70% of these mutant channels can be
mias.93 In a retrospective study of 24 patients, plasma BIN1 rescued to the plasma membrane by antiarrhythmic drugs such
predicted cardiac function status and incidence of ventricular as E4031.96,97 The precise mechanism by which drugs decrease
arrhythmias. BIN1 has high specificity and sensitivity in distin- the cell surface expression of the hERG channel protein is uncer-
guishing ARVC patients with severe disease from those with tain. These drugs likely act chronically to stabilize misfolded
milder symptoms. In addition, BIN1 levels correlated inversely protein through facilitation of quality control machinery in the
with disease progression in serial blood draws and predicted endoplasmic reticulum to facilitate its maturation and export
future arrhythmias in patients without severe heart failure with from the endoplasmic reticulum.98,99 Importantly, these studies
82% accuracy.92 Current studies are focused on why BIN1, which give credence to the idea that antiarrhythmic drugs may be devel-
is an intracellular yet membrane-attached protein, is also avail- oped to manipulate specific ion channel trafficking pathways as a
able in the blood. novel therapeutic approach for treating cardiac arrhythmias.
Recently, a previously unrecognized mechanism of antiar-
rhythmic drug action in the acute modulation of Kv1.5 channel
trafficking was reported.94 Using quinidine, an antiarrhythmic
Pharmacologic Manipulation of Channel agent that has both class Ia actions100,101 and class III actions in
Trafficking and Implications for Cardiac/ mammalian atrium and ventricle, it demonstrated that channel
Antiarrhythmic Therapy blockers can both inhibit ion conduction and regulate the stabil-
ity of the channel protein within the membrane. In this study,
quinidine resulted in a dose- and time-dependent internalization
New therapeutic strategies that focus on the regulation of ion of Kv1.5, concomitant with channel block. Interestingly, this
channel surface density are emerging.94,95 Traditional antiar- quinidine-induced internalization of Kv1.5 was found to be
rhythmic drugs target the ion permeability of channels; however, subunit-dependent and stereospecific,94 which highlights the pos-
this approach has not yet yielded a satisfactory outcome. There sibility for the development of atrial selective agents that specifi-
are two ways to decrease channel current: either through a direct cally modulate surface density. Although much work needs to be
effect on the conduction properties (classically pore block) of done to identify the precise cellular trafficking machinery and the
channel subunits or through alterations in surface density of the mechanisms regulating ion channel surface density, the recent
protein. The concept of drugs modulating ion conduction and/ studies highlighted in this chapter suggest that modulation of ion
or surface density of channels it not new. Research into the channel trafficking pathways may be an alternative or comple-
therapeutic potential of antiarrhythmic drugs that alter channel mentary strategy for treating cardiac arrhythmias.
potassium channels. Biochem Soc Trans 35:1069– 24. Hong TT, et al: BIN1 is reduced and Cav1.2
References 1073, 2007. trafficking is impaired in human failing cardio-
13. Deutsch C: The birth of a channel. Neuron myocytes. Heart Rhythm 9:812–820, 2012.
1. Ellgaard L, Helenius A: Quality control in 40:265–276, 2003. 25. Smyth JW, et al: Limited forward trafficking of
the endoplasmic reticulum. Nat Rev 4:181–191, 14. Deutsch C: Potassium channel ontogeny. Annu connexin 43 reduces cell-cell coupling in stressed
2003. Rev Physiol 64:19–46, 2002. human and mouse myocardium. J Clin Invest
2. Olson TM, et al: Kv1.5 channelopathy due to 15. Staudacher I, Schweizer PA, Katus HA, et al: 120:266–279, 2010.
KCNA5 loss-of-function mutation causes human hERG: protein trafficking and potential for 26. Smyth JW, et al: Actin cytoskeleton rest stops
atrial fibrillation. Hum Mol Genet 15:2185–2191, therapy and drug side effects. Curr Opin Drug regulate anterograde traffic of connexin 43 vesi-
2006. Disc Dev 13:23–30, 2010. cles to the plasma membrane. Circ Res 110:978–
3. London B, et al: Targeted replacement of KV1.5 16. Hamm-Alvarez SF, Sheetz MP: Microtubule- 989, 2012.
in the mouse leads to loss of the 4-aminopyridine- dependent vesicle transport: modulation of 27. Smyth JW, Shaw RM: The gap junction life cycle.
sensitive component of I(K,slow) and resistance to channel and transporter activity in liver and Heart Rhythm 9:151–153, 2012.
drug-induced qt prolongation. Circ Res 88:940– kidney. Physiol Rev 78:1109–1129, 1998. 28. Sohl G, Willecke K: An update on connexin genes
946, 2001. 17. Gaietta G, et al: Multicolor and electron micro- and their nomenclature in mouse and man. Cell
4. Marks DL, Pagano RE: Endocytosis and sorting scopic imaging of connexin trafficking. Science Commun Adhes 10:173–180, 2003.
of glycosphingolipids in sphingolipid storage 296:503–507, 2002. 29. van Kempen MJ, Fromaget C, Gros D, et al:
disease. Trends Cell Biol 12:605–613, 2002. 18. Shaw RM, et al: Microtubule plus-end-tracking Spatial distribution of connexin43, the major
5. Maxfield FR, McGraw TE: Endocytic recycling. proteins target gap junctions directly from the cell cardiac gap junction protein, in the developing
Nat Rev Mol Cell Biol 5:121–132, 2004. interior to adherens junctions. Cell 128:547–560, and adult rat heart. Circ Res 68:1638–1651,
6. Prekeris R: Rabs, Rips, FIPs, and endocytic mem- 2007. 1991.
brane traffic. Sci World J 3:870–880, 2003. 19. Johnson RG, et al: Gap junctions assemble in the 30. Fromaget C, el Aoumari A, Gros D: Distribution
7. Brown SS: Cooperation between microtubule- presence of cytoskeletal inhibitors, but enhanced pattern of connexin 43, a gap junctional protein,
and actin-based motor proteins. Annu Rev Cell assembly requires microtubules. Exp Cell Res during the differentiation of mouse heart myo-
Dev Biol 15:63–80, 1999. 275:67–80, 2002. cytes. Differentiation 51:9–20, 1992.
8. Dantzig JA, Liu TY, Goldman YE: Functional 20. Lauf U, et al: Dynamic trafficking and delivery of 31. Eckardt D, et al: Functional role of connexin43
studies of individual myosin molecules. Ann N Y connexons to the plasma membrane and accretion gap junction channels in adult mouse heart
Acad Sci 1080:1–18, 2006. to gap junctions in living cells. Proc Natl Acad Sci assessed by inducible gene deletion. J Mol Cell
9. Langford GM: Myosin-V, a versatile motor for U S A 99:10446–10451, 2002. Cardiol 36:101–110, 2004.
short-range vesicle transport. Traffic 3:859–865, 21. Zadeh AD, et al: Kif5b is an essential forward 32. Danik SB, et al: Modulation of cardiac gap junc-
2002. trafficking motor for the Kv1.5 cardiac potassium tion expression and arrhythmic susceptibility.
10. Seabra MC, Coudrier E: Rab GTPases and channel. J Physiol 587:4565–4574, 2009. Circ Res 95:1035–1041, 2004.
myosin motors in organelle motility. Traffic 22. Zhang SS, et al: Iroquois homeobox gene 3 estab- 33. Gutstein DE, et al: Conduction slowing and
5:L393–399, 2004. lishes fast conduction in the cardiac His-Purkinje sudden arrhythmic death in mice with cardiac-
11. Steele DF, Eldstrom J, Fedida D: Mechanisms of network. Proc Natl Acad Sci U S A 108:13576– restricted inactivation of connexin43. Circ Res
cardiac potassium channel trafficking. J Physiol 13581, 2011. 88:333–339, 2001.
582:17–26, 2007. 23. Hong TT, et al: BIN1 localizes the L-type 34. Shaw RM, Rudy Y: Ionic mechanisms of propaga-
12. Steele DF, Zadeh AD, Loewen ME, et al: Local- calcium channel to cardiac T-tubules. PLoS Biol tion in cardiac tissue. Roles of the sodium and
ization and trafficking of cardiac voltage-gated 8:e1000312, 2010. L-type calcium currents during reduced excitabil-
ity and decreased gap junction coupling. Circ Res 56. Folco EJ, Liu GX, Koren G: Caveolin-3 and normal myocardium and human ischaemic
81:727–741, 1997. SAP97 form a scaffolding protein complex that heart disease. Histochemistry 99:105–120,
35. Falk MM, Baker SM, Gumpert AM, et al: Gap
junction turnover is achieved by the internaliza-
tion of small endocytic double-membrane vesi-
regulates the voltage-gated potassium channel
Kv1.5. Am J Physiol Heart Circ Physiol
287:H681–H690, 2004.
77.
1993.
Peters AM, Bertram P, Gahr M, et al: Reduced
secretion of interleukin-1 and tumor necrosis
17
cles. Mol Biol Cell 20:3342–3352, 2009. 57. Chen L, Kass RS: A-kinase anchoring protein 9 factor-alpha by neonatal monocytes. Biol Neonate
36. Piehl M, et al: Internalization of large double- and IKs channel regulation. J Cardiovasc Pharma- 63:157–162, 1993.
membrane intercellular vesicles by a clathrin- col 58:459–413, 2011. 78. Smith JH, Green CR, Peters NS, et al: Altered
dependent endocytic process. Mol Biol Cell 58. Abi-Char J, et al: The anchoring protein SAP97 patterns of gap junction distribution in ischemic
18:337–347, 2007. retains Kv1.5 channels in the plasma membrane heart disease. An immunohistochemical study of
37. Rhett JM, Jourdan J, Gourdie RG: Connexin 43 of cardiac myocytes. Am J Physiol Heart Circ human myocardium using laser scanning confocal
connexon to gap junction transition is regulated Physiol 294:H1851–H1861, 2008. microscopy. Am J Pathol 139:801–821, 1991.
by zonula occludens-1. Mol Biol Cell 22:1516– 59. Abi-Char J, et al: Membrane cholesterol modu- 79. Beardslee MA, et al: Dephosphorylation and
1528, 2011. lates Kv1.5 potassium channel distribution and intracellular redistribution of ventricular con-
38. Theiss C, Meller K: Microinjected anti-actin anti- function in rat cardiomyocytes. J Physiol nexin43 during electrical uncoupling induced by
bodies decrease gap junctional intercellular com- 582:1205–1217, 2007. ischemia. Circ Res 87:656–662, 2000.
mmunication in cultured astrocytes. Exp Cell Res 60. Eldstrom J, Choi WS, Steele DF, et al: SAP97 80. Ai X, Pogwizd SM: Connexin 43 downregulation
281:197–204, 2002. increases Kv1.5 currents through an indirect and dephosphorylation in nonischemic heart
39. Thomas T, Jordan K, Laird DW: Role of cyto- N-terminal mechanism. FEBS Lett 547:205–211, failure is associated with enhanced colocalized
skeletal elements in the recruitment of Cx43-GFP 2003. protein phosphatase type 2A. Circ Res 96:54–63,
and Cx26-YFP into gap junctions. Cell Commun 61. Folco EJ, Liu GX, Koren G: Caveolin-3 and 2005.
Adhes 8:231–236, 2001. SAP97 form a scaffolding protein complex that 81. Miyoshi J, Takai Y: Structural and functional asso-
40. Dai S, Hall DD, Hell JW: Supramolecular assem- regulates the voltage-gated potassium channel ciations of apical junctions with cytoskeleton.
blies and localized regulation of voltage-gated ion Kv1.5. Am J Physiol Heart Circ Physiol Biochim Biophys Acta 1778:670–691, 2008.
channels. Physiol Rev 89:411–452, 2009. 287:H681–690, 2004. 82. Giepmans BN, Moolenaar WH: The gap junc-
41. Benitah JP, Alvarez JL, Gomez AM: L-type 62. Martens JR, Sakamoto N, Sullivan SA, et al: tion protein connexin43 interacts with the second
Ca(2+) current in ventricular cardiomyocytes. J Isoform-specific localization of voltage-gated K+ PDZ domain of the zona occludens-1 protein.
Mol Cell Cardiol 48:26–36, 2010. channels to distinct lipid raft populations. Target- Curr Biol 8:931–934, 1998.
42. Fang K, Colecraft HM: Mechanism of auxiliary ing of Kv1.5 to caveolae. J Biol Chem 276:8409– 83. Toyofuku T, et al: Direct association of the gap
beta-subunit-mediated membrane targeting of 8414, 2001. junction protein connexin-43 with ZO-1 in
L-type (Ca(V)1.2) channels. J Physiol 589:4437– 63. Maruoka ND, et al: Alpha-actinin-2 couples to cardiac myocytes. J Biol Chem 273:12725–12731,
4455, 2011. cardiac Kv1.5 channels, regulating current density 1998.
43. Yasuda T, et al: Auxiliary subunit regulation of and channel localization in HEK cells. FEBS Lett 84. Zhu C, et al: Quantitative analysis of ZO-1 colo-
high-voltage activated calcium channels expressed 473:188–194, 2000. calization with Cx43 gap junction plaques in cul-
in mammalian cells. Eur J Neurosci 20:1–13, 64. Mathur R, et al: A specific N-terminal residue in tures of rat neonatal cardiomyocytes. Microsc
2004. Kv1.5 is required for upregulation of the channel Microanal 11:244–248, 2005.
44. Simms BA, Zamponi GW: Trafficking and stabil- by SAP97. Biochem Biophys Res Commun 85. Hunter AW, Barker RJ, Zhu C, et al: Zonula
ity of voltage-gated calcium channels. Cell Mol 342:1–8, 2006. occludens-1 alters connexin43 gap junction size
Life Sci 69:843–856, 2012. 65. McEwen DP, et al: Rab-GTPase-dependent and organization by influencing channel accre-
45. O’Connell KM, Whitesell JD, Tamkun MM: endocytic recycling of Kv1.5 in atrial myocytes. J tion. Mol Biol Cell 16:5686–5698, 2005.
Localization and mobility of the delayed-rectifer Biol Chem 282:29612–29620, 2007. 86. Bruce AF, Rothery S, Dupont E, et al: Gap
K+ channel Kv2.1 in adult cardiomyocytes. Am J 66. Choi WS, et al: Kv1.5 surface expression is modu- junction remodelling in human heart failure is
Physiol Heart Circ Physiol 294:H229–H237, lated by retrograde trafficking of newly endocy- associated with increased interaction of con-
2008. tosed channels by the dynein motor. Circ Res nexin43 with ZO-1. Cardiovasc Res 77:757–765,
46. Maruoka ND, et al: Alpha-actinin-2 couples to 97:363–371, 2005. 2008.
cardiac Kv1.5 channels, regulating current density 67. Balse E, et al: Cholesterol modulates the recruit- 87. Splawski I, et al: Ca(V)1.2 calcium channel dys-
and channel localization in HEK cells. FEBS Lett ment of Kv1.5 channels from Rab11-associated function causes a multisystem disorder including
473:188–194, 2000. recycling endosome in native atrial myocytes. arrhythmia and autism. Cell 119:19–31, 2004.
47. Mason HS, Latten MJ, Godoy LD, et al: Modula- Proc Natl Acad Sci U S A 106:14681–14686, 88. Splawski I, et al: Severe arrhythmia disorder
tion of Kv1.5 currents by protein kinase A, tyro- 2009. caused by cardiac L-type calcium channel muta-
sine kinase, and protein tyrosine phosphatase 68. Seebohm G, et al: Regulation of endocytic recy- tions. Proc Natl Acad Sci U S A 102:8089–8096,
requires an intact cytoskeleton. Mol Pharmacol cling of KCNQ1/KCNE1 potassium channels. discussion 8086–8088, 2005.
61:285–293, 2002. Circ Res 100:686–692, 2007. 89. Yang Y, et al: L-type Ca2+ channel alpha 1c
48. Nakahira K, Matos MF, Trimmer JS: Differential 69. Delisle BP, et al: Small GTPase determinants for subunit isoform switching in failing human ven-
interaction of voltage-gated K+ channel beta- the Golgi processing and plasmalemmal expres- tricular myocardium. J Mol Cell Cardiol 32:973–
subunits with cytoskeleton is mediated by unique sion of human ether-a-go-go related (hERG) K+ 984, 2000.
amino terminal domains. J Mol Neurosci 11:199– channels. J Biol Chem 284:2844–2853, 2009. 90. Antzelevitch C, et al: Loss-of-function mutations
208, 1998. 70. Zupkovitz G, et al: Negative and positive regula- in the cardiac calcium channel underlie a new
49. Shi G, et al: Beta subunits promote K+ channel tion of gene expression by mouse histone deacety- clinical entity characterized by ST-segment eleva-
surface expression through effects early in biosyn- lase 1. Mol Cell Biol 26:7913–7928, 2006. tion, short QT intervals, and sudden cardiac
thesis. Neuron 16:843–852, 1996. 71. Van Wagoner DR, Pond AL, McCarthy PM, et death. Circulation 115:442–449, 2007.
50. Manganas LN, Trimmer JS: Subunit composition al: Outward K+ current densities and Kv1.5 91. Lyon AR, et al: Plasticity of surface structures and
determines Kv1 potassium channel surface expression are reduced in chronic human atrial beta2-adrenergic receptor localization in failing
expression. J Biol Chem 275:29685–29693, fibrillation. Circ Res 80:772–781, 1997. ventricular cardiomyocytes during recovery from
2000. 72. Brundel BJ: Alterations in potassium channel heart failure. Circ Heart Fail 5(3):357–365, 2012.
51. Mathur R, et al: A specific N-terminal residue in gene expression in atria of patients with persistent 92. Hong TT, et al: Plasma BIN1 correlates with
Kv1.5 is required for upregulation of the channel and paroxysmal atrial fibrillation: differential heart failure and predicts arrhythmia in patients
by SAP97. Biochem Biophys Res Commun regulation of protein and mRNA levels for K+ with arrhythmogenic right ventricular cardiomy-
342:1–8, 2006. channels. J Am Coll Cardiol 37:926–932, 2001. opathy. Heart Rhythm 9:961–967, 2012.
52. Murata M, et al: SAP97 interacts with Kv1.5 73. Michelakis ED, Weir EK: The pathobiology of 93. Sen-Chowdhry S, Morgan RD, Chambers JC, et
in heterologous expression systems. Am J pulmonary hypertension. Smooth muscle cells al: Arrhythmogenic cardiomyopathy: etiology,
Physiol Heart Circ Physiol 281:H2575–H2584, and ion channels. Clin Chest Med 22:419–432, diagnosis, and treatment. Annu Rev Med 61:233–
2001. 2001. 253, 2010.
53. Li H, Guo W, Mellor RL, et al: KChIP2 modu- 74. Tanabe Y, et al: Over-expression of Kv1.5 in rat 94. Schumacher SM, et al: Antiarrhythmic drug-
lates the cell surface expression of Kv 1.5-encoded cardiomyocytes extremely shortens the duration induced internalization of the atrial-specific k+
K(+) channels. J Mol Cell Cardiol 39:121–132, of the action potential and causes rapid excitation. channel kv1.5. Circ Res 104:1390–1398, 2009.
2005. Biochem Biophys Res Commun 345:1116–1121, 95. van der Heyden MA, Smits ME, Vos MA: Drugs
54. McEwen DP, Li Q, Jackson S, et al: Caveolin 2006. and trafficking of ion channels: a new pro-
regulates kv1.5 trafficking to cholesterol-rich 75. London B, et al: Long QT and ventricular arrhythmic threat on the horizon? Br J Pharmacol
membrane microdomains. Mol Pharmacol arrhythmias in transgenic mice expressing the N 153:406–409, 2008.
73:678–685, 2007. terminus and first transmembrane segment of a 96. Ficker E, Dennis A, Kuryshev Y, et al: HERG
55. Abi-Char J, et al: Membrane cholesterol modu- voltage-gated potassium channel. Proc Natl Acad channel trafficking. Novartis Found Symp
lates Kv1.5 potassium channel distribution and Sci U S A 95:2926–2931, 1998. 266:57–69, discussion 70–54, 95–59, 2005.
function in rat cardiomyocytes. J Physiol 76. Green CR, Severs NJ: Robert Feulgen Prize 97. Ficker E, Obejero-Paz CA, Zhao S, et al: The
582:1205–1217, 2007. Lecture. Distribution and role of gap junctions in binding site for channel blockers that rescue
misprocessed human long QT syndrome type 2 99. Walker VE, Atanasiu R, Lam H, et al: 101. Slawsky MT, Castle NA: K+ channel blocking
ether-a-gogo-related gene (HERG) mutations. Co-chaperone FKBP38 promotes HERG traf- actions of flecainide compared with those of
J Biol Chem 277:4989–4998, 2002. ficking. J Biol Chem 282:23509–23516, 2007. propafenone and quinidine in adult rat ventricular
98. Ficker E, Dennis AT, Wang L, et al: Role of the 100. Colatsky TJ: Mechanisms of action of lidocaine myocytes. J Pharmacol Exp Ther 269:66–74,
cytosolic chaperones Hsp70 and Hsp90 in matu- and quinidine on action potential duration in 1994.
ration of the cardiac potassium channel HERG. rabbit cardiac Purkinje fibers. An effect on steady
Circ Res 92:e87–e100, 2003. state sodium currents? Circ Res 50:17–27, 1982.
Macromolecular Complexes
and Regulation of the
Sodium Channel Nav1.5 18
Hugues Abriel
cardiomyocytes.4 Recent studies have presented evidence sup-

CHAPTER OUTLINE
porting this multiple-pool model of Nav1.5. Roden’s group inves-
Nav1.5 and Interacting Proteins 179 tigated a knock-in mouse model harboring the p.D1275N
mutation in the gene SCN5A, which codes for Nav1.5 and is
Localization of Nav1.5 in Cardiac Cells: Evidence
found in patients with dilated cardiomyopathy.12 They observed
for Distinct Pools 179 a marked reduction in the expression of Nav1.5 exclusively at the
Proteins Interacting with Nav1.5 without lateral membrane.12 More recently, Lin et al.13 obtained func-
Demonstrated Roles in Arrhythmias 179 tional evidence for two pools of cardiac Na+ channels with dif-
ferent biophysical properties by performing macropatch
Proteins Interacting with Nav1.5 That Are Linked experiments at different locations in cardiac cells. It is very pos-
to Cardiac Arrhythmias 184 sible that this two-pool expression model of Nav1.5 is an over-
Conclusions and Perspectives 186 simplification because there is both functional14 and morphologic
evidence6,15 of a T-tubular population of Nav1.5 (Figure 18-3). In
a recent paper, Nav1.5 was found to be colocalized with SAP97
at the T-tubules.16 Our recent findings suggest that this third pool
Nav1.5 and Interacting Proteins does not depend on the syntrophin/dystrophin complex because
it is well distinguished in dystrophin-deficient (mdx) myocytes
Among the many ionic currents known to be involved in the (see Figure 18-3).
genesis of the cardiac action potential (AP), the Na+ current (INa) The relationship between the expression of Nav1.5 at the
has been studied for more than 50 years1 and remains a major intercalated discs and other disc proteins has recently been
focus of research. The basic structure and function of the cardiac studied. It was observed that the intercalated disc pool of Nav1.5
Na+ channel Nav1.5 and its central role in cardiac pathologies are is dependent on the expression of key proteins known to be well
covered in Chapters 1, 9, and 50 of this book. This chapter sum- expressed at the discs, such as connexin43,17,18 plakophilin2,19-21
marizes recently published data on the proteins interacting with and desmoglein2.22 Thus, Nav1.5 not only interacts with many
Nav1.5 that form distinct macromolecular and multiprotein com- different partners, but its localization in different cellular and
plexes. The roles of the four β-subunits are not reviewed because membrane compartments is also very diverse. One of the obvious
that topic is covered in other chapters of this book. Over the past conclusions of these observations is that there is not one cardiac
several years, Nav1.5 has been shown to interact with a growing Na+ channel Nav1.5, but rather a multiplicity of them with vari-
list of regulatory proteins (Figure 18-1, A and B; Table 18-1). able functions and regulatory mechanisms.
The genes coding for several of these interacting proteins have
been found in patients with inherited arrhythmias, such as con-
genital long QT syndrome (LQTS)2 and Brugada syndrome
(BrS).3 The proteins interacting with Nav1.5 have been classified Proteins Interacting with Nav1.5 Without
as (1) anchoring/adaptor proteins involved in trafficking, target- Demonstrated Roles in Arrhythmias
ing, and anchoring the channel protein to specific membrane
compartments; (2) enzymes interacting with and modifying the This section lists the proteins that interact with Nav1.5 that cur-
channel structure via posttranslational modifications such as rently have no demonstrable pathologic roles. These interacting
protein kinases or ubiquitin ligases; and (3) proteins modulating proteins were discovered by either performing protein-protein
the biophysical properties of Nav1.5 upon binding (see Table interaction screens, such as yeast two-hybrid assays, or by using
18-1). These classifications are not mutually exclusive. proteomic-based protein identification assays. The sites of inter-
action, often protein-protein interaction domains, were mapped
on the sequence of Nav1.5 as described in Figure 18-1, A. This
list does not follow any specific logic related to importance but
Localization of Nav1.5 in Cardiac Cells: is dictated by the chronologic order of the published studies.
Evidence for Distinct Pools
Immunofluorescence staining experiments have shown that Ubiquitin-Protein Ligases of the Nedd4/
Nav1.5 is expressed in distinct membrane compartments (i.e., the Nedd4-like Family
intercalated discs and the “lateral membrane” of cardiac cells)
(Figure 18-2).4-6 This notion has been under debate because Ubiquitylation of target proteins tags them for proteasome- or
earlier studies described a quasi-exclusive localization of Nav1.5 lysosome-dependent degradation23 but also serves multiple non-
at the intercalated discs.7,8 Our group proposed a model of Nav1.5 degradative functions, in particular the trafficking of membrane
localization at two distinct pools based on strong molecular and proteins.24 Ubiquitin is a small protein of 76 amino acids found
in vivo evidence showing that Nav1.5 belongs to the dystrophin in all animal cells.25 It binds covalently to lysine residues of target
multiprotein complex.9 Dystrophin and syntrophin proteins are proteins. This protein ubiquitylation is performed by E3
not expressed at the intercalated discs of human,10 rat,11 or mouse ubiquitin-protein ligases.25 Ubiquitylated membrane proteins are
179
NH2 DI DII DIII DIV
+ + + +
+ + + +
β 12 3+ 5 6 12 3 + 5 6 12 3 + 5 6 12 3 +5 6
+ + + +
+ + + +
+ + + +
14-3-3η FGF13
IFM
CaMKIIδc
COOH
α-actinin-2
PTP-H1
MOG1 Calmodulin
Plakophilin-2 ?
NH2
Ankyrin-G nedd4-like Desmoglein-2 ?
Syntrophin COOH
Dystrophin Caveolin 3 ?
Utrophin
Telethonin ?
SAP97
A B
GPD1-L ?
Figure 18-1. Schematic representation of the α- and β-subunits of the Nav1.5 and interacting proteins. A, The predicted membrane topology of the α-subunit of Nav1.5
is illustrated. DI-DIV indicates the four homologous domains of the α-subunit; segments S5 and S6 are the pore-lining segments and the S4 helices (yellow) serve as voltage
sensors. In the connecting loop between DIII and DIV, the IFM red symbol presents the three residues (isoleucine, phenylalanine, and methionine) known to play a key role
in the fast inactivation process. The two intracellular loops display many sites for phosphorylation, whereas the C-terminal domain bears many protein-protein interaction
motifs. Only one β-subunit of the four known to be able to interact with sodium channels is shown (purple = transmembrane domain protein on the left side).The eleven
proteins, in addition to the β-subunits, are reported to interact with Nav1.5, for which a binding site to one of the intracellular domains is represented schematically. B, For
five proteins, the interaction has been shown only by coimmunoprecipitation experiments, but the binding sites are still unknown (question marks). These interactions may
be indirect (i.e., requiring adaptor proteins).
Nav1.5 pan-syntrophin merge
Figure 18-2. Stainings of cardiac sections leading to the concept of multiple pools of Nav1.5 channels in cardiac cells. Sections of mouse ventricular myocardium with
anti-Nav1.5 staining (left panel), anti-pan-syntrophin staining (middle panel), and overlay (right panel, merge). From these stainings, it is clear that syntrophin proteins are not
expressed at the intercalated discs (white arrowhead) where Nav1.5 is present. This defines the intercalated disc pool of Nav1.5 that has been proposed to interact with
SAP97 (see Figure 1, A, and Figure 4). The yellow arrowhead shows the lateral membrane pool of Nav1.5 colocalized with syntrophin proteins; white bar = 25 µm.
subsequently internalized and can be targeted for lysosomal or Nedd4-2–mediated ubiquitylation.34 Our group investigated the
proteasomal degradation. Alternatively, they can also be deubiq- molecular determinants of this regulation35 and found that the
uitylated by specific proteases and recycled back to the mem- ubiquitin-protein ligase Nedd4-2 binds directly to the PY motif
brane.26,27 Many different ion channels have been reported of Nav1.5 and ubiquitylates the channel in mammalian cells.
recently to be regulated by the Nedd4-like family of E3 ubiquitin- Because there was no reduction in the total level of Nav1.5
protein ligases. Nedd4-like enzymes bind specifically to target protein upon coexpression with Nedd4-2, we concluded that this
proteins that have consensus domains known as PY motifs with was most likely caused by an increased internalization rate rather
the sequence [L/P]PxY.28 Such PY motifs are found in the than by Nav1.5 degradation.29 Inhibition of the proteasome has
C-termini of all voltage-gated Na+ channels (with the exception been linked recently to an increase in INa and Nav1.5 expression
of Nav1.4, Nav1.9, and Nax29,30), as well as in other cardiac ion in neonatal rat cardiomyocytes.36 Ubiquitylated Nav1.5 was found
channels.27,31,32 Nedd4/Nedd4-like enzymes harbor several WW to be present in mouse cardiac tissue,35 further suggesting that
domains33 that can interact with these PY motifs. When expressed membrane turnover or the stability of Nav channels can be regu-
in Xenopus oocytes, Nav1.5-mediated INa is decreased by lated in vivo via ubiquitylation. Nine of such Nedd4-like E3
MACROMOLECULAR COMPLEXES AND REGULATION OF THE SODIUM CHANNEL NaV1.5 181
Table 18-1 The 16 Proteins (or Families of Proteins) Reported to Interact and Regulate Nav1.5
Protein
Type of
Protein Main Effects on Nav1.5
Mutated in
Cardiac Disorder
Interaction Domain
on Nav1.5
UniProt
Reference References
18
Ankyrin-G Anchoring- Trafficking and anchoring Nav1.5 binding VPIAxxSD motif in Q12955 15, 114
adaptor to the cell membrane site mutated in intracellular loop (ANK3_HUMAN)
one BrS patient DII-III
Syntrophin Anchoring- Adaptation to dystrophin Gene mutated in PDZ domain–binding Q13424 9, 97, 98
proteins adaptor and utrophin complex and LQTS and SIDS motif in C-terminus (SNTA1_HUMAN)
stabilization at the lateral
membrane of the
myocytes
MOG1 Anchoring- Involved in trafficking of Gene mutated in Intracellular loop Q9HD47 106, 107
adaptor Nav1.5 by unknown BrS between DII-III (MOG1_HUMAN)
mechanisms
Desmoglein2 Anchoring- Overexpression of ARVC Gene mutated in Not determined Q14126 22
adaptor mutants in mice reduces ARVC (DSG2_HUMAN)
Ina
α–Actinin-2 Anchoring- Involved in trafficking of n/a Intracellular loop P35609 84
adaptor Nav1.5 by unknown between DIII-IV (ACTN2_HUMAN)
mechanisms
SAP97 Anchoring- Trafficking and anchoring n/a PDZ domain–binding Q12959 4
adaptor to the cell membrane by motif in C-terminus (DLG1_HUMAN)
unknown mechanisms
Plakophilin2 Anchoring Silencing reduces INa, Gene mutated in Not determined Q99959 19–21
and alteration steady-state inactivation ARVC (PKP2_HUMAN)
of biophysical negatively, and slows
properties recovery from inactivation
14-3-3 h (eta) Alteration of Modulation of steady-state n/a Intracellular loop Q04917 40
biophysical inactivation between DI-II (1433F_HUMAN)
properties
Caveolin-3 Alteration of Mutant of caveolin-3 Gene mutated in Not determined P56539 92, 93
biophysical induces persistent current LQTS and SIDS (CAV3_HUMAN)
properties
Calmodulin Alteration of Many discrepant effects n/a IQ-motif (1900-1920) P62158 53–56, 60
biophysical but may confer in C-terminus and (CALM_HUMAN)
properties intracellular calcium loop between DIII-IV
sensitivity to Nav1.5
FGF proteins Alteration of Modulation of steady-state n/a Residues 1773-1832 Q92913 42, 46, 47
biophysical inactivation, recovery from in C-terminus (FGF13_HUMAN)
properties inactivation, and density at
cell membrane
Telethonin Alteration of Modulation of voltage Gene mutated in Not determined O15273 81
biophysical dependence of activation patient with GI (TELT_HUMAN)
properties disorder
GPD1-L Alteration of Loss-of-function variants Gene mutated in Not determined Q8N335 102–104
biophysical reduce INa by modulating BrS (GPDlL_HUMAN)
properties PKC-dependent
phosphorylation of Nav1.5
Nedd4-like E3 Enzyme Ubiquitylation and n/a PY-motif in Q96PU5 29, 35
ubiquitin ligases internalization C-terminus (NED4L_HUMAN)
Calmodulin Enzyme Phosphorylation of n/a Intracellular loop Q13557 67, 69, 70
kinase II dc residues in intracellular between DI-II and via (KCC2D_HUMAN)
loop I and modulation of βIV-spectrin
biophysical activity
Protein-tyrosine- Enzyme Phosphorylation (site n/a PDZ domain–binding P26045 76
phosphatase-H1 unknown) and modulation motif in C-terminus (PTN3_HUMAN)
of biophysical activity
This table does not take into account the four β sodium channel subunits.
LQTS, long QT syndrome; SIDS, sudden infant death syndrome; BrS, Brugada syndrome; ARVC, arrhythmogenic right ventricular cardiomyopathy; FGF, fibroblast growth
factor; GI, gastrointestinal; PKC, protein kinase C.
WT
Nav 1.5 Dystrophin WT
20 µm
mdx
mdx
5 nA
2 ms
A B
Figure 18-3. Reduction of Nav1.5 and sodium current in dystrophin-deficient mouse cardiomyocytes. A, Isolated mouse cardiomyocytes from wild type and mdx
(dystrophin-deficient) mice with Nav1.5 staining (green) and dystrophin staining (red). Whereas Nav1.5 channels are found at the intercalated discs and at the lateral mem-
brane compartment, it is apparent that dystrophin is excluded from the intercalated discs. The Nav1.5 pool at the discs has been shown to colocalize with SAP97. B, Sodium
current recordings from whole-cell patch-clamp experiments of freshly isolated wild type and mdx mouse cardiomyocytes. It can be inferred that the reduction in current
is caused by the specific loss of Nav1.5 channels from the lateral membrane pool.
(A, From Petitprez S, Zmoos AF, Ogrodnik J, et al: SAP97 and dystrophin macromolecular complexes determine two pools of cardiac sodium channels Nav1.5 in cardiomyocytes.
Circ Res 108:294–304, 2011; with permission from Wolters Kluwer Health. B, Modified from Gavillet B, Rougier JS, Domenighetti AA, et al: Cardiac sodium channel Nav1.5 is regu-
lated by a multiprotein complex composed of syntrophins and dystrophin. Circ Res 18:407–414, 2006; with permission from Wolters Kluwer Health.)
enzymes are present in the human genome,37 at least eight of The FGF homologous factors FHF3, FHF1, FHF2, and FHF4
which are expressed at the RNA level in human cardiac tissue.31 are also called FGF11, FGF12, FGF13, and FGF14, respectively.
It remains to be determined which of these Nedd4/Nedd4-like One of the best characterized functions of the FHFs is their
proteins regulates Nav1.5 in a physiologic context and whether ability to bind to and modulate voltage-gated Na+ channels. Liu
the PY motif plays a role in cardiac disease. Altogether these et al.42 were the first to demonstrate that FGF12 interacts with
results suggest that the ubiquitin-proteasome system is involved the proximal portion of the C-terminus of Nav1.5 (see Figure
in several aspects of Nav1.5 regulation, the intricacies of which 18-1, A). This finding was then confirmed by Goetz et al.,43 who
remain to be elucidated. also showed that FGF11 and FGF14 interact with the same
domain of Nav1.5. In HEK293 cells, coexpression of FGF12 with
Nav1.5 shifted the steady-state inactivation relationship toward
14-3-3η (eta) Protein hyperpolarized values without affecting the other parameters
studied.42 Interestingly, several LQTS type 3 and BrS mutations
The 14-3-3 protein family is composed of dimeric cytosolic are located in the domain that interacts with FHF1B (see Table
adaptor ubiquitous proteins.38 The members of this family are 18-1). The SCN5A mutation D1790G44 disrupted the binding of
involved in many cellular functions such as the binding and regu- this protein with Nav1.5 and abolished the FHF 12-induced shift
lation of trafficking of various membrane proteins.39 Allouis et al. of steady-state inactivation. Another member of the FGF family,
performed yeast two-hybrid and coimmunoprecipitation experi- FGF14, has been shown to downregulate the activity of Nav1.5.45
ments40 showing that the isoform 14-3-3η interacts with the Because this regulatory protein is expressed in the brain, and not
N-terminal part of the intracellular loop linking domains I to II in the heart,45 the significance of this interaction is likely restricted
(see Figure 18-1, A and Table 18-1). Furthermore, it was observed to the central nervous system. A recent study46 demonstrated that
that 14-3-3 and Nav1.5 are colocalized at the intercalated discs FGF13 is expressed in mouse cardiac myocytes. FGF13 was
of myocytes. No influence on the Nav1.5-mediated peak current shown to directly bind to and colocalize with Nav1.5 in mouse
was observed when Nav1.5 and 14-3-3η were coexpressed in COS cardiac cells. Reducing the expression of FGF13 decreased the
cells, suggesting that this protein does not influence Nav1.5 traf- Na+ current, shifted the availability curve toward more negative
ficking. In this expression system, 14-3-3 shifted the inactivation potentials, and slowed recovery from inactivation. Furthermore,
curve toward negative potentials and delayed recovery from inac- decreased membrane expression of Nav1.5 was observed in cells
tivation, illustrating that 14-3-3 proteins are able to modify the with reduced expression of FGF13. These results underscore the
biophysical properties of ion channels. Because different isoforms specific role of FGF13 in modulating Nav1.5 function and suggest
of 14-3-3 proteins are expressed in cardiac cells,40 their exact roles that this gene may be a susceptibility gene for cardiac arrhyth-
in normal cardiac function and their implications in disease states mias. A crystal structure of the Nav1.5 C-terminal domain com-
require further investigation. plexed with FGF13 and calmodulin was recently published.47 The
functional relationship between FGFs and calmodulin remains to
be studied in more detail.
Fibroblast Growth Factor Homologous Factors
Although fibroblast growth factor homologous factor (FHF) Calmodulin
family members are highly homologous in sequence and struc-
ture to fibroblast growth factors (FGF), their functional roles are Intracellular Ca2+ has been shown to modulate the function of
quite different. FHFs are found in the cytosol because their many ion channels, including the voltage-gated Na+ channels.48,49
N-termini lack signal sequences that are required for secretion.41 Many cardiac ion channels use calmodulin (CaM), a ubiquitous
intracellular Ca2+-binding protein involved in many different cel- phosphorylation of Tyr-1495.74 This residue is close to the Ile-
18
lular processes,50 as a Ca2+-sensing partner. The Nav1.5 C-terminal Phe-Met cluster (IFM) in the intracellular loop linking domains
domain has an IQ motif with a consensus sequence of IQxxxRxxxxR III-IV (see Figure 18-1, A), which is known to mediate the rapid
(see Table 18-1), which is very similar to that found in voltage- inactivation process of voltage-gated Na+ channels.75 Fyn coex-
gated calcium channels.51 The IQ motif is also found in all iso- pression in HEK293 cells shifted the steady-state inactivation
forms of the Nav family.52 Several studies47,53-55 have shown a curve toward depolarized potentials and accelerated the recovery
direct interaction between CaM and the IQ motif of Nav1.5. from Nav1.5 inactivation.74 The site of interaction of Fyn with
Recent work by Wang et al. described a crystal structure with an Nav1.5 remains to be investigated. We reported76 that the protein
IQ motif ternary complex of Nav1.5, FGF13, and CaM. The tyrosine phosphatase, PTPH1, interacts with the PDZ-domain
functional consequences of this CaM-Nav1.5 interaction are con- binding motif of Nav1.5, which is also known to interact with two
troversial. Several studies52,56,57 have shown inconsistent results other PDZ-bearing proteins, syntrophin9 and SAP97.4 PTPH1
that have been difficult to reconcile. A few groups53,55-57 reported coexpression in HEK293 cells shifted the availability curve of
that the voltage dependence and stability of the inactivated state wild-type Nav1.5 toward hyperpolarized potentials. This effect
were dependent on CaM and the IQ motif. In addition, it has was abolished when the PDZ domain–binding motif of the
been proposed that CaM may not be the only sensor for the Nav1.5 C-terminus was removed by an early stop mutation.
Ca2+-dependent regulation of Nav1.5, because the proximal part These results suggest that tyrosine phosphorylation of Nav1.5
of the Nav1.5 C-terminal domain is similar to an EF-hand motif, modulates the stability of the inactivated state. The fact that
which is known to bind Ca2+.58,59 Whether this domain binds several proteins (e.g., PTPH1, syntrophin proteins and SAP97)
intracellular calcium is still under debate.47,60-62 The role of intra- interact with the same binding domain supports the coexistence
cellular Ca2+ in Nav1.5 regulation was recently investigated by of different multiprotein complexes with Nav1.5 in cardiac cells.
Casini et al.63 They demonstrated that Nav1.5 single-channel
conductance is decreased (most likely directly) with an increase
in the intracellular Ca2+ concentration. Intracellular Ca2+ appears Telethonin
to be an important regulator of Nav1.5 function, but more studies
are needed to clarify its role. Telethonin is a small 19-kDa protein expressed in striated muscle
cells, including cardiac cells. Mutations in the gene coding for
telethonin, TCAP, lead to hypertrophic and dilated cardiomyopa-
Ca2+/Calmodulin-dependent Protein Kinase II thy and limb-girdle muscular dystrophy.77,78 This protein has
been shown to interact with the sarcomeric protein titin,79 as well
Ca2+/calmodulin-dependent protein kinases II (CaMKII) are as the β-subunit of the KCNQ1 channel (KCNE1), which medi-
serine/threonine protein kinases expressed in many cell types, ates the cardiac IKs current.80 Telethonin can be coprecipitated
where they transduce intracellular Ca2+ increases into the phos- with Nav1.5 from mouse cardiac tissue, and the two proteins
phorylation of target proteins, including cardiac ion channels.64 colocalize in cardiomyocytes.81 The site of interaction on Nav1.5
CaMKIIδc is the predominant cardiac isoform upregulated in has not yet been mapped. When the endogenous expression of
human and animal heart failure models.65,66 Nav1.5 has been telethonin in HEK293 cells was reduced, the voltage-dependent
found to colocalize and coimmunoprecipitate with CaMKIIδc.67-69 activation of the Nav1.5-mediated current shifted toward positive
Ashpole et al.69 have shown that CaMKIIδc interacts with the potential values. In addition, coexpression experiments of tele-
first intracellular loop of Nav1.5 (see Figure 18-1) and that the thonin and Nav1.5 in HEK293 cells altered several of the kinetic
residues Ser-516 and Thr-594 may be phosphorylated by this properties of the channel.81 Furthermore, a point mutation in
kinase. Hund et al.70 found that Ser-571 was also a target of TCAP was found in a patient with abnormal gut motility,81 con-
CaMKII within the same intracellular loop of Nav1.5, and that sistent with the fact that Nav1.5 and telethonin are expressed in
this regulation was dependent on the correct expression of βIV- the neurons of the intestinal mucosa.82 The precise role of tele-
spectrin, which interacts with ankyrin-G at the intercalated discs. thonin in the regulation Nav1.5 in the heart (and other tissues)
Overexpression of the CaMKIIδc enzyme in rabbit myocytes and remains to be clarified.
transgenic mice induced a Ca2+-dependent hyperpolarizing
shift of the steady-state inactivation curve, slowed the recovery
from inactivation, and increased the persistent INa. These α–Actinin-2
biophysical property alterations of INa are similar to some
congenital LQTS type 3 mutation-dependent modifications.71 α–Actinin-2 is a protein of the F actin–cross-linking protein
In mice, transgenic overexpression of CaMKIIδc leads to family (similar to spectrin and dystrophin) that interacts with and
chronic heart failure and episodes of ventricular tachycardia. In regulates the trafficking of several potassium channels,83 in par-
dog ventricular myocytes, it has also been shown that ticular Kv1.5. A recent study used84 pull-down experiments to
CaMKII activity increases the persistent/late INa,72 but it is show that α–actinin-2 interacts with the intracellular III-IV loop
unclear whether these arrhythmias are the direct consequence of of Nav1.5. Coexpression of α–actinin-2 in tsA201 cells increased
the Nav1.5 biophysical alterations or are related to other heart the measured INa densities as well as the Nav1.5 protein at the cell
failure mechanisms. Nevertheless, CaMKII is an important com- surface. Colocalization of Nav1.5 and α–actinin-2 was found
ponent in the pathogenesis of arrhythmias and potentially a new mainly at the lateral membrane at the level of the Z-lines. α–
drug target.64 actinin-2 seems to be an alternative to the dystrophin/syntrophin
complex and ankyrin-g for the connection of Nav1.5 to the cyto-
skeletal network. Whether these anchoring proteins have over-
Proteins Tyrosine Phosphatase PTPH1 lapping or clearly distinct functions remains to be investigated.
Ion channels are also regulated by the phosphorylation of tyro-

sine residues, a process that depends on both the phosphorylation Synapse-associated Protein 97
activity of tyrosine protein kinases and the dephosphorylation
activity of phosphatases.73 It was demonstrated that the overex- Proteins of the membrane-associated guanylate kinase (MAGUK)
pression of the protein tyrosine kinase Fyn in HEK293 cells family are expressed mainly at the cell-cell junctions and are
altered several of the biophysical properties of Nav1.5 via the characterized by numerous protein-protein interaction domains,
Lateral membrane Nav 1.5 Intercalated disc Nav 1.5
NH2 COOH NH2 COOH
Ankyrin-G Ankyrin-G
PDZ
Dystrophin Syntrophin Intermediate protein?
PDZ
PDZ
PDZ
SAP97
Utrophin Plakophilin-2
Desmoglein-2
Figure 18-4. Schematic presentation of the proteins defining the two proposed pools of Nav1.5. The findings of our group have led to the concept of two distinct mac-
romolecular complexes with Nav1.5: (left) at the lateral membrane with the dystrophin/syntrophin complex and (right) at the intercalated discs with the MAGUK protein
SAP97. Note that when dystrophin is absent (as in mdx mouse myocytes), utrophin may also interact with Nav1.5 (see Figure 18-1, A).
including PDZ domains.85 The MAGUK proteins regulate the muscle cells, including cardiomyocytes. Mutations in CAV3 have
function and localization of many membrane proteins (including been linked to limb-girdle muscular dystrophy, rippling muscle
ion channels) in neurons, epithelial cells, and cardiomyocytes. disease, and familial hypertrophic cardiomyopathy.91 CAV3 has
Synapse-associated protein 97 (SAP97) and zonula occludens 1 also recently been shown to be mutated in patients with congeni-
are the predominant MAGUK proteins expressed in cardiac tal LQTS type 992 and sudden infant death syndrome (SIDS).93
cells.86 SAP97 regulates the targeting, localization and function Caveolin-3 was coimmunoprecipitated with Nav1.5 in rat cardiac
of cardiac K+ channels, such as Kir2.x,87 Kv1.5,88 and Kv4.x89 via tissue94 and HEK293 cells.92 The site of interaction between
their PDZ domain–binding motifs located in the C-termini. We caveolin-3 and Nav1.5 has not yet been determined (see Figure
recently obtained evidence for the coexistence of at least two 18-1, B). Immunohistostaining of cardiac cells demonstrated that
pools of Nav1.5 channels in cardiomyocytes: one located at the these two proteins are mainly colocalized at the lateral mem-
lateral membrane with the dystrophin multiprotein complex, and brane.92,94,95 Dystrophin is also a component of caveolae,96 sug-
the other with SAP97 at the intercalated discs (Figure 18-4).4 We gesting that the interaction between caveolin-3 and Nav1.5 could
demonstrated that the interaction of Nav1.5 and SAP97 was be indirect via proteins of the dystrophin multiprotein complex.9
dependent on the PDZ domain–binding motif that most likely The precise role of the Nav1.5/caveolin-3 interaction in normal
interacts with one of the PDZ domains of SAP97. Reduction of physiology remains to be clarified. The coexpression of Nav1.5
the expression of SAP97 in both HEK293 and atrial cardiac cells and the mutants of caveolin-3 found in patients with LQTS and
led to a decreased INa without modification of any biophysical SIDS in HEK293 cells was shown to increase the inward Na+-
properties, suggesting that SAP97 may play a role in controlling persistent current.92,93 In an earlier study, it was reported that
the density of Nav1.5 at the cell surface of cardiac cells. The β-adrenergic stimulation by isoproterenol led to a rapid increase
detailed mechanisms underlying this regulation have not yet been of peak INa in rat cardiac myocytes.94 This phenomenon is most
investigated. SAP97 expression has been found at the intercalated likely independent of protein kinase A, because a protein kinase A
discs4,88 and the T-tubule compartment of cardiac cells.16 Milstein inhibitor did not reduce this effect. The increase was, however,
et al.16 recently demonstrated that Nav1.5 expression is coregu- completely abolished by antibodies against caveolin-3. The
lated with the potassium channel Kir2.1, and both interact with precise molecular and cellular mechanisms underlying these
SAP97 (see Chapter 21 of this book). observations require further investigation.
α-1 Syntrophin and the Dystrophin/

Proteins Interacting With Nav1.5 That Are Utrophin Complex
Linked to Cardiac Arrhythmias
Earlier work has shown that Nav1.5 is part of the dystrophin mul-
This section summarizes the proteins interacting with Nav1.5 for tiprotein complex.97 Our group demonstrated that Nav1.5 inter-
which genes have been found to be mutated in patients with acts with dystrophin via adaptor syntrophin proteins.9 Similar to
cardiac arrhythmias. the binding with SAP97 and PTPH1, this interaction is depen-
dent on the C-terminal PDZ domain–binding motif of Nav1.5
(see Figure 18-1, A), which is composed of the last three residues
Caveolin-3 of the protein. In the hearts of dystrophin-deficient mice (mdx),
the best-studied animal model of Duchenne muscular dystrophy,
Caveolae are plasma membrane invaginations with an enrichment the protein level of Nav1.5 was decreased.9 The decreased expres-
of signaling molecules and ion channels.90 Caveolin proteins are sion of the Na+ channel resulted in reduced cellular INa and con-
important constituents of these caveolae. Caveolin-3, encoded by duction defects, which were reflected on the electrocardiogram by
the gene CAV3, is the predominant isoform expressed in striated a prolongation of the QRS complex duration, as well as
conduction slowing in optical mapping experiments.4 The reduced mRNA is well expressed in cardiac tissue.105 Wu et al.106 demon-
18
expression of the Nav1.5 protein could not be explained by a strated that MOG1 interacts with the intracellular loop between
decrease in the SCN5A mRNA level,9 suggesting that a defect in domains II and III of Nav1.5 (see Figure 18-1, A). This interac-
the translational process or a lack of dystrophin may reduce the tion was first described by performing a yeast two-hybrid screen,
stability of the Nav1.5 protein. Albesa et al.98 showed that utro- followed by pull-down and coimmunoprecipitation experi-
phin, a homologue protein of dystrophin, was upregulated in mdx ments.106 The two proteins were also shown to colocalize in
hearts and interacted with Nav1.5 similarly to dystrophin. Cardiac mouse ventricular cells, mostly at the intercalated discs. MOG1
cells deficient in both dystrophin and utrophin displayed a larger coexpression in HEK293 cells increased the Nav1.5-mediated
decrease in Nav1.5 expression and INa.98 Two missense mutations current without altering its biophysical properties, suggesting
in SNTA1, which encodes α1-syntrophin, have been described in that MOG1 is a cofactor for optimal channel expression at the
patients with congenital LQTS,99,100 further supporting the cell membrane. The role of MOG1 in regulating the surface
important role of this multiprotein complex in the regulation of expression of Nav1.5 was confirmed in a more recent study.107
Nav1.5. The SNTA1 mutation, p.A390V,99 was reported to disrupt This study described two genetic variants of MOG1 found in
a presently undescribed macromolecular complex comprising patients with BrS that led to reduced expression of Nav1.5 at the
neuronal nitric oxide synthase, plasma membrane Ca-ATPase cell membrane of rat atrial cardiomyocytes and a decreased
type 4b with syntrophin, and Nav1.5. The overexpression of the current. This work puts RANGRF on the increasing long list of
mutant syntrophin protein in cardiac myocytes increased the susceptibility genes for BrS and illustrates the heterogeneity of
persistent Na+ current, a finding consistent with the LQTS phe- mechanisms found in this syndrome.
notype in affected individuals. The authors also described
increased nitrosylation of Nav1.5 when mutant syntrophin was
coexpressed in HEK293 cells, suggesting that under basal condi- Plakophilin2
tions with wild-type syntrophin, Nav1.5 nitrosylation is low
because of the inhibition of neuronal nitric oxide synthase activity Plakophilin2 is a desmosomal protein found in the intercalated
by Ca-ATPase. The syntrophin mutation may disrupt this discs of cardiac cells. The human gene (PKP2), which encodes
complex and lead to increased nitrosylation of the channel, thus plakophilin2, is mutated in patients with arrhythmogenic right
increasing the Nav1.5-persistent current. Because syntrophin ventricular cardiomyopathy.108 By performing pull-down and
seems to be excluded from the intercalated discs,4 syntrophin- coimmunoprecipitation experiments, Delmar’s group19,20 demon-
dependent regulation of the persistent current is most likely strated that Nav1.5 interacts not only with plakophilin2 but also
exclusively related to the lateral pool of Nav1.5, suggesting that with ankyrin-G and connexin43.20 Whether the interactions
Nav1.5-dependent late currents at the lateral membrane may be between these different proteins of the intercalated discs are
different from those at the intercalated discs. direct or indirect, as well as the site of interaction with Nav1.5,
remains to be determined. It was, however, convincingly shown
that these proteins are colocalized at the intercalated discs, and
Glycerol-3-Phosphate Dehydrogenase-like Protein that silencing of plakophilin2 in cardiomyocytes reduces the INa
and alters some of its biophysical properties. The influence of
Mutations in the gene coding for Nav1.5, SCN5A, are found in plakophilin2 integrity on Nav1.5 was demonstrated in vivo by
approximately 20% of patients with BrS. Other possible causative studying a mouse model expressing only one allele of pla-
genes are still being investigated (see Chapter 92 ). London et al. kophilin2.21 Reduced expression of plakophilin2 decreased INa
described a locus on chromosome 3101 in a large family with BrS without changing the localization or total expression of Nav1.5
but simultaneously excluded SCN5A. A missense mutation of the in the mouse heart. These findings clearly support the notion of
gene coding for the glycerol-3-phosphate dehydrogenase-like the coregulation of desmosomal proteins and the Nav1.5 complex
protein (GPDlL) was later found.102 GPDlL is expressed in at the intercalated discs. The molecular and cellular mechanisms
cardiac tissue and coexpression experiments using HEK293 cells underlying these observations are still unclear.
showed that mutant GPDlL reduced the INa. Three more muta-
tions of the GPDlL gene have been found in infants who died of
SIDS.103 Expression of these variants in neonatal mouse cardio- Desmoglein2
myocytes decreased INa, suggesting that a proportion of SIDS
patients may have decreased INa, as is observed in BrS patients. The most recently described protein interacting with Nav1.5 is
Valdivia et al.104 observed an interaction between Nav1.5 and desmoglein2, another protein of the desmosomal complex of
GPDlL by performing pull-down experiments. The site of inter- cardiac cells found to be mutated in patients with arrhythmogenic
action has not yet been investigated. The mechanisms by which right ventricular cardiomyopathy (ARVC).109 Desmoglein2 was
the genetic variants of GPDlL reduce the Na+ current have been shown22 to coimmunoprecipitate with Nav1.5 in mouse cardiac
studied in expression systems.104 A pathway has been proposed tissue. The study of a transgenic mouse model overexpressing
whereby Ser-1503 can be phosphorylated by protein kinase C one mutant of desmoglein2 (p.N271S) revealed a reduction of
(PKC) and leads to reduced INa. It has been shown that the activ- the INa and electrical impulse propagation. There was, however,
ity of PKC depends on GPDlL function and that the mutant no reduction in Nav1.5 total expression. These results are very
GPDlL variants lead to a further decrease of the INa. These similar to those seen with the plakophilin2 mutants (discussed
observations linking the redox state of the cells with the activity earlier) and further support a cross-talk mechanism between the
of Nav1.5 are very interesting and should be investigated using desmosomal proteins and Nav1.5, whose molecular details remain
native cardiac cells and tissues. to be investigated.
Multicopy Suppressor of gsp1 (MOG1) Ankyrin-G

MOG1 is a small 29-kDa protein that is encoded by the RANGRF The ankyrin proteins organize, transport, and anchor membrane
gene. MOG1 binds to the GTP-binding nuclear protein Ran and proteins to the actin and spectrin cytoskeleton of cells.110 In the
is involved in regulating nuclear protein trafficking.105 MOG1 is human genome, ANK1-3 are the three genes encoding ankyrins.
found in the nucleus and the cytosol of cardiac cells, and its The expression of ankyrin-B (ANK2) and ankyrin-G (ANK3) has
been demonstrated in the heart.110 Mutations of ANK2 have been arrhythmias, clearly illustrating their important roles in the
linked to congenital LQTS type 4,111 sinoatrial node dysfunction, pathophysiology of disease. Most of the detailed molecular cel-
atrial fibrillation, conduction slowing, and sudden cardiac death, lular mechanisms involved in the regulation of Nav1.5 are still
the collection of which define “ankyrin-B cardiac syndrome.”112 very poorly understood, thus providing many challenges for
There is currently no evidence that Nav1.5 is directly regulated future studies in this field. It is clear that an important concept
by ankyrin-B, even though in ankyrin-B–deficient mouse cardiac is emerging from the reviewed findings. The cardiac Na+ channel
cells, Na+ channels display late openings similar to those seen in Nav1.5 is most likely part of several multiprotein macromolecular
Nav1.5 congenital LQTS type 3 mutant channels.113 On the other complexes that define multiple populations of Nav1.5 channels in
hand, ankyrin-G was shown to directly interact with the ankyrin- cardiac cells. These associated proteins may interact at different
binding motif of the linker loop between domains II and III of life cycle stages of the Nav1.5 subunit and in many different
Nav1.5114 (see Figure 18-1, A, and Table 18-1). Ankyrin-G is intracellular compartments. Nav1.5 may encounter hundreds of
predominantly located at the intercalated discs, where it interacts proteins during its lifespan, from biosynthesis until degradation.
with not only Nav1.5 but also the actin-associated protein βIV- Among the most intriguing questions that remain to be answered
spectrin and CaMKII.70 A clear T-tubular localization of are: (1) Where are the other pools of Nav1.5 located in cardiac
ankyrin-G has also been observed.70 A SCN5A mutation, cells, and what are their molecular determinants?; (2) What are
p.E1053K, in this motif was found in a BrS patient and was shown the specific roles and cross-talk mechanisms between the proteins
to disrupt the interaction between Nav1.5 and ankyrin-G.15 of the intercalated discs and Nav1.5?; and (3) What are the spe-
Expression of wild type and mutant Nav1.5 channels in adult rat cific functions of the different populations of Nav1.5 in cardiac
myocytes15 was achieved using viral vectors, and although wild cells?
type Nav1.5 channels were correctly transported to the interca-
lated discs and lateral membranes, the E1053K channels remained
in the cytoplasm of the transduced cells.15 These findings were
recently confirmed115 by silencing the expression of ankyrin-G in Acknowledgments
neonatal cardiomyocytes, which resulted in the reduction of
Nav1.5 expression and incorrect trafficking of most of the chan- The author thanks Dr. A. Felley for her useful comments on this
nels to the cell membrane. chapter and Dr. L. Gillet for his help in collecting information.
We also thank J. Gramling for the preparation of preliminary
versions of Figure 18-3 and 18-4.
This work was supported by Swiss National Science Founda-
Conclusions and Perspectives tion grant 310030B_135693 (to H.A.). The research leading to
part of the results of the research group of H. Abriel has received
This chapter summarizes the most recent findings related to the funding from the European Community’s Seventh Framework
rapidly growing list of Nav1.5-associated proteins. Some of these Program FP7/2007-2013 under grant agreement no.
proteins were found as mutated in patients with genetic forms of HEALTH-F2-2009-241526, EUTrigTreat.
10. Kaprielian RR, Stevenson S, Rothery SM, et al: 18. Desplantez T, McCain M, Beauchamp P, et al:
References Distinct patterns of dystrophin organization in Connexin 43 ablation in fetal atrial myocytes
myocyte sarcolemma and transverse tubules of decreases electrical coupling, partner connexins
1. Fozzard HA: Cardiac sodium and calcium chan- normal and diseased human myocardium. Circu- and sodium current. Cardiovasc Res 94(1):58–65,
nels: A history of excitatory currents. Cardiovasc lation 101(22):2586–2594, 2000. 2012.
Res 55(1):1–8, 2002. 11. Stevenson SA, Cullen MJ, Rothery S, et al: High- 19. Sato PY, Musa H, Coombs W, et al: Loss of
2. Cerrone M, Priori SG: Genetics of sudden death: resolution en-face visualization of the cardiomyo- plakophilin-2 expression leads to decreased
focus on inherited channelopathies. Eur Heart J cyte plasma membrane reveals distinctive sodium current and slower conduction velocity in
32(17):2109–2118, 2011. distributions of spectrin and dystrophin. Eur J cultured cardiac myocytes. Circ Res 105(6):523–
3. Mizusawa Y, Wilde AAM: Brugada syndrome. Cell Biol 84(12):961–971, 2005. 526, 2009.
Circ Arrhythmia Electrophysiol 5(3):606–616, 12. Watanabe H, Yang T, Stroud DM, et al: Striking 20. Sato PY, Coombs W, Lin X, et al: Interactions
2012. in vivo phenotype of a disease-associated human between ankyrin-G, plakophilin-2, and con-
4. Petitprez S, Zmoos AF, Ogrodnik J, et al: SAP97 SCN5A mutation producing minimal changes in nexin43 at the cardiac intercalated disc. Circ Res
and dystrophin macromolecular complexes deter- vitro. Circulation 124(9):1001–1011, 2011. 109(2):193–201, 2011.
mine two pools of cardiac sodium channels 13. Lin X, Liu N, Lu J, et al: Subcellular heterogene- 21. Cerrone M, Noorman M, Lin X, et al: Sodium
Nav1.5 in cardiomyocytes. Circ Res 108:294–304, ity of sodium current properties in adult cardiac current deficit and arrhythmogenesis in a murine
2011. ventricular myocytes. Heart Rhythm 8(12):1923– model of plakophilin-2 haploinsufficiency. Car-
5. Haufe V, Camacho JA, Dumaine R, et al: Expres- 1930, 2011. diovasc Res 95(4):460–468, 2012.
sion pattern of neuronal and skeletal muscle 14. Brette F, Orchard CH: Density and sub-cellular 22. Rizzo S, Lodder EM, Verkerk AO, et al:
voltage-gated Na+ channels in the developing distribution of cardiac and neuronal sodium Intercalated disc abnormalities, reduced Na+
mouse heart. J Physiol 564(3):683–696, 2005. channel isoforms in rat ventricular myocytes. current density and conduction slowing in
6. Leoni AL, Gavillet B, Rougier JS, et al: Variable Biochem Biophys Res Commun 348(3):1163– desmoglein-2 mutant mice prior to cardiomyop-
Na(v)1.5 protein expression from the wild-type 1166, 2006. athic changes. Cardiovasc Res 95(4):409–418,
allele correlates with the penetrance of cardiac 15. Mohler PJ, Rivolta I, Napolitano C, et al: Nav1.5 2012.
conduction disease in the Scn5a mouse model. E1053K mutation causing Brugada syndrome 23. Glickman MH, Ciechanover A: The ubiquitin-
PLoS ONE 5(2):e9298, 2010. blocks binding to ankyrin-G and expression of proteasome proteolytic pathway: destruction for
7. Maier SKG, Westenbroek RE, McCormick KA, Nav1.5 on the surface of cardiomyocytes. Proc the sake of construction. Physiol Rev 82(2):373–
et al: Distinct subcellular localization of different Natl Acad Sci U S A 101(50):17533–17538, 2004. 428, 2002.
sodium channel α and β subunits in single ven- 16. Milstein ML, Musa H, Balbuena DP, et al: 24. Staub O, Rotin D: Role of ubiquitylation in cel-
tricular myocytes from mouse heart. Circulation Dynamic reciprocity of sodium and potassium lular membrane transport. Physiol Rev 86(2):669–
109(11):1421–1427, 2004. channel expression in a macromolecular complex 707, 2006.
8. Kucera JP, Rohr S, Rudy Y: Localization of controls cardiac excitability and arrhythmia. Proc 25. Hershko A, Ciechanover A: The ubiquitin system.
sodium channels in intercalated disks modulates Natl Acad Sci U S A 109(31):E2134–E2143, 2012. Annu Rev Biochem 67:425–479, 1998.
cardiac conduction. Circ Res 91(12):1176–1182, 17. Jansen JA, Noorman M, Musa H, et al: Reduced 26. Hicke L, Dunn R: Regulation of membrane
2002. heterogeneous expression of Cx43 results in protein transport by ubiquitin and ubiquitin-
9. Gavillet B, Rougier JS, Domenighetti AA, et al: decreased Nav1.5 expression and reduced sodium binding proteins. Annu Rev Cell Dev Biol
Cardiac sodium channel Nav1.5 is regulated by a current which accounts for arrhythmia vulnerabil- 19:141–172, 2003.
multiprotein complex composed of syntrophins ity in conditional Cx43 knockout mice. Heart 27. Abriel H, Staub O: Ubiquitylation of ion chan-
and dystrophin. Circ Res 18(99):407–414, 2006. Rhythm 9(4):600–607, 2011. nels. Physiology 20(6):398–407, 2005.
28. Rotin D, Kumar S: Physiological functions of the 49. Pitt GS: Calmodulin and CaMKII as molecular 71. Kass RS: The channelopathies: Novel insights
HECT family of ubiquitin ligases. Nat Rev Mol switches for cardiac ion channels. Cardiovasc Res into molecular and genetic mechanisms of human
Cell Biol 10(6):398–409, 2009.
29. Rougier J-S, van Bemmelen MX, Bruce MC,
et al: Molecular determinants of voltage-gated
73(4):641–647, 2007.
50. Chin D, Means AR: Calmodulin: A prototypical
calcium sensor. Trends Cell Biol 10(8):322–328,
disease. J Clin Invest 115(8):1986–1989, 2005.
72. Maltsev VA, Reznikov V, Undrovinas NA, et al:
Modulation of the late sodium current by Ca2+,
18
sodium channel regulation by the Nedd4/Nedd4- 2000. calmodulin, and CaMKI I in normal and failing
like proteins. Am J Physiol Cell Physiol 51. Pitt GS, Zuhlke RD, Hudmon A, et al: Molecular dog cardiomyocytes: Similarities and differences.
288(3):C692–C701, 2005. basis of calmodulin tethering and Ca2+-depen- Am J Physiol Heart Circ Physiol 294(4):1597–
30. Fotia AB, Ekberg J, Adams DJ, et al: Regulation dent inactivation of L-type Ca2+ channels. J Biol 1608, 2008.
of neuronal voltage-gated sodium channels by the Chem 276(33):30794–30802, 2001. 73. Davis MJ, Wu X, Nurkiewicz TR, et al: Regula-
ubiquitin-protein ligases Nedd4 and Nedd4-2. 52. Herzog RI, Liu C, Waxman SG, et al: Calmodulin tion of ion channels by protein tyrosine phos-
J Biol Chem 279(28):28930–28935, 2004. binds to the C terminus of sodium channels phorylation. Am J Physiol Heart Circ Physiol
31. Jespersen T, Membrez M, Nicolas CS, et al: The Nav1.4 and Nav1.6 and differentially modulates 281(5):H1835–H1862, 2001.
KCNQ1 potassium channel is down-regulated by their functional properties. J Neurosci 74. Ahern CA, Zhang JF, Wookalis MJ, et al: Modula-
ubiquitylating enzymes of the Nedd4/Nedd4like 23(23):8261–8270, 2003. tion of the cardiac sodium channel NaV1.5 by
family. Cardiovasc Res 74:64–74, 2007. 53. Tan HL, Kupershmidt S, Zhang R, et al: A Fyn, a Src family tyrosine kinase. Circ Res
32. Albesa M, Grilo LS, Gavillet B, et al: Nedd4-2- calcium sensor in the sodium channel modulates 96(9):991–998, 2005.
dependent ubiquitylation and regulation of the cardiac excitability. Nature 415(6870):442–447, 75. Hartmann HA, Tiedeman AA, Chen SF, et al:
cardiac potassium channel hERG1. J Mol Cell 2002. Effects of III-IV linker mutations on human heart
Cardiol 51(1):90–98, 2011. 54. Deschenes I, Neyroud N, DiSilvestre D, et al: Na+ channel inactivation gating. Circ Res
33. Staub O, Rotin D: WW domains. Structure Isoform-specific modulation of voltage-gated 75(1):114–122, 1994.
4(5):495–499, 1996. Na(+) channels by calmodulin. Circ Res 76. Jespersen T, Gavillet B, van Bemmelen MX, et al:
34. Abriel H, Kamynina E, Horisberger JD, 90(4):E49–E57, 2002. Cardiac sodium channel Nav1.5 interacts with
et al: Regulation of the cardiac voltage-gated 55. Kim J, Ghosh S, Liu H, et al: Calmodulin medi- and is regulated by the protein tyrosine phospha-
Na+ channel (H1) by the ubiquitin-protein ates Ca2+ sensitivity of sodium channels. J Biol tase PTPH1. Biochem Biophys Res Commun
ligase Nedd4. FEBS Lett 466(2-3):377–380, Chem 279(43):45004–45012, 2004. 348:1456–1463, 2006.
2000. 56. Young KA, Caldwell JH: Modulation of skeletal 77. Faulkner G, Lanfranchi G, Valle G: Telethonin
35. van Bemmelen MX, Rougier J-S, Gavillet B, et al: and cardiac voltage-gated sodium channels by and other new proteins of the Z-disc of skeletal
Cardiac voltage-gated sodium channel Nav1.5 is calmodulin. J Physiol 565(2):349–370, 2005. muscle. IUBMB Life 51(5):275–282, 2001.
regulated by Nedd4-2 mediated ubiquitination. 57. Motoike HK, Liu H, Glaaser IW, et al: The Na+ 78. Hayashi T, Arimura T, Itoh-Satoh M, et al: Tcap
Circ Res 95(3):284–291, 2004. channel inactivation gate is a molecular complex: gene mutations in hypertrophic cardiomyopathy
36. Kang L, Zheng M, Morishima M, et al: Bepridil A novel role of the COOH-terminal domain. and dilated cardiomyopathy. J Am Coll Cardiol
up-regulates cardiac Na(+) channels as a long- J Gen Physiol 123(2):155–165, 2004. 44(11):2192–2201, 2004.
term effect by blunting proteasome signals 58. Shah VN, Wingo TL, Weiss KL, et al: Calcium- 79. Mues A, van der Ven PFM, Young P, et al: Two
through inhibition of calmodulin activity. Br J dependent regulation of the voltage-gated sodium immunoglobulin-like domains of the Z-disc
Pharmacol 157(3):404–414, 2009. channel hH1: Intrinsic and extrinsic sensors use a portion of titin interact in a conformation-
37. Ingham RJ, Gish G, Pawson T: The Nedd4 common molecular switch. PNAS 103(10):3592– dependent way with telethonin. FEBS Lett 428(1-
family of E3 ubiquitin ligases: Functional diver- 3597, 2006. 2):111–114, 1998.
sity within a common modular architecture. 59. Wingo TL, Shah VN, Anderson ME, et al: An 80. Furukawa T, Ono Y, Tsuchiya H, et al: Specific
Oncogene 23(11):1972–1984, 2004. EF-hand in the sodium channel couples intracel- interaction of the potassium channel beta-subunit
38. Morrison DK: The 14-3-3 proteins: integrators lular calcium to cardiac excitability. Nat Struct minK with the sarcomeric protein T-cap suggests
of diverse signaling cues that impact cell fate and Mol Biol 11(3):219–225, 2004. a T-tubule-myofibril linking system. J Mol Biol
cancer development. Trends Cell Biol 19(1):16– 60. Potet F, Chagot B, Anghelescu M, et al: Func- 313(4):775–784, 2001.
23, 2009. tional interactions between distinct sodium 81. Mazzone A, Strege PR, Tester DJ, et al: A muta-
39. Mrowiec T, Schwappach B: 14-3-3 proteins in channel cytoplasmic domains through the action tion in telethonin alters Nav1.5 function. J Biol
membrane protein transport. Biol Chem of calmodulin. J Biol Chem 284(13):8846–8854, Chem 283(24):16537–16544, 2008.
387(9):1227–1236, 2006. 2009. 82. Ou Y, Gibbons SJ, Miller SM, et al: SCN5A is
40. Allouis M, Le Bouffant F, Wilders R, et al: 14-3-3 61. Chagot B, Potet F, Balser JR, et al: Solution NMR expressed in human jejunal circular smooth
is a regulator of the cardiac voltage-gated sodium structure of the C-terminal EF-hand domain of muscle cells. Neurogastroenterol Motil 14(5):
channel Nav1.5. Circ Res 98(12):1538–1546, human cardiac sodium channel NaV1.5. J Biol 477–486, 2002.
2006. Chem 284(10):6436–6445, 2008. 83. Steele DF, Eldstrom J, Fedida D: Mechanisms of
41. Olsen SK, Garbi M, Zampieri N, et al: Fibroblast 62. Miloushev VZ, Levine JA, Arbing MA, et al: Solu- cardiac potassium channel trafficking. J Physiol
growth factor (FGF) homologous factors share tion structure of the NaV1.2 C-terminal EF-hand (Lond) 582(Pt 1):17–26, 2007.
structural but not functional homology with domain. J Biol Chem 284(10):6446–6454, 2009. 84. Ziane R, Huang H, Moghadaszadeh B, et al: Cell
FGFs. J Biol Chem 278(36):34226–34236, 2003. 63. Casini S, Verkerk AO, van Borren MM, et al: membrane expression of cardiac sodium channel
42. Liu CJ, Dib-Hajj SD, Renganathan M, et al: Intracellular calcium modulation of voltage-gated Na(v)1.5 is modulated by alpha-actinin-2 interac-
Modulation of the cardiac sodium channel Nav1.5 sodium channels in ventricular myocytes. Cardio- tion. Biochemistry 49(1):166–178, 2010.
by fibroblast growth factor homologous factor 1B. vasc Res 81(1):72–81, 2009. 85. Funke L, Dakoji S, Bredt DS: Membrane-
J Biol Chem 278(2):1029–1036, 2003. 64. Couchonnal LF, Anderson ME: The role of associated guanylate kinases regulate adhesion
43. Goetz R, Dover K, Laezza F, et al: Crystal struc- calmodulin kinase II in myocardial physiology and and plasticity at cell junctions. Annu Rev Biochem
ture of a fibroblast growth factor homologous disease. Physiology 23(3):151–159, 2008. 74:219–245, 2000.
factor (FHF) defines a conserved surface on 65. Maier LS: Role of CaMKII for signaling and 86. Godreau D, Vranckx R, Maguy A, et al: Expres-
FHFS for binding and modulation of voltage- regulation in the heart. Front Biosci 14:486–496, sion, regulation and role of the MAGUK protein
gated sodium channels. J Biol Chem 2009. SAP-97 in human atrial myocardium. Cardiovasc
284(26):17883–17896, 2009. 66. Maier LS, Bers DM: Calcium, calmodulin, and Res 56(3):433–442, 2002.
44. Wehrens XHT, Abriel H, Cabo C, et al: Arrhyth- calcium-calmodulin kinase II: Heartbeat to heart- 87. Leonoudakis D, Conti LR, Anderson S, et al:
mogenic mechanism of an LQT-3 mutation of the beat and beyond. J Mol Cell Cardiol 34(8):919– Protein trafficking and anchoring complexes
human heart Na(+) channel alpha-subunit: A 939, 2002. revealed by proteomic analysis of inward rectifier
computational analysis. Circulation 102(5):584– 67. Wagner S, Dybkova N, Rasenack EC, et al: Ca/ potassium channel (Kir2.x) associated proteins.
590, 2000. calmodulin-dependent protein kinase II regulates J Biol Chem 279(21):22331–22346, 2004.
45. Lou JY, Laezza F, Gerber BR, et al: Fibroblast cardiac Na channels. J Clin Invest 116(12):3127– 88. Godreau D, Vranckx R, Maguy A, et al: Different
growth factor 14 is an intracellular modulator of 3128, 2006. isoforms of synapse-associated protein, SAP97,
voltage-gated sodium channels. J Physiol 569(Pt 68. Yoon JY, Ho WK, Kim ST, et al: Constitutive are expressed in the heart and have distinct effects
1):179–193, 2005. CaMKII activity regulates Na+ channel in rat on the voltage-gated K+ channel Kv1.5. J Biol
46. Wang C, Hennessey JA, Kirkton RD, et al: Fibro- ventricular myocytes. J Mol Cell Cardiol Chem 278(47):47046–47052, 2003.
blast growth factor homologous factor 13 regu- 47(4):475–484, 2009. 89. El-Haou S, Balse E, Neyroud N, et al: Kv4 potas-
lates Na+ channels and conduction velocity in 69. Ashpole NM, Herren AW, Ginsburg KS, et al: sium channels form a tripartite complex with the
murine hearts. Circ Res 109(7):775–782, 2011. Ca2+/calmodulin-dependent protein kinase II anchoring protein SAP97 and CaMKII in cardiac
47. Wang C, Chung BC, Yan H, et al: Crystal struc- (CaMKII) regulates cardiac sodium channel myocytes. Circ Res 104(6):758–769, 2009.
ture of the ternary complex of a NaV C-terminal NaV1.5 gating by multiple phosphorylation sites. 90. Williams TM, Lisanti MP: The Caveolin genes:
domain, a fibroblast growth factor homologous J Biol Chem 287(24):19856–19869, 2012. From cell biology to medicine. Ann Med
factor, and calmodulin. Structure 20(7):1167– 70. Hund TJ, Koval OM, Li J, et al: A bIV-spectrin/ 36(8):584–595, 2004.
1176, 2012. CaMKII signaling complex is essential for mem- 91. Cohen AW, Hnasko R, Schubert W, et al: Role of
48. Saimi Y, Kung C: Calmodulin as an ion channel brane excitability in mice. J Clin Invest caveolae and caveolins in health and disease.
subunit. Annu Rev Physiol 64:289–311, 2002. 120(10):3508–3519, 2010. Physiol Rev 84(4):1341–1379, 2004.
92. Vatta M, Ackerman MJ, Ye B, et al: Mutant 100. Wu G, Ai T, Kim JJ, et al: Alpha-1-syntrophin 108. Gerull B, Heuser A, Wichter T, et al: Mutations
caveolin-3 induces persistent late sodium current mutation and the long QT syndrome: A disease in the desmosomal protein plakophilin-2 are
and is associated with long-QT syndrome. Circu- of sodium channel disruption. Circ Arrhythm common in arrhythmogenic right ventricular car-
lation 114(20):2104–2112, 2006. Electrophysiol 1:193–201, 2008. diomyopathy. Nat Genet 36(11):1162–1164, 2004.
93. Cronk LB, Ye B, Kaku T, et al: Novel mechanism 101. Weiss R, Barmada MM, Nguyen T, et al: Clinical 109. Pilichou K, Nava A, Basso C, et al: Mutations in
for sudden infant death syndrome: Persistent late and molecular heterogeneity in the Brugada syn- desmoglein-2 gene are associated with arrhyth-
sodium current secondary to mutations in drome: A novel gene locus on chromosome 3. mogenic right ventricular cardiomyopathy. Circu-
caveolin-3. Heart Rhythm 4(2):161–166, 2007. Circulation 105(6):707–713, 2002. lation 113(9):1171–1179, 2006.
94. Yarbrough TL, Lu T, Lee HC, et al: Localization 102. London B, Michalec M, Mehdi H, et al: Mutation 110. Cunha SR, Mohler PJ: Cardiac ankyrins:
of cardiac sodium channels in caveolin-rich mem- in glycerol-3-phosphate dehydrogenase 1 like Essential components for development and main-
brane domains: Regulation of sodium current gene (GPDl-L) decreases cardiac Na+ current and tenance of excitable membrane domains in heart.
amplitude. Circ Res 90(4):443–449, 2002. causes inherited arrhythmias. Circulation Cardiovasc Res 71(1):22–29, 2006.
95. Shibata EF, Brown TL, Washburn ZW, et al: 116:2260–2268, 2007. 111. Mohler PJ, Schott JJ, Gramolini AO, et al:
Autonomic regulation of voltage-gated cardiac 103. Van Norstrand DW, Valdivia CR, Tester DJ, et al: Ankyrin-B mutation causes type 4 long-QT
ion channels. J Cardiovasc Electrophysiol Molecular and functional characterization of cardiac arrhythmia and sudden cardiac death.
17(Suppl 1):S34–S42, 2006. novel glycerol-3-phosphate dehydrogenase 1 like Nature 421(6923):634–639, 2003.
96. Doyle DD, Goings G, Upshaw-Earley J, et al: gene (GPDl-L) mutations in sudden infant death 112. Mohler PJ, Le Scouarnec S, Denjoy I, et al:
Dystrophin associates with caveolae of rat cardiac syndrome. Circulation 116:2253–2259, 2007. Defining the cellular phenotype of “ankyrin-B
myocytes: Relationship to dystroglycan. Circ Res 104. Valdivia CR, Ueda K, Ackerman MJ, et al: GPDlL syndrome” variants. Human ANK2 variants asso-
87(6):480–488, 2000. links redox state to cardiac excitability by PKC- ciated with clinical phenotypes display a spectrum
97. Gee SH, Madhavan R, Levinson SR, et al: Inter- dependent phosphorylation of the sodium channel of activities in cardiomyocytes. Circulation
action of muscle and brain sodium channels SCN5A. Am J Physiol Heart Circ Physiol 115(4):432–441, 2007.
with multiple members of the syntrophin family 297(4):H1446–H1452, 2009. 113. Chauhan VS, Tuvia S, Buhusi M, et al: Abnormal
of dystrophin-associated proteins. J Neurosci 105. Marfatia KA, Harreman MT, Fanara P, et al: cardiac Na(+) channel properties and QT heart
18(1):128–137, 1998. Identification and characterization of the human rate adaptation in neonatal ankyrin(B) knockout
98. Albesa M, Ogrodnik J, Rougier J-S, et al: Regula- MOG1 gene. Gene 266(1-2):45–56, 2001. mice. Circ Res 86(4):441–447, 2000.
tion of the cardiac sodium channel Nav1.5 by 106. Wu L, Yong SL, Fan C, et al: Identification of a 114. Lemaillet G, Walker B, Lambert S: Identification
utrophin in dystrophin-deficient mice. Cardio- new co-factor, MOG1, required for the function of a conserved ankyrin-binding motif in the
vasc Res 89(2):320–328, 2011. of cardiac sodium channel Nav1.5. J Biol Chem family of sodium channel alpha subunits. J Biol
99. Ueda K, Valdivia C, Medeiros-Domingo A, et al: 283(11):6968–6978, 2008. Chem 278(30):27333–27339, 2003.
Syntrophin mutation associated with long QT 107. Kattygnarath D, Maugenre S, Neyroud N, et al: 115. Lowe JS, Palygin O, Bhasin N, et al: Voltage-
syndrome through activation of the nNOS- MOG1: A new susceptibility gene for Brugada gated Nav channel targeting in the heart requires
SCN5A macromolecular complex. PNAS syndrome. Circ Cardiovasc Genet 4(3):261–268, an ankyrin-G dependent cellular pathway. J Cell
105(27):9355–9360, 2008. 2011. Biol 180(1):173–186, 2008.
Calmodulin and CaMKII as Ca2+
Switches for Cardiac Ion Channels 19
Geoffrey S. Pitt and Steven O. Marx
the cardiac action potential and thereby for controlling CaV1.2

CHAPTER OUTLINE
gating and the consequent intracellular Ca2+ signal. Growing
Ca2+: The Final Signal of Electrical Activity and evidence over the past decade has demonstrated that Ca2+ can
Regulator of Ion Channels 189 regulate cardiac ion channels at many levels. For example, Ca2+
can control channel transcription, biosynthesis, trafficking, or
Unique Features of Ca2+ as a Signaling Ion in Myocytes 189 gating of mature channels at the sarcolemma.
2+
Translating Changes in Ca into a Cellular Response:
CaM as the Prototypical Cardiac Ca2+ Sensor 189
CaM Effector Functions on Channels: Direct Binding Unique Features of Ca2+ as a Signaling
and Indirect Actions through CaMKII 190 Ion in Myocytes
Calmodulin Regulation of Cardiac Channel Gating 190
The efficacy of Ca2+ as an intracellular signal derives from specific
CaMKII Regulation of Cardiac Channels: Indirect properties that set it apart from other intracellular ions. First, Ca2+
Regulation by CaM 192 signals have a broad dynamic range. Global intracellular Ca2+ can
rapidly increase more than tenfold, and even more in certain sub-
cellular locations such as near the mouth of an ion channel. This
feature results from the concerted actions of several cellular
Abstract mechanisms designed to keep free cytoplasmic Ca2+ concentra-
tions low and thereby prevent precipitation with organic phos-
Changes in intracellular Ca2+ are among the most diverse and phates, a major intracellular counter anion. These Ca2+-limiting
important means of cell signaling. In the heart, signaling path- mechanisms maintain a free intracellular Ca2+ concentration
ways from excitation-contraction coupling to humoral activation ([Ca2+]i) close to 100 nM during diastole, which rises to upwards
of hypertrophic responses all rely on changes in the concentra- of 1 µM during the plateau phase of the action potential, in con-
tion of intracellular Ca2+. Ion channels belong to a class of signal- trast to an extracellular free Ca2+ concentration of 1 to 2 mM.
ing proteins that is particularly sensitive to a change in intracellular Thus, there is a strong chemical driving force for Ca2+ entry.
Ca2+, which is the final signal of all coordinated ion channel activ- Second, Ca2+ entry is fast. When CaV1.2 Ca2+ channels (the major
ity. Thus, regulation of ion channel function by Ca2+ provides an route for Ca2+ entry) open during the action potential plateau,
essential feedback mechanism for cellular electrical activity. Two Ca2+ flows down a large electrochemical gradient into the cell at
intracellular proteins activated by Ca2+, the ubiquitous Ca2+ rates approaching 106 ions per second. Several pumps and trans-
binding protein calmodulin (CaM), and the Ca2+/CaM-dependent porters remove Ca2+ from the cytoplasm almost as rapidly, return-
protein kinase CaMKII, dictate most of the actions of intracel- ing [Ca2+]i levels back to baseline in time for the next heartbeat.
lular Ca2+ on cardiac ion channels. In this chapter, we highlight These swift influx and efflux pathways allow the Ca2+ signal to
key roles for CaM and CaMKII in cardiac myocytes as down- achieve its large (tenfold) dynamic range within a single beat and
stream integrators of Ca2+ signals in the regulation of cardiac ion trigger downstream responses within the same time scale.
channels and consequent actions on myocyte excitability.
Translating Changes in Ca2+ Into a

2+
Ca : The Final Signal of Electrical Activity Cellular Response: CaM as the
and Regulator of Ion Channels Prototypical Cardiac Ca2+ Sensor
In adult cardiomyocytes, voltage-gated ion channels exist solely Several Ca2+ binding proteins serve as the Ca2+ sensors to trans-
to change intracellular Ca2+. Whether regulating action poten- late changes in intracellular Ca2+ into cellular actions. For
tials in myocytes, neurons, or hormone-secreting cells, ion chan- example, Ca2+ binding to troponin C causes dissociation of the
nels control membrane voltage and thereby either activate or troponin complex from the active site on actin, allowing myosin
inhibit voltage-gated Ca2+ channels to control Ca2+ entry into the interaction and force generation. For regulation of ion channels,
cells. K+ channels can control resting membrane potential and the best-characterized Ca2+ sensor is the ubiquitous Ca2+-binding
drive action potential repolarization, but they do so to limit Ca2+ protein CaM, a 16.8 kDa protein that binds four moles of Ca2+
entry through L-type CaV1.2 Ca2+ channels. Voltage-gated Na+ per mole of protein. CaM is highly abundant in cardiac myocytes,
channels could be essential for the initiation of action potentials but more than 98% CaM is apoCaM (Ca2+-free) sequestered by
in the working myocardium, but the net effect of action potentials binding proteins only to be released upon a significant increase
is to open CaV1.2 Ca2+ channels to translate electrical activity into in [Ca2+]i.1 The affinity of CaM for Ca2+ varies significantly
cellular contraction. Not only does Ca2+ serve as the final signal depending on whether CaM is free in solution or bound in its
of cellular electrical activity, but Ca2+ also actively participates in apo-state to a target protein, such as an ion channel. Ca2+ binding
the regulation of many of the channels responsible for controlling by CaM occurs in the context of 1 to 2 mM intracellular Mg2+,
189
the major competing divalent cation. The structural motif in therefore ideally suited to respond to changes in cardiac rhythm
CaM capable of distinguishing Ca2+ at levels less than 1 : 1000 of and the accompanying change in frequency of Ca2+ influx through
Mg2+ is the “EF hand,” a helix-loop-helix domain also found in the L-type CaV1.2 Ca2+ channel with every heartbeat. Thus,
many other Ca2+-binding proteins, including troponin C, in CaMKII phosphorylation of the L-type CaV1.2 Ca2+ channel
which it was originally identified. CaM has two EF hands in an provides a direct Ca2+-dependent means to regulate the major
N-terminal lobular domain and two more in a C-terminal lobular source of Ca2+ influx. Finally, because gating of CaV1.2 Ca2+
domain. The two Ca2+-binding domains are connected by an channels depends on the concerted actions of Na+ and K+ chan-
α-helical segment. Upon Ca2+ binding, CaM undergoes signifi- nels and the consequent regulation of membrane voltage,
cant conformational changes that expose a hydrophobic surface, CaMKII phosphorylation of any of these channels can provide
which can then interact with target proteins in a Ca2+-dependent additional, indirect regulation of CaV1.2 Ca2+ channels and sub-
manner. Within most well-characterized target proteins, such as sequent Ca2+ influx, thereby leading back to CaMKII to generate
CaMKII, the CaM interaction domain is an amphipathic helix a web of feedback or feed forward mechanisms to control intra-
for which the hydrophobic amino acid side chains become buried cellular Ca2+.
within the hydrophobic surface exposed in Ca2+-saturated CaM. Reactive oxygen species provide an additional means to acti-
In CaMKII, this amphipathic helix blocks access to the kinase’s vate CaMKII. Oxidation of two methionine residues, adjacent to
constitutively active site; CaM binding to this autoinhibitory the site of transphosphorylation, endows CaMKII with Ca2+-
domain reveals the active site and thereby endows the kinase with independent activity in a manner similar to transphosphoryla-
a Ca2+-dependent response. In many of the cardiac ion channels tion.2 As with transphosphorylation, Ca2+/CaM binding must
to which CaM binds directly, the CaM binding motif has a similar occur first, exposing the target methionines to oxidation. Given
amphipathic pattern. The actions of CaM on the channels, the association of CaMKII with several adverse cardiac outcomes,
however, are less well understood, and this will be discussed next. including arrhythmias, and an increased redox state with those
outcomes, this means of activating CaMKII may have important
detrimental effects on cardiac ion channels.
CaM Effector Functions on Channels:
Direct Binding and Indirect Actions
Through CaMKII Calmodulin Regulation of Cardiac
Channel Gating
In the last decade there has been an explosion of functional, bio-
chemical, and structural information about how CaM interacts CaM Regulation of Ca2+-Dependent Inactivation
directly with multiple ion channels to affect their function. Details and Membrane Targeting of CaV1.2 Ca2+ Channels
about how CaM regulates the voltage-gated Na+ channel NaV1.5,
the L-type voltage-gated Ca2+ channel CaV1.2, the ryanodine Ca2+-dependent inactivation (CDI) of Ca2+ channels serves as a
receptor type 2 (RyR2) Ca2+ release channel, and several different classic example of Ca2+/CaM regulation of ion channel function
voltage-gated K+ channels, including KCNQ1, will be presented (Figure 19-1). CDI denotes the accelerated channel inactivation
in specific parts of this chapter. The regulatory effects of Ca2+ on seen in experiments in which Ca2+ is used as the charge carrier
cardiac ion channels are not mediated solely by CaM interaction rather than another permeant divalent cation, such as Ba2+. That
with the channels, however, but also indirectly through the actions the permeant ion regulates channel gating sets CaV1.2 apart from
of other Ca2+-regulated proteins, the most prominent of which is other voltage-gated cardiac channels, in which gating is solely
CaMKII. In fact, activation of CaMKII results in the phosphory- voltage-dependent. CDI of CaV1.2 in myocytes is critical for
lation of many of the same channels targeted by CaM, and the regulating Ca2+ entry and for controlling the length of the plateau
resulting protein modulation can have profound effects on channel phase of the cardiac action potential. CaM, bound to the “IQ”
activity, sometimes in ways that oppose the effects of direct CaM motif in the C-terminus of the CaV1.2 pore-forming α1C-subunit,
interaction with the same channel. Which Ca2+-dependent regu- serves as the Ca2+ sensor for CDI of CaV1.2. Interaction between
lator dominates depends on the subcellular location, amplitude, CaM and the IQ motif appears essential: homozygous mice with
frequency, and duration of the intracellular Ca2+ signals, largely a knock-in mutation in the IQ motif that disrupted CaM interac-
because of properties of CaMKII that allow it to target to specific tion died during embryogenesis.3 Adult mice with an IQ motif
subcellular locations and independently modulate its kinase activ- mutation, obtained with an inducible Cre-recombinase strategy,
ity depending on timing of repetitive Ca2+ stimuli. Thus, in the died within three weeks after inducing the mutant allele.
heart, CaMKII has the capacity for graded effects that correlate An interesting phenotype observed in these mice was a reduc-
with changes in heart rate. tion in the number of channels, as indicated by reduced CaV1.2
Several features of CaMKII endow the kinase with this ability Ca2+ current density and α1C protein. This finding confirmed
to regulate activity depending on the timing of the Ca2+ stimuli. previous reports showing that CaM interaction with the α1C IQ
These include the holoenzyme’s structure and the fact that motif regulates trafficking of the α1C protein to the plasma mem-
CaMKII is itself a CaMKII substrate. The kinase is a homodo- brane.4 Thus, regulation of channel biosynthesis demonstrates
decamer, in which the 12 subunits are arranged in two stacked another means by which Ca2+, via CaM, can affect cellular electri-
hexameric rings with the active site and the CaM-binding auto- cal activity. Although CaM control of channel biosynthesis is best
inhibitory helix of each subunit closest to the periphery. If the studied for CaV1.2 Ca2+ channels, CaM appears to play a similar
Ca2+ signal is large enough, or if two Ca2+ signals are temporally role for other cardiac channels, such as the KCNQ1 K+ channel
spaced so that CaM binds simultaneously to two adjacent sub- and the NaV1.5 Na+ channel, as discussed next.
units in a ring, then one of the activated subunits can phosphory- Recent data suggest that CaM interaction with the α1C
late the other CaM-bound subunit at a specific Thr residue N-terminus also contributes to CDI.5 The mechanisms by which
(Thr287 in the cardiac CaMKIIδ isoform) that is adjacent to the CaM accelerates channel inactivation, and whether the CaM
CaM binding helix. This phosphorylation event increases bound to the α1C C-terminus is the same molecule bound to the
the kinase’s affinity for CaM more than 10,000-fold, causing the N-terminus, are not clear.
kinase to become persistently activated even after the activating Crystal structures that identify putative interaction residues
Ca2+ signal has subsided. This capacity to respond differentially for CaM on the α1C C-terminus6,7 provide a framework for future
to changes in Ca2+ influx intensity or frequency make CaMKII investigations. These models show an unexpected dimerization
CALMODULIN AND CaMKII AS Ca2+ SWITCHES FOR CARDIAC ION CHANNELS 191
K+ 4
Open
Closed
1
Na+ 2 19
Open
2
Closed
Availability
Ca++
1 2
2 T-tubule
More
protein
Voltage
1 Ca++ RyR2 Ca++
Ca++ RyR2 Ca++
Sarcoplasmic
reticulum
Ca++
1 +CaM
Figure 19-1. CaM regulation of cardiac ion channels. CaM controls Ca2+-dependent inactivation of CaV1.2 L-type Ca2+ channels (1), increases availability of NaV1.5 Na+ chan-
nels (2), decreases open probability of RyR2 receptors (3), and increases the abundance of KCNQ1 K+ channels (4). The traces for RyR2 were adapted from Yamaguchi et al.54
of two α1C C-termini through simultaneous binding of multiple channel’s C-terminus and the III-IV linker. Interestingly, an x-ray
CaMs to the two C-termini. In cardiac myocytes, the stoichiom- crystal structure presented evidence for an interaction between
etry of CaM with the α1C C-terminus is not known, but experi- Ca2+-loaded CaM C-lobe and the channel’s III-IV intracellular
mental evidence from a heterologous expression system suggests linker.14 In the context of biochemical and structural data showing
that a single CaM can regulate CDI.8 Moreover, these structural interaction of the apoCaM C-lobe with the channel’s C-terminus,
models do not account for how any of these CaMs could also these data suggest that Ca2+ may induce a movement of the
bind to the α1C N-terminus. Therefore, these models might not C-lobe from the C-terminus to the III-IV linker and that the
accurately reflect how CaM interacts with α1C within the context CaM N-lobe may instead bind the IQ motif to act as a bridge.
of the intact channel. Indeed, the dimerization of α1C C-termini Whether the CaM N-lobe interacts with the channel’s C-terminus
has been suggested to represent a crystallographic artifact.6 On in the presence of Ca2+ simultaneously with the CaM C-lobe
the other hand, recent work presents an intriguing possibility that binding to the III-IV linker has not been demonstrated. Under-
native α1C subunits can multimerize through their C-termini via standing how Ca2+/CaM regulates NaV1.5 Na+ channels will
a CaM-dependent mechanism.7,9 likely be greatly aided as more structural information becomes
available.
As with CaV1.2 Ca2+ channels, CaM interaction with NaV1.5
CaM Regulation of NaV1.5 Na+ Channels may also participate in channel biosynthesis based on analogies
from data focused on noncardiac Na+ channels. For the skeletal
Ca2+ also regulates NaV1.5 Na+ channels (see Figure 19-1). Na+ muscle NaV1.4 and the neuronal NaV1.6 Na+ channels, mutations
channels initiate the action potential, and mutations that affect disrupting the interaction between CaM and the channel’s
NaV1.5 Na+ channel function have been linked to multiple C-terminus almost completely reduced current amplitude.15
arrhythmogenic disorders, most prominently long QT syndrome With new structural information, these results should be revisited
and Brugada syndrome. Elevated Ca2+ increases channel avail- with a specific focus on NaV1.5 Na+ channels in cardiomyocytes
ability,10 but whether CaM affects inactivation of NaV1.5 in myo- because of recent data showing that fibroblast growth factor
cytes has not yet been demonstrated. Moreover, the mechanism homologous factors, which bind to NaV1.5 in the region adjacent
by which Ca2+ affects NaV1.5 Na+ channel function is not clear. to the CaM-binding IQ motif,16 help traffic NaV1.5 to the sarco-
As with CaV1.2 Ca2+ channels, CaM binds to an IQ motif in the lemma of cardiomyocytes.17
NaV1.5 Na+ channel’s C-terminus, and thus one candidate Ca2+
sensor is this CaM. Mutations in NaV1.5 designed to disrupt
CaM binding to the IQ motif decrease channel availability CaM Regulation of Voltage-gated K+ Channels
(hyperpolarize steady-state inactivation) when studied in heter-
ologous systems.10 A recent crystal structure shows how, in the Another voltage-gated channel in the heart that contains a CaM-
absence of Ca2+, the CaM’s C-lobe wraps around the NaV1.5 Na+ binding IQ motif is the KCNQ1 K+ channel, the most commonly
channel’s IQ motif.11 As additional structures with Ca2+/CaM are mutated locus in long QT syndrome. KCNQ1 K+ channels
reported, critical insights into the actions of CaM on NaV1.5 are control repolarization of the cardiac action potential; loss-of-
likely to be revealed. CaM can also bind to a channel’s putative function mutations hinder repolarization and lengthen the
inactivation gate, the III-IV intracellular linker,12 and a Ca2+- cellular action potential and the electrocardiographic QT inter-
dependent complex containing the NaV1.5 C-terminus, CaM, val, providing a basis for life-threatening arrhythmias. As with
and the III-IV can be formed,13 suggesting a model by which CaV1.2 Ca2+ channels and NaV1.5 Na+ channels, CaM interaction
Ca2+/CaM can affect channel inactivation by bridging the with the KCNQ1 C-terminus appears to participate in channel
biosynthesis (see Figure 19-1). In contrast to CaV1.2 Ca2+ chan- changes in cardiac rhythm, phosphorylates CaV1.2 and thereby
nels and NaV1.5 Na+ channels, CaM interaction with the KCNQ1 increases Ca2+ influx, a process called Ca2+-dependent facilitation.
C-terminus requires two distinct helices that are separated by At the single-channel level, CaMKII phosphorylation induces the
approximately 80 amino acids. Biochemical and functional studies channel to open frequently and to remain open longer. This
implicate that CaM is essential for generating KCNQ1 K+ cur- increase in so-called “mode 2” gating raises intracellular Ca2+ and
rents. The significance of these findings is underlined by the initiates a chain of downstream events that increase Ca2+ efflux
association between various long QT syndrome mutations and from internal stores and reduce Ca2+ reuptake in the sarcoplasmic
loss of CaM binding to channel’s C-terminus.18,19 How CaM is reticulum, providing elevated levels of Ca2+ for the contractile
deemed necessary was elucidated by studies that implicated a role apparatus. Although the end results on cardiac contractility are
for CaM in channel trafficking to the plasma membrane rather well established, detailed mechanistic insight into how CaMKII
than in channel assembly.20 Through CaM, Ca2+ also appears to affects Ca2+ influx through CaV1.2 Ca2+ channels, where the
participate in channel gating,18,19 but the molecular mechanisms process initiates, is lacking. The substrates for CaMKII within
have not yet been clarified. CaV1.2 has been proposed as Ser1512 and Ser1570 in the α1C-
subunit28 or Thr498 in the auxiliary β2-subunit.29 Recent knock-in
data question the role of the phosphorylation sites in the
CaM Regulation of Ryanodine Receptors β-subunit.30 The experimental approaches (two Ser to Ala muta-
tions in a knock-in mouse for the α1C sites; adenoviral overexpres-
CaM regulates the SR Ca2+ release channel RyR2, the internal sion of a Thr to Ala mutant for β2 sites) behind each of those
source of the Ca2+ that triggers contraction with each heartbeat conclusions have their limitations, however, and perhaps phos-
(see Figure 19-1). How CaM binds to RyR2 and the consequent phorylation of both α1C- and β2-subunits are required for CaMKII
regulation of channel function are different from the manner potentiation.31 Similarly, separate studies have shown that
described before for CaM interaction with, and regulation of, CaMKII can bind directly to the CaV1.2 Ca2+ channel α1C- or
plasma membrane channels. In vitro experiments show that β2-subunit.32,33 The advantages of a dedicated Ca2+-frequency
apoCaM and Ca2+/CaM have a single binding site on the large detector (CaMKII) at the source of Ca2+ entry (CaV1.2) that is
RyR2 cytoplasmic domain that does not resemble a consensus IQ capable of modulating the amount of Ca2+ that enters are obvious,
motif.21 especially in the context of the broad array of CaMKII substrates
CaM inhibits cardiac muscle RyR2 at greater than 1 µM Ca2+ and the myriad consequences of CaMKII activity in the heart.34
and either inhibits RyR222 or has no effect23 on RyR2 at less than Moreover, mutation of the CaMKII binding site for either α1C
1 µM Ca2+. In single-channel measurements, CaM decreased or β2 has been shown to ablate Ca2+-dependent facilitation, sug-
RyR2 channel open probability by decreasing the number of gesting that this association is critical. Similar to the debate about
channel events and increasing the duration of close times.24 In which subunit is the “true” CaMKII substrate, the answer of
permeabilized myocytes, the addition of exogenous CaM which subunit is the CaMKII scaffold is also not clearly defined;
decreased Ca2+ spark frequency.25 The physiological importance both α1C and β2 may be important for proper channel
of CaM binding to RyR2 was shown using mice harboring tar- regulation.
geted RyR2 mutations in the CaM binding site. These mutant Although CaMKII appears to be essential for Ca2+-dependent
mice demonstrated an increased ratio of heart weight to body facilitation of CaV1.2 Ca2+ channels, the CaM molecule that
weight and greatly reduced fractional shortening of the left ven- interacts with the channel’s C-terminus also is required. Muta-
tricle and lethality at 9 to 16 days of age.26 Interestingly, in a tions that disrupt CaM interaction, and thereby affect CDI, also
rabbit heart failure model, less CaM coimmunoprecipitated with abolish CaV1.2 Ca2+ channel facilitation in cardiomyocytes.3 How
RyR2, despite unaltered expression.27 Taken together, altered these two separate Ca2+ sensors cooperate to produce facilitation
CaM bound to RyR2 may contribute to both the enhanced SR is not known, although it has been speculated that the CaM
Ca2+ leak in heart failure and to arrhythmogenesis. bound to the channel’s IQ motif may also be able to serve as the
CaM that regulates the CaMKII bound to α1C or β2, providing a
model that incorporates the binding of CaMKII to both sub-
CaMKII Regulation of Cardiac Channels: units.35 This model is not yet supported by biochemical or struc-
tural evidence.
Indirect Regulation by CaM
Heart Rate–dependent Regulation of Cardiac CaMKII Regulation of NaV1.5 Na+ Channels

Ion Channels
CaMKII activity decreases NaV1.5 Na+ channel availability and
2+
CaMKII adds additional dimensions to Ca /CaM signaling. By enhances the Na+ “late” or “persistent” current in cardiomyocytes
providing a means by which the frequency of Ca2+ signals can be (see Figure 19-2).36 Because CaMKII protein and activity are
translated into changes in kinase activity, CaMKII is a link increased in heart failure, these effects upon NaV1.5 Na+ channels
between physiological changes in heart rate, such as during exer- may be particularly relevant for the arrhythmogenesis that
cise, and alterations in channel function. Several channels are accompanies heart failure. In contrast to CaV1.2 Ca2+ channels
CaMKII substrates, including most of those that are also regu- on which CaMKII binds directly, the CaMKII relevant for
lated directly by CaM. NaV1.5 Na+ channel modulation is instead bound to βIV-spectrin,37
an actin-associated protein that has previously been shown to
cluster ion channels in neuronal axons. In the heart, βIV-spectrin
CaMKII Regulation of CaV1.2 Ca2+ Channels localizes to the intercalated discs, where a fraction of NaV1.5 Na+
channels is enriched. CaMKII directly phosphorylates NaV1.5,
CaMKII regulation of CaV1.2 Ca2+ channels is a major contribu- but the specific sites targeted are controversial. Hund et al identi-
tor to the positive force-frequency relationship of cardiac con- fied Ser571 within the channel’s I-II intracellular linker as the
traction (Figure 19-2). Cardiac output, a product of stroke critical site; mutation of this Ser to Ala abolished the effects of
volume and heart rate, increases during exercise not only because CaMKII on channel inactivation.37 In contrast, Ashpole et al
of an increase in heart rate but also because each contraction found that the CaMKII targets in the I-II linker were Ser483/
produces more force. CaMKII, ideally suited to respond to Ser484, Ser516, and Thr594, and that Ala-substitution of both
K+ 4
Open
Closed
1
Na+ 2
19
Availability
Open 2 1
2
Closed
Voltage
1 Ca++
T-tubule
CaMKII
Less 2
protein Ca++ RyR2 Ca++
Ca++ RyR2 Ca++
Sarcoplasmic
reticulum
Ca++
1
1
2
Figure 19-2. CaMKII regulation of cardiac ion channels. CaMKII elicits Ca2+-dependent facilitation of CaV1.2 L-type Ca2+ channels (1), decreases availability of NaV1.5 Na+
channels (2), increases open probability of RyR2 receptors (3), and decreases the abundance of Kv4 and Kir2.1 K+ channels (4). The traces for RyR2 were adapted from Wehrens
et al.55
Ser516 and Thr594 abolished the CaMKII mediated shift in INa CaMKII is associated with the RyR2 macromolecular complex42
availability to a more negative membrane potential.38 Neither set and phosphorylates the channel in vitro and in vivo.43-45 Although
of experiments was performed in cardiomyocytes, so the relevant Ser2809 on canine and human RyR2 (Ser2808 in murine RyR2)
sites within their endogenous milieu remain to be determined. was originally identified as a protein kinase A and CaMKII phos-
phorylation site,43,45 site-directed mutagenesis studies and
knock-in mice have shown that Ser2809 is exclusively phosphory-
CaMKII Regulation of Kv4 and Kir2.1 K+ Channels lated by protein kinase A,46 and Ser2815 (Ser2814 in murine
RyR2) is phosphorylated by CaMKII.44,47,48 Other sites for
The K+ channels responsible for the resting membrane potential CaMKII phosphorylation on RyR2 may also exist, and the iden-
(IK1, mediated by Kir2.1) and for the transient outward current tity of the relevant sites is controversial. CaMKII phosphoryla-
(Ito, which produces the “notch” in the ventricular action poten- tion of RyR2 increases the open probability of RyR2 by sensitizing
tial, mediated by Kv4.2 and Kv4.3) are regulated by CaMKII. the channel to cytosolic Ca2+.44,49 CaMKII phosphorylation of
Chronic increased CaMKII activity, such as during heart failure, RyR2 by endogenous CaMKII increases resting SR Ca2+ release
reduced Ito current and reduced Kv4.2 and Kv4.3 protein, sug- by increasing Ca2+ spark frequency and duration.25 CaMKII phos-
gesting that CaMKII can affect channel biosynthesis (see Figure phorylation of RyR2 plays an important role in mediating the
19-2). Transcription of Kv4.3 in canine myocytes has been shown positive force-frequency relationship in the heart. Mice engi-
to be under the control of CaMKII by a downstream mechanism neered with a RyR2-S2814A mutation have RyR2 channels that
dependent on the nuclear factor of activated T cells transcription cannot be phosphorylated by CaMKII and exhibit a blunted posi-
factor.39 In addition to these effects on channel transcription, tive force-frequency relationship.47 The positive force-frequency
CaMKII has direct effects on channel function. CaMKII phos- relationship is blunted in heart failure, and in a rat model of heart
phorylation slowed Kv4.3 channel inactivation.40 The current failure, a rate-dependent increase in CaMKII phosphorylation of
regulating IK was also reduced by chronic CaMKII expression RyR2-Ser2814 was not observed.44
and, similar to Kv4 K+ channels, the effect is through transcrip- Increased activity of CaMKII is believed to promote heart
tional regulation.41 failure progression and arrhythmogenesis, including atrial fibril-
lation.50,51 Heart failure and arrhythmias are associated with
abnormal Ca2+ handling. What is the role of CaMKII phosphory-
CaMKII Regulation of RyR2 lation of RyR2 in these processes? In patients with nonischemic
cardiomyopathy, phosphorylation of RyR2-Ser2814 is increased.48
CaMKII regulation of RyR2 channel function plays an important Mice engineered with a RyR2-S2814A mutation were relatively
role in modulating cardiac contractility and arrhythmogenesis. protected from heart failure development after transverse aortic
constriction compared with wild type littermates. These protec-

tive effects on cardiac contractility were not observed, however, Conclusion
after myocardial infarction in the S2814A mice.47 The RyR2-
S2814A mice are also protected from the development of Modulating the interactions of CaM and CaMKII with ion chan-
atrial fibrillation51,52 and pacing-induced arrhythmias after nels and the phosphorylation of ion channels by CaMKII is likely
transverse aortic constriction.53 Conversely, mice engineered to be an important future target for therapies for heart failure
with a constitutively activated CaMKII phosphorylation site and arrhythmias. These two Ca2+-dependent sensors provide, for
(S2814D) develop sustained ventricular tachycardia and sudden several discussed cases, opposing effects on channel function.
cardiac death with caffeine and epinephrine or programmed The capacity of CaMKII to respond to change in heart rate or
electrical stimulation.53 These studies demonstrate that CaMKII to reactive oxygen species provides additional means by which
phosphorylation of RyR2, although important for the positive CaMKII can affect ion channel function. Understanding the inte-
force-frequency response, plays a role in the progression of heart gration of these various signals, not only at the level of a target
failure and the development of atrial and ventricular ion channel but also within the context of integrated cellular
arrhythmias. activity, will likely be a focus of future research.
16. Wang C, Wang C, Hoch EG, et al: Identification Cav1.2 channel. J Biol Chem 287(27):22584–
References of novel interaction sites that determine specificity 22592, 2012.
between fibroblast growth factor homologous 31. Sun AY, Pitt GS: Pinning down the CaMKII
1. Wu X, Bers DM: Free and bound intracellular factors and voltage-gated sodium channels. J Biol targets in the L-type Ca2+ channel: An essential
calmodulin measurements in cardiac myocytes. Chem 286(27):24253–24263, 2011. step in defining CaMKII regulation. Heart Rhythm
Cell Calcium 41(4):353–364, 2007. 17. Wang C, Hennessey JA, Kirkton RD, et al: Fibro- 8(4):631–633, 2011.
2. Erickson JR, Joiner ML, Guan X, et al: A dynamic blast growth factor homologous factor 13 regulates 32. Koval OM, Guan X, Wu Y, et al: CaV1.2 beta-
pathway for calcium-independent activation of Na+ channels and conduction velocity in murine subunit coordinates CaMKII-triggered cardio-
CaMKII by methionine oxidation. Cell 133(3):462– hearts/novelty and significance. Circ Res myocyte death and afterdepolarizations. Proc Natl
474, 2008. 109(7):775–782, 2011. Acad Sci U S A 107(11):4996–5000, 2010.
3. Poomvanicha M, Wegener JW, Blaich A, et al: 18. Shamgar L, Ma L, Schmitt N, et al: Calmodulin is 33. Hudmon A, Schulman H, Kim J, et al:
Facilitation and Ca2+-dependent inactivation are essential for cardiac IKS channel gating and assem- CaMKII tethers to L-type Ca2+ channels, estab-
modified by mutation of the Cav1.2 channel IQ bly: Impaired function in long-QT mutations. Circ lishing a local and dedicated integrator of Ca2+
motif. J Biol Chem 286(30):26702–26707, 2011. Res 98(8):1055–1063, 2006. signals for facilitation. J Cell Biol 171(3):537–547,
4. Wang HG, George MS, Kim J, et al: Ca2+/ 19. Ghosh S, Nunziato DA, Pitt GS: KCNQ1 assem- 2005.
calmodulin regulates trafficking of Ca(V)1.2 Ca2+ bly and function is blocked by long-QT syndrome 34. Erickson JR, He BJ, Grumbach IM, et al: CaMKII
channels in cultured hippocampal neurons. J Neu- mutations that disrupt interaction with calmodulin. in the cardiovascular system: Sensing redox states.
rosci 27(34):9086–9093, 2007. Circ Res 98(8):1048–1054, 2006. Physiol Rev 91(3):889–915, 2011.
5. Dick IE, Tadross MR, Liang H, et al: A modular 20. Wiener R, Haitin Y, Shamgar L, et al: The 35. Pitt GS: Calmodulin and CaMKII as molecular
switch for spatial Ca2+ selectivity in the calmodulin KCNQ1 (Kv7.1) COOH terminus, a multitiered switches for cardiac ion channels. Cardiovasc Res
regulation of CaV channels. Nature 451(7180):830– scaffold for subunit assembly and protein interac- 73(4):641–647, 2007.
834, 2008. tion. J Biol Chem 283(9):5815–5830, 2008. 36. Wagner S, Dybkova N, Rasenack EC, et al: Ca2+/
6. Kim EY, Rumpf CH, Van Petegem F, et al: Mul- 21. Balshaw DM, Xu L, Yamaguchi N, et al: Calmodu- calmodulin-dependent protein kinase II regulates
tiple C-terminal tail Ca2+/CaMs regulate CaV1.2 lin binding and inhibition of cardiac muscle cardiac Na+ channels. J Clin Invest 116(12):3127–
function but do not mediate channel dimerization. calcium release channel (ryanodine receptor). 3138, 2006.
EMBO J 29(23):3924–3938, 2010. J Biol Chem 276(23):20144–20153, 2001. 37. Hund TJ, Koval OM, Li J, et al: A beta(IV)-
7. Fallon JL, Baker MR, Xiong L, et al: Crystal struc- 22. Deleted in proof. spectrin/CaMKII signaling complex is essential for
ture of dimeric cardiac L-type calcium channel 23. Fruen BR, Bardy JM, Byrem TM, et al: Differen- membrane excitability in mice. J Clin Invest
regulatory domains bridged by Ca2+⋅calmodulins. tial Ca(2+) sensitivity of skeletal and cardiac muscle 120(10):3508–3519, 2010.
Proc Natl Acad Sci U S A 106(13):5135–5140, ryanodine receptors in the presence of calmodulin. 38. Ashpole NM, Herren AW, Ginsburg KS, et al:
2009. Am J Physiol Cell physiol 279(3):C724–C733, Ca2+/calmodulin-dependent protein kinase II
8. Mori MX, Erickson MG, Yue DT: Functional stoi- 2000. (CaMKII) regulates cardiac sodium channel
chiometry and local enrichment of calmodulin 24. Xu L, Meissner G: Mechanism of calmodulin inhi- NaV1.5 gating by multiple phosphorylation sites.
interacting with Ca2+ channels. Science bition of cardiac sarcoplasmic reticulum Ca2+ J Biol Chem 287(24):19856–19869, 2012.
304(5669):432–435, 2004. release channel (ryanodine receptor). Biophys J 39. Xiao L, Coutu P, Villeneuve LR, et al: Mechanisms
9. Navedo MF, Cheng EP, Yuan C, et al: Increased 86(2):797–804, 2004. underlying rate-dependent remodeling of transient
coupled gating of L-type Ca2+ channels during 25. Guo T, Zhang T, Mestril R: Bers DM. Ca2+/ putward potassium current in canine ventricular
hypertension and Timothy syndrome. Circ Res Calmodulin-dependent protein kinase II phos- myocytes. Circ Res 103(7):733–742, 2008.
106(4):748–756, 2010. phorylation of ryanodine receptor does affect 40. Sergeant GP, Ohya S, Reihill JA, et al: Regulation
10. Biswas S, DeSilvestre D, Tian Y, et al: calcium sparks in mouse ventricular myocytes. Circ of Kv4.3 currents by Ca2+/calmodulin-dependent
Calcium-mediated dual-mode regulation of cardiac Res 99(4):398–406, 2006. protein kinase II. Am J Physiol Cell Physiol
sodium channel gating. Circ Res 104(7):870–878, 26. Yamaguchi N, Takahashi N, Xu L, et al: 288(2):C304–C313, 2005.
2009. Early cardiac hypertrophy in mice with 41. Wagner S, Hacker E, Grandi E, et al: Ca/
11. Wang C, Chung BC, Yan H, et al: Crystal structure impaired calmodulin regulation of cardiac muscle calmodulin kinase II differentially modulates
of the ternary complex of a NaV C-terminal Ca release channel. J Clin Invest 117(5):1344– potassium currents. Circ Arrhythm Electrophysiol
domain, a fibroblast growth factor homologous 1353, 2007. 2(3):285–294, 2009.
factor, and calmodulin. Structure 20(7):1167–1176, 27. Ai X, Curran JW, Shannon TR, et al: Ca2+/ 42. Maier LS, Zhang T, Chen L, et al: Transgenic
2012. calmodulin-dependent protein kinase modulates CaMKIIdeltaC overexpression uniquely alters
12. Potet F, Chagot B, Anghelescu M, et al: Functional cardiac ryanodine receptor phosphorylation and cardiac myocyte Ca2+ handling: reduced SR Ca2+
interactions between distinct sodium channel cyto- sarcoplasmic reticulum Ca2+ leak in heart failure. load and activated SR Ca2+ release. Circ Res 92(8):
plasmic domains through the action of calmodulin. Circ Res 97(12):1314–1322, 2005. 904–911, 2003.
J Biol Chem 284(13):8846–8854, 2009. 28. Blaich A, Welling A, Fischer S, et al: Facilitation 43. Witcher DR, Kovacs RJ, Schulman H, et al:
13. Kim J, Ghosh S, Liu H, et al: Calmodulin mediates of murine cardiac L-type Cav1.2 channel is modu- Unique phosphorylation site on the cardiac ryano-
Ca2+ sensitivity of sodium channels. J Biol Chem lated by calmodulin kinase II-dependent phos- dine receptor regulates calcium channel activity.
279(43):45004–45012, 2004. phorylation of S1512 and S1570. Proc Natl Acad J Biol Chem 266(17):11144–11152, 1991.
14. Sarhan MF, Tung CC, Van Petegem F, et al: Crys- Sci U S A 107(22):10285–10289, 2010. 44. Wehrens XH, Lehnart SE, Reiken SR, et al: Ca2+/
tallographic basis for calcium regulation of sodium 29. Grueter CE, Abiria SA, Dzhura I, et al: L-type calmodulin-dependent protein kinase II phosphor-
channels. Proc Natl Acad Sci 109(9):3558–3563, Ca2+ channel facilitation mediated by phosphory- ylation regulates the cardiac ryanodine receptor.
2012. lation of the beta subunit by CaMKII. Mol Cell Circ Res 94(6):e61–e70, 2004.
15. Herzog RI, Liu C, Waxman SG, et al: Calmodulin 23(5):641–650, 2006. 45. Rodriguez P, Bhogal MS, Colyer J: Stoichiometric
binds to the C terminus of sodium channels Nav1.4 30. Brandmayr J, Poomvanicha M, Domes K, et al: phosphorylation of cardiac ryanodine receptor on
and Nav1.6 and differentially modulates their func- Deletion of the C-terminal phosphorylation sites serine 2809 by calmodulin-dependent kinase II and
tional properties. J Neurosci 23(23):8261–8270, in the cardiac beta subunit does not affect the protein kinase A. J Biol Chem 278(40):38593–
2003. basic β-adrenergic response of the heart and the 38600, 2003.
46. Wehrens XH, Lehnart SE, Reiken S, et al: Ryano- release channel of sarcoplasmic reticulum from calmodulin-dependent protein kinase II promotes
dine receptor/calcium release channel PKA phos- cardiac muscle. J Biol Chem 270(5):2074–2081, life-threatening ventricular arrhythmias in mice
phorylation: a critical mediator of heart failure
progression. Proc Natl Acad Sci U S A 103(3):511–
518, 2006.
50.
1995.
Anderson ME, Brown JH, Bers DM: CaMKII in
myocardial hypertrophy and heart failure. J Mol
with heart failure. Circulation 122(25):2669–2679,
2010.
54. Yamaguchi N, Takahashi N, Xu L, et al: Early
19
47. Kushnir A, Shan J, Betzenhauser MJ, et al: Role of Cell Cardiol 51(4):468–473, 2011. cardiac hypertrophy in mice with impaired calmod-
CaMKIIdelta phosphorylation of the cardiac 51. Chelu MG, Sarma S, Sood S, et al: Calmodulin ulin regulation of cardiac muscle Ca2+ release
ryanodine receptor in the force frequency relation- kinase II-mediated sarcoplasmic reticulum Ca2+ channel. J Clin Invest 117(5):1344–1353, 2007.
ship and heart failure. Proc Natl Acad Sci U S A leak promotes atrial fibrillation in mice. J Clin 55. Wehrens XHT, Lehnart SE, Reiken SR, et al:
107(22):10274–10279, 2010. Invest 119(7):1940–1951, 2009. Ca2+/calmodulin-dependent protein kinase II
48. Respress,JL, van Oort RJ, Li N, et al: Role of 52. Li N, Wang T, Wang W, et al: Inhibition of phosphorylation regulates the cardiac ryanodine
RyR2 phosphorylation at S2814 during heart CaMKII phosphorylation of RyR2 prevents induc- receptor. Circ Res 94(6):e61–e70, 2004.
failure progression. Circ Res 110(11):1474–1483, tion of atrial fibrillation in FKBP12.6 knockout
2012. mice. Circ Res 110(3):465–470, 2012.
49. Hain J, Onoue H, Mayrleitner M, et al: Phos- 53. van Oort RJ, McCauley MD, Dixit SS, et al:
phorylation modulates the function of the calcium Ryanodine receptor phosphorylation by calcium/
Macromolecular Complexes and
Cardiac Potassium Channels 20
Stéphane Hatem and Elise Balse
interacts with the α-subunit (stoichiometry of 1α-1β) to regulate

CHAPTER OUTLINE
channel trafficking. However, this chaperone-like property of
The Four General Classes of Accessory Subunits 197 the Kvβ-subunits has not been found for all Kv channels6,7
(Figure 20-1).
The MAGUK Proteins 199
The most dramatic effect of Kvβ on the voltage-dependent
Membrane Lipids and Potassium Channel Complexes 200 outward current is to accelerate its rate of inactivation. Kv chan-
nels inactivate through two main mechanisms: the slow C-type
Conclusion 201
inactivation and the rapid N-type inactivation called the ball-and-
chain mechanism.8 An example of a Kv channel undergoing C-type
inactivation is Kv1.5, which does not have an N-terminal domain
In the heart, the main repolarizing currents are carried by potas- acting as an open channel blocker. Kv1.5 is inactivated through
sium ions, a number of potassium currents activate during the a mechanism that involves slow conformational changes of the
action potential (AP). The inward rectifier K+ current is crucial outer mouth of the pore, referred to as C-type inactivation. The
for maintaining the membrane potential near the equilibrium coexpression of Kvβ1 with Kv α-subunits with C-type inactiva-
potential for potassium, which is approximately –90 mV during tion results in a fast inactivating current with increased mem-
diastole. The voltage-dependent outward currents regulate the brane depolarization. For example, the Kv1.5-encoded current,
activation and termination of the plateau phase of the AP while one main component of IKur in atrial myocytes, is transformed by
delayed rectifier currents control the final repolarization phase of Kvβ1.3 subunits in a transient outward Ito-type current.9 This
the action potential. The molecular bases of potassium currents effect on current kinetic is associated with a shift of the voltage-
are diverse, with more than 40 genes encoding potassium chan- dependent activation and inactivation to more negative poten-
nels expressed in the heart. In addition, a number of auxiliary tials. The effect of Kvβ1 on channel inactivation is mediated by
units and protein partners contribute to the trafficking and their N-terminal domain that blocks the inner cavity of the pore
anchoring of potassium channels at the plasma membrane, and of the α-subunit resembling the ball-and-chain process.8,10 In
to their organization in macromolecular complexes. These part- addition, by binding to the C-terminus of the channels, Kvβ can
ners confer important properties to potassium currents contrib- accelerate the C-type inactivation.11 In heterologous expression
uting to the plasticity of cardiac electrical properties both in systems, coexpression of Kvβ1.3 with Kv1.5 is necessary for
normal conditions and during cardiac diseases. Proteins and the PKA (cAMP-dependent protein kinase)-mediated increase in
structural lipids of the plasma membrane appear as major part- K+ current. Moreover, PKC has little effect on Kv1.5 channels
ners for potassium channels. This chapter will focus on the alone; however, when coexpressed with the Kvβ1.2, the
description of partners involved in the formation of macromo- protein kinase reduces the K+ current.12 These observations
lecular potassium channel complexes and their role in cardiac are consistent with the presence of multiple phosphorylation sites
excitability. on the β- and α-subunits13 and could provide an explanation
for the effects of β-adrenergic or PKC stimulation on IKur in
human atrial myocytes.14 The characteristics of repetitive mem-
brane depolarizations (i.e., their duration and frequency)
The Four General Classes can modify persistently the rate of inactivation of Ikur in human
of Accessory Subunits atrial myocytes. This effect is modulated by the activation of
CaMKII (calcium-calmodulin dependent protein kinase-II) and
Kvβ Family could also involve the interaction between Kvβ and the Kvα1.5
subunits.15
The first identified auxiliary subunits of potassium channels are Another illustration of the important role played by the regu-
the so-called Kvβ–subunits, which associate with the pore- lation of Kv channels by Kvβ is provided by the study on phar-
forming Kv channels (referred as α-subunits). Initially purified macological properties of the Kv1.5 channel. The contribution
from bovine brain,1,2 nine β-subunits encoded by four genes have of Ikur to shortening of the AP during atrial fibrillation16-19and the
been identified. In mammals, three genes are expressed in the fact that Kv1.5 channel is more abundantly expressed in atrial
heart and variants from alternative splicing are also found. Kvβ- than ventricle myocardium, explain the effort to develop selective
subunits are localized in the cytosol with a conserved carboxy- IKur blockers as potential atria-specific antiarrhythmic agents.
terminal and a variable amino-terminal. Only the Kvβ1.1, Kvβ1.2, Some of these molecules compete with β-subunits to bind to the
and Kvβ1.3 are expressed in the heart.3 inner cavity of the Kv1.5 channel pore and are responsible for an
Two main roles have been described for Kvβ subunits. It has apparent open channel block of Ikur. For example, Kvβ1.3
been reported that Kvβ1 and Kvβ3 associate early with α-subunits decreases drug affinity of Kv1.5 for local anesthetic and antiar-
during their biosynthesis in the endoplasmic reticulum and rhythmic drugs (e.g., bupivacaine,20,21 AVE0118,22 vernakalant23).
exert a chaperone-like effect, facilitating their stable expression In atrial myocytes as well, the analog of tedisamil (i.e., bertosamil)
at the plasma membrane.4,5 The C-terminal domain of Kvβ accelerates the rate of inactivation of the outward current with
197
+
+
MIRP1
DPPX
S1 S2 S3 S4 S5 S6 Sarcolemma
+
+
KChAP
Kvβ
SAP97
KChIP
DPPX MIRP1
Golgi Network
KChIP
KChAP
PSD95 Kvβ
ER
Figure 20-1. Main partners of Kv4 channels from their synthesis and assembly to their organization in macromolecular complexes at the plasma membrane of cardiac
myocytes. DPP, Dipeptidyl aminopeptidase-protein; ER, endoplasmic reticulum; KChAP, K channel associated protein; KChIP, K channel interacting protein; MIRP1, MinK related
peptide.
membrane depolarization, consistent with a competition between delayed-type current—Iks.27,28 MinK can interact with other
the drug and endogenous Kvβ at the inner face of the pore.24 channels like hERG (or Kv11.1), which is responsible for the
Finally, an oxidoreductase activity has been reported for Kvβ activation of the rapid delayed rectifier, Ikr, but this observation
subunits as indicated by the observation that Kvβ2.1 confers was obtained only in heterologous expression systems. In mice,
sensitivity to oxygen to Kv4.2.25 This enzymatic activity is likely deletion of KCNE1 is associated with an impaired QT-RR adapt-
due to the presence of a binding site for the cofactor NADP+ and ability on the electrocardiogram (ECG), prolongation of epicar-
catalytic domains; however, in vivo data are lacking. dial AP duration, and frequent episodes of atrial tachycardia29,30
indicating the physiologic importance of this ancillary subunit in
cardiac electrophysiology.
MinK and MinK-Related Proteins Four other peptides belonging to the minK family have been
identified and called MinK- and MinK-related protein (MiRP;
Originally thought to be the α-subunit of the delayed potassium for MinK related peptide, KCNE2-5).7 In the heart, MiRP1
current Iks, minK (for minimal K channel protein)26 is a small (KCNE2) is expressed mainly in the nodal tissue and Purkinje
transmembrane protein (14–20 kDa) encoded by KCNE1. The cells. In vitro, MiRP1 can interact with several potassium chan-
first interaction described for minK is with the Kv7.1 channel nels, including Kv7.1,31 Kv4,32 and hERG.33 The hERG channel
(KvLQT1). Expressed alone, Kv7.1 underlies a voltage-dependent is regulated by cyclic adenosine monophosphate (cAMP)-
outward current of low amplitude. In contrast, when coexpressed dependent protein kinase, PKA, which increases the current
with minK, this channel is responsible for a large, slow, amplitude through the shift of the voltage-dependent activation
Macromolecular Complexes and Cardiac Potassium Channels 199
of the channels. This pathway is an important target for sympa- A third partner that contributes to the formation of the Kv4
20
thetic regulation and the subsequent adaptation of cardiac repo- macromolecular complex is the transmembrane dipeptidyl
larization to increased heart rate. It has been reported that aminopeptidase-protein (DPP). This integral transmembrane
KCNE2 facilitates the PKA regulation of Ikr by stabilizing hERG protein interacts with the extracellular matrix, and either sepa-
in its phosphorylated form.34 Transgenic models have contributed rately or jointly with KChIP associates with Kv4 subunits at the
to establish the physiologic role of MiRP1 in the heart. Mice with N-terminus to modulate the trafficking and targeting of the
specific deletion of the KCNE2 gene show prolonged ventricular channel46 (see Figure 20-1).
APs and reduced density of the fast and slow components of the
outward potassium current in keeping with KCNE2 interacting
with various potassium channels.7,35 K Channel Associated Protein
Genetic studies of patients suffering from inherited or drug-
induced long QT syndrome have provided also important infor- Only a few studies have been conducted on the K channel associ-
mation on the physiologic role of this family of potassium ated protein (KChAP). This small protein of 574 amino acids
partners. Mutations of genes encoding minK and MiRP1 have expresses in the heart. It is a cytoplasmic subunits that interact
been found during the familial form of long QT syndrome, transiently with several voltage-dependent Kv channels, includ-
LQT5 and LQT6, respectively.33 Some mutations of KCNE2 are ing Kv1.x, Kv2.1, and Kv4.3.47,48 In vitro, KChAP upregulates Kv
responsible for the decrease in Ikr due the acceleration of hERG channel’s surface expression through a chaperone effect47,48 (see
inactivation. This results in less repolarizing current during the Figure 20-1).
termination of the AP causing the prolongation of the phase 3
and a risk of triggered activities.36 Other mutations have been
shown to modify the pharmacologic profile of hERG providing
a molecular explanation for iatrogenic long QT syndrome and The MAGUK Proteins
torsade de pointe33,37,38 (see Figure 20-1).
In 1995, Kim et al.49 reported that neuronal Shaker channels are
clustered by a protein (PSD95) belonging to a family of multi-
K+ Channel Interacting Protein domain membrane proteins, called the membrane-associated
guanylate kinase homologs (MAGUK proteins). Since that early
K+ channel interacting protein (KChIP) is another important study was published, a number of channels have been found to
family of ancillary subunits of K+ channels. They are encoded by interact with MAGUK proteins in different tissues, including the
four genes, KChIP1-4 generated and alternative splicing. KChIP heart.49a These anchoring proteins are viewed as central organiz-
isoforms differ by their N-terminus, whereas they share 70% of ers of specialized plasma membrane domains, such as the inter-
homology, notably with four conserved “EF-hand-like” domains calated disc of cardiac myocytes.
that bind Ca2+.39 Only KChIP2 is expressed in the heart. The The presence of several protein-protein interacting domains
C-terminus of KChIP2 interacts directly at the cytosolic face of is the hallmark of MAGUK proteins such as an Src homology 3
the membrane with the N-terminus of Kv4 channels, character- domain, a GUK domain with homology to the enzyme guanylate
ized by a rapid N-terminal inactivation. In expression systems, kinase, and one or several Post Synaptic Density Protein, Dro-
the interaction of KChIP with Kv4 α-subunits is associated with sophila disc large tumor suppressor, Zonula occludens (PDZ)
a marked increase in the amplitude of Ito,f,39,40 an effect attributed domain(s). Most MAGUK proteins contain three PDZ domains.
to the capacity of KChIP to prevent the retention of Kv4 The PDZ domain of MAGUK proteins expressed in the heart
α-subunits in the endoplasmic reticulum (ER).40 The other effect recognizes a C-terminal–Ser/thr-X-ψ–Val sequence, where X
of KChIP on Kv4 channels is to regulate their rate of inactivation refers to any amino acid and ψ refers to an hydrophobic amino
and recovery from inactivation.39,40 Of note, the three cardiac acid50,51 (Figure 20-2).
isoforms of KChIP2 differ in their effects on the functional Several MAGUK proteins are expressed in the heart: SAP97,
expression and gating properties of potassium channels and ZO-1, CASK, and MAGI3. For example, the gap junction
might not have the same regulatory effects on cardiac electro- channel, connexin 43, interacts with zonula occludens-1 (ZO-1),
physiology7 (see Figure 20-1). Endocardial myocytes are charac- which regulates its localization at the intercalated disc and prob-
terized by APs of longer duration than epicardial APs. This ably also its internalization.52-54 An additional cardiac MAGUK
gradient of repolarization from endocardium to epicardium is protein that has been studied extensively is SAP97. In atrial and
essential for the oriented propagation of the depolarization wave, ventricular myocytes, SAP97 is predominantly localized at the
the prevention of retrograde activation of the myocardium, and intercalated disc.55-60 In vitro, it has been shown that SAP97 is
the risk of arrhythmogenic reentry. In mammals, the epicardial- targeted exclusively at the level of myocyte-myocyte contacts in
to-endocardiac gradient is due to regional differences in density membrane domains, such as adherens and gap junctions.61 SAP97
of the fast component of the transient outward current, Ito,f, the is also present at the adrenergic synapse in myocytes, facing the
highest density of Ito,f has been observed in myocytes of the epi- nerve endings; therefore, it is part of the β-adrenergic signaling
cardial layer.41,42 In large mammals such as humans, Kv4.3 is the complex.62 These results indicate that SAP97 is an important
predominant molecular basis of Ito,f. This channel is homoge- determinant of the targeting of potassium channels in specialized
nously distributed in the ventricular wall and cannot explain the cell-cell contact domains of the sarcolemma.
gradient of Ito,f.41,42 In contrast, both at the transcript and protein The MAGUK SAP97 interacts with several cardiac ion chan-
levels, the concentration of KChIP2 increases from endocardium nels including Nav1.5,60 Kir2.x, Kv4.x, Kv1.5, HCN-2, and
to epicardium. This gradient is likely to contribute to the gradual HCN-4.* In most cases, SAP97 interacts with the consensus
expression of Ito,f in human and dog ventricles.43,44 In rodents, the PDZ binding domain located at the C-terminus of ion channels.
gradient of repolarization is not due to KChIP, but to the gradual Besides this classical interaction, indirect interactions involving
expression of Kv4.2 between the endocardium and epicardium. the N-terminus of the channel, SAP97, and α-actinin have
However, despite such species specificities, mouse models have been reported and can also contribute to the role of SAP97 in
contributed to establish the importance of KChIP2 in the normal
activation of voltage-dependent outward current Ito, which is
suppressed in KChIP2 knockout mice.45 *References 49, 55, 56, 59, 60, 63.
Kv1.5
Control myocyte
L27 PDZ PDZ PDZ SH3 Guk
A
Kv1.5 SAP97 Merge
Ad-SAP97-infected myocyte
GuK HK SH3 SH3
4.1 PDZ 4.1

4.1 PDZ
4.1
C
SH3
GuK 4.1 4.1

120
100
Relative fluorescence
Kv1.5
80
Spectrin/actin Spectrin/actin
intensity
60
B Kv1.5 + Ad-SAP97
40
20
0
D 0 50 100 150 200 250 300
Figure 20-2. The anchoring protein SAP97 of the MAGUK family. A, Molecular organization of MAGUK proteins. B, Scheme of possible interactions of MAGUK protein and
with potassium channels. C, The overexpression of SAP97 (Ad-SAP97) is associated with the accumulation of GFP-tagged Kv1.5 channels at the level of plasma membrane.
D, The immobilization of Kv1.5 channel in cardiac myocytes.
(From Abi-Char J, El-Haou S, Balse E, et al: The anchoring protein SAP97 retains Kv1.5 channels in the plasma membrane of cardiac myocytes. Am J Physiol Heart Circ Physiol
294:H1851–H1861, 2008.)
the organization of potassium channels in macromolecular By its capacity to oligomerize, SAP97 contributes to the for-
complexes.64,65 mation of channel networks that can gather different ion channel
In both native myocytes and heterologous expression systems, families, as suggested by the recent observations indicating the
it has been shown that SAP97 enhances the functional expression role of SAP97 in the formation of Nav1.5/Kir2.1 complexes.68-70
of ion channels, resulting in the upregulation of corresponding Finally, the ability to interact with multiple proteins and to
currents. By means of single-channel recording, it has been pos- form networks confers SAP97 a role of linker between ion
sible to directly record the increased activity of single Kv1.5 channels or receptors and signaling pathways71-73 Concerning
channel activity in cells overexpressing SAP97.61 In keeping with the potassium channels, such a role has been described for the
the accumulation of α-subunits in the membrane protein fraction G-protein regulation of the inward rectifier K+ channel74 and
of cardiac myocytes overexpressing SAP97, these data indicate the regulation of Kv4 channels by CaMKII.55
that SAP97 enhances the density of channels at the plasma
membrane.66
SAP97 can enhance the functional expression of ion channels Membrane Lipids and Potassium
through various mechanisms, including retention, increased Channel Complexes
forward trafficking, aggregation, and formation of signalosomes.
For example, in cardiac myocytes overexpressing SAP97, potas- Potassium channels are embedded in a lipid environment, which
sium channels are immobilized and concentrated at the level of can constitute an important modulator for channels activity.
myocyte-myocyte contacts (see Figure 20-2). Another study also Lipids can regulate potassium channels through several mecha-
reported that SAP97 inhibits the endocytosis of potassium chan- nisms, including the conformation changes of the protein, the
nels, thereby promoting their accumulation at the plasma modulation of the delivery of channels to the plasma membrane,
membrane.67 and their clustering in macromolecular complexes.
However, it is possible that in native cardiac myocytes, Kv1.5

Lipids and the Biophysical Properties
20
channels are only transiently localized in lipid rafts, during their
of Potassium Channels trafficking from delivery sites in the plasma membrane to their
final targeting into specialized domains.87,90
There is an interrelation among membrane stiffness, bilayer
deformation energy, and changes in channel conformation.75 For
example, the effects of cholesterol on the open probability of Kir Membrane Cholesterol and the Delivery of
or Ca2+ channels are due to an increase in bilayer stiffness and in Potassium Channels
the energy necessary for the transition of channels from the
closed to the open state.75-77 Using the paddle chimera voltage- In atrial myocytes, the reduction of membrane cholesterol
dependent (Kv) channel, it has been shown that the rate of content using β–MCD (β-methylcyclodextrin) is associated with
channel opening, the voltage dependence of open probability, a slow increase in the outward potassium current. This cholesterol-
and the maximum open-probability level achievable vary with the sensitive current, inhibited by low concentration of 4-
lipid membrane composition.78 The explanation is that move- aminopyridine (a blocker of Kv channels), activates concomitantly
ment of the voltage sensor of the Kv channel depends on the with the progressive increase in the activity of single Kv1.5 chan-
fluidity or on the mechanical state of the plasma membrane, nels.66,91 Moreover, the depletion of membrane cholesterol is
hence indirectly on lipid composition.78,79 associated with the appearance of Kv1.5 channel clusters at the
There are also reports on direct modulation of channel func- plasma membrane, suggesting an effect of the lipid on channel
tion by cholesterol. For example, cholesterol inhibits Kir2 chan- delivery.66,91,92 These results suggest that cholesterol depletion
nels by stabilizing them in the closed state, rendering them silent. causes the recruitment of channels at the plasma membrane. This
This regulation is mediated via an interaction of cholesterol with suggestion is supported by the finding that cholesterol depletion
a specific region of the cytosolic C-terminal domain of the causes the accumulation of active channels in the plasma
channel. The optical isomer of cholesterol—epicholesterol, membrane. These newly recruited channels might arise from
whose effects on membrane biophysical properties are similar to endogenous pools of channels localized beneath the plasma
cholesterol—fails to modulate Kir2 channels, indicating specific- membrane.91 Ion channels and receptors are delivered to the
ity in cholesterol-channel interactions.80 This finding is reminis- plasma membrane from intracellular compartments exiting the
cent of observations that other membrane lipids, such as Golgi network (de novo synthesis) or from recycling pathways.
arachidonic acid, can induce rapid inactivation in otherwise non Several Rab GTP-ases regulate the fate of endosome trafficking,
inactivating Kv channels by binding to a specific site near the such as recycling or degradation. For example, Rab11 regulates
selectivity filter.81 the recycling endosome and has been shown to be involved in
the effect of cholesterol on Kv1.5 channels recruitment.91 Rab11-
mediated recycling has been described for other potassium chan-
Lipid Rafts and the Clustering nels, such as KCNQ1/KCNE193 or cardiac pacemaker channels
of Potassium Channels HCN2 and HCN4,94 as well as for the glucose transporter
GluT4.95 In addition, the fusion of endosome vesicles with the
Cholesterol and sphingolipid, two structural lipids of the plasma plasma membrane (exocytosis) requires the formation of soluble
membrane, pack together to form cholesterol-enriched domains N-ethylmaleimide–sensitive factor attachment protein receptor
referred as lipid rafts. A subset of lipid rafts that form small (SNARE) complexes. Indeed, the inhibition of SNARE complex
invaginations of the plasma membrane are named caveolae and formation suppresses the effects of cholesterol of Kv1.5 encoded
contain the scaffolding protein caveolin. Caveolae are particu- current in atrial myocytes.91 In pancreatic α-cells, cholesterol
larly abundant in cardiac myocytes.82 Lipid rafts are considered depletion causes the dissociation between Kv4.x channels and
as platforms where protein are delivered and clustered together SNAREs, leading to a decrease in the A-type potassium current.96
with signaling pathways to form dynamic macromolecular com- Less known is the role of cholesterol as a regulator of the
plexes. Several potassium channels are localized in lipid rafts, endocytosis of ion channels. Caveolae is an important route of
notably in caveolae.83-85 The importance of lipid raft integrity for endocytosis for many proteins in the myocardium. As already
the normal functional expression of potassium channels is illus- described, cholesterol content regulates the shape and size of
trated by the activation of the acetylcholine-dependent inward lipid rafts; for example, the depletion of cholesterol causes the
rectifier potassium current, IKACh, formed of GIRK (G-protein flattening of caveolae.97-100 This finding provides an explanation
gated inwardly-rectifying potassium) subunits. Attachment of of why drugs that sequester cholesterol, such as filipin, nystatin,
acetylcholine to its inhibitory G-protein (Gi)-coupled muscarinic and cyclodextrin, block the endocytosis mediated by caveolae.
2 (M2) receptor leads to the release of Gβγ subunits and the activa- More recently, cholesterol has been reported to play a role in
tion of GIRK channels.86 Other receptors coupled to stimulatory internalization through the clathrin-mediated endocytosis path-
G-protein (Gs), such as the β-adrenergic receptor, fail to activate ways.101,102 Given that the endocytosis rate would control the
GIRK channels. In contrast to Gs, which localizes exclusively in number of potassium channels in the sarcolemma, any alteration
raft fractions, the M2 signaling pathway is localized in both lipid of the endocytosis step should profoundly affect the properties
raft and non–lipid raft fractions. Therefore, one explanation for of potassium currents in cardiac myocytes (Figure 20-3).
the specific effect of Gi on GIRK could be that distinct localiza-
tion in lipid raft of signaling cascades acts as a mechanism of
selectivity, such that free Gβγ released from lipid rafts is much
less efficient in activating GIRK. However, there is controversy Conclusion
between studies conducted in heterologous expression systems
versus native myocytes on the localization of potassium The classical view of well-individualized potassium currents being
channels in lipid rafts and notably caveolae. In the atrial and the functional expression of distinct potassium channels constitu-
ventricular myocardium, Kv1.5 channel is not targeted to caveo- tively expressed at the surface of the cells is challenged by the
lae contrary to chinese hamster ovary (CHO) cells. This channel molecular dissection of potassium channel complexes. The emerg-
is present predominantly at the level of the intercalated discs ing picture is that potassium channels activity is the result of con-
where caveolin-3 is expressed poorly.66,87,88 Moreover, the two tinuing and dynamic processes of protein and lipid interaction
proteins cannot be coprecipitated from cardiac protein samples.89 going from the synthesis of the α-subunit to its specific targeting
Exocytosis Cholesterol
PM
t-SNAREs
v-SNARE SE/EE
Rab11
Kv1.5 Recycling Rab4/5 Endocytosis
Rab8
RE
Rab11
TGN LE
Rab7/9
ER
de novo synthesis Degradation
Figure 20-3. The recycling of Kv channels in cardiac myocytes. EE, Early endosome; ER, endoplasmic reticulum; PM, plasma membrane; RE, recycling endosome; SE, sorting
endosome; TGN, trans-Golgi network.
(From Balse E, El-Haou S, Dillanian G, et al: Cholesterol modulates the recruitment of Kv1.5 channels from Rab11-associated recycling endosome in native atrial myocytes. Proc
Natl Acad Sci U S A 106:14681–14686, 2009.)
into submembrane domains of the sarcolemma for interaction their role in activation and propagation of the electrical activity.
with macromolecular complexes and proper function. This This concept is suggested by the recent discovery of Nav1.5/Kir
complex regulation over time (turnover) and space (tethering) sug- complexes at the lateral membrane68 and Cx43/Nav1.5 complexes
gests an important plasticity for these protein complexes, which at the level of the intercalated disc.105 Many more studies con-
could be a major determinant of physiologic adaptation of cardiac ducted on cardiac myocytes or using experimental models are
excitability and of pathogenesis of electrical disorders. It is possible needed to establish new schemes of cardiac electrophysiology
that potassium channels are recruited or recycled in response to integrating this large diversity of molecular actors involved in the
various physiologic stimuli or during pathologic conditions, as in formation of potassium channels. Progress also will come from
other excitable tissues. For example, Guo et al.103 reported that the the development of new tools to study protein processing, assem-
external potassium concentration can modulate the availability of bly, and interactions in native myocytes.
hERG channel; for example, low external potassium increases
channel internalization, resulting in the suppression of IKr.103 This
phenomenon could contribute to the prolongation of the QT
interval during hypokalemia. Antiarrhythmic agents such as quini- Acknowledgments
dine can also accelerate the internalization of Kv1.5 channels
through a dynamin- and dynein-dependent process.104 This work was supported by Fondation Leducq “Structural
Another emerging concept is the heterogeneity of macromo- Alterations in the Myocardium and the Substrate for Cardiac
lecular channel complexes, which can be composed of α-subunits Fibrillation” and the European Network for Translational
of different types depending on their membrane localization and Research in Atrial Fibrillation (EUTRAF-261057).
7. Pongs O, Schwarz JR: Ancillary subunits associ- 14. Li GR, Yang B, Feng J, et al: Transmembrane ICa
References ated with voltage-dependent K+ channels. Physiol contributes to rate-dependent changes of action
Rev 90:755–796, 2010. potentials in human ventricular myocytes. Am J
1. Parcej DN, Scott VE, Dolly JO: Oligomeric 8. Rasmusson RL, Morales MJ, Wang S, et al: Inac- Physiol 276:H98–H106, 1999.
properties of alpha-dendrotoxin-sensitive potas- tivation of voltage-gated cardiac K+ chan- 15. Tessier S, Godreau D, Vranckx R, et al: Cumula-
sium ion channels purified from bovine brain. nels. Circ Res 82:739–750, 1998. tive inactivation of the outward potassium current:
Biochemistry 31:11084–11088, 1992. 9. Uebele VN, England SK, Gallagher DJ, et al: a likely mechanism underlying electrical memory
2. Scott VE, Rettig J, Parcej DN, et al: Primary Distinct domains of the voltage-gated K+ channel in human atrial myocytes. J Mol Cell Cardiol
structure of a beta subunit of alpha-dendrotoxin- Kv beta 1.3 beta-subunit affect voltage-dependent 33:755–767, 2001.
sensitive K+ channels from bovine brain. Proc gating. Am J Physiol 274:C1485–95, 1998. 16. Dobrev D, Carlsson L, Nattel S: Novel molecular
Natl Acad Sci U S A 91:1637–1641, 1994. 10. Rettig J, Heinemann SH, Wunder F, et al: Inac- targets for atrial fibrillation therapy. Nat Rev
3. England SK, Uebele VN, Kodali J, et al: A novel tivation properties of voltage-gated K+ channels Drug Discov 11:275–291, 2012.
K+ channel beta-subunit (hKv beta 1.3) is pro- altered by presence of beta-subunit. Nature 17. Hatem SN, Coulombe A, Balse E: Specificities of
duced via alternative mRNA splicing. J Biol Chem 369:289–294, 1994. atrial electrophysiology: Clues to a better under-
270:28531–28534, 1995. 11. Morales MJ, Wee JO, Wang S, et al: The standing of cardiac function and the mechanisms
4. Shi G, Nakahira K, Hammond S, et al: Beta sub- N-terminal domain of a K+ channel beta subunit of arrhythmias. J Mol Cell Cardiol 48:90–95, 2010.
units promote K+ channel surface expression increases the rate of C-type inactivation from the 18. Le Grand BL, Hatem S, Deroubaix E, et al:
through effects early in biosynthesis. Neuron cytoplasmic side of the channel. Proc Natl Acad Depressed transient outward and calcium cur-
16:843–852, 1996. Sci U S A 93:15119–15123, 1996. rents in dilated human atria. Cardiovasc Res
5. Nagaya N, Papazian DM: Potassium channel 12. Williams CP, Hu N, Shen W, et al: Modulation 28:548–556, 1994.
alpha and beta subunits assemble in the endoplas- of the human Kv1.5 channel by protein kinase C 19. Ravens U, Wettwer E: Ultra-rapid delayed recti-
mic reticulum. J Biol Chem 272:3022–3027, activation: role of the Kvbeta1.2 subunit. J Phar- fier channels: molecular basis and therapeutic
1997. macol Exp Ther 302:545–550, 2002. implications. Cardiovasc Res 89:776–785, 2011.
6. Leicher T, Roeper J, Weber K, et al: Structural 13. Kwak YG, Hu N, Wei J, et al: Protein kinase A 20. Arias C, Guizy M, David M, et al: Kvbeta1.3
and functional characterization of human potas- phosphorylation alters Kvbeta1.3 subunit- reduces the degree of stereoselective bupivacaine
sium channel subunit beta 1 (KCNA1B). Neuro- mediated inactivation of the Kv1.5 potassium block of Kv1.5 channels. Anesthesiology 107:641–
pharmacology 35:787–795, 1996. channel. J Biol Chem 274:13928–13932, 1999. 651, 2007.
21. Gonzalez T, Arias C, Caballero R, et al: Effects 41. Nerbonne JM, Guo W: Heterogeneous expres- Nav1.5 in cardiomyocytes. Circ Res 108:294–304,
of levobupivacaine, ropivacaine and bupivacaine sion of voltage-gated potassium channels in the 2011.
on HERG channels: stereoselective bupivacaine
block. Br J Pharmacol 137:1269–1279, 2002.
22. Decher N, Kumar P, Gonzalez T, et al: Binding
heart: roles in normal excitation and arrhythmias.
J Cardiovasc Electrophysiol 13:406–409, 2002.
42. Oudit GY, Kassiri Z, Sah R, et al: The molecular
61. Abi-Char J, El-Haou S, Balse E, et al: The
anchoring protein SAP97 retains Kv1.5 channels
in the plasma membrane of cardiac myocytes. Am
20
site of a novel Kv1.5 blocker: a “foot in the door” physiology of the cardiac transient outward potas- J Physiol Heart Circ Physiol 294:H1851–H1861,
against atrial fibrillation. Mol Pharmacol sium current (I(to)) in normal and diseased 2008.
70:1204–1211, 2006. myocardium. J Mol Cell Cardiol 33:851–872, 62. Shcherbakova OG, Hurt CM, Xiang Y, et al:
23. Eldstrom J, Wang Z, Xu H, et al: The molecular 2001. Organization of beta-adrenoceptor signaling
basis of high-affinity binding of the antiarrhyth- 43. Deschenes I, DiSilvestre D, Juang GJ, et al: Regu- compartments by sympathetic innervation of
mic compound vernakalant (RSD1235) to lation of Kv4.3 current by KChIP2 splice variants: cardiac myocytes. J Cell Biol 176:521–533,
Kv1.5 channels. Mol Pharmacol 72:1522–1534, a component of native cardiac I(to)? Circulation 2007.
2007. 106:423–429, 2002. 63. Leonoudakis D, Mailliard W, Wingerd K, et al:
24. Tessier S, Rucker-Martin C, Mace L, et al: The 44. Rosati B, Grau F, Rodriguez S, et al: Concordant Inward rectifier potassium channel Kir2.2 is asso-
antiarrhythmic agent bertosamil induces inactiva- expression of KChIP2 mRNA, protein and tran- ciated with synapse-associated protein SAP97.
tion of the sustained outward K+ current in sient outward current throughout the canine ven- J Cell Sci 114:987–998, 2001.
human atrial myocytes. Br J Pharmacol 122:291– tricle. J Physiol 548:815–822, 2003. 64. Eldstrom J, Choi WS, Steele DF, et al: SAP97
301, 1997. 45. Kuo HC, Cheng CF, Clark RB, et al: A defect in increases Kv1.5 currents through an indirect
25. Perez-Garcia MT, Lopez-Lopez JR, Gonzalez C: the Kv channel-interacting protein 2 (KChIP2) N-terminal mechanism. FEBS Lett 547:205–211,
Kvbeta1.2 subunit coexpression in HEK293 cells gene leads to a complete loss of I(to) and confers 2003.
confers O2 sensitivity to kv4.2 but not to Shaker susceptibility to ventricular tachycardia. Cell 65. Mathur R, Choi WS, Eldstrom J, et al: A specific
channels. J Gen Physiol 113:897–907, 1999. 107:801–813, 2001. N-terminal residue in Kv1.5 is required for
26. Takumi T, Ohkubo H, Nakanishi S: Cloning of a 46. Nadal MS, Ozaita A, Amarillo Y, et al: The upregulation of the channel by SAP97. Biochem
membrane protein that induces a slow voltage- CD26-related dipeptidyl aminopeptidase-like Biophys Res Commun 342:1–8, 2006.
gated potassium current. Science 242:1042–1045, protein DPPX is a critical component of neuronal 66. Abi-Char J, Maguy A, Coulombe A, et al: Mem-
1988. A-type K+ channels. Neuron 37:449–461, 2003. brane cholesterol modulates Kv1.5 potassium
27. Barhanin J, Lesage F, Guillemare E, et al: K(V) 47. Kuryshev YA, Gudz TI, Brown AM, et al: KChAP channel distribution and function in rat cardio-
LQT1 and lsK (minK) proteins associate to form as a chaperone for specific K(+) channels. Am J myocytes. J Physiol 582:1205–1217, 2007.
the I(Ks) cardiac potassium current. Nature Physiol Cell Physiol 278:C931–41, 2000. 67. Ting YK, Morikawa K, Kurata Y, et al: Transcrip-
384:78–80, 1996. 48. Wible BA, Yang Q, Kuryshev YA, et al: Cloning tional activation of the anchoring protein SAP97
28. Sanguinetti MC, Curran ME, Zou A, et al: Coas- and expression of a novel K+ channel regulatory by heat shock factor (HSF)-1 stabilizes K(v) 1.5
sembly of K(V)LQT1 and minK (IsK) proteins to protein, KChAP. J Biol Chem 273:11745–11751, channels in HL-1 cells. Br J Pharmacol 162:1832–
form cardiac I(Ks) potassium channel. Nature 1998. 1842, 2011.
384:80–83, 1996. 49. Kim E, Niethammer M, Rothschild A, et al: Clus- 68. Milstein ML, Musa H, Balbuena DP, et al:
29. Drici MD, Arrighi I, Chouabe C, et al: Involve- tering of Shaker-type K+ channels by interaction Dynamic reciprocity of sodium and potassium
ment of IsK-associated K+ channel in heart rate with a family of membrane-associated guanylate channel expression in a macromolecular complex
control of repolarization in a murine engineered kinases. Nature 378:85–88, 1995. controls cardiac excitability and arrhythmia. Proc
model of Jervell and Lange-Nielsen syndrome. 49a. Balse E, Steele D, Abriel H, et al: Dynamic of Natl Acad Sci U S A 2012.
Circ Res 83:95–102, 1998. ion channel expression at the plasma membrane 69. Kim E, Sheng M: Differential K+ channel cluster-
30. Temple J, Frias P, Rottman J, et al: Atrial fibrilla- of cardiomyocytes. Physiol Revue 92(3):1317– ing activity of PSD-95 and SAP97, two related
tion in KCNE1-null mice. Circ Res 97:62–69, 1358, 2012. membrane-associated putative guanylate kinases.
2005. 50. Nourry C, Grant SG, Borg JP: PDZ domain Neuropharmacology 35:993–1000, 1996.
31. Tinel N, Diochot S, Borsotto M, et al: KCNE2 proteins: plug and play! Sci STKE 2003:RE7, 70. Godreau D, Vranckx R, Maguy A, et al: Different
confers background current characteristics to the 2003. isoforms of synapse-associated protein, SAP97,
cardiac KCNQ1 potassium channel. EMBO J 51. Songyang Z, Fanning AS, Fu C, et al: Recognition are expressed in the heart and have distinct effects
19:6326–6330, 2000. of unique carboxyl-terminal motifs by distinct on the voltage-gated K+ channel Kv1.5. J Biol
32. Zhang M, Jiang M, Tseng GN: minK-related PDZ domains. Science 275:73–77, 1997. Chem 278:47046–47052, 2003.
peptide 1 associates with Kv4.2 and modulates its 52. Barker RJ, Price RL, Gourdie RG: Increased 71. Bassand P, Bernard A, Rafiki A, et al: Differential
gating function: potential role as beta subunit of association of ZO-1 with connexin43 during interaction of the tSXV motifs of the NR1 and
cardiac transient outward channel? Circ Res remodeling of cardiac gap junctions. Circ Res NR2A NMDA receptor subunits with PSD-95
88:1012–1019, 2001. 90:317–324, 2002. and SAP97. Eur J Neurosci 11:2031–2043, 1999.
33. Abbott GW, Sesti F, Splawski I, et al: MiRP1 53. Itoh M, Morita K, Tsukita S: Characterization of 72. Mauceri D, Gardoni F, Marcello E, et al: Dual
forms IKr potassium channels with HERG and is ZO-2 as a MAGUK family member associated role of CaMKII-dependent SAP97 phosphoryla-
associated with cardiac arrhythmia. Cell 97:175– with tight as well as adherens junctions with a tion in mediating trafficking and insertion of
187, 1999. binding affinity to occludin and alpha catenin. NMDA receptor subunit NR2A. J Neurochem
34. Cui J, Kagan A, Qin D, et al: Analysis of the cyclic J Biol Chem 274:5981–5986, 1999. 100:1032–1046, 2007.
nucleotide binding domain of the HERG potas- 54. Jesaitis LA, Goodenough DA: Molecular charac- 73. Gardoni F, Schrama LH, van Dalen JJ, et al:
sium channel and interactions with KCNE2. J terization and tissue distribution of ZO-2, a tight AlphaCaMKII binding to the C-terminal tail of
Biol Chem 276:17244–17251, 2001. junction protein homologous to ZO-1 and the NMDA receptor subunit NR2A and its modula-
35. Roepke TK, Kontogeorgis A, Ovanez C, et al: Drosophila discs-large tumor suppressor protein. tion by autophosphorylation. FEBS Lett 456:394–
Targeted deletion of kcne2 impairs ventricular J Cell Biol 124:949–961, 1994. 398, 1999.
repolarization via disruption of I(K,slow1) and 55. El-Haou S, Balse E, Neyroud N, et al: Kv4 potas- 74. Vaidyanathan R, Taffet SM, Vikstrom KL, et al:
I(to,f). FASEB J 22:3648–3660, 2008. sium channels form a tripartite complex with the Regulation of cardiac inward rectifier potassium
36. Isbrandt D, Friederich P, Solth A, et al: Identifica- anchoring protein SAP97 and CaMKII in cardiac current (I(K1)) by synapse-associated protein-97.
tion and functional characterization of a novel myocytes. Circ Res 104:758–769, 2009. J Biol Chem 285:28000–28009, 2010.
KCNE2 (MiRP1) mutation that alters HERG 56. Godreau D, Vranckx R, Maguy A, et al: Expres- 75. Lundbaek JA, Birn P, Girshman J, et al: Mem-
channel kinetics. J Mol Med (Berl) 80:524–532, sion, regulation and role of the MAGUK protein brane stiffness and channel function. Biochemis-
2002. SAP-97 in human atrial myocardium. Cardiovasc try 35:3825–3830, 1996.
37. Lu Y, Mahaut-Smith MP, Huang CL, et al: Res 56:433–442, 2002. 76. Andersen OS, Koeppe RE, 2nd: Bilayer thickness
Mutant MiRP1 subunits modulate HERG K+ 57. Leonoudakis D, Conti LR, Radeke CM, et al: A and membrane protein function: an energetic per-
channel gating: a mechanism for pro-arrhythmia multiprotein trafficking complex composed of spective. Annu Rev Biophys Biomol Struct
in long QT syndrome type 6. J Physiol 551:253– SAP97, CASK, Veli, and Mint1 is associated with 36:107–130, 2007.
262, 2003. inward rectifier Kir2 potassium channels. J Biol 77. Elinder F, Madeja M, Arhem P: Surface charges
38. Sesti F, Abbott GW, Wei J, et al: A common Chem 279:19051–19063, 2004. of K channels. Effects of strontium on five cloned
polymorphism associated with antibiotic-induced 58. Murata M, Buckett PD, Zhou J, et al: SAP97 channels expressed in Xenopus oocytes. J Gen
cardiac arrhythmia. Proc Natl Acad Sci U S A interacts with Kv1.5 in heterologous expression Physiol 108:325–332, 1996.
97:10613–10618, 2000. systems. Am J Physiol Heart Circ Physiol 78. Long SB, Tao X, Campbell EB, et al: Atomic
39. An WF, Bowlby MR, Betty M, et al: Modulation 281:H2575–H2584, 2001. structure of a voltage-dependent K+ channel in a
of A-type potassium channels by a family of 59. Peters CJ, Chow SS, Angoli D, et al: In situ lipid membrane-like environment. Nature
calcium sensors. Nature 403:553–556, 2000. co-distribution and functional interactions of 450:376–382, 2007.
40. Bahring R, Dannenberg J, Peters HC, et al: Con- SAP97 with sinoatrial isoforms of HCN channels. 79. Schmidt D, MacKinnon R: Voltage-dependent
served Kv4 N-terminal domain critical for effects J Mol Cell Cardiol 46:636–643, 2009. K+ channel gating and voltage sensor toxin sensi-
of Kv channel-interacting protein 2.2 on channel 60. Petitprez S, Zmoos AF, Ogrodnik J, et al: SAP97 tivity depend on the mechanical state of the lipid
expression and gating. J Biol Chem 276:23888– and dystrophin macromolecular complexes deter- membrane. Proc Natl Acad Sci U S A 105:19276–
23894, 2001. mine two pools of cardiac sodium channels 19281, 2008.
80. Epshtein Y, Chopra AP, Rosenhouse-Dantsker A, myocyte sarcolemma. FEBS Lett 580:6039–6046, cholesterol in MDCK cells. J Lipid Res 39:369–
et al: Identification of a C-terminus domain criti- 2006. 379, 1998.
cal for the sensitivity of Kir2.1 to cholesterol. 90. Schubert AL, Schubert W, Spray DC, et al: Con- 98. Rothberg KG, Ying YS, Kamen BA, et al:
Proc Natl Acad Sci U S A 106:8055–8060, 2009. nexin family members target to lipid raft domains Cholesterol controls the clustering of the
81. Oliver D, Lien CC, Soom M, et al: Functional and interact with caveolin-1. Biochemistry 41: glycophospholipid-anchored membrane receptor
conversion between A-type and delayed rectifier 5754–5764, 2002. for 5-methyltetrahydrofolate. J Cell Biol 111:
K+ channels by membrane lipids. Science 91. Balse E, El-Haou S, Dillanian G, et al: Choles- 2931–2938, 1990.
304:265–270, 2004. terol modulates the recruitment of Kv1.5 chan- 99. Rubinstein RA, Jr., Shelbourne KD, McCarroll
82. Palade GE: An electron microscope study of the nels from Rab11-associated recycling endosome JR, et al: The accuracy of the clinical examination
mitochondrial structure. J Histochem Cytochem in native atrial myocytes. Proc Natl Acad Sci in the setting of posterior cruciate ligament inju-
1:188–211, 1953. U S A 106:14681–14686, 2009. ries. Am J Sports Med 22:550–557, 1994.
83. Balijepalli RC, Kamp TJ: Caveolae, ion channels 92. O’Connell KM, Rolig AS, Whitesell JD, et al: 100. Schnitzer JE, Oh P, Pinney E, et al: Filipin-
and cardiac arrhythmias. Prog Biophys Mol Biol Kv2.1 potassium channels are retained within sensitive caveolae-mediated transport in endothe-
98:149–160, 2008. dynamic cell surface microdomains that are lium: reduced transcytosis, scavenger endocytosis,
84. Dart C: Lipid microdomains and the regulation defined by a perimeter fence. J Neurosci 26:9609– and capillary permeability of select macromole-
of ion channel function. J Physiol 588:3169–3178, 9618, 2006. cules. J Cell Biol 127:1217–1232, 1994.
2010. 93. Seebohm G, Strutz-Seebohm N, Birkin R, et al: 101. Rodal SK, Skretting G, Garred O, et al: Extrac-
85. Maguy A, Hebert TE, Nattel S: Involvement Regulation of endocytic recycling of KCNQ1/ tion of cholesterol with methyl-beta-cyclodextrin
of lipid rafts and caveolae in cardiac ion KCNE1 potassium channels. Circ Res 100:686– perturbs formation of clathrin-coated endocytic
channel function. Cardiovasc Res 69:798–807, 692, 2007. vesicles. Mol Biol Cell 10:961–974, 1999.
2006. 94. Hardel N, Harmel N, Zolles G, et al: Recycling 102. Subtil A, Gaidarov I, Kobylarz K, et al: Acute
86. Schwarzer S, Nobles M, Tinker A: Do caveolae endosomes supply cardiac pacemaker channels for cholesterol depletion inhibits clathrin-coated pit
have a role in the fidelity and dynamics of receptor regulated surface expression. Cardiovasc Res budding. Proc Natl Acad Sci U S A 96:6775–
activation of G-protein-gated inwardly rectifying 79:52–60, 2008. 6780, 1999.
potassium channels? J Biol Chem 285:27817– 95. Uhlig M, Passlack W, Eckel J: Functional role of 103. Guo J, Massaeli H, Xu J, et al: Extracellular K+
27826, 2010. Rab11 in GLUT4 trafficking in cardiomyocytes. concentration controls cell surface density of IKr in
87. Locke D, Liu J, Harris AL: Lipid rafts prepared Mol Cell Endocrinol 235:1–9, 2005. rabbit hearts and of the HERG channel in human
by different methods contain different connexin 96. Xia F, Leung YM, Gaisano G, et al: Targeting of cell lines. J Clin Invest 119:2745–2757, 2009.
channels, but gap junctions are not lipid rafts. voltage-gated K+ and Ca2+ channels and soluble 104. McEwen DP, Schumacher SM, Li Q, et al: Rab-
Biochemistry 44:13027–13042, 2005. N-ethylmaleimide-sensitive factor attachment GTPase-dependent endocytic recycling of Kv1.5
88. Yarbrough TL, Lu T, Lee HC, et al: Localization protein receptor proteins to cholesterol-rich lipid in atrial myocytes. J Biol Chem 282:29612–29620,
of cardiac sodium channels in caveolin-rich mem- rafts in pancreatic alpha-cells: effects on glucagon 2007.
brane domains: regulation of sodium current stimulus-secretion coupling. Endocrinology 105. Delmar M: Connexin43 regulates sodium current;
amplitude. Circ Res 90:443–449, 2002. 148:2157–2167, 2007. ankyrin-G modulates gap junctions: the interca-
89. Eldstrom J, Van Wagoner DR, Moore ED, et al: 97. Hailstones D, Sleer LS, Parton RG, et al: lated disc exchanger. Cardiovasc Res 93:220–222,
Localization of Kv1.5 channels in rat and canine Regulation of caveolin and caveolae by 2012.
Reciprocity of Cardiac Sodium and
Potassium Channels in the Control of
Excitability and Arrhythmias 21
José Jalife and Michelle Lynne Milstein
of Kir2.1 plus NaV1.5 than when Kir2.1 was overexpressed alone.

CHAPTER OUTLINE
In NRVM monolayers, co-overexpression of NaV1.5 with Kir2.1
Intermolecular Interactions Involving NaV1.5 and increased CV, abbreviated action potential duration (APD) and
Kir2.x Channels 207 increased rotor frequency beyond those produced by Kir2.1 OE
alone.5 Furthermore, recent data in the literature suggest that
Reciprocal Regulation of NaV1.5 and Kir2.1 in conditions that result in Nav1.5 protein reduction, such as that
Adult Rat Ventricular Myocytes 208 which occurs in dystrophin-deficient mdx5cv mice, are accompa-
SAP97 and Syntrophin Are Involved in NaV1.5-Kir2.1 nied by a concomitant reduction in Kir2.1 protein levels.6 Impor-
Interactions 208 tantly, the finding that coexpression of NaV1.5 can reduce
internalization of Kir2.1 is a central mechanistic observation.5
NaV1.5 -Kir2.1 Interactions Are Posttranslational The purpose of this chapter is to discuss those results in the
and Model-Independent 208 context of cardiac excitability and mechanisms of reentrant
NaV1.5-Kir2.1 Interactions Involve Membrane arrhythmias. It will be shown that sodium and potassium channel
interactions depend on more than membrane voltage alone. Alto-
Trafficking 209
gether, the evidence that will be discussed suggests that cardiac
Reciprocal NaV1.5 -Kir2.1 Interactions Control cells undergo model-independent co-regulation involving post-
Reentry Frequency 211 translational mechanisms of Kir2.1 and NaV1.5, with important
functional consequences for myocardial excitation, impulse
velocity, and arrhythmogenesis. Moreover, the evidence suggests
that similar interactions might apply to other sarcolemmal ion
Current understanding of the relationship between the sodium channels as well, which could themselves have unique effects on
current (INa) and the inward rectifier potassium current (IK1), the myocardial function.
two most important ionic currents controlling ventricular excit-
ability, derives primarily from traditional electrophysiology. It
derives also from basic and clinical studies on arrhythmogenesis Sodium Channels and Cardiac Excitation
in ion channel diseases and heart failure, which have demon-
strated that modification in the peak density of either INa or IK1 In the heart, INa is the major current that excites cardiac cells in
changes cell excitability and conduction velocity (CV). However, the atria, the ventricles, and the Purkinje fibers. Normally closed
the pathophysiologic consequences of a molecular interplay at the resting potential (approximately –85 mV in the adult
between the individual channels at the center of such diseases working myocardium), sodium channels open upon depolariza-
have not been investigated. In the heart, IK1 is the major current tion beyond threshold, allowing an influx of Na+ ions down their
responsible for the maintenance of the resting membrane poten- electrochemical gradient. This inward current causes “all-or-
tial (RMP), whereas INa provides the largest fraction of the inward none” membrane depolarization at a rapid rate (~500 V/s in the
depolarizing current that flows during an action potential.1 It is Purkinje fibers) in a process that moves the membrane potential
well known that by controlling the RMP, IK1 modifies Na+ channel to positive values. The rapid voltage-dependent activation of the
availability and therefore, cell excitability.2 In addition, IK1-INa sodium channel is immediately followed by an inactivation
interactions are important in stabilizing and controlling the fre- process that is also initiated by the initial depolarization. The
quency of the electrical rotors that are responsible for the most inactivation process causes INa to be brief and results in the ter-
dangerous cardiac arrhythmias, including ventricular tachycardia mination of the current. These properties are important for the
(VT) and ventricular fibrillation (VF).3,4 rapid (≥1 m/s) conduction of the electrical impulse in the
Recent data obtained from adult transgenic mice, single adult myocardium.
rat ventricular myocytes (ARVMs), neonatal rat ventricular From the clinical standpoint, multiple mutations in the
myocyte (NRVM) monolayers, and human embryonic kidney SCN5A gene coding for NaV1.5 have been identified in associa-
(HEK 293) cells, have demonstrated that the INa-IK1 interplay is tion with the long QT syndrome (LQTS), Brugada syndrome,
much more complex than previously thought. It comprises a idiopathic ventricular fibrillation, cardiac conduction defects, and
model independent, reciprocal modulation of expression of their dilated cardiomyopathy associated with atrial fibrillation.7 Such
respective channel proteins (Nav1.5 and Kir2.1) within a macro- mutations illustrate the pathophysiologic importance of these
molecular complex that involves the MAGUK-type protein channels.
SAP97,5 and possibly additional scaffolding proteins. In adult Homozygous knockout (KO) Scn5a-/- mouse embryos die
transgenic mice overexpressing Kir2.1 (Kir2.1 OE), peak INa during mid-gestation, most likely because of severe defects in
density is twice as large as that measured in cells from control ventricular morphogenesis,8 which provides evidence for an
hearts. In heterozygous Kir2.1 knockout mice (Kir2.1-/+), NaV1.5 essential role of Nav1.5 in cardiac development.7 However, het-
protein and INa are significantly reduced. Similarly, in single erozygous KO mice (Scn5a+/−) show normal survival but exhibit
ARVMs, IK1 increased significantly more upon adenoviral transfer slow atrial, atrioventricular (AV) and intraventricular conduction,
205
prolonged ventricular refractoriness, and enhanced ventricular or pathophysiological modulation of IK1 will have a significant
arrhythmia inducibility.7,8 Ventricular myocytes isolated from effect on excitability. The Kir channels show strong rectification
adult Scn5a+/− mice demonstrate an approximately 50% reduction between –50 and 0 mV, which means that they remain closed
in sodium conductance.8 However, an important question that during the AP plateau; they only open when the membrane
has never been addressed, is whether, as has been demonstrated potential repolarizes to levels between –30 and –80 mV, which
in other model systems, changes in Nav1.5 expression alter the occurs during the late phases of the action potential.12 Rectifica-
level of functional expression of other membrane ion channels, tion is achieved by a voltage-dependent blockade by intracellular
particularly Kir2.1.5 Based on recent results,5 it would be reason- magnesium and/or one of the polyamines (putrescine, spermine,
able to expect a significant reduction in Kir2.1 (and its respective spermidine),12 which are known to interact with at least three
ionic current, IK1) and in the RMP of adult ventricular cardio- amino acid residues located inside the pore of the channel
myocytes as a consequence of one SCN5A allele being absent. complex.13 Investigators have used several strategies to modify
Should that be the case, one would predict that IK1, the RMP and and study Kir2.1, including a knockout mouse,14 antisense oligo-
excitability would be rescued to normal levels by virally mediated nucleotide targeting,15 and DNA transfection of a dominant-
gene transfer of Nav1.5 into adult cardiomyocytes isolated from negative construct.16 These studies have helped to define the role
these mice. of IK1 in cardiac excitability.17,18 As shown by Zartizky et al.,
In a recent study using a transgenic mouse line overexpressing ventricular myocytes from Kir2.1 knockout (Kir2.1−/−) mice
the human SCN5A gene,9 levels for the cardiac sodium mRNA lack IK1 in whole-cell recordings under physiologic conditions,
transcript (Scn5a) and protein (Nav1.5) were each increased by which demonstrated that Kir2.1 is the major determinant of
approximately tenfold. However, no abnormalities were found in IK1.14 In that model, sustained outward K+ currents and Ba2+
the electrophysiology of the ventricles. The QRS duration and currents through L- and T-type channels were not significantly
the corrected QT interval (QTc) in SCN5A overexpressing mice altered by the mutation. However, the direct consequences
were indistinguishable from their nontransgenic littermates. In of Kir2.1 disruption on Nav1.5 function were never studied
addition, no ventricular arrhythmias were detected in the trans- in the homozygous Kir2.1 KO mice. Recently, the authors
genic mice.9 The sodium current densities and APDs from trans- took advantage of the availability of the heterozygous Kir2.1
genic ventricular cardiomyocytes were nearly identical to those KO (Kir2.1−/+) mouse model to study the functional consequences
of nontransgenic cells. Equally striking were the similarities of reducing Kir2.1, and therefore IK1, at both cellular and
between transgenic and normal hearts with respect to the sodium organismal levels. In particular, as discussed in detail below, the
current densities and APDs found in atrial cells. However, base- Kir2.1−/+ mouse was used to address the question of whether
line ECG recordings by telemetry revealed a shortened PR inter- reduced Kir2.1 protein is associated with reduced expression of
val and P wave duration in the transgenic mice compared to their Nav1.5.5
littermate controls, which the authors interpreted as being the Loss of function mutations in the KCNJ2 gene, which codes
result of altered AV nodal conduction.9 However, the possibility for Kir2.1, have been identified in patients affected by Andersen-
that His-Purkinje conduction was enhanced and thus caused the Tawil syndrome (ATS), also known as LQTS type 7, which is
PR shortening was not considered. Nevertheless, it was clear that characterized by prolonged repolarization.19 In addition to being
overexpression of SCN5A did not significantly increase the expressed in the heart, Kir2.1 is expressed in other organs, such
surface density of sodium channels in ventricular or atrial myo- as skeletal muscle. As a result, ATS is associated with hypokalemic
cytes.9 This finding is different from what was found in the cul- periodic paralysis and skeletal developmental abnormalities,
tured neonatal rat myocyte. In these cells, INa is inherently including clinodactyly, low-set ears, mandibular hypoplasia, short
relatively small, but viral transfer (and therefore overexpression) stature, and scoliosis.19 In the heart, reduction of IK1 leads to QT
of Nav1.5 is accompanied by a significant increase in Nav1.5 and prolongation and predisposes to arrhythmias; however, QT pro-
INa, excitability, and CV. As suggested by Abriel,7 one possible longation is less prominent in patients with ATS than in those
explanation to account for the disparate results of the two studies with other types of LQTS.20 Moreover, although ATS patients
may be that the sodium channels are binding in a macromolecular develop ventricular tachyarrhythmias, including torsades de
complex. Based on this premise, and the assumption that there pointes, sudden cardiac death is rare.20
are a limited number of docking sites or complexes to which the In patients with short QT syndrome type 3, a gain-of-function
sodium channel binds in the adult heart, then this complex (i.e., mutation (D172N) in the KCNJ2 gene was demonstrated.21 The
including its localization, stoichiometry, available components) D172N mutation causes a significant increase in the outward
could provide an upper limit as to how many functional sodium component of the I-V relation of IK1, and it is associated with
channels can be brought to the membrane in the adult mouse an accelerated repolarization, which can be arrhythmogenic.
myocyte.7 However, such an idea might not be tenable in light However, the direct involvement of IK1 in arrhythmia mecha-
of recent experiments demonstrating that transgenic overexpres- nisms was not demonstrated in the affected patients; therefore,
sion of Kir2.1 increases INa density in the adult mouse heart the mouse model of IK1 overexpression22 was used to study the
(discussed later). effect of IK1 increase on VF at the molecular level.23 An increase
of IK1 was shown to shorten repolarization and the QT interval
and to exert a proarrhythmic effect in both the atria and the
The Inward Rectifying Potassium Current (IK1) ventricles of this transgenic mouse model.22,23 Optical mapping
and numerical studies in these mice demonstrated that, by
Among the three strong inward-rectifying potassium channels increasing the RMP, IK1 overexpression enhances the availability
(Kir2.1, 2.2 and 2.3) that are expressed in the heart, Kir2.1 is the of sodium channels during sustained reentry, which contributed
most abundant in the ventricles. Kir channels are responsible to the observed increase in the frequency and stability of rotors
for IK1, and they are involved in the depolarization, repolariza- and ventricular fibrillation.23 Recent experiments with the same
tion, and the resting phases of the cardiac action potential.10,11 Kir2.1 overexpressing mouse model have shown that Kir2.1, and
It is usually accepted that, near the resting potential, the therefore IK1, upregulation leads to a significant increase in the
ventricular IK1 conductance is much larger than that of any density of INa. These results strongly support the hypothesis that
other current, with the exception of the adenosine triphosphate a change in the functional expression of Nav1.5 could be the
(ATP)-sensitive potassium current (IKATP), which, however, is result of protein-protein–type interactions with Kir2.1. Such
generally not present since the KATP channels are not open interactions can be mediated through common partners in a
under normal conditions. It is thus likely that physiological and/ macromolecular protein complex.5
Reciprocity of Cardiac Sodium and Potassium Channels in the Control of Excitability and Arrhythmias 207
a defect in ankyrin-mediated Nav1.5 trafficking and localization

Intermolecular Interactions Involving Nav1.5
21
on the membrane could alter Kir2.1 functional expression. This
and Kir2.x Channels idea has not yet been explored in any detail.
PDZ domain-mediated interactions are among the most fre-

quently encountered protein-protein interactions in cell biology.24 Nav1.5 Is Regulated by SAP97
PDZ stands for postsynaptic density protein (PSD-95), discs-
large (the drosophila septate junction protein), and zona The C-terminus of Nav1.5 contains 244 amino acid residues that
occludens-1 (the epithelial tight junction protein).25 The primary have been shown to interact with a number of proteins, including
function of PDZ domains is to mediate protein-protein interac- fibroblast growth factor-homologous factor 1B, calmodulin,
tions by recognizing a consensus sequence (Thr/Ser-X-Val/Leu) Nedd4-like ubiquitin-protein ligase, dystrophin and syntrophin,
usually located at the C-terminus of their target proteins.26-28 Fyn, and PTPH1 (protein-tyrosine phosphatase).49-51 Of impor-
PDZs can combine with other interaction modules (such as WW, tance, the three last three amino-acids (Ser-Ilc-Val, SIV) of the
SH3, and PTB domains) and help to direct the specificity of Nav1.5 C-terminus constitute a PDZ-domain binding motif that
receptor tyrosine kinases, establish and maintain cell polarity, is known to interact with PDZ domains found in proteins of
direct protein trafficking, and coordinate synaptic signaling.29-32 the membrane-associated guanylate kinase (MAGUK) family.52
Their pathophysiologic importance is highlighted by significant MAGUK proteins act as scaffolding proteins involved in inter-
neuronal and developmental defects observed in PDZ knockout molecular interactions and protein trafficking to the cell mem-
mice33-37 and by their implication in human inherited diseases.38-40 brane. Petitprez et al.53 reported their study on the interaction
More than 70 PDZ domain–containing proteins (hereafter between SAP97, one of the cardiac MAGUK proteins, and
referred to as PDZ domain proteins) have been identified that Nav1.5, with the idea of demonstrating that the Nav1.5-SAP97
interact with different ion channels, receptors, and signaling mol- interaction is involved in the turnover and regulation of the
ecules.29 PDZ domain proteins are multidomain proteins that biophysical properties of Nav1.5. Using Nav1.5 C-terminal fusion
serve to link different proteins to form macromolecular com- proteins in pull-down experiments with human and mouse heart
plexes via interactions with their various domains. For example, protein extracts, these investigators demonstrated that the asso-
the protein structure of synapse-associated protein 97 (SAP97) ciation between SAP97 and Nav1.5 depended on the PDZ-
contains three PDZ domains and an SH3 domain, HOOK domain binding motif of Nav1.5. The interaction appeared to be
domain, I3 domain, and an inactive guanylate kinase (GK) specific for SAP97 and Nav1.5, because the Nav1.5 constructs did
domain. Interactions via its HOOK domain enable SAP97 to not pull down either PSD95 or ZO-1, two other MAGUK pro-
bind protein 4.1; and in so doing, SAP97 is able to link proteins teins that are expressed in the human heart.54 In patch-clamp
bound to its PDZ domains to the actin–spectrin membrane cyto- experiments, Petitprez et al.53 further demonstrated that silenc-
skeleton or to protein components of the actin–spectrin mem- ing SAP97 expression reduced the whole-cell sodium current
brane cytoskeleton, such as protein 4.1.41 The complement of without decreasing the total protein content in HEK293 cells
interacting proteins varies among the different PDZ domain pro- stably expressing Nav1.5 channels. Taken together, with the dem-
teins, and this provides a mechanism to recruit ion channel onstrated colocalization of NaV1.5 and SAP97 at both interca-
proteins into distinct macromolecular complexes depending on lated discs53 and T-tubules,5 these findings support the existence
which scaffolding proteins they bind (e.g., see Tiffany et al.42). of an interaction between Nav1.5 and SAP97 in cardiac tissue.
The cellular localization of PDZ domain proteins can also vary, This interaction could have a role in determining the channel
and it has been suggested that ion channel–PDZ domain protein density at the plasma membrane. Therefore, Petitprez et al.53
interactions might be an important mechanism for plasma mem- have suggested that there are at least two pools of Nav1.5 chan-
brane expression and distribution of ion channels.42-44 For nels in cardiomyocytes: one targeted to lateral membranes by the
example, ZO-1 and SAP97 are found in overlapping but distinct syntrophin-dystrophin complex, and another targeted to the
subcellular locations in the heart. ZO-1 is located exclusively in intercalated discs by SAP97.
the intercalated disc, whereas SAP97 is found in the intercalated
disc, lateral membranes,44-46 and at the T-tubules.5
Nav1.5 Interacts with Syntrophins
Ankyrin-G and Nav1.5 Trafficking Like other ion channels, Nav1.5 has been reported to be part of
the dystrophin multiprotein complex.55 Gavillet et al.6 demon-
A number of accessory proteins have been shown to interact and strated that Nav1.5 interacts with dystrophin via syntrophin
form a multiprotein complex with Nav1.5.7,47 Ankyrin-G and adaptor proteins through the PDZ domain-binding motif at the
Nav1.5 are both localized at the intercalated disc and at T-tubule Nav1.5 C-terminus.50 Dystrophin-deficient mice (mdx5cv) have
membranes in cardiomyocytes, and Nav1.5 coimmunoprecipitates reduced protein levels of Nav1.5 in ventricular lysates, and this is
with the 190-kDa ankyrin-G in detergent-soluble lysates from rat directly associated with reduced INa in isolated cardiomyocytes
heart.47 The two proteins interact through a 9-amino acid motif and with conduction defects documented on an electrocardio-
in the Nav1.5 cytoplasmic loop II between DII and DIII, which gram. Immunostaining of frozen mouse heart slices demonstrated
helps to promote the localization of sodium channels to the cell colocalization of Nav1.5 and dystrophin specifically at lateral
membrane in cardiomyocytes.47 In 2004, Mohler et al.47 identified membranes, but not at the intercalated discs.6 However, mecha-
a point mutation (E1053K) in the ankyrin-binding motif of Nav1.5 nisms by which a lack of dystrophin reduces Nav1.5 protein
that is associated with Brugada syndrome. The E1053K mutation expression without altering the mRNA level6 have not been elu-
abolishes binding of Nav1.5 to ankyrin-G and prevents the accu- cidated. A study by Ueda et al.56 demonstrated that α1-syntrophin
mulation of Nav1.5 at cell surface sites in ventricular cardiomyo- (encoded by the SNTA1 gene), connects Nav1.5 to the neuronal
cytes.47 Those data suggested that association of Nav1.5 with nitric oxide synthase (nNOS) – plasma membrane Ca-ATPase
ankyrin-G is required for Nav1.5 localization, and therefore func- (PMCA) complex in the heart. These results implicated SNTA1
tion, at excitable membranes in cardiomyocytes. Concerted as a susceptibility gene for inherited LQTS. The cardiac isoform
ankyrin-G interaction with potassium (Kv7) and Nav channels has of PMCA, PMCA4b, participates in the nNOS complex to
also been demonstrated in neurons,48 but Kir2.1 has not been inhibit nitric oxide synthesis.57 Furthermore, the study by Ueda
shown to interact with ankyrin-G. There is still the possibility that et al.56 also identified a rare missense mutation (A390V-SNTA1)
in α1-syntrophin that was found in an LQTS patient. The A390V adenoviral (Ad) construct (Ad-Nav1.5)5; a control group of myo-
mutation selectively disrupted association of α1-syntrophin with cytes was infected with an adenoviral construct encoding green
PMCA4b, increased Nav1.5 nitrosylation, and increased late fluorescence protein (Ad-GFP). Average INa-density–voltage
sodium current, which all suggested that the SNTA1 mutation (I-V) plots are shown in Figure 21-1, A. There was a significant
was pathogenic. increase (p < 0.005) in INa density compared with control cells.
There was no significant difference for V1/2 values between
control and Ad-Nav1.5–treated cells.5 Overexpression of Nav1.5
Kir2.1 Interacts with Both SAP97 and α-Syntrophin also resulted in a significant increase in IK1 density in both the
inward (p < 0.01) and outward (p < 0.05) directions (see Figure
Members of the Kir2.x subfamily (Kir2.1, Kir2.2 and Kir2.3) each 21-1, B). Next, the possibility was considered that Kir2.1 over-
have a C-terminal motif that enables interaction with PDZ expression would increase Nav1.5 functional expression.5 Infec-
domain proteins. However, the C-terminal motifs are not identi- tion with Ad-Kir2.1 resulted in an increase of IK1 in ARVMs
cal between the isoforms. RRESEI is found in Kir2.1 and Kir2.2, (Figure 21-1, C). Peak inward (–100 mV) and outward (–60 mV)
whereas RRESAI is found in Kir2.3.58 It is noteworthy that the currents were significantly larger (p < 0.01) than in control (Ad-
PDZ binding motif sequence on Kir2 overlaps a consensus GFP) cells. As hypothesized, peak INa density was also increased
sequence (RRxS) for protein kinase A–dependent phosphoryla- (p < 0.005; Figure 21-1, D) with no changes in voltage depen-
tion. Consistent with this finding, Kir2.3 binding to PDZ domain dence of activation or inactivation.5 These results suggest that
proteins is modulated in vitro by phosphorylation of Kir2.3 common molecular mechanisms are involved in the regulation of
protein.59 Kir2.1 and Nav1.5 functional expression.5
A number of PDZ domain-containing proteins have already
been shown to bind Kir2 channels. In two recent studies,60, 61
affinity purification was used to isolate PDZ domain proteins
from cardiac and brain tissue extracts. Identification of these SAP97 and Syntrophin Are Involved
isolated proteins demonstrated that cardiac SAP97, CASK, Veli3, in NaV1.5-Kir2.1 Interactions
and Mint1 are found in a complex with Kir2.2 channels. For the
most part, the consequences (e.g., channel expression and bio- SAP97 is able to link proteins bound to its PDZ domains, and
physics of the interactions between Kir2.x channels with the PDZ because both Nav1.5 and Kir2.1 have PDZ binding domains at
domain proteins) are unknown. An association of PDZ domain their respective C-terminus, the potential role of SAP97 in the
proteins with many other ion channels or receptors has been reciprocal INa-IK1 interactions, discussed in the previous section,
shown to affect various aspects of protein trafficking in the secre- was examined. As described elsewhere in detail,5 adenoviral trans-
tory and endocytic pathways. In general, PDZ domain–mediated fer was used to knock down SAP97 expression (Ad-shSAP97) in
protein-protein interactions can alter the rate of channel/receptor ARVMs, and then properties of IK1 and INa were studied under
trafficking to the plasma membrane or alter the rate at which a these conditions. As illustrated in Figure 21-2, an approximate
given cell surface channel and/or receptor is endocytosed.62-64 It 56% reduction in the relative levels of SAP97 (day 3 postinfec-
has also been suggested that binding to PDZ domain proteins is tion) (see Figure 21-2, B) was accompanied by an approximately
important for anchoring the Kir2 channels at the plasma mem- 50% reduction in IK1 density compared with control cells (see
brane46; however, until recently, this hypothesis had not been Figure 21-2, C ). Similarly, silencing SAP97 expression also
tested directly, with the exception of the interaction of PSD-95 reduced whole-cell INa density at several tested voltages (see
(a neuronal scaffolding protein) and Kir2.1 (discussed later).65 Figure 21-2, D). Peak inward INa density was reduced by approxi-
In 2004, Leonoudakis et al.61 used a proteomics approach to mately 60% compared with the control cells (p < 0.005).
identify proteins associated with Kir2 channels via the channel’s An investigation of whether Kir2.1 and Nav1.5 also associate
C-terminal PDZ binding motif. Based on immunoaffinity puri- with syntrophin is demonstrated in Figure 21-2, E, F. Syntrophin
fication and affinity chromatography from skeletal and cardiac was detected in ventricular membrane fractions following immu-
muscle and brain, they demonstrated that, in addition to noprecipitation with antibodies raised to Kir2.1 (see Figure 21-2,
MAGUK-type proteins, Kir2 channels interact with various E) or to Nav1.5 (see Figure 21-2, F).
components of the dystrophin-associated protein complex,
including α1-, β1-, and β2-syntrophin, dystrophin, and dystro-
brevin. Their affinity pull-down experiments revealed that Kir2.1,
Kir2.2, Kir2.3, and Kir4.1 all bind to scaffolding proteins but NaV1.5-Kir2.1 Interactions Are
with different affinities for the dystrophin-associated protein Posttranslational and Model Independent
complex, including dystrophin and dystrobrevin,60 as well as the
MAGUK proteins such as SAP97, CASK, and Veli. Moreover, The availability of a transgenic mouse model overexpressing
their immunofluorescent localization studies demonstrated that Kir2.123 gave us the opportunity to ensure that the interactions
Kir2.2 colocalizes with syntrophin, dystrophin, and dystrobrevin were not an artifact of adenoviral infection and to discard any
at skeletal muscle neuromuscular junctions. Overall, the results species-related differences in the interactions. Therefore, addi-
of Leonoudakis et al.60,61 suggested that Kir2 channels associate tional patch-clamp experiments were conducted in ventricular
with protein complexes that might be important to target and myocytes isolated from these adult transgenic mice, which are
traffic channels to specific subcellular locales, as well as to anchor known to have approximately twelvefold greater IK1 density than
and stabilize channels in the plasma membrane. littermate wild type (WT) mice.66 A comparison of INa density in
WT and Kir2.1 overexpressing (OE) mice is shown in Figure
21-3. Representative INa tracings from each genotype are pre-
sented in the left inset (see Figure 21-3, A). Compared with WT,
Reciprocal Regulation of NaV1.5 and Kir2.1 the I-V relationship for INa in myocytes from Kir2.1 OE mice
in Adult Rat Ventricular Myocytes (ARVMs) shows significantly larger INa density at several tested voltages
(see Figure 21-3, A, right inset; p < 0.005). Consistent with the
Voltage-clamp experiments presented in Figure 21-1 show functional consequences observed with Kir2.1 overexpression
a reciprocal regulation of Nav1.5 and Kir2.1 in adult rat ventricu- (see Figure 21-3, A), increased Kir2.1 expression in the mouse
lar myocytes (ARVMs). Nav1.5 was overexpressed using the heart leads to altered membrane protein expression of SAP97,
INa pA/pF Vm (mV)

#
–100 –80 –60 –40 –20 20 40
–120 –80 –40 2 40 80
21
Vm (mV)
–40
IK1 (pA/pF)
–8 # 1.5
–80
* ***
Ad-Nav1.5 –120 –90 –45
Ad-GFP –1.5
A B ** –16
Vm (mV) δ INa pA/pF

20
20
–120 –80 –40 40 80 –120 –100 –80 –60 –40 –20 20 40
Vm (mV) –20
IK1 (pA/pF)
–40
–40 –60
δ Ad-Kir2.1 –80
*
***
Ad-GFP
–100
C D –120
–80
Figure 21-1. Reciprocal regulation of NaV1.5 and Kir2.1 in adult rat ventricular myocytes (ARVMs). A and B, NaV1.5 overexpression increases both INa and IK1 densities.
A, Superimposed INa density–voltage relationships (5 mmol/L [Na+]o) for Ad-GFP (black; N = 1, n = 7) and Ad-NaV1.5 (red; N = 2, n = 6) infected cells. B, Superimposed IK1
density–voltage relationship for Ad-GFP (black; N = 3, n = 8) and Ad-NaV1.5 (red; N = 3, n = 8) infected cells. The inset shows the magnification of the outward component
of the IK1 I-V relationship. C and D, Kir2.1 overexpression increases both IK1 and INa densities. C, Superimposed IK1 density–voltage relationships for Ad-GFP (black; N = 3,
n = 8) and Ad-Kir2.1 (red; N = 2, n = 5). D, Superimposed INa density–voltage relationships (20 mmol/L [Na+]0) for Ad-GFP (black; N = 2, n = 6) and Ad-Kir2.1 (red; N = 4, n =
7). #p < 0.05; δp < 0.01; *p < 0.005 unpaired t test with Welch’s correction.
(Modified from Milstein ML, Musa H, Balbuena DP, et al. Dynamic reciprocity of sodium and potassium channel expression in a macromolecular complex controls cardiac excit-
ability and arrhythmia. Proc Natl Acad Sci U S A 109:E2134–E2143, 2012.)
syntrophin, and Nav1.5 in the transgenic Kir2.1 OE mouse heart results suggest that regardless of whether these reciprocal
(see Figure 21-3, B, C). Relative levels of SAP97, syntrophin, and changes lead to downregulation or upregulation of channel pro-
Nav1.5 were significantly increased by factors of 1.6, 1.9, and 2.2 teins, they appear to involve posttranscriptional or posttransla-
respectively, compared with levels in WT littermates (p < 0.01; tional mechanisms.
see Figure 21-3, B, C). These data demonstrate that reciprocal
regulation between Kir2.1 and Nav1.5 (and their respective cur-
rents) also occurs in the mouse heart, and is thus not model
specific or unique to ARVMs. It also provides assurance that NaV1.5-Kir2.1 Interactions Involve
the observed reciprocal regulation in rat myocytes was not Membrane Trafficking
a virally mediated artifact of overexpression. However, real-time
polymerase chain reaction (RT-PCR) analyses showed that The results discussed in the previous section demonstrate that
although the gene encoding Kir2.1 (KCNJ2) was significantly the synergistic and reciprocal effects of Kir2.1 and NaV1.5 expres-
elevated in the transgenic, Kir2.1 overexpressing mice, mRNA sion on channel current density occur at posttranslational levels.
levels for the genes coding for SAP97 (DLG1), NaV1.5 (SCN5A), Recently, it was suggested that these channel proteins share a
and syntrophin (SNTA1) were unchanged in the hearts of trans- common trafficking pathway where the synergistic effects act to
genic Kir2.1 overexpressing mice, compared with WT litter- modulate the surface levels of Kir2.1 and NaV1.5 channels.
mates (Figure 21-4, A). However, the mechanisms controlling Kir2.1 or NaV1.5 plasma
Additional RT-PCR experiments (see Figure 21-4, B) demon- membrane targeting or localization remain poorly explored. The
strated that the expression levels of the genes encoding for balance between anterograde and retrograde trafficking pathways
NaV1.5, SAP97, and syntrophin mRNA were unchanged in ven- determines steady-state cell surface expression of channel pro-
tricles of heterozygous KCNJ2 KO (Kir2.1–/+) mice14 compared teins. Disease-associated mutations in both Kir2.1 and NaV1.5
with WT littermates. However, as shown in Figure 21-5, 50% have been shown to affect anterograde trafficking by inhibiting
allelic reduction of KCNJ2 gene expression by homologous steps early in the secretory pathway to cause intracellular reten-
recombination resulted in a significant decrease in relative mem- tion. However, it is unknown whether retention of Kir2.1 results
brane protein levels of NaV1.5, SAP97, and syntrophin (p < 0.05). in retention of NaV1.5, or vice versa. In contrast, once at the
These data demonstrate that reciprocal regulation between plasma membrane, endocytosis is the initial step in retrograde
NaV1.5 and Kir2.1 also occurs in the mouse heart and that it was movement, after which internalized proteins can follow multiple
observed in two independent mouse models. Furthermore, the routes to different intracellular fates.67 One well-recognized
1.5
Ad-515 (control virus) Ad-shSAP97 (silencing virus)
Relative levels of SAP97

(normalized to GAPDH)
148
SAP97 98 1.0
SAP97ψ δ
0.5
GAPDH 37
0.0
A B Ad-515 Ad-shSAP97
Ad-515 (control) Ad-515 (control)

pA/pF
Ad-shSAP97 Ad-shSAP97
# 2 pA/pF
Vm (mV) –100 –80 –60 –40 –20 20
–120 –100 –80 –60 –40 –20 20 40

Vm (mV)
–2 –10
+30 mV –20
+20 mV –4
–50 mV –30
–6
–100 mV –100 mV
–120 mV
–40
–8
δ
C –10 D
t
t
Ab
Ab
pu
pu
Ab
Ab
in
in
e
.5
un
un
e
e
.1
v1
an
an
m
m
r2
a
-im
-im
br
br
Ki
N
em
ti-
em
ti-
on
on
An
An
M
M
N
N
IB: syntrophin IB: syntrophin
IB: syntrophinψ IB: syntrophinψ
E F
Figure 21-2. SAP97 and syntrophin are involved in NaV1.5-Kir2.1 interactions. A–D, Role of SAP97 in the reciprocal modulation of IK1 and INa. A–B, Ad-shSAP97 infection
silences SAP97 expression in adult rat myocytes by day 3 in culture compared with control infected cells (Ad-515). A, Immunoblot showing immunodetection of SAP97
expression (top and middle) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (loading control; bottom) from cells infected with either a control (Ad-515) or shSAP97
(Ad-shSAP97) adenovirus. Middle (SAP97) shows a longer exposure of the same immunoblot shown in the top (denoted by ψ). The overexposure confirms expression of
SAP97 in the Ad-SAP97–silenced cells and shows the contrasting intensities and apparent protein levels between control and SAP97-silenced ARVMs. B, Densitometric
analysis of SAP97 expression normalized to GAPDH levels in control (black) and shSAP97-silenced (red) cells. Values represent data (mean ± SEM) from two different prepara-
tions harvested 3 days after infection. SAP97 expression was effectively knocked down by approximately 56% on day 3. C, IK1 density is reduced following SAP97 silencing
in adult rat ventricular myocytes. Peak inward current density at –100 mV (control, –7.67 ± 1.52 pA/pF; Ad-shSAP97, –4.55 ± 0.47 pA/pF) and peak outward current density
at –60 mV (control, 1.03 ± 0.46 pA/pF; ad-shSAP97, 0.34 ± 0.03 pA/pF) were significantly different (p = 0.02 and p = 0.04, respectively) between myocytes infected with
shSAP97 (N = 4, n = 11) and those infected with Ad-515 (control; N = 2, n = 6). The inset shows the protocol used to measure the current. D, Effects of SAP97 knockdown
on sodium channel current density in adult rat ventricular myocytes. Superimposed INa density–voltage relationships in control (Ad-515) and SAP97-silenced (Ad-shSAP97)
cells. A significant reduction in peak INa was observed for SAP97-silenced cells at several tested voltages. For both control and silenced conditions, N = 2 and n = 11. The
inset shows the voltage clamp step protocol. #p < 0.05; δp < 0.01; *p < 0.005 unpaired t test with Welch’s correction. E and F, Kir2.1 and NaV1.5 each associate with syntrophin
in rat heart ventricle. Syntrophin is detected following immunoprecipitation with specific antibodies raised to Kir2.1 (E) or NaV1.5 (F). Coimmunoprecipitation reactions
used membrane-enriched preparations generated from the ventricles of rat heart. (ψ images were digitally enhanced for clarity). IB, Antibody used for immunoblotting; IP,
antibody used for immunoprecipitation.
(Modified from Milstein ML, Musa H, Balbuena DP, et al. Dynamic reciprocity of sodium and potassium channel expression in a macromolecular complex controls cardiac excit-
ability and arrhythmia. Proc Natl Acad Sci U S A 109:E2134–E2143, 2012.)
outcome is the targeting of internalized proteins to lysosomes myocytes, which contributes to the control of Kir2.1 steady-state
followed by degradation. Alternatively, trafficking through recy- surface density.5 In addition, it was demonstrated that NaV1.5
cling endosomes allows proteins to return to the plasma mem- promotes cell surface expression of Kir2.1, at least in part, by
brane and thereby protects them from degradation.68 Data reducing its internalization.5 However, it is unclear how NaV1.5
presented recently in collaboration with Dr. Jeffrey Martens or macromolecular complex formation with SAP97 affect this
shows that Kir2.1 undergoes constitutive internalization in HL-1 process. Elucidation of the mechanisms controlling the surface
INa pA/pF
Normalized change in mRNA expression

**
21
20 100
WT –120 –100 –80 –60 –40 –20 20 40
WT
80 Kir2.1 overexpressor
Vm (mV)
Kir2.1 OE –20
60
5
–40
20 pA/pF 4
Kir2.1 OE
3
WT * –60
A
5 ms *** 2
1
0
KCNJ2 DIg1 SCN5a SNTA1
WT Kir2.1 OE WT Kir2.1 OE A
IB:
Nav1.5
Normalized change in mRNA expression

WT
1.0 Kir2.1–/+
SAP97
Syntrophin
Protein stain Immunoblot
0.5
2.5 Control (WT)

**
Relative protein levels (au)
δ
Kir2.1 OE
2 δ
δ
1.5 0.0
KCNJ2 SCN5a DIg1 SNTA1
1 B
Figure 21-4. SAP97, syntrophin, and NaV1.5 mRNA levels are unchanged in the
0.5 hearts of transgenic mouse models of Kir2.1 overexpression and underexpression.
Total RNA was extracted from ventricular tissue that was harvested from Kir2.1-
0 overexpressing (OE) mice, heterozygous Kir2.1 knockout (Kir2.1–/+), or the respective
SAP97 Syntrophin Nav1.5
littermate wild type mice. The mRNA levels for the genes encoding for Kir2.1
B (KCNJ2), SAP97 (Dlg1), and NaV1.5 (SCN5A) were determined by real-time poly-
merase chain reaction for all genotypes. Glyceraldehyde-3-phosphate dehydroge-
Figure 21-3. Reciprocal regulation of INa and NaV1.5 in Kir2.1 overexpressing mice.
nase was used as an endogenous control for data normalization in each sample.
A, Superimposed INa density–voltage relationships in wild type (WT; black) and
Relative quantification of mRNA expression was calculated using the 2−ΔΔCt method.
Kir2.1-overexpressing (OE, red) mice. The left inset shows representative examples
A, The bar graph compares messenger RNA (mRNA) expression for all genes
of INa traces in each group. The dotted line denotes 0 pA. (WT N = 4, n = 10; Kir2.1
between control and transgenic Kir2.1-OE ventricles. B, NaV1.5, SAP97, syntrophin
OE N = 2, n = 6. *p < 0.005, unpaired t test with Welch’s correction. B, Relative levels
mRNA levels are unchanged in the hearts of heterozygous Kir2.1 knockout (Kir2.1−/+)
of syntrophin, SAP97, and NaV1.5 are significantly increased in hearts of transgenic
mice. N = 4 for all genotypes. **p < 0.005.
mice overexpressing Kir2.1. Crude membrane vesicles were prepared from the
ventricles of control (WT) and Kir2.1-OE mice. Samples (16 µg/lane) were analyzed (Modified from Milstein ML, Musa H, Balbuena DP, et al. Dynamic reciprocity of
with sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and
sodium and potassium channel expression in a macromolecular complex controls
immunoblotted using specific antibodies for syntrophin, SAP97, or NaV1.5, as indi-
cardiac excitability and arrhythmia. Proc Natl Acad Sci U S A 109:E2134–E2143, 2012.)
cated. B, Representative immunoblots following detection of protein immunore-
activity with horseradish peroxidase–conjugated secondary antibodies and
chemiluminescence (right). The corresponding Amido black nitrocellulose (protein
stain) is shown on the left to demonstrate analysis of equal total protein. Protein stability and frequency of reentry and VT/VF.3,23 In fact, IK1
concentrations were also verified by Lowry assay. (Bottom) Densitometric analysis upregulation is a substrate for very fast electrical rotors in the
of data shown in B (top) comparing relative protein levels between WT and structurally normal ventricles.23 IK1 upregulation hyperpolarizes
Kir2.1-OE mice. Results are expressed as mean signal intensity and represent data the resting membrane potential and thus increases sodium
from three animals for each genotype (N = 3 per genotype; δp < 0.01, mean ± SEM). channel availability during reentry in a voltage-dependent
(Modified from Milstein ML, Musa H, Balbuena DP, et al. Dynamic reciprocity of manner. Recently, Milstein et al.5 demonstrated in neonatal rat
sodium and potassium channel expression in a macromolecular complex controls ventricular myocyte (NRVM) monolayers that voltage depen-
cardiac excitability and arrhythmia. Proc Natl Acad Sci U S A 109:E2134–E2143, 2012.) dency is not the only mechanism by which NaV1.5-Kir2.1 inter-
actions control reentry dynamics and lead to faster rotors. Figure
21-6 shows reproduced phase maps and rotation frequency plots
expression of Kir2.1 and NaV1.5 will contribute significantly to obtained from optical mapping experiments that investigated the
the currently limited understanding of protein transport in the electrophysiological consequences of the NaV1.5-Kir2.1 molecu-
heart and to channelopathies involving altered trafficking. lar interplay when one or both protein channels are overex-
pressed.5 A single stationary rotor maintained the electrical
activity in each monolayer. It was previously shown in single
NRVMs that Nav1.5 overexpression alone prolonged the APD,
Reciprocal NaV1.5 -Kir2.1 Interactions whereas Kir2.1 overexpression had the opposite effect.5 In Figure
Control Reentry Frequency 21-6, A, APD prolongation induced by NaV1.5 overexpression
in the monolayer was manifest as a lengthening of the
The electrophysiological interplay between INa and IK1 is essential wavelength and slowing of the reentry frequency. However,
in the control of cardiac excitability and in determining the the shortened APD of Kir2.1 overexpressing monolayers
GFP Nav1.5 Kir2.1 Kir2.1 + Nav1.5

WT Kir2.1–/+ π
250
Nav1.5
148
SAP97 98 –π
A
60 δ
Syntrophin
δ
50
Rotation frequency (Hz)

Immunoblot
40
δ
WT Kir2.1–/+ 30
δ
20
10
0
GFP Nav1.5 Kir2.1 Kir2.1 + Nav1.5
B
Figure 21-6. Molecular NaV1.5–Kir2.1 interactions modulate reentry frequency in
neonatal rat ventricular myocyte (NRVM) monolayers. A, Phase maps for single
rotations obtained from representative optical mapping movies of monolayers
A Protein stain infected with Ad-GFP, Ad-NaV1.5, Ad-Kir2.1, or Ad-Kir2.1 plus Ad-NaV1.5. The color
bar indicates the phase in the excitation–recovery cycle. B, Reentry frequencies in
monolayers infected with Ad-GFP (black; n = 11), Ad-NaV1.5 (blue; n = 13), Ad-Kir2.1
Control (WT) (yellow; n = 11), or Ad-Kir2.1 plus Ad-NaV1.5 (red; n = 13). δp < 0.01, analysis of
1.5 variance.
Kir2.1–/+
Relative protein levels (a.u.)
(Modified from Milstein ML, Musa H, Balbuena DP, et al. Dynamic reciprocity of
# sodium and potassium channel expression in a macromolecular complex controls
#
1.0 cardiac excitability and arrhythmia. Proc Natl Acad Sci U S A 109:E2134–E2143, 2012.)
#
Nav1.5 Kir2.1
0.5
0.0
Syntrophin Nav1.5 SAP97 PDZ PDZ
B Syn Syn SH3
Figure 21-5. Transgenic reduction of Kir2.1 gene expression leads to a significant

PDZ PDZ PDZ GK
decrease in relative protein levels of NaV1.5 and SAP97, and syntrophin. Crude SAP97
membrane vesicles were prepared from the ventricles of control (wild type [WT]) H1 H1
and Kir2.1−/+ mice. See Figure 21-3 for methods describing SDS-PAGE and H2 H2
immunoblotting. A, Representative immunoblots following detection of protein
immunoreactivity with horseradish peroxidase–conjugated secondary antibodies
and chemiluminescence (top). The corresponding Amido black nitrocellulose Figure 21-7. Schematic diagram of possible interactions between Kir2.1, NaV1.5,
(protein stain) is shown on the bottom to demonstrate analysis of equal total syntrophin, and SAP97 as well as their possible connection with the dystrophin-
protein. Protein concentrations were also verified by Lowry assay. B, Densitometric associated protein complex. The subcellular localization and channel activity of
analysis comparing relative protein levels between WT and Kir2.1−/+ mice. Results both NaV1.5 and Kir2.1 are regulated by protein–protein interactions via their
are expressed as mean signal intensity (N = 7 per genotype. #p < 0.05, mean ± SEM. respective C-terminal PDZ-binding motifs with such PDZ-domain–containing pro-
teins as SAP97 and syntrophin. As shown here, the C-terminus of individual NaV1.5
(Modified from Milstein ML, Musa H, Balbuena DP, et al. Dynamic reciprocity of and Kir2.1 channels may each bind to the same SAP97 or syntrophin molecule, but
sodium and potassium channel expression in a macromolecular complex controls at different PDZ domains. It is possible that these interactions occur as part of a
cardiac excitability and arrhythmia. Proc Natl Acad Sci U S A 109:E2134–E2143, 2012.) macromolecular complex (a “channelosome”) and result in changes in the expres-
sion of NaV1.5 or Kir2.1 and thereby influence their function in the cell membrane.
GK, Guanylate kinase-like domain of SAP97; SH3, src kinase homology domain of
SAP97.
significantly decreased the wavelength and increased the rotation

frequency (see Figure 21-6, B). Most remarkable was the fact that cardiac electrical function, NaV1.5 and Kir2.1, are part of a
combined overexpression of NaV1.5+Kir2.1 hyperpolarized the common macromolecular complex that involves at least two dis-
RMP, resulting in a shortened APD and a faster conduction tinct scaffolding proteins, SAP97 and syntrophin (Figure 21-7).
velocity.5 Consequently, the frequency of reentry was even higher Most likely, their interactions provide a means for their reciprocal
than that produced by Ad-Kir2.1 infection alone. These results regulation, with vital functional consequences for myocardial
suggest that the reciprocal intermolecular interplay of Kir2.1 and excitation, conduction velocity, and arrhythmogenesis.5 It is
NaV1.5 could have profound consequences on the frequency and tempting to speculate that the complex formed by these two
stability of reentry in the heart. major ion channels, along with their protein partners, represents
In summary, the results discussed in this chapter provided the a molecular sensing system in the cell that might enable it to
first evidence that two major ion channel proteins that control monitor, and adjust accordingly, expression levels of proteins
involved in generating and maintaining the cardiac impulse as Kir2.1-NaV1.5 interactions, arrhythmogenesis and heart failure
21
well as adjusting to pathophysiological conditions. Most exciting, is further underscored by the demonstration that INa density
the demonstrated intermolecular interaction between these two decreases in chronic heart failure.71 The regulation of expression
essential channels controlling cardiac excitability opens a new and mechanisms of interactions between Kir2.1 and NaV1.5 dis-
pathway in the study of the molecular mechanisms underlying cussed here could contribute to the development of more effica-
sudden cardiac death in highly prevalent heart diseases, including cious treatments for arrhythmias in heart disease, particularly in
heart failure, and with inherited cardiac arrhythmias in which heart failure and inherited arrhythmogenic diseases in which a
defects in the functional expression of Kir2.1 have been demon- defect in the membrane expression of NaV1.5 channel might
strated clearly.69,70 Moreover, a mechanistic link between affect the functional expression of Kir2.1, and vice versa.
17. Miake J, Marban E, Nuss HB: Functional role of 36. Laverty HG, Wilson JB: Murine cask is disrupted
References inward rectifier current in heart probed by kir2.1 in a sex-linked cleft palate mouse mutant. Genom-
overexpression and dominant-negative suppres- ics 53:29–41, 1998.
1. Fozzard HA, hanck DA: Structure and function of sion. J Clin Invest 111:1529–1536, 2003. 37. Caruana G, Bernstein A: Craniofacial dysmorpho-
voltage-dependent sodium channels: Comparison 18. McLerie M, Lopatin A: Dominant-negative sup- genesis including cleft palate in mice with an inser-
of brain II and cardiac isoforms. Physiol Rev pression of ik1 in the mouse heart leads to altered tional mutation in the discs large gene. Mol Cell
76:887–926, 1996. cardiac excitability. J Mol Cell Cardiol 35:367–378, Biol 21:1475–1483, 2001.
2. Lopatin AN, Nichols CG: Inward rectifiers in the 2003. 38. Boeda B, El-Amraoui A, Bahloul A, et al: Myosin
heart: An update on i(k1). J Mol Cell Cardiol 19. Plaster NM, Tawil R, Tristani-Firouzi M, et al: viia, harmonin and cadherin 23, three usher i gene
33:625–638, 2001. Mutations in kir2.1 cause the developmental and products that cooperate to shape the sensory hair
3. Noujaim SF, Berenfeld O, Kalifa J, et al: Universal episodic electrical phenotypes of Andersen’s syn- cell bundle. EMBO J 21:6689–6699, 2002.
scaling law of electrical turbulence in the mam- drome. Cell 105:511–519, 2001. 39. Verpy E, Leibovici M, Zwaenepoel I, et al: A defect
malian heart. Proc Natl Acad Sci U S A 104:20985– 20. Anderson CL, Delisle BP, Anson BD, et al: Most in harmonin, a pdz domain-containing protein
20989, 2007. lqt2 mutations reduce kv11.1 (herg) current by a expressed in the inner ear sensory hair cells, under-
4. Noujaim SF, Kaur K, Milstein M, et al: A null class 2 (trafficking-deficient) mechanism. Circula- lies usher syndrome type 1c. Nat Genet 26:51–55,
mutation of the neuronal sodium channel nav1.6 tion 113:365–373, 2006. 2000.
disrupts action potential propagation and 21. Priori SG, Pandit SV, Rivolta I, et al: A novel form 40. Boerkoel CF, Takashima H, Stankiewicz P, et al:
excitation-contraction coupling in the mouse of short qt syndrome (sqt3) is caused by a mutation Periaxin mutations cause recessive dejerine-sottas
heart. FASEB J 26:63–72, 2012. in the kcnj2 gene. Circ Res 96:800–807, 2005. neuropathy. Am J Hum Genet 68:325–333,
5. Milstein ML, Musa H, Balbuena DP, et al: 22. Li J, McLerie M, Lopatin AN: Transgenic upregu- 2001.
Dynamic reciprocity of sodium and potassium lation of ik1 in the mouse heart leads to multiple 41. Marfatia SM, Byron O, Campbell G, et al: Human
channel expression in a macromolecular complex abnormalities of cardiac excitability. Am J Physiol homologue of the drosophila discs large tumor
controls cardiac excitability and arrhythmia. Heart Circ Physiol 287:H2790–H2802, 2004. suppressor protein forms an oligomer in solution.
Proc Natl Acad Sci U S A 109(31):E2134–43, 23. Noujaim SF, Pandit SV, Berenfeld O, et al: Identification of the self-association site. J Biol
2012. Up-regulation of the inward rectifier k+ current (i Chem 275:13759–13770, 2000.
6. Gavillet B, Rougier JS, Domenighetti AA, et al: k1) in the mouse heart accelerates and stabilizes 42. Tiffany AM, Manganas LN, Kim E, et al: Psd-95
Cardiac sodium channel nav1.5 is regulated by a rotors. J Physiol 578:315–326, 2007. and sap97 exhibit distinct mechanisms for regulat-
multiprotein complex composed of syntrophins 24. Stiffler MA, Grantcharova VP, Sevecka M, et al: ing k(+) channel surface expression and clustering.
and dystrophin. Circ Res 99:407–414, 2006. Uncovering quantitative protein interaction net- J Cell Biol 148:147–158, 2000.
7. Abriel H: Roles and regulation of the cardiac works for mouse pdz domains using protein micro- 43. Melnyk P, Zhang L, Shrier A, et al: Differential
sodium channel na v 1.5: Recent insights from arrays. J Am Chem Soc 128:5913–5922, 2006. distribution of kir2.1 and kir2.3 subunits in canine
experimental studies. Cardiovasc Res 76:381–389, 25. Stanfield PR, Nakajima S, Nakajima Y: Constitu- atrium and ventricle. Am J Physiol Heart Circ
2007. tively active and g-protein coupled inward rectifier Physiol 283:H1123–H1133, 2002.
8. Papadatos GA, Wallerstein PM, Head CE, et al: k+ channels: Kir2.0 and kir3.0. Rev Physiol 44. Murata M, Buckett PD, Zhou J, et al: Sap97 inter-
Slowed conduction and ventricular tachycardia Biochem Pharmacol 145:47–179, 2002. acts with kv1.5 in heterologous expression systems.
after targeted disruption of the cardiac sodium 26. Kim E, Niethammer M, Rothschild A, et al: Clus- Am J Physiol Heart Circ Physiol 281:H2575–
channel gene scn5a. Proc Natl Acad Sci U S A tering of shaker-type k+ channels by interaction H2584, 2001.
99:6210–6215, 2002. with a family of membrane-associated guanylate 45. Tiffany AM, Manganas LN, Kim E, et al: Psd-95
9. Zhang T, Yong SL, Tian XL, et al: Cardiac-specific kinases. Nature 378:85–88, 1995. and sap97 exhibit distinct mechanisms for regulat-
overexpression of scn5a gene leads to shorter p 27. Kornau HC, Schenker LT, Kennedy MB, et al: ing k+ channel surface expression and clustering.
wave duration and pr interval in transgenic mice. Domain interaction between nmda receptor sub- J Cell Biol 148:147–158, 2000.
Biochem Biophys Res Commun 355:444–450, units and the postsynaptic density protein psd-95. 46. Leonoudakis D, Mailliard W, Wingerd K, et al:
2007. Science 269:1737–1740, 1995. Inward rectifier potassium channel kir2.2 is associ-
10. Nichols CG, Makhina EN, Pearson WL, et al: 28. Sato T, Irie S, Kitada S, et al: Fap-1: A protein ated with synapse-associated protein sap97. J Cell
Inward rectification and implications for cardiac tyrosine phosphatase that associates with fas. Sci 114:987–998, 2001.
excitability. Circ Res 78:1–7, 1996. Science 268:411–415, 1995. 47. Mohler PJ, Rivolta I, Napolitano C, et al: Nav1.5
11. Shimoni Y, Clark RB, Giles WR: Role of an 29. Garner CC, Nash J, Huganir RL: Pdz domains in e1053k mutation causing Brugada syndrome blocks
inwardly rectifying potassium current in rabbit synapse assembly and signalling. Trends Cell Biol binding to ankyrin-g and expression of nav1.5 on
ventricular action potential. J Physiol 448:709–727, 10:274–280, 2000. the surface of cardiomyocytes. Proc Natl Acad Sci
1992. 30. Bilder D: Pdz proteins and polarity: Functions U S A 101:17533–17538, 2004.
12. Nichols CG, Lopatin AN: Inward rectifier potas- from the fly. Trends Genet 17:511–519, 2001. 48. Pan Z, Kao T, Horvath Z, et al: A common
sium channels. Annu Rev Physiol 59:171–191, 31. Sheng M, Sala C: Pdz domains and the organiza- ankyrin-g-based mechanism retains kcnq and nav
1997. tion of supramolecular complexes. Annu Rev Neu- channels at electrically active domains of the axon.
13. Yang J, Jan YN, Jan LY: Control of rectification and rosci 24:1–29, 2001. J Neurosci 26:2599–2613, 2006.
permeation by residues in two distinct domains in 32. Nourry C, Grant SG, Borg JP: Pdz domain 49. Wu L, Yong SL, Fan C, et al: Identification of a
an inward rectifier k+ channel. Neuron 14:1047– proteins: Plug and play! Sci STKE 2003:RE7, new co-factor, mog1, required for the full function
1054, 1995. 2003. of cardiac sodium channel nav 1.5. J Biol Chem
14. Zaritsky JJ, Redell JB, Tempel BL, et al: The con- 33. Hildebrand JD, Soriano P: Shroom, a pdz domain- 283:6968–6978, 2008.
sequences of disrupting cardiac inwardly rectifying containing actin-binding protein, is required for 50. Abriel H, Kass RS: Regulation of the voltage-gated
k(+) current (i(k1)) as revealed by the targeted dele- neural tube morphogenesis in mice. Cell 99:485– cardiac sodium channel nav1.5 by interacting pro-
tion of the murine kir2.1 and kir2.2 genes. J Physiol 497, 1999. teins. Trends Cardiovasc Med 15:35–40, 2005.
533:697–710, 2001. 34. Zhadanov AB, Provance DW, Jr., Speer CA, et al: 51. Jespersen T, Gavillet B, van Bemmelen MX, et al:
15. Nakamura TY, Artman M, Rudy B, et al: Inhibition Absence of the tight junctional protein af-6 dis- Cardiac sodium channel na(v)1.5 interacts with and
of rat ventricular ik1 with antisense oligonucle- rupts epithelial cell-cell junctions and cell polarity is regulated by the protein tyrosine phosphatase
otides targeted to kir2.1 mrna. Am J Physiol during mouse development. Curr Biol 9:880–888, ptph1. Biochem Biophys Res Commun 348:1455–
274:H892–H900, 1998. 1999. 1462, 2006.
16. Preisig-Muller R, Schlichthorl G, Goerge T, et al: 35. Bladt F, Tafuri A, Gelkop S, et al: Epidermolysis 52. Gardoni F, Marcello E, Di Luca M: Postsynaptic
Heteromerization of kir2.X potassium channels bullosa and embryonic lethality in mice lacking the density-membrane associated guanylate kinase
contributes to the phenotype of Andersen’s syn- multi-pdz domain protein grip1. Proc Natl Acad proteins (psd-maguks) and their role in CNS dis-
drome. PNAS 99:7774–7779, 2002. Sci U S A 99:6816–6821, 2002. orders. Neuroscience 158:324–333, 2009.
53. Petitprez S, Zmoos AF, Ogrodnik J, et al: subunit of n-methyl-d-aspartate receptors. J Biol suppressed by clustering interactions with psd-95.
Sap97 and dystrophin macromolecular complexes Chem 280:26992–26996, 2005. J Biol Chem 275:1357–1364, 2000.
determine two pools of cardiac sodium channels 59. Cohen NA, Sha Q, Makhina EN, et al: Inhibition 65. Leyland ML, Dart C: An alternatively spliced
nav1.5 in cardiomyocytes. Circ Res 108:294–304, of an inward rectifier potassium channel (kir2.3) by isoform of psd-93/chapsyn 110 binds to the
2011. g-protein betagamma subunits. J Biol Chem inwardly rectifying potassium channel, kir2.1.
54. Godreau D, Vranckx R, Maguy A, et al: Expres- 271:32301–32305, 1996. J Biol Chem M407575200, 2004.
sion, regulation and role of the maguk protein 60. Leonoudakis D, Conti LR, Radeke CM, et al: A 66. Piao L, Li J, McLerie M, et al: Transgenic upregu-
sap-97 in human atrial myocardium. Cardiovascu- multiprotein trafficking complex composed of lation of ik1 in the mouse heart is proarrhythmic.
lar Research 56:433–442, 2002. sap97, cask, veli, and mint1 is associated with Basic Res Cardiol 102:416–428, 2007.
55. Gee SH, Madhavan R, Levinson SR, et al: Interac- inward rectifier kir2 potassium channels. J Biol 67. Jordens I, Marsman M, Kuijl C, et al: Rab proteins,
tion of muscle and brain sodium channels with Chem 279:19051–19063, 2004. connecting transport and vesicle fusion. Traffic
multiple members of the syntrophin family of 61. Leonoudakis D, Conti LR, Anderson S, et al: 6:1070–1077, 2005.
dystrophin-associated proteins. J Neurosci 18:128– Protein trafficking and anchoring complexes 68. Mellman I: Endocytosis and molecular sorting.
137, 1998. revealed by proteomic analysis of inward rectifier Annu Rev Cell Dev Biol 12:575–625, 1996.
56. Ueda K, Valdivia C, Medeiros-Domingo A, et al: potassium channel (kir2.X)-associated proteins. 69. Bendahhou S, Donaldson MR, Plaster NM, et al:
Syntrophin mutation associated with long qt syn- J Biol Chem 279:22331–22346, 2004. Defective potassium channel kir2.1 trafficking
drome through activation of the nnos-scn5a mac- 62. Hruska-Hageman AM, Benson CJ, Leonard AS, et underlies Andersen-Tawil syndrome. J Biol Chem
romolecular complex. Proc Natl Acad Sci U S A al: Psd-95 and lin-7b interact with acid-sensing ion 278:51779–51785, 2003.
105:9355–9360, 2008. channel-3 and have opposite effects on h+- gated 70. Li GR, Lau CP, Leung TK, et al: Ionic current
57. Oceandy D, Cartwright EJ, Emerson M, et al: current. J Biol Chem 279:46962–46968, 2004. abnormalities associated with prolonged action
Neuronal nitric oxide synthase signaling in the 63. Lin Y, Skeberdis VA, Francesconi A, et al: Postsyn- potentials in cardiomyocytes from diseased human
heart is regulated by the sarcolemmal calcium aptic density protein-95 regulates nmda channel right ventricles. Heart Rhythm 1:460–468, 2004.
pump 4b. Circulation 115:483–492, 2007. gating and surface expression. J Neurosci 71. Maltsev VA, Sabbab HN, Undrovinas AI: Down-
58. Wang L, Piserchio A, Mierke DF: Structural char- 24:10138–10148, 2004. regulation of sodium current in chronic heart
acterization of the intermolecular interactions of 64. Jugloff DG, Khanna R, Schlichter LC, et al: Inter- failure: Effect of long-term therapy with carvedilol.
synapse-associated protein-97 with the nr2b nalization of the kv1.4 potassium channel is Cell Mol Life Sci 59:1561–1568, 2002.
The Intercalated Disc: A Molecular
Network That Integrates Electrical
Coupling, Intercellular Adhesion, and
Cell Excitability
22
Marina Cerrone, Esperanza Agullo-Pascual, and Mario Delmar
other ultrastructural observations,4 the intercalated disc has been

CHAPTER OUTLINE
recognized as an area of specialization that provides a physical
Historical Perspective 215 continuum between cardiac cells through mechanical junctions
(desmosomes, adherens junctions) and intercellular channels (gap
Intercalated Disc Proteins in Inherited and
junctions).
Acquired Diseases 215 The availability of immunofluorescence microscopy allowed
Structural Features of the Intercalated Disc 216 the demonstration that other molecular complexes, not detect-
able by electron microscopy, are also present in the intercalated
Ion Channel Complexes That Reside at the disc. Of particular relevance to this chapter is the fact that channel
Intercalated Disc 217 protein complexes involved in both depolarization and repolar-
Noncanonical Functions of Intercalated Disc Molecules 218 ization localize preferentially to the intercalated disc. This physi-
cal proximity allows for a key functional consequence; molecules
The Perinexus as a Site of Functional Integration 223 traditionally defined as junctional, such as connexin43 (Cx43) and
Subcellular Localization of Ion Channels and Cell-Cell plakophilin-2 (PKP2), actually regulate the function of ion chan-
Propagation: Gap Junction–Independent Electrical nels responsible for the action potential. In turn, molecule acces-
Coupling and the Importance of the Intercellular Cleft 224 sories to ion channels are also relevant for cell adhesion and gap
junction function.5-7 These data support the notion that the inter-
calated disc is not just the site of residence of independent junc-
tional and nonjunctional complexes that are oblivious to the
presence and function of the others. It is, rather, the home of a
Historical Perspective protein interacting network (an interactome) where molecules
multitask to achieve jointly, intimately related functions: the
The heartbeat results from the added output of millions of cells entry and exit of charge into the cell, the transfer of charge
that contract in synchrony. To achieve this function, complex between cells, and the anchoring of cells to each other, which
molecular networks work in concert, with exquisite temporal provides a mechanically stable environment critical to ion channel
precision. The accurate timing of the molecular events demands function.
a comparable precision on the location of each molecule within The following sections contain an update of current knowl-
the cell. Indeed, molecular networks organize within well- edge on the composition of selected molecular complexes of the
confined microdomains, where physical proximity allows for intercalated disc, their interactions, and the possible mechanisms
prompt and efficient interaction. In turn, loss of molecular orga- by which dysfunction of intercalated disc molecules may lead to
nization in the nanoscale can be a core component in the patho- arrhythmia disease. This discussion converges with other inves-
physiology of disease. tigators to challenge the notions that: (1) connexins are only
This chapter focuses on the intercalated disc, a region of involved in the formation of gap junctions, (2) sodium channels
specialization formed at the end-end site of contact between are only important for single cell excitability, (3) desmosomal
cardiac myocytes. When first observed through light microscopy molecules are only relevant to cell adhesion, and (4) it is only
(in 1866), the intercalated disc was considered “a cementing through modifications of those functions that these proteins par-
material” at cardiac cell boundaries. However, the scientific com- ticipate in the genesis of lethal cardiac arrhythmias, or are poten-
munity at the time was divided on whether cardiac cells were tially valuable as targets for antiarrhythmic therapy.
separate from each other or fused into a single syncytium. The
latter hypothesis was in fact favored by most during the early
twentieth century. The advent of electron microscopy eventually
settled this debate. The studies of Sjostrand and Andersson1 and Intercalated Disc Proteins in Inherited
others showed that the intercalated disc consisted of a double and Acquired Diseases
membrane, flanked by the termination of myofibrils in dense
material. Their observations led Muir2 to conclude that “the The function of intercalated disc components is relevant not only
discs represent the junctions between neighboring cardiac to normal physiology, but also to the understanding of disease. It
muscle cells.” He wrote that “there is no valid evidence to is not the purpose of this chapter to review clinical aspects of
contest the statement that the intercalated discs are specialized arrhythmias, but it seems worth mentioning at the outset selected
regions of cellular adhesion.” Since then, and as a result of the examples where novel findings regarding intercalated disc biology
pioneering electrophysiology experiments of Weidmann3 and can provide insight into arrhythmia mechanisms.
215
Initial studies on the relation between desmosome integrity nisms of ventricular arrhythmias between patients with ARVC
and cardiac electrophysiology were propelled by the finding and Brugada syndrome.”
that most familial cases of arrhythmogenic right ventricular car-
diomyopathy (ARVC) in which a genetic link has been found
associate with mutations in genes coding for desmosomal mole-
cules.8 The latter brought forth the question of how a molecule Structural Features of the Intercalated Disc
considered purely relevant to cell adhesion altered the electrical
behavior of the heart. The associations between desmosomal In its classic definition, the intercalated disc is composed of three
proteins and ion channel function (particularly the sodium current electron-dense structures: adherens junctions, desmosomes, and
INa) are extensively reviewed in this chapter. (Recent publications gap junctions (Figure 22-1). Additional reviews on the character-
refer to this disease as “arrhythmogenic cardiomyopathy,” to note istics of these structures can be found elsewhere.24,25 These struc-
the occurrence of left ventricular involvement.8) tures are described briefly in the following sections, and the
A number of studies have demonstrated remodeling of gap structural and molecular definitions of the intercalated disc are
junction proteins (in particular Cx43) in a number of inherited expanded to include the area composita, the intercellular space,
and acquired arrhythmia-related diseases.9-13 Recently, the and the nonjunctional ion channel complexes.
Fishman laboratory provided evidence that aberrant posttransla-
tional phosphorylation of Cx43 could be the common pathway
leading to pathologic gap junction remodeling and arrhyth- Adherens Junctions
mias.10,14 Although there is a relationship between Cx43 remodel-
ing and arrhythmias, it is unclear whether these arrhythmias are Adherens junctions are specialized structures essential for the
exclusively consequent to changes in gap junction formation.15 mechanical coupling between neighboring cells. The three mor-
Recent data show that reduced Cx43 expression alters sodium phologically different forms of adherens junctions are puncta
and potassium current function, and these changes could become adherentia, zonula adherens, and fascia adherens, with the last
part of the arrhythmogenic substrate.16-18 name corresponding to the morphology found in the cardiac
Disruption of the voltage-gated sodium channel complex is intercalated disc.26 Cell-cell mechanical anchoring occurs at two
considered an important molecular substrate for arrhythmogen- crucial points: the extracellular space, within which cadherins
esis. Extensive reviews on the relation between mutations in tightly bind to each other, and the intracellular space, within
proteins of the voltage-gated sodium channel (VGSC) complex which the cytoplasmic end of cadherin is indirectly attached to
and arrhythmias are available.19 Of particular interest in the the actin cytoskeleton. The association between cadherin and the
context of this review is the observation that haploinsufficiency cytoskeleton involves at least two molecular “hinges”; cadherin
of a desmosomal protein, or overexpression of a mutant desmo- binds to β-catenin and plakoglobin, and both molecules in turn
somal protein, leads to INa deficit and increased arrhythmia sus- bind to α-catenin (among others), the latter being in direct
ceptibility.20,21 The latter recalls the previously formulated contact with actin. This is only a simplified description, because
concept that ARVC and Brugada syndrome (a channelopathy other interactions are likely to occur.27 This string of intermo-
caused by mutations on genes of the VGSC complex) share lecular interactions provides mechanical continuity between cells,
common features22 and that, as Corrado et al.23 stated, there is allowing for the mechanical work of individual myocytes to inte-
the possibility of “an overlap in clinical manifestation and mecha- grate into the pumping function of the heart.
GJ
A 2 µm B 200 nm
D
GJ
M
GJ
C 200 nm D 200 nm E
Figure 22-1. Transmission electron micrographs of intercalated disc in adult murine heart. Tissue prepared by high-pressure freezing and freeze substitution, thus greatly
improving structural preservation. A, Low magnification to demonstrate the entire length of the intercalated disc as it meanders between two cells. B and C, Proximity
(and contact in C, yellow arrow) between mitochondria, gap junctions, and desmosomes. D, Example of close proximity between a gap junction and a desmosome.
E, Compiled three-dimensional tomographic electron microscopy reconstruction of the intercalated disc. Notice the extensive vesicular activity between the two cells.
D, Desmosomes; GJ, gap junction; M, mitochondria.
(Reproduced from Delmar M, Liang FX: Connexin43 and the regulation of intercalated disc function. Heart Rhythm 9:835–838, 2012.)
THE INTERCALATED DISC: A MOLECULAR NETWORK 217
Desmosomes heart, the Wnt signaling and the activation of genes by β-catenin/
22
Tcf/Lef has been associated with the regulation of physiologic
The desmosome (macula adherens) appears as two parallel tri- and pathologic growth of the cardiomyocytes.44 Plakoglobin,
partite plaques containing an intercellular gap of approximately another protein of the armadillo family, shows high homology
30 nm bisected by a distinct line, parallel to the apposed cell with β-catenin and has been associated with the Wnt signaling.
membranes (see Figure 22-1).28 Desmosomes contribute to Different studies have shown that plakoglobin interacts and com-
mechanical continuity between cardiac cells. Whereas adherens petes with β-catenin at multiple levels, acting as an antagonist of
junctions link the actin cytoskeleton of adjacent cells, desmo- the Wnt/β-catenin signaling.45-47 The fact that desmosomal pro-
somes provide continuity to the intermediate filament network teins are involved in the regulation of the Tcf/Lef complex has
(mainly desmin, in the case of heart).28,29 In the extracellular been invoked as a possible mechanism for the fibrofatty infiltra-
space, desmosomal cadherins (desmocollins and desmogleins) tion common in hearts affected with ARVC.48-50
bind tightly to each other. In the intracellular space, the inter-
mediate filaments bind to desmoplakin. The interaction between
desmoplakin and the desmosomal cadherin can be in some cases Gap Junctions
direct, but it mostly occurs through their association with pla-
kophilin and plakoglobin.28,29 The topologic organization of des- In 1958, Sjostrand et al.4 described an area of specialization in
mosomal molecules was studied by North et al.30 using the cardiac intercalated disc composed of “three dark lines with
quantitative immunogold electron microscopy. More recently, two intervening less dense lines”. This structure, which was
Al-Amoudi et al.31 solved the three-dimensional molecular struc- similar to the one previously identified in the giant axon of the
ture of the desmosomal plaque. Overall, structural and biochemi- crayfish, was named the “longitudinal connexion” by these inves-
cal evidence combined show that desmoplakin binds to plakophilin tigators. Years later, Revel coined the term gap junctions, thus
through their N-terminal domains,28,32 whereas desmoplakin emphasizing two key features: a gap between the cells and a junc-
binds to the intermediate filament by way of its C-terminal tion between them.
domain,28,31 yielding a highly organized structure. Gap junctions form intercellular channels that provide a low-
Hatsell and Cowin29 once described the desmosome as “a resistance pathway for direct cell-to-cell passage of electrical
system as staid and solid as the queen’s corsets.”29 With that charge between cardiac myocytes. Each gap junction channel is
analogy, it is easy to imagine the loss of containment that would composed of two hexameric structures called connexons that dock
follow in its absence. Mice deficient in plakoglobin, desmoplakin, across the extracellular space and form a permeable pore isolated
or plakophilin-2 (PKP2), die during embryonic development as from the extracellular space. Each connexon results from oligo-
a result of severe myocardial rupture.33-35 More relevant from the merization of an integral membrane protein, connexin. The most
point of view of clinical cardiology, ARVC in humans has been abundant connexin isotype in the heart, brain and other tissues
linked to mutations in desmosomal proteins.36 The relation is the 43-kD protein, connexin43 (Cx43).
between various complexes of the intercalated disc and ARVC is The importance of Cx43 in the propagation of the cardiac
discussed later in this chapter and in other review articles.24,37 action potential is well established. If Cx43 channels are not
present, normal propagation is disrupted and lethal arrhythmias
can ensue.51,52 Gap junction remodeling has been studied for
The Area Composita various inherited and acquired diseases.10,11,14,53-55 The implica-
tions of connexin remodeling to electrophysiology are discussed
Recent immunoelectron microscopy studies revealed the pres- later in this chapter.
ence of a structure with mixed features of desmosomes and adhe-
rens junctions, dubbed the area composita.38 This structure is
found only in the heart of higher vertebrate species including Intercellular Space
mouse and humans.39,40 The combination of components of the
desmosomes and the adherens junctions allows anchorage of The size of the space separating two cardiac cells at the interca-
actin and desmin filaments to the same point, perhaps providing lated disc changes depending on the proximity to the various
additional strength and flexibility to the muscle cell. Knockdown structures, as well as the vesicular activity between the two cells
of PKP2 in neonatal cardiomyocytes leads to remodeling of the (see Figure 22-1). It is a common view that the intercellular space
area composita.41 Additional studies suggest a role for α-catenin is not relevant for electrophysiology. This view, however, is
in the maintenance of these hybrid junctions.42 Interestingly, changing. Mathematical modeling studies56,57 and experimental
knockdown of α-catenin leads to PKP2 decrease only at the area evidence58 support the idea that the intercellular space is critical
composita and not at the desmosomes, suggesting that the molec- to propagation via an electric field mechanism.59 This model will
ular composition of the area composita, and its regulation, can be discussed later. Of note, increased size of the intercellular
be independent from that of other junctions.42 The area com- space has been reported in animal models of ARVC.20,21,60,61
posita may represent a physical space where mechanical junction
proteins interact with ion channel complexes.
Transcriptional Regulation by Mechanical Ion Channel Complexes That Reside at the

Junction Proteins Intercalated Disc
Catenins are a part of both adherens junctions and desmosomes, Voltage-Gated Sodium Channel Complex
thus participating in cell adhesion; however, these proteins also
act as transcriptional activators. A prominent example is the par- In 1996, Cohen62 showed that cardiac sodium channel proteins
ticipation of β-catenin in canonical Wnt signaling.43 Binding of were preferentially localized at the intercalated disc,62 although
Wnt to its Frizzled receptor leads to an increase in levels of they are also present over the cell surface, following a striated
cytosolic β-catenin and a consequent translocation of the protein pattern.63 Recent data emphasize how the VGSC interacts also
to the nucleus, where it binds to Tcf/Lef complex and promotes with scaffolding, anchoring, and adhesions proteins, which regu-
the expression of various genes such as c-Myc or c-Fos. In the late its function.
α-Subunit NaV1.5 where NaV1.5 interacts with the MAGUK proteins SAP97 and
The cardiac, pore-forming, α-subunit NaV1.5 contains four ZO1, as well as with PKP2.72 Lin et al. recently demonstrated
homologous transmembrane domains DI-DIV, linked by intra- that this subcellular distribution correlates with differences in
cellular loops (IDI-II, IDII-III, IDIII-IV). Each domain is formed function.65 Using cell-attached macropatch recordings, these
by the six transmembrane segments S1 to S6, and is involved in authors showed that the magnitude of INa is larger at the inter-
the voltage-dependent activation of the channel. The channel calated disc than in the midsection of the cell. Furthermore,
pore conducting Na+ is lined by the S6 segment and the S5-S6 TTX-resistant channels in the midsection showed a significant
pore loops in each domain.19 The inactivation gate is a complex negative shift in the steady-state inactivation curve, suggesting
formed by the DIII-IV loop and the C-terminus.64 This channel that these channels are mostly in an inactivated state at the
is defined as TTX-resistant, in contrast with other Nav channels. normal resting potential, and that the burden of excitation is on
Of note, TTX-sensitive channel proteins, and currents, have also the channels at the intercalated disc. These authors also demon-
been found in cardiac myocytes, with the current localized pri- strated that the amplitude of current is larger if cells remain
marily to the midsection of the cell.63,65 paired, strongly suggesting that cell adhesion preserves sodium
The SCN5A gene, on chromosome 3p21, codes for the NaV1.5 channel function. The interaction between sodium channels and
subunit. Mutations resulting in increased INa are associated with proteins of the intercalated disc is discussed extensively later.
long QT syndrome type 3. Mutations that cause a reduction in
INa are responsible for a series of different diseases, such as
Brugada syndrome, progressive cardiac conduction defect, sick Potassium Channels at the Intercalated Disc
sinus syndrome, and a form of inherited atrial fibrillation. Some
mutations are associated with a clinical spectrum encompassing In 1995, Mays et al.73 described for the first time that the potas-
more than one of those phenotypes and can manifest differently sium channel protein KV1.5 localizes to the intercalated disc of
among carriers, even within the same family.19 Interestingly, some adult myocardial cells. Later studies showed that KV1.5 associates
SCN5A mutations have been linked to a form of dilated cardio- with SAP97.74 Interestingly, SAP-97 also associates with NaV1.5,71
myopathy with a high incidence of atrial and ventricular arrhyth- making this protein a candidate for mutual regulation of both
mias.66 It is not yet clear how mutations in the sodium channel depolarization and repolarization channels, as is the case of
could lead to structural damage of the myocardium. Based on the NaV1.5 and Kir2.1.75 Additional studies have shown that the func-
crosstalk between intercalated disc structures described in this tion of KV1.5 depends on the expression of N-cadherin,76 an
chapter, we are tempted to speculate that the integrity of the interesting parallelism to the interaction between PKP2 (a des-
sodium channel complex is also relevant to intercellular adhesion mosomal molecule), and NaV1.5.72 Cheng et al.76 also showed
strength.20,21,67 that cortactin is required for the N-cadherin–dependent regula-
tion of KV1.5. Of note, mice deficient in kcne2, an ancillary
β-Subunits of the Sodium Channel potassium channel subunit, display an impaired ventricular repo-
The NaV β-subunit family consists of four proteins: β1-4, coded larization because of inhibition in the trafficking to the mem-
by genes SCN1B to SCN4B, respectively. These are single-span brane of KV1.5.77
transmembrane proteins oriented with the amino terminus facing Another voltage-gated potassium channel that preferentially
the extracellular space. The extracellular domain presents a con- (though not exclusively) localizes to the intercalated disc is KV4.2;
served immunoglobulin domain, homologous to the one in cell it is responsible for the rapid repolarization phase of the cardiac
adhesion molecules.68 The carboxyl terminus associates with action potential.78,79 Finally, regarding Kir channels, two different
cytoskeletal and scaffolding proteins. β1 and β2 are localized at subunits localize at the transverse tubules and at the intercalated
the T-tubules–Z lines and at the intercalated discs in rat cardiac disc in canine myocytes: Kir2.1 and Kir2.3.80
myocytes. β3 colocalizes with β1 at the T-tubules and β4 colocal-
izes with β2 at the intercalated disc. β1 and β2 associate with
ankyrin-G and ankyrin-B in both brain and heart, and their
interaction is critical for channel surface expression and modu- Noncanonical Functions of Intercalated
lates the channel function in vivo. β1 and β2 associate with Disc Molecules
N-cadherin and with Cx43 at the intercalated disc.69 Altogether,
β-subunits have a key role in the interactions between the VGSC The various complexes described in the preceding sections do
multiple proteins at the intercalated disc, including those relevant not exist in isolation. Rather, they dynamically interact with each
for cell adhesion and for electrical coupling between cells.69 Fur- other. Through these interactions, they exert unconventional
thermore, work from the Isom lab has demonstrated that null (i.e., noncanonical) functions. Some of those functions are
mice for the β1 subunit show a significant increase in SCN5A described as follows.
mRNA in cardiac myocytes.70 Further research is needed to elu-
cidate the role of β-subunits in transcription regulation in the
heart.70 Overall, the data show that, as in the case of catenin and Desmosomal Proteins Are Necessary for the
plakoglobin, β-subunits can have distal, contact-dependent Formation of Functional Gap Junctions
effects, including regulation of gene transcription, with conse-
quences to the function and structure of the heart. Studies on the relation between desmosome integrity and cardiac
electrophysiology were motivated by the finding that most famil-
Subcellular Heterogeneity of Voltage-Gated Sodium Channel ial cases of ARVC where a genetic link has been found, result
Recent studies have shown that not all sodium channels on the from mutations in genes coding for desmosomal molecules. Dr.
surface of the cardiac myocyte are equal. Instead, the molecular Saffitz and his colleagues were first to propose a link between
composition and the function of a sodium channel are different mechanical and electrical junctions.54,81 In a series of seminal
depending on whether NaV1.5 is localized to the intercalated disc studies, these investigators first provided evidence that the
or to the midsection of the cell. Petiprez et al.71 described two molecular phenotype of the ARVC-afflicted heart involves not
separate pools of VGSC in ventricular cardiomyocytes. One of only desmosomes, but other molecules of the intercalated disc as
these two subpopulations localizes at the lateral membrane of the well.53,54 In particular, the authors showed a significant decrease
myocytes, where NaV1.5 interacts with dystrophin and syntro- in the abundance of Cx43 immunoreactive protein in the inter-
phin; the second subpopulation of VGSC localizes at the ID, calated disc region of heart tissue obtained from affected patients.
The hypothesis of a desmosome–gap junction crosstalk was con-

Desmosomal Molecules Are Necessary for Sodium
22
firmed in vitro by Oxford et al.,82 who used RNA silencing
technology to reduce PKP2 expression in cardiac ventricular Channel Function
myocytes, as well as in epicardium-derived cells obtained from
neonatal rat hearts. Their data showed that the loss of PKP2 The observations described earlier led to the question of whether
expression led to a decrease in total Cx43 content, a significant Cx43 is the only molecule relevant to electrophysiology that
redistribution of Cx43 to the intracellular space, and a decrease interacts with PKP2. Given the preferential localization of the
in dye coupling between cells. Separately, they demonstrated that VGSC complex at the intercalated disc, we explored its possible
Cx43 and PKP2 coexist in the same macromolecular complex. cross-talk with desmosomal molecules. Single, adult cardiac myo-
Follow up studies confirmed this observation,41,83,84 giving support cytes were treated with small interfering RNA (siRNA) to prevent
to the notion that two complexes previously considered indepen- expression of PKP2. Control cells were treated with a nonsilenc-
dent are in fact, functionally and molecularly interactive. Recent ing construct. As shown in Figure 22-2, A, the amplitude of INa
studies on samples obtained from hearts affected with ARVC was significantly reduced in cells lacking PKP2 expression. This
have confirmed the notion that in most (though not all) cases, reduction in amplitude was observed across the voltage range (see
there is a significant loss of Cx43 immunoreactive signal associ- Figure 22-2, B). In addition, loss of PKP2 expression caused a
ated with the loss of desmosomal integrity in the cardiac inter- negative shift in the voltage dependence of steady-state inactiva-
calated disc.85,86 tion (see Figure 22-2, C), and a slowing of recovery from inactiva-
Quantitative analysis in experimental models indicated that tion (see Figure 22-2, D). These changes in the major excitatory
complete loss of PKP2 expression led to an approximate 50% current were reflected in a decrease in the velocity of action
decrease in gap junction–mediated cell-cell coupling.82 Previous potential propagation in monolayers of neonatal rat ventricular
studies had shown, however, that a 50% reduction in electrical myocytes (Figure 22-3). Interestingly, computer simulations
coupling does not lead to significant changes in conduction showed that the changes in amplitude and kinetics of the INa rep-
velocity.87-89 We therefore speculated that in addition to interact- resented the key substrate for the generation of reentrant activity
ing with gap junctions, desmosomal molecules can also interact in a two-dimensional model of cardiac cells.90 Overall, these data
with ion channel complexes that reside at the intercalated disc. led us to propose that there is a functional crosstalk between a
INa I-V RELATIONSHIP

UNT pA/pF
20
–100 –80 –60 –40 –20 20 40
Vm (mV)
KD –20
–40
UNT
KD –60
KD
φKD
20 pA/pF –80
A 10 msec B
INA INACTIVATION RECOVERY FROM INACTIVATION
1.0 UNT 1.0

KD
0.8 φKD 0.8
Normalized INa
0.6 0.6
I/Imax
0.4 0.4
0.2 0.2 UNT

KD
0.0 0.0 φKD
–140 –120 –100 –80 –60 –40 –20 0 20 40 60 80

C Time (msec) D Time (msec)
Figure 22-2. Voltage clamp data obtained from adult cardiomyocytes after knockdown of PKP2. Cells were untreated (UNT) or treated with a nonsilencing construct (ϕKD)
or with oligonucleotides that prevent PKP2 expression (KD). A, Examples of sodium currents recorded under the three different conditions. B, Complete voltage dependence
of sodium current density. C, Voltage dependence of steady-state inactivation. D, Recovery from inactivation. Notice that the loss of PKP2 expression decreased sodium
current density, caused a negative shift in steady-state inactivation, and prolonged recovery from inactivation. Further details in.72
A B
10
φ shRNA
30 12
11 Untreated
11 shRNA
12
Mean velocity (cm/s)
12 10 9
25 8 6 +π
9
9
9
20 7 9
7 3
6 7 6 6
7 6 5
6 2
5
15 5
10
1 2 3 4 5 6 7 8 9 10
–π
C Frequency (Hz) D
Figure 22-3. Optical mapping of action potential propagation in monolayers of neonatal rat ventricular myocytes. Isochrome maps from monolayers were treated with a
nonsilencing construct (A) or silenced for PKP2 (B). C, Conduction velocity as a function of pacing frequency. D, A phase map of spontaneous reentrant activity in a PKP2-
deficient monolayer.
(Reproduced from Sato PY, Musa H, Coombs W, et al: Loss of plakophilin-2 expression leads to decreased sodium current and slower conduction velocity in cultured cardiac
myocytes, Circ Res 105:523–526, 2009.)
protein defined in the context of intercellular junctions (PKP2), et al.,20 flecainide provoked ventricular arrhythmias and death in
and another complex primarily involved in supporting cell excit- PKP2-Hz animals, but not in wild type. These results showed
ability (the VGSC complex). The results supported the hypoth- that PKP2 haploinsufficiency leads to INa deficit in murine hearts,
esis that arrhythmias that occur in patients with ARVC may have thus documenting for the first time the relation between the
as a substrate not only changes in the macroscopic architecture of “desmosomal molecule” PKP2 and the VGSC complex in a living
the tissue (as it would be expected once the fibrofatty infiltrations heart. Our results supported the contention that INa dysfunction
populate the heart), or in the integrity of intercellular coupling, contributes to generation and maintenance of arrhythmias in
but also changes in the electrical properties of the cardiac myo- patients with desmosomal deficiency. It remains unclear whether
cytes. Of relevance, these results revealed that a property “of the pharmacologic challenges could help to unveil arrhythmia risk in
single cell” (excitability; inward sodium current [INa]) is in fact patients with mutations or variants in PKP2.
subject to modulation by proteins classically defined as belonging The relation between desmosomal molecules and the VGSC
to the group of intercellular junction molecules. has also been demonstrated in mice overexpressing a desmoglein-2
PKP2 mutations associated with ARVC have all been found in (DSG-2) mutation.21 This model is interesting given that severe
only one allele. We therefore characterized the relation between arrhythmias and sudden death in these mice often occur before
PKP2 abundance and sodium current function in mice that were structural damage; the latter mimics a phenomenon often
haploinsufficient for the pkp2 gene (PKP2-Hz).20 Of note, one of observed in humans affected with ARVC, where arrhythmias and
the most common mutations in PKP2 is the presence of a stop electrocardiographic changes have been described early in the
codon at amino acid position 79 (R79x).91 This early truncation history of the disease, before overt structural changes in the
is functionally equivalent to haploinsufficiency.92 Patch clamp myocardium.24,54,93 Electrophysiological analysis of these DSG-2
experiments showed a decreased amplitude and a shift in gating transgenic mutant mice revealed prolonged ventricular activation
and kinetics of INa in PKP2-Hz myocytes, compared with control time, decreased conduction velocity in both longitudinal and
myocytes. To further unmask INa deficiency, we exposed myo- transverse directions, and increased arrhythmia susceptibility,
cytes, Langendorff-perfused hearts and anesthetized animals to a prior to signs of fibrosis or necrosis in the myocardium. The
pharmacologic challenge (flecainide). In PKP2-Hz hearts, the authors also observed a decrease in maximal upstroke velocity and
extent of flecainide-induced INa block, impaired ventricular con- decreased INa amplitude. Taken together, the data demonstrate
duction, and altered electrocardiographic parameters were larger that a decrease in PKP2 abundance,20 as well as the overexpres-
than controls. As shown in Figure 22-4 and described by Cerrone sion of a DSG-2 mutation,21 lead to impaired INa function in
Baseline
WT
Flecainide (10')
22
50 mV
500 ms
PKP2-Hz
100 mV 100 mV
A 20 ms B 500 ms
Figure 22-4. Electrocardiographic (ECG) features of PKP2-Hz mice at baseline, and in response to flecainide. A, Examples of ECG traces from wild type (top) and PKP2-Hz
mice (bottom). Recordings obtained at baseline (left) and 10 minutes after flecainide (40 mg/kg intraperitoneally, right). B, ventricular tachycardia (VT) in PKP2-Hz mice.
Overall, the flecainide caused a prolongation of the P and QRS durations, and of the PR and QTc intervals that were significantly more pronounced in PKP2-Hz animals than
in control. Six of 12 PKP2-Hz showed ventricular arrhythmias. None of the 11 wild type mice tested presented ventricular arrhythmias. Arrhythmic death occurred in three
PKP2-Hz animals and in none of the control animals.
(Reproduced from Cerrone M, Noorman M, Lin X, et al: Sodium current deficit and arrhythmogenesis in a murine model of plakophilin-2 haploinsufficiency. Cardiovasc Res
95:460–468, 2012.)
hearts with no histologic features of ARVC. These data provide Cx43 expression in adult ventricular myocytes leads to a decrease
a demonstration in vivo of the interaction between desmosomal in the amplitude of the INa. The functional effect coincided with
molecules and the VGSC, and they suggest impaired sodium decreased colocalization of Cx43 and NaV1.5 at the intercalated
current as a substrate for lethal arrhythmias in the concealed disc. A similar decrease in INa was later reported by Desplantez
phase of ARVC, as proposed earlier.72 et al.18 in fetal atrial myocytes of Cx43-deficient mice.18 Overall,
The question remains as to whether other desmosomal pro- the data demonstrate that Cx43 expression is necessary for proper
teins also interact with NaV1.5, and whether these interactions sodium current function, and for the accumulation of NaV1.5 at
occur in the human heart. Recently, Gomes et al.94 reported that the cardiac intercalated disc. Interestingly, the regulation of INa
mice that are haploinsufficient for desmoplakin present average by Cx43 is reciprocated by the fact that ankyrin-G, a molecule
peak INa density similar to control; however, careful analysis of that is well characterized as a component of the VGSC complex,
their results suggests the possibility of technical limitations in is necessary to preserve gap junction–mediated coupling between
their voltage clamp recordings, which could have masked small neonatal myocytes.67
differences between the groups.20 However, in the same study, The VGSC is not the only electrically functional complex of
the authors showed that patients with heterozygous mutations in the intercalated disc that is disrupted consequent to loss of Cx43
desmoplakin and without overt structural disease had significant expression. In fact, the first report correlating Cx43 expression
regional conduction delays and heterogeneous NaV1.5 distribu- to nonjunctional currents was by Danik et al.16 These authors
tion.94 The possibility of changes in the abundance or colocaliza- noted that the action potential duration recorded from the ven-
tion of NaV1.5 in the intercalated disc area of the hearts of tricle of Cx43 conditional knockout animals were significantly
patients with ARVC is a matter of current investigation. shorter than control. This shortening associated with higher
levels of sustained repolarizing current and higher levels of
inward rectifier current in myocytes from the right ventricle.
Connexin43 Regulates Sodium Overall, the data show that Cx43 is not only a gap junction–
and Potassium Currents forming molecule in the heart but also, a component of a molecu-
lar network that regulates excitability and repolarization.
Loss of Cx43 expression leads to propagation block and to
arrhythmias. Interpretation of this result has centered mostly on
the role of Cx43 as the pore-forming subunit of gap junctions. AnkG and Cx43 Are Necessary to Maintain
However, the latter does not exclude the possibility that Cx43 Intercellular Adhesion
could also interact with other channel complexes, and affect their
function. In fact, there is no reason to confine Cx43 to a single The previous sections have described that molecules of the des-
task, in a single structure. mosome, relevant to intercellular adhesion (namely PKP2 and
The first indication of an interaction between Cx43 and the DSG-2), are actually important to preserve INa amplitude and
VGSC came from the work of Malhotra et al., showing copre- electrical coupling between cardiac cells. We then speculate that,
cipitation of NaV1.5 with Cx43.95 The physical proximity of these if the system works as a unit, decreased protein levels of either
molecules was recently confirmed by Rhett et al.96 Not only are Cx43 or a component of the VGSC would affect intercellular
Cx43 and NaV1.5 in close proximity, but these two molecules are adhesion strength. Consistent with this hypothesis, we have
also functionally intertwined. Indeed, as shown in Figure 22-5, shown that intercellular adhesion strength is decreased in mono-
Jansen et al.17 recently reported that siRNA-mediated loss of layers of neonatal rat ventricular myocytes treated with siRNA
INa I-V RELATIONSHIP

UT φKD KD pA/pF
10
–100 –80 –60 –40 –20 20
Cx43 –10
Vm (mV)
–20
–30
Cx43sil
Cx43scr
–40
–50
–60
α-actinin
–70
–80
A B
INa INACTIVATION RECOVERY FROM INACTIVATION

1.0
1.0 Cx43sil
Cx43scr
0.8
0.8
Normalized INa
0.6
0.6
I/Imax
0.4
0.4
Cx43sil
Cx43scr
0.2 0.2
0.0 0.0
–140 –120 –100 –80 –60 –40 0 10 20 30 40 50

C Voltage (mV) D Time (msec)
Figure 22-5. Decreased connexin43 (Cx43) expression leads to reduced INa in isolated adult rat ventricular myocytes. A, Western blot for Cx43 in cells untreated (UT), treated
with an oligonucleotide that prevented Cx43 expression (KD) or a non-targeting construct (ϕKD). B, Peak sodium current density was lower in cells lacking Cx43. C and D,
Loss of Cx43 expression did not affect steady-state inactivation or recovery from inactivation kinetics.
(Reproduced from Jansen JA, Noorman M, Musa H, et al: Reduced heterogeneous expression of Cx43 results in decreased Nav1.5 expression and reduced sodium current that
accounts for arrhythmia vulnerability in conditional Cx43 knockout mice. Heart Rhythm 9:600–607, 2012.)
to prevent expression of AnkG.67 Similarly, loss of Cx43 expres- that followed demonstrated their critical role in cancer.103 Inter-
sion in cultured cells significantly impaired intercellular adhesion estingly, the fact that β-subunits are actually independent adhe-
strength,7 a result consistent with previous observations.97 sion molecules has, for the most part, missed the attention of
cardiac electrophysiologists. Another case of a molecule’s identity
being boxed into the function related to its original discovery. If
The “α Personality” of the β-Subunit: Intercellular that molecule had been found through cell adhesion assays
Adhesion and the Sodium Current and called adherin, perhaps its role as a modulator of INa could
have gone unnoticed for years. The fact is that the sodium
The finding that mutations in desmosomal molecules associate channel β-subunits are, together with PKP2, DSG-2, AnkG, and
with familial cases of ARVC has highlighted the important link Cx43, members of a growing family of molecules that regulate
between cell adhesion and sodium channel function. In retro- sodium channel function and intercellular adhesion strength.
spect, the link between these two seemingly unrelated functions Whether β-subunits regulate gap junction–mediated coupling
was first established several years ago by the Isom lab. In 1981, (as the other molecules listed earlier) is a matter of future
Hartshorne and Caterall98 purified “the saxitoxin receptor of the investigation.
sodium channel from rat brain” and identified two polypeptides,
which they referred to as “α” and “β.”98 They proposed that these
two subunits conformed the functional sodium channel. In 1992, Other Intercellular Adhesion Molecules That Cross-
Isom et al.99 isolated the cDNA, sequenced and functionally Talk with Cardiac Ion Channels
expressed the β-1 subunit, concluding that this protein is
crucial to the overall function of the sodium channel. In this Two other adhesion molecules have been associated with cardiac
manner, this 22,581-d protein was labeled as a “β” for its “α,” a electrical function: N-cadherin and the coxsackievirus and adeno-
subunit merely accessory to sodium channel function. It was 8 virus receptor (CAR). N-Cadherin is well recognized as critical
years later (in 2000) that the Isom lab demonstrated that “sodium to the mechanical coupling between cells. However, Li et al.104
channel beta subunits” also mediate cell adhesion,100 an important showed that restricted cardiac deletion of N-cadherin also leads
fact in the formation of the sodium channel complex100,101 and in to severe arrhythmias, and a loss of Cx43 from the intercalated
sodium channel–independent functions such as cell migration, disc. More recently, the Radice lab reported that in addition to
cell aggregation, and interaction with the cytoskeleton.102 Studies effects on coupling, loss of N-cadherin leads to decreased density
of the repolarizing current IK.slow, concurrent with decreased junction–forming Cx43 connexons. Their studies showed that
22
expression of KV1.5 and its accessory protein Kcne2.76 The Cx43 in this area, which they dubbed the perinexus, closely associ-
properties of the INa in cadherin-deficient hearts remain ates with the scaffolding protein zonula occludens-1 (ZO-1);
undefined. more importantly, they demonstrated that the Cx43–ZO-1 asso-
CAR is another case of a molecule whose identity is burdened ciation in the perinexus limits the abundance of Cx43 within the
by its birth name. Undoubtedly a receptor for both coxsackievirus gap junction plaque (Figure 22-6). As such, loss of the Cx43–
and adenovirus, studies have demonstrated that the immuno- ZO-1 interaction increases the proportion of Cx43 involved in
globulin extracellular domains of CAR are capable of homophilic gap junction channel formation at the expense of a non–channel-
binding and participate in intercellular adhesion in epithelial forming pool.
cells. The role of this molecule on cell adhesion in the heart is The finding of a separate pool of Cx43 in the intercalated disc,
less defined. Interestingly, cardiac-restricted deletion of CAR invites speculation as to its possible function. It is unlikely that
causes significant slowing of A-V propagation and disruption of the Cx43 molecules in the perinexus are simply in standby mode,
gap junctions at the intercalated disc.105,106 Research is in progress waiting idle for the signal to move into the gap junction plaque.
to determine whether, as other molecules with immunoglobulin It is also unlikely that Cx43 and ZO-1 are the only inhabitants
extracellular domains (such as the sodium channel β-subunits), of the perinexal space. Rather, this pool of Cx43 is likely to be
CAR expression modulates sodium current function. exposed to a variety of other molecular complexes that approach,
but are not components of, the gap junction plaque. In this space,
Cx43 can act as accessory to the function of other molecules, such
as NaV1.5. Biochemical evidence of a physical proximity between
The Perinexus as a Site Cx43 and NaV1.5 was first reported by Malhotra et al.95 Recently,
of Functional Integration Rhett et al.96 used a proximity ligation assay (PLA; also known
by the proprietary name of Duolink) to demonstrate, in neonatal
The electron micrographic images shown in Figure 22-1 dem- rat ventricular myocytes, that Cx43 and NaV1.5 co-inhabit the
onstrate the proximity between electron-dense structures (e.g., perinexus. Loss of Cx43 in this space can lead to reduction in
gap junctions to desmosomes). Distances could be even shorter INa.17,18,65 Overall, although still speculative, there is reason to
for molecules in the perimeter of either a gap junction or a des- believe that the perinexus represents the physical space where
mosome plaque. In that context, it is important to mention that Cx43-dependent, gap junction–independent functions take place.
Cx43 at the intercalated disc is not exclusively circumscribed to The data draw a portrait of Cx43 as the highly regulatable mol-
the gap junctions. Rhett et al.107 demonstrated that the area ecule that has been extensively described and as a regulator of the
surrounding a gap junction plaque is populated by non–gap function of other molecular complexes.
PERINEXUS MODEL
Gap junction Connexon hemichannel

Actin fiber
ZO-1
Intercellular channel
GJ edge
us
ex
rin
Pe
Figure 22-6. Diagram of the perinexus, described by Rhett et al.107 According to the model, connexons gather in the periphery of the gap junction. The interaction with
ZO-1 prevents the transfer of the connexon into the gap junction plaque. Importantly, the connexon in the perinexus is not forming gap junctions. Instead, the perinexus
may be the region where Cx43 interacts with other intercalated disc proteins, such as NaV1.5, or PKP2.
(Reproduced from Rhett JM, Gourdie RG: The perinexus: a new feature of Cx43 gap junction organization. Heart Rhythm 9:619–623, 2012.)
Subcellular Localization of Ion Channels maintained via a separate “electric field mechanism.”56,57,59,109
and Cell-Cell Propagation: Gap Junction– This alternative postulates that the large INa in the proximal side
of an intercellular cleft generates a negative extracellular poten-
Independent Electrical Coupling and the tial within the cleft, which depolarizes the distal membrane and
Importance of the Intercellular Cleft activates its sodium channels. Thus, propagation can continue
downstream in the absence of gap junctions if there is a large INa
Several mathematical models of cardiac action potential propaga- at the intercalated disc and a narrow intercellular cleft separating
tion assume that gap junctions are the only path for transfer of the two opposing cells.
charge between cells. Accordingly, they predict that decreases in The model described here indicates the importance of the
junctional conductance bring about decreases in conduction intercellular space as a critical element of cell-cell propagation.
velocity. This notion contrasts sharply with actual data showing Indeed, if the width of the intercellular cleft increases, the model
that only extreme reductions in Cx43 abundance (and electrical predicts that propagation would fail. Interestingly, separation of
coupling) lead to significant changes in conduction velocity.51,88,108 the intercellular space is one of the features observed in murine
These results have given new impetus to the notion that, under models of desmosomal protein deficiency, even when overt struc-
poor gap junction–mediated coupling, propagation can be tural damage is not apparent.20,21 Figure 22-7 shows an example
A 200 nm B 200 nm
1.0 mm
C 200 nm D 200 nm
F
1.0 mm E 200 nm F 200 nm

Figure 22-7. Tomographic electron microscopy images of intercalated disc in a PKP2-Hz adult heart. Outlined regions in the left panels are expanded for insets a-f. Notice
the enlarged intercellular space that bulges into the cell, leaving large clefts between cells in some planes. Whether this increased gap between cells impairs electric field-
mediated propagation, remains to be defined.
(Reproduced from Cerrone M, Noorman M, Lin X, et al: Sodium current deficit and arrhythmogenesis in a murine model of plakophilin-2 haploinsufficiency. Cardiovasc Res
95:460–468, 2012.)
of tomographic electron microscopy of adult heart tissue obtained that intercalated disc molecules are not necessarily constrained
22
from a mouse heterozygous for PKP2. The image is reproduced to a single function (e.g., Cx43 is not only limited to making gap
from Cerrone et al.20 These images can be compared with those junctions). We propose that molecules of the intercalated disc
shown in Figure 22-1, obtained from a control animal. The multitask within a protein interacting network, working in
results in Figure 22-7 show an expanded intercellular space that concert toward one common function: the propagation of excit-
coincides with the presence of membrane invaginations in one atory current from one cell to the next. From this perspective,
side of the intercalated disc. Planes of the same section (see Cx43 is a molecule that is relevant to cell excitability (by modulat-
Figure 22-3, A–F) reveal that the invaginations extended several ing INa),17,18 sodium channels can support cell-cell electrical cou-
nanometers into the intracellular space; in some planes, the pling,56,57 and “adhesion molecules” are actually required for
invaginations seemed to “pinch off,” leaving a healed membrane proper function of electrical complexes20 and for propagation
continuum facing the intercellular cleft. This observation was across an intercellular cleft. In the balance, multitasking canpro-
confirmed in three separate samples analyzed by tomographic vide functional overlap, thus enhancing the safety factor for
electron microscopy (T-EM), and not found in controls. Clearly, propagation. Future research will discern the spatial relations
this membrane separation would represent a barrier for transfer between these molecules. The novel concept of the perinexus110
of charge mediated via an electrical field mechanism. The impor- can be extended to include the neighboring areas of mechanical
tance of the intercellular cleft has also been highlighted by the junctions, or clusters of NaV1.5 molecules. Just as a mitochon-
recent work of Veeraraghavan et al.58 and his collaborators, drion is not just a conglomerate of independent molecules, but
showing that changes in intercellular volume affect action poten- is an organelle sharing common functions, the intercalated disc
tial propagation in the heart. is a single functional unit composed of molecules that interact
with each other. The implication of these interactions to the
understanding of arrhythmia mechanisms and arrhythmia treat-
ment emerges as an exciting area of future investigation.
Conclusions
This chapter has presented evidence showing that molecules clas-
sically defined as belonging to the desmosome, the VGSC, and Acknowledgments
the gap junctions, interact with each other. We have also described
how, through these interactions, the function of one complex This work was supported by National Institutes of Health grants
(e.g., the VGSC) is altered by changes in the expression of mol- RO1-HL106632, PO1-HL087226, and RO1-GM67691 and the
ecules of a different group (e.g., Cx43). The evidence suggests Foundation Leducq Transatlantic Network.
12. Zhang Y, Wang H, Kovacs A, et al: Reduced density and conduction slowing in desmoglein-2
References expression of Cx43 attenuates ventricular remod- mutant mice prior to cardiomyopathic changes.
eling after myocardial infarction via impaired Cardiovasc Res 2012; July 31st epub ahead of
1. Sjostrand FS, Andersson E: Electron microscopy TGF-beta signaling. Am J Physiol Heart Circ print.
of the intercalated discs of cardiac muscle tissue. Physiol 298(2):H477–H487, 2010. 22. Corrado D, Buja G, Basso C, et al: What is the
Experientia 10(9):369–370, 1954. 13. Smyth JW, Hong TT, Gao D, et al: Limited Brugada syndrome? Cardiol Rev 7(4):191–195,
2. Muir AR: An electron microscope study of the forward trafficking of connexin 43 reduces cell- 1999.
embryology of the intercalated disc in the heart cell coupling in stressed human and mouse myo- 23. Corrado D, Basso C, Buja G,et al: Right bundle
of the rabbit. J Biophys Biochem Cytol 3(2):193– cardium. J Clin Invest 120(1):266–279, 2010. branch block, right precordial st-segment eleva-
202, 1957. 14. Remo BF, Qu J, Volpicelli FM, et al: Phosphatase- tion, and sudden death in young people. Circula-
3. Weidmann S: The electrical constants of resistant gap junctions inhibit pathological tion 103(5):710–717, 2001.
Purkinje fibres. J Physiol 118(3):348–360, remodeling and prevent arrhythmias. Circ Res 24. Delmar M, McKenna WJ: The cardiac desmo-
1952. 108(12):1459–1466, 2011. some and arrhythmogenic cardiomyopathies:
4. Sjostrand FS, Andersson-Cedergren E, Dewey 15. Chkourko HS, Guerrero-Serna G, Lin X, et al: from gene to disease. Circ Res 107(6):700–714,
MM: The ultrastructure of the intercalated discs Remodeling of mechanical junctions and of 2010.
of frog, mouse and guinea pig cardiac muscle. J microtubule-associated proteins accompany 25. Bass-Zubek AE, Godsel LM, Delmar M, et al:
Ultrastruct Res 1(3):271–287, 1958. cardiac connexin43 lateralization. Heart Rhythm Plakophilins: multifunctional scaffolds for adhe-
5. Delmar M: Connexin43 regulates sodium current; 9(7):1133–1140, 2012. sion and signaling. Curr Opin Cell Biol 21(5):708–
ankyrin-G modulates gap junctions: the interca- 16. Danik SB, Rosner G, Lader J, et al: Electrical 716, 2009.
lated disc exchanger. Cardiovasc Res 93(2):220– remodeling contributes to complex tachyarrhyth- 26. Gottardi CJ, Niessen CM, Gumbiner BM: The
222, 2012. mias in connexin43-deficient mouse hearts. adherens junction. In Beckerle MC, editor: Cell
6. Delmar M, Liang FX: Connexin43 and the regu- FASEB J 22(4):1204–1212, 2008. Adhesion, ed 1, Oxford, 2001, Oxford University
lation of intercalated disc function. Heart Rhythm 17. Jansen JA, Noorman M, Musa H, et al: Reduced Press, pp 259–287.
9(5):835–838, 2012. heterogeneous expression of Cx43 results in 27. Gates J, Peifer M: Can 1000 reviews be wrong?
7. Agullo-Pascual E, Delmar M: The Noncanonical decreased Nav1.5 expression and reduced sodium Actin, alpha-Catenin, and adherens junctions.
Functions of Cx43 in the Heart. J Membr Biol current that accounts for arrhythmia vulnerability Cell 123(5):769–772, 2005.
245(3):477–482, 2012. in conditional Cx43 knockout mice. Heart rhythm 28. Bannon LJ, Goldfinger LE, Jones JCR, et al:
8. Basso C, Bauce B, Corrado D, et al: Pathophysiol- 9(4):600–607, 2012. Desmosomes and hemidesmosomes. In
ogy of arrhythmogenic cardiomyopathy. Nat Rev 18. Desplantez T, McCain ML, Beauchamp P, et al: Beckerle MC, editor: Cell Adhesion, ed 1,
Cardiol 9(4):223–233, 2011. Connexin43 ablation in foetal atrial myocytes Oxford, 2001, Oxford University Press, pp 324–
9. Fontes MS, van Veen TA, de Bakker JM, et al: decreases electrical coupling, partner connexins, 368.
Functional consequences of abnormal Cx43 and sodium current. Cardiovasc Res 94(1):58–65, 29. Hatsell S, Cowin P: Deconstructing desmoplakin.
expression in the heart. Biochim Biophys Acta 2012. Nat Cell Biol 3(12):E270–272, 2001.
2011. 19. Wilde AA, Brugada R: Phenotypical manifesta- 30. North AJ, Bardsley WG, Hyam J, et al: Molecular
10. Qu J, Volpicelli FM, Garcia LI, et al: Gap junctions of mutations in the genes encoding subunits map of the desmosomal plaque. J Cell Sci 112( Pt
tion remodeling and spironolactone-dependent of the cardiac sodium channel. Circ Res 23):4325–4336, 1999.
reverse remodeling in the hypertrophied heart. 108(7):884–897, 2011. 31. Al-Amoudi A, Castano-Diez D, Devos DP, et al:
Circ Res 104(3):365–371, 2009. 20. Cerrone M, Noorman M, Lin X, et al: Sodium The three-dimensional molecular structure of the
11. Akar FG, Nass RD, Hahn S, et al: Dynamic current deficit and arrhythmogenesis in a desmosomal plaque. Proc Natl Acad Sci U S A
changes in conduction velocity and gap junction murine model of plakophilin-2 haploinsufficiency, 108(16):6480–6485, 2011.
properties during development of pacing-induced Cardiovasc Res 95:460–468, 2012. 32. Chen X, Bonne S, Hatzfeld M, et al: Protein
heart failure. Am J Physiol Heart Circ Physiol 21. Rizzo S, Lodder EM, Verkerk AO et al: Interca- binding and functional characterization of pla-
293(2):H1223–H1230, 2007. lated disc abnormalities, reduced Na+ current kophilin 2. Evidence for its diverse roles in
desmosomes and beta -catenin signaling. J Biol cardiomyopathy. Circ Res 109(12):1342–1353, 71. Petitprez S, Zmoos AF, Ogrodnik J, et al: SAP97
Chem 277(12):10512–10522, 2002. 2011. and dystrophin macromolecular complexes deter-
33. Gallicano GI, Kouklis P, Bauer C, et al: Desmo- 51. Danik SB, Liu F, Zhang J, et al: Modulation of mine two pools of cardiac sodium channels
plakin is required early in development for assem- cardiac gap junction expression and arrhythmic Nav1.5 in cardiomyocytes. Circ Res 108(3):294–
bly of desmosomes and cytoskeletal linkage. J Cell susceptibility. Circ Res 95(10):1035–1041, 304, 2011.
Biol 143(7):2009–2022, 1998. 2004. 72. Sato PY, Musa H, Coombs W, et al: Loss of
34. Grossmann KS, Grund C, Huelsken J, et al: 52. Yao JA, Gutstein DE, Liu F, et al: Cell coupling plakophilin-2 expression leads to decreased
Requirement of plakophilin 2 for heart morpho- between ventricular myocyte pairs from sodium current and slower conduction velocity in
genesis and cardiac junction formation. J Cell Biol connexin43-deficient murine hearts. Circ Res cultured cardiac myocytes. Circ Res 105(6):523–
167(1):149–160, 2004. 93(8):736–743, 2003. 526, 2009.
35. Ruiz P, Brinkmann V, Ledermann B, et al: Tar- 53. Kaplan SR, Gard JJ, Carvajal-Huerta L, et al: 73. Mays DJ, Foose JM, Philipson LH, et al: Local-
geted mutation of plakoglobin in mice reveals Structural and molecular pathology of the heart ization of the Kv1.5 K+ channel protein in
essential functions of desmosomes in the embry- in Carvajal syndrome. Cardiovasc Pathol explanted cardiac tissue. J Clin Invest 96(1):282–
onic heart. J Cell Biol 135(1):215–225, 1996. 13(1):26–32, 2004. 292. PMCID: 185199, 1995.
36. Sen-Chowdhry S, Syrris P, McKenna WJ: Genet- 54. Kaplan SR, Gard JJ, Protonotarios N, et al: 74. Murata M, Buckett PD, Zhou J, et al: SAP97
ics of right ventricular cardiomyopathy. J Cardio- Remodeling of myocyte gap junctions in arrhyth- interacts with Kv1.5 in heterologous expression
vasc Electrophysiol 16(8):927–935, 2005. mogenic right ventricular cardiomyopathy due to systems. Am J Physiol Heart Circ Physiol
37. Delmar M: Desmosome-ion channel interactions a deletion in plakoglobin (Naxos disease). Heart 281(6):H2575–H2584, 2001.
and their possible role in arrhythmogenic cardio- Rhythm 1(1):3–11, 2004. 75. Milstein ML, Musa H, Balbuena DP, et al:
myopathy. Pediatr Cardiol 2012. 55. Oxford EM, Everitt M, Coombs W, et al: Molec- Dynamic reciprocity of sodium and potassium
38. Franke WW, Borrmann CM, Grund C, et al: The ular composition of the intercalated disc in a channel expression in a macromolecular complex
area composita of adhering junctions connecting spontaneous canine animal model of arrhythmo- controls cardiac excitability and arrhythmia.
heart muscle cells of vertebrates. I. Molecular genic right ventricular dysplasia/cardiomyopathy. Proc Natl Acad Sci U S A 109(31):E2134–2143,
definition in intercalated disks of cardiomyocytes Heart Rhythm 4(9):1196–1205, 2007. 2012.
by immunoelectron microscopy of desmosomal 56. Mori Y, Fishman GI, Peskin CS: Ephaptic con- 76. Cheng L, Yung A, Covarrubias M, et al: Cortactin
proteins. Eur J Cell Biol 85(2):69–82, 2006. duction in a cardiac strand model with 3D elec- is required for N-cadherin regulation of Kv1.5
39. Borrmann CM, Grund C, Kuhn C, et al: The area trodiffusion. Proc Natl Acad Sci U S A channel function. J Biol Chem 286(23):20478–
composita of adhering junctions connecting heart 105(17):6463–6468, 2008. 20489. PMCID: 3121477, 2011.
muscle cells of vertebrates. II. Colocalizations of 57. Tsumoto K, Ashihara T, Haraguchi R, et al: Roles 77. Roepke TK, Kontogeorgis A, Ovanez C, et al:
desmosomal and fascia adhaerens molecules in the of subcellular Na+ channel distributions in the Targeted deletion of kcne2 impairs ventricular
intercalated disk. Eur J Cell Biol 85(6):469–485, mechanism of cardiac conduction. Biophys J repolarization via disruption of I(K,slow1) and
2006. 100(3):554–563, 2011. I(to,f). FASEB Journal 22(10):3648–3660.
40. Pieperhoff S, Franke WW: The area composita 58. Veeraraghavan R, Salama ME, Poelzing S: Inter- PMCID: 2537427, 2008.
of adhering junctions connecting heart muscle stitial volume modulates the conduction velocity- 78. Barry DM, Trimmer JS, Merlie JP, et al: Differ-
cells of vertebrates. VI. Different precursor struc- gap junction relationship. Am J Physiol Heart ential expression of voltage-gated K+ channel
tures in non-mammalian species. Eur J Cell Biol Circ Physiol 302(1):H278–H286, 2012. subunits in adult rat heart. Relation to functional
87(7):413–430, 2008. 59. Sperelakis N: An electric field mechanism for K+ channels? Circ Res 77(2):361–369, 1995.
41. Pieperhoff S, Schumacher H, Franke WW: The transmission of excitation between myocardial 79. Takeuchi S, Takagishi Y, Yasui K, et al: Voltage-
area composita of adhering junctions connecting cells. Circ Res 91(11):985–987, 2002. gated K(+)Channel, Kv4.2, localizes predomi-
heart muscle cells of vertebrates. V. The impor- 60. Kant S, Krull P, Eisner S, et al: Histological and nantly to the transverse-axial tubular system of
tance of plakophilin-2 demonstrated by small ultrastructural abnormalities in murine desmo- the rat myocyte. J Mol Cell Cardiol 32(7):1361–
interference RNA-mediated knockdown in cul- glein 2-mutant hearts. Cell Tissue Res 348(2):249– 1369, 2000.
tured rat cardiomyocytes. Eur J Cell Biol 259, 2012. 80. Melnyk P, Zhang L, Shrier A, et al: Differential
87(7):399–411, 2008. 61. Kostetskii I, Li J, Xiong Y, et al: Induced deletion distribution of Kir2.1 and Kir2.3 subunits in
42. Li J, Goossens S, van Hengel J, Gao E, et al: Loss of the N-cadherin gene in the heart leads to dis- canine atrium and ventricle. Am J Physiol Heart
of alphaT-catenin alters the hybrid adhering junc- solution of the intercalated disc structure. Circ Circ Physiol 283(3):H1123–H1133, 2002.
tions in the heart and leads to dilated cardiomy- Res 96(3):346–354, 2005. 81. Saffitz JE: Dependence of electrical coupling on
opathy and ventricular arrhythmia following 62. Cohen SA: Immunocytochemical localization of mechanical coupling in cardiac myocytes: insights
acute ischemia. J Cell Sci 125(Pt 4):1058–1067, rH1 sodium channel in adult rat heart atria and gained from cardiomyopathies caused by defects
2012. ventricle. Presence in terminal intercalated disks. in cell-cell connections. Annals of the New York
43. Clevers H: Wnt/beta-catenin signaling in devel- Circulation 94(12):3083–3086, 1996. Academy of Sciences 1047:336–344, 2005.
opment and disease. Cell 127(3):469–480, 63. Maier SK, Westenbroek RE, Schenkman KA, 82. Oxford EM, Musa H, Maass K, et al: Connexin43
2006. et al: An unexpected role for brain-type sodium remodeling caused by inhibition of plakophilin-2
44. Chen X, Shevtsov SP, Hsich E, et al: The beta- channels in coupling of cell surface depolarization expression in cardiac cells. Circ Res 101(7):703–
catenin/T-cell factor/lymphocyte enhancer factor to contraction in the heart. Proc Natl Acad Sci 711, 2007.
signaling pathway is required for normal and U S A 99(6):4073–4078, 2002. 83. Fidler LM, Wilson GJ, Liu F, et al: Abnormal
stress-induced cardiac hypertrophy. Mol Cell Biol 64. Motoike HK, Liu H, Glaaser IW, et al: The Na+ connexin43 in arrhythmogenic right ventricular
26(12):4462–4473, 2006. channel inactivation gate is a molecular complex: cardiomyopathy caused by plakophilin-2 muta-
45. Chen S, Guttridge DC, You Z, et al: Wnt-1 sig- a novel role of the COOH-terminal domain. J tions. J Cell Mol Med 13(10):4219–4228, 2009.
naling inhibits apoptosis by activating beta- Gen Physiol 123(2):155–165, 2004. 84. Li MW, Mruk DD, Lee WM, et al: Connexin 43
catenin/T cell factor-mediated transcription. J 65. Lin X, Liu N, Lu J, et al: Subcellular heterogene- and plakophilin-2 as a protein complex that regu-
Cell Biol 152(1):87–96, 2001. ity of sodium current properties in adult cardiac lates blood-testis barrier dynamics. Proc Natl
46. Longo KA, Kennell JA, Ochocinska MJ, et al: ventricular myocytes. Heart Rhythm 8(12):1923– Acad Sci U S A 106(25):10213–10218. PMCID:
Wnt signaling protects 3T3-L1 preadipocytes 1930, 2011. 2700929, 2009.
from apoptosis through induction of insulin-like 66. McNair WP, Sinagra G, Taylor MR, et al: SCN5A 85. Asimaki A, Tandri H, Huang H, et al: A new
growth factors. J Biol Chem 277(41):38239– mutations associate with arrhythmic dilated car- diagnostic test for arrhythmogenic right ventricu-
38244, 2002. diomyopathy and commonly localize to the lar cardiomyopathy. N Engl J Med 360(11):1075–
47. Ross SE, Hemati N, Longo KA, et al: Inhibition voltage-sensing mechanism. J Am Coll Cardiol 1084, 2009.
of adipogenesis by Wnt signaling. Science 57(21):2160–2168, 2011. 86. Asimaki A, Tandri H, Duffy ER, et al: Altered
289(5481):950–953, 2000. 67. Sato PY, Coombs W, Lin X, et al: Interactions desmosomal proteins in granulomatous myocar-
48. Garcia-Gras E, Lombardi R, Giocondo MJ, et al: between ankyrin-G, Plakophilin-2, and Con- ditis and potential pathogenic links to arrhythmo-
Suppression of canonical Wnt/beta-catenin sig- nexin43 at the cardiac intercalated disc. Circ Res genic right ventricular cardiomyopathy. Circ
naling by nuclear plakoglobin recapitulates phe- 109(2):193–201, 2011. Arrhythm Electrophysiol 4(5):743–752, 2011.
notype of arrhythmogenic right ventricular 68. Patino GA, Isom LL: Electrophysiol and beyond: 87. Eckardt D, Theis M, Degen J, et al: Functional
cardiomyopathy. J Clin Invest 116(7):2012–2021, multiple roles of Na+ channel beta subunits in role of connexin43 gap junction channels in adult
2006. development and disease. Neurosci Lett mouse heart assessed by inducible gene deletion.
49. Lombardi R, Dong J, Rodriguez G, et al: Genetic 486(2):53–59, 2010. J Mol Cell Cardiol 36(1):101–110, 2004.
fate mapping identifies second heart field pro- 69. Meadows LS, Isom LL: Sodium channels as mac- 88. Morley GE, Vaidya D, Jalife J: Characterization
genitor cells as a source of adipocytes in arrhyth- romolecular complexes: implications for inherited of conduction in the ventricles of normal and het-
mogenic right ventricular cardiomyopathy. Circ arrhythmia syndromes. Cardiovasc Res 67(3):448– erozygous Cx43 knockout mice using optical
Res 104(9):1076–1084, 2009. 458, 2005. mapping. J Cardiovasc Electrophysiol 11(3):375–
50. Lombardi R, da Graca Cabreira-Hansen M, Bell 70. Lopez-Santiago LF, Meadows LS, Ernst SJ, et al: 377, 2000.
A, et al: Nuclear plakoglobin is essential for dif- Sodium channel Scn1b null mice exhibit pro- 89. Thomas SP, Kucera JP, Bircher-Lehmann L, et al:
ferentiation of cardiac progenitor cells to adipo- longed QT and RR intervals. J Mol Cell Cardiol Impulse propagation in synthetic strands of neo-
cytes in arrhythmogenic right ventricular 43(5):636–647, 2007. natal cardiac myocytes with genetically reduced
levels of connexin43. Circ Res 92(11):1209–1216, 95. Malhotra JD, Thyagarajan V, Chen C, et al: 103. Brackenbury WJ, Isom LL: Na channel beta sub-
2003. Tyrosine-phosphorylated and nonphosphorylated units: overachievers of the ion channel family.
90. Deo M, Sato PY, Musa H, et al: Relative contribu-
tion of changes in sodium current versus intercel-
lular coupling on reentry initiation in
sodium channel beta1 subunits are differentially
localized in cardiac myocytes. J Biol Chem
279(39):40748–40754, 2004.
Front Pharmacol 2:53. PMCID: 3181431, 2011.
104. Li J, Patel VV, Kostetskii I, et al: Cardiac-specific
loss of N-cadherin leads to alteration in connexins
22
2-dimensional preparations of plakophilin-2-defi- 96. Rhett JM, Ongstad EL, Jourdan J, et al: Cx43 with conduction slowing and arrhythmogenesis.
cient cardiac cells. Heart Rhythm 8(11):1740– Associates with Na(v)1.5 in the Cardiomyocyte Circ Res 97(5):474–481, 2005.
1748, 2011. Perinexus. J Membr Biol 245(7):411–422, 2012. 105. Lim BK, Xiong D, Dorner A, et al: Coxsackievi-
91. van der Zwaag PA, Cox MG, van der Werf C, 97. Meyer RA, Laird DW, Revel JP, et al: Inhibition rus and adenovirus receptor (CAR) mediates
et al: Recurrent and founder mutations in of gap junction and adherens junction assembly atrioventricular-node function and connexin 45
the Netherlands: Plakophilin-2 p.Arg79X muta- by connexin and A-CAM antibodies. J Cell Biol localization in the murine heart. J Clin Invest
tion causing arrhythmogenic right ventricular 119(1):179–189, 1992. 118(8):2758–2770, 2008.
cardiomyopathy/dysplasia. Neth Heart J 18(12): 98. Hartshorne RP, Catterall WA: Purification of the 106. Lisewski U, Shi Y, Wrackmeyer U, et al: The
583–591, 2010. saxitoxin receptor of the sodium channel from rat tight junction protein CAR regulates cardiac con-
92. Joshi-Mukherjee R, Coombs W, Musa H, et al: brain. Proc Natl Acad Sci U S A 78(7):4620–4624, duction and cell-cell communication. J Exp Med
Characterization of the molecular phenotype of P, 1981. 205(10):2369–2379, 2008.
two arrhythmogenic right ventricular cardiomy- 99. Isom LL, De Jongh KS, Patton DE, et al: Primary 107. Rhett JM, Jourdan J, Gourdie RG: Connexin 43
opathy (ARVC)-related plakophilin-2 (PKP2) structure and functional expression of the beta 1 connexon to gap junction transition is regulated
mutations. Heart Rhythm 5(12):1715–1723, subunit of the rat brain sodium channel. Science by zonula occludens-1. Mol Biol Cell 22(9):1516–
2008. 256(5058):839–842, 1992. 1528, 2011.
93. Bauce B, Basso C, Rampazzo A, et al: Clinical 100. Malhotra JD, Kazen-Gillespie K, Hortsch M, et 108. Kirchhoff S, Kim JS, Hagendorff A, et al: Abnor-
profile of four families with arrhythmogenic right al: Sodium channel beta subunits mediate homo- mal cardiac conduction and morphogenesis in
ventricular cardiomyopathy caused by dominant philic cell adhesion and recruit ankyrin to points connexin40 and connexin43 double-deficient
desmoplakin mutations. Eur Heart J 26(16):1666– of cell-cell contact. J Biol Chem 275(15):11383– mice. Circ Res 87(5):399–405, 2000.
1675, 2005. 11388, 2000. 109. Kucera JP, Rohr S, Rudy Y: Localization of
94. Gomes J, Finlay M, Ahmed AK, et al: 101. Dhar Malhotra J, Chen C, Rivolta I, et al: Char- sodium channels in intercalated disks modulates
Electrophysiological abnormalities precede acterization of sodium channel alpha- and beta- cardiac conduction. Circ Res 91(12):1176–1182,
overt structural changes in arrhythmogenic right subunits in rat and mouse cardiac myocytes. 2002.
ventricular cardiomyopathy due to mutations in Circulation 103(9):1303–1310, 2001. 110. Rhett JM, Gourdie RG: The perinexus: a new
desmoplakin-A combined murine and human 102. Isom LL: Sodium channel beta subunits: anything feature of Cx43 gap junction organization. Heart
study. Eur Heart J 2012. but auxiliary. Neuroscientist 7(1):42–54, 2001. Rhythm 9(4):619–623, 2012.
Scaffolding Proteins and Ion
Channel Diseases 23
Ravi Vaidyanathan and Jonathan C. Makielski
among others; (2) the ankyrin family of proteins, specifically

CHAPTER OUTLINE
ankyrin-2 or ankyrin-B, which is implicated in LQTS-4 and
Scaffolding Proteins in the Heart 229 ankyrin syndrome; (3) the caveolin family of proteins, specifically
caveolin-3, which is implicated in LQTS-9 and sudden infant
Scaffolding Proteins Involved in Long QT Syndrome 229
death syndrome (SIDS); (4) the A-kinase anchoring protein
Summary 233 (AKAP) family specifically AKAP9/yotiao which is implicated in
LQTS-11; and (5) the syntrophin family of proteins, specifically
SNTA1 or α-1-syntrophin, which is implicated in LQTS-12.
Scaffolding proteins are modular proteins that assemble multi- In this chapter, our focus is specially limited to the scaffolding
molecular signaling complexes or macromolecular signaling proteins that effect ion channel function and are involved in
complexes and that also modulate and regulate the function of LQTS, which have been identified as ankyrin-B, caveolin-3,
associated proteins. The modular nature of their structure gives AKAP9, and α-1-syntrophin. The structure, associations, and
them the ability to associate with a multitude of signaling pro- mechanism of action of each of these scaffolding proteins are
teins via protein-protein interaction domains. In the heart, scaf- discussed in the following sections.
folding proteins have also been shown to interact with ion
channels and to modulate their function. Compelling evidence in
the literature shows that scaffolding proteins associate with ion
channels responsible for depolarization (sodium channels and
Scaffolding Proteins Involved
both L-type and T-type calcium channels), ion channels respon- in Long QT Syndrome
sible for repolarization potassium channels (time-independent,
time-dependent, and voltage-gated delayed rectifiers), and gap Ankyrin-B (LQT4)
junctions (connexin40 and connexin43) that connect cells electri-
cally. The pore-forming protein of an ion channel is generally Structure and Associations
called the α-subunit, and other proteins thought to be associated Ankyrins are a family of scaffolding proteins that link the
with specific α-subunits are β-, δ-, or γ-subunits, or have other membrane-bound proteins to the underlying cytoskeleton. This
names such as minK. Although these subunits can assist in function of ankyrins contributes to the expression, specific local-
forming the complex and attaching the complex to the cytoskel- ization, and overall stability of membrane proteins within the
eton and extracellular matrix and regulate channel surface expres- plasma membrane.1,2 The ankyrin family of proteins consists of
sion or regulation, and although their associations may be less three family members, ankyrin-R, ankyrin-B, and ankyrin-G
than specific in some instances, we have not included these sub- encoded by genes ANK1, ANK2, and ANK3, respectively.
units as a part of this review of scaffolding proteins that have Ankyrin-B and ankyrin-G have been detected in a variety of
more general interactions. The importance of scaffolding pro- tissue. In contrast, ankyrin-R is detected only in erythrocytes,
teins and their role in the regulation of associated ion channels striated muscle, and some neurons.3 In general, an ankyrin
in physiology and pathophysiology has been highlighted by the protein has four functional domains, a membrane-binding
discovery of mutations on scaffolding proteins that alter ion domain (MBD), spectrin-binding domain (SBD), death domain,
channel function and regulation. The focus of this chapter is and a C-terminal regulatory domain (Figure 23-2).
specifically on the scaffolding proteins, mutants of which have The N-terminal MBD contains 24 tandem ankyrin repeats;
been shown to induce long QT syndrome (LQTS). As an these ankyrin repeats have inherent spring-link qualities, and it
example, a schematic drawing of a macromolecular complex for has been suggested that this elastic property of the ankyrin repeat
the cardiac Na channel is depicted in Figure 23-1. could have a role in mechanotransduction and confer pliancy to
the MBD from mechanical perturbations.4 In addition, ankyrin
repeats are also a common motif for protein-protein interactions
found in many membrane proteins, such as sodium-calcium
Scaffolding Proteins in the Heart exchanger (NCX), sodium-potassium ATPase (NKA), ATP-
sensitive potassium channels (Kir6), inositol 1,4,5 trisphosphate
The functional roles of scaffolding proteins were first identified receptor (IP3R), and anion exchangers (Table 23-1). The SBD
and elucidated in the nervous system and were subsequently has been shown to bind to spectrin via its ZU5 domain (i.e.,
extended into the cardiac field. In the heart, the functional role domains that are found in zona-occludens-1 [ZO-1] and the
of scaffolding proteins is an emerging and exciting area of uncoordinated protein 5) that composes the minimal binding
research. Scaffolding proteins currently identified in the heart domain for spectrin. In addition to binding to spectrin, SBD
to be associated with ion channels include (1) the membrane- associates with the regulatory subunit of protein phosphatase 2A
associated guanylate kinase family of proteins, which includes (B56a), and dynactin4 an adaptor protein that links the dynein
synapse-associated proteins-97 (SAP97), postsynaptic density-95 motor to membrane cargo.5,6 The death domain of ankyrin-G has
(PSD95), calcium-calmodulin–dependent serine protein kinase been shown to interact with proteins involved in apoptosis,7 but
229
Syntrophin
SAP97
Cholesterol
nNOS
PKA
Phosphatase
Caveolin 3
dimer
Nav1.5 β subunit PMCA-4B Calmodulin

channel
Figure 23-1. A macromolecular complex representing channel-associated proteins, in this example the NaV1.5 sodium channel in a caveolar lipid membrane. The drawings
are roughly representational of relative size, shape, and membrane associations. Protein-protein interactions are represented based on the literature, but no implication for
where or how they are attached is intended.
Ankyrin-B
Membrane binding domain Spectrin binding domain Death domain C-terminal domain
1 Ankyrin repeat domain ZU5 domain Death domain 1872

aa 63–822 aa 966–1103 aa 1482–1565
Caveolin 3
N-terminal domain Scaffolding domain Membrane domain C-terminal domain

1 54 55 73 74 106 107 151
Alpha-1-syntrophin
PH1a PDZ domain PH1b PH2 Syntrophin unique domain

1 85 86 167 168 296 297 399 400 505
A-kinase anchoring protein-9
A-kinase anchoring protein 9
1 KCNQ1 Adenylyl cyclase PKA binding Protein KCNQ1 PDE4D3 3907

binding site-1 binding domain domain phosphatase 1 binding site-2 aa 1710–2872
aa 29–46 aa 808–957 aa 1140–1457 binding domain aa 1574–1643
aa 1171–1229
Figure 23-2. Schematic of scaffolding proteins (1) ankyrin-B, (2) caveolin-3, (3) a kinase anchoring protein-9 and (4) α1-syntrophin involved in LQTS 4, 9, 11, and 12,
respectively.
Scaffolding Proteins and Ion Channel Diseases 231
classes of variants included bradycardia, sinus arrhythmia, delayed

Table 23-1. Associations of Ankyrin B
23
conduction–conduction block, idiopathic ventricular fibrillation,
Domain on and catecholaminergic polymorphic ventricular tachycardia.
Protein Ankyrin References The severity of the loss of function of ankyrin variants has
been directly linked to the inability of ankyrin to target NCX and
Ion channels and exchangers NKA to the membrane.11 In a recent study, Camors et al.13, using
an ankyrin-B heterozygous knockout (ankyrin-B+/–) mouse, dem-
IP3R MBD Mohler et al.,47 onstrated that myocytes from ankyrin-B+/– mice had approxi-
Mohler et al.48 mately 20% reduced NCX and NKA protein expression with no
Cav1.3 ? Cunha et al.,49 Le significant change in the mRNA levels.13 They also reported that
Scouarnec et al.50 the calcium transients, SR calcium content, and fractional SR
calcium release were significantly larger in the ankyrin-B+/– myo-
Kir6.2 MBD Li et al.51
cytes compared with wild type myocytes. Based on their experi-
NCX MBD Mohler et al.8 mental results, they concluded that the reason for the higher
NKA MBD Mohler et al.8 frequency of spontaneous diastolic calcium sparks in ankyrin-B+/–
myocytes was caused by an alteration of the transport of sodium
Cytoskeletal/structure and calcium and enhancement of the coupled opening of ryano-
dine receptors. The results presented in the literature and dis-
β-Spectrin SBD Bhasin et al.6
cussed here suggest that additional work is necessary to elucidate
Obscurin CTD Cunha et al.52 the molecular mechanism underlying phenotypic variability of
Dystrophin ? Ayalon et al.1 each human ankyrin-B variant.
Nav channel β-subunit MBD Malhotra et al.53
Intracellular transport Caveolin 3 (LQT9)

Tubulin MBD Ayalon et al. 1
Structure and Associations
EHD1-4 MBD Gudmundsson et al. 54 Caveolin is a protein that is essential to the formation of cavelike
membrane structures called caveolae. Typically, caveolae are
Dynactin-4 SBD Ayalon et al.5 surface plasma-membrane invaginations or flask-shaped struc-
MBD, Membrane-binding domain; SBD, spectrin-binding domain; CTD, tures with a diameter of 50 to 100 nm. The caveolin family of
C-terminal domain. proteins is encoded by three genes and consists of six known
caveolin subtypes: caveolin-1α, caveolin-1β, caveolin-2α,
caveolin-2β, caveolin-2γ, and caveolin-3. Of the six known caveo-
lins, caveolin-3 is specifically expressed in muscle tissue, includ-
the exact function of the death domain of ankyrin-B is currently ing the heart. Most of the information regarding the molecular
unknown. The C-terminal regulatory domain, region of ankyrin structure of caveolins comes from the early studies on caveolin-1.
proteins has the least homology, and this region confers different Given that caveolin-1 and caveolin-3 are approximately 65%
ankyrin family members their functional specificity and distinct identical and 85% similar,14 it is necessary to discuss the structure
subcellular localization. In cardiomyocytes, Mohler et al.8 identi- of caveolin-1 and point out key differences between caveolin-1
fied an ankyrin-B based macromolecular complex comprising and -3. The molecular structure of caveolin-3 is composed of 151
Na/K-ATPase (alpha 1 and alpha 2 isoforms), Na/Ca exchanger amino acids divided into four domains: the N-terminal domain
1, and InsP3 receptor that were specifically localized in cardio- (residues 1 to 54), the scaffolding domain (residues 55 to 73), the
myocyte T-tubules in discrete microdomains distinct from classic membrane domain (residues 74 to 106), and the C-terminal
dihydropyridine receptor/ryanodine receptor dyads. domain (residues 107 to 151; see Figure 23-2). The N-terminal
domain of caveolin-1 is longer than that of caveolin-3 by 27
Mechanism of Action amino acid residues. The scaffolding domain, which is next to the
A mutation in ankyrin-B that caused LQTS was first described N-terminal domain, is essential to the formation of caveolin
in 1995 by Schott et al.9 and was further defined in 2003 by oligomers and for associating with other proteins (Table 23-2). A
Mohler et al.,10 who later in 2007 further identified ankyrin-B W-W–like domain, a protein-protein interaction domain, over-
loss-of-function variants.11,12 These mutations caused a disrup- laps the scaffolding domain and the membrane domain and is
tion in the localization of the sodium pump, the sodium–calcium followed by a highly conserved proline residue.15 The W-W–like
exchanger, and inositol-1,4,5-trisphosphate receptors, which domain and the proline residues are conserved between caveolin-1
reduces the total protein level and targeting to T-tubules. In adult and -3, but absent in caveolin-2.15 The membrane domain of
cardiac myocytes, ankyrin-B mutations altered Ca2+ signaling, caveolin-1 and -3 consists of hydrophobic amino acids, and it has
caused extrasystoles, and provided a rationale for QT prolonga- been suggested that this high density of hydrophobic residues
tion and associated arrhythmia. The variants of ankyrin-B were aids in forming a hairpinlike loop into the membrane. It is sug-
categorized into three distinct functional classes based on severity gested that two different domains, the membrane domain and the
of clinical phenotype in associated patients and degree of ankyrin C-terminal domain, could aid in membrane attachment of caveo-
dysfunction in cardiomyocytes. The variants (G1406C, R1450W, lin. The C-terminal domain of both caveolin-1 and caveolin-3
and L1503V) associated with less severe clinical phenotype (or are palmitoylated on three cysteine residues.16 These palmitate
asymptomatic individuals) and a lesser degree of ankyrin dysfunc- chains aid in anchoring the C-terminal domain to the membrane,
tion were put into class 1.11,12 The variants (T1404I, T1552N and and it has been suggested that they could increase the stability of
V1777M) associated with more severe clinical phenotype and the caveolar structure.16 The palmitate chains have also been
severe degree of dysfunction in cardiomyocyte were put into class suggested to aid in binding and transport of cholesterol, another
2. The variants (E1425G, V1516D, and R1788W) with the most component key in caveolar structures.17 Caveolin-3 with its
severe clinical phenotypes and loss of function in cardiomyocyte modular structure can associate with ion channels and other
were put into class 3.9-11 The phenotypes associated with these signaling molecules listed in Table 23-2. In cardiac myocytes,
Table 23-2. Associations of Caveolin3 in Cardiac Myocytes Table 23-3. Associations of AKAP9
Domain/ Proteins Residues on AKAP9 References

Protein residues of Cav3 References
Ion channels
Ion channel
KCNQ1 (IKs) 29-46 and 1574-1643 Chen et al.31
55
Pacemaker channels ? Ye et al.
Signaling molecules
(HCN4)
L-Type calcium ? Balijepalli et al.18 PKA 1140-1457 Tasken et al.27
channels (Cav1.2) Adenylyl cyclase 808-957 Piggott et al.26
56
T-Type calcium NT Markendaya et al. Phosphodiesterase 1710-2872 Tasken et al.29
channels (Cav3.2) (PDE4D3)
Kv1.5 ? Folco et al.57 Protein phosphatase 1 1171-1229 Redden et al.28
58
Kir6.2 ? Garg et al.
Kir2.1 ? Vaidyanathan et al.21
Nav1.5 ? Yarbrough et al.59 increase in L-type current density by S141R–caveolin-3 com-
+ + 60 pared with wild type caveolin-3. Therefore, caveolin-3 mutations
Na /Ca exchanger ? Bossuyt et al.
could prolong QT by affecting at least three ion currents.
(NCX1)
Signaling molecules
AKAP9 (LQTS11)
nNos aa 69-84 and Venema et al.61
aa104-130 Structure and Associations
G-proteins Scaffolding Head et al.62 Signal transduction events at specific sites within the cell are
domain primarily due to subcompartmentalization of protein kinases and
phosphatases through their association with scaffolding proteins
Protein kinase C ? Markendaya et al.56
called AKAPs. AKAPs are a family of more than 50 anchoring
β2-adrenergic receptor ? Balijepalli et al.18 proteins. This family of anchoring proteins are structurally
diverse, but have a common ability to bind and target PKA.23
Cytoskeletal, structure, scaffolding proteins
There are four known AKAPs in the heart: AKAP5, AKAP6,
β-dystroglycan WW domain Sotgia et al.15 AKAP8, and AKAP9.24 AKAPs also have unique protein-lipid or
protein-protein interaction domains that aid in specific subcel-
SAP97 ? Folco et al.57
lular localization of PKA-AKAP complex.23-25 Using the protein-
protein interaction domain AKAPs can associate with signaling
molecules as detailed in Table 23-3. The AKAP of interest in this
chapter is AKAP9, also known as AKAP350 / AKAP450 / yotiao
caveolin-3 has been shown specifically to localize to the interca- / centrosomal and Golgi N-kinase anchoring protein (see Figure
lated disc and the lateral membrane and to show a striated pattern 23-2). AKAP9 specially associates with adenylyl cyclase,26 PKA,27
reminiscent of T-tubular localization.18 protein phosphatase 1 (PP1),28 and phosphodiesterase, specifi-
cally PDE4D3.29 It has been suggested that the presence of sig-
Mechanism of Action naling molecules such as kinases, phosphatases, and diesterase in
Mutations in caveolin-3 cause multiple phenotypes, including close proximity could have a critical role in tight regulation of
limb girdle muscular dystrophy, rippling muscle disease, distal posttranslational modification of associated proteins and even
myopathy, hyperCKemia, LQTS-9, and SIDS. The LQTS-9– AKAP, which can also be a substrate for posttranslational
causing mutations include T78M, A85T, F97C, and S141R. The modification.24
SIDS-causing mutations include V14L, T78M, and L79R.19,20
Mutations of caveolin-3 that cause LQTS and SIDS induced a Mechanism of Action
gain-of-function increase in the late sodium current,19,20 which is Marx et al.30 have reported previously that yotiao/AKAP9 associ-
a plausible QT-prolonging mechanism. The molecular mecha- ates with the leucine zipper domain of KCNQ1 using its N- and
nism for the increase in the late sodium current could be through C-terminal domains.30 The mutation S1570L on yotiao is in close
a loss of suppression of nitrosylation, similar to the mechanism proximity to the C-terminal binding domain of yotiao. In addi-
proposed for LQT12. In addition, given the plethora of ion chan- tion, it has been illustrated that S1770L-yotiao alters but does
nels and signaling molecules (see Table 23-2) with which not eliminate the association between yotiao and KCNQ1.31
caveolin-3 associates, it would not be surprising that LQTS-9 They have reported that the alterations in association between
mutants affect other ion channels. Recent studies have implicated KCNQ1 and yotiao reduced cAMP/PKA-induced phosphoryla-
other ion channels such as Kir2.1 and CaV1.2.21,22 Recently, Vaid- tion of KCNQ1, and this affects regulation of KCNQ1. The
yanathan et al.21 have demonstrated that F97C–caveolin-3 reduction in phosphorylation by PKA eliminates the functional
decreased the current density of Kir2.1 channels compared with response of KCNQ1 channels to cAMP/PKA stimulation during
wild type caveolin-3 and that this loss of function was caused by sympathetic nervous system activation. Under normal physiolog-
a decrease in the cell surface expression of Kir2.1 channels in a ical conditions during sympathetic nervous system activation
heterologous expression system. In addition, Foell et al.22 pre- cAMP/PKA has two important roles: increasing L-type calcium
sented compelling evidence suggesting F97C–caveolin-3 and current and increasing IKs and IKr currents. This increase in
S141R–caveolin-3, induce a gain of function effect on L-type inward and outward currents helps to keep the action potential
calcium channel perhaps by slowing of the calcium-dependent duration smaller and aids in sustaining higher heart rates. Chen
inactivation for F97C–caveolin-3 and S141R–caveolin-3 and an et al.31 illustrated that S1570L-yotiao reduces this functional
Scaffolding Proteins and Ion Channel Diseases 233
response of KCNQ1/IKs and hence aids in the prolongation of

Table 23-4. Associations of α1-Syntropin in Cardiac Myocytes
23
action potential duration to cause LQTS.31
Protein Domain of SNTA1 References
α1-Syntrophin (LQTS12) Ion channel and pumps
Structure and Associations Nav1.5 PDZ Gavillet et al.34

Syntrophins are a family of homologous protein members include Kir2.1 ? Leonoudakis et al.63
α1, β1, β2, γ1, and γ2 and show approximately 50% homology
Aquaporin (AQP4) PDZ Neely et al.,35 Adams
between the different syntrophins. α1-Syntrophin also belongs
et al.36
to the family of dystrophin-associated protein complexes. α1-
Syntrophin is primarily expressed in striated muscle and is the Sarcolemmal Linker region Williams et al.37
focus of this section. The β-syntrophin has been shown to express calcium pump between PH2 and
ubiquitously in mammalian tissue. (PMCA4B) PDZ (aa 399-447)
All syntrophins share a similar modular domain structure with
Signaling molecules
two pleckstrin homology (PH1 and PH2) domains (domain first
identified on pleckstrin), a PDZ domain (domain found in nNos PDZ Brennan et al.39
PSD95, disc-large, and ZO-1) and the syntrophin unique domain Hillier et al.38
(see Figure 23-2). In α1-syntrophin there are two tandem PH
G proteins PDZ Chen et al.64
domains that occupy the N-terminal part of the protein. Some
Okamura et al.65
of the key properties associated with PH domains include asso-
ciating with lipid bilayers containing inositol-4,5-bisphosphate; Kinases PDZ Hasegawa et al.66
they also interact with the βγ-subunit of G-proteins and protein Phosphatidylinositol PH1 Chockalingam et al.67
kinases (Table 23-4). In α1-syntrophin, the PH1 domain is split 4,5-bisphosphate Yan et al.32
into PH1a and PH1b by a PDZ domain. The presence of the
PDZ domain does not seem to affect the function of the split Calmodulin PH1a and PDZ Iwata et al.33
PH1 domain and the PDZ domain.32 The two split domains can Structural proteins
bind to each other to form an intact PH domain, which can then
F-actin PH2, SU Iwata et al.33
interact with inositol-phospholipids.32 On the other hand, PH2,
the other PH domain, has been shown to bind to the structural Dystrophin SU Kramarcy et al.40
protein, F-actin.33 The PDZ domain has been shown to associate Utrophin ? Kramarcy et al.40
with the ion channels, such as the cardiac sodium channel
SCN5A,34 Aquaporin AQP4,35, 36 and the sarcolemmal calcium β1-Dystroglycan SU Suzuki et al.41
pump PMCA4B.37 The PDZ domain has also been shown to
associate with neuronal nitric oxide synthase (nNos), a scaffold-
ing protein via PDZ-PDZ interaction.38,39 The C-termini of all
syntrophins have an approximately 50-aa stretch that is unique currents. Both the nitrosylation and the increased current were
to all syntrophins and are therefore called the syntrophin unique prevented by nNOS inhibitors.43 α1-Syntrophin also associates
domain. It has been shown that syntrophin associates via this with other ion channels, but it is not known whether the
domain with the dystrophin family of proteins, including dystro- LQTS12-causing syntrophin mutation affects the function or
brevin and β-dystroglycan.40,41 α1-Syntrophin in cardiac myo- regulation of other proteins or ion channels.
cytes tends to localize to the lateral membrane of myocytes.42
Mechanism of Action
A mutation A390V in α1-syntrophin was reported in a patient Summary
with syncope, a prolonged QT interval greater than 600 ms, and
designated LQT12.43 The mechanism was postulated to be an This chapter summarizes the role of some scaffolding proteins in
increased late sodium current associated with increased channel the physiology and pathophysiology of the heart using the
nitrosylation involving a complex consisting of α1-syntrophin, LQTSs as examples. It is not an exhaustive review of all scaffold-
SCN5A, nNos, and PMCA4B.43 A390V was shown to selectively ing proteins nor all inherited arrhythmia syndromes. More work
disrupt the association of PMCA4B with the complex. PMCA4B is warranted in this area of research to address questions such as,
was a known inhibitor of nNOS activity,44,45 and the absence of are there other scaffolding proteins in the heart that contribute
PMCA4B in the complex would favor increased nitrosylation of to arrhythmia? If so, what are their associations, functions, and
SCN5A by nNOS, which was also known to increase the late mechanisms of action? How are the molecular complexes assem-
sodium current.43,46 Consistent with this idea, the mutant was bled and trafficked to the membrane? These questions certainly
shown to increase nitrosylation of SCN5A and late sodium will be the focus of additional investigations in the near future.
exchanger in the inner segment of rod photorecep- are required for dystrophin-based protection of
References tors. Exp Eye Res 88:57–64, 2009. skeletal muscle from exercise injury. J Biol Chem
3. Bennett V, Chen L: Ankyrins and cellular targeting 286:7370–7378, 2011.
1. Ayalon G, Davis JQ, Scotland PB, et al: An of diverse membrane proteins to physiological 6. Bhasin N, Cunha SR, Mudannayake M,
ankyrin-based mechanism for functional organiza- sites. Curr Opin Cell Biol 13:61–67, 2001. et al: Molecular basis for pp2a regulatory
tion of dystrophin and dystroglycan. Cell 135: 4. Bennett V, Baines AJ: Spectrin and ankyrin-based subunit b56alpha targeting in cardiomyocytes. Am
1189–1200, 2008. pathways: Metazoan inventions for integrating J Physiol Heart Circ Physiol 293:H109–H119,
2. Kizhatil K, Sandhu NK, Peachey NS, et al: cells into tissues. Physiol Rev 81:1353–1392, 2001. 2007.
Ankyrin-b is required for coordinated expression 5. Ayalon G, Hostettler JD, Hoffman J, et al: 7. Del Rio M, Imam A, DeLeon M, et al: The death
of beta-2-spectrin, the Na/K-atpase and the Na/Ca Ankyrin-b interactions with spectrin and dynactin-4 domain of kidney ankyrin interacts with fas and
promotes fas-mediated cell death in renal epithelia. 28. Redden JM, Dodge-Kafka KL: Akap phosphatase 48. Mohler PJ, Davis JQ, Davis LH, et al: Inositol
J Am Soc Nephrol 15:41–51, 2004. complexes in the heart. J Cardiovasc Pharmacol 1,4,5-trisphosphate receptor localization and sta-
8. Mohler PJ, Davis JQ, Bennett V: Ankyrin-B coor- 58:354–362, 2011. bility in neonatal cardiomyocytes requires interac-
dinates the Na/K atpase, Na/Ca exchanger, and 29. Tasken KA, Collas P, Kemmner WA, et al: Phos- tion with ankyrin-B. J Biol Chem 279:12980–12987,
insp3 receptor in a cardiac T-tubule/sr microdo- phodiesterase 4d and protein kinase A type ii con- 2004.
main. PLoS Biol 3:e423, 2005. stitute a signaling unit in the centrosomal area. 49. Cunha SR, Hund TJ, Hashemi S, et al: Defects in
9. Schott JJ, Charpentier F, Peltier S, et al: Mapping J Biol Chem 276:21999–22002, 2001. ankyrin-based membrane protein targeting path-
of a gene for long qt syndrome to chromosome 30. Marx SO, Kurokawa J, Reiken S, et al: Require- ways underlie atrial fibrillation. Circulation
4q25-27. Am J Hum Genet 57:1114–1122, 1995. ment of a macromolecular signaling complex for 124:1212–1222, 2011.
10. Mohler PJ, Schott JJ, Gramolini AO, et al: beta adrenergic receptor modulation of the 50. Le Scouarnec S, Bhasin N, Vieyres C, et al: Dys-
Ankyrin-b mutation causes type 4 long-qt cardiac KCNQ1-KCNE1 potassium channel. Science function in ankyrin-B-dependent ion channel and
arrhythmia and sudden cardiac death. Nature 295:496–499, 2002. transporter targeting causes human sinus node
421:634–639, 2003. 31. Chen L, Marquardt ML, Tester DJ, et al: Mutation disease. Proc Natl Acad Sci U S A 105:15617–
11. Mohler PJ, Le Scouarnec S, Denjoy I, et al: Defin- of an A-kinase-anchoring protein causes long-qt 15622, 2008.
ing the cellular phenotype of “Ankyrin-B syn- syndrome. Proc Natl Acad Sci U S A 104:20990– 51. Li J, Kline CF, Hund TJ, et al: Ankyrin-b regulates
drome” Variants: Human ANK2 variants associated 20995, 2007. kir6.2 membrane expression and function in heart.
with clinical phenotypes display a spectrum of 32. Yan J, Wen W, Xu W, et al: Structure of the split J Biol Chem 285:28723–28730, 2010.
activities in cardiomyocytes. Circulation 115:432– ph domain and distinct lipid-binding properties of 52. Cunha SR, Mohler PJ: Obscurin targets ankyrin-b
441, 2007. the ph-pdz supramodule of alpha-syntrophin. and protein phosphatase 2A to the cardiac M-line.
12. Mohler PJ, Splawski I, Napolitano C, et al: A EMBO J 24:3985–3995, 2005. J Biol Chem 283:31968–31980, 2008.
cardiac arrhythmia syndrome caused by loss of 33. Iwata Y, Pan Y, Yoshida T, et al: Alpha1-syntrophin 53. Malhotra JD, Koopmann MC, Kazen-Gillespie
ankyrin-b function. Proc Natl Acad Sci U S A has distinct binding sites for actin and calmodulin. KA, et al: Structural requirements for interaction
101:9137–9142, 2004. FEBS Lett 423:173–177, 1998. of sodium channel beta 1 subunits with ankyrin.
13. Camors E, Mohler PJ, Bers DM, et al: Ankyrin-b 34. Gavillet B, Rougier JS, Domenighetti AA, et al: J Biol Chem 277:26681–26688, 2002.
reduction enhances Ca spark-mediated Sr Ca Cardiac sodium channel NaV1.5 is regulated by a 54. Gudmundsson H, Hund TJ, Wright PJ, et al: Eh
release promoting cardiac myocyte arrhythmic multiprotein complex composed of syntrophins domain proteins regulate cardiac membrane
activity. J Mol Cell Cardiol 52:1240–1248, 2012. and dystrophin. Circ Res 99:407–414, 2006. protein targeting. Circ Res 107:84–95, 2010.
14. Krajewska WM, Maslowska I: Caveolins: Structure 35. Neely JD, Amiry-Moghaddam M, Ottersen OP, 55. Ye B, Balijepalli RC, Foell JD, et al: Caveolin-3
and function in signal transduction. Cell Mol Biol et al: Syntrophin-dependent expression and local- associates with and affects the function of
Lett 9:195–220, 2004. ization of aquaporin-4 water channel protein. Proc hyperpolarization-activated cyclic nucleotide-
15. Sotgia F, Lee JK, Das K, et al: Caveolin-3 directly Natl Acad Sci U S A 98:14108–14113, 2001. gated channel 4. Biochemistry 47:12312–12318,
interacts with the c-terminal tail of beta-dystrogly- 36. Adams ME, Mueller HA, Froehner SC: In vivo 2008.
can. Identification of a central ww-like domain requirement of the alpha-syntrophin pdz domain 56. Markandeya YS, Fahey JM, Pluteanu F, et al:
within caveolin family members. J Biol Chem for the sarcolemmal localization of nnos and Caveolin-3 regulates protein kinase a modulation
275:38048–38058, 2000. aquaporin-4. J Cell Biol 155:113–122, 2001. of the Ca(V)3.2 (alpha1h) T-type Ca2+ channels.
16. Dietzen DJ, Hastings WR, Lublin DM: Caveolin 37. Williams JC, Armesilla AL, Mohamed TM, et al: J Biol Chem 286:2433–2444, 2011.
is palmitoylated on multiple cysteine residues. Pal- The sarcolemmal calcium pump, alpha-1 syntro- 57. Folco EJ, Liu GX, Koren G: Caveolin-3 and sap97
mitoylation is not necessary for localization of phin, and neuronal nitric-oxide synthase are parts form a scaffolding protein complex that regulates
caveolin to caveolae. J Biol Chem 270:6838–6842, of a macromolecular protein complex. J Biol Chem the voltage-gated potassium channel Kv1.5. Am J
1995. 281:23341–23348, 2006. Physiol Heart Circ Physiol 287:H681–H690,
17. Uittenbogaard A, Smart EJ: Palmitoylation of 38. Hillier BJ, Christopherson KS, Prehoda KE, et al: 2004.
caveolin-1 is required for cholesterol binding, Unexpected modes of PDZ domain scaffolding 58. Garg V, Jiao J, Hu K: Regulation of atp-sensitive
chaperone complex formation, and rapid transport revealed by structure of nnos-syntrophin complex. k+ channels by caveolin-enriched microdomains in
of cholesterol to caveolae. J Biol Chem 275:25595– Science 284:812–815, 1999. cardiac myocytes. Cardiovasc Res 82:51–58,
25599, 2000. 39. Brenman JE, Chao DS, Gee SH, et al: Interaction 2009.
18. Balijepalli RC, Foell JD, Hall DD, et al: Localiza- of nitric oxide synthase with the postsynaptic 59. Yarbrough TL, Lu T, Lee HC, et al: Localization
tion of cardiac l-type Ca(2+) channels to a caveolar density protein PSD-95 and alpha1-syntrophin of cardiac sodium channels in caveolin-rich mem-
macromolecular signaling complex is required for mediated by PDZ domains. Cell 84:757–767, 1996. brane domains: Regulation of sodium current
beta(2)-adrenergic regulation. Proc Natl Acad Sci 40. Kramarcy NR, Vidal A, Froehner SC, et al: Asso- amplitude. Circ Res 90:443–449, 2002.
U S A 103:7500–7505, 2006. ciation of utrophin and multiple dystrophin short 60. Bossuyt J, Taylor BE, James-Kracke M, et al: The
19. Cronk LB, Ye B, Kaku T, et al: Novel mechanism forms with the mammalian m(r) 58,000 dystrophin- cardiac sodium-calcium exchanger associates with
for sudden infant death syndrome: Persistent late associated protein (syntrophin). J Biol Chem caveolin-3. Ann N Y Acad Sci 976:197–204, 2002.
sodium current secondary to mutations in 269:2870–2876, 1994. 61. Venema VJ, Ju H, Zou R, et al: Interaction of
caveolin-3. Heart Rhythm 4:161–166, 2007. 41. Suzuki A, Yoshida M, Ozawa E: Mammalian alpha neuronal nitric-oxide synthase with caveolin-3 in
20. Vatta M, Ackerman MJ, Ye B, et al: Mutant 1- and beta 1-syntrophin bind to the alternative skeletal muscle. Identification of a novel caveolin
caveolin-3 induces persistent late sodium current splice-prone region of the dystrophin cooh termi- scaffolding/inhibitory domain. J Biol Chem 272:
and is associated with long-qt syndrome. Circula- nus. J Cell Biol 128:373–381, 1995. 28187–28190, 1997.
tion 114:2104–2112, 2006. 42. Petitprez S, Zmoos AF, Ogrodnik J, et al: Sap97 62. Head BP, Patel HH, Roth DM, et al: G-protein-
21. Vaidyanathan R, Vega LA, Zhou Q, et al: Mecha- and dystrophin macromolecular complexes deter- coupled receptor signaling components localize in
nisms of lqts9 associated mutant, caveolin3-f97c, mine two pools of cardiac sodium channels NaV1.5 both sarcolemmal and intracellular caveolin-3-
regualtion of Kir2.1 channels. Biophys J 102:538a, in cardiomyocytes. Circ Res 108:294–304, 2011. associated microdomains in adult cardiac myo-
2012. 43. Ueda K, Valdivia C, Medeiros-Domingo A, et al: cytes. J Biol Chem 280:31036–31044, 2005.
22. Foell JD, Best JM, Tester DJ, et al: Long qt Syntrophin mutation associated with long qt syn- 63. Leonoudakis D, Conti LR, Radeke CM, et al: A
syndrome-associated f97c and s141r cav-3 muta- drome through activation of the Nnos-SCN5A multiprotein trafficking complex composed of
tions exert pleiotropic effects on cardiac l-type macromolecular complex. Proc Natl Acad Sci U S SAP97, CASK, VELI, and MINT1 is associated
calcium channels. Biophys J 102:127a, 2012. A 105:9355–9360, 2008. with inward rectifier KIR2 potassium channels. J.
23. Wong W, Scott JD: Akap signalling complexes: 44. Schuh K, Uldrijan S, Telkamp M, et al: The plas- Biol. Chem 279:19051–19063, 2004.
Focal points in space and time. Nat Rev Mol Cell mamembrane calmodulin-dependent calcium 64. Chen Z, Hague C, Hall RA, et al: Syntrophins
Biol 5:959–970, 2004. pump: A major regulator of nitric oxide synthase i. regulate alpha1d-adrenergic receptors through a
24. McConnachie G, Langeberg LK, Scott JD: J Cell Biol 155:201–205, 2001. PDZ domain-mediated interaction. J Biol Chem
Akap signaling complexes: Getting to the heart 45. Oceandy D, Cartwright EJ, Emerson M, et al: 281:12414–12420, 2006.
of the matter. Trends Mol Med 12:317–323, Neuronal nitric oxide synthase signaling in the 65. Okumura A, Nagai K, Okumura N: Interaction of
2006. heart is regulated by the sarcolemmal calcium alpha1-syntrophin with multiple isoforms of het-
25. Trotter KW, Fraser ID, Scott GK, et al: Alternative pump 4b. Circulation 115:483–492, 2007. erotrimeric g protein alpha subunits. FEBS J
splicing regulates the subcellular localization of 46. Ahern GP, Hsu SF, Klyachko VA, et al: Induction 275:22–33, 2008.
a-kinase anchoring protein 18 isoforms. J Cell Biol of persistent sodium current by exogenous and 66. Hasegawa M, Cuenda A, Spillantini MG, et al:
147:1481–1492, 1999. endogenous nitric oxide. J Biol Chem 275:28810– Stress-activated protein kinase-3 interacts with the
26. Piggott LA, Bauman AL, Scott JD, et al: The 28815, 2000. pdz domain of alpha1-syntrophin. A mechanism
a-kinase anchoring protein yotiao binds and regu- 47. Mohler PJ, Gramolini AO, Bennett V: The for specific substrate recognition. J Biol Chem
lates adenylyl cyclase in brain. Proc Natl Acad Sci ankyrin-b c-terminal domain determines activity of 274:12626–12631, 1999.
U S A 105:13835–13840, 2008. ankyrin-b/g chimeras in rescue of abnormal inosi- 67. Chockalingam PS, Gee SH, Jarrett HW: Pleckstrin
27. Tasken K, Aandahl EM: Localized effects of camp tol 1,4,5-trisphosphate and ryanodine receptor homology domain 1 of mouse alpha 1-syntrophin
mediated by distinct routes of protein kinase A. distribution in ankyrin-B (-/-) neonatal cardiomyo- binds phosphatidylinositol 4,5-bisphosphate. Bio-
Physiol Rev 84:137–167, 2004. cytes. J Biol Chem 277:10599–10607, 2002. chemistry 38:5596–5602, 1999.
Epigenetics in Cardiac
Rhythm Diseases 24
Adam B. Stein
homeotic genes are expressed in one body segment and not in a

CHAPTER OUTLINE
neighboring segment. By using genetic approaches, it was dem-
Definition of Epigenetics 235 onstrated that the spatial and temporal expression of the homeo-
tic genes was enhanced by the Trithorax group of genes and
Historical Experiments That Illustrate the Salient
repressed by the Polycomb group of genes in a segment-specific
Features of Epigenetics 235 manner.1 However, the manner in which the Trithorax and Poly-
Basis of Epigenetic Phenomena 235 comb groups of proteins controlled homeotic gene expression
was not initially understood. Investigators using yeast and Tetra-
Histone Tail Modifications 237 hymena discovered that enzymes that modify histones (the pack-
ATP-Dependent Chromatin Remodeling Complexes 237 aging structure for DNA) could silence specific genes.2-4
Biochemistry work revealed that the histone-modifying enzymes
Epigenetics in Induced Pluripotent Stem Cells that repress gene expression contain a SET domain that catalyzes
and Direct Reprogramming 237 histone methylation. The importance of histone methylation
Heart Muscle Development 238 in Drosophila species development became apparent when the
Polycomb protein complex was shown to have a SET domain–
Epigenetics in Disease States 239 containing enzyme capable of methylating H3K9, a repressive
Conclusion 240 epigenetic mark.5,6 Similarly, the Trithorax proteins TRX and
Ash1 were found to contain SET domains and capable of meth-
ylating histones with a mark that facilitates active gene
expression.7-9 Taken together, these studies reveal that histone
Definition of Epigenetics modifications are well-conserved phenomena that are critical for
regulating gene expression profiles during development.
Genetic mutations are well-recognized causes of heritable phe- The creation of Dolly the sheep also revealed important fea-
notypic shifts. The completion of the Human Genome Project tures of epigenetic mechanisms. Wilmut et al.10 sought to clone
and technical advances that allow for high-throughput sequenc- a sheep by taking the nucleus from a differentiated cell (mammary
ing have advanced the exploration for genetic signatures that gland cell) and placing that nucleus into an unfertilized oocyte
associate with and, perhaps, cause disease states. However, it is (developing egg cell). Although they were eventually successful,
very possible that many disease states result from environmental this process of nuclear transfer was extremely inefficient. This
influences that affect the expression or repression of different work revealed that although the mammary gland cell DNA con-
genes without altering the underlying DNA sequence. By defini- tains the blueprint necessary to develop an entire sheep, it was
tion, an epigenetic phenomenon is a heritable alteration in phe- not easy to revert a differentiated nucleus back to a pluripotent
notype without an underlying DNA mutation. Epigenetic state. It became apparent that although the blueprint of DNA is
phenomena regulate gene expression profiles by controlling the same in an embryonic stem (ES) cell and a differentiated cell,
genes in a binary manner such that genes are actively expressed nonmutation-based changes have occurred to the DNA in the
(i.e., “on”) or repressed (i.e., “off”). In addition, true epigenetic differentiated cell that prevent the simple reversion or manipula-
changes that occur as a result of a stimulus should be heritable tion of a cell’s identity. This study supports the idea that, as cells
and stable, even in the subsequent absence of that stimulus. differentiate, they accumulate stable and heritable changes in
Although the concept of epigenetics seems simple, it differs dis- their nucleus that restrict gene expression patterns and define
tinctly from the classic notion that only mutations at the DNA identity. These changes are, by definition, epigenetic.
level can cause heritable phenotypic shifts (Figure 24-1). Clini-
cians and scientists are exploring epigenetic mechanisms and
their contribution to the development of disease states, because
these phenomena provide a rational mechanistic explanation for Basis of Epigenetic Phenomena
how environmental factors can alter phenotypes in a heritable
manner in the absence of DNA mutations. The study of epigenetic phenomena presupposes that genes can
be marked as transcriptionally “on” or “off” by modifications that
are independent of the primary nucleotide coding sequence. The
DNA in each cell is composed of approximately 3 × 109 nucleo-
Historical Experiments That Illustrate the tide bases that could presumably stretch to 2 m in length. To
Salient Features of Epigenetics facilitate efficient packaging, DNA is wrapped in a threadlike
fashion around protein spools called histones. The histone octamer
Early work in Drosophila species development revealed that the consists of two molecules of each histone: H2A, H2B, H3, and
homeotic genes define segment identity and the position of H4.11-13 The histone octamer and the 146 bp of DNA looped
appendages along the body axes. Thus, developmental biologists around it forms a structure called a nucleosome, the functional unit
became interested in mechanisms to explain why specific of chromatin. In addition to serving as packaging facilitators,
235
Repressed Active
Repressed Active
B Repressed Active
Figure 24-1. Heritability of phenotypic changes at the level of DNA (A) and by epigenetic mechanisms (B). A, The genetic code is defined by the DNA template (orange
lines). In this model, DNA mutations (red X) can alter the expression of genes and the resultant phenotype. DNA replication results in mutations being passed from a mother
cell to her progeny. B, The process by which epigenetic mechanisms can regulate gene expression profiles in a heritable manner without mutating the underlying DNA.
The expression of genes can be regulated by interactions between nucleosome components (blue cylinders) and the DNA (orange strand) such that certain genes are
repressed (red X) and certain genes are active (green +). The DNA and the nucleosome components are both heritable so that the active and repressed gene patterns are
similar in the mother cells and daughter progeny.
ATP ADP Histone

modifier
Histone
DNA modifier modification
Chromatin
Me Me remodeler
B
Figure 24-2. A, DNA provides the template that encodes all potential genes. Not all genes are expressed in all cells. Chromatin provides a way of partitioning this genetic
information into active and repressed components. The three mechanisms whereby chromatin can modify the expression or repression of genes includes DNA methylation
(Me), chromatin remodelers (blue oval), and histone modifications (e.g., phosphorylation, methylation, acetylation) imparted by histone modifiers.
(Adapted from Chang CP, Bruneau BG: Epigenetics and cardiovascular development. Annu Rev Physiol 74:41–68, 2012.)
nucleosomes are amenable to modifications and dynamic remod- The regulation of gene expression at the chromatin level
eling, which makes them an active participant in many occurs by: (1) covalent histone tail modifications; (2) adenosine
chromosomal processes including transcription, replication, triphosphate (ATP)-dependent chromatin remodeling; and (3)
DNA repair, kinetochore and centromere construction, and telo- DNA methylation (Figure 24-2).14 This chapter focuses on cova-
mere maintenance. lent histone tail modifications and ATP-dependent chromatin
Epigenetics in Cardiac Rhythm Diseases 237
remodeling. The amino terminal tails of histones protrude and These proteins contain a SET domain and form HMT protein
24
are subjected to various covalent posttranslational modifications complexes. In some cases, these HMT and acetylase complexes
(e.g., methylation, acetylation, sumoylation, phosphorylation).11 are directed to specific sites of DNA by their interaction with
These histone tail modifications regulate whether the neighbor- transcription factors (Figure 24-4).19
ing DNA is transcriptionally active (euchromatin) or repressed
(heterochromatin). Activating histone modifications attract ATP-
dependent nucleosome remodeling factors that can reposition or
eject histones, thus promoting transcription or repression of the ATP-Dependent Chromatin
neighboring DNA. Thus, complex interplay between covalent Remodeling Complexes
histone modifications, ATP-dependent chromatin remodeling
enzymes, and the transcriptional machinery (i.e., DNA binding Chromatin-modifying enzymes use the energy derived from ATP
factors, transcription initiation, elongation complexes) deter- to alter the chromatin structure in a manner that permits (euchro-
mines whether genes are actively expressed or repressed. matin) or represses gene expression (heterochromatin) by restruc-
turing, mobilizing, and ejecting nucleosomes. There are four
families of SWI-like, ATP-dependent chromatin remodeling
complexes: the switching defective/sucrose nonfermenting (SWI/
Histone Tail Modifications SNF); imitation switch (ISWI); chromodomain, helicase, DNA-
binding (CHD); and inositol requiring 80 complexes (INO80).20
From an epigenetic standpoint, the two most important tail These chromatin-remodeling complexes share a well-conserved
modifications are acetylation and methylation.15 Histone tail SWI-like ATPase catalytic domain in combination with a unique
acetylation is associated with transcriptionally active chromatin. functional domain. The ATPase domain uses ATP hydrolysis to
In contrast, histone tail methylation can be associated with either remodel nucleosomes and render the neighboring DNA acces-
euchromatin or heterochromatin, depending upon the extent of sible or repressed.
methylation (mono- [me], di- [me2], or tri- [me3] methylation) The vertebrate SWI/SNF-type, ATP-dependent, chromatin-
and the specific tail residues that are modified. For example, remodeling complexes are known as Brg1/Brahma-associated
lysine (K) residues available for methylation include K4, K9, K27, factor (BAF) complexes. The BAF complexes use Brg1 or Brahma
and K36 of histone H3, and K20 of histone H4. Methylation of as their ATPase subunits. These BAF complexes are well studied
H3K9 and H3K27 is associated with gene repression gene and have been shown to play an important role in many aspects
expression,16 whereas dimethyl (H3K4me2) and trimethyl of development, including the maintenance of stem cell pluripo-
(H3K4me3) marks at H3K4 are associated with actively expressed tency and tissue-specific differentiation. Studies in mice have
genes.17 As shown in Figure 24-3,18 the extent, specific lysine shown that Brg1 is critical to normal heart development and
residue, and location of different histone methylation marks interacts with developmental cardiac transcription factors.
around a gene provide a signature for the transcriptional state of
that gene. For example, H3K27me3 marks (black line) are located
around the 5′ regulatory region of silenced genes. In contrast,
H3K4me3 marks are highly enriched at the 5′ transcription start Epigenetics in Induced Pluripotent Stem
site of actively expressed genes. In contrast, H3K4me3 marks are Cells and Direct Reprogramming
not as enriched around poised and silenced genes.
Histone methylation marks and acetylation marks are imparted Cells accumulate epigenetic marks that restrict their cell fates
and removed by enzymes that are part of protein complexes. during development. Pluripotent ES cells have an epigenome
Histone acetyltransferases, acetylate histones, and histone that permits the expression and repression of all genes that are
deacetylases reverse this acetylation process. Similarly, histone necessary for development and differentiation. During develop-
methyltransferase (HMT) complexes impart histone methylation ment, ES cells respond to external cues and begin to differentiate.
marks, and histone demethylases remove histone methylation During differentiation, ES cells accumulate stable and heritable
marks. HMT complexes often consist of a SET domain– epigenetic marks that restrict their fate and the fate of their
containing HMT enzyme that is capable of adding methyl groups daughter cells. Nuclear transfer experiments demonstrated that
to specific lysine residues. The mammalian homologues to the removing a nucleus from fully a differentiated cell, transferring
yeast Set1 and the Drosophila Trithorax proteins are the mixed that nucleus into a denucleated zygote, and developing a pluripo-
lineage leukemia family of proteins, now known as KMTA-D. tent stem is a very inefficient process, partly because of an
Extent of modification
H3K27me3 H3K4me3 H3K36me3
5′ Silenced Poised Active 3′
Figure 24-3. Epigenetic signatures differ around genes that are repressed (red rectangle), poised (black rectangle), and active (green rectangle). H3K27me3 marks (red line)
are enriched at the 5′ region of silenced genes H3K4me3 marks (green line) are enriched at actively expressed genes. Thus active genes show enrichment of H3K4me3
marks and a paucity of H3K27me3 marks at the 5′ region. H3K36me3 (yellow line) marks are enriched at active genes. This illustrates the histone code hypothesis, which
suggests that the epigenetic signature of individual genes, determines whether that gene is actively expressed or repressed.
(Adapted from Lee BM, Mahadevan LC: Stability of histone modifications across mammalian genomes: implications for “epigenetic” marking. J Cell Biochem 108:22–34, 2009.)
TF
HMT
TF binding site
Me
HMT
TF
TF binding site
HMT Me
Pol II
TF
TF binding site
C
Figure 24-4. How are certain histones marked with activating methylation marks? A, Transcription factors (red rectangle), histone methyltransferase (HMT) complexes
(orange octagon), DNA (orange line), histones (blue cylinders), and the histone tails (black lines) are present in the cell. Upon activation, a transcription factor (TF) binds to the
DNA at defined binding sites (red line). B, These transcription factors interact with histone methyltransferase complexes (orange hexagon), which impart activating methyla-
tion marks on neighboring histones (yellow hexagon). As shown in (C), these active histone methylation marks then interact with the transcription machinery (green square)
to facilitate transcription (black arrow).
inability to reset the epigenetic marks that are present in the different chromatin-modifying enzymes affect cellular repro-
differentiated cell nucleus.10 Despite the stability of these epigen- gramming.25 They demonstrated that inhibition of components
etic marks, investigators have more recently discovered that they of PRC1 and PRC2 reduces reprogramming efficiency. In com-
can revert differentiated fibroblasts into induced pluripotent stem parison, inhibition of SUV39H1, YY1, and the H3K79 methyl-
cells (iPSCs) by introducing a cocktail of transcription factors transferase DOT1L enhanced reprogramming.
that revert the epigenome to a more pluripotent state.21,22 Thus, To harness the ability of transcription factors to reprogram
although epigenetic stability is important for establishing cell cells in a clinically relevant manner, Ieda et al.26 demonstrated
identity during development in vivo, the ability to revert differ- that the combination of three developmental cardiac transcrip-
entiated cells to pluripotent cells by manipulating the epigenome tion factors (Gata4, Mef2c, and Tbx5) could be used to directly
in vitro has garnered much excitement. reprogram dermal and postnatal cardiac fibroblasts into differen-
Manipulating the epigenome to develop iPSCs has significant tiated beating cardiaclike cells without first developing iPSCs.
clinical relevance. iPSCs can serve as a source of pluripotent cells They showed that direct reprogramming of fibroblasts resulted
for regenerative medicine, and they provide a method of studying in the stable shift of the epigenetic signature of several cardiac
genetic-based disease states in vitro. For example, Yazawa et al. genes—Actn2, Ryr2, and Tnnt2.
used fibroblasts from patients with Timothy syndrome to gener-
ate iPSCs and differentiate them into cardiac myocytes in vitro.23
Using this model, they studied the electrical and calcium-handling
properties of myocytes that possess the Timothy syndrome muta- Heart Muscle Development
tion in vitro. Although the forced expression of a cocktail of
transcription factors reverts differentiated cells to pluripotent Successful development of the cardiovascular system requires a
iPSCs that share many features of ES cells, recent work complex integration of multiple inputs that ultimately results in
has demonstrated that iPSCs have incomplete epigenetic repro- stable gene expression profiles that determine the identity of
gramming.24 To determine the importance of epigenetic mecha- the differentiated cell. Transcription factors (e.g., Nkx2.5, Tbx5,
nisms on cellular reprogramming, Onder et al. investigated how Tbx1) have a critical role in defining cardiac identity by inducing
the expression of cardiac genes in a timed manner. Transcription Mef2cAHF:Cre driver that is expressed in cardiac progenitors of
24
factors interact with epigenetic factors to define how develop- the anterior heart field, they demonstrated that the absence of an
mental cues can be translated into stable and heritable cardiac Ezh2 results in attenuation of the observed increase that accom-
gene expression profiles. panies normal H3K27me3 marks development and an inability
to repress Six1 gene expression. The inappropriate expression of
Six1 resulted in massive right ventricular hypertrophy and mild
SWI/SNF Complexes in Heart Development pulmonary stenosis in adult mice.29
Many components of the BAF complex are necessary for the

development of the heart chambers and cardiac muscle. Brg1, an
adenosine triphosphatase (ATPase)-containing protein that is a Epigenetics in Disease States
member of the BAF complex has been well studied in myocardial
development and provides an example of the importance of Diabetes and Epigenetics
chromatin-remodeling factors in heart development. Brg1 is nec-
essary for early cardiomyocyte cell proliferation as well as the It is well established that patients with type I and type 2 diabetes
differentiation into mature cardiac myocytes. Murine models have a substantially increased risk of developing cardiovascular
with cardiac-specific deletion of Brg1 have a thinned myocar- complications.31,32 Thus, investigators have explored the effect
dium and absent interventricular septums.27 Work by Hang that tight glucose control has on diabetic complications. Basic
et al.27 revealed that Brg1 is necessary for the effective expression science and human clinical trials have demonstrated that a tran-
of Bmp10 and the repression of p57kip2, a cyclin dependent kinase sient exposure to hyperglycemia results in a higher incidence of
inhibitor, such that the deletion of Brg1 results in a lack of diabetic vascular disease despite a return to normal glucose
Bmp10 and ectopic expression of p57kip2 that represses cell cycle levels.33 This “legacy effect” is referred to as metabolic memory.
progression. Because epigenetic mechanisms can induce stable changes in
In addition to regulating cell cycle progression, Brg1 is also gene expression that persist despite the subsequent absence of the
necessary for the differentiation of cardiac myocytes. Beta- inciting stimulus, investigators have explored epigenetic mecha-
myosin heavy chain (β-MHC) is the primary MHC isoform nisms to explain the legacy effect.34 In corroboration with these
expressed in fetal myocytes whereas alpha-myosin heavy chain clinical observations, in vitro work on human aortic endothelial
(α-MHC) is the primary MHC expressed in adult myocytes. Brg1 cells has demonstrated that transient exposure to hyperglycemia
plays an important role in MHC isoform switching. In the results in sustained activation of NF-κB, a proinflammatory tran-
absence of Brg1, embryonic myocytes prematurely switch from scriptional regulator.35 A hyperglycemic exposure recruits epi-
β-MHC to α-MHC. Thus, Brg1 is important in the normal genetic machinery to the promoter region of the p65 subunit of
progression of differentiation from fetal to adult cardiac NF-κB. The epigenetic machinery facilitates a decrease in
myocytes.27 H3K9me repressive marks, an increase in H3K4me activating
Brg1 has also been used to demonstrate the importance of the marks, and a sustained transcriptional increase in the p65 subunit
relationship between cardiac transcription factors and epigenetic of NF-κB. The increase in p65 results in the persistent expression
complexes. Brg1+/– mice have variable cardiac phenotypes includ- of a proinflammatory gene expression profile. This in vitro work
ing ventricular septal defects (VSDs), patent foramen ovales demonstrates how environmental factors (e.g., hyperglycemia)
(PFO), conduction abnormalities, and cardiac dilatation.27 Brg1 can be translated into a stable shift in disease-defining gene
has been shown to interact with the key cardiac transcription expression profiles by epigenetic mechanisms.
factors Tbx5, Nkx2-5, and Gata4 and potentiate transcription
factor–mediated transcription.28 Furthermore, the combination
of Brg1+/– haploinsufficiency with NKX2-5 or Tbx20 heterozy- Cardiac Hypertrophy and Cardiac Failure
gous mice results in the development of severe lethal cardiac
defects. These studies suggest that the effective development of The development of cardiac hypertrophy and cardiac failure is a
cardiac myocytes requires an interaction between transcription result of a stimulus-induced shift in the cardiac phenotype.
factors and epigenetic complexes. Kaneda et al. investigated which shifts in histone tail modifica-
tions accompany rat and human models of heart failure.36 By
using genome-wide approaches, they demonstrated that there is
Ezh2 in Cardiac Development a significant shift in the profile of H3K4me3 (activating) and
H3K9me3 (repressive) marks in Dahl salt–sensitive rats with
Polycomb repressive complexes (PRCs) play a role in repressing heart failure. Next, using human heart failure samples, they con-
gene expression profiles. Ezh2 in the is a key histone methyltrans- firmed that human heart failure is also associated with a signifi-
ferase of PRC2 complex which trimethylates histone 3 at lysine cant shift in genome-wide H3K4me3 and H3K9me3 profiles.
27 (H3K27). H3K27 methylation marks are associated with tran- They found that the activating H3K4me3 marks are associated
scriptionally repressed chromatin. As cardiac myocytes develop, with many important canonical signaling pathways (e.g., calcium
they undergo an increase in H3K27me3 levels.29 The deletion of signaling, nitric oxide signaling, cAMP-mediated signaling,
Ezh2 during cardiac development in murine models has resulted G-protein–coupled receptor signaling, actin signaling) that are
in disparate results. He et al.30 deleted a floxed Ezh2 allele in early important in heart failure. Taken together, these results suggest
cardiac development in mice using an NKX2.5:Cre driver. These that there is an association between H3K4me3 marks and
mice demonstrated lethal congenital heart malformations includ- H3K9me3 marks and the alteration in gene expression profiles
ing compact myocardial hypoplasia, hypertrabeculation, and ven- that accompanies cardiac disease states.
tricular septal defects. Deleting Ezh2 with Tnnt:Cre, which To investigate the mechanistic role of histone-modifying pro-
comes on after NKX2-5:Cre, attenuated the phenotype observed teins in the development of cardiac hypertrophy, Zhang et al.
with NKX2-5:Cre, suggesting that the establishment of looked at the importance of Jmjd2a in the development of cardiac
H3K27me3 epigenetic marks occurs during a finite window of hypertrophy.37 Jmjd2a is a member of the JmjC domain–
time during cardiac development.30 In a different study, Delgado- containing family JMJD2 of histone demethylases. The JMJD2
Olguin et al. reported that deletion of Ezh2 using the NKX2.5:Cre family of proteins demethylate H3K9me3 and H3K36me3.
mice resulted in mild cardiac hypertrophy. However, using a H3K9me3 is associated with heterochromatin; thus, Jmjd2a acts
as an activator by removing repressive epigenetic marks. By using disease-like phenotype that included significant attenuation in
murine models of pressure overload–induced hypertrophy, Zhang Kcnip2 gene expression, a decrease in the transient outward
et al. demonstrated that Jmjd2a deletion attenuates the develop- potassium current, an increase in calcium transients, and a pro-
ment of cardiac hypertrophy. They also showed that cardiac- pensity to develop premature ventricular beats. This work dem-
specific overexpression of Jmjd2a resulted in an exaggerated onstrated that it is necessary to maintain epigenetic profiles in
hypertrophic response. fully differentiated tissues, and that an inability to maintain these
Studies in cardiac development revealed that Brg1 is critical marks over time can be a mechanistic cause of disease.
for the switch from β-MHC to α-MHC. Interestingly, the devel-
opment of cardiac hypertrophy is known to be accompanied by
a switch from the adult α-MHC isoform to the fetal β-MHC
isoform. Human studies reveal that Brg1 is upregulated in cardiac Conclusion
tissue from patients with hypertrophic cardiomyopathy. The
degree of Brg1 expression correlates with the extent of the hyper- Epigenetic phenomenon in cardiac development and disease are
trophic phenotype. Brg1 is normally downregulated in adult currently under intense investigation because epigenetic phenom-
heart tissue; however, pressure overload–induced hypertrophy in ena provide a mechanistic explanation for how environmental
murine models results in re-expression of Brg1. Deletion of Brg1 factors can cause changes in the cardiac phenotype. As outlined
in murine models attenuates the hypertrophic response, blunts earlier, considerable evidence supports the important role of epi-
the development of fibrosis, and blocks the switch in MHC genetic mechanism in the developing heart. These same epigen-
isoforms.27 etic mechanisms that establish stable and heritable gene expression
profiles during development also seem to be important contribu-
tors to the development of disease. Epigenetic mechanisms may
Electromechanical Function and Arrhythmias be particularly important in the development of adult cardiac
disease states because most cardiac myocytes are post-mitotic
Although epigenetic mechanisms are important for defining a terminally differentiated cells. Therefore, an inability to maintain
cell’s identity during development, it remains unclear whether an the epigenome over the life of a cell may alter gene expression
inability to maintain the epigenome once a cell is fully differenti- profiles in aged myocytes. These shifts in the epigenome could
ated can be a causative mechanism of disease. To test this hypoth- predispose to or initiate the development of disease. In addition,
esis, Stein et al. deleted PTIP, a protein that is part of a histone epigenetic mechanisms may provide a rational explanation for
methyltransferase complex that enriches at actively expressed why some disease processes appear reversible and some are irre-
genes, in fully differentiated murine adult myocytes using an versible. In addition to potentially causing disease, manipulating
inducible Cre driver.38 These studies revealed that deleting a the epigenome to derive iPSCs or to direct differentiation into
component of the histone methyltransferase complex in a fully other cell types (i.e., fibroblasts to myocytes) may serve a thera-
differentiated myocyte results in an inability to maintain peutic purpose. In any case, a deeper understanding of epigenetic
H3K4me3 marks. This epigenetic instability resulted in a mechanisms will provide us with new insights into cardiac biology.
10. Campbell KH, McWhir J, Ritchie WA, et al: Sheep 21. Takahashi K, Yamanaka S: Induction of pluripotent
References cloned by nuclear transfer from a cultured cell line. stem cells from mouse embryonic and adult fibro-
Nature 380:64–66, 1996. blast cultures by defined factors. Cell 126:663–676,
1. Ringrose L, Paro R: Polycomb/trithorax response 11. Luger K, Rechsteiner TJ, Flaus AJ, et al: Charac- 2006.
elements and epigenetic memory of cell identity. terization of nucleosome core particles containing 22. Park IH, Zhao R, West JA, et al: Reprogramming
Development 134:223–232, 2007. histone proteins made in bacteria. J Mol Biol of human somatic cells to pluripotency with
2. Nakayama J, Rice JC, Strahl BD, et al: Role of 272:301–311, 1997. defined factors. Nature 451:141–146, 2008.
histone h3 lysine 9 methylation in epigenetic 12. Germond JE, Hirt B, Oudet P, et al: Folding of the 23. Yazawa M, Hsueh B, Jia X, et al: Using induced
control of heterochromatin assembly. Science DNA double helix in chromatin-like structures pluripotent stem cells to investigate cardiac pheno-
292:110–113, 2001. from simian virus 40. Proc Natl Acad Sci U S A types in timothy syndrome. Nature 471:230–234,
3. Rea S, Eisenhaber F, O’Carroll D, et al: Regulation 72:1843–1847, 1975. 2011.
of chromatin structure by site-specific histone 13. Olins AL, Olins DE: Spheroid chromatin units 24. Lister R, Pelizzola M, Kida YS, et al: Hotspots of
h3 methyltransferases. Nature 406:593–599, (V bodies). Science 183:330–332, 1974. aberrant epigenomic reprogramming in human
2000. 14. Chang CP, Bruneau BG: Epigenetics and cardio- induced pluripotent stem cells. Nature 471:68–73,
4. Strahl BD, Ohba R, Cook RG, et al: Methylation vascular development. Annu Rev Physiol 74:41–68, 2011.
of histone h3 at lysine 4 is highly conserved and 2012. 25. Onder TT, Kara N, Cherry A, et al: Chromatin-
correlates with transcriptionally active nuclei in 15. Kouzarides T: Chromatin modifications and their modifying enzymes as modulators of reprogram-
tetrahymena. Proc Natl Acad Sci U S A 96:14967– function. Cell 128:693–705, 2007. ming. Nature 483:598–602, 2012.
14972, 1999. 16. Litt MD, Simpson M, Gaszner M, et al: Correla- 26. Ieda M, Fu JD, Delgado-Olguin P, et al: Direct
5. Czermin B, Melfi R, McCabe D, et al: Drosophila tion between histone lysine methylation and devel- reprogramming of fibroblasts into functional car-
enhancer of zeste/esc complexes have a histone h3 opmental changes at the chicken beta-globin locus. diomyocytes by defined factors. Cell 142:375–386,
methyltransferase activity that marks chromosomal Science 293:2453–2455, 2001. 2010.
polycomb sites. Cell 111:185–196, 2002. 17. Santos-Rosa H, Schneider R, Bannister AJ, et al: 27. Hang CT, Yang J, Han P, et al: Chromatin regula-
6. Muller J, Hart CM, Francis NJ, et al: Histone Active genes are tri-methylated at k4 of histone h3. tion by brg1 underlies heart muscle development
methyltransferase activity of a Drosophila poly- Nature 419:407–411, 2002. and disease. Nature 466:62–67, 2010.
comb group repressor complex. Cell 111:197–208, 18. Lee BM, Mahadevan LC: Stability of histone 28. Lickert H, Takeuchi JK, Von Both I, et al: Baf60c
2002. modifications across mammalian genomes: Impli- is essential for function of baf chromatin remodel-
7. Beisel C, Imhof A, Greene J, et al: Histone meth- cations for “epigenetic” marking. J Cell Biochem ling complexes in heart development. Nature
ylation by the drosophila epigenetic transcriptional 108:22–34, 2009. 432:107–112, 2004.
regulator ash1. Nature 419:857–862, 2002. 19. Patel SR, Kim D, Levitan I, et al: The brct-domain 29. Delgado-Olguin P, Huang Y, Li X, et al:
8. Milne TA, Briggs SD, Brock HW, et al: Mll targets containing protein ptip links pax2 to a histone h3, Epigenetic repression of cardiac progenitor
set domain methyltransferase activity to hox gene lysine 4 methyltransferase complex. Dev Cell gene expression by ezh2 is required for postnatal
promoters. Mol Cell 10:1107–1117, 2002. 13:580–592, 2007. cardiac homeostasis. Nat Genet 44:343–347,
9. Nakamura T, Mori T, Tada S, et al: All-1 is a 20. Saha A, Wittmeyer J, Cairns BR: Chromatin 2012.
histone methyltransferase that assembles a super- remodelling: The industrial revolution of DNA 30. He A, Ma Q, Cao J, et al: Polycomb repressive
complex of proteins involved in transcriptional around histones. Nat Rev Mol Cell Biol 7:437–447, complex 2 regulates normal development of the
regulation. Mol Cell 10:1119–1128, 2002. 2006. mouse heart. Circ Res 110:406–415, 2012.
31. Retinopathy and nephropathy in patients with type 33. Nathan DM, Cleary PA, Backlund JY, et al: Inten- 36. Kaneda R, Takada S, Yamashita Y, et al: Genome-
1 diabetes four years after a trial of intensive sive diabetes treatment and cardiovascular disease wide histone methylation profile for heart failure.
therapy. The diabetes control and complications
trial/epidemiology of diabetes interventions and
complications research group. N Engl J Med
in patients with type 1 diabetes. N Engl J Med
353:2643–2653, 2005.
34. Siebel AL, Fernandez AZ, El-Osta A: Glycemic
Genes Cells 14:69–77, 2009.
37. Zhang QJ, Chen HZ, Wang L, et al: The histone
trimethyllysine demethylase jmjd2a promotes
24
342:381–389, 2000. memory associated epigenetic changes. Biochem cardiac hypertrophy in response to hypertrophic
32. Writing Team for the Diabetes Control and Com- Pharmacol 80:1853–1859, 2010. stimuli in mice. J Clin Invest 121:2447–2456, 2011.
plications Trial/Epidemiology of Diabetes Inter- 35. Brasacchio D, Okabe J, Tikellis C, et al: Hypergly- 38. Stein AB, Jones TA, Herron TJ, et al: Loss of h3k4
ventions and Complications Research Group: cemia induces a dynamic cooperativity of histone methylation destabilizes gene expression patterns
Sustained effect of intensive treatment of type 1 methylase and demethylase enzymes associated and physiological functions in adult murine cardio-
diabetes mellitus on development and progression with gene-activating epigenetic marks that coexist myocytes. J Clin Invest 121:2641–2650, 2011.
of diabetic nephropathy: The epidemiology of dia- on the lysine tail. Diabetes 58:1229–1236, 2009.
betes interventions and complications (EDIC)
study. JAMA 290:2159–2167, 2003.
Cell Biology of Cardiac Impulse PART IV
Initiation and Propagation
Cardiac Impulse Is Initiated by a

Coupled System of Membrane Ion
Channels and Ca2+ Cycling Proteins
25
Edward G. Lakatta, Yael Yaniv, and Victor A. Maltsev
electrotonic impulses, neurotransmitter or hormonal stimulation

CHAPTER OUTLINE of surface membrane receptors).
Realization of Importance of Ca2+ Signaling for Cardiac The present perspective views SANC, per se, as a system
Pacemaker Function 243 comprising several levels of complexity and integrated compo-
2+
nents. Sarcolemmal ion currents (Figure 25-1 and Table 25-1)
Ca -Clocks in Pacemaker Cells 244 have been extensively studied in the past (see review1), and the
Crosstalk of Ryr and NCX to Transfer Intracellular Ca 2+ ensemble of these currents re-created in silico from experimental
Signals to M-Clock 245 voltage-clamp data can generate spontaneous APs.2 The subsys-
tem of sarcolemmal molecules forming a voltage membrane
The Essence of Pacemaker Function Is a Coupled oscillator was conceptualized as a membrane clock3 (M-clock, for
Function of an Intracellular Ca2+-Clock and Membrane short) and has been the predominant feature in at least 12 SANC
Ion Channel Clock 245 numerical models.2 During rhythmic spontaneous AP firing by
SANC, the M-clock is dynamically coupled to Ca2+ cycling (see
Autonomic Receptor Modulation of the
Figure 25-1, gray area) via multiple voltage-, Ca2+-, cyclic adenos-
Coupled-Clock System 249 ine monophosphate (cAMP)-, and phosphorylation-dependent
Mitochondrial Function in SANC 250 mechanisms (coupling factors).4 SANC sarcoplasmic reticulum
(SR), however, in isolation from the M-clock can generate its own
An Additional Level of Complexity of Heart spontaneous local (submembrane) rhythmic Ca2+ signals. Thus,
Pacemaker Function Arises Within the SAN Tissue 251 similar to the M-clock, the SR is intrinsically also an almost
perfectly periodic oscillator that can be envisioned as a Ca2+-
clock.3 During spontaneous AP firing by SANC, the M- and
Different types of heart cells (i.e., pacemaker cells within the Ca2+-clocks do not exist as separate entities: They mutually
sinoatrial node [SAN] and contractile myocytes within myocar- entrain each other via aforementioned coupling factors. These
dium) determine how fast and strong, the heart beats. The heart’s coupled oscillators that differ in nature thus represent heteroge-
pacemaker cells normally generate spontaneous rhythmic changes neous redundancy within SANC that confers a physiological
in their membrane potential, thereby producing almost perfectly robust, yet extremely flexible, united pacemaker cell system.4,5
periodic spontaneous action potential (APs; i.e., normal automa-
ticity). Pacemaker cells within the heart having “clocks” with the
briefest rhythmic periods “capture” or trigger other excitable
cells. SAN cells (SANC) are the dominant cardiac pacemaker Realization of Importance of Ca2+ Signaling
cells, because they exhibit shorter periods between spontaneous for Cardiac Pacemaker Function
APs than do atrioventricular nodal cells or His-Purkinje cells.
Thus, SANC normally initiate the cardiac impulse by generating In the late 1970s to the early 1980s, as the importance of an
spontaneous APs that are conducted to the ventricle and entrain intracellular Ca2+ transient (as the trigger for ventricular myocyte
the duty cycle of ventricular myocytes. contraction) came into focus, the idea that an intracellular Ca2+
The essence of cardiac pacemaker function is diastolic depo- oscillator could also drive membrane excitations in Purkinje
larization (DD) that is absent in ventricular myocytes. DD rep- fibers was suggested.6 But because oscillatory current in Purkinje
resents a slow spontaneous decrease in membrane polarization fibers became manifest only in a Ca2+ overload state, such current
until it reaches the membrane excitation threshold, culminating fluctuations were interpreted in the context of “abnormal auto-
in generation of an AP. SANC robustness (i.e., “fail safe” proper- maticity” (i.e., they were attributed to a disturbance in normal
ties conserved during evolution of the animal kingdom) and flex- cardiac function7).
ibility (the ability to react to demands for faster or slower AP In the late 1980s, afterdepolarizations and aftercontractions
firing rate) result from the timely integration of signaling events were recorded in Purkinje fibers under normal Ca2+ loading con-
at multiple levels within the SAN, including subcellular, cellular ditions,8 and these fibers demonstrated localized, spontaneous
surface membrane, and tissue architecture. The interactive myofilament motion caused by spontaneous local Ca2+ oscilla-
network of mechanisms intrinsic to SANC must also interpret tions9 in the absence of Ca2+ overload. A cytosolic Ca2+ transient
and react to signals arising extrinsic to the cell (e.g., stretch, is evoked by each spontaneous SANC AP (see Figure 25-2, D),
243
244 CELL BIOLOGY OF CARDIAC IMPULSE INITIATION AND PROPAGATION
β-Adrenergic Cholinergic
Receptor Receptor IKACh If (HCN4/2) ICaL IKs INaK
gs P (+) P
giα giαβγ gβγ (+) Basal
A Cyclase (+) (+) (-)
(+) (-)
GPCR
(+) cAMP(+) PKA
Activated (+) (-)
A Cyclase PHOSPHO-
CaM Ca2+clock DIESTERASES
INCX (-) (+)
RyR SR P (+)
Ca2+
IK
(+)
(+) [Ca2+] P
Ca2+
PLB
P
(-)
ICaT SERCA2a
PHOSPHATASES
Ca2+ SR CyclingCa2+
ICaL CaM
(+) Kinetics & Ca2+
Ca2 Release Amplitude
(+) (-)
(-) (+) + RyR, PLB, SERCA2a

P Protein
(+) CaM phosphatase
inhibitor
M clock
(-) CaMK-II (+)
P P P
(-) P (-)
(+)
(+)
1
Cell Surface Membrane

Figure 25-1. Schematic illustration of interactions of identified molecules comprising the fully coupled-pacemaker cell clock. Note that common regulatory factors (purple
large lettering) govern the function of both the SR Ca2+-clock (gray intracellular area) and the M-clock (light-blue cell membrane area with blue labels depicting electrogenic
proteins). These common factors act as nodes within the system to couple the function of the activities of both clocks. The system is balanced as illustrated by traffic
light–like colors: Green arrows designate signaling driving AP firing, but red arrows show suppression, maintaining a given steady state level of cAMP and protein phos-
phorylation. G protein–coupled receptors (green and red shapes within the membrane) modulate both Ca2+-clock and M-clock function via those same crucial signaling
nodes of the system.
(Modified from Lakatta EG, Maltsev VA, Vinogradova TM: A coupled SYSTEM of intracellular Ca2+ clocks and surface membrane voltage clocks controls the timekeeping mechanism
of the heart’s pacemaker. Circ Res 106:659–673, 2010.)
and that Ca2+ influx via L-type Ca2+ channels (LCCh) (see Figure venosus cells19 has documented the occurrence of spontaneous,
25-1) affects SR Ca2+ loading. Chelation of intracellular Ca2+ in localized, diastolic Ca2+ release in pacemaker cells in the absence
SANC10-12 markedly slowed or abolished spontaneous AP firing. of Ca2+ overload. These local Ca2+ releases (LCRs) during DD
The importance of SR Ca2+ release and Na+/Ca2+ exchanger are initiated beneath the cell surface membrane via spontaneous
(NCX) current (INCX) for pacemaker cell pacemaker function was activation of RyR (Figure 25-2A,B). SANC exhibit robust
also demonstrated directly by effects of ryanodine on Ca2+ SERCA2 and RyR immunolabeling,20,21 and although SERCA2
cycling.13,14 Ryanodine (by disabling the ryanodine receptor immunolabeling in SANC occurs diffusely throughout the cyto-
[RyR] and the SR release channel [Figure 25-2], and by depleting plasm and in perinuclear area, RyR immunolabeling is most
SR Ca2+ content] had a profound negative chronotropic effect on intense in the subsarcolemmal space.20,21 LCRs appear as 4- to
the automaticity of subsidiary atrial pacemakers15 and in 10-µm Ca2+ wavelets in confocal line-scan images of spontane-
SANC.13,14 Additional voltage-clamp experiments in perforated ously firing rabbit SANC and emerge following dissipation of the
patch configuration in isolated cat atrial latent pacemaker cells global systolic transient effected by the previous AP; they cre-
demonstrated the occurrence of an inward current during late scendo during DD, peaking during late DD, as they merge into
DD that was sensitive to ryanodine, and presumably was linked the global cytosolic Ca2+ transient triggered by the next AP (see
to INCX activated by SR Ca2+ release.16 Figure 25-2A).18,22 LCRs and the ensemble LCR signal (i.e., inte-
gral of all LCRs), that is, late diastolic Ca2+ elevations (LDCaE),
have now been documented in numerous species (see review4)
(see Figure 25-2, B-D).
Ca2+-Clocks in Pacemaker Cells LCR occurrence does not require triggering by depolarization
of the surface membrane: Persistent rhythmic oscillatory mem-
Confocal imaging of Ca2+ in mammalian SANC and atrial sub- brane currents can be activated by rhythmic LCRs during
sidiary pacemaker cells combined with noninvasive perforated voltage-clamp (at potentials that prevent cell Ca2+ depletion, e.g.,
patch-clamp electrophysiology17,18 and imaging of toad sinus −10 mV). Both persistent LCRs and the currents activate simul-
CARDIAC IMPULSE IS INITIATED BY A COUPLED SYSTEM OF MEMBRANE ION CHANNELS AND Ca2+ CYCLING PROTEINS 245
Table 25-1. Major Ion Currents Reported in SA Node and SANC Crosstalk of Ryanodine Receptor and Na+/
of Various Species
Group of Currents Ion Current

Ca2+ Exchanger to Transfer Intracellular
Ca2+ Signals to M-Clock 25
Voltage-gated Na+ current INa (absent in primary SANC) NCX is not an ion channel, and its amplitude almost instantly
Voltage-gated Ca 2+
ICaL, high voltage–activated, L-type Ca follows changes in membrane potential or intracellular [Ca2+].
currents (ICa= ICaL + ICaT ) current NCX function is both voltage- and Ca2+-dependent. When intra-
cellular [Ca2+] is high as the result of AP-induced Ca2+ transient,
ICaT, low voltage–activated, T-type Ca AP repolarization activates NCX forward mode, generating
current inward current INCX by exchanging one Ca2+ (going out) to 3 Na+
Voltage-gated K+ currents Ito, 4-AP–sensitive transient K+ current (going into the cell). In contrast to hyperpolarization-activated
(I4-AP = Ito + Isus and IK = IKr + “funny” current, If (see Table 25-1), INCX is activated earlier, as it
Isus, 4-AP–sensitive sustained K+ current
IKs) almost instantly follows membrane hyperpolarization (see Figure
(the sustained part of initially
25-2, D). Furthermore while the non-selective If is decreasing and
discovered Ito or I4-AP)
reversing during later DD, INCX is substantially increasing further
IKr, the rapidly activating component of by waxing diastolic Ca2+ release from the SR (see Figure 25-1 and
IK exhibiting strong inward rectification Figure 25-2, D, described later in detail). Thus, INCX has been
(mouse, rat, guinea pig, rabbit) realized as a major transduction mechanism of intracellular Ca2+
IKs, the slowly activating component of signaling that crucially contributes to DD and normal automatic-
IK, exhibiting only weak rectification ity of SANC. The NCX is one of the earliest functional genes that
(guinea pig, pig) exist during early embryonic heart development,27 and NCX1 is
required for normal pacemaker activity. Atrial-specific NCX
Small IK1, inwardly rectifying K+ current knockout mice that lack NCX1 but have intact If, and express
(mouse, rat, and monkey) NCX1 in ventricular myocytes,28 live to adulthood, and exhibit a
Voltage-gated Hyperpolarization-activated, “funny” markedly reduced heart rate. However APs can be evoked under
monovalent cation current, If or Ih current-clamp conditions in SANC isolated from these mice,
nonselective currents indicating that knockout cells retain electrical excitability.
Steady nonselective current Ist
At the resolution of the confocal microscope21 NCX and
ACh-activated K+ current IKACh RyRs, molecules appear co-localized across the ≈12-nm subsar-
Background and ion Store-operated Ca2+ current (mouse)
colemmal gap between RyRs on the SR and NCX molecules on
transporter currents the sarcolemma. The crosstalk of LCRs and NCX activates a
Background Na+ current, Ib,Na forward mode NCX operation that generates local inward INCX
Na+/K+ pump current, INaK or Ip currents, which produce miniature membrane voltage fluctua-
tions.18,29,30 The net late DD inward current initiated by the LCR
Na+/Ca2+ exchanger current, INaCa or INCX ensemble (measured as ryanodine-sensitive current) in rabbit
Cl− current, ICl SANC varies from 0.3 pA/pF31 to 1.6 pA/pF,18 yielding a whole-
cell INCX range from 9.6 to 51.2 pA for a 32-pF SANC (see Figure
25-2, D). Because an extremely small net ion current change (a
few pA in rabbit SANC) has marked effects on membrane poten-
tial, an inward current of this magnitude is sufficient to broadly
taneous fluctuations of the same frequency,22,23 and both modulate the DD.5,30,31 For comparison, the peak of If achieved
are abolished by ryanodine. Sustained LCR activity is also during DD in SANC ranges from 0.02 to 0.23 pA/pF as predicted
observed in chemically “skinned” SANC (i.e., having a detergent- by 12 different numerical models.2 LCR-activated inward INCX,
permeabilized cell surface membrane) bathed in a physiological together with other DD mechanisms (such as L-type Ca2+
[Ca2+] of 100 nM.22,23 Because LCRs are generated as rhythmic current, ICaL), imparts the exponentially rising phase to the
events at rates of 1 to 5 Hz (i.e., encompassing those of spontane- DD18,29,30 that initiates the subsequent AP (see Figure 25-2, A-C).
ous AP firing in SANC), SR was conceptualized as an intracel- Although voltage-activated L-type Ca2+ currents surely contrib-
lular “Ca2+-clock.”3 The ability of SANC SR to generate sustained ute to late DD, and L-type Ca2+ currents are required for AP
intracellular Ca2+ oscillations under normal physiological condi- initiation, failure to generate diastolic INCX signals results in pace-
tions within this broad frequency range has also been demon- maker failure when ICaL density is normal (e.g., when RyRs are
strated in numerical model simulations that have embraced the disabled ryanodine or during short-term NCX blockade18). Thus
occurrence of experimentally determined LCRs.5 available experimental and numerical modeling data clearly indi-
RyR2 knockout not only causes failure of spontaneous Ca2+ cate an important role of Ca2+ cycling molecules and their “cross-
releases in embryonic cardiac cells and suppresses spontaneous talk” in normal physiological function.
cell beating, it also leads to a marked depression of the obligatory
developmental increase in heart rate and cardiac output that, in
the absence of knockout, support continued cardiac embryonic The Essence of Pacemaker Function Is a
differentiation.24 As in adult SANC, a crucial regulatory role of Coupled Function of an Intracellular Ca2+-
RyR2-mediated Ca2+ releases in pacemaker function is demon-
strated in RyR2 knockout cells, as β-adrenergic receptor (β-AR)
Clock and Membrane Ion Channel Clock
stimulation is ineffective in these cells.25 An essential role of
RyR2 in pacemaker function has been recently demonstrated in The LCR Period Reports the Functional State of the
tissue-specific RyR2 knockout mice with acute ≈50% loss of Coupled System (Not That of the Ca2+-Clock Per Se)
RyR2 protein in the heart, but not in other tissues.26 The RYR2
loss-of-function causes bradycardia and arrhythmia. Further- The LCR period is the delay between the AP-triggered global
more, cardiac RyR2 knockout mice exhibit some functional and cytosolic Ca2+ transient and LCR emergence during DD (Figure
structural hallmarks of heart failure, including sudden cardiac 25-2, A). It reports the time at which AP ignition is prompted
death. via INCX activation, and it closely predicts the AP cycle length in
Dynamic interactions of a system of electrogenic and Ca2+ cycling molecules

Experiment Numerical model prediction
Sarcolemma (M-clock) Action

Potential (AP)
35 Vm
10
60 mV
DD
acceleration LCR
10µm
–15
–40 AP Ignition
I kr
20 pA
LCR period 1 3.5
–65 mV 500 ms
A Cycle length
F/F0
If
10 pA 5 pA
I NCX activation
by hyperpolarization
Superimposed raw traces of
2
LCRs
100 ms I NCX
F/Fo
INCX
activation
by LCRs
Ca2+ signal
AP-induced
100 pA
2 Ca2+ transient I CaL
Ensemble LCR INCX activation
INCX activation
signal (LDCaE) by Ca2+ transient
F/Fo
by Ca2+ transient
during AP upstroke
1 100 ms
B
Yin Yang
i Rabbit SANC AP-induced
Ca2+ transient
LDCaE
(noisy curve)
Ca2+ signal
Sarcoplasmic reticulum ICal-induced

Ca2+ -clock) Ca2+ release
(CICR)
[Ca] Submembrane space
LCR period Ca2+ transient
Vm
AP-induced LDCaE
ii Guinea-pig SANC Ca2+ transient
1 µM
(noisy curve)
LDCaE
Ca2+ signal
20 mV
500 ms SR depletion/
[Ca]SR
300 µM
“resetting” + “refueling”
Refilling
C Vm D SR with Ca2+
Figure 25-2. A, Line-scan image of LCRs (white arrowheads) with superimposed spontaneous APs recorded in rabbit SANC. B, Upper panel, LCRs imaged by confocal
microscopy; lower panel, Temporal average of LCRs creates ensemble LCR signal or late diastolic Ca2+ elevation (LDCaE) that precedes AP-induced Ca2+ transient.29 C, LDCaE
in single SANC of rabbit57 and guinea pig.20 D, Novel numerical model predicts “yin-yang” types of interactions within the coupled-clock system of SANC (see text for details).
D, Diastolic depolarization; SR, sarcoplasmic reticulum.
(Modified from Maltsev VA, Lakatta EG: Synergism of coupled subsarcolemmal Ca2+ clocks and sarcolemmal voltage clocks confers robust and flexible pacemaker function in a
novel pacemaker cell model. Am J Physiol Heart Circ Physiol 296:H594–H615, 2009.)
200
CCh
ISO
300
Control
Suppression
Activation
β-AR
25
250
ChR
LCR period, % control
IBMX
Cycle length % Control

SERCA
150
Milrinone 200 PKA
PKI PP1 & PP2
Control 150 PDE

Mito-NCX
100 100 Uniporter
50
50 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
A Phospholamban phosphorylation % control B LCR period, % control
Figure 25-3. A, PKA-dependent changes in phosphorylation of phospholamban predict changes in the SANC LCR period. B, Changes in the steady state LCR period in
response to numerous perturbations of the coupled-clock system predict concomitant changes in the steady state spontaneous AP cycle length.
(Modified from Lakatta EG, Maltsev VA, Vinogradova TM: A coupled SYSTEM of intracellular Ca2+ clocks and surface membrane voltage clocks controls the timekeeping mechanism
of the heart’s pacemaker. Circ Res 106:659-673, 2010. Updated with data from Vinogradova TM, Brochet DX, Sirenko S, et al: Sarcoplasmic reticulum Ca2+ pumping kinetics regu-
lates timing of local Ca2+ releases and spontaneous beating rate of rabbit sinoatrial node pacemaker cells. Circ Res 107:767–775, 2010.)
numerous experimental perturbations of SANC function (Figure the cell via NCX earlier in the prior cycle. At the same time, SR
25-3, B). Over the wide range of conditions and widely varying begins to refill Ca2+ (via sarco/endoplamic reticulum Ca2+-ATPase
AP firing rates from about 1 to 4 Hz, the relationship of LCR [SERCA] Ca2+pumping and, likely, inter-SR Ca2+ diffusion toward
periods to the spontaneous cycle lengths of SANC is nearly junctional SR) and then generates LCRs again when restitution is
identical, with a slope being close to 1 and an offset “ignition” completed with the achievement of a threshold intra-SR [Ca2+]
time of 10 to 100 ms (review4). The LCR period and the cycle and removal of RyR inactivation, both of which are required
length remain strongly coupled, not only in the steady state for subsequent spontaneous RyR activation. The growth of
beating, but also during stringent transitions, such as the tran- submembrane LCRs activates INCX (i.e., the electrogenic
sient state after removal of voltage-clamp at the maximum component of the coupled system), which drives DD to reach a
diastolic potential (MDP),22 and during intrinsic beat-to-beat threshold of LCCh activation that continues to further accelerate
variations in AP cycle length.32 DD, ultimately leading to the generation of rapid upstroke of
Although LCRs are indeed generated by SR, the LCR period the next AP. Of note, LCCh activation accelerates the DD of a
is not an exclusive property of the SR function per se, but rather given cycle not only directly via its inward current, but also
is the result of the mutual entrainment of both Ca2+- and M- indirectly via its Ca2+ influx (during DD, before the AP upstroke),
clocks regulating Ca2+ balance of the system (including Ca2+ avail- resulting in Ca2+-induced Ca2+ release and, hence, in additional
able for pumping by the SR) via multiple Ca2+-, voltage-, and LCRs and greater NCX activation, creating (for a short time)
enzyme-dependent processes (described in detail later). Thus, the a regenerative, positive-feedback loop. These chain reaction
LCR period characterizes the integration and kinetics of the types of interactions (revealed by numerical modeling5; see
clock-system, rather than just the function of a Ca2+-clock alone. Figure 25-2D) that become apparent during late DD may
These ideas have recently evolved into a concept of a coupled- be construed as being akin to a “rocket launch” (i.e., likely
clock pacemaker system4,5 featuring a “yin-yang”–type coexis- important for robust late DD acceleration and spontaneous AP
tence via interdependence33 (see Figure 25-2, D). generation).
Hence, during spontaneous AP firing, the SR does not oscil-
late Ca2+ independently, as it can in the absence of M-clock, but
rather operates as a “stop-watch” to generate a timely crucial
Symmetrical Importance of Ca2+- and LCR/INCX signal during DD that is likely further amplified by
M-Clocks: Clocks Coupling Guarantees Each the Ca2+-induced Ca2+ release. After each successful “rocket
Other’s Timely Existence and Confers Robustness launch” (i.e., AP ignition), SR function is reset by the AP. Thus,
to SANC Normal Automaticity in a given steady state, the electrochemical interactions of the
ensemble proteins of the system recur during each cycle (i.e.,
Although spontaneous Ca2+ release is intrinsically stochastic, its M-clock [or Ca2+-clock]), initiating Ca2+-clock (or M-clock); this
potentially deleterious, unlimited regenerative nature is quenched guarantees its own future periodic existence and thereby robust
by induction of an AP, thus shifting Ca2+ release properties from spontaneous AP firing. Thus, the coupled-clock concept and the
stochastic to periodic. The SR Ca2+ release periodicity is entrained symmetrical importance of both clocks have actually “married”
by rhythmically occurring APs via rhythmically triggered deple- the efforts of two earlier reductionist approaches to explain
tion of the SR. While the M-clock–generated AP initially cardiac pacemaker function based predominantly on either
depresses SR Ca2+ release by effecting global Ca2+ SR depletion, Hodgkin-Huxley–type ion current kinetics or on the periodic
this Ca2+ depletion simultaneously initiates the coupled-clock Ca2+ pumping-release paradigm of the SR function, igniting APs
restitution process: Ca2+ influx via LCCh restores Ca2+ lost from simply via an INCX electrogenic link.
local CaMKII activity and AC activities and LCRs. CaMKII

Beat-to-Beat Ca2+-Dependent Regulation of SAN inhibition reduces ICaL in isolated SANC, thereby reducing the
Pacemaker Cell Rate and Rhythm rate of AP firing.11 Increased CaMKII activity is also essential for
the increase in SR Ca2+ release in SANC in response to β-AR
If the Ca2+- and M-clocks were indeed coupled, as has been stimulation.37 BayK 8644 (a LCCh agonist) requires CaMKII
described, then Ca2+ changes must regulate the AP firing rate on activation to increase SANC AP firing.38 CaMKII is activated in
a beat-to-beat basis. Evidence for the role of Ca2+ in beat-to-beat rabbit SANC during AP firing in the basal state (likely as the
regulation stems from experiments in rabbit SANC in which result of active basal intracellular Ca2+ cycling).
spontaneous AP firing is interrupted by a voltage-clamp to The likelihood for spontaneous SR Ca2+ release to occur
−65 mV.22 As voltage-clamp is applied, the ensemble LCR signal increases as a function of the SR Ca2+ load, determined, in a
or LCR signal mass (a product of LCR amplitude, size, and dura- physiological context, by the Ca2+ that is available for pumping
tion) during the time interval, corresponding to the inter-AP into the SR and the phosphorylation status of the Ca2+ cycling
interval (i.e., “would-be” cycle length), remains as during spon- proteins, including phospholamban (which regulates the activity
taneous AP firing but then wanes after a few “would-be” cycles of SR Ca2+ pumping protein SERCA), RyR (the SR release
as SR Ca2+ becomes depleted in the absence of regularly occur- channel), and LCCh (see review39). SR Ca2+ refilling kinetics
ring APs. Upon removal of the voltage-clamp, spontaneous AP regulates the LCR period and the spontaneous beating rate of
firing is restored, and this produces Ca2+ influx for pumping into rabbit SANC.40 PKA- and CaMKII-dependent phosphorylation
the Ca2+-depleted SR. Beat-to-beat Ca2+-dependent regulation is of SR cycling proteins and LCCh leads to synchronization of
revealed during the transition process, as SR Ca2+ load increases: spontaneous RyR activation. The schematic in Figure 25-4A
The beat-to-beat increase in spontaneous AP firing rate closely depicts the concept that the restitution process that determines
follows the concomitant beat-to-beat increase in diastolic LCR the LCR period is regulated (1) by the kinetics of SR Ca2+ cycling
activity (assessed as the LCR signal mass22). (i.e., by the rate of Ca2+ pumping into the SR); and (2) by the
A more direct type of evidence for beat-to-beat Ca2+ depen- threshold of SR Ca2+ load required for spontaneous RyR activa-
dent regulation of SANC has been recently provided by acute tion. This concept is also supported by numerical model simula-
photolysis of the intracellular caged Ca2+ compound.34 In response tions of the SANC coupled-clock system (Figure 25-4B).
to loading of the caged Ca2+ buffer, the kinetics of Ca2+ removal Experimental data clearly support the critical importance of
from the cytosol is markedly slowed, the integral of Ca2+ release Ca2+-dependent, basal protein phosphorylation for normal SANC
in the form of LCRs becomes substantially reduced, and LCRs automaticity. Reducing the phosphorylation state of these pro-
become uncoupled from AP generation, which becomes mark- teins in rabbit or guinea pig SANC by buffering intracellular
edly slowed and then dysrhythmic. When Ca2+ is acutely released [Ca2+] or directly inhibiting AC, or basal PKA or CaMKII activi-
from the caged compound by flash photolysis, intracellular Ca2+ ties (including by cholinergic receptor [ChR] stimulation, see
dynamics are acutely restored, and rhythmic APs resume imme- below), has marked effects in slowing or even abolishing rhyth-
diately at a normal rate. Thus, when Ca2+ is photo-released, a small mic APs (see review4). And vice versa—increasing PKA signaling
increase in Ca2+ acutely affects SANC function: The initial effect of via stimulation of β-ARs, or exposure to a membrane-permeable
the flashes markedly accelerates DD and acutely reduces sponta- analog of cAMP, or inhibition of phosphodiesterase or phospha-
neous cycle length and synchronous LCRs. After a few rhythmic tase activity (see later) accelerates the AP firing rate. It is impor-
cycles, however, these effects of the flash wane as interference tant to note that independent of the type of perturbation of
with Ca2+ dynamics by the caged buffer is reestablished. protein phosphorylation, the LCR period always closely predicts
Another important piece of evidence for the beat-to-beat the cycle length (see Figure 25-3B). Maneuvers that increase
Ca2+-dependent regulation of SANC function is provided by protein phosphorylation require an intact Ca2+-clock to effect an
two-dimensional (2D)-camera measurements of the whole-cell increase in the spontaneous AP firing rate.23,41 Normal automatic-
ensemble of LCRs simultaneously with perforated-patch AP ity of SANC, in fact, can be envisioned as a phenomenon that
recordings.32 These experiments demonstrate that the beat-to- emerges from the phosphorylation state of the coupled-clock
beat variations in average LCR period (measured each cycle) of systems proteins; this synchronizes their functions to result in
the whole-cell LCR ensemble predict variations in AP cycle rhythmic spontaneous AP firing.
length.
Mechanisms Intrinsic to the Coupled-Clock System

That Restrain Its Basal AP Firing Rate
Basal Levels of cAMP and Protein
Phosphorylation by PKA and CaMKII Drive The coupled-pacemaker system dynamics are driven by a positive-
the Coupled-Clock System feedback Ca2+ signaling, that is, Ca2+release begets more Ca2+
release via PKA and CaMKII pathways (see Figure 25-1, green
SANC have a high constitutive level of activation of adenylyl arrows):
cyclases (ACs) that results in a high level of basal cAMP23 and
PKA-dependent protein phosphorylation. Although SANC, like Ca 2+ → Ca 2+ -activated ACs → cAMP → PKA
ventricular myocytes, express high levels of Ca2+-inhibited AC → phospholamban &RyR &LCCh phosphorylation
types 5 and 6, the discoveries of Ca2+-activated AC types (i.e., AC1 → more Ca 2+ → calmodulin → CaMKII
and AC8), in rabbit and guinea pig SANC35,36 and localization of → SERCA &phospholamban &RyR &LCCh
the basal Ca2+-activated AC activity within lipid raft microdo- phosphorylation → more Ca 2+ → etc.
mains36 link Ca2+ to localized cAMP production (see Figure 25-1).
Ca2+ binds to calmodulin to activate AC, leading to a high basal
level of cAMP-mediated, PKA-dependent phosphorylation of This positive-feedback signaling scheme obviously would drive
surface membrane and intracellular proteins involved in cell Ca2+ the AP firing rate to its maximum. Thus, mechanisms to damp
balance and SR Ca2+ cycling23,36 (see Figure 25-1). this signaling (see Figure 25-1, red arrows) must function continu-
Confocal imaging of immunolabeled proteins demonstrates ously and concurrently to keep in check cell cAMP level and the
that active CaMKII is highly localized beneath the surface mem- phosphorylation of molecules that drive AP firing. Basal phos-
brane.11 Thus, a geographical association has been noted between phodiesterase activity is one such restraining mechanism to
[Ca] in Various SR Ca2+ pumping rate, mM/s
25
junctional SR 30 Various
Action potential Various SR Ca 20
Ca2+ release
LCR-induced refilling rate 15
0.5 mM
depolarizations 12 threshold
10
1 2 3 7.5
5 3
MEMBRANE 2 1
POTENTIAL
Ca2+ release Late Diastolic
Global flux
Ca2+ transient LCR period Ca2+ Elevations
100 µM/ms
LOCAL (LDCAE)
SUB-SARCOLEMMAL LCR2 LCR3
LCR1
FREE CALCIUM
SR FREE CALCIUM
Same threshold 2 Various predicted LCR periods
1 3 INCX
for spontaneous
20 pA
Ca2+ release;
variable Ca2+
pumping rate
Diastolic
Variable SR Ca2+ refilling rate INCX
SR FREE CALCIUM
Action Cycle length Basal Cycle length
3 potentials decrease firing increase
Same Ca2+
pumping rate; Vm Late DD
variable release 2 acceleration Excitation
60 mV
threshold threshold
1
SR Ca2+ depletion
100 ms
A B
Figure 25-4. A, Schematic illustration of the concept of how the rate of SR Ca2+ refilling and the Ca2+ release threshold determine the LCR period, as well as the timing of
the LCR-induced DD.4 B, A new numerical of the SANC model5 predicts the wide range of pacemaker rate modulation via variations in SR Ca2+ pumping rate (1 to 30 mM/s),
mimicking various degrees of PKA-dependent phospholamban phosphorylation.
reduce the high constitutive AC activity in SANC.31 Such a the system’s ability to function as a clock (i.e., despite substantial
system of a high rate of cAMP production by Ca2+-activated ACs transient cell Ca2+ changes during each AP cycle, the steady state
is balanced to a high rate of cAMP degradation by phosphodies- average cell Ca2+ balance and levels of phosphorylation of M- and
terase to still maintain high basal levels of cAMP and, hence, basal Ca2+-clock proteins remain stable), so that all events recur during
phosphorylation of the signaling proteins of both clocks. Basal each cycle in the same sequence and magnitude, ensuring a stable
activation of phosphoprotein phosphatases in SANC42 is another AP cycle length.
mechanism that limits PKA- and CaMKII-dependent phosphor-
ylation. The net result of basal activation of phosphodiesterase
and phosphoprotein phosphatases is that the basal LCR period
and the AP cycle length are maintained near the midpoint of their Autonomic Receptor Modulation of the
functional ranges (see Figure 25-3A,B). Coupled-Clock System
Other restraining mechanisms of the system limit Ca2+ influx
and hence cell Ca2+ load, for example, calmodulin-mediated Nature imparts flexibility to the coupled clock of the cardiac
LCCh inactivation, which limits Ca2+ influx via LCCh during pacemaker to vary its AP firing rate over a wide physiological
each AP. It is interesting to note that membrane ion channels and range (i.e., from 60 to 240 bpm in humans) via variation in
transporters regulate Ca2+ balance of the coupled system not only GPCR signaling (see Figure 25-1, green and red shapes) that links
directly, by impacting on Ca2+ influx/efflux of LCCh/NCX, but both adrenergic and cholinergic receptors to the very same nodes
also indirectly, for example, by IK activation, which repolarizes of the coupled-clock system (see Figure 25-1, purple labels) that
AP, thereby limiting Ca2+ influx of LCCh and simultaneously regulate the basal state LCR period4: Stimulation of sympathetic
increasing Ca2+ efflux via voltage-dependent activation of the β-ARs in SANC, via Gs activation, increases the spontaneous AP
NCX forward mode. Even If activation regulates cell Ca2+ balance, firing rate via effects on the proteins of both clocks. It is well
by limiting the MDP, thereby limiting voltage-dependent Ca2+ established that β-AR stimulation can modulate ion channels of
efflux via NCX. the M-clock,1 specifically, IK and ICaL (via PKA-dependent phos-
Thus, the robustness of the coupled-clock system is imparted phorylation) and If (via direct effects of cAMP; see Figure 25-1).
by its numerous functional redundancies and by dual regulation However, β-AR stimulation, via changes in cAMP/PKA-depen-
of SR Ca2+cycling and membrane ion current generators by a dent phosphorylation due to AC activation and phosphatase inhi-
common set of factors (coupling factors or nodes [see Figure bition, can also alter SR Ca2+ efflux via RyR, as well as Ca2+
25-1, purple labels]) kept in check by phosphodiesterase and phos- sequestration via SERCA.14,20,41 Thus, β-AR stimulation modu-
phatase activities, voltage-dependent negative feedback on ion lates cAMP and protein phosphorylation levels of both clocks,
fluxes, and calmodulin-dependent modulation of LCCh. These simultaneously increases Ca2+ within the system via more fre-
complex interactions (depicted in Figure 25-1) are the essence of quent and larger ICaL activation, and accelerates refilling of SR
with Ca2+ to generate stronger LCRs with shorter LCR ICaL activation that further decelerates refilling of SR with Ca2+
periods23,40,41 that herald stronger INCX, thus accelerating DD.43 to generate weaker LCRs with longer LCR periods,44,45 decreas-
CaMKII-dependent phosphorylation of phospholamban and ing and delaying INCX activation and decelerating DD. Thus, the
RyR also increases in response to β-AR stimulation (likely as the feed-forward mechanism of the AP firing rate reduction rein-
result of changes in cell Ca2+ dynamics and concomitant calmod- forces the initial effect of ChR stimulation on cell Ca2+ balance,
ulin activation, effected primarily by PKA activation). Thus, which likely extends the regulation range and confers robustness
CaMKII-dependent phosphorylation likely has an important role to the ChR-mediated bradycardic effect.
as a Ca2+-calmodulin–dependent amplifier of the initial effect of The ideas of the Ca2+-clock and the coupled-clock system have
β-AR stimulation. served as the basis for a novel type of robust and flexible biologi-
ChR act as a physiological “break” on the heart rate. ChR cal pacemakers.46-48 One way to actually mimic the clock coupling
signaling activates Gi protein that couples to several downstream that naturally occurs in SANC is to express Ca2+-activated AC
targets: (1) Gi-βγ activation of IKACh (see Table 25-1; see review1 type 1 or 835,36 to activate proteins of both clocks via cAMP and
for details), and (2) a Gi-αs–coupled reduction in AC activity that cAMP-activated PKA-dependent phosphorylation.47,49 Another
limits cAMP/protein kinase A (PKA)-dependent phosphoryla- interesting approach is to reactivate within ventricular myocytes
tion. The lack of effect of pertussis toxin (PTX), which prevents the genetic program of the SAN.48
Gi activation, on basal AP firing rate indicates an absence of any
significant basal Gi activity that relates to AP firing. In isolated,
single rabbit SANC, the threshold for a cholinergic agonist, car-
bachol, to reduce the AP firing rate is ≈10 nM, and half maximal Mitochondrial Function in SANC
inhibition is achieved at 100 nM. IKACh activation becomes appar-
ent at higher concentrations of carbachol (>30 nM).44 A reduc- Rabbit SANC specifically had been referred to as “empty” cells
tion in AP firing rate in response to carbachol concentrations with sparse and scattered myofilaments and mitochondrial densi-
between 10 and 30 nM theoretically can be explained by a reduc- ties compared with atrial and ventricular cells. These conclusions,
tion in If via reduced cAMP, but stimulation remains intact in however, were based on rough-cut sections of the SAN visualized
If-deficient mice. Alternative powerful, cAMP-dependent mecha- by electron microscopy requiring a thin tissue slice, in which
nisms include a concurrent reduction in PKA-dependent phos- mitochondria can be damaged, and on the fact that artifacts can
phorylation of Ca2+ cycling proteins (e.g., phospholamban, be created, in part as the result of different cut angles, which can
RyR)44,45 and likely the M-clock (e.g., LCCh and IK channels). change the tissue morphology, giving rise to the impression of
Thus, in contrast to β-AR stimulation, ChR stimulation decreases “empty” space. Mitochondrial density visualized in isolated intact
the Ca2+ balance of the system, which leads to a reduction in AP rabbit SANC by specific mitochondrial labeling revealed the
firing rate. It is important to note that the transition to a new presence of abundant mitochondria and mitochondrial density in
steady state of the reduced AP firing rate is a feed-forward SANC that is similar to that in atrial and ventricular myocytes50;
process. Specifically, during this transition, each spontaneous hence, mitochondria are an important part of the coupled-clock
cycle is affected by prior cycles featuring less frequent and smaller system (Figure 25-5).
ATP ADP Na+

Ca2+
SR NCX
L type Ca2+ SERCA Ca2+ RyR Ca2+
Ca2+
ChR PLB ATP
Gi
PDE
Ca2+
AC ATP cAMP PKA
Myofilaments
Gs
AR Ca2+
ATP
Ca2+
If Uniporter Na/K K+
* NCX ATPase
ADP Na+
ADP
Na+
Mitochondria Ca2+ ATP
*Molecules involved in ATP production

Figure 25-5. Illustration of the interplay of AC- and cAMP/PKA-dependent signaling to the myofilaments, SR Ca2+-cycling proteins, ion channels, and basal state ATP demand
and supply in SANC.
The respiration or oxygen consumption rate in a physiologi- is limited), glycolysis may have an important role in ATP produc-
25
cally coupled mitochondrial system reflects the rate of mitochon- tion in SANC.
drial adenosine triphosphate (ATP) production. The respiration In ventricular myocytes, mitochondrial Ca2+ flux plays a fun-
rate of SANC is comparable with that of unloaded, ventricular damental role in buffering cytosolic Ca2+ and modulates the SR
myocytes that are electrically stimulated at 3Hz (i.e., similar to Ca2+ load in both normal and pathological conditions. Similar to
spontaneous SANC AP firing rate).50 Therefore, the ability of the ventricular myocytes, mitochondrial-SR Ca2+ crosstalk is present
mitochondria to produce ATP in SANC is comparable with that within SANC, and mitochondrial Ca2+ flux in SANC affects the
of ventricular myocytes, but “the energy budget” is spent differ- SR Ca2+ load, thus indirectly affecting SR Ca2+ release.56An
ently in the two cell types. The AP-initiated cytosolic Ca2+ tran- increase in mitochondrial Ca2+ attained by inhibition of mito-
sient delivers Ca2+ to the contractile proteins to induce chondrial NCX decreases the SR Ca2+ load and reduces the
myofilament force production and displacement, and myofila- ensemble LCR Ca2+ signal. In contrast, a reduction in mitochon-
ments are the major ATP consumers in ventricular myocytes. But drial Ca2+ by inhibition of mitochondrial uniporter increases the
SANC myofilament density is relatively low compared with that SR Ca2+ load and increases the ensemble LCR Ca2+ signal. Basal
of ventricular myocytes.51 In SANC, ATP consumption by spon- coupled-clock system automaticity is modulated by changes in
taneous myofilament contraction is lower than in ventricular Ca2+ cycling into and out of mitochondria in SANC, impacting
myocytes, and most ATP in SANC is consumed in the synthesis on cytosolic Ca2+, SR Ca2+ loading, and, indirectly, SR Ca2+
of cAMP and Ca2+-cycling maintenance of cell ionic homeostasis release characteristics. Changes in the spontaneous AP cycle
(see Figure 25-5). Short-term metabolic inhibition by cyanide, length in response to changes in mitochondrial Ca2+ are predicted
which inhibits cytochrome oxidase in the mitochondria, reduces by concurrent changes in the periodicity of the spontaneous
the intracellular Ca2+ transient amplitude and the AP firing rate rhythmic LCR Ca2+ signal. Because basal Ca2+ is also linked to
in primary pacemaker cells.50,52 cAMP-PKA signaling within SANC that regulates ATP supply/
Cytosolic Ca2+ enters mitochondria via the mitochondrial uni- demand matching, mitochondrial Ca2+-cytosolic Ca2+ crosstalk
porter and is extruded by the mitochondrial NCX (see Figure indirectly links to ATP supply and demand.
25-5). In ventricular myocytes, changes in mitochondrial Ca2+
alter the activity of several mitochondrial enzymes that take part
in ATP production (see review cf53). In SANC, however, reducing
Ca2+ influx into the mitochondria does not significantly alter the An Additional Level of Complexity of
ATP level. These results strongly suggest that mitochondrial Ca2+ Heart Pacemaker Function Arises Within
does not directly regulate ATP supply to demand. In contrast to the SAN Tissue
ventricular myocytes, Ca2+ activation of cAMP/PKA signaling in
SANC has a major role in linking basal ATP utilization and
mitochondrial ATP production. A gradual reduction in cAMP/ SAN tissue is a highly heterogeneous structure that is character-
PKA signaling is accompanied by a gradual reduction in ATP and ized by complex cell-to-cell interactions and extremely robust
a reduction in AP firing rate, suggesting that the Ca2+-activated impulse generation. It has been well established that electrical
cAMP/PKA signaling cascade that drives spontaneous AP firing activity of each beat is initiated within the SAN central area,
of SANC, as has been described, is a unique core feed-forward which contains mostly small SANC, and that the shape of the
system that not only drives basal ATP consumption but also membrane potential impulses varies between the SAN center and
regulates ATP production. cAMP/PKA signaling can effect ATP periphery.
production by phosphorylation of several mitochondrial proteins In the final analysis, a requirement for the supremacy of the
(e.g., voltage-dependent anion carrier [VDAC]) and complexes I central SAN area in initiating the SAN impulse is that the mutu-
through V (see review54). An increase in phosphorylation of these ally entrained rate of its cells, determined by the net balance of
complexes in the electron transport chain by cAMP/PKA intrinsic mechanisms and their modulation by extrinsic factors,
enhances the flux of protons and drives them to complex V to exceeds that of other neighborhoods to which the impulse
increase the rate of ATP production. Phosphorylation of VDAC spreads. A phase shift in mutually entrained APs, generated
can also increase the delivery rate of ATP from the mitochondria within a SAN neighborhood, must occur between neighborhoods
to the cytosol. Therefore, a decrease in basal cAMP/PKA signal- to account for the often observed shift in the leading pacemaker
ing in SANC appears to signal to mitochondria that ATP produc- site after interventions to alter intrinsic SANC properties (e.g.,
tion should be reduced. vagal or sympathetic stimulation [see review4]).
ATP can be produced in the cytosol by glycolysis. In SANC,
inhibition of glycolysis does not affect the basal spontaneous AP
firing rate.50,55 Moreover, inhibition of glycolysis in SANC does
not affect either ATP level or O2 consumption; therefore under Acknowledgments
basal conditions, glycolysis is not the major mechanism to
produce ATP in SANC. However, during high demand, or This research was supported by the Intramural Research Program
hypoxic conditions (i.e., when the amount of oxygen to the cell of the National Institutes of Health, National Institute on Aging.
strength of the heartbeat. J Pharmacol Sci 100:338– Am J Physiol Heart Circ Physiol 296:H594–H615,
References 369, 2006. 2009.
4. Lakatta EG, Maltsev VA, Vinogradova TM: A 6. Tsien RW, Kass RS, Weingart R: Cellular and sub-
1. Mangoni ME, Nargeot J: Genesis and regulation coupled SYSTEM of intracellular Ca2+ clocks and cellular mechanisms of cardiac pacemaker oscilla-
of the heart automaticity. Physiol Rev 88:919–982, surface membrane voltage clocks controls the time- tions. J Exp Biol 81:205–215, 1979.
2008. keeping mechanism of the heart’s pacemaker. Circ 7. Kass RS, Tsien RW: Fluctuations in membrane
2. Wilders R: Computer modelling of the sinoatrial Res 106:659–673, 2010. current driven by intracellular calcium in cardiac
node. Med Biol Eng Comput 45:189–207, 5. Maltsev VA, Lakatta EG: Synergism of coupled Purkinje fibers. Biophys J 38:259–269, 1982.
2007. subsarcolemmal Ca2+ clocks and sarcolemmal 8. Lipsius SL, Gibbons WR: Membrane currents,
3. Maltsev VA, Vinogradova TM, Lakatta EG: The voltage clocks confers robust and flexible pace- contractions, and aftercontractions in cardiac Pur-
emergence of a general theory of the initiation and maker function in a novel pacemaker cell model. kinje fibers. Am J Physiol 243:H77–H86, 1982.
9. Lakatta EG, Lappe DL: Diastolic scattered light rhythmicity in adult mice. Cardiovasc Res 96:372– heart’s pacemaker node cells by site-specific
fluctuation, resting force and twitch force in mam- 380, 2012. modulation of phospholamban phosphorylation
malian cardiac muscle. J Physiol 315:369–394, 27. Reppel M, Sasse P, Malan D, et al: Functional that regulates spontaneous subsarcolemmal local
1981. expression of the Na+/Ca2+ exchanger in the Ca2+ releases. Circulation 122:A21546, 2010.
10. Li J, Qu J, Nathan RD: Ionic basis of ryanodine’s embryonic mouse heart. J Mol Cell Cardiol Abstract.
negative chronotropic effect on pacemaker cells 42:121–132, 2007. 43. Maltsev VA, Lakatta EG: A novel quantitative
isolated from the sinoatrial node. Am J Physiol 28. Groenke S, Larson DL, Nakano H, et al: Atrial- explanation for autonomic modulation of cardiac
273:H2481–H2489, 1997. specific NCX KO mice reveal dependence of sino- pacemaker cell automaticity via a dynamic system
11. Vinogradova TM, Zhou YY, Bogdanov KY, et al: atrial node pacemaker activity on sodium-calcium of sarcolemmal and intracellular proteins. Am J
Sinoatrial node pacemaker activity requires Ca2+/ exchange. Biophys J 102:663a, 2012. Physiol Heart Circ Physiol 298:H2010–H2023,
calmodulin-dependent protein kinase II activation. 29. Maltsev VA, Vinogradova TM, Bogdanov KY, et 2010.
Circ Res 87:760–767, 2000. al: Diastolic calcium release controls the beating 44. Lyashkov AE, Vinogradova TM, Zahanich I, et al:
12. Sanders L, Rakovic S, Lowe M, et al: Fundamental rate of rabbit sinoatrial node cells: Numerical mod- Cholinergic receptor signaling modulates sponta-
importance of Na+-Ca2+ exchange for the pacemak- eling of the coupling process. Biophys J 86:2596– neous firing of sinoatrial nodal cells via integrated
ing mechanism in guinea-pig sino-atrial node. J 2605, 2004. effects on PKA-dependent Ca2+ cycling and IKACh.
Physiol 571:639–649, 2006. 30. Bogdanov KY, Maltsev VA, Vinogradova TM, Am J Physiol Heart Circ Physiol 297:H949–H959,
13. Rigg L, Terrar DA: Possible role of calcium release et al: Membrane potential fluctuations resulting 2009.
from the sarcoplasmic reticulum in pacemaking in from submembrane Ca2+ releases in rabbit sino- 45. van Borren MM, Verkerk AO, Wilders R, et al:
guinea-pig sino-atrial node. Exp Physiol 81:877– atrial nodal cells impart an exponential phase Effects of muscarinic receptor stimulation on Ca2+
880, 1996. to the late diastolic depolarization that controls transient, cAMP production and pacemaker fre-
14. Ju YK, Allen DG: How does beta-adrenergic stim- their chronotropic state. Circ Res 99:979–987, quency of rabbit sinoatrial node cells. Basic Res
ulation increase the heart rate? The role of intra- 2006. Cardiol 105:73–87, 2010.
cellular Ca2+ release in amphibian pacemaker cells. 31. Vinogradova TM, Sirenko S, Lyashkov AE, et al: 46. Zahanich I, Sirenko SG, Maltseva LA, et al: Rhyth-
J Physiol 516(Pt 3):793–804, 1999. Constitutive phosphodiesterase activity restricts mic beating of stem cell-derived cardiac cells
15. Rubenstein DS, Lipsius SL: Mechanisms of auto- spontaneous beating rate of cardiac pacemaker cells requires dynamic coupling of electrophysiology
maticity in subsidiary pacemakers from cat right by suppressing local Ca2+ releases. Circ Res and Ca cycling. J Mol Cell Cardiol 50:66–76,
atrium. Circ Res 64:648–657, 1989. 102:761–769, 2008. 2011.
16. Zhou Z, Lipsius SL: Na+-Ca2+ exchange current in 32. Monfredi OJ, Maltseva LA, Boyett MR, et al: Sto- 47. Boink GJ, Nearing BD, Shlapakova IN, et al: Ca2+-
latent pacemaker cells isolated from cat right chastic beat-to-beat variation in periodicity of local stimulated adenylyl cyclase AC1 generates efficient
atrium. J Physiol 466:263–285, 1993. calcium releases predicts intrinsic cycle length vari- biological pacing as single gene therapy and in
17. Huser J, Blatter LA, Lipsius SL: Intracellular Ca2+ ability in single sinoatrial node cells. Biophys J combination with HCN2. Circulation 126:528–
release contributes to automaticity in cat atrial 100:558a, 2011. Abstract. 536, 2012.
pacemaker cells. J Physiol 524(Pt 2):415–422, 33. Maltsev VA, Lakatta EG: The funny current in the 48. Kapoor N, Liang W, Marbán E, Cho HC: Direct
2000. context of the coupled clock pacemaker cell system. conversion of quiescent cardiomyocytes to pace-
18. Bogdanov KY, Vinogradova TM, Lakatta EG: Heart Rhythm 9:302–307, 2012. maker cells by expression of Tbx18. Nat Biotech-
Sinoatrial nodal cell ryanodine receptor and Na+- 34. Yaniv Y, Maltsev VA, Escobar AL, et al: Beat-to- nol 31:54–62, 2013.
Ca2+ exchanger: Molecular partners in pacemaker beat Ca2+-dependent regulation of sinoatrial nodal 49. Maltsev VA, Lakatta EG, Zahanich I, Sirenko SG:
regulation. Circ Res 88:1254–1258, 2001. pacemaker cell rate and rhythm. J Mol Cell Cardiol Engineered biological pacemakers (U.S. provi-
19. Ju YK, Allen DG: The distribution of calcium in 51:902–905, 2011. sional patent application no. 61/180,491). Federal
toad cardiac pacemaker cells during spontaneous 35. Mattick P, Parrington J, Odia E, et al: Ca2+-stimu- Register 74:53268, 2009.
firing. Pflugers Arch 441:219–227, 2000. lated adenylyl cyclase isoform AC1 is preferentially 50. Yaniv Y, Juhaszova M, Lyashkov AE, et al: Ca2+-
20. Rigg L, Heath BM, Cui Y, et al: Localisation and expressed in guinea-pig sino-atrial node cells and regulated-cAMP/PKA signaling in cardiac pace-
functional significance of ryanodine receptors modulates the If pacemaker current. J Physiol maker cells links ATP supply to demand. J Mol
during beta-adrenoceptor stimulation in the 582:1195–1203, 2007. Cell Cardiol 51:740–748, 2011.
guinea-pig sino-atrial node. Cardiovasc Res 36. Younes A, Lyashkov AE, Graham D, et al: Ca2+- 51. Boyett MR, Honjo H, Kodama I: The sinoatrial
48:254–264, 2000. stimulated basal adenylyl cyclase activity localiza- node, a heterogeneous pacemaker structure. Car-
21. Lyashkov AE, Juhaszova M, Dobrzynski H, et al: tion in membrane lipid microdomains of cardiac diovasc Res 47:658–687, 2000.
Calcium cycling protein density and functional sinoatrial nodal pacemaker cells. J Biol Chem 52. Ju YK, Allen DG: Early effects of metabolic inhibi-
importance to automaticity of isolated sinoatrial 283:14461–14468, 2008. tion on intracellular Ca2+ in toad pacemaker cells:
nodal cells are independent of cell size. Circ Res 37. Wu Y, Gao Z, Chen B, et al: Calmodulin kinase II Involvement of Ca2+ stores. Am J Physiol Heart
100:1723–1731, 2007. is required for fight or flight sinoatrial node physi- Circ Physiol 284:H1087–H1094, 2003.
22. Vinogradova TM, Zhou YY, Maltsev V, et al: ology. Proc Natl Acad Sci U S A 106:5972–5977, 53. Yaniv Y, Juhaszova M, Nuss HB, et al: Matching
Rhythmic ryanodine receptor Ca2+ releases during 2009. ATP supply and demand in mammalian heart: In
diastolic depolarization of sinoatrial pacemaker 38. Gao Z, Singh MV, Hall DD, et al: Catecholamine- vivo, in vitro, and in silico perspectives. Ann N Y
cells do not require membrane depolarization. Circ independent heart rate increases require Ca2+/ Acad Sci 1188:133–142, 2010.
Res 94:802–809, 2004. calmodulin-dependent protein kinase II. Circ 54. Covian R, Balaban RS: Cardiac mitochondrial
23. Vinogradova TM, Lyashkov AE, Zhu W, et al: Arrhythm Electrophysiol 4:379–387, 2011. matrix and respiratory complex protein phosphory-
High basal protein kinase A-dependent phosphory- 39. Lakatta EG: Functional implications of spontane- lation. Am J Physiol Heart Circ Physiol 303:H940–
lation drives rhythmic internal Ca2+ store oscilla- ous sarcoplasmic reticulum Ca2+ release in the H966, 2012.
tions and spontaneous beating of cardiac pacemaker heart. Cardiovasc Res 26:193–214, 1992. 55. Senges J, Brachmann J, Pelzer D, et al: Effect of
cells. Circ Res 98:505–514, 2006. 40. Vinogradova TM, Brochet DX, Sirenko S, et al: glycolytic inhibitors on the sinoatrial node, atrium
24. Yang HT, Tweedie D, Wang S, et al: The ryano- Sarcoplasmic reticulum Ca2+ pumping kinetics and atrioventricular node in the rabbit heart. J Mol
dine receptor modulates the spontaneous beating regulates timing of local Ca2+ releases and sponta- Cell Cardiol 13:811–821, 1981.
rate of cardiomyocytes during development. Proc neous beating rate of rabbit sinoatrial node pace- 56. Yaniv Y, Spurgeon HA, Lyashkov AE, et al: Cross-
Natl Acad Sci U S A 99:9225–9230, 2002. maker cells. Circ Res 107:767–775, 2010. talk between mitochondrial and sarcoplasmic
25. Fu JD, Li J, Tweedie D, et al: Crucial role of the 41. Vinogradova TM, Bogdanov KY, Lakatta EG: reticulum Ca2+ cycling modulates cardiac pace-
sarcoplasmic reticulum in the developmental regu- Beta-adrenergic stimulation modulates ryanodine maker cell automaticity. PLoS One 7:e37582,
lation of Ca2+ transients and contraction in cardio- receptor Ca2+ release during diastolic depolariza- 2012.
myocytes derived from embryonic stem cells. tion to accelerate pacemaker activity in rabbit sino- 57. Lakatta EG, Maltsev VA, Bogdanov KY, et al:
FASEB J 20:181–183, 2006. atrial nodal cells. Circ Res 90:73–79, 2002. Cyclic variation of intracellular calcium: A critical
26. Bround MJ, Asghari P, Wambolt R, et al: Cardiac 42. Zahanich I, Li Y, Lyashkov AE, et al: Protein phos- factor for cardiac pacemaker cell dominance. Circ
ryanodine receptors control heart rate and phatase 1 regulates normal automaticity of the Res 92:e45–e50, 2003.
Biological Pacing 26
Michael R. Rosen, Peter R. Brink, Ira S. Cohen, and Richard B. Robinson
CHAPTER OUTLINE The Normal Cardiac Pacemaker

The Natural Pacemaker 253 The mammalian sinoatrial node was identified structurally in
Strategies, Successes, and Failures in Building 1907 by Keith and Flack,5 and its function as the primary source
Biological Pacemakers 255 of the cardiac impulse was demonstrated electrophysiologically
by Lewis et al in 1910.6 These investigators recognized that they
Challenges 258 were dealing with a heterogeneous structure that appeared to
Conclusions 261 serve as the dominant site of origin of the normal heartbeat. Yet
the mechanisms determining sinoatrial node function remained
a mystery for decades. One school of thought held that the
node—and indeed the entire cardiac conduction system—
consisted of specialized neural fibers, with a neural impulse ini-
The Natural Pacemaker tiating the heartbeat.7 That sinus rhythm persisted after cardiac
denervation8 argued against central neural origin, although a role
The natural pacemaker of the heart, the sinoatrial node (the true for local neuronal activity as the cause of the heartbeat was a
“biological pacemaker”), is a complex structure whose anatomy continuing belief.7
facilitates the transmission of impulses to the rest of the heart The modern era of molecular biophysics has determined the
while protecting it from excitation by impulses arising elsewhere. why and the wherefore of the origin of cardiac impulses, although
Its unique ion channel complement initiates impulses rhythmi- the origin of the sinoatrial node cells themselves remains a subject
cally throughout the life cycle, while its autonomic neural supply of active investigation.9 As is shown in Figure 26-1, A, multiple
ensures ready adjustment to physiological demands for altering currents contribute to the sinoatrial node action potential.10-13
heart rate. That the resting membrane potentials of normal sinoatrial node
Like any biological system, the sinoatrial node is affected by cells are more depolarized than those of most other cardiac myo-
aging and pathology in ways that can lead to dysfunction. For cytes is the result of a small inward rectifier current, IK1. Action
much of human history, this dysfunction, expressed as sinoatrial potentials initiated at these voltages have calcium rather than
arrest or block or atrioventricular (AV) block, was accompanied sodium as the inward charge carrier. The result is a slowly propa-
by syncope, a marginal quality of life, and—over varying time gating impulse similar to that in the normal AV node, the AV
spans—death.1,2 The major therapy until the mid-20th century valves, and the coronary sinus.
was a stopgap: ephedrine or sublingual isoproterenol, adminis- As sinoatrial node cells repolarize to their maximum diastolic
tered every 2 hours. potentials, their hyperpolarization permits the opening of HCN
In the 1960s, electronic pacing became available as devices (for Hyperpolarization-activated Cyclic Nucleotide gated)
that could be implanted transvenously and with little risk.3,4 Early channels, through which an inward sodium current, referred to
units were relatively massive—rather like hockey pucks in as If, flows and begins to depolarize the membrane (Figure
appearance—but they were lifesaving. Technology improved 26-1, A, B).10-12 A major contributor to pacemaker activity is the
and the size of units diminished over the second half of the calcium clock mechanism, which depends on calcium cycling
20th century; improvement and innovation continue in the between intracellular uptake and release sites, as well as calcium
current era. entry via L- and T-type Ca channels.13,14 This affects the activity
However, electronic pacing is not perfect: Although it restores of the Na/Ca exchanger, which directly modulates phase 4 depo-
individuals to their lives, their families, and their society, it is not larization.13,14 All this activity occurs against the backdrop of
the seamlessly functioning structure exemplified by the sinoatrial outward (repolarizing) potassium current, such that any increase
node. And this has led investigators to explore gene and cell in inward current and/or decrease in outward current will increase
therapies that replace or mimic the sinoatrial node as possible the slope of phase 4 depolarization and the pacemaker rate.
alternatives to electronic pacing. The intent is to replace the The sympathetic and parasympathetic nervous systems are
function of the sinoatrial node, albeit not its structure. the principal modulators of phase 4 depolarization. Their
Seen in this light, cardiac pacemaking appears to be coming respective neurohumors, norepinephrine and acetylcholine, act
full circle—from biological to electronic and back to biological— via β-adrenergic or muscarinic receptor-G protein-linked path-
but success is not yet within our grasp. To provide a perspective ways to increase (β-adrenergic) or decrease (muscarinic) cyclic
and a view of the future, we will first discuss the mechanisms adenosine monophosphate (cAMP) synthesis (see Figure 26-1, B).
underlying and the pathologies influencing sinoatrial node func- Cyclic AMP binding to a site near the carboxy terminus of the
tion and the rationale for developing biological pacemakers. We HCN channel shifts channel activation positively, resulting in an
will then consider strategies for building biological pacemakers increase in phase 4 depolarization.12
and successes and failures to date. We will finish with a discussion Whereas all the currents described contribute to phase 4
of challenges for the future. depolarization and to pacemaker rate, the initiator of the process
appears to be If.10,11 That all pacemaker activity is not attributable
to If has been demonstrated in experiments using If-blocking
These studies were supported by USPHS-NHLBI grant HL-28958. drugs. In both experimental models and the clinic, the result is
253
0 500 ms
mV
50
pA If
50
50
IK
ICaL
50
ICaT
50
INa/Ca
50
Na

HCN channel (If)
P
1–3
4 5 6
cAMP
G Gi () () GB G
M2 AC 1-AR
ACh NE
B
Figure 26-1. A, The sinoatrial node action potential and the currents contributing to impulse generation. B, The HCN channel. See text for discussion.
(Reproduced and modified with permission from DiFrancesco D, Camm AJ: Heart rate lowering by specific and selective lf current inhibition with ivabradine: A new therapeutic
perspective in cardiovascular disease. Drugs 64:1757–1765, 2004; and Biel M, Schneider A, Wahl C: Cardiac HCN channels: Structure, function, and modulation. Trends Cardiovasc
Med 12:202–216, 2002.)
significant slowing of sinus rate but not termination of

the rhythm.15,16 Indeed, the inability to suppress sinus node Pathologies Influencing Normal Pacemaker
function using any one blocker highlights the redundancy within Function and a Brief History of Therapies
the system such that it continues to function even under
challenge. In 1827 and 1846, respectively, Adams1 and Stokes2 noted the
The HCN channels that initiate pacemaker function have 6 clinical characteristics of an event that had likely afflicted mankind
transmembrane-spanning domains, are ubiquitous in the heart, for centuries: They described patients who became pale, whose
and are found elsewhere in the body as well (see Figure 26-1, B). pulse became slow or was absent, and who then collapsed. We
Four channel isoforms, HCN1-4, are present, with 4 and 1 in the now recognize that the pathology was a high-degree heart block.
sinoatrial node, 2 in much of the conducting system and the In the 20th century, the availability of synthetic catecholamines
myocardium, and 3 largely in neural tissues.12 Each has differing and later of β-agonists afforded one form of treatment: sublingual
activation and deactivation kinetics and differing cAMP sensitiv- isoproterenol every 2 hours.17 Although this therapy increased
ity. The critical aspects of these channels with regard to pace- the rate of idioventricular pacemakers to a more physiological
maker activity include their activation on hyperpolarization and range in many patients, it also could and did cause ventricular
their modulation by autonomic neurohumors. tachycardia. Moreover, this stopgap therapy could not ensure any
To summarize, the primary biological pacemaker is structur- long-term survival.
ally and functionally complex. In settings of disordered sinoatrial The answer to this dilemma faced by patients and physicians
node impulse initiation or propagation that interfere with normal was the electronic pacemaker. Discovery of this device is traced
cardiac function, one has the choice of attempting to remake or to the 1889 report of McWilliam, who described a means of
repair the sinoatrial node—a difficult and as yet impossible task— delivering shocks at about 60 to 70 times per minute to a patient
or to reproduce aspects of its function. The overall goals are to in whom the heartbeat had failed.18 In 1928, Lidwell and Booth
restore the quality of life and to prolong life itself. used an electrical power source attached to a needle plunged into
the heart to revive “a stillborn infant.”19 The term cardiac pace- the phasing out of electronic pacing. Therefore, biological pacing
26
maker is attributed to Hyman, who used a hand-cranked device represents a disruptive technology.
to generate electrical shocks in 1932.20 However, the modern era
in pacing commenced in the early 1950s, when Hopps21 in
Canada and then Zoll22 in the United States reported devices that
delivered transcutaneous shocks. Pacing via shocks delivered Strategies, Successes, and Failures
through an intramyocardial needle in the setting of postsurgical in Building Biological Pacemakers
complete heart block was reported by Weirich et al in 195723—
the same year in which Bakken fabricated the first portable exter- In the late 1980s, biological pacing was a pipe dream occasionally
nal pacemaker.24 discussed around coffee tables among investigators working on
Initially, pacemaker implantation in the human heart required mechanisms of impulse initiation. These discussions were likely
a thoracotomy for intramyocardial electrode placement.4 The stimulated by Dario DiFrancesco’s discovery of the pacemaker
first report of temporary transvenous approaches was published current If,10,11 and usually began with statements like, “Wouldn’t
by Furman and Schwedel in 1959.3 The early 1960s then saw the it be wonderful if we could put pacemaker currents into diseased
rapid adoption of transvenous pacing with implanted units. Early sinus nodes or ventricles and create a site of normal impulse
pacemakers had no sensing function, paced at a fixed rate, and, initiation?” Surely, this was the stuff of science fiction, but as
although internally implanted, were cumbersome and required we’ve learned time and again, science fiction often explores desir-
frequent power pack changes because of their use of mercury able outcomes before the needed technology has been developed;
batteries. Moreover, their failure to sense often led to competi- the works of Jules Verne are a prime example.
tion between implanted and idioventricular pacemakers and, in By the late 1990s, advances in genetic manipulation and stem
far too many instances, arrhythmogenesis. But the 1960s and the cell science brought the concept of biological pacing to the realm
1970s saw rapid improvement in batteries, electrodes, and soft- of the possible, as we and others ventured to state in lectures and
ware, such that pacemakers could be placed in a demand mode, in print rather than simply at the coffee table. Several templates
sensing spontaneously occurring heartbeats and re-setting for exploration of biological pacing are suggested by Figure 26-1.
appropriately.4,24,25 For example, β-adrenergic stimulation can modulate If to increase
Advances continued through the 20th century and into the pacemaker rate, and an initial strategy for biological pacing was
current decade, bringing AV sequential pacing that synchronized to overexpress β2-adrenergic receptors in murine, then pig,
the sequence of atrial and ventricular activation in patients having hearts.28,29 Injecting a naked plasmid while incorporating the β2
normal sinus node function and AV block; exploration of the receptor into pig atria led to increased basal rate and catechol-
epicardial veins of the heart as sites for biventricular pacing to amine response. Although overexpressing β-adrenergic receptors
improve cardiac output; and the evolution of cardioverter- augmented sinus node function and increased the rate of normal
defibrillators that could sense, shock, and then pace if needed, for rhythm, the following concerns arose: (1) In a diseased heart in
those patients who had potentially lethal tachyarrhythmias. which the sinus node is not functioning well and/or AV block is
Innovation continues through the present time. For example, present, β-receptor overexpression can be arrhythmogenic;
units are being tested that vary heart rate according to the body’s (2) upregulating β receptors implies that pacemaker cells already
physiological needs,26 and leadless electrodes are being developed in the heart will respond normally to increased catecholaminergic
that hold the promise of stimulating the heart without incorpo- effect—which may or may not be the case; and (3) transfection
rating a catheter.27 So the history of electronic pacing has been using naked plasmids has marginal success in the heart in situ.
meteoric, and its future is bright. Hence, although early studies provided proof of concept for
biological pacing, the strategy was not one that could be advanced.
Subsequent strategies have depended on the manipulation of ion
Rationale for Developing Biological Pacemakers channels delivered via viral vectors or stem cell platforms, or on
the properties of channels resident in stem cells or other cell
Given the previous description of electronic pacemakers, why types, to perform biological pacing.
bother with biological pacing—or indeed with any other
approach? Simply because the electronic pacemaker, although a
wonderful treatment, is a panacea, not a cure. Table 26-1 sum- Viral Vector-Based Delivery of Constructs
marizes the reasons why biological pacemakers are being inves-
tigated to normalize cardiac rhythm in a way that more nearly As can be seen in Figure 26-1, A, the initial ion channel-based
approximates the physiological state when compared with elec- approach reported transfection of guinea pig ventricles with an
tronic pacemakers. If biological pacing succeeds, the two tech- adenoviral vector incorporating a dominant negative Kir2.1 con-
nologies would likely be used in tandem for some time to struct to reduce the repolarizing current, IK1.30 As is shown on
maximize patient safety. But the eventual goal would be to see electrocardiogram (ECG) in the intact heart and with microelec-
trode and ion current recordings, an ectopic ventricular rhythm
was expressed that was associated with phase 4 depolarization and
reduced IK1 in myocytes. A shortcoming was that reducing repo-
larizing current increases action potential duration, and a subse-
Table 26-1. Limitations of Cardiac Pacemakers quent publication showed that the repolarization characteristics
• Not responsive to the autonomic nervous system—to the demands of Anderson’s syndrome were created by this approach to biologi-
of exercise and emotion cal pacing,31 such that its continued development was
• Require monitoring and maintenance, including at times battery undesirable.
and/or electrode replacement Our group has worked primarily with means to modify and
• Not optimal for pediatric patients magnify the pacemaker current, If, as originally reported by
• Problems with infection, interference from other devices DiFrancesco.10,11 We first demonstrated in neonatal rat myocytes
• Often cannot implant at sites that optimize contraction in in cell culture that using an adenoviral vector to overexpress the
individual patients HCN2 channel isoform resulted in a robust and rapid pacemaker
• Palliate but do not cure that far exceeded in rate and stability the rhythm normally
recorded in non-transfected cultures.32,33 For this reason, we
adopted the strategy of using HCN genes to build biological For the 2 weeks during which stable HCN2 expression
pacemakers and centered the approach on HCN2.34 An addi- occurred, about 70% of beats originated from the injection site.
tional attraction of working with HCN channels was that their Basal rates averaged 50 to 60 beats/minute (bpm).36 The remain-
inward current flows only during diastole; hence, although we ing 30% of beats were provided by the demand electronic pace-
were increasing an inward current, this would not prolong action maker, set at an escape rate of 45 bpm. A control group of
potential duration, thereby avoiding the problem that had saline-injected animals with electronic pacemakers, also set at
afflicted IK1 downregulation. 45 bpm, had 90% of beats electrically stimulated, with the
Our initial in vivo study involved injecting adenoviral HCN2 remainder arising from various idioventricular sites. Finally, the
and green fluorescent protein (GFP) constructs into canine left biological pacemaker manifested an approximate 50% rate
atria, and several days later using vagal stimulation to suppress increase in response to catecholamine injection. Taking this
sinus rhythm.34 The result consisted of escape beats originating together with our earlier report,34 we concluded that both vagal
from the implant site. Cells disaggregated from this site showed and sympathetic (or at least β-adrenergic) modulation of biologi-
If (actually IHCN2) about 100 times greater in magnitude than that cal pacing was possible.
in native atrial myocytes. Additional vagal stimulation terminated
firing of this pacemaker locus, consistent with parasympathetic
control. Optimizing Pacemaker Gene Constructs
A follow-up study used a custom-designed electrode catheter
to locate the left bundle branch system in dogs and to inject the Attempts are ongoing to replicate or improve on the function of
same adenoviral-HCN2 complex into the bundle branch.35 When HCN2 using designer-based K channel genes,37 mutant or chi-
the vagi were stimulated to terminate sinus rhythm and induce meric HCN channels,36,38,39 and combinations of two channels.
AV block, a rhythm emerged from the injection site. Microelec- We first studied a mutant HCN2 (E324A) that showed some
trode studies demonstrated spontaneous and rapid impulse initia- enhanced catecholamine responsiveness, but only modest
tion at these bundle branch sites. Subsequently, we found that in improvement over wild type HCN2.36 We then designed a
dogs with radiofrequency-induced complete heart block, a bio- chimera, HCN212, whose transmembrane portion included the
logical pacemaker inserted into the left bundle branch system pore of HCN1 (which has more positive activation than HCN2)
drove the heart regularly and stably.36 These rhythms were and the amino- and carboxy-termini (the latter incorporating the
mapped to their bundle branch origins (Figure 26-2). cAMP-binding site) of HCN2.39 We hypothesized that this
Figure 26-2. Noncontact mapping of left ventricle (LV) with CARTO system in dog with complete heart block, an electronic right ventricular (RV) apical endocardial pace-
maker, and an HCN2 adenoviral construct administered into the proximal left bundle branch. Panels demonstrate four projections, showing early (red) through late (blue)
activation. Upper panels show an impulse activating LV endocardium at several sites simultaneously, reflecting arrival of an impulse initiated in the left bundle branch
system. Lower panels show early activation of LV septum via electronic pacemaker.
channel would generate faster basal pacemaker rates than HCN2 manifest typical pacemaker characteristics, including action
26
while exhibiting greater autonomic responsiveness than HCN1 potentials that show prominent phase 4 depolarization.47 The
alone. However, HCN212 resulted in an overshoot, such that underlying populations of ion channels include a very weak IK1
ventricular tachycardias having rates in excess of 200 bpm and a robust If, which together would tend to generate depolar-
occurred. Although this result was unfortunate with regard to ization during phase 4.48 Unlike in a sinoatrial node, however, the
construct design, it permitted us to test whether If blockade might major contributor of inward current is reportedly sodium, rather
be useful in suppressing arrhythmias generated by biological than L-type Ca current.48
pacemakers. The If blocker ivabradine was highly effective here.39 Clusters of these cells in culture have been shown to originate
Another mutant gene was reported by Kashiwakura et al,37 and propagate spontaneous rhythms, and, when implanted into
who used a K channel engineered to carry inward sodium current the ventricles of pigs in complete heart block, they have initiated
that had many of the properties of If. The shortcoming was the stable spontaneous rhythms that persist for up to 3 months.47
absence of a cAMP-binding site, without which there is no auto- Major issues regarding this approach have been and remain the
nomic modulation. Although these and other constructs so far need to use immunosuppression and questions regarding the
developed have not offered advantages over the wild type, it is eventual fate of the cells (e.g., Will they persist as pacemaker
anticipated that continued efforts at discovery will lead to robust cells? Will they mature into nonbeating ventricular myocytes or
alternatives. into other inimical cell types?).
In contrast to this, combinations of two genes have been Other attempts to work with cells containing full ion channel
promising. One such approach uses the dominant negative Kir2.1 complements have included the grafting of fetal49 or neonatal50
construct previously described.40 This was tested in an adenoviral pacemaker cells. These research directions have not yet been
vector together with an HCN gene to increase inward current. published in sufficient detail to permit objective evaluation of the
The intent was to decrease hyperpolarizing current and increase possibilities they offer. An alternative cell therapy approach has
inward current simultaneously in a small region of the ventricle. involved the generation of induced pluripotent stem cells (IPSCs),
This was done successfully in pigs40 and is being explored by initially from dermal fibroblasts51 and more recently from kera-
investigators as a potential treatment in settings where electronic tinocytes.52 At the cellular level, good pacemaker function has
pacemakers have become infected, necessitating hardware extrac- been demonstrated. The advantage of IPSCs is that they permit
tion from the heart. The concern here is that the adenoviral pacemaker cells to be developed that have the full complement
vector, which itself can cause inflammation, may contribute to of pacemaker genes, such as the embryonic stem cell. An advan-
the inflammatory process in the heart. Therefore, it is likely that tage over the embryonic stem cell is that IPSCs can be used
this approach will be tested in a large animal model of endocar- autologously, thereby creating no immune response. Questions
ditis before the therapy is advanced. remain regarding the possibility of tumor development, although
Another approach uses HCN2 together with the gene for the the oncogenes originally used to create IPSCs no longer need be
skeletal muscle sodium channel, SkM1.41,42 SkM1 not only used.53,54 In addition, questions continue regarding whether
improves propagation of the cardiac impulse at depolarized mem- autologous stem cells developed from tissues of aging populations
brane potentials,43,44 it also moves threshold potential to more (such as most patients requiring pacemakers) will be as robust as
negative voltages.42 The net result is that impulse initiation those from young individuals. Animal studies with biological
occurs earlier in phase 4, speeding pacemaker rate. When tested pacemakers developed from IPSCs are awaited.
in a canine model of complete heart block, the outcome has been The second broad approach to cell therapy uses cell types that
basal rates of approximately 80 bpm, with responses to catechol- do not have the proper ion channel population to generate pace-
amine or exercise of up to 140 to 150 bpm. In addition, in this maker function but have other characteristics that make them
setting, there has been no need for the backup electronic pace- candidates for platform therapy, that is, they can be loaded with a
maker to fire.41 gene or genes of interest that result in delivery of a pacemaker
A different attempt to improve function uses the Ca2+-stimu- signal to adjacent cells. Although some work has been reported
lated adenylyl cyclase AC1 gene expressed alone or in combina- using fibroblasts as platforms and fusing them with ventricular
tion with HCN2.45,46 Basal beating rates were excessively myocytes,55 the more completely reported approach to date has
high—around 140 bpm with HCN2/AC1, and only in the 50 to been our own—using adult human mesenchymal stem cells
65-bpm range with AC1 alone. Escape times in both groups were (hMSCs).56-58 Gene chip analysis of hMSCs has shown that they
within the 1- to 2-second range following overdrive pacing, and have no appreciable signal for the pacemaker genes, HCN1-4, but
the percent of electronic beats was in the 2% to 7% range. they do have a robust signal for the cardiac connexins, Cx40 and
Instantaneous and long-term heart rate variability and circadian Cx43.57 In deciding to work with hMSCs as biological pacemak-
rhythm on 24-hour recordings showed greater sensitivity to para- ers, we hypothesized that if we could load them with a pacemaker
sympathetic modulation in animals injected with AC1 and a high gene, they might couple adequately to myocytes such that the
degree of sympathetic modulation in those injected with HCN2/ depolarizing signal generated by If would be transmitted to adja-
AC1. In vitro and in vivo data suggested that enhancement not cent myocytes, causing them to depolarize (Figure 26-3).59-61
only of If but of other, Ca-based pacemaker mechanisms occurs Using both dye transfer and current injection experiments in cell
with AC1, and this is a matter of concern.46 But given the robust pairs, we showed that hMSCs could indeed communicate with
autonomic response to this intervention, its continued explora- myocytes.57 This prompted us to load them with HCN2 via elec-
tion is warranted, although better control of basal rates will be troporation (avoiding any viral vector) (Figure 26-4, A-C) and to
required. implant them in canine left ventricle.56,58 In an initial study, we
used vagal stimulation to suppress sinoatrial node impulse initia-
tion and AV conduction. Here, an escape rhythm that pace-
Cell Therapies mapped to the injection site was generated by the hMSC-based
pacemakers (Figure 26-4, D). We also demonstrated connexins at
Two differing philosophies have provided the framework for interstices between hMSCs and cardiac myocytes (Figure 26-4, E).
using cell therapies in biological pacing. The first uses cells that In a subsequent study of dogs in complete heart block, hMSCs
have a normal complement of ion channel genes to develop a generated pacemaker function for at least 6 weeks when 700,000
biological pacemaker. The most advanced therapy reported uses or more cells, approximately 50% of which were loaded with
human embryonic stem cells that have been “forced” along a HCN2, were implanted at a single left ventricular site.58 Longer-
cardiogenic lineage.47 Within this lineage a subset of cells term follow-up has revealed an issue of major concern, that is, by
Rationale for Stem Cell Based Pacemaker

Natural pacemaker
Challenges
Some of the challenges we will discuss are unique to biological
HCN pacing; others hold for a variety of other approaches to gene and
cell therapy. A major challenge is to identify the optimal con-
struct as the groups working in this area continue attempts at
optimizing mutants and chimeras. HCN2 alone manifests prop-
SA node, native pacemaker Atrial myocyte
erties that likely are adequate for first-generation biological
A Gap junction pacing. However, it might be preferable to have a construct that
maintains basal rates in the 60 to 70-bpm range and reaches peaks
hMSC as platform for a biological pacemaker of 120 to 140 bpm on catecholamine stimulation, as has been
reported for the HCN2/SkM1 combination.41,42
Although we have demonstrated that vagal and β-adrenergic
HCN responsiveness is a characteristic of HCN2-based biological
pacemakers, further understanding of autonomic control is
required. Use of HCN2 in both viral and hMSC constructs has
shown that vagal control depends on the site at which the pace-
maker is implanted,67 that is, in vagally innervated regions, Poin-
Stem cell Myocyte caré plots of heart rate variability are consistent with
vago-sympathetic input. In contrast, on the lateral free wall of
B Gap junction
the left ventricle, where little to no vagal innervation has been
Figure 26-3. Initiation of spontaneous rhythms by wild type pacemaker cells and noted, no evidence of vagal input has been found, and only sym-
by genetically engineered stem cell pacemakers. A, In a native pacemaker cell (or pathetic responsiveness is seen.
in a myocyte engineered to incorporate pacemaker current via gene transfer), Preclinical and clinical testing of biological pacing is also a
action potentials (inset) are initiated via inward current flowing through transmem- major challenge. At the preclinical level, questions that
brane HCN channels. These open when the membrane repolarizes to its maximum
arise relate largely to safety, efficacy, and duration of action.
diastolic potential and close when the membrane has depolarized during the
action potential. Current flowing via gap junctions to adjacent myocytes results in
Another challenge involves effective utilization of the relation-
their excitation and the propagation of impulses through the conducting system. ship between biological and electronic pacing. When and if
B, A stem cell has been engineered to incorporate HCN channels in its membrane. biological pacing is ready for clinical trials, it cannot simply be
These channels can open, and current can flow through them (inset) only when administered to human subjects in need of support of their
the membrane is hyperpolarized; such hyperpolarization can be delivered only if ventricular rate. Rather, it will have to be used together with
an adjacent myocyte is tightly coupled to the stem cell via gap junctions. In the electronic pacemakers, because these represent the current
presence of such coupling and opening of the HCN channels to induce local state of the art. For this reason, we have tested integration of
current flow, the adjacent myocyte will be excited and will initiate an action poten- biological and electronic pacemakers in tandem operation.36 We
tial that then propagates through the conducting system. Depolarization of the
found that tandem biological-electronic pacemaker therapy man-
action potential will result in closing of the HCN channels until the next repolariza-
tion restores a high negative membrane potential. In summary, wild type and
ifests (1) a seamless interface between biological and electronic
genetically engineered pacemaker cells incorporate in each cell all the machinery components, maintaining heart rate greater than 45 bpm;
needed to initiate and propagate action potentials. In contrast, in the stem cell– (2) conservation of total electronic beats delivered (which should
myocyte pairing, two cells together work as a single functional unit whose opera- prolong battery life); (3) the possibility of using the memory
tion is critically dependent on the gap junctions that form between the two function of the electronic pacemaker to track the function of the
disparate cell types. biological component, and (4) a more physiological and catechol-
amine responsive heart rate than is provided by electronic pace-
(Reproduced with permission from Rosen MR, Brink PR, Cohen IS, et al: Genes, stem
makers alone.
cells and biological pacemakers. Cardiovasc Res 64:12–23, 2004.)
The need to reclaim normal cardiac activation in the setting
of normal sinus rhythm and complete heart block has led inves-
tigators to engineer bypass tracts such that sinus impulses can
gain access to the ventricles. This design would facilitate AV
8 to 10 weeks, most of the hMSCs have migrated from the site conduction and would potentially rival the function of AV
of administration. For this reason, attempts are now being made sequential electronic pacing. However, this work is still in its
to encapsulate them in nanofabrics that will maintain them at the infancy as it relates to biological pacemaking. Through one inno-
site of administration while still permitting gap junction forma- vative approach, rat skeletal myoblasts were placed in an artificial
tion and transmission of signals across the fabric. matrix and were inserted into the AV groove of rats whose adja-
Critical questions regarding these—and other—stem cells cent atrium and ventricle had been denuded of epicardium. AV
involve not only the possibility of evolution into inimical or inef- conduction and survival of transplanted cells were then docu-
fective cell types, but also the possibility of rejection or apopto- mented.68 Replication of results and data regarding long-term
sis.60,61 In a 6-week trial of hMSCs xenotransplanted into canine success of this approach are awaited.
hearts, no rejection was reported (Figure 26-5).58 This finding is What do we and others need to show to allow biological
consistent with the hypothesis that hMSCs are immunopro- pacing to move into clinical testing? We need to report results
tected, most likely by one of several humoral factors that they from long-term trials of viral or stem cell constructs that show
release.62-66 However, we have seen robust rejection of occasional no toxicity while demonstrating reliable effect. The hallmark for
lots of hMSCs by canine hearts, and this is not unexpected with comparison is the electronic pacemaker. Trials should show that
xenotransplantation. As for the eventual human application of regardless of whether a viral vector or a cell platform delivery
hMSCs, studies of their allogeneic administration to human subsystem or a pacemaker cell is used, it provides function that sup-
jects for myocardial repair have reported no rejection response, ports life with an adequate basal rate and adequate responses to
but a high degree of cell loss,62,63 mirroring our own experience autonomic input. This activity must emanate uniquely from the
in the dog. Clearly, issues remain regarding how to best bring site implanted, with no evidence of wandering of constructs or
this therapy to human patients. cells to other sites in the heart that might compete.
0 0 26
35pA/pF
1s
A B
1.0
n9
.8
.6
15pA/pF
I/Imax
.4
.2 0.5 s
0.0
160 140 120 100 80 60 40
C Vm (mV) D Injection Spontaneous Overdrive
E
Figure 26-4. Example of loading human mesenchymal stem cell (hMSC) with a gene of interest, studying it biophysically, and determining its effect on the heart.
A, Pacemaker current, If, is not seen in a non-electroporated hMSC. B, Functional expression of If in hMSC transfected with the mHCN2 gene. C, Fit by the Boltzmann equa-
tion to the normalized tail currents of If gives a midpoint of −91.8 ± 0.9 mV and a slope of 8.8 ± 0.5 mV (n=9). If is fully activated around −140 mV with an activation threshold
of −60 mV. Inset shows representative tail currents used for activation curves. Voltage protocol: Hold at −30 mV, hyperpolarize × 1.5 s to −40 to −160 mV in 10-mV incre-
ments, followed by 1.5 s voltage step to +20 mV. D, Pacemaker function in canine heart in situ. Top to bottom, Electrocardiogram (ECG) leads I, II, III, AVR, AVL, and AVF.
Left, Pacing from hMSC injection site showing ECG configuration. Middle, Spontaneous rhythm pace-maps to site of injection, suggesting that it is initiated at the injection
site; Right, Last two beats of overdrive pacing (black bar) at 80 bpm followed by escape rhythm. Escape time = 1.3 sec. E, Immunostaining for cardiac connexin (Cx)43 in a
region of interface between an injection site and myocardium. DAPI staining reveals nuclei. Purple arrows are intercalated discs; white arrows show Cx43 staining between
hMSCs; red arrows show Cx43 staining between hMSCs and myocytes. (Modified with permission from Potapova I, Plotnikov A, Lu Z, et al: Human mesenchymal stem cell
as a gene delivery system to create cardiac pacemakers. Circ Res 94:841–959, 2004.)
Satisfying these demands is no mean feat: It requires that we We have worked with quantum dots as a tracking method
follow the beat-by-beat function of the heart for 6- to 12-month (Figure 26-6).69 These nanoparticles offer the advantages of
periods while concurrently ensuring that viral vectors are not passive loading into cells, inability to pass through gap junctions,
causing untoward effects on the heart or on other tissues, and uptake and removal by the reticuloendothelial system if host cells
that cells—whether hMSCs or embryonic stem cell (ESC)- die, and a strong emission signal that can be readily detected.
derived or other—are not being destroyed and are not being Although they are readily identified in tissue sections, tracing the
transformed into forms that are malignant, or if nonmalignant, fate of cells loaded with quantum dots using standard imaging
then nonfunctional. techniques in vivo remains a major challenge.
A variety of methods are available for tracking cells (Table Another issue important to large animal research and to future
26-2). Ideally, tracking should use markers that stay with the cell applicability to human subjects is construct delivery. Although
forever and cannot pass from cell to cell, and whose signal can catheters are available to deliver viral vectors in specific settings,
be read from the body surface via imaging techniques. To date, the ideal flexible catheter with electrode recording capability at
no method has been adequate. Variations on magnetic resonance its tip and a bore large enough to prevent inflicting damage on
imaging (MRI) have been used most frequently in in vivo studies stem cells has not been reported.
to track iron or ferritin particles loaded into stem cells. The Viral vectors remain a problem in the following sense: Adeno-
shortcoming is that if a cell or cells die, these particles can be viral vectors express only episomally and are of use largely for
taken up by or can reside in other cells and/or can reside extracel- proof of concept, or transient gene delivery. Adeno-associated
lularly. Upon being read on MRI, they may be correctly identi- virus provides more long-term expression, but episomal versus
fied as a persistent signal, thus giving rise to a false-positive genomic expression remains an issue, and the adeno-associated
conclusion of the continued presence of stem cells. virus cannot carry large constructs like SkM1. Lentivirus, which
H&E CD44
A B
GFP DOG
DOG IGG
IGG
C D
CD3 CASPASE–3
E F
Figure 26-5. Human mesenchymal stem cells (hMSCs) 6 weeks after left ventricular (LV) anterior wall intramyocardial injection. Hematoxylin and eosin (H&E) identifies
basophilic cells (upper left) that are CD44 positive (upper right) and GFP positive (peroxidase stain, middle left). hMSCs do not display labeling/binding of dog-IGG to their
surface (middle right)—evidence against humoral rejection. CD3-positive T lymphocytes are rarely noted in association with clusters of hMSCs (lower left)—evidence against
cellular rejection. Staining is negative for activated CASPASE 3 (lower right)—evidence against apoptosis. (Original magnification, ×400.)
(Reproduced with permission from Plotnikov AP, Shlapakova I, Szaboks MJ, et al: Xenografted adult human mesenchymal stem cells provide a platform for sustained biological
pacemaker function in canine heart. Circulation 116:706–713, 2007.)
assurance of quality control from patient to patient would be

Table 26-2. Methods of Cell Tracking nightmarish in comparison with electronic pacing. It has been
shown that hMSCs can be administered allogeneically to human
Agent Limitations
subjects with no adverse reactions, although long-term persis-
Reporter Proteins tence remains an issue.61 These cells may ultimately offer a
readily standardized cell type, but more research is needed to test
LacZ/b-gal Require secondary detection
this hypothesis. Human ESCs have required the use of immuno-
GFP Unstable, may be weaker than suppression in experiments to date.47 Whether this will be the
autofluorescence case in human experimentation remains to be seen, but a lifetime
of electronic pacing appears far preferable to a lifetime of
Ex Vivo Staining
immunosuppression.
FISH Expensive, false-positives The choice of promoter and the regulation of nucleic acids
are additional issues. To date, many of us have worked with a
Surface markers Lost/unexpressed over time; require
cytomegalovirus (CMV) promoter. However, temporal changes
secondary detection
in the expression of promoters may be noted,70 such that CMV
Inorganic may not be the best long-term choice. Moreover, because of
safety concerns, the idea of a cardiac-specific (and maybe even a
Radiometals Toxicity, uniformity of loading cardiac region–specific) promoter would be of importance. This
Fluorescent microspheres Significant aggregation, unstable would seem especially relevant with regard to viral delivery.
signal Duration of effect and robustness of expression will be influ-
enced by processes that regulate protein expression. The discov-
Nanoparticles (i.e., quantum New technology, many unknowns
ery that microRNA may reduce expression and/or that alterations
dots)
in protein degradation may alter expression of proteins can have
FISH, Fluorescent in situ hybridization; GFP, green fluorescent protein. an impact on RNA/DNA expression, DNA regulatory sequences,
and modulation of expression. All of these possibilities will con-
tinue to be explored over the coming years.
One issue on which some progress has been made is the design
leads to genomic incorporation, would appear to be the vector of of preclinical and clinical trials of biological pacemakers. One
choice for long-term maintenance of an effect, although safety could not envisage a standard phase 1 trial in which healthy
issues remain a concern. And although lentivirus readily supports volunteers are sought. Rather, a reasonable trial design in human
gene expression in cell culture, achieving consistent expression subjects would enroll patients in chronic atrial fibrillation and
in vivo has been a challenge. with complete heart block or significant bradycardia, in whom
Cell systems must be optimized and their persistence docu- demand ventricular pacing is needed. The canine equivalent of
mented. Autologous cells elicit no immune response when this trial would include animals in ablation-induced complete
re-administered to the donor. However, they would make bio- heart block. Either setting would incorporate implantation of a
logical pacing into a designer therapy whose expense and right ventricular apical endocardial demand pacemaker, as well
26
A
B
7.24
1.46 base
mm
6.24 base
endo epi
mm
1.06
epi endo
5.24 apex 0.66 apex
1.0
1.0 0.5 mm
0 1.0 2.0 1.0
C mm D mm 1.5
100
distance x
75
% of cells at 50
25 1 day after injection

1 hour after injection
0
0.5 1 1.5 2 2.5
E
F Distance from centroid, mm
Figure 26-6. Quantum dots (QD) are used to identify single human mesenchymal stem cells (hMSCs) after injection into the rat heart and are further used to reconstruct
the three-dimensional (3D) distribution of all delivered cells. Rat hearts were injected with QD-hMSCs. Fixed, frozen sections were cut transversely (plane shown in B, inset)
at 10 µm and were mounted onto glass slides. Sections were imaged for QD fluorescence emission (655 nm) with phase overlay for visualization of tissue borders. QD-hMSCs
can be visualized at (A) low power, and (A, inset) high power (Hoechst 33342 dye used to stain nuclei blue). In (A, inset), endogenous nuclei can be seen adjacent to the
delivered cells in the mid-myocardium (arrowheads). Serial low-power images were registered with respect to one another, and (B) binary masks were generated, where
white pixels depict QD-positive zones in the images. The vertical line in (B, inset) represents the z-axis, which has a zero value at the apex of the heart. The binary masks
for all QD-positive sections of the heart were compiled and were used to generate the 3D reconstruction of delivered cells in the tissue. QD-hMSCs remaining in the tissue
adhesive on the epicardial surface (not within the cardiac syncytium) were excluded from the reconstruction. (C) QD-hMSC reconstruction in an animal that was terminated
1 hour after injection. (D) Reconstruction from an animal euthanized 1 day after injection, with orientation noted in the inset. Our reconstructions in (C) and (D) do not
account for all of the approximately 100,000 hMSCs delivered through the needle. Some of these cells undoubtedly leaked out of the needle track, and others may not
have survived the injection protocol. Views of both reconstructions—(C) and (D)—are oriented for optimal static visualization (and have different scales and are situated
at different positions along the z-axis, depending on the distance of the injection site from the apex of the heart). (E) One day after injection into the heart, the pattern of
QD-hMSCs is well organized and appears to mimic the endogenous myocardial orientation (dotted white line highlights myofibril alignment). Complete representations
of the spatial localization of QD-hMSCs in the heart permit further quantitative analyses. (F) One parameter that can be computed is the distance of individual cells from
the centroid of the total cell mass. The plots show the percentage of cells at a distance less than or equal to x for both 1-hour and 1-day rats. At both time points, most of
the cells are within 1.5 mm of the centroid.
500 µm, inset = 20 µm Scale bar on B, inset Scale bar on a = 1 cm Scale bar on e = 500 µm
(Reproduced with permission from Rosen AB, Kelly DJ, Schuldt AJ, et al: Finding fluorescent needles in the cardiac haystack: Tracking human mesenchymal stem cells labeled
with quantum dots for quantitative in vivo 3-D fluorescence analysis. Stem Cells 25:2128–2138, 2007.)
as implantation of the biological pacemaker. The latter would be have shown no competition between biological and electronic
introduced at a site that optimizes cardiac output, such as the units.36
high interventricular septum. The trial would proceed in such a
way that the electronic pacemaker would provide a safety net
should temporary or permanent failure of the biological occur. Conclusions
In addition, the electronic pacemaker can be used to track and
record the function of the biological. The biological unit, in turn, Modern media mania expects this morning’s ideas to be this
would likely conserve battery power in the electronic unit—our afternoon’s cures. But it would be irresponsible to try to make
work to date has shown that tandem biological-electronic opera- biological pacing fit this mold. Because electronic pacing pro-
tion sees the electronic unit firing only 30% of the time, which vides the luxury of a competent backstop, biological pacing
would indeed prolong the life of the battery.36 In addition, the should be optimized and adequately tested before human experi-
biological component would be the heart’s primary pacemaker, mentation begins. The advantage of this cautious approach is that
would allow autonomic responsiveness, and would provide the difficult questions that must be asked and the difficult experi-
cardiac activation and output characteristics superior to those ments that must be done not only should help to advance our
provided by the electronic unit. As reviewed previously, proof of knowledge in this field but likely will find beneficial application
concept experiments of tandem pacing have been promising and in a variety of areas of gene and cell therapy.
24. Nelson G: A brief history of cardiac pacing. Texas arrhythmia: An in silico, in vivo, in vitro study.
References Heart Inst J 20:12–18, 1993. Circulation 119:19, 2009.
25. Zivin A, Mehra R, Bardy GH: Cardiac pacemakers. 45. Kryukova Y, Protas L, Robinson RB: Ca2+-acti-
1. Adams R: Cases of diseases of the heart, accompa- In Spooner PM, Rosen MR, editors: Foundations vated adenylyl cyclase 1 introduces Ca2+-
nied with pathological observations. Dublin Hos- of Cardiac Arrhythmias, New York, 2001, Marcel dependence to beta-adrenergic stimulation of
pital Reports 4:353–453, 1827. Dekker Inc, pp 571–598. HCN2 current. J Mol Cell Cardiol 52:1233–1239,
2. Stokes W: Observations on some cases of perma- 26. Alt E, Matula M, Theres H, et al: The basis for 2012.
nently slow pulse. Dublin Quarterly Journal of activity controlled rate variable cardiac pacemak- 46. Boink GJJ, Kryukova Y, Lau DH, et al: Introducing
Medical Science 2:73–85, 1846. ers: An analysis of mechanical forces on the human the Ca2+-stimulated adenylyl cyclase AC1 into
3. Furman S, Schwedel JB: An intracardiac pacemaker body induced by exercise and environment. Pac HCN2-based biological pacemakers enhances
for Stokes-Adams seizures. N Engl J Med 261:943– Clin Electrophysiol 12:1667–1680, 1989. their function. Circulation 122:A16415, 2010.
948, 1959. 27. Lee KL: In the wireless era: Leadless pacing. 47. Kehat I, Khimovich L, Caspi O, et al: Electrome-
4. Jeffrey K: The invention and reinvention of cardiac Expert Rev Cardiovasc Ther 8:171–174, 2010. chanical integration of cardiomyocytes derived
pacing. Cardiol Clin 10:561–571, 1992. 28. Edelberg JM, Aird WC, Rosenberg RD: Enhance- from human embryonic stem cells. Nat Biotechnol
5. Keith A, Flack MW: The form and nature of the ment of murine cardiac chronotropy by the molec- 22:1282–1289, 2004.
muscular connections between the primary divi- ular transfer of the human β2-adrenergic receptor 48. Satin J, Kehat I, Caspi O, et al: Mechanism of
sions of the vertebrate heart. J Anat Physiol cDNA. J Clin Invest 101:337–343, 1998. spontaneous excitability in human embryonic stem
41:172–189, 1907. 29. Edelberg JM, Huang DT, Josephson ME, et al: cell derived cardiomyocytes. J Physiol 559:479–
6. Lewis T, Oppenheimer BS, Oppenheimer A: The Molecular enhancement of porcine cardiac chro- 496, 2004.
site of origin of the mammalian heart-beat; the notropy. Heart 86:559–562, 2001. 49. Lin G, Cai J, Jiang H, et al: Biological pacemaker
pacemaker in the dog. Heart 2:147–169, 1910. 30. Miake J, Marbán E, Nuss HB: Gene therapy: Bio- created by fetal cardiomyocyte transplantation.
7. Fye WB: The origin of the heart beat: a tale of logical pacemaker created by gene transfer. Nature J Biomed Sci 12:513–519, 2005.
frogs, jellyfish, and turtles. Circulation 76:493– 419:132–133, 2002. 50. Cai J, Lin G, Jiang H, et al: Transplanted neonatal
500, 1987. 31. Miake J, Marban E, Nuss HB: Functional role of cardiomyocytes as a potential biological pacemaker
8. Geison G: The Royal Institution Lectures of 1869. inward rectifier current in heart probed by Kir2.1 in pigs with complete atrioventricular block.
In Michael Foster and the Cambridge School of overexpression and dominant-negative-suppres- Transplantation 81:1022–1026, 2006.
Physiology, Princeton, NJ, 1978, Princeton Uni- sion. J Clin Invest 111:1529–1536, 2003. 51. Zhang J, Wilson GF, Soerens AG, et al: Functional
versity Press, pp 200. 32. Yu H, Wu J, Potapova I, et al: MinK-related cardiomyocytes derived from human induced
9. Mommersteeg MT, Hoogaars WM, Prall OW, et peptide 1: A β subunit for the HCN ion channel pluripotent stem cells. Circ Res 104:e30–e41,
al: Molecular pathway for the localized formation subunit family enhances expression and speeds 2009.
of the sinoatrial node. Circ Res 100:354–362, 2007. activation. Circ Res 88:E84–E87, 2001. 52. Novak A, Shtrichman R, Germanguz I, et al:
10. DiFrancesco D: A study of the ionic nature of the 33. Qu J, Barbuti A, Protas L, et al: HNC2 overexpres- Enhanced reprogramming and cardiac differentia-
pacemaker current in calf Purkinje fibres. J Physiol sion in newborn and adult ventricular myocytes: tion of human keratinocytes derived from plucked
314:377–393, 1981. Distinct effects on gating and excitability. Circ Res hair follicles, using a single excisable lentivirus.
11. DiFrancesco D: Block and activation of the pace- 89:E8–E14, 2001. Cell Reprogram 12:665–678, 2010.
maker channel in calf Purkinje fibres: Effects of 34. Qu J, Plotnikov AN, Danilo P, Jr, et al: Expression 53. Efe JA, Hilcove S, Kim J, et al: Conversion of
potassium, caesium and rubidium. J Physiol and function of a biological pacemaker in canine mouse fibroblasts into cardiomyocytes using a
222:329–347, 1982. heart. Circulation 107:1106–1109, 2003. direct reprogramming strategy. Nat Cell Biol
12. Biel M, Schneider A, Wahl C: Cardiac HCN chan- 35. Plotnikov AN, Sosunov EA, Qu J, et al: A biologi- 13:215–222, 2011.
nels: Structure, function, and modulation. Trends cal pacemaker implanted in the canine left bundle 54. Warren L, Manos PD, Ahfeldt T, et al: Highly
Cardiovasc Med 12:202–216, 2002. branch provides ventricular escape rhythms having efficient reprogramming to pluripotency and
13. Lakatta EG, Maltsev VA, Vinogradova TM: A physiologically acceptable rates. Circulation directed differentiation of human cells with syn-
coupled SYSTEM of intracellular Ca2+ clocks and 109:506–512, 2004. thetic modified mRNA. Cell Stem Cell 7:618–630,
surface membrane voltage clocks controls the time- 36. Bucchi A, Plotnikov AN, Shlapakova I, et al: Wild- 2010.
keeping mechanism of the heart’s pacemaker. Circ type and mutant HCN channels in a tandem 55. Cho HC, Kashiwakura Y, Marban E: Creation of
Res 106:659–673, 2010. biological-electronic cardiac pacemaker. Circula- a biological pacemaker by cell fusion. Circ Res
14. Lakatta EG, DiFrancesco D: What keeps us tion 114: 992–999, 2006. 100:1112–1115, 2007.
ticking: A funny current, a calcium clock, or both? 37. Kashiwakura Y, Cho HC, Barth AS, et al: Gene 56. Potapova I, Plotnikov A, Lu Z, et al: Human mes-
J Mol Cell Cardiol 47:157–170, 2009. transfer of a synthetic pacemaker channel into the enchymal stem cell as a gene delivery system to
15. Thollon C, Bedut S, Villeneuve N, et al: Use- heart: A novel strategy for biological pacing. Cir- create cardiac pacemakers. Circ Res 94:841–959,
dependent inhibition of hHCN4 by ivabradine and culation 114:1682–1686, 2006. 2004.
relationship with reduction in pacemaker activity. 38. Tse HF, Xue T, Lau CP, et al: Bioartificial sinus 57. Valiunas V, Doronin S, Valiuniene L, et al: Human
Br J Pharmacol 150: 37–46, 2007. node constructed via in vivo gene transfer of an mesenchymal stem cells make cardiac connexins
16. Borer JS, Fox K, Jaillon P, et al: for the engineered pacemaker HCN channel reduces the and form functional gap junctions. J Physiol
Ivabradine Investigators Group: Antianginal and dependence on electronic pacemaker in a sick-sinus 555:617–626, 2004.
antiischemic effects of ivabradine, an If inhibitor, in syndrome model. Circulation 114:1000–1011, 58. Plotnikov AP, Shlapakova I, Szabolcs MJ, et al:
stable angina: A randomized, double-blind, multi- 2006. Xenografted adult human mesenchymal stem cells
centered, placebo-controlled trial. Circulation 39. Plotnikov AN, Bucchi A, Shlapakova I, et al: provide a platform for sustained biological pace-
107:817–823, 2003. HCN212-channel biological pacemakers manifest- maker function in canine heart. Circulation
17. Scherf D, Schott A: Extrasystoles and Allied ing ventricular tachyarrhythmias are responsive to 116:706–713, 2007.
Arrhythmias, ed 2, Chicago, 1973, Year Book treatment with If blockade. Heart Rhythm 5:282– 59. Rosen MR, Brink PR, Cohen IS, et al: Genes, stem
Medical Publishers. 288, 2008. cells and biological pacemakers. Cardiovasc Res
18. McWilliam JA: Electrical stimulation of the heart 40. Cingolani E, Yee K, Shehata M, et al: Biological 64:12–23, 2004.
in man. BMJ 1:348–350, 1889. pacemaker created by percutaneous gene delivery 60. Rosen M: Biological pacemaking: In our lifetime?
19. Lidwell MC: Cardiac disease in relation to anaes- via venous catheters in a porcine model of Heart Rhythm 2:418–428, 2005.
thesia. In Transactions of the Third Session, Aus- complete heart block. Heart Rhythm 8:S112, 61. Rosen MR, Robinson RB, Brink PR, et al: The
tralasian Medical Congress, September 2-7, 1929, 2011. road to biological pacing. Nat Rev Cardiol 8:656–
Sydney, Australia, p 160. 41. Boink GJJ, Kryukova Y, Lau DH, et al: HCN2/ 666, 2011.
20. Furman S, Szarka G, Layvand D: Reconstruction SkM1 gene transfer into the canine left bundle 62. Zimmett JM, Hare JM: Emerging role for bone
of Hyman’s second pacemaker. Pacing Clin Elec- branch induces highly stable biological pacing at marrow derived mesenchymal stem cells in myo-
trophysiol 28:446–453, 2005. physiological beating rates. Heart Rhythm 8:S54, cardial regenerative therapy. Basic Res Cardiol
21. Bigelow WG, Callaghan JC, Hopps JA: General 2011. 100:471–481, 2005.
hypothermia for experimental intracardiac surgery: 42. Boink GJJ, Sosunov EA, Shlapakova IN, et al: 63. Rosen MR: Are stem cells drugs? The regulation
The use of electrophrenic respirations, an artificial SkM1 gene transfer into the canine ventricular of stem cell research and development. Circulation
pacemaker for cardiac standstill, and radio- myocardium decreases the threshold potential for 114:1992–2000, 2006.
frequency rewarming in general hypothermia. action potential initiation. Circulation 123:A4013, 64. Groh ME, Maitra B, Szekely E, et al: Human mes-
Trans Meet Am Surg Assoc Am Surg Assoc 68:211– 2011. enchymal stem cells require monocyte-mediated
219, 1950. 43. Protas L, Dun W, Lu J, et al: Expression of skeletal activation to suppress alloreactive T cells. Exp
22. Zoll PM: Resuscitation of the heart in ventricular but not cardiac Na+ channel isoform preserves Hematol 33:928–934, 2005.
standstill by external electric stimulation. N Engl J normal conduction in a depolarized cardiac syncy- 65. Di Nicola M, Carlo-Stella C, Magni M, et al:
Med 247:768–771, 1952. tium. Cardiovasc Res 81:528–535, 2009. Human bone marrow stromal cells suppress T lym-
23. Weirich W, Gott V, Lillehei C: The treatment of 44. Lau DH, Clausen C, Sosunov EA, et al: Epicardial phocyte proliferation induced by cellular or nonspe-
complete heart block by the combined use of a border zone overexpression of skeletal muscle cific mitogenic stimuli. Blood 99:3838–3843, 2002.
myocardial electrode and an artificial pacemaker. sodium channel, SkM1, normalizes activation, pre- 66. Tse WT, Pendleton JD, Beyer WM, et al: Suppres-
Surg Forum 8:360–363, 1957. serves conduction and suppresses ventricular sion of allogeneic T-cell proliferation by human
marrow stromal cells: Implications in transplanta- 68. Choi YH, Stamm C, Hammer PE, et al: Cardiac quantum dots for quantitative in vivo 3-D fluores-
tion. Transplantation 75:389–397, 2003. conduction through engineered tissue. Am J Pathol cence analysis. Stem Cells 25:2128–2138, 2007.
67. Shlapakova IN, Nearing BD, Lau DH, et al: Bio-
logical pacemakers in canines exhibit positive chro-
notropic response to emotional arousal. Heart
169:72–85, 2006.
69. Rosen AB, Kelly DJ, Schuldt AJ, et al: Finding
fluorescent needles in the cardiac haystack: Track-
70. Reinhard E, Nedivi E, Wegner J, et al: Neural
selective activation and temporal regulation of a
mammalian GAP-43 promoter in zebrafish. Devel-
26
Rhythm 7:1835–1840, 2010. ing human mesenchymal stem cells labeled with opment 120:1767–1775, 1994.
Intercellular Communication and
Impulse Propagation 27
André G. Kléber
CHAPTER OUTLINE and are continuously internalized from the plaque center along
microfilaments.7 Proteins, such as ZO-1 and CAR, are taking part
Cardiac Intercellular Communication by Gap Junctions 265 in these processes and are involved in interactions with other
Myocardial Connexins Cx43, Cx40, and Cx45 265 proteins of the intercellular junction and plaque size control.9
The total of these behaviors indicates that gap junctions repre-
Role of Gap Junctions in Electrical Impulse Transfer sent a highly dynamic, regulated, and interactive system and
and Propagation 266 explains why changes in composition and function can contribute
importantly to disease phenotypes.
Effect of Heterogeneous Expression of Connexins
on Propagation 266
Interaction of Cell-to-Cell Coupling, Tissue
Architecture, and Ion Currents 268 Myocardial Connexins Cx43, Cx40, and Cx45
Ephaptic Impulse Transmission: Potential Alternative Cx43, Cx40, and Cx45 are the three major connexins of the
Mechanism of Electrical Impulse Transfer 269 myocardium.10-12 A fourth connexin, Cx32.1, has been described
Remodeling of Gap Junctions: Only a Part of in murine nodal tissues.13 Cx43 is the dominating connexin in
Remodeling of the Whole Intercalated Disc? 271 ventricular myocardium, Cx40 dominates in the ventricular con-
duction system, and Cx45 is present in the atrioventricular and
Summary 273 sinoatrial nodes and in lower amounts in atrial and ventricular
myocardium.
Atrial tissue shows a wide spectrum of electrical propagation
velocities (0.2 to 1.6 m/s).14 This raises the question to what
Cardiac Intercellular Communication extent Cx43 and Cx40, which are both abundantly expressed and
by Gap Junctions colocalize in intercalated discs (the ratio of Cx40:Cx43 expression
is larger in the right than the left atrium15), contribute to this
The existence of low-resistance pathways between cardiac cells variability. In human atria, an important factor in determining
was postulated from experimental assessment of electronic inter- propagation is the ratio of Cx40:Cx43 expression, with Cx40
actions (cable analysis) between cardiac Purkinje fibers before the dominance decreasing and a Cx43 dominance increasing local
existence of gap junction channels was known.1 During the sub- velocity.16 A similar behavior was observed in vitro in atrial
sequent decades ion channels between cardiac cells were identi- strands cultured from myocytes with genetic Cx43 and Cx40
fied, formed from two connexon hemichannels, each being ablation.17 In contrast, measurements of macroscopic excitation
composed of six connexin molecules. The main connexins of the in murine atria with Cx40 ablation in vivo yielded slowed or
myocardium—Cx43, Cx40, and Cx45—have been defined in unchanged propagation.17 It cannot be excluded that macroscopic
terms of genetic code, amino acid sequence, and molecular func- measurements in whole atria with genetic Cx40 ablation are
tion. The introduction of the whole-cell dual voltage clamp tech- affected by malformations.18 Colocalization of Cx43 and Cx40
nique made it possible to measure directly the intercellular raises the question of whether these connexins form separate gap
electric conductance and single channel conductance between a junctions or whether mixed Cx40/Cx43 gap junction channels
pair of cells obtained from enzymatic disaggregation of cardiac exist in the real atrial myocardium. This question is especially
tissue or after neoformation in cultured cells. Later, transfection interesting regarding their effect on propagation, because
techniques made it possible to study the biophysical properties mixed Cx40/Cx3 channels could have an electrical conductance
of specific cardiac connexins in heterologous expression systems different from pure homomeric or homotypic Cx43 or Cx40
and to define the properties of pure or mixed connexons and channels.2,19 In heterologous expression systems, controversial
connexins (heteromeric vs. homomeric connexons and homolo- results about the function of mixed Cx40/Cx3 channels have been
gous vs. heterologous connexin channels) in terms of biophysical reported.2,19-20
and diffusive properties.2-5 Cx45 is a dominating connexin in early stages of cardiac devel-
Similar to ion channels, the electrical behavior of gap junc- opment and in the atrioventricular and sinoatrial nodes. In atrial
tions channels is sensitive to the electrical field across the channel and ventricular tissue from human, rat, and murine hearts, Cx45
and to variations of the intercellular milieu (Ca2+, pH, Mg2+, has been detected in small amounts,15,17,21-22 where it is likely to
ATP).6 Gap junction channels have a high turnover rate with a form mixed channels with Cx43 and/or Cx40. Heteromeric
half-time of only a few hours.7 Trafficking of connexins to the connexins have been described in heterologous expression
membrane along microtubular highways requires the presence of systems, if Cx45 was coexpressed with Cx43 or Cx40.2 Germline
proteins specifically linking the microtubular ends to the scaffold ablation of Cx43 in murine ventricular pairs produces a greater
of mechanical junctions.8 The hemichannels (connexons) are than 90% reduction in cell-to-cell coupling with persistence of
transported by the microtubules to the periphery of the gap junc- Cx45 expression.22 This finding indicates that Cx45 contributes
tion plaques, shift to the center of the plaques in a short time, little to electrical coupling; however, a regulatory role has been
265
attributed to Cx45 in cardiac failure, in which Cx45 is upregu- Several distinct features specifically related to cellular uncoupling
lated together with a decrease in Cx43 and a reduction in gap can be derived from this figure. First, changes in cell-to-cell
junction size.23 This could indicate that Cx45 is involved in regu- coupling close to the normal level of coupling produce relatively
lation of gap junction size, as also suggested from coexpression small changes in propagation velocity. Second, cell-to-cell uncou-
of Cx43 with Cx45 in a rat liver epithelial cell line.24 Recently, it pling can produce extremely slow propagation (on the order of
has been shown that genetic deletion of the Coxsackie-adenovirus 1 cm/s) if the conductance between cells is reduced more than
receptor (CAR) is associated with marked reduction or deletion 100-fold. This behavior also predicts preservation of propaga-
of Cx45. Conditional ablation of CAR increased cell-to-cell dye tion, albeit slow, even at extreme levels of uncoupling. This pre-
diffusion between ventricular cells, further suggesting that Cx45 diction made from theoretical studies has been verified in
is a significant modulator of ventricular cell-to-cell experimental work using engineered strands of neonatal rat ven-
communication.25 tricular myocytes either exposed to an uncoupling agent or using
cells with germline ablation of connexin43 (Cx43).22,28 The expla-
nation for this behavior, which is typical for cell-to-cell uncou-
pling and contrasts to the behavior during block of depolarizing
Role of Gap Junctions in Electrical Impulse ion currents (INa, IL,Ca),27-28 is given by the biphasic behavior of
Transfer and Propagation the margin of safety of propagation (see Figure 27-1). The bipha-
sic time course of SF is determined by two processes with oppos-
The role of gap junction channels in cardiac propagation as one ing effects. First, the axial currents flowing into a given cell from
of the key elements is well acknowledged. Early experimental and excited tissue upstream decreases because of an increase in the
theoretical studies examined the effect of a decrease in gap junc- resistance between the cells. This decrease (so-called source effect)
tion conductance on propagation velocity in a general way by causes the membrane capacitance of a given cell to charge more
implicating or inducing a uniform change of intercellular electri- slowly and to reach the excitation threshold later than normal,
cal conductance in the tissue.26 The effect of a uniform decrease and therefore, propagation to slow. Second, the excitatory electri-
in intercellular electrical conductance is best understood by con- cal charge flowing downstream, because of the increased intercel-
sidering the principle of conduction safety. The formalism lular resistance between the downstream cells, is distributed over
describing conduction safety (SF) states that27: fewer cells and is therefore conserved at the site of excitation
within the wavefront (decrease of so-called sink effect). In other
SF = (∫ I
A C ⋅dt + ∫ I
A OUT ⋅dt ) (∫ i ⋅dt ); A |Qm > 0
A M
words, cell-to-cell uncoupling protects the excitatory current
from downstream dissipation and therefore makes propagation
safer (increase of SF). Only at extreme levels of uncoupling does
In simple terms, the numerator of this equation equals the electri- failure occur and propagation eventually gets blocked. The bio-
cal charge produced by a given cell in the propagating wavefront, physical principle of source-sink interaction is not only valid for
and it consists of the sum of electrical charges needed to produce states of cell-to-cell uncoupling; it is a general rule independent
the action potential (capacitive component, ∫AIC dt) and the
charge flowing to excite the cells downstream (axial component,
⋅ of scale and holds also for propagation in the presence of discon-
tinuities in tissue structure (e.g., infarct scars26). Importantly, slow
⋅
∫AIout dt). The denominator corresponds to the electrical charge propagation is only preserved if sites of increased coupling resis-
⋅
(∫AIm dt) exciting this same cell, which in turn is produced by
the excited cells upstream of the propagating front. From this
tance are closely spaced and produce an effective decrease in the
electrotonic sink. If this effect is absent, such as at the transition
formalism, it follows that propagation becomes decremental and between tissue of decreased to normal coupling, propagation
eventually gets blocked, when SF decreases less than 1, or when block is observed.29 This likely explains some observations made
more electrical charge is needed to excite a given cell than this in experimental models of heterogeneous connexin expression,30
same cell can produce by its machinery of excitation.27 The theo- which is discussed as follows.
retical dependence of propagation velocity on cell-to-cell cou-
pling and the associated change in SF is illustrated in Figure 27-1.
Effect of Heterogeneous Expression
of Connexins on Propagation
Heterogeneous expression of connexins, leading to heteroge-
neous propagation and arrhythmogenesis, has been implicated in
the genesis of atrial fibrillation and ventricular arrhythmias.31
However, in only two studies has a direct association between the
incidence of ventricular arrhythmias in heart failure patients with
heterogeneous connexin expression been demonstrated.32-33
Interestingly, Cx43 immunofluorescence measured in specimens
from such patients with ventricular arrhythmias showed a heter-
ogenous pattern with scattered absence of Cx43 signals. Similarly,
in a mouse model of conditional cardiac Cx43 ablation, it has been
shown that the occurrence of ventricular arrhythmias and sudden
cardiac death is associated with a marked heterogeneity in con-
nexin expression and a decrease in macroscopic propagation
velocity less than 50% of normal (Figure 27-2).34 Recent work
using cell engineering techniques made an attempt to quantita-
Figure 27-1. Dependence of safety and velocity of propagation on intercellular
coupling. Note logarithmic scale on abscissa.
tively study electrical cell-to-cell coupling and propagation in
tissue with reproducible degrees of connexin heterogeneity.35
(From Shaw RM, Rudy Y: Ionic mechanisms of propagation in cardiac tissue. Roles of Heterogeneity was produced in engineered cell pairs and strands
the sodium and l-type calcium currents during reduced excitability and decreased from defined mixtures of wild type (WT) neonatal ventricular
gap junction coupling. Circ Res 81:727–741, 1997.) myocytes expressing a green fluorescent protein (GFP) tag
Intercellular Communication and Impulse Propagation 267
100
Control
27
75
O-CKO
Survival (%)
50
25
0
0 50 100 150 200 250 300
Time (days)
A 25d B 35d C 45d
Control Control Control
D 25d E 35d F 45d
O-CKO O-CKO O-CKO
G 45d H 45d I 45d
Cx43
O-CKO O-CKO O-CKO
Figure 27-2. Mouse model of heterogeneous Cx43 expression with delayed conditional Cx43 knock out (O-CKO). Top, Kaplan-Meier survival curve showing onset of sudden
death at approximately 30–40 days. Bottom, Cx43 immunofluorescence signals in ventricle between 25 and 45 days after birth in control (A–C), and O-CKO mice
(D–F). Inhomogeneity of Cx43 expression by comparison of three different regions of interests (G–I).
(From Danik SB, Liu F, Zhang J, et al: Modulation of cardiac gap junction expression and arrhythmic susceptibility. Circ Res 95:1035–1041, 2004.)
(WT-GFP) and Cx43–/– ventricular myocytes (Figure 27-3). typical of pairs with Cx45 expression in one cell and WT Cx43/
Immunofluorescence analysis showed coexistence of Cx43 and Cx45 expression in the other.36-37 The absence of a Cx43 signal in
Cx45 at the cell interface in the WT/WT pairs. In contrast, no the presence of reduced persistent electrical coupling indicated
Cx45 or Cx43 signals were observed in mixed WT/CX43–/– pairs, that the electrical conductance is a more sensitive indicator of the
and no Cx45 signals in pairs consisting of Cx43–/– cells. In the level of cell-to-cell coupling than immunofluorescence, and it
same pairs, intercellular electrical conductance (gj) decreased confirmed recent measurements correlating electrical conduc-
markedly to less than 90% of WT in the mixed pairs (WT/ tance to gap junction size as assessed by immunofluorescence.38
Cx43–/–) and Cx43–/–/Cx43–/– pairs. However, all three cell inter- The experimental model of engineered cell strands consisting of
faces showed consistent electrical coupling. For the mixed cell controlled mixtures of WT-GFP and Cx43–/– cells was used to
pairs, the dependence of gj on the transjunctional voltage, Vj, was measure the effect of Cx43 inhomogeneity on propagation
Cx43 +/+ Cx43 +/+
** **
A 80 19
B
Cx43 +/+ Cx43 –/– 60
Intercellular conductance (nS)

C
40
D
Cx43 –/– Cx43 –/–
20 n.s.
15
E 12
0
WT - WT WT - KO KO - KO
F
Figure 27-3. Cell-to-cell coupling at the interface between GPF-tagged Cx43+/+ (wild type [WT]) and Cx43–/– (knockout [KO]) neonatal rat ventricular myocytes. Left, Top
and side view of WT/WT cell pair (A, B); WT/KO cell pair (C, D) and KO/KO cell pair (E, F). Cx43 immunofluorescence is in orange; Cx45 immunofluorescence is in white.
Note superimposition of Cx43 and Cx45 immunofluorescence at the cell-to-cell interface in A and B and the absence of immunofluorescence in C–F. Right, Intercellular
conductance between the pairs is illustrated on the left. Note that a small, finite electrical conductance is present in WT/KO and KO/KO cells, despite the absence of Cx43
and Cx45 immunofluorescence signals.
(From Beauchamp P, Desplantez T, McCain ML et al: Electrical coupling and propagation in engineered ventricular myocardium with heterogeneous expression of connexin43.
Circ Res 110;1445–1453, 2012.)
velocity (Figure 27-4).35 The change in macroscopic propagation Cx43+/+ host blastocysts.30 The hearts of these surviving chimeric
velocity along engineered strands revealed a relatively preserved mice, characterized by a “patchy” or mosaic-like regional ablation
velocity in mixtures containing ≥50% WT-GFP cells, and a of Cx43 with macroscopic interfaces, were highly arrhythmo-
marked decrease in cell mixtures with ≤50% WT-GFP cells. The genic, and showed an irregular epicardial activation pattern with
explanation for preserved propagation at a relatively fast level, decreased contractility.30 Comparison with the other mouse
even in the presence of 50% of cells devoid of Cx43 expression, is models therefore suggests that not only the average degree of
illustrated in Figure 27-5, which depicts the pattern of cell-to-cell Cx43 ablation but also the scale of heterogeneity determine the
propagation recorded at high spatial and temporal resolution. effects on electrical and contractile behavior. A biophysical expla-
Both the sequence of activation times and the shapes of the action nation for irregular electrical activation with macroscopic,
potential upstrokes reveal a mixed type of propagation, with fast mosaic-like Cx43 ablation was given by theoretical work, which
continuous conduction26 meandering through the clusters with simulated propagation at an interface between two large-strand
normal Cx43 expression, and delayed activation of Cx43–/– cells segments with different degrees of cell-to-cell coupling.29 In such
showing a prolonged foot potential before the rapid action poten- a case, propagation from a segment of low intercellular coupling
tial upstroke, characteristic of discontinuous conduction.26 to a segment of normal intracellular coupling encounters a large
Although local propagation delays at cell borders can be as long downstream sink at the transition. As a consequence, there is a
as several milliseconds, they are too short to produce microentry large local decrease of safety factor and a conduction delay, which
in markedly uncoupled tissue.22,28,35 Restriction of Cx43 ablation is nearly one order of magnitude larger than the delay observed
to small cell clusters and eventual excitation of all cells probably with uniform cell-to-cell uncoupling. This increased delay, which
explains why optical recordings of macroscopic velocities in most likely accounts for the highly irregular propagation patterns
the mouse model with conditional Cx43 ablation show relatively and the increased arrhythmogenicity in mice with mosaic-
smooth isochronal activation.39 Furthermore, the observation chimeric Cx43 ablation, is due to the absence of the protective
that Cx43 immunosignals are absent from the interface between effect of low resistance coupling downstream.
Cx43-expressing and nonexpressing cells suggests that the assess-
ment of the degree of Cx43 ablation (proportion of Cx43–/–
cell) is overestimated if calculated from the percent loss of
Cx43–/– immunosignal. Interaction of Cell-to-Cell Coupling, Tissue
In contrast to mouse models of microscopic connexin inho- Architecture, and Ion Currents
mogeneity (conditional Cx43 ablation or mixed Cx43+/+/Cx43–/–
engineered strands35,39) macroscopic connexin inhomogeneity The morphology of atrial and ventricular tissues is highly discon-
was produced by injecting Cx43–/– embryonic stem cells into tinuous at several levels. Ventricular tissue is organized in layers,
source (strand), which emerges into a large sink (bulk). As a
27
consequence, unidirectional block occurs at the transition, if the
source is too small (narrow strand diameter) to excite the large
sink (see Figure 27-6, B). In contrast, propagation from the bulk
to the strand is safe. In such a situation, the degree of cell-to-cell
coupling acts as a modulator of propagation and a decrease in
cell-to-cell coupling can restore unidirectional block to bidirec-
tional propagation. The exact mechanism is illustrated in Figure
27-6, A, in which the strand segment and the bulk are simulated
by a network of excitable cells interconnected by longitudinal and
transverse resistors, mimicking cell-to-cell coupling.48 Reduction
of cell-to-cell coupling reduces the loss of local electrotonic
current into the sink at the transition site and restores propaga-
tion. This explanation becomes evident from the fact that only
the increase of lateral elements of cell-to-cell coupling contrib-
utes to the decrease of the sink effect and restoration of propaga-
tion.48 In the extreme hypothetical case, when the longitudinal
1 mm resistors are left unchanged and the transverse resistors are made
A
infinite, propagation into the lateral sink is no more possible and
propagation gets linear—that is, the source-sink mismatch is
40 eliminated. These examples make it clear that there is no essential
9
** ** difference in biophysical mechanisms between source-to-sink
mismatch at a transition of linear strand segments with different
degrees of cell-to-cell coupling and a transition of a small to a
Propagation velocity (cm/s)
30
10
* ** large tissue structure.29 Repetitive alternations of sink and sources
due to tissue structure, such as branching tissue fibers (scars), can
cause a decrease of the downstream sink, a high margin of safety
20
** for propagation and slow propagation of approximately 1 cm/s.49
20 The propagation behavior at macroscopic alternations is similar
to the situation of uniform cell-to-cell uncoupling illustrated in
Figure 27-1, and it underlines the prior statement that the rules
of source-to-sink mismatch apply to both the cellular and network
10 scales.26
12
In normal linear propagation, Na inward current, which pro-
10 duces the upstroke of the cardiac action potential, drives propaga-
tion by contributing the major portion of charge flowing
0 downstream. The L-type Ca2+ current, which is activated later
WT-GFP/Cx43+/– WT-GFP/Cx43–/– Cx43–/–
50%:50% 50%:50% 100% and more slowly, does not contribute significantly to propagation
WT-GFP WT-GFP/Cx43–/– (<1% to 2%) in the normal condition.27 Because a decrease in
WT-GFP/Cx43–/–
B 100% cell-to-cell coupling resistance is associated with a delay at the
80%:20% 20%:80%
cell border, the action potential of the driven cells is created with
Figure 27-4. Macroscopic impulse propagation in engineered strands with mixed delay with respect to the driver cell. This delay can become so
Cx43 expression. Top, Cell culture consisting of wild type (WT) GFP (50%) and long that the upstroke of the driven cell falls into the late upstroke
knockout (KO) cells (50%) at low magnification. Propagation velocity is measured
or early plateau phase of the driver cell. As a consequence, the
from the conduction time and the distance between two regions of interest (quad-
rangles) by multisite high-resolution optical mapping of membrane potential.
L-type Ca2+ current becomes a main charge carrier for propaga-
Bottom, Dependence of macroscopic velocity on the proportions of WT-GPF and tion in uncoupled tissue.29
KO cells in the strand. Note the rapid decrease in velocity, if the relative amount of This discussion illustrates that the three main functional ele-
WT-GFP cells is less than 50%. ments determining cardiac propagation—the action potential,
the degree of cell-to-cell coupling, and the cellular network
(From Beauchamp P, Desplantez T, McCain ML et al: Electrical coupling and propaga- structure—show a high degree of interdependence at sites of
tion in engineered ventricular myocardium with heterogeneous expression of con- source-to-sink mismatch. Therefore, effects of changes in these
nexin43. Circ Res 110;1445–1453, 2012.) determinants (i.e., in disease) should be considered within the
context of a whole complex system. Interestingly, interacting ion
currents, cell-to-cell coupling, and structural determinants of
which rotate with respect to their longitudinal axis of heart tissue do not only affect cardiac propagation via biophysical
anisotropy.40-42 Atrial and subendocardial ventricular tissues are interaction; they also form a highly interactive system at the level
trabeculated,43 and age or disease can cause additional microfi- of molecular expression in the cell, as discussed in the preceding
brosis.44,45 There is a common consequence of these structural paragraph.
discontinuities for impulse propagation, as recognized early on
by the seminal work of M.E. Spach,26,46 namely that propagating
waves encounter sites of structural and electrical mismatch. In Ephaptic Impulse Transmission:
the context of this chapter, it is important to mention that local
changes in cell-to-cell coupling can have a marked influence on Potential Alternative Mechanism
propagation at such sites. Figure 27-6, A, illustrates a simple of Electrical Impulse Transfer
engineered tissue structure, “archetypical” for source-sink mis-
match, and consisting of a cardiac strand connected and forming The observation that cardiac tissue consists of individual cells,
a transition to a large area (bulk).47 In such structures, propaga- and functions at the same time as a syncytium, raised the question
tion from the strand to the large area is determined by a small of how the electrical impulse would be transferred between cells.
100 100 100 100

APA (%)
APA (%)
APA (%)
APA (%)
50 50 50 50
0 0 0 0
0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30
Activation time (ms) Activation time (ms) Activation time (ms) Activation time (ms)
100 100
APA (%)
APA (%)
50 50
0 0
0 10 20 30 0 10 20 30
Activation time (ms) Activation time (ms)
100 100
APA (%)
APA (%)
50 50
0 0
0 10 20 30 0 10 20 30
Activation time (ms) Activation time (ms)
50 µm
100 100
APA (%)
APA (%)
50 50
0 0
0 10 20 30 0 10 20 30
A Activation time (ms)
100
Activation time (ms)
APA (%)
50
0 10 20 30 40 50 60
B Activation time (ms)
Figure 27-5. Microscopic impulse propagation in engineered strands with mixed Cx43 (50% of cells) expression. The center shows a strand segment measuring 100 µm
wide, with wild type (WT)-GFP cells emitting green native fluorescence. A, Action potentials with smooth upstrokes illustrating rapid sequential activation (velocity, 32 cm/s)
of the WT-GFP cells, starting from the filled white circle. B, Activation of two KO cell areas, with a region (blue) preceding the activation of the WT-GFP cells and a region
(red) with markedly delayed activation. The bar in the middle of B corresponds to the activation window of the WT-GFP cell cluster. Note that the “foot” of the action potentials
in the KO cell is typical for discontinuous activation. Although activation was microscopically highly irregular, no local propagation blocks were observed.
(From Beauchamp P, Desplantez T, McCain ML et al: Electrical coupling and propagation in engineered ventricular myocardium with heterogeneous expression of connexin43.
Circ Res 110;1445–1453, 2012.)
transmission between cells implicates that the flow of inward Na+
27
current during depolarization creates an electrical field across the
cleft and radial resistances with respect to the extracellular space.
This transient intracleft potential is negative regarding the extra-
cellular reference and thus acts to depolarize the membrane
potential of the downstream cell. In turn, this depolarization
leads to local activation of juxtaposed Na+ channels located in the
cleft-membrane of the downstream cell. In essence, two variables
affecting ephaptic impulse transmission are of interest: the cleft
width and the fraction of total Na+-channels located in the inter-
A calated disc. Two models have been used thus far (one- and
three-dimensional54,55), both of which agree with respect to the
main results. Figure 27-7, B, illustrates the complex dependence
of propagation velocity on cleft width at different degrees of
resistive cell-to-cell coupling via gap junctions and with the
assumption that all Na+ channels are clustered at the cell junction.
At normal to moderate degrees of cell-to-cell uncoupling decreas-
100 µm
ing the cleft width decreases propagation velocity significantly.
B This effect occurs because a negative cleft potential will decrease
Figure 27-6. Effect of change in cell-to-cell coupling at a site of source-to-sink
the driving force for Na+ ions. Moreover, if appropriately
mismatch because of a discontinuity in the geometry of the cellular network. Top, modeled, [Na+] will decrease in the intercellular cleft because of
Schematic presentation of an abrupt transition in tissue geometry with the longi- the high diffusion barrier between the cleft and the normal extra-
tudinal cell-to-cell coupling resistors shown in black and the transverse resistors cellular space during inward Na+ current flow.55 At high degrees
shown in red. The excitable cell elements (not shown) reside in the notches con- of cell-to-cell uncoupling, decreasing cleft width below a certain
necting the resistors. Bottom, Source-to-sink mismatch in a culture of neonatal rat threshold increases propagation velocity. This effect is due to the
ventricular myocytes (produced by patterned growth) with a small strand (50 µm negative electrical field accumulating in the cleft (the field effect),
wide) emerging into a large bulk (left). During normal perfusion (middle left) the which depolarizes the membrane of the juxtaposed cell to thresh-
excitation wave (red) is blocked at the entrance of the bulk, which remains in its
old for activation of Na+ channels. Importantly, cell-to-cell trans-
resting state (blue). Application of an uncoupling agent (arachidonic acid) reduces
the source-to-sink mismatch and establishes bidirectional conduction (middle
fer of the electrical impulse without any resistive coupling is
right). Restoration of normal coupling reestablishes the unidirectional block (right). theoretically possible only if 100% of all Na+ channels are clus-
The experiment in the lower panel demonstrates that moderate cell-to-cell uncou- tered in the cleft, a postulate that seems unlikely in the light of
pling can favor electrical propagation at a site of current-to-load mismatch. recent findings of molecular compartmentation of Na+ channels
into an intercalated disc and a surface pool (see next section).56 If
(From Rohr S, Kucera JP, Fast VG, Kléber AG: Paradoxical improvement of impulse approximately 50% of Na channels are located at the intercalated
conduction in cardiac tissue by partial cellular uncoupling. Science 275:841–844, disc, field effect transmission could support slow impulse propa-
1997.) gation at high degrees of cell-to-cell uncoupling.54 However,
impulse transmission in absence of residual resistive coupling is
not possible. Currently, it seems difficult to estimate the impor-
Already in the 1960s two opposite hypotheses prevailed and tance and contribution of ephaptic transmission in experimental
became a matter of friendly controversy among investigators.50 settings. Although impulse transmission is maintained at low
The first hypothesis postulated low-resistance pathways between degrees of gap junction coupling,22,35 total inhibition of resistive
cardiac cells, based on measurements of passive electrical proper- coupling by means of a gap junction channel blocker produces
ties, the diffusion constant for K+, and the observation that a full rapid propagation block.22
electrical resistive local circuit was necessary for propagation to
occur, implicating resistive connections between cardiac cells.51
The second hypothesis stated that electrical impulse transfer
would be possible without direct electrical connections between Remodeling of Gap Junctions: Only a Part of
cardiac cells.52 These two hypotheses, which from today’s point Remodeling of the Whole Intercalated Disc?
of view could represent complementary rather than opposite
mechanisms, have attracted new attention with the observation Most experimental and theoretical studies involving the effect of
that propagation in cardiac ventricular tissue with conditional changes in intercellular electrical coupling, because of connexin
Cx43 ablation was maintained at approximately 50% of a normal remodeling or effects of drugs and metabolites on the conduc-
level, despite an approximately 90% reduction of Cx43 immuno- tance of gap junction channels, regarded these changes as being
fluorescence signal.39,53 The opportunities to perform theoretical independent of changes in other proteins of the cardiac interca-
modeling of the cardiac cell junction offered the possibility to lated disc, including proteins linking the cytoskeleton of neigh-
selectively assess the role of the two postulated hypotheses using boring cells (microfilament and intermediate filament system)
settings of critical biophysical parameters within a wide range.54,55 and ion channels. However, recent work indicates that proteins
Impulse transmission involving ephaptic transmission at a cardiac of the intercalated disc involved in intercellular mechanical and
cell border is schematically illustrated in Figure 27-7, A. In con- electrical functions share common regulatory mechanisms. These
trast to models, which represent the cell junction by a simple mechanisms appear to be complex and have as yet only partially
resistor, the cleft model of a cardiac cell junction contains an been elucidated.57-58 A first indication of a relationship between
excitable element that faces the cell junction, a high intercellular regulation of proteins of desmosomal or adherence junctions and
cleft resistance, and a radial resistance connecting the cleft to connexins was suggested from the observation that patients with
the normal extracellular space. In the model, the high cleft resis- inherited arrhythmogenic right ventricular cardiomyopathy show
tance is varied by changing the cleft width. Inward sodium a decrease in immunofluorescence signals for plakoglobin and
current flowing during depolarization through the Na+ channels Cx43 in the intercalated discs.59 This observation is independent
in the cleft is forced to flow through the narrow intercellular of the mutated protein (desmoglein, desmoplakin, plakophylin-2,
cleft toward the extracellular space. Field effect or ephaptic plakoglobin) and therefore suggests that plakoglobin (γ-catenin)
Cell 1 Cell 2
Rgap
Rmyo Rmyo Rmyo Rmyo Rmyo Rmyo
Rcleft
Rradial
LRd
LRd
LRd LRd LRd LRd
60
100%
Gap junctional coupling (% of normal)

50
Conduction velocity [cm/s]
50%
40
30%
30
20 10%
5%
10 3%
1%
0 0.3%
10 100 1000 Non-cleft
model
B Cleft width [µm]
Figure 27-7. Ephaptic impulse transmission. A, Schematic presentation of the electrical circuitry between two cells (green) used in the model. The classical electrical circuitry
usually used to represent a cell is drawn in black and consists of myoplasmic (myo) resistors and excitable elements (LRd) in parallel with a membrane capacitor. The cells
are linked by a resistor (blue), representing electrical coupling by gap junctions. For simulation of ephaptic transmission excitable elements are added to each cell (red),
which face the intercellular cleft and are connected to the extracellular space by a cleft and a radial resistor (red). B, Dependence of propagation velocity on cleft width at
different degrees of gap junction coupling. In the presence of normal gap junction coupling, the field effect caused by current flow through the Na+ channels facing the
cleft produces propagation slowing. At a large degree of cell-to-cell uncoupling (0.3% to 5%), propagation is markedly slowed or blocked at a cleft width greater than 40
to 50 µm. At a smaller width, propagation resumes because of ephaptic transmission. This dependence occurs if 100% of the Na channels reside in the intercellular cleft.
(From Kucera JP, Rohr S, Rudy Y: Localization of sodium channels in intercalated disks modulates cardiac conduction. Circ Res 91:1176–1182, 2002.)
might have a pivotal role in the determination of the disease complexes (dystrophin-syntrophin vs. SAP97). Whereas some
phenotype. Although the mechanism of associated modulation of controversies exist whether these pools all contain active chan-
Cx43 has not yet been clarified, earlier and recent work suggests nels65 and the observed decrease in INa is due to changes in
that the presence of an intact fascia adherens and desmosomal inactivation kinetics rather a decrease in Nav1.5 expression,64
apparatus is a prerequisite for gap junction formation. Thus, these observations suggest that modification of connexins in the
neoformation of gap junctions in cultured adult rat cardiomyo- intercalated disc might have an effect on expression or function
cytes is preceded by the formation of adherens junctions in a of Na+ channels in the intercalated disc. Indeed, such a relation-
spatiotemporally defined manner.60-62 This observation is prob- ship was recently observed in both ventricular and atrial tissues.32,66
ably explained by the finding that delivery of connexin hemichan- In the ventricle, tamoxifen-induced ablation of Cx43 was associ-
nels trafficking along microtubules63 to the periphery of the gap ated with local absence of Nav1.5 immunosignal, and silencing
junction plaques7 requires tethering of microtubules to β-catenin, of Cx43 led to a marked decrease of inward Na current.32 A
via the microtubule plus end-tracking protein EB1 and the inter- similar finding was obtained in cell cultures of neonatal mouse
mediate glue protein p150.8 The high probability that mechani- atrial myocytes, where ablation of Cx43 showed an approximately
cal, gap junction, and ion channel proteins are at least partially 50% decrease of INa—that is, a decrease similar to the ablation
regulated in common is also supported by the observation that of the intercalated disc pool of Nav1.5.66 The observation that,
silencing of plakophilin-2 is associated with an approximately at least in some conditions, remodeling of connexins can be
50% decrease in flow of Na inward current.64 Nav1.5 channels associated with a decrease in Na channels, makes it difficult to
in adult rat ventricular myocytes reside within two different correlate changes in Cx expression directly to changes in propa-
membrane compartments: the intercalated disc and the surface gation velocity, without information about concomitant changes
membrane. Each of these pools has different anchoring in Nav1.5. Some studies involving marked electrical uncoupling
of myocyte strands with genetic ablation or drug inhibition of eling with changes in the intercellular Kir2.1 pool, and accord-
27
gap junction channel conductance have shown that a marked ingly an effect on propagation velocity seems most likely.
decrease in gap junction coupling is necessary to observe signifi-
cant changes in propagation velocity22,28 in accordance with theo-
retical experiments.27 However, other studies have suggested that
relatively moderate changes in local Cx expression lead to a Summary
measurable decrease in propagation velocity.67 Therefore, the
possibility needs to be considered that Na+ channel expression, The recent findings of interactions among and common regula-
which might vary according to experimental and pathologic con- tion of mechanical junction proteins, ion channels, and connex-
ditions, might accompany changes in connexin expression and ins, possibly at the level of trafficking, indicate the difficulty in
associated changes in propagation velocity. An additional poten- explaining the importance of remodeling of connexins and gap
tial complexity related to the finding that compartmentation has junction channels for propagation, within a complex molecular
also been observed for the inward rectifier channel Kir2.1,68 phenotype. Moreover, at least one level of complexity is added
which is a determinant of the late phase of repolarization of the by the functional interaction of electrical cell-to-cell coupling,
action potential and largely determines the resting electrical con- cellular network architecture and ion current flow in producing
ductance, a further determinant of excitability and propagation.14 linear or circulating electrical propagation waves. Further experi-
Both Nav1.5 and Kir2.1 appear to be associated with SAP97 mental and theoretical experiments will be necessary to integrate
in the intercalated disc.68 Thus, a correlation of connexin remod- these findings into a common dynamic system.
propagation in synthetic strands of neonatal and tional cx43 knockout mice. Heart Rhythm 9:600–
References fetal murine cardiomyocytes. Circ Res 99:1216– 607, 2012.
1224, 2006. 33. Kitamura H, Ohnishi Y, et al: Heterogeneous loss
1. Weidmann S: Heart: Electrophysiology. Annu Rev 18. Gu H, Smith FC, et al: High incidence of cardiac of connexin43 protein in nonischemic dilated car-
Physiol 36:155–169, 1974. malformations in connexin40-deficient mice. Circ diomyopathy with ventricular tachycardia. J Car-
2. Cottrell GT, Burt JM: Functional consequences of Res 93:201–206, 2003. diovasc Electrophysiol 13:865–870, 2002.
heterogeneous gap junction channel formation and 19. Valiunas V, Gemel J, et al: Gap junction channels 34. Danik SB, Liu F, et al: Modulation of cardiac gap
its influence in health and disease. Biochim Biophys formed by coexpressed connexin40 and conjunction expression and arrhythmic susceptibility.
Acta 1711:126–141, 2005. nexin43. Am J Physiol Heart Circ Physiol Circ Res 95:1035–1041, 2004.
3. Harris AL: Emerging issues of connexin channels: 281:H1675–H1689, 2001. 35. Beauchamp P, Desplantez T, McCain ML, et al:
Biophysics fills the gap. Q Rev Biophys 34:325– 20. Cottrell GT, Burt JM: Heterotypic gap junction Electrical coupling and propagation in engineered
472, 2001. channel formation between heteromeric and ventricular myocardium with heterogeneous
4. Moreno AP: Biophysical properties of homomeric homomeric CX40 and CX43 connexons. Am J expression of connexin43. Circ Res 110;1445–
and heteromultimeric channels formed by cardiac Physiol Cell Physiol 281:C1559–67, 2001. 1453, 2012.
connexins. Cardiovasc Res 62:276–286, 2004. 21. Johnson CM, Kanter EM, et al: Redistribution of 36. Elenes S, Martinez AD, et al: Heterotypic docking
5. Spray DC, Bennett MV: Physiology and pharma- connexin45 in gap junctions of connexin43- of cx43 and cx45 connexons blocks fast voltage
cology of gap junctions. Annu Rev Physiol 47:281– deficient hearts. Cardiovasc Res 53:921–935, 2002. gating of cx43. Biophys J 81:1406–1418, 2001.
303, 1985. 22. Beauchamp P, Choby C, et al: Electrical propaga- 37. Desplantez T, Halliday D, et al: Cardiac connexins
6. Spray DC, White RL, et al: Regulation of gap tion in synthetic ventricular myocyte strands from cx43 and cx45: Formation of diverse gap junction
junctional conductance. Am J Physiol 248:H753– germline connexin43 knockout mice. Circ Res channels with diverse electrical properties. Pflugers
H764, 1985. 95:170–178, 2004. Arch 448:363–375, 2004.
7. Gaietta G, Deerinck TJ, et al: Multicolor and elec- 23. Yamada KA, Rogers JG, et al: Up-regulation of 38. McCain ML, Desplantez T, et al: Cell-to-cell
tron microscopic imaging of connexin trafficking. connexin45 in heart failure. J Cardiovasc Electro- coupling in engineered pairs of rat ventricular car-
Science 296:503–507, 2002. physiol 14:1205–1212, 2003. diomyocytes: Relation between cx43 immunofluo-
8. Shaw RM, Fay AJ, et al: Microtubule plus-end- 24. Grikscheit K, Thomas N, et al: Coexpression of rescence and intercellular electrical conductance.
tracking proteins target gap junctions directly from connexin 45 with connexin 43 decreases gap junc- Am J Physiol Heart Circ Physiol 302:H443–H450,
the cell interior to adherens junctions. Cell tion size. Cell Commun Adhes 15:185–193, 2008. 2012.
128:547–560, 2007. 25. Lisewski U, Shi Y, et al: The tight junction protein 39. Gutstein DE, Morley GE, et al: Conduction
9. Kleber AG: Gap junctions and conduction of cardiac car regulates cardiac conduction and cell-cell com- slowing and sudden arrhythmic death in mice with
excitation. Heart Rhythm 8:1981–1984, 2011. munication. J Exp Med 205:2369–2379, 2008. cardiac-restricted inactivation of connexin43. Circ
10. Davis LM, Rodefeld ME, et al: Gap junction protein 26. Kleber AG, Rudy Y: Basic mechanisms of cardiac Res 88:333–339, 2001.
phenotypes of the human heart and conduction impulse propagation and associated arrhythmias. 40. Caldwell BJ, Trew ML, et al: Three distinct direc-
system. J Cardiovasc Electrophysiol 6:813–822, 1995. Physiol Rev 84:431–488, 2004. tions of intramural activation reveal nonuniform
11. Saffitz JE, Davis LM, et al: The molecular basis of 27. Shaw RM, Rudy Y: Ionic mechanisms of propaga- side-to-side electrical coupling of ventricular myo-
anisotropy: Role of gap junctions. J Cardiovasc tion in cardiac tissue. Roles of the sodium and cytes. Circ Arrhythm Electrophysiol 2:433–440,
Electrophysiol 6:498–510, 1995. l-type calcium currents during reduced excitability 2009.
12. van Veen AA, van Rijen HV, et al: Cardiac gap and decreased gap junction coupling. Circ Res 41. Gilbert SH, Benoist D, et al: Visualization and
junction channels: Modulation of expression and 81:727–741, 1997. quantification of whole rat heart laminar structure
channel properties. Cardiovasc Res 51:217–229, 28. Rohr S, Kucera JP, et al: Slow conduction in using high-spatial resolution contrast-enhanced
2001. cardiac tissue, i: Effects of a reduction of excit- mri. Am J Physiol Heart Circ Physiol 302:H287–
13. Bukauskas FF, Kreuzberg MM, et al: Properties of ability versus a reduction of electrical coupling on H298, 2012.
mouse connexin 30.2 and human connexin 31.9 microconduction. Circ Res 83:781–794, 1998. 42. Pope AJ, Sands GB, et al: Three-dimensional
hemichannels: Implications for atrioventricular 29. Wang Y, Rudy Y: Action potential propagation in transmural organization of perimysial collagen in
conduction in the heart. Proc Natl Acad Sci U S A inhomogeneous cardiac tissue: Safety factor con- the heart. Am J Physiol Heart Circ Physiol
103:9726–9731, 2006. siderations and ionic mechanism. Am J Physiol 295:H1243–H1252, 2008.
14. Kleber AG, Janse MJ, et al: Normal and abnormal Heart Circ Physiol 278:H1019–H1029, 2000. 43. Spach MS, Miller WT, 3rd, et al: The functional
conduction in the heart. The handbook of physiol- 30. Gutstein DE, Morley GE, et al: Heterogeneous role of structural complexities in the propagation of
ogy. 455–529, 2000. expression of gap junction channels in the heart depolarization in the atrium of the dog. Cardiac
15. Vozzi C, Dupont E, et al: Chamber-related differ- leads to conduction defects and ventricular dys- conduction disturbances due to discontinuities of
ences in connexin expression in the human heart. function. Circulation 104:1194–1199, 2001. effective axial resistivity. Circ Res 50:175–191,
J Mol Cell Cardiol 31:991–1003, 1999. 31. Severs NJ, Bruce AF, et al: Remodelling of gap 1982.
16. Kanagaratnam P, Rothery S, et al: Relative expres- junctions and connexin expression in diseased myo- 44. Dolber PC, Spach MS: Thin collagenous septa in
sion of immunolocalized connexins 40 and 43 cor- cardium. Cardiovasc Res 80:9–19, 2008. cardiac muscle. Anat Rec 218:45–55, 1987.
relates with human atrial conduction properties. 32. Jansen JA, Noorman M, et al: Reduced heteroge- 45. Spach MS, Heidlage JF, et al: Mechanism of origin
J Am Coll Cardiol 39:116–123, 2002. neous expression of cx43 results in decreased of conduction disturbances in aging human atrial
17. Beauchamp P, Yamada KA, et al: Relative contribu- nav1.5 expression and reduced sodium current that bundles: Experimental and model study. Heart
tions of connexins 40 and 43 to atrial impulse accounts for arrhythmia vulnerability in condi- Rhythm 4:175–185, 2007.
46. Spach MS, Kootsey JM: The nature of electrical 55. Mori Y, Fishman GI, et al: Ephaptic conduction in 62. Kostin S, Hein S, et al: Spatiotemporal develop-
propagation in cardiac muscle. Am J Physiol a cardiac strand model with 3d electrodiffusion. ment and distribution of intercellular junctions in
244:H3–22, 1983. Proc Natl Acad Sci U S A 105:6463–6468, 2008. adult rat cardiomyocytes in culture. Circ Res
47. Rohr S, Kucera JP, Fast VG, et al: Paradoxical 56. Petitprez S, Zmoos AF, et al: Sap97 and dystrophin 85:154–167, 1999.
improvement of impulse conduction in cardiac macromolecular complexes determine two pools of 63. Smyth JW, Vogan JM, et al: Actin cytoskeleton rest
tissue by partial cellular uncoupling. Science cardiac sodium channels nav1.5 in cardiomyocytes. stops regulate anterograde traffic of connexin 43
275:841–844, 1997. Circ Res 108:294–304, 2011. vesicles to the plasma membrane. Circ Res
48. Fast VG, Kleber AG: Block of impulse propagation 57. Delmar M: Connexin43 regulates sodium current; 110:978–989, 2012.
at an abrupt tissue expansion: Evaluation of the ankyrin-g modulates gap junctions: The interca- 64. Sato PY, Musa H, et al: Loss of plakophilin-2
critical strand diameter in 2- and 3-dimensional lated disc exchanger. Cardiovasc Res 93:220–222, expression leads to decreased sodium current and
computer models. Cardiovasc Res 30:449–459, 2012. slower conduction velocity in cultured cardiac
1995. 58. Delmar M, Liang FX: Connexin43 and the regula- myocytes. Circ Res 105:523–526, 2009.
49. Kucera JP, Kleber AG, et al: Slow conduction in tion of intercalated disc function. Heart Rhythm 65. Lin X, Liu N, et al: Subcellular heterogeneity of
cardiac tissue, II: Effects of branching tissue geom- 9:835–838, 2012. sodium current properties in adult cardiac ventricu-
etry. Circ Res 83:795–805, 1998. 59. Asimaki A, Tandri H, et al: A new diagnostic test lar myocytes. Heart Rhythm 8:1923–1930, 2011.
50. Weidmann S: Cardiac action potentials, membrane for arrhythmogenic right ventricular cardiomyopa- 66. Desplantez T, McCain ML, et al: Connexin43
currents, and some personal reminiscences. Annu thy. N Engl J Med 360:1075–1084, 2009. ablation in foetal atrial myocytes decreases electri-
Rev Physiol 55:1–14, 1993. 60. Geisler SB, Green KJ, et al: Ordered assembly of cal coupling, partner connexins, and sodium
51. Barr L, Dewey MM, et al: Propagation of action the adhesive and electrochemical connections current. Cardiovasc Res 94:58–65, 2012.
potentials and the structure of the nexus in cardiac within newly formed intercalated disks in primary 67. Eloff BC, Lerner DL, et al: High resolution optical
muscle. J Gen Physiol 48:797–823, 1965. cultures of adult rat cardiomyocytes. J Biomed Bio- mapping reveals conduction slowing in connexin43
52. Sperelakis N, Hoshiko T, et al: Nonsyncytial technol 2010:624719, 2010. deficient mice. Cardiovasc Res 51:681–690, 2001.
nature of cardiac muscle: Membrane resistance of 61. Hertig CM, Butz S, et al: N-cadherin in adult rat 68. Milstein ML, Musa H, et al: Dynamic reciprocity
single cells. Am J Physiol 198:531–536, 1960. cardiomyocytes in culture. Ii. Spatio-temporal of sodium and potassium channel expression in a
53. Danik SB, Rosner G, et al: Electrical remodeling appearance of proteins involved in cell-cell contact macromolecular complex controls cardiac excit-
contributes to complex tachyarrhythmias in and communication. Formation of two distinct ability and arrhythmia. Proc Natl Acad Sci U S A
connexin43-deficient mouse hearts. FASEB J n-cadherin/catenin complexes. J Cell Sci 109(Pt 2012.
22:1204–1212, 2008. 1):11–20, 1996.
54. Kucera JP, Rohr S, Rudy Y: Localization of sodium
channels in intercalated disks modulates cardiac
conduction. Circ Res 91:1176–1182, 2002.
Mechanisms of Atrioventricular Nodal
Excitability and Propagation 28
Hye Jin Hwang, Fu Siong Ng, and Igor R. Efimov
profoundly complex structure. The successful optical imaging of

CHAPTER OUTLINE
the AV junction (AVJ) in the human heart has provided important
Anatomy and Molecular Characteristics insights into its electrophysiology. Three-dimensional histologic
of the Mammalian Atrioventricular Node 275 and immunofluorescence reconstructions of the AVJ, as well as
molecular biological studies looking into gene and protein
Transitional Cells 275 expression, have helped to demonstrate the significant structural
Inferior Nodal Extension 276 and molecular heterogeneity of the AVJ. This chapter reviews the
current state of knowledge of AVJ structure and function with
Molecular Characteristics of Transitional Cells specific reference to these recent discoveries in the human heart.
and Inferior Nodal Extension Cells 277
Anatomy of the Compact Atrioventricular Node
and the Bundle of His 277 Anatomy and Molecular Characteristics
Molecular Characteristics of the Compact of the Mammalian Atrioventricular Node
Atrioventricular Node and Bundle of His 277
Functional Heterogeneity of the Atrioventricular
Anatomy of the Triangle of Koch and
Junction: Atrionodal Cells, Nodal Cells, and Atrioventricular Junction Structures
Nodo-His Cells 277 The cardiac pacemaking and conduction system comprises spe-
Dual-Pathway Electrophysiology of the cialized cells composed of pacemaker cells that generate electrical
Atrioventricular Node: Slow and Fast Pathways 277 impulses as well as a His-Purkinje system that rapidly conducts
the electrical impulses to enable properly timed synchronous
His Bundle Activation: Longitudinal Dissociation myocardial contraction. The electrical impulse generated from
of Atrioventricular Junctional Conduction 278 the sinoatrial node (SAN), embedded at the junction of superior
Antegrade and Retrograde Conduction Properties vena cava and the right atrium, exits the SAN via sinoatrial exit
of the Atrioventricular Junction 280 pathways, and travels through the right atrium via the crista
terminalis and atrial septum to reach the compact AVN located
Decremental Conduction and Wenckebach Periodicity 281 at the apex of the triangle of Koch. As shown in Figure 28-1, A,
Atrioventricular Junctional Pacemaker 282 the triangle of Koch is defined as the area enclosed by the septal
leaflet of the tricuspid valve, the ostium of coronary sinus, and
Autonomic Innervation of the Atrioventricular Junction 282 the tendon of Todaro. The apex point of the triangle of Koch is
Optical Mapping of the Human Atrioventricular Junction 282 formed by the membranous portion of the ventricular septum,
which is another anatomic landmark for the AVN.5 Cells of the
Conclusions 284 compact AVN at the apex of the triangle of Koch are small and
spindle shaped, with no clear cellular orientation. In contrast,
atrial myocardial cells are large, densely packed, and oriented
The atrioventricular node (AVN) is a critical component of the parallel with one another. Two distinct cell populations exist
cardiac conduction system, primarily responsible for the conduc- between cells of the atrial myocardium and those of the compact
tion of electrical impulses from the atria to the ventricles, and AVN: (1) transitional cells with intermediate features between the
has a secondary function as an escape pacemaker. Since its dis- atrial myocytes and spindle-shaped compact AVN, and (2) cells
covery by Tawara a little more than a century ago, many impor- of one or two inferior nodal extension(s), which are regarded as
tant advances have been made in terms of understanding the extension(s) of the compact AVN. These two cell types have
complex structure and function of the AV node, as well as its role traditionally been regarded as being responsible for the two atrial
in cardiac physiology and pathophysiology. In 1906, Tawara first electrical inputs into the compact AVN; transitional cells have
discovered the spindle-shaped compact network of small cells, been thought to compose the “fast pathway,”6,7 whereas cells of
which he called a knoten (node).1 Subsequently, Mines provided inferior nodal extensions constitute the “slow pathway.”8,9
the first clear and elegant description of reentry involving the
AVN in 1916.2 More than half a century later, Moe et al. described
the existence of dual-pathway AV nodal physiology and demon-
strated the slow and fast pathways with AV nodal reentrant tachy- Transitional Cells
cardia (AVNRT) in the dog,3 and Janse et al.4 subsequently
confirmed the same in the rabbit. The transitional cells extend anteriorly from the anterior limbus
Despite a long history of research, many aspects of AV nodal of fossa ovalis, located just above the noncoronary cusp of the
physiology remain unclear, and some controversies surrounding aortic valve, and cover the middle and anterior part of the triangle
its nature remain unresolved. Over the past decade, important of Koch after crossing the tendon of Todaro, before connecting
strides have been made to enhance our understanding of this to the proximal portion of the compact AVN. Such a distribution
275
AVJ 3D model Histology sections
S
1 2 3
IA
IAS IAS IAS
s
IAS
Hi
Aortic
FP
valve
LN N
C
B
LE AV nodal
VS
TT LE
artery
TC
His CN
pid
RE
His TA 1 mm 1 mm
us
1 mm
Tric
TA RE
TA
S
C
TA
IVS
SP
IVS IVS IVS
A B Atria Connective tissue His bundle Ventricle
40-year-old CFB 70-year-old Connective

male 58-year-old VS male VS tissue
female S His
Tricuspid valve
Tricuspid valve
IAS P A Compact AVN/
Tricuspid valve
IAS CFB IAS CFB
1.2 cm
I leftward extension
1.2 cm
S LNB/rightward
1.2 cm
VS
TT
P A extension
TT
I Inferior transitional
TT
cells
S Artery
3 mm P A 3 mm 3 mm Vein
C CS
I
CS
CS
Figure 28-1. Triangle of Koch and AVJ three-dimensional model of the human heart. A, Schematic of the triangle of Koch and AVJ structures. The AV nodal portion is
composed of the compact AV node (CN) and the lower nodal bundle (LNB). The right inferior nodal extension (RE) extends proximally from the lower nodal bundle and
the left inferior nodal extension (LE) extends from the compact AV node. B, Three-dimensional anatomic reconstruction of the AVJ and histologic sections through the
bundle of His, the compact AV node, and extensions (RE and LE). C, Three-dimensional reconstruction of the AVJ, conduction system. AVJ, Atrioventricular junction; TA,
tricuspid annulus; IAS, interatrial septum; VS, ventricular septum; FP, fast pathway; SP, slow pathway; CN, compact AV node; CS, coronary sinus; CFB, central fibrous body; TT,
tendon of Todaro; A-P, anterior–posterior; S-I, superior–inferior.
(A, Reproduced with permission from Kurian T, Ambrosi C, Hucker W, et al: Anatomy and electrophysiology of the human AV node. Pacing Clin Electrophysiol 33:754–762, 2010.
B, Reproduced with permission from Fedorov VV, Ambrosi CM, Kostecki G, et al: Anatomic localization and autonomic modulation of atrioventricular junctional rhythm in failing
human hearts. Circ Arrhythm Electrophysiol 4:515–525, 2011. C, Reproduced with permission from Hucker WJ, McCain ML, Laughner JI, et al: Connexin 43 expression delineates
two discrete pathways in the human atrioventricular junction. Anat Rec [Hoboken] 291:204–215, 2008.)
of transitional cells enables the electrical impulse to spread inside potential in the ablation of AVNRT. The site where the double
the triangle of Koch without conduction block by the tendon of potential consists of a low-frequency deflection (nodal compo-
Todaro.10-12 The area of transitional cells distributed at the right nent) followed by high-frequency deflection (atrial component),
anterior interatrial septum just outside the tendon of Todaro and as described by Jackman and colleagues,19 corresponds approxi-
behind the bundle of His corresponds approximately to the fast mately to the distal part of the INE beneath the orifice of the
pathway site that was targeted during early attempts to cure coronary sinus.20 The double potential of high-frequency fol-
AVNRT using radiofrequency ablation.6,13 Transitional cells have lowed low-frequency deflections described by Haissaguerre
also been observed in the left atrial septum,14 supporting clinical et al.21 may be generated by the two populations of cells in the
observations of earliest atrial activation at that site during retro- middle and anterior parts of the triangle of Koch.20
grade fast pathway conduction. Although transitional cells have There are important species differences in the anatomy of the
traditionally been considered to constitute the “fast pathway,” AVJ. In contrast to rabbits, which have only one INE, humans
experimental observations do not confirm fast conduction veloc- typically have two INEs,4,8 which are referred to as the rightward
ity in this pathway.15-17 A shorter conduction delay in this pathway and leftward extensions. As shown in a study of the anatomy of
is primarily caused by anatomically shorter distance compared the human AVJ, there is significant anatomic variability in
with the “slow pathway.” human.8 The majority of humans have two INEs (13 of 21
hearts), some have only the rightward INE (7 of 21 hearts), and,
rarely, others have only the leftward INE.8 The leftward INE in
humans extends from the compact AVN and lies more superiorly
Inferior Nodal Extension to and is usually shorter than the rightward extension (see Figure
28-1, B).8 Interestingly, three-dimensional reconstruction of the
The inferior nodal extension (INE) lies parallel to the superior human AVJ shows age-related changes of the INEs and the
portion of the tricuspid annulus within the triangle of Koch, compact AVN (see Figure 28-1, C). Compared with the left
extends towards the coronary sinus, and then merges with the extension, the length of the right extension increases with age,
compact AVN (see Figure 28-1, B). The distal portion of the INE accompanied by a widening of transitional cell zone.22 Evidence
contacts atrial myocytes directly. In the middle of the triangle of of age-related change of INEs may explain the high incidence of
Koch, loosely packed transitional cells and the INE converge, AVNRT within an older aged population compared with children
and transitional cells overlie the INE at the anterior part of the (i.e., >10 years of age compared with <5 years of age). The right-
triangle of the Koch.14,18 The anatomic relationship between the ward INE in humans has been regarded to constitute the slow
transitional cells and the INE within the triangle of Koch may pathway, yet the existence of the leftward extension should also
explain the two different types of double potentials observed on be noted, because it is associated with the atypical left variant of
intracardiac electrograms while targeting the slow pathway AVNRT.23
Mechanisms of Atrioventricular Nodal Excitability and Propagation 277
known to change from an oval shape to a spindle shape as age

Molecular Characteristics of Transitional Cells
28
increases (see Figure 28-1, C).22
and Inferior Nodal Extension Cells
Histologically, the transitional cells and cells of the INE are rela-
tively small and dispersed among connective tissues, and although Molecular Characteristics of the Compact
these two populations of cells share some similar features, there Atrioventricular Node and Bundle of His
are important molecular and electrophysiologic differences
between these cells. It has recently been demonstrated in human Both the compact AVN and His bundle express high quantities
tissue that the transitional cells and the INE cells both express of connexin40 mRNA and protein, which are gap junction pro-
the intermediate levels of Tbx3, a transcription factor that regu- teins that form large conductance channels, whereas Cx43 is
lates the development of the cardiac conduction system com- scantly expressed in the compact AVN and expressed at an inter-
pared with the compact AVN.14,24 However, in rabbit, transitional mediate level in the lower nodal bundle/His bundle (Figure
cells stain negative for neurofilament, and INE cells are 28-2).24,28 mRNA levels for Nav1.5, the cardiac sodium channel
neurofilament positive, suggesting that these two cell populations protein, and a determinant of conduction velocity are scant in the
have different embryologic origins.14,16 An important electro- compact AVN (Figure 28-3, A) compared with their high expres-
physiologic difference is that, compared with atrial cells, transi- sion levels in the His bundle.24 The abundant expression of con-
tional cells express connexin43 (Cx43), an important cardiac gap nexin proteins and Nav1.5 mRNA in His bundle enables the
junction protein, at the intermediate level, whereas the cells of electrical wave to pass rapidly through the His bundle.
the compact AVN and the leftward INE scantly express Cx43.
However, detailed histologic and molecular discrimination
between the rightward and the leftward INE in the human AVJ Functional Heterogeneity of the
is currently lacking.
Another important difference is that pacemaker activity is Atrioventricular Junction: Atrionodal Cells,
observed in INE cells but not in transitional cells in rabbit Nodal Cells, and Nodo-His Cells
studies.25 This is supported by gene expression data from humans
showing that messenger RNA (mRNA) levels for HCN4, respon- As described in the previous section, the AVJ is composed of
sible for the hyperpolarization-activated “funny” (If) current, and numerous components, with different cell types that possess dif-
Cav1.3, an alternative L-type calcium-channel isoform, are ferent cell morphologies and varying expression of sarcolemmal
highly abundant in the INE, whereas expression of SCN5A, the ion channels and cardiac connexins. The effect of this variability
gene encoding for the cardiac sodium channel protein Nav 1.5, of cell types is the marked functional heterogeneity within the
is low.24 In contrast, the expression of HCN4 and Cav1.3 is lower AVJ. Characteristics of the action potential for the different cells
in transitional cells, and the expression of SCN5A is high.24 The in the AVJ have previously been demonstrated in the rabbit heart
INE cells have slower Ca2+-dependent action potential upstrokes, through the combined study of microelectrodes and histology.30
like the nodal cells, whereas transitional cells have fast sodium Significant heterogeneity of action potential morphology in the
current–dependent upstrokes. This may account for relatively rabbit AVJ was found,31-33 and three main types of cells have been
slow conduction across the INE. These differences in gene described: atrionodal (AN), nodal (N), and nodo-His (NH) cells,
expression may explain why sodium-channel blocking Class IA according to the morphology of the action potential (see Figure
and IC antiarrhythmic drugs are effective in blocking the fast 28-3, B). Compared with characteristic atrial myocardial cells, N
pathway, whereas L-type calcium-channel blockers such as vera- cells in the compact AVN have less negative resting potentials,
pamil are effective in blocking the slow pathway.26 The presence smaller amplitudes, and slower upstrokes, and they possess pace-
of nodal-like cells in the INE could also explain the occurrence maker activity. AN cells have intermediate features between the
of junctional rhythms during radiofrequency application at the atrial cells and the N cells—that is, more negative resting mem-
inferior triangle of Koch.27 brane, steeper action potential upstroke, and large amplitude
compared with the N cells. NH cells also show intermediate elec-
trophysiologic characteristics between N cells and the His bundle
cells. These cells also have more negative resting potentials and
Anatomy of the Compact Atrioventricular steeper action potential upstrokes compared with N cells.
Node and the Bundle of His Attempts have been made to correlate the functional characteris-
tics to the various cell populations within the AVJ with specific
The AVN and the His bundle are demarcated by the central molecular and morphological characteristics, and it appears that
fibrous body, as initially described by Tawara.1 Morphologically18 AN cells are the transitional cells, N cells are in the compact AVN,
and based on Cx43 expression,28,29 the AVN can be subdivided into and NH cells are in the lower nodal bundle. To date, these obser-
the lower nodal bundle and the compact AVN. The compact node vations of variability of action potential morphology in animal
is the superior portion of the AVN connected to transitional tissue models remain unconfirmed in the human AVJ.
and the INEs, which was previously referred to as the “open
node” by earlier investigators.18 The lower nodal bundle indicates
the inferior portion that is in some cases enveloped by connective Dual-Pathway Electrophysiology
tissue, which was previously referred to as the “closed node.” Cells
of the lower nodal bundle are longer and arranged more parallel of the Atrioventricular Node:
to each other than in the compact AVN and possess an intermedi- Slow and Fast Pathways
ate functional phenotype between that of the compact AVN and
the His bundle. Some investigators have recently described the It has been demonstrated both in experimental studies3 and in
lower nodal bundle as the penetrating bundle of His because it is the clinical setting that there are often two conduction pathways
enclosed by connective tissue.5 The lower nodal bundle has been through the AVJ, known as the “fast pathway” and the “slow
shown to connect to the rightward INE, whereas the leftward pathway,” which reference the conduction delay with which these
INE and the compact AVN are a continuous structure (see Figure pathways conduct an electrical impulse. It has been thought that
28-1, C).28 Regarding age-related change, the compact AVN is conduction longitudinally along the anterior interatrial septum
IAS
Left atrial D Compact
TCs AVN
Endocardial E
TCs
CFB Cx43 Cx43
α-actinin
vimentin
AVN F
P
LNB
A
1 mm
G
VS 300 µm
A 300 µm
B C
Left TCs Endo TCs Compact AVN LNB
100 µm
D E F G
Figure 28-2. Connexin43 (Cx43) expression of the AVJ. A, Masson’s trichrome stain of the AV node. The outlined area surrounding the AV node corresponds to immuno-
histochemistry shown in B and C. B, Immunohistochemistry of the AV node showing α-actinin in red, vimentin in blue, and Cx43 in green. C, Cx43 expression in the AV
node. D-G, Higher magnification of Cx43, vimentin, and α-actinin expression in various areas of the AV node region.
(Reproduced with permission from Hucker WJ, McCain ML, Laughner JI, et al: Connexin 43 expression delineates two discrete pathways in the human atrioventricular junction.
Anat Rec [Hoboken] 291:204–215, 2008.)
represents the fast pathway (anterior input) because earliest ret- anterior septum and INE are thought to be the fast and slow path-
rograde atrial activation is observed at the anterior interatrial ways, respectively. Clinically, there are two distinct sites of earliest
septum or near the bundle of His during right ventricular pacing retrograde atrial activation (the anterior septum near His and near
and during slow-fast AVNRT.13 Transverse conduction along the the ostium of coronary sinus) during AVNRT or ventricular
triangle of Koch has been thought to represent the slow pathway pacing, which has contributed to the concept that the input
(posterior input) because the inferior triangle of Koch is often through the anterior atrial septum represents the fast pathway and
the site of successful slow pathway ablation to eliminate typical the input through the inferior triangle of Koch represents the slow
AVNRT.19 pathway. Functionally, the traditional terminology of the slow
However, attempting to correlate these functional pathways pathway (posterior input) and the fast pathway (anterior input)
to specific cell types and to clearly identify them on histology has may not necessarily reflect differences in conduction velocity.
proved to be challenging. Optical coherence tomography of Visualization of electrical propagation by optical mapping has
intact rabbit AVJ has demonstrated a great degree of variability revealed that the wavefront during sinus rhythm or electrical
of cell orientation at different depths.29 Because the transitional stimulation at the right atrium propagates broadly toward the
cells were found at the anterior interatrial septum, it has been AVN rather than proceeding via two electrically distinguishable,
thought that the transitional cells are responsible for the fast isolated pathways.12,15,17 This finding suggests that, at least during
pathway. Similarly, the INE cells found within the triangle of AVNRT, the atrial myocardium/transitional cells enveloping the
Koch have been thought to represent the slow pathway. In reality, areas of slow conduction, in which the entry or exit sites are the
transitional cells are widely distributed across the triangle of anterior interatrial septum or the inferior triangle of Koch, may
Koch and coexist with the INEs within the triangle of Koch and play the role of true fast pathway.16 Furthermore, it should be
are not located exclusively at the anterior interatrial septum. noted that because techniques of visualizing electrical propagation
Similarly, correlation of the functional slow pathway to specific have developed and improved, they may be better considered as
cells or histology may be difficult because the posterior inputs to different distances of electrical activation pathways.
the INE exhibit profound functional complexity, and the triangle
of Koch includes not only INE but also transitional cells covering
INE and the lower nodal bundle beneath INE.14,28 Such complex His Bundle Activation: Longitudinal
distributions of different cell types surrounding the AVJ may give
rise to the conceptual confusion of what constitutes the slow and Dissociation of Atrioventricular
the fast pathways. Junctional Conduction
Histologically, two inputs (atrial cells and transitional cells), the
INEs and the compact AVN, and two His compartments (the Previous rabbit and human AVN studies have demonstrated that
lower nodal bundle and His bundle, respectively) exist in the fast pathway and the slow pathway differentially activate the
the human heart. As described before, the transitional cells in the His bundle.29,34 Fast pathway activation has been shown to engage
Lipofuscin protein HCN4 Control junctional rhythm 30 bmp

Compact node Ventricular muscle IAS
s
Retrograde
1-CN
Anterograde 28
FP Hi
5-FP
CN
IVS 2-LE
LE
RE
CS SP 3-RE
A TA
4-SP
B 0 500 1000 ms
Nav1.5
IAS Retrograde IAS Anterograde 900
Compact node Atrial muscle 250
800
His 200 His
700
150
600
IVS 100 IVS 500

CS 50 CS 400
C TA D TA 300
Cav3.1
Compact node Atrial muscle SCL: 840 ms, AH: 80 ms, HH’: 40 ms–90 ms, H’V: 40 ms 500 ms
Lead II
HRA
HH’
HBE
75 80 85 85 90
20 µm 500 ms
A
Lead II
HRA H H’ H HH’
Atrial AN N NH His HBE
90 95 H - no H’ 40 85
E
B Figure 28-4. Atrioventricular (AV) junctional beat with a retrograde reentrant beat
and intra-His Wenckebach block. A, Photograph of the coronary-perfused human
Figure 28-3. Action potentials and immunofluorescence labeling of ion channels.
AVJ preparation with an optical field of view that included compact node (CN),
A, Immunofluorescence labeling of HCN4, Nav1.5, and Cav3.1 of human atrioven-
leftward (LE), and rightward extensions (RE), and the bundle of His (His). B, Optical
tricular (AV) node and working myocardium. HCN4 (top; green signal) is expressed
action potentials, from the five locations indicated in A, showing retrograde and
at the compact AV node (primarily within the cell membrane) but is less expressed
anterograde activation separated by the vertical dashed line. C and D, Activation
in the ventricular muscle. Nav1.5 (middle; red signal) is expressed in the atrial muscle
maps of retrograde and anterograde activations. Dashed arrows show the main
(primarily within the cell membrane) but is less expressed at the compact AV node.
conduction directions. E, Intracardiac electrograms recorded at the high right
Cav3.1 (bottom; green signal) is expressed at the compact AV node but is less
atrium (HRA) and His bundle (HBE), showing intra-hisian Wenckebach block with
expressed in the atrial muscle. Scale bars in each panel are shown in the bottom
progressive increase of the H-H′ interval, followed by block of the H′ potential. The
right corner. B, Action potential recordings from atrial, atrionodal, nodal, nodo-His,
lead II surface electrocardiogram is also shown.
and His cells at the rabbit AV junction.
(A-D, Reproduced with permission from Fedorov VV, Ambrosi CM, Kostecki G, et al:
(A, Reproduced with permission from Greener ID, Monfredi O, Inada S, et al: Molecular
Anatomic localization and autonomic modulation of atrioventricular junctional
architecture of the human specialised atrioventricular conduction axis. J Mol Cell
rhythm in failing human hearts. Circ Arrhythm Electrophysiol 4:515–525, 2011.
Cardiol 50:642–651, 2011. B, Reproduced with permission from De Carvalho AP, De
E, Provided by S.S. Kim, Yonsei University College of Medicine, Seoul, Korea.)
Almeida DF: Spread of activity through the atrioventricular node. Circ Res 8:801–809,
1960.)
the superior portion of the His bundle, whereas slow pathway pathway activation because the rightward INE connects directly
activation engages the inferior portion of the His bundle, sug- to the lower nodal bundle that lies beneath the compact AVN.
gesting the possible existence of functional longitudinal dissocia- Therefore, there is the possibility that activation of the slow
tion of conduction through the AVJ. Furthermore, it has been pathway can bypass the compact AVN and proceed directly to
shown that the time interval between pacing and His activation the lower nodal bundle and His bundle. Further evidence comes
by direct pacing of the slow pathway is shorter than by fast from optical mapping studies in human hearts, where two differ-
pathway activation involving transitional cells in animal studies, ent amplitudes of the bipolar His electrogram are seen depending
and it was in reverse relationship to the distance from His.29 Some on whether the slow pathway or the fast pathway is activated,
investigators postulate that, although conduction through the indicating that the His is differentially activated by these two
slow pathway is slower than that of the fast pathway, as supported pathways.17 Figure 28-4, B-D, shows reentrant echo beats using
by the gene expression profiles described before,35 direct pacing two dissociated pathways for antegrade conduction and retro-
of the slow pathway may activate the His bundle earlier than fast grade conduction. These experimental findings suggest that there
may be two functionally dissociated axes of the AV conduction observation may explain why a phenomenon known as the “AH
system: (1) the fast pathway: transitional cells–(leftward INE)— jump,” defined as the abrupt prolongation of the atrial-His (AH)
the compact AVN-His bundle; (2) the slow pathway: the atrial interval as a result of the large difference in conduction delay
cells/transitional cells–rightward INE—the lower nodal bundle– between the two pathways, is not common (<10% of the
His bundle.24,29,36-39 Intra-His Wenckebach block may be a clinical population).
manifestation of longitudinal dissociation of AV nodal conduc- Rate-dependent decreases in excitability have been observed
tion (see Figure 28-4, E). It should be noted that interactions in transitional cells located at the anterior interatrial septum.6,41
between two functionally dissociated axes may exist and these two Microelectrode experimental studies in animal models demon-
axes may not be completely electrically isolated because of their strated that as cycle length was decreased, the transitional cells
anatomic proximity. The presence of two longitudinal axes of AV in the anterior interatrial septum proximal to the AVN had less
conduction gives rise to potential clinical strategies to modulate negative resting membrane potentials and reduced action poten-
ventricular rate in atrial fibrillation.40 tial amplitude and upstroke compared with those of transitional
cells distal to the AVN.6,41 Thus, the anterior interatrial septum
(anterior input) has significantly steeper conduction velocity res-
titution compared with the posterior input.41 These differences
Antegrade and Retrograde Conduction in rate-dependent excitability and conduction velocity restitution
Properties of the Atrioventricular Junction between the anterior and posterior inputs cause conduction to
delay or block in the “fast pathway” at short cycle lengths,
Antegrade Atrioventricular Conduction enabling conduction across the slow pathway to reach the AVN
and thereby allowing AVNRT to initiate. These experimental
Antegrade AV conduction describes the propagation of electrical observations may explain how AVNRT can be initiated without
impulse from the atria to the ventricles through the AVJ. Experi- an abrupt conduction delay during continuous rapid pacing.
mental optical imaging studies, which can generate thousands of
simultaneous recordings, have helped to improve our under-
standing of AV conduction. Recently, the successful optical Retrograde VA Conduction
mapping of human AVJs has allowed, for the first time, the visu-
alization of the electrical activation patterns through the “slow” Retrograde VA conduction refers to conduction from the ven-
and “fast” pathways during atrial pacing (Figure 28-5).39 Com- tricles to the atria through the AVJ. Retrograde VA conduction
pared with the rapid electrical conduction over atrial tissue is not common during ventricular pacing in the electrophysiology
(up to 117 cm/s), conduction was significantly slower laboratory and is also infrequently seen during ventricular
across both the slow and fast pathways, and the action potential arrhythmias such as premature ventricular beats, ventricular
upstrokes in both pathways were slower than that of atrial tissue tachycardia, and accelerated idioventricular rhythm. One signifi-
(see Figure 28-5, B). Conduction is slower in transitional cells, cant difference between retrograde AV conduction and antegrade
thought to constitute the “fast pathway,” than in atrial myocar- AV conduction is that the retrograde electrical impulse poten-
dium because they express lower levels of Cx43 and Nav1.5, tially encounters significant source-sink mismatch42 when it exits
which are responsible for the action potential upstroke.24,35 This the insulated His bundle and compact AVN and encounters the
Control atrial pacing 60 bpm Activation maps, ms

Atrial CN His Atrial CN His Atrium Pathways - CN CN-His
200
OAPs IAS IAS IAS 255
18
His His 160 His
5 - Atria FP 14
FP4.4 245
cm/s
FP 3c
m/
s
120 13
10
36 c 235
4 - SP m/s
/s
/s
80
cm
IVS 6 IVS IVS

m
3 - FP
7
117 c
3.
SP SP 225
2 - His-CN B SP TA 2
TA 40 TA
1 - His dV/dt max maps, units/ms
dV/dt Atrium Pathways - CN CN-His

2.2 0.6 1.8
IAS IAS IAS
His His 0.5 His 1.4
Units/ms
1.5 1.8
FP FP FP
1
0.4 1
0.5 1.4
0
0.3 0.6
IVS 1 IVS IVS
0.2
SP SP SP 0.2
TA TA TA
C 0.6
A 0 500 ms
Figure 28-5. Optical mapping of atrioventricular junction during atrial pacing. A, Optical action potentials (OAPs) and their derivatives recorded from sites 1 to 5, as shown
in B and C, during atrial pacing at 60 beats/min (cycle length = 1000 ms). Red dots on OAP upstrokes correspond to dV/dt peaks. B, Atrial, atrioventricular node, and His
bundle activation maps superimposed on the optical field of view (30 × 30 mm2). C, dV/dtmax maps corresponding to the atrial, atrionodal, and His activation from B. The
black line demarcates the tricuspid valve annulus (TA). FP, fast pathway; SP, slow pathway; IAS, interatrial septum; IVS, intraventricular septum.
(Reproduced with permission from Fedorov VV, Ambrosi CM, Kostecki G, et al: Anatomic localization and autonomic modulation of atrioventricular junctional rhythm in failing
human hearts. Circ Arrhythm Electrophysiol 4:515–525, 2011.)
large mass atrial tissue, which increases the likelihood of conduc- at which this occurs is referred to as the “Wenckebach
28
tion block. Block of retrograde conduction through the AVJ has periodicity.”
been demonstrated in human optical mapping studies,39 where Hoffman and Cranefield43 explained the mechanism of slow
electrical impulses originating from the His bundle were seen to AV nodal conduction using a decremental conduction hypothesis
be delayed at distal sites of the fast pathway and the slow pathway, or the “decremental driving force hypothesis.” In this hypothesis,
ultimately resulting in exit block (Figure 28-6, A). Interestingly, decremental conduction is a result of increased rates causing
as is shown in Figure 28-6, retrograde conduction of AV junc- reduced excitability and a reduction in action potential upstroke
tional ectopic beats originating from the NH region can follow velocity and amplitude. However, this hypothesis does not ade-
different patterns of electrical wave propagation.39 One junctional quately explain the Wenckebach phenomenon. A different
beat originating from the NH region can be seen to be conducted hypothesis regarding the AVN properties was put forward by
preferentially through the fast pathway (see Figure 28-6, A), Rosenblueth, who postulated that a step-delay in conduction at
whereas another beat is conducted through the slow pathway, a particular location in the AVN could explain the appearance of
prominent with a shift in the electrical wave propagation pattern Wenckebach cycles at fast atrial rates, also referred to as the
(see Figure 28-6, B). This finding potentially supports the pos- “constant driving force hypothesis.”44 It was postulated that
sibility of two longitudinally dissociated retrograde conduction during rapid constant continuous stimuli, incomplete recovery
pathways. Similarly, Figure 28-4 shows a junctional beat originat- from excitation by the previous beat creates a conduction
ing in the NH region, conducting retrogradely via the fast delay for the next beat, and thus progressively accumulating step-
pathway, then turning and reentering anterograde via the slow delays ultimately result in block, providing time for full recovery
pathway. of excitation. A theoretical study demonstrated that a reduction
in intercellular conductance (such as that seen in the AV node)
could produce a step delay in conduction, reduce conduction
velocity, and paradoxically increase the safety factor of conduc-
Decremental Conduction tion.45 In a microelectrode experiment of single AV nodal cell of
and Wenckebach Periodicity rabbit,46 Wenckebach periodicity was found to be associated with
a gradual loss of diastolic potential, resulting in progressive
Decremental conduction is the electrophysiologic phenomenon impairment of cellular recovery from beat to beat, and thereby
whereby conduction delay is increased when the pacing cycle progressive conduction slowing because of postrepolarization
length is progressively shortened. As the cycle length is further refractoriness.
shortened, a point is reached where, at a constant cycle length, Although most theoretical and experimental studies have
conduction across the AVJ slows progressively, culminating in implicated the compact AVN as playing a key role in the decre-
a nonconducted impulse, with subsequent restoration of conduc- mental properties of the AV conduction and the Wenckebach
tion for the next impulse, known as the “Wenckebach phenom- phenomenon, some investigators have demonstrated a role for
enon.” This is another physiologic property and the cycle length the fast and slow pathways. Transitional cells have also been
OAPs 1 - CN-His 2 - FP 3 - SP 4 - TC Fast pathway

a. His-AVJ b. Atrium c. Exit block from AVJ
a b C IAS 300 IAS 650 IAS 160
His 250 His His
FP FP
200 550 120
IVS 150 IVS IVS
80
SP 450
100 SP
40
50
CS TA CS TA 350 CS TA
0 2000 4000 6000
A
Slow pathway Fast pathway
d. His-AVJ e. atrium f. AVJ-His
550 1140
d e f IAS IAS IAS
1000
His 450 His His 1130
FP FP
350 900
1120
IVS IVS IVS
250 800 1100
SP SP
150 700 1090
CS TA 50 TA CS TA
CS 600 1080
10000 12000 14000 16000 ms
B
Figure 28-6. Optical mapping of atrioventricular junction during intrinsic rhythm under control conditions. A, Optical action potentials (OAPs) from the nodal-His region
(CN-His), fast and slow conduction pathways (FP and SP), and the transitional cell (TC) region near the coronary sinus (CS) recorded from sites 1 to 4 in (B) during intrinsic
AVJ rhythm. B, Activation maps of the different tissue layers during FP conduction between atrioventricular node (AVN) (a) and atria (b); exit block from the AVN (c); and
during SP conduction between AVN (d) and atria (e), and then AVJ again (f). White arrows indicate the wave propagation directions from the different tissue layers.
shown to exhibit less negative resting membrane potentials, and described before, and contributes to the pacemaking activity of
a reduction in AP amplitude and upstroke as cycle lengths were the His bundle and the NH region. Another molecular mecha-
decreased, contributing to decremental conduction.41 Some nism of pacemaking activity of the SAN relates to the “calcium
studies have also demonstrated the occurrence of 2 : 1 AV block clock” hypothesis, whereby sarcoplasmic reticulum calcium
in the fast pathway and 3 : 2 Wenckebach block in the slow release and the related activation of sodium-calcium exchanger
pathway conduction. Recent studies have shown that the decre- (NCX) current play roles in diastolic depolarization, especially
mental conduction and Wenckebach block in the fast and slow during β-adrenergic stimulation.50,51 This mechanism may also
pathways can also affect the Wenckebach phenomenon of the play a substantial role in the human AVN. In humans, the mRNA
compact AVN.36,37 These experiments showed that electrical levels of NCX1 were shown to be elevated in the compact AVN
input from the slow pathway affects the compact AVN, thereby compared with the INE.24 This observation may support the
slowing the electrical conduction emanating from the fast superior shift of the AVJ pacemaker from the His bundle to
pathway, ultimately creating complex Wenckebach block. This compact node during β-adrenergic stimulation.
suggests a dynamic electrotonic interaction of the longitudinally
dissociated functional pathways.
Autonomic Innervation of the
Atrioventricular Junction
Atrioventricular Junctional Pacemaker
In addition to modifying pacemaker activity, autonomic innerva-
In addition to playing a critical role in conducting electrical tion of the AVJ can also modulate AVN function and affect the
impulses from the atria to the ventricles, cells in the AVJ can also effective refractory period.52,53 In the clinical setting, alterations
play a pacemaking role. Clinically, AV junctional rhythm is com- of autonomic tone in the AVJ may also be related to initiation of
monly observed during periods of sinus node pauses, when it AVNRT54 or transient AV block, and may be important in deter-
fulfills the role of an escape pacemaker. Accelerated junctional mining the AVN conduction properties and thus the ventricular
rhythm can also occur during acute illnesses, postoperative rate during atrial tachyarrhythmias such as atrial fibrillation. The
cardiac surgery, and sympathetic overdrive. The notion that the epicardial ganglionated plexus embedded in the atrioventricular
AVJ has a pacemaking function is not new. In fact, when Tawara fat pad located at the junction of both atria and inferior vena cava
first published his discovery of the AVN, his mentor Ludwig innervates the AVJ, and high-frequency nerve stimulation has
Aschoff suggested that the AVN may be the pacemaker of the been shown to slow AVN conduction without altering sinus rate,
heart.1 leading to a slowing of ventricular rate, or transient complete
The pacemaking function of the AVJ has been clearly demon- heart block.55-57 Autonomic nerve stimulation at the ganglionated
strated in optical mapping studies in animals and in humans. In plexus innervating the AVJ was also shown to shorten the local
the rabbit, the dominant AV junctional pacemaker was identified effective refractory period of the atrial tissue of the AVJ.55
in the INE, where HCN4 is abundantly expressed, with activa- Attempts have also recently been made to understand the func-
tion spreading toward the bundle of His without a significant tional complexity and the distribution of the autonomic innerva-
delay.25 Recently, we have demonstrated that the AV junctional tion of the AVJ. Studies in rabbits have demonstrated that the
pacemaker rhythm originates from the NH region or His bundle inferior nodal extension expresses higher levels of parasympa-
in optical mapping studies on the failing human heart.39 As shown thetic innervation, and the His bundle was innervated with both
in the example of a human AVJ in Figures 28-6 and 28-7, pace- sympathetic and parasympathetic neurons, in contrast to the
making activity originates from the NH/His bundle, which is sparsely innervated compact AVN region.58 Despite a lack in
electrically isolated from the ventricular myocardium, and then understanding of autonomic innervation, modulation of AVN
spreads retrogradely into the atrium. Diastolic depolarization conduction may be an emerging therapeutic target, because atrial
preceded each action potential upstroke in the NH/His region, fibrillation with rapid rate caused by high AVN conduction may
demonstrating the pacemaker function of the proximal NH/His lead to fatal hemodynamic instability, especially in heart failure.
bundle. In response to isoproterenol, the AV junctional rate
increased from 41 bpm to 80 bpm. Interestingly, β-adrenergic
stimulation also shifted the location of pacemaking activity from
the proximal His to the AVN, accompanied by the movement of Optical Mapping of the Human
the predominant site of diastolic depolarization (see Figure 28-7, Atrioventricular Junction
A and B). Moreover, isoproterenol also altered the preferential
retrograde conduction pattern of the junctional impulses (see For many years, it has been difficult to study the electrophysiol-
Figure 28-7, C). Retrograde atrial activation, which occurs pre- ogy of the AVJ in detail because it is a complex and deep struc-
dominantly via the fast pathway in intact heart, occurred simul- ture, with multiple histologically and electrophysiologically
taneously through both the slow and fast pathways during distinct cell types and components. Although His bundle–
β-adrenergic stimulation. Junctional pacemaker activity was sup- potential recordings using bipolar electrodes in clinical settings
pressed by acetylcholine. Acetylcholine in a second human heart had advanced our understanding of the functional properties of
resulted in a shift of preferential conduction toward the slow the AVJ, integrated and simultaneous recording of electrical
pathway, with no shift of pacemaker site of the NH/His bundle activity of the compact AVN, slow pathway, and fast pathway in
(see Figure 28-7, C). These data suggest that modulation of the intact human heart had remained elusive. However, recent
adrenergic and cholinergic tone can affect the preferential con- advances in optical imaging capabilities, such as the design of
duction pathway as well as the location of the dominant pace- high-resolution complementary metal-oxide-semiconductor
maker within the AVJ. cameras and the advent of novel near-infrared potentiometric
It is currently thought that the voltage-dependent “funny dyes, such as di-4-ANBDQBS, have made possible the simulta-
current” (If current) and the “calcium clock” are two important neous imaging of transmembrane voltage transients in the differ-
molecular mechanisms involved in the spontaneous diastolic ent components of the intact human AVJ.39 Optical mapping
depolarization of pacemaking cells.47-49 HCN4, which encodes using di-4-ANBDQBS allows the imaging of electrical activity
for the If channel, is expressed abundantly in the His bundle in from deeper structures such as the INE, in contrast to traditional
the human heart,24 as well as the compact AVN and INE as potentiometric dyes such as di-4-ANEPPS, which only allows for
Junctional rhythm Activation maps, ms
Control, 40 bpm OAPs

His-CN-pathways
IAS 70
His 60
IAS
Atrial
His
180
28
1- His-CN 160
FP 50
2- CN 40 140
3- Atrial IVS 30 IVS 120
10 mm 20
SP 100
CS TA 10 CS TA 80
Iso, 81 bpm CN-pathways Atrial

24 65
IAS IAS
His 20 His 60
FP FP 55
16
50
IVS 12
IVS 45
8 40
SP SP 35
4
CS TA CS TA 30
A
Pacemaker sites AVJ rhythm (bpm)
Control 80
Iso Iso, 69 ± 12
IAS
ACh
60
His
FP
40
Control, 29 ± 11
CS IVS
SP 20
ACh, 18 ± 4
0
TA AVJ# 1N 2I 3I 4N 5I 6N Avr
B
N - Nonischemic CM I-Ischemic CM
His-AVN Iso 76 bpm Atrial activation

IAS 160 IAS
OAPs 450
FP His 140 His 400
1 - CN-His 120
2 - FP 100 350
Iso 3 - RE 80
4 - TC IVS 60 IVS 300
SP
40 250
CS TA 20 CS TA
His-AVN Ach 18 bpm Atrial activation

160 500
IAS IAS
140
Ach FP His 120 His
450
100 400
80
IVS 350
SP 60 IVS
40 300
C 0 200 400 600 800 1000 1200 1400 ms TA 20 TA
CS CS 250
Figure 28-7. Optical mapping of atrioventricular junction (AVJ) during autonomic stimulation. A, Optical action potentials (OAPs) and activation maps of the different
tissue layers during spontaneous AVJ rhythm in control (upper panel) and after a 10-minute perfusion of isoproterenol (Iso) (lower panel). OAPs were recorded from the
nodo-His (NH) region (1, His-CN, blue), compact node (2, CN, green), and interatrial septum (IAS) (3, atria, red). White ovals show atrial breakthrough sites as recorded by
optical mapping. B, Summary of anatomic locations of leading pacemaker sites in the human AVJ (left panel). Individual for each heart and average values for AVJ rhythm
in control and during perfusion of Iso and acetylcholine (Ach; right panel). C, OAPs from sites 1 to 4 during Iso and ACh perfusion (left panel). His-AVN and atrial activation
maps showing the leading pacemaker remaining in the same location (NH region) but different wave propagation during both Iso and ACh perfusion (right panel). CM,
cardiomyopathy.
imaging of superficial, subendocardial human AVJ structures pathways, suggesting the presence of longitudinal dissociation
such as the His bundle and parts of the fast and slow pathways, (see Figure 28-6), shift of AV junctional pacemaker activity by
but not the complete AV conduction pathway.17 modulation by autonomic tone (see Figure 28-7), and reentry
Using combined optical mapping, microelectrode recordings, within the AVJ structures (see Figure 28-4, A).
and surface electrogram recordings, it has been clearly demon-
strated that optical action potentials from the superficial and
deeper structure can be distinguished in the rabbit heart.59 With
optical mapping, we were able to map electrical activation in Conclusions
human AVJs. As shown in Figure 28-5, action potentials from
different tissues (i.e., atria, transitional cells, INEs, compact The recent advances in optical imaging and molecular biological
AVN, His) with different morphologies could be recorded, allow- techniques have helped to significantly enhance our understand-
ing detailed activation patterns to be discerned. Interestingly, the ing of the AVJ. We have learned much about its electrophysiology
first derivative traces of optical AP (dV/dt) displayed very similar from optical imaging studies of the human AVJ, including its
morphologies to that of bipolar electrogram recordings at those dual-pathway physiology, its pacemaker activity, and its response
sites (see Figure 28-5, A). For example, the first derivative of the to autonomic modulation. Simultaneously, molecular biology
optical signal recorded at the site of the slow pathway has a high- studies have shed light on the significant heterogeneity of gene
amplitude deflection followed by a low-amplitude deflection as a and protein expression within the different components of the
result of atrial and rightward INE activation, respectively, which AVJ that underlie its functional heterogeneity. In the coming
is similar to that of slow pathway potential in bipolar electrode decade, improved multiparametric fluorescence imaging, high-
recording. throughput microelectrode recordings, and molecular imaging
Using such optical mapping techniques, we have been able to techniques will allow us to probe the AVJ with ever greater detail,
demonstrate many of the physiologic properties of the intact and as a result, some of the many controversies surrounding the
human AVJ. This includes demonstrating antegrade conduction complex and heterogeneous structure and function of the AVJ
through the slow pathway and the fast pathway (see Figure 28-5), that have arisen since its first description by Tawara more than a
two retrograde conduction patterns through two distinct century ago may finally begin to be resolved.
electrodes in humans with atrioventricular junc- tachycardia: Electrophysiological characteristics

References tional (av nodal) reentrant tachycardia. Circulation and effect of slow pathway ablation within coro-
88:2315–2328, 1993. nary sinus. J Cardiovasc Electrophysiol 17:1177–
1. Tawara S: Das reizleitungssystem des saugertier- 13. Jazayeri MR, Hempe SL, Sra JS, et al: Selective 1183, 2006.
herzens (the conduction system of the mammalian transcatheter ablation of the fast and slow pathways 24. Greener ID, Monfredi O, Inada S, et al: Molecular
heart), (Translated into English by Kozo Suma and using radiofrequency energy in patients with atrio- architecture of the human specialised atrioven-
Munehiro Shimada) London, 2000, Imperial ventricular nodal reentrant tachycardia. Circula- tricular conduction axis. J Mol Cell Cardiol
College Press, 1906. tion 85:1318–1328, 1992. 50:642–651, 2011.
2. Mines GR: On dynamic equilibrium in the heart. 14. Li J, Greener ID, Inada S, et al: Computer three- 25. Dobrzynski H: Site of origin and molecular sub-
J Physiol 46:349–383, 1913. dimensional reconstruction of the atrioventricular strate of atrioventricular junctional rhythm in the
3. Moe GK, Preston JB, Burlington H: Physiologic node. Circ Res 102:975–985, 2008. rabbit heart. Circ Res 93:1102–1110, 2003.
evidence for a dual a-v transmission system. Circ 15. Efimov IR, Fahy GJ, Cheng Y, et al: High- 26. Goy JJ, Fromer M: Antiarrhythmic treatment of
Res 4:357–375, 1956. resolution fluorescent imaging does not reveal a atrioventricular tachycardias. J Cardiovasc Phar-
4. Janse MJ, van Capelle FJ, Freud GE, et al: Circus distinct atrioventricular nodal anterior input macol 17 (Suppl 6):S36–40, 1991.
movement within the av node as a basis for supra- channel (fast pathway) in the rabbit heart during 27. Nikoo MH, Emkanjoo Z, Jorat MV, et al: Can
ventricular tachycardia as shown by multiple sinus rhythm. J Cardiovasc Electrophysiol 8:295– successful radiofrequency ablation of atrioventricu-
microelectrode recording in the isolated rabbit 306, 1997. lar nodal reentrant tachycardia be predicted by
heart. Circ Res 28:403–414, 1971. 16. Efimov IR, Nikolski VP, Rothenberg F, et al: pattern of junctional ectopy? J Electrocardiol
5. Anderson RH, Yanni J, Boyett MR, et al: The Structure-function relationship in the av junction. 41:39–43, 2008.
anatomy of the cardiac conduction system. Clin Anat Rec 280A:952–965, 2004. 28. Hucker WJ, McCain ML, Laughner JI, et al: Con-
Anat 22:99–113, 2009. 17. Hucker WJ, Fedorov VV, Foyil KV, et al: Optical nexin 43 expression delineates two discrete path-
6. Antz M, Scherlag BJ, Patterson E, et al: Electro- mapping of the human atrioventricular junction. ways in the human atrioventricular junction. Anat
physiology of the right anterior approach to the Circulation 117:1474–1477, 2008. Rec 291:204–215, 2008.
atrioventricular node: Studies in vivo and in the 18. Meijler FL, Janse MJ: Morphology and electro- 29. Hucker WJ, Sharma V, Nikolski VP, et al: Atrio-
isolated perfused dog heart. J Cardiovasc Electro- physiology of the mammalian atrioventricular ventricular conduction with and without av nodal
physiol 8:47–61, 1997. node. Physiol Rev 68:608–647, 1988. delay: Two pathways to the bundle of his in the
7. Patterson E, Scherlag BJ: Fast pathway-his bundle 19. Jackman WM, Beckman KJ, McClelland JH, et al: rabbit heart. Am J Physiol: Heart Circ Physiol
connections in the rabbit heart. J Interv Card Elec- Treatment of supraventricular tachycardia due to 293:H1122–H1130, 2007.
trophysiol 10:121–129, 2004. atrioventricular nodal reentry, by radiofrequency 30. Anderson RH, Janse MJ, van Capelle FJ, et al: A
8. Inoue S, Becker AE: Posterior extensions of the catheter ablation of slow-pathway conduction. N combined morphological and electrophysiological
human compact atrioventricular node: A neglected Engl J Med 327:313–318, 1992. study of the atrioventricular node of the rabbit
anatomic feature of potential clinical significance. 20. McGuire MA, de Bakker JM, Vermeulen JT, et al: heart. Circ Res 35:909–922, 1974.
Circulation 97:188–193, 1998. Origin and significance of double potentials near 31. Billette J: Atrioventricular nodal activation during
9. Moe GK, Preston JB, Burlington H: Physiologic the atrioventricular node. Correlation of extracel- periodic premature stimulation of the atrium. Am
evidence for a dual a-v transmission system. Circ lular potentials, intracellular potentials, and histol- J Physiol 252:H163–177, 1987.
Res 4:357–375, 1956. ogy. Circulation 89:2351–2360, 1994. 32. De Carvalho AP, De Almeida DF: Spread of activ-
10. Chang BC, Schuessler RB, Stone CM, et al: Com- 21. Haissaguerre M, Gaita F, Fischer B, et al: Elimina- ity through the atrioventricular node. Circ Res
puterized activation sequence mapping of the tion of atrioventricular nodal reentrant tachycardia 8:801–809, 1960.
human atrial septum. Ann Thorac Surg 49:231– using discrete slow potentials to guide application 33. Hoffman BF, De Carvalho AP, Mello WC, et al:
241, 1990. of radiofrequency energy. Circulation 85:2162– Electrical activity of single fibers of the atrioven-
11. Pandozi C, Ficili S, Galeazzi M, et al: Propagation 2175, 1992. tricular node. Circ Res 7:11–18, 1959.
of the sinus impulse into the koch triangle and 22. Waki K, Kim JS, Becker AE: Morphology of the 34. Zhang Y, Bharati S, Mowrey KA, et al: His elec-
localization, timing, and origin of the multicompo- human atrioventricular node is age dependent: A trogram alternans reveal dual-wavefront inputs
nent potentials recorded in this area. Circ Arrhythm feature of potential clinical significance. J Cardio- into and longitudinal dissociation within the
Electrophysiol 4:225–234, 2011. vasc Electrophysiol 11:1144–1151, 2000. bundle of his. Circulation 104:832–838, 2001.
12. McGuire MA, Bourke JP, Robotin MC, et al: High 23. Otomo K, Okamura H, Noda T, et al: “Left- 35. Greener ID, Tellez JO, Dobrzynski H, et al: Ion
resolution mapping of Koch’s triangle using sixty variant” atypical atrioventricular nodal reentrant channel transcript expression at the rabbit atrio-
ventricular conduction axis. Circ Arrhythm Elec- 44. Rosenblueth A: Mechanism of the Wenckebach- 52. Lazzara R, Scherlag BJ, Robinson MJ, et al: Selec-
trophysiol 2:305–315, 2009. Luciani cycles. Am J Physiol 194:491–494, tive in situ parasympathetic control of the canine
36. Zhang Y, Mazgalev TN: Av nodal dual pathway
electrophysiology and Wenckebach periodicity.
J Cardiovasc Electrophysiol 22:1256–1262, 2011.
1958.
45. Shaw RM, Rudy Y: Ionic mechanisms of propaga-
tion in cardiac tissue. Roles of the sodium and
sinoatrial and atrioventricular nodes. Circ Res
32:393–401, 1973.
53. O’Toole MF, Ardell JL, Randall WC: Functional
28
37. Patterson E, Scherlag BJ, Lazzara R: Stable pat- l-type calcium currents during reduced excitability interdependence of discrete vagal projections to sa
terns of ah block arising from longitudinal dissocia- and decreased gap junction coupling. Circ Res and av nodes. Am J Physiol 251:H398–404,
tion and reentry within the superfused rabbit av 81:727–741, 1997. 1986.
junction. J Interv Card Electrophysiol 28:5–18, 46. Hoshino K, Anumonwo J, Delmar M, et al: 54. Stellbrink C, Diem B, Schauerte P, et al: Differen-
2010. Wenckebach periodicity in single atrioventricular tial effects of atropine and isoproterenol on induc-
38. Patterson E, Scherlag BJ: Longitudinal dissocia- nodal cells from the rabbit heart. Circulation ibility of atrioventricular nodal reentrant
tion within the posterior av nodal input of the 82:2201–2216, 1990. tachycardia. J Interv Card Electrophysiol 5:463–
rabbit : A substrate for av nodal reentry. Circula- 47. Bogdanov KY, Vinogradova TM, Lakatta EG: 469, 2001.
tion 99:143–155, 1999. Sinoatrial nodal cell ryanodine receptor and 55. Quan KJ, Lee JH, Van Hare GF, et al: Identifica-
39. Fedorov VV, Ambrosi CM, Kostecki G, et al: Ana- na(+)-ca(2+) exchanger: Molecular partners in tion and characterization of atrioventricular para-
tomic localization and autonomic modulation of pacemaker regulation. Circ Res 88:1254–1258, sympathetic innervation in humans. J Cardiovasc
atrioventricular junctional rhythm in failing human 2001. Electrophysiol 13:735–739, 2002.
hearts. Circ Arrhythm Electrophysiol 4:515–525, 48. Baruscotti M, Bucchi A, Difrancesco D: Physiol- 56. Hou Y, Scherlag BJ, Lin J, et al: Ganglionated plexi
2011. ogy and pharmacology of the cardiac pacemaker modulate extrinsic cardiac autonomic nerve input:
40. Zhang Y, Bharati S, Mowrey KA, et al: His elec- (“funny”) current. Pharmacol Ther 107:59–79, Effects on sinus rate, atrioventricular conduction,
trogram alternans reveal dual atrioventricular 2005. refractoriness, and inducibility of atrial fibrillation.
nodal pathway conduction during atrial fibrillation: 49. Hata T, Noda T, Nishimura M, et al: The role of J Am Coll Cardiol 50:61–68, 2007.
The role of slow-pathway modification. Circula- ca2+ release from sarcoplasmic reticulum in the 57. Rossi P, Bianchi S, Monari G, et al: Vagal tone
tion 107:1059–1065, 2003. regulation of sinoatrial node automaticity. Heart augmentation to the atrioventricular node in
41. Patterson E, Scherlag BJ: Decremental conduction Vessel 11:234–241, 1996. humans: Efficacy and safety of burst endocardial
in the posterior and anterior av nodal inputs. J 50. Vinogradova TM, Bogdanov KY, Lakatta EG: stimulation. Heart Rhythm 7:683–689, 2010.
Interv Card Electrophysiol 7:137–148, 2002. Beta-adrenergic stimulation modulates ryanodine 58. Hucker WJ, Nikolski VP, Efimov IR: Autonomic
42. Rohr S, Kucera JP, Fast VG, et al: Paradoxical receptor ca(2+) release during diastolic depolariza- control and innervation of the atrioventricular
improvement of impulse conduction in cardiac tion to accelerate pacemaker activity in rabbit sino- junctional pacemaker. Heart Rhythm 4:1326–1335,
tissue by partial cellular uncoupling. Science atrial nodal cells. Circ Res 90:73–79, 2002. 2007.
275:841–844, 1997. 51. Joung B, Tang L, Maruyama M, et al: Intracellular 59. Efimov IR, Mazgalev TN: High-resolution, three-
43. Hoffman BF, Cranefield PF: Electrophysiology calcium dynamics and acceleration of sinus rhythm dimensional fluorescent imaging reveals multilayer
of the heart, New York, 1960, McGraw-Hill, by beta-adrenergic stimulation. Circulation conduction pattern in the atrioventricular node.
Blakiston Division. 119:788–796, 2009. Circulation 98:54–57, 1998.
Cell Biology of the Specialized Cardiac
Conduction System 29
David S. Park and Glenn I. Fishman
conduction system and the transcriptional networks that govern

CHAPTER OUTLINE
its formation.
Histologic Analysis of the Developing Mammalian
Cardiac Conduction System 287
Cellular Origins of the Cardiac Conduction System 288 Histologic Analysis of the Developing
Models of Cardiac Conduction System Development 288 Mammalian Cardiac Conduction System
Molecular Markers of the Cardiac Conduction System 289 Viragh and Challice3-5 performed meticulous histologic analysis
Transcription Factor Regulatory Networks 290 of the developing cardiac conduction system in mouse embryos
between 8 and 12 days after coitus (E8-E12). Conduction cells
Conclusion 295 were distinguished from working cardiomyocytes by the follow-
ing characteristics: (1) periodic acid–Schiff (PAS) positive stain-
ing, (2) poorly organized contractile apparatus; (3) enriched
The cardiac conduction system (CCS) consists of the slowly glycogen content, and (4) reduced number of T-tubules. Using
conducting sinoatrial node (SAN) and atrioventricular node these features, the temporal-spatial distribution of conduction
(AVN) and the rapidly conducting ventricular conduction system cells was then tracked during cardiac development.3-5
(VCS). The SAN is the dominant pacemaker and is situated In the developing mouse embryo, the contractile sequence of
between the superior vena cava and the right atrium. The AVN the heart is established by E9, well before the appearance of the
persists as the last point of communication between the atria and sinoatrial node primordium at E11.5 The origin of contraction
ventricles and conducts slowly, giving the ventricles adequate was noted to be in the right sinus horn. Within the dorsolateral
time to fill during diastole. The VCS, which consists of the His wall of the sinus horns, loose mesenchymal cells were noted to
bundle, bundle branches, and the Purkinje fiber network, con- transform into the early sinus musculature, which extends along
ducts rapidly and coordinates cardiac chamber contraction in the the sinus side of the sinoatrial venous valves. This aggregation of
interventricular axis. The insulated bundle branches ensure that early sinus muscle tissue was the presumed site of SAN develop-
impulse propagation will proceed from an apical-to-basal direc- ment and corresponded to the site that Wenink termed the “sino-
tion. This sequence of ventricular activation optimizes blood atrial (SA) ring.”6 The SAN primordium was recognizable at E11
propulsion toward the aorta and pulmonary artery. in the medioanterior wall of the right superior vena cava within
The framework of the cardiac conduction system is laid down the early sinus muscle. A left-sided SAN develops simultaneously
early during heart development. Paff et al.1 noted that the chick in the medioanterior region of the left common cardinal vein,
electrocardiogram (ECG) transforms from a sinusoidal waveform but ultimately resorbs and incorporates into the wall of the left
to the mature configuration before identifiable components of atrium.5 In humans, the SAN primordium is a long structure
the cardiac conduction system are formed (Figure 29-1).1 They located within the subepicardial cleft at the site of invagination
also noted that atrioventricular (AV) block was achievable with of the sinus venosus and the right atrium, which forms the right
digitalis at the 18-somite stage before a discernible PR interval venous valve.5
was evident, suggesting that the AV nodal primordium develops The sinoatrial (SA) and AV conduction system develop simul-
within the early heart tube (see Figure 29-1, B).1 After the taneously. At E9 to E10, the AV canal (AVC; i.e., AV ring) is a
18-somite stage, cardiac chamber formation ensues and the ECG well-defined constriction, and the inner cell layer of the AVC
begins to manifest evidence of fast conduction by the presence makes numerous interconnections with the trabecular compart-
of high frequency P waves and QRS complexes (see Figure 29-1, ment, which is the source of the His-Purkinje system (HPS).3,4
C and D). The evolution of the chick ECG clearly demonstrates At E11, the primordium of the AVN was identified as a PAS+ cell
that the slow conducting nodal elements are present in the early cluster in the inner, dorsal AVC. These PAS+ AVC cells were
looped heart and that the fast conducting elements are incorpo- contiguous with the crest of the developing interventricular
rated during chamber formation. The attainment of a mature septum (primary ring), positioning the AV nodal anlage in direct
ECG configuration before structural maturation of the CCS is communication with the primordial His bundle and bundle
achieved by interposing slowly conducting AV canal myocardium branches. During this time, the outer cell layer of the AVC was
between the fast conducting atrial and ventricular chamber myo- undergoing apoptosis. In the trabecular region, glycogen-rich,
cardium (Figure 29-1, E). The fast-conducting components are PAS+ cells were seen immediately subjacent to the endocardium;
enriched in high-conductance gap junction proteins, connexin40 these nascent Purkinje cells formed extensive connections with
(Cx40), and the α-subunit of the cardiac sodium channel, Nav1.5 the developing bundle branches.3,4 Therefore, all components of
(encoded by Scn5a), whereas the AV canal expresses low- the AV conduction system are in contact with each other through-
conductance gap junction proteins, Cx30.2 and Cx45.2 How out cardiogenesis, indicating that an initial framework for the
these electrophysiologically distinct regions are specified and mature conduction system is in place in the early heart.
what defines the boundaries between slow and fast conduction The work by Viragh and Challice3-5 clearly demonstrated that
have been the focus of intense research over the past 50 years. conduction system development is inextricably linked to cardio-
This chapter discusses the developmental origins of the cardiac genesis. Yet, significant questions remained regarding the cellular
287
11-S: Heart tube stage

Models of Cardiac Conduction
System Development
A
The prevailing models of conduction system development are the
ring model,6 the inductive recruitment or in-growth model,7 the
early specification or outgrowth model,12 and the biphasic model
(Figure 29-2). The ring hypothesis was based on early observa-
18-S: Looping stage P tions that the specialized conduction system formed within four
Primordial constriction points in the D-looped heart (see Figure 29-2, A).6
vent Four rings were noted to form from the venous pole to the arte-
Primordial rial pole: the SA ring, AV ring, primary (interventricular) ring,
atrial B and ventriculoarterial (VA) ring. These constriction areas are
QRS created by differential proliferation rates that exist between ring
AV canal P
T myocardial cells and flanking chamber cardiomyocytes. Although
these areas of constriction exist within the developing heart
20-S: Chamber stage during chamber formation, the ring hypothesis has largely been
discredited, as there is no evidence to support a preexisting tem-
C
P plate for ring formation within the tubular heart. It is now known
QRS
that the linear heart tube is composed of clonally related myo-
T cardial cells from the first heart field.13
Vent Cheng et al.7 put forth the recruitment model or “in-growth
model” based on several observations noted during chick VCS
Atria
development (see Figure 29-2, B).7 First, the proliferative capac-
D PR ity of developing VCS components was found to be significantly
QRS lower than working myocytes based on pulse labeling experi-
P ments using [3H]-thymidine. Once specified, conduction cells
33-S: Septaion stage
T appeared to exit the cell cycle and become quiescent. Second,
lineage tracing studies showed that individually labeled myocyte
clones gave rise to conduction cells and working myocytes.
Atria Third, cell birth dating experiments demonstrated that new con-
ductive cells were added to the developing His bundle in lamellar
AV canal PR fashion, analogous to tree rings. These observations led the
Vent authors to conclude that the specialized conduction system
QRS expands through a process of inductive recruitment of neighbor-
E ing myocytes.7 However, it was not speculated what constituted
the early framework upon which new conduction cells were
Figure 29-1. Schematic of chick heart development with corresponding electro-
cardiograms at the somite stage. A, 11. B, 18. C and D, 20. E, 33.1
added or the nature of the molecular signal used for inductive
recruitment.
The early specification model, or outgrowth model, states that
conduction cells expand from a progenitor pool that retains its
specialized conduction phenotype (see Figure 29-2, C).12,14 The
conduction gene programming is retained by the expression of
origins of the conduction system, the mechanism by which the transcriptional repressors that suppress a working myocardial
pool of conduction cells expands, and the factors that dictate CCS phenotype, which is the default pathway. In support of this
specification and patterning. hypothesis, persistent expression of repressive transcription
factors (Tbx2, Tbx3, Msx2, and Id2) has been identified within
primordial conduction regions.12,14,15 Tbx3 is expressed as a con-
tinuous band linking the SAN and AVN, internodal tracts, and
Cellular Origins of the Cardiac the proximal ventricular conduction system.12 Heterologous
Conduction System expression of Tbx2 or Tbx3 is able to suppress chamber-type
myocardial genes (Nppa [ANF], Gja1 [Cx43], Gja5 [Cx40]),
The neuronal qualities of the cardiac conduction system led many inhibit cardiac chamber formation, and in the case of Tbx3 elicit
to believe that its cellular origins were from neural crest deriva- ectopic pacemaker formation.12,14-16 Consistent with these find-
tives. However, lineage-tracing studies in the chick and mouse ings, Tbx3 and Cx43 exhibit complementary expression patterns
demonstrated that all conductive components of the CCS are in the developing heart.12
myocardial in origin.7-9 Using a replication-defective retrovirus Most recently a biphasic model of conduction system develop-
expressing LacZ, individual myocytes were labeled in developing ment has been proposed (Figure 29-2D).9 In this model, once
chick hearts before neural crest immigration, and the expression conduction cells are recruited from myocardial precursors, they
of β-galactosidase (β-gal) was traced in daughter cells.7 Single retain the capacity to undergo limited rounds of cell division.
myocyte clones gave rise to both conduction cells and working Analysis of labeled myocyte clones revealed two classes of con-
cardiomyocytes. Labeling of neural crest cells, however, failed to ductive clusters, mixed and unmixed (see Figure 29-2, E). The
show any incorporation into the CCS.7,8 Similar findings were mixed clusters represented single myocyte clones that gave rise
reported in the murine heart using retrospective clonal analysis, to both conductive and working myocytes (recruitment). The
where individual myocyte progenitor clones gave rise to conduc- unmixed clones were composed of either working myocytes or
tion and nonconduction cells.9 Similar to the chick, lineage- conduction myocytes, but not both. Conduction-only, unmixed
tracing studies in mice have failed to identify a direct contribution clones were identified throughout the central and peripheral
of neural crest cells to the specialized conduction system.10,11 VCS, indicating that once specified, all components of the
Cell Biology of the Specialized Cardiac Conduction System 289
29
Tbx3+
His CCS
VA ring
bundle
Primary ring
AV ring
No cell
SA ring
division
Tbx3-
working CM
A B C
Progenitor cell
D E Mixed clusters Unmixed Unmixed

working CM CCS
Figure 29-2. Models of cardiac conduction system development. A, Ring model. Prespecification of conduction system components within the linear heart tube. B, Induc-
tive recruitment or ingrowth model. Undefined inductive signals recruit proliferating cardiomyocytes to a conduction lineage, then cease to proliferate (red arrow). C, Early
specification or outgrowth model. Tbx3+ CCS cells expand from primitive myocardium that retains a conduction phenotype. Loss of Tbx3 (or Tbx2) expression results in a
phenotypic change from a CCS cell to a working cardiomyocyte, the default pathway (red arrow). D, The biphasic model incorporates both the ingrowth and outgrowth
models. E, Retrospective clonal analysis of LacZ-labeled, cardiomyocyte clones (blue nuclei). Mixed clusters of conduction and working cardiomyocytes are consistent with
a common myocardial progenitor. Unmixed clusters of CCS-only cells demonstrate the potential for limited rounds of cell proliferation.
(From Miquerol L, Beyer S, Kelly RG: Establishment of the mouse ventricular conduction system. Cardiovasc Res 91:232–242, 2011.)
HB
HB
LBB
LBB LVW LF LVW
LBB
PF
IVS LVW PF
PF PF
A B C
Figure 29-3. Cardiac conduction system reporter mice. A, CCS-LacZ. B, Contactin2-eGFP. C, Connexin-40–eGFP.20
conduction system are capable of approximately four to five

rounds of cell division. Based on these findings, the authors Molecular Markers of the Cardiac
concluded that mammalian VCS development appears to use Conduction System
both in-growth and outgrowth modes of expansion.9 However,
these findings are not incompatible with the early specification/ Visualization of the developing CCS has been greatly enhanced
outgrowth model, because mixed clusters might represent con- by the development of conduction system reporter mice (Figure
duction cells that have lost Tbx2/Tbx3 expression and have 29-3). Each reporter mouse delineates different components of
defaulted to the chamber pathway. the CCS at various developmental time points using LacZ or
green fluorescent protein (GFP) expression. The CCS-LacZ and with Holt-Oram syndrome manifest variable degrees of CCS
minK-LacZ mouse lines are representative examples of well- dysfunction, including sinus bradycardia and AV block, even in
established markers of the specialized conduction system.17-19 the absence of overt structural heart disease.22 Analysis of Tbx5
The CCS-LacZ mouse was created serendipitously through a transgenic mice demonstrated an exquisitely sensitive, dose-
complex genomic rearrangement involving the MC4/engrailed-2- dependent correlation of Tbx5 expression level with severity of
LacZ cassette (see Figure 29-3, A).18 CCS-LacZ reporter expres- congenital heart defects and conduction disease.34
sion can first be detected at E8.5 in the SAN primordium within Interestingly, the temporal-spatial expression of Tbx5 corre-
the venous pole. At subsequent stages, β-gal expression is detected lates well with developmental timing of conduction system
in the developing and mature AVN and His-Purkinje system. All components.35 At E8.5, Tbx5 expression is enriched in a posterior-
CCS reporter lines have some degree of cardiac expression anterior axis, with highest levels of expression in the atria and
outside of the conduction system. In the adult CCS-LacZ heart, sinus venosus, the site of SA node specification. During chamber
significant β-gal expression is seen within the right atrium.18 formation, Tbx5 expression is graded in a left-right axis with
The minK-LacZ reporter mouse was created by replacing the highest levels in the left ventricle, particularly in the AV canal,
minK gene with a nuclear-targeted LacZ cassette.17 Early in ventricular conduction system and the trabecular myocardium
development, β-gal expression was noted in the SA ring, AV ring, (the site of Purkinje fiber development). Tbx5 expression is scant
interventricular ring, and the VA ring. Subsequently, β-gal in the right ventricle except in the developing AV bundle, right
expression was confined to the AVN and the proximal conduction bundle branch, and trabecular region.35
system, as well as in the venous valves, AV ring, and VA valves.17,19 Both the dose–phenotype correlation and the temporospatial
The Cx40-eGFP reporter mouse has become a widely used expression pattern of Tbx5 implicate it as a master regulator of
tool to characterize normal and abnormal patterning of the conduction system development. However, the broad expression
mature His-Purkinje system (see Figure 29-3, C).20 Developmen- pattern of Tbx5 beyond the borders of the CCS, and the ability
tally, Cx40 expression is not restricted to the VCS, with signifi- of Tbx5 with its binding partners, Nkx2-5 and Gata4, to elicit
cant expression in the trabecular myocardium. In addition, Cx40 chamber myocardial gene programming suggests that it functions
is not expressed in the distal AVN or His bundle before E14.5. with CCS-restricted cofactors to elicit conduction specifica-
In the mature heart, Cx40 is enriched in atrial myocardium and tion.26,28 Numerous regulatory feedback loops have been identi-
in coronary endothelial cells.20 fied in CCS components that enhance the conduction phenotype
Contactin-2 (Cntn-2) was recently identified as a CCS- while simultaneously repressing chamber-specific gene program-
enriched factor using differential gene profiling of adult mouse ming (see Figure 29-4). These conduction-restricted cofactors
Purkinje fibers versus working myocytes (see Figure 29-3, B).21 will be discussed in their regional context.
Cntn-2 is a cell adhesion molecule that has a role in neuronal
patterning and ion channel clustering. Both Cntn2-LacZ
knock-in mice and Cntn2-EGFP BAC transgenic reporter mice The Sinoatrial Node
delineated the entire cardiac conduction system in postnatal
hearts. Currently, a functional role for Cntn-2 in the CCS has The mammalian sinoatrial node is a large comma-shaped struc-
not been identified.21 ture with its head region located at the junction between the right
superior vena cava and the right atrium, and the tail region situ-
ated along the crista terminalis. These regions represent distinct
cellular lineages, as evidenced by their unique expression profiles.
Transcription Factor Regulatory Networks The SAN head, which constitutes up to 75% of SAN volume,
develops from sinus venosus (SV) myocardium and retains the
A rich hierarchy of gene networks dictates the specification and SV signature, Shox2+;Tbx3+;Tbx18+;Nkx2-5−.36,37 In contrast,
patterning of conduction components (Figure 29-4). Unifying all working atrial myocytes derive from second heart field mesoder-
these networks is the balance struck between prochamber myo- mal progenitors and express Nkx2-5+;Shox2−;Tbx3−;Tbx18−. The
cardial programming versus antichamber programming. As men- SAN tail domain has an expression profile in between that of
tioned previously, the T-box transcription factors dictate much the head region and the atrial myocardium, expressing
of this equilibrium, tilting the scales toward or away from a con- Shox2+;Tbx3+;Tbx18−;Nkx2-5weakly +. Lineage tracing studies have
duction lineage. The T-box factors can function as transcrip- shown that the SAN tail originates from SV myocardium, but
tional activators or repressors and are known to be critical loses Tbx18 expression during development.37
regulators of cardiac specification and differentiation.22 Seven Proper SAN development is dependent on the appropriate
TBX family members are expressed in the developing heart, four expression of Tbx5, Shox2, Tbx18, and Tbx3 (see Figure 29-4,
of which (TBX1, TBX5, TBX20, TBX3) have been linked to B). Tbx5 expression in the sinus venosus is a critical regulator of
human congenital heart disease.23-25 The major cardiac transcrip- the SAN signature through its actions on the transcriptional
tional activators, Tbx5 and Tbx20, act through Nkx2-5 and repressor, Tbx3, and the homeobox transcription factor,
Gata4 to drive prochamber myocardial gene expression, such as Shox2.34,38 Homozygous Tbx5del/del mice die embryonically at
Nppa (ANF) and Gja5 (Cx40).26-28 The transcriptional repressors E10.5 because of severe hypoplasia of the sinoatrial region and
Tbx2, Tbx3, and Tbx18 compete with Tbx5 for Nkx2-5 binding of the primitive LV.27 Microarray analysis of Tbx5 heterozygous
to suppress chamber-specific gene expression, thus maintaining hearts, identified Tbx3 and Shox2 as significantly downregulated
a conduction gene profile.12,14,15,29,30 targets, and both factors showed reduced expression in the sino-
The delicate balance between Tbx activators and repressors atrial region.34,38
and the expression of cofactors dictates the rich conduction As stated earlier, Tbx3 is expressed throughout the developing
system phenotypes seen in the mature CCS. Sitting on top of the and mature CCS (except the Purkinje network) and represses
CCS specification hierarchy is Tbx5, a critical determinant of chamber-specific programming. Like Tbx5, Tbx3 displays criti-
many elements of the slow and fast conduction system.31-33 Muta- cal dose dependency for proper differentiation and homeostatic
tions in Tbx5 result in Holt-Oram syndrome, an autosomal maintenance of the conduction system.39 Analysis of Tbx3 mutant
dominant condition characterized by preaxial radial ray limb mice revealed a dose-dependent SAN phenotype with variable
deformities and cardiac septation defects.22 The septal defects are degrees of sinus node dysfunction and inappropriate expression
typically ostium secundum atrial septal defects, muscular ven- of chamber-specific genes (Cx43, Cx40, Nppa, Scn5a) within the
tricular septal defects, and atrioventricular canal defects. Patients SAN region. Although the overall structure of the SAN is normal,
Nkx2-5 Tbx5
Pitx2c
29
Tbx5 Cx30.2
Shox2
Tbx3 Cx40
Shox2 ANF Tbx5 Nkx2-5
Tbx18 Cx43 Cx40
Nkx2-5 Nav1.5 Cx43
HCN4 Nav1.5
ANF
Tbx3+
Tbx18–
Shox2+
A B
Notch2
Nav1.5 Hey1 Tbx20
ANF
Cx40 Cx43 Tbx2
Nkx2-5 Msx1/2
Tbx3 Bmp2+Alk3
Tbx5
Tbx2 Foxn4
Notch Tbx2
Cx30.2 Msx1/2
Cx40 Cx43 Hey2 Tbx20
ANF Hey1
Nav1.5
Notch2
C
ET-1 NRG-1
Id2
?
Tbx5 Nkx2-5 Prox1
Prox1 Nkx2-5 Tbx5
Cx40 Hop
Nav1.5
Hf-1b Cx40 Nav1.5
Tbx3 Irx3
Cx43
ANF E
D
Figure 29-4. Transcription factor regulatory networks. A, Schematic of the cardiac conduction system. B, Right sinoatrial node (head and tail domains), atrial myocardium,
and left sinoatrial node (stippled circle). C, Atrioventricular canal/node region flanked by atrial and ventricular myocardium. D, His bundle and bundle branches and ventricular
myocardium. E, Purkinje fiber.
SAN volume was reduced by approximately 45% to 60%, expression, thereby preventing atrial myocardialization of the
although this phenotype has not been consistently seen in all SAN region.41
Tbx3 mutants when adjusted for weight.16,29,37,39 Ectopic overex- The left-sided sinoatrial node (L-SAN) forms in parallel with
pression of Tbx3 in atrial myocardium results in inappropriate right-sided SAN (R-SAN), but regresses during development and
suppression of chamber genes and upregulation of the SAN gene incorporates into the left atrium.5 Abnormal persistence of
profile.29 These data suggest that Tbx3 is not essential for SAN L-SAN remnants has been implicated as a source for left atrial
formation, but is important for establishing and maintaining arrhythmic triggers.42 The homeodomain transcription factor
proper pacemaker gene programming. Pitx2c, which regulates left-right asymmetry, directly regulates
Shox2 is essential for formation of the sinoatrial valves and L-SAN resorption.43 Mice deficient in Pitx2c invariably develop
development of the sinoatrial node.40 Mice deficient in Shox2 are a persistent L-SAN, which shares an identical gene expression
embryonic lethal because of severe bradycardia in the setting of profile with the R-SAN.16 Therefore, L-SAN specification is
SAN hypoplasia.40,41 Evaluation of the Shox2−/− SAN region dependent on Shox2 signaling. Pitx2c directly represses Shox2
revealed reduced levels of Tbx3 and the pacemaker channel, expression in the developing sinus venous myocardium and in the
HCN4, indicating a failure of differentiation.41 Downregulation adult left atrium, leading to L-SAN regression.42 Genome-wide
of Tbx3 resulted in ectopic expression of chamber myocardial association studies (GWAS) have identified a region near PITX2
genes (Nppa, Cx43, Cx40) within the SAN. Furthermore, Nkx2-5, as a key susceptibility locus for atrial fibrillation.44 Interestingly,
which is normally absent in SAN myocardium, was expressed Pitx2c haploinsufficiency increased the propensity for atrial
ectopically in the SAN region.40 Luciferase reporter assays arrhythmias, corroborating the GWAS data.42
revealed that Shox2 negatively regulates expression of Nkx2-5, The importance of Tbx18 in SAN formation was identified
a known inhibitor of Tbx3 and Hcn4 expression.16 Thus, in knockout mice, which revealed a marked reduction in the
Shox2 promotes SAN development by repressing Nkx2-5 volume of the SAN head, because of either failure to expand the
mesenchymal precursor pool or failure to differentiate into SAN conduction time, as measured by prolonged Atrio-His (A-H)
cells.37 The tail region, however, was unaffected.37 Despite this intervals during invasive EP testing.33
reduction in SAN volume, sinus node function was reportedly Nkx2-5, a member of the homeodomain family, plays a central
normal owing to residual SAN tail tissue.37 Tbx3 and Tbx18 role in cardiac development. Loss of the Nkx2-5 homolog,
double heterozygous embryos demonstrated normal SAN devel- tinman, in the fruit fly results in failure of cardiogenesis.51 Mice
opment, indicating a lack of interaction on a genetic level. In deficient in Nkx2-5 die in utero at E9 to E10 because of arrested
addition, Tbx3 deficiency had no effect on Tbx18 and Shox2 cardiac development in the linear heart tube stage. The hearts of
levels, suggesting that Tbx3 functions downstream to these these mice undergo partial looping morphogenesis, lack endocar-
factors.45 dial cushions and trabeculae, and have underdeveloped AV
Similar to Tbx3, heterologous expression of Tbx18 leads to canals.52 Nkx2-5 null embryos lacked minK-LacZ reporter stain-
transcriptional repression of Cx43.45 In summary, Shox2 and ing in the region of AVN primordium.53 Nkx2-5+/neo mice exhib-
Tbx18 are critical regulators of SAN formation, whereas Tbx3 ited an overall reduction in the size of the VCS from the AVN
together with Tbx18 maintain the appropriate pacemaker signa- to the distal Purkinje network. Histologic evaluation of the
ture, assuring optimal function. Nkx2-5+/neo AVN revealed that compact nodal cells (N region,
Cx40−/Cx45+) are markedly reduced whereas the nodo-His (NH)
region (Cx40+/Cx45+) remains intact.53 Consistent with these
Atrioventricular Canal and Atrioventricular Node findings, Nkx2-5+/neo mice had prolonged PR intervals on surface
ECG and displayed abnormal AVN physiology on EP testing.53
The mammalian AVN is a complex, heterogeneous structure that Tbx2 and Tbx3 are essential for maintaining the slow conduc-
serves as the only conduction pathway between the atria and tion phenotype of the AV canal/AVN.12,14,15 Tbx2 and Tbx3 are
ventricles. Histologic and electrophysiologic (EP) evaluation of differentially expressed in the AVC, with Tbx2 being more domi-
the AVN revealed three distinct layers, from subendocardial to nant in the left AVC and Tbx3 being more dominant in the right
deep layers, termed atrionodal (AN), compact nodal (N), and nodo- AVC.54 Evaluation of a Tbx3 allelic series demonstrated an exqui-
His (NH).46 The compact or true nodal region has characteristic site dose–phenotype correlation, with diminishing Tbx3 levels
nodal action potentials, such as slow upstroke of phase 0, lower directly correlating with the severity of AV conduction disease
peak amplitudes, and diastolic depolarization of phase 4. The AN and fetal demise.39 AV block was exclusively seen in embryos with
and NH transition regions exhibit hybrid action potential mor- the lowest Tbx3 gene dosage. Consistent with its known repres-
phologies intermediate to nodal cells and atrial myocytes or His sive function, Tbx3-deficient embryos showed ectopic expression
bundle, respectively.47 These transition zones serve a similar of chamber genes within the SAN, AVC, AV bundle, and bundle
purpose as in SAN cells, to insulate core nodal cells from the branches.39,55 Regional deletion of Tbx3 within the AV bundle
hyperpolarizing influence of working myocytes. The AN and primordium or the AV canal caused AV block. Therefore, Tbx3
NH regions are enriched in Scn5a (Nav1.5), Cx40, and Cx43, expression is required throughout the multitiered AV conduction
whereas the N region has low levels of Nav1.5 and high levels of axis to maintain normal AV conduction. Furthermore, condi-
Cx45/30.2 resulting in the different action potential morpholo- tional deletion of Tbx3 at an adult time point resulted in AV
gies and conduction velocities (CV).46 block, indicating that Tbx3 is critical for proper maintenance of
Lineage tracing studies and differential gene expression analy- the AV conduction axis.39
sis indicate that the compact AVN (N) derives from AV canal Tbx2 deficiency results in embryonic lethality because of
myocardium, whereas the lower nodal cells (NH) are ventricular defects in AVC differentiation and outflow tract septation.56 Inap-
in origin.48,49 In the adult heart, remnants of the AVC persist as propriate expression of chamber-type genes was also noted in the
rings of myocardial tissue that possess distinct electrophysiologi- AVC of Tbx2-deficient hearts, with the left AVC being more
cal layers, mirroring that seen in the AVN.50 These results lend affected than the right.54,56 Accordingly, Tbx2 deficiency had a
support to the histologic observations that the AVN derives from minimal impact on AVN formation, which is predominantly a
dorsal AV canal cells.3,4 right-sided structure. However, loss of myocardial Tbx2 resulted
During development, the AVC can be distinguished from in structural defects in the left annulus fibrosus resulting in acces-
flanking atrial and ventricular myocardium by the enrichment of sory pathway formation and ventricular preexcitation.54 There-
Tbx2, Tbx3, and Bmp2 (see Figure 29-4, C). This transcriptional fore, Tbx2 and Tbx3 share nonredundant functions in AVC
signature maintains the slow conduction properties (Cx45+; development and maturation, which may be in part due to
Cx30.2+) and low proliferative capacity of the AVC. As the unequal distribution of the T-box proteins. Tbx2 expression
embryonic heart matures, the outer cell layer of the AVC under- appears to be critical for maturation of the left annulus fibrosus,
goes apoptosis. Paracrine signals from AVC myocytes induce which is the most common location of accessory pathways in
epicardial cells to undergo epithelial-to-mesenchymal transfor- humans.
mation and invade the AV junction to create the annulus fibrosus. A transcriptional regulatory network involving Bmp2, Foxn4,
This process electrically isolates the atrium and ventricle, except Tbx20, and Notch2/Hey signaling has been identified that regu-
at the dorsal wall of the AVC where the lower AV nodal transi- lates the expression of Tbx2 within the AVC. Bmp2 and Foxn4
tional cells penetrate the fibrous insulation. Defective gene function as activators of Tbx2 within the AVC, whereas Tbx20
dosing of Tbx2, Tbx3, and Bmp2 leads to developmental abnor- and Notch2/Hey signaling are negative regulators of Tbx2 that
malities of the AVC, which can manifest as AV block or as ven- define the boundaries of the AVC. The bone morphogenetic
tricular pre-excitation (i.e., Wolff-Parkinson-White) because of factor Bmp2 is enriched in the AVC during development and
defective annulus fibrosus formation. drives the local expression of Tbx2.57 Bmp2 upregulates Tbx2
The AVC is established in part by the expression of Tbx5 and promoter activity through a Bmp receptor (Alk3)/Smad depen-
Nkx2-5, whereas Notch signaling delimits the AVC to its con- dent pathway.58 Deletion of Bmp2 in the AV canal resulted in loss
fined region within the developing heart. Tbx5 is enriched in the of Tbx2 expression and ectopic expression of chamber genes.57
AVC during development and drives the local expression of Tbx3 Furthermore, AVC-restricted loss of Alk3 led to abnormal
and Cx30.2.31,34 Mice haploinsufficient for Tbx5 exhibit sinus annulus fibrosus formation, ventricular preexcitation, and struc-
node dysfunction, P wave widening, and first- and second-degree tural and functional defects in the AVN.59,60 A human correlate
AV block.27,33 Significant maturation defects were noted in the has been identified in a rare familial form of Wolff-Parkinson-
AVC and AVN of Tbx5+/− mice.33 The immature configuration White syndrome associated with a microdeletion involving the
of the AV canal resulted in significant slowing of AV nodal BMP2 gene.61
Foxn4, an upstream regulator of tbx2b, was recently identified interval prolongation.53,68,69 Ventricular-restricted, Nkx2-5
29
from a zebrafish mutagenesis screen.62 Foxn4 mutants displayed knockout mice display progressive AV conduction disease,
structural and functional defects in the AV canal. Several AV advancing to complete heart block by 6 months to 1 year. Histo-
canal restricted genes were mislocated, including bmp4 and logic analysis revealed a diminutive AVN and an atrophic His-
endocardial notch1b, and tbx2b was completely absent from the Purkinje system that worsened with age.53,68,70,71 These mice
AVC. Highly conserved Foxn4 and Tbx5 binding sites were iden- accurately phenocopy the progressive postnatal AV conduction
tified in the tbx2b promoter, and tbx2b expression in the AVC disease seen in patients with NKX2-5 mutations.68 The pheno-
proved highly sensitive to mutagenesis of the Foxn4 or Tbx5 typic manifestations of Nkx2-5 haploinsufficiency are highly
binding sites.62 pleotropic in man and mouse models, which suggests that Nkx2-5
In the chick heart, Notch2, Hey1, and Hey2 signaling delimits is subject to upstream regulatory control by genetic modifiers.
the AV canal myocardium. Hey1 (expressed in atrium and ven- One such modifier is prospero-related homeobox protein 1
tricle) and Hey2 (expressed in ventricle only) are expressed in (Prox-1), which functions in concert with HDAC3 to regulate
complementary fashion to Bmp2 in the looped heart.63 Notch2 Nkx2-5 expression.72 Combined haploinsufficiency of Nkx2-
acts directly through Hey1 to suppress Bmp2, whereas Hey2 5cre/+;Prox1loxP/+ rescued the Nkx2-5cre/+ conduction phenotype on
suppresses Bmp2 in a Notch-independent manner.63 Mouse and a structural and functional level. Compound heterozygotes
zebrafish hearts deficient in Hey2 displayed abnormally expanded rescued the hypoplastic phenotype of the AVN and significantly
AVC regions that were enriched in Bmp2/bmp4, respectively.63 restored cellularity of the His-Purkinje system, resulting in nor-
Tbx2 expression in the AVC is also delimited by atrial and ven- malization of ECG parameters.72 In light of these findings, Prox1
tricular chamber expression of Tbx20. Observations of Tbx20 appears to function as an upstream regulator of Nkx2-5 gene
knockout embryos showed precocious upregulation and ectopic dosage, which ensures accurate gene expression within the ven-
expression of Tbx2 throughout the cardiac crescent and early tricular conduction system.
heart tube. In vitro reporter assays demonstrated that Tbx20 Mice haploinsufficient for Tbx5 exhibited patterning and
inhibits Tbx2 promoter activity through its actions on the Alk3/ maturation defects of the ventricular conduction system. The His
Smad signaling pathway.57 bundle and left bundle branches remained immature in all
Notch signaling in endocardial cells is critical for the forma- Tbx5del/+ mice, and many of the mice had absent right bundle
tion of slow conducting AVC myocytes in zebrafish and trabecu- branches. The expression of Tbx5-responsive genes, Nppa and
lar myocardium in mice.64,65 Notch2 signaling in chamber Cx40, were significantly downregulated.33 Nkx2-5 and Tbx5
myocytes restricts Tbx2 expression within the AVC.63 Whether coordinately drive the expression of Nppa, Cx40, and inhibitor of
Notch signaling has a role within developing conduction cells differentiation 2 (Id2) within the proximal VCS.32 Id2 is believed
was explored using loss-of-function and gain-of-function Notch to function as an inhibitor of muscle differentiation allowing
mutants overexpressed under the control Mlc2vCre/+.66 The domi- specification towards a conduction lineage.32 Id2 null mice dis-
nant negative Notch mutant, DNMAML, driven by Mlc2vCre/+ played structural and functional VCS abnormalities similar to
displayed reduced AV nodal volumes owing to loss of Cx30.2- those seen in Tbx+/− mice, suggesting that they function within
expressing cells. Mlc2vCre/+/DNMAML had shortened PR and the same transcriptional regulatory network. Indeed, combined
AH intervals on EP testing, consistent with loss of slow AV nodal haploinsufficiency of Tbx5+/−;Id2+/− resulted in developmental
conduction.66 Constitutive Notch activation, however, resulted in failure of the His bundle and bundle branches.68 Therefore,
enlarged AVNs and accessory pathway formation because of Nkx2-5, Tbx5, and Id2 function as a ventricular conduction
abnormal boundary formation between the AVC and ventricular system transcriptional unit that imparts fast conduction proper-
myocardium.66 Subsequent work using gain-of-function Notch ties while also inhibiting myocardial gene programming.
mutants demonstrated a robust upregulation of conduction Tbx5 is essential for maintaining the fast conduction proper-
system markers, such as Cntn-2, Tbx5, and Nkx2-5. Notch acti- ties of the ventricular conduction system. Using a VCS-specific,
vation also upregulated nodal genes Gjc3 (Cx30.2) and HCN1 tamoxifen-inducible Cre driver (minKCreERT2), Tbx5 was
and Purkinje gene Scn5a. Notch activation was able to reprogram knocked out of the adult mouse AVN and HPS, resulting
ventricular myocytes in vivo and newborn ventricular myocytes in sudden death as early as 5 weeks after Cre induction.73
in vitro into a Purkinje-type lineage. These results suggest that Tbx5minKCreERT2 VCS conduction was severely slowed, leading to
Notch signaling fate-restricts cardiomyocytes toward a conduc- spontaneous arrhythmias, including Mobitz type II AV block and
tion lineage.67 ventricular tachycardia. Selective ablation of Tbx5 in the VCS
resulted in a corresponding loss of Gja5 (Cx40) and Scn5a expres-
sion.73,74 SCN5a mutations have been identified in patients with
Proximal Ventricular Conduction System progressive cardiac conduction disease that can exist alone or as
overlap syndromes with Brugada or LQT3.75 Inherited progres-
The proximal ventricular conduction system consists of the His sive cardiac conduction disease is associated with a high risk of
bundle and bundle branches. These fast conducting components complete AV block and Stoke-Adams syncope.75-77
of the CCS derive from ventricular myocardial precursors. For- GWAS focused on CCS parameters have identified several
mation of the proximal VCS is critically dependent on the coex- loci (TBX5, TBX3, NKX2-5, SCN10A, and SCN5A) that modu-
pression of Nkx2-5 and Tbx5, as combined haploinsufficiency of late PR and QRS duration in the general population.78-82 Given
these two factors results in developmental failure of the His the importance of Tbx5 and Tbx3 in CCS specification and func-
bundle and bundle branches.32 During development, the expres- tion, the Scn5a locus was screened for Tbx-responsive elements.
sion levels of Nkx2-5 and Tbx5 increase significantly in the proxi- A Tbx-responsive enhancer was identified approximately 15 kb
mal VCS, which drives the expression of a unique set of genes downstream of Scn5a and mirrored Scn5a expression in the
that impart fast conduction properties (Cx40+, Nav1.5+) while AV bundle and bundle branches.74 Using another approach, the
also maintaining phenotypic distinctiveness from working myo- genome-wide occupancy profile of Tbx3 was performed using
cytes (Tbx3+, Id2+, Cx43−; see Figure 29-4, D). Both Nkx2-5 and chromatin immunoprecipitation-massive parallel sequencing
Tbx5 have cell-autonomous roles in VCS development and per- (ChIP-Seq) analysis.83 Two Tbx3/Tbx5 responsive enhancers
turbation of either factor leads to significant CCS structural and were located within the Scn5a/Scn10a locus. The orthologous
functional abnormalities. human fragments had expression patterns similar to their
Nkx2-5 haploinsufficient mice exhibit marked hypoplasia of mouse counterparts within the developing ventricular conduc-
the ventricular conduction system resulting in PR and QRS tion system, providing evidence that these enhancers were
functionally conserved in humans.83 Furthermore, the GWAS PFN derives from the trabecular myocardium likely through a
single nucleotide polymorphism, rs6801957, was positioned in a process of recruitment; therefore, the complex patterning of the
conserved enhancer site in the SCN10A locus. Interestingly, the PFN directly mirrors the complexity of the trabecular network.8
single nucleotide polymorphism itself alters a highly conserved Paracrine signals from the overlying endocardium, such as
residue in the consensus T-box binding site, altering its ability to endothelin-1 (ET-1) and neuregulin-1 (NRG-1), direct underly-
respond to Tbx5 activation and Tbx3 repression.83 These data ing trabecular myocytes towards a conduction path (see Figure
reaffirm the fidelity of GWAS in identifying gene regulatory 29-4, E).91 The involvement of ET-1 was discovered in the chick,
networks that underlie phenotypic variations in humans on a where unlike in mammals, avian Purkinje fibers are found endo-
population scale. cardially and periarterially.8 Periarterial Purkinje fibers (PFs)
The maintenance of Cx40 in the proximal VCS is vital for develop in a pattern that strictly follows coronary vascular
proper function of the His bundle and bundle branches. Mice anatomy. It was shown subsequently that exposure of cultured
deficient in Cx40 exhibit slowed conduction through the His- embryonic myocytes to ET-1 was able to convert these cells to
Purkinje system and display bundle branch block.84,85 During a PF lineage.8,91 Exposure of mouse embryonic stem cells to
development, Cx40 is expressed in the ventricular trabeculae, the ET-1, but not NRG-1, was able to upregulate CCS markers.92
site of His-Purkinje formation.3,4,86 Cx40 expression in the ven- Retroviral-mediated coexpression of both the endothelin-1 pre-
tricular trabeculae dictates the switch in ventricular activation cursor (preproendothelin-1) and ECE-1 in the chick embryonic
from a pattern of activation to the mature apical-to-basal myocardium induced myocytes to express PF genes in vivo.93
sequence.86 The regional localization of Cx40 within the proxi- ET-1 is converted to its active form by endothelin-converting
mal VCS is dependent on several transcription factors that func- enzyme (ECE), which is highly expressed in endothelial cells and
tion downstream of Nkx2-5 and Tbx5: Hf-1b, Hop, and Irx3. in developing PFs.94 ECE-1 expression is regulated by hemody-
Hf-1b is a zinc-finger transcription factor with enriched namic load. Conotruncal banding, a model of ventricular pres-
expression within the compact myocardial layers and the ven- sure overload, resulted in significant upregulation of ECE-1
tricular conduction system.87 Mice deficient in Hf-1b manifest expression and concomitant precocious expansion of Cx40+ PFs.95
sinus node dysfunction, intermittent AV block, and sudden As discussed before, the shift in ventricular activation to an
cardiac death due to spontaneous ventricular arrhythmias. apex-to base pattern directly correlates with Cx40 expression.
Hf-1b−/− hearts demonstrated a reduced number of Cx40+ distal Banding resulted in an earlier shift in ventricular activation to an
Purkinje fibers and as well as an abnormal intracellular distribu- apex-to-base pattern, whereas pressure unloading delayed this
tion of Cx40, suggesting a defect in trafficking. This phenotype process. Therefore, ventricular hemodynamics is a key epigenetic
was seen predominantly in the ventricular apex; apical working factor in the regulation of HPS development.94-96
myocytes were also smaller and had significantly reduced levels Neuregulin-1 (Nrg-1) is another endothelial-derived factor
of Cx43.88 In addition, the coronary arterial structure and func- with a significant role in PF specification. Using an in vitro
tion were also perturbed within the apical region.88 Therefore, embryonic culture system, the response of developing CCS-LacZ
the defects in Cx40 and Cx43 expression within the ventricular hearts to Nrg-1 and ET-1 treatment was studied.97 Nrg-1 induced
apex could be multifactorial and will need further investigation. a profound upregulation of LacZ+ cells in the developing heart
Homeodomain-only protein (Hop) is a unique member of the in a time window between E8.5 and E10.5. The increase in LacZ+
homeobox transcription factors that does not directly bind cells could not be explained by changes in proliferation or apop-
DNA.89 It is known to function downstream of Nkx2-5 and tosis, and suggested that Nrg-1 induced embryonic cardiomyo-
inhibits serum response factor (SRF)-dependent transcription.68 cytes toward a conduction phenotype.97 The marked upregulation
Hop is expressed in the cardiac conduction system, with signifi- of CCS-LacZ–positive cells in Nrg-1–treated hearts was associ-
cant enrichment beyond the neonatal time period. Hop−/− mice ated with changes in the ventricular activation pattern on optical
have structurally normal hearts and CCS anatomy but manifest mapping. The effect of ET-1 on CCS-LacZ upregulation,
slowed conduction in the HPS on electrophysiological testing. however, was more modest.97
Expression of Cx40 was significantly reduced in the atria and the The ability of NRG-1 and ET-1 to induce a Purkinje gene
HPS, whereas levels of Cx43 remained normal.89 Therefore, Hop profile was evaluated in dissociated mouse embryonic ventricular
is not required for CCS specification and patterning, but it has myocytes.98 Both Nrg-1 and ET-1 increased the expression of
an important role in optimizing fast conduction in the HPS Nkx2-5, Gata4, Irx4, Hop, HF-1b, minK, Cx40, and Cx45.
through its action on Cx40 expression. There was no cumulative effect of coadministration of NRG-1
The Iroquois homeobox 3 (Irx3) transcription factor is and ET-1 on Purkinje gene programming.98 The ability of ET-1
expressed in the VCS and regulates the fast conduction proper- and NRG-1 to activate Nkx2-5, Gata4, HF-1b, and Hop suggests
ties of the His-Purkinje system.90 Irx3 knockout mice exhibit that these secreted factors activate a potent gene regulatory
prolonged QRS intervals on surface ECG. Bundle branch cells network for PF specification.
deficient in Irx3 had reduced levels of Cx40 and ectopic expres- Nkx2-5 gene dosage is critical for proper maturation of the
sion of Cx43. The direct coupling of His-bundle cells to working PFN.99 Inappropriately high or low levels of Nkx2-5 can nega-
myocytes presumably resulted in source-sink mismatch, leading tively affect proper Purkinje development.99 Consistent with
to charge dissipation and conduction block. Irx3 represses Cx43 other models of Nkx2-5 deficiency, adult Nkx2-5+/−/Cx40eGFP/+
expression by directly binding a putative Irx3 responsive element mice showed marked reduction in HPS cellularity, despite
that overlaps with an Nkx2-5 binding site near a conserved T-box the trabecular region appearing qualitatively normal during
binding element. The activation of Cx40, on the other hand, is development.69 Evaluation of the remaining PFs revealed that
through an indirect mechanism.90 Therefore, Hf-1b, Hop, and they were of normal size and had normal electrophysiologic
Irx3 function in a nonredundant manner to enhance Cx40 expres- properties.69 Therefore, the conduction defects seen in
sion in the VCS, whereas simultaneously Irx3 and Tbx3 repress Nkx2-5+/−/Cx40eGFP/+ hearts appears to be due to HPS patterning
Cx43 expression in specialized conduction cells. abnormalities rather than electrophysiological defects at the cel-
lular level.69 The reduction of Purkinje fibers in Nkx2-5+/− hearts
occurs mainly after birth and is due to either reduced PF recruit-
Purkinje Fiber Network ment or loss of PF in the postnatal period.69 Based on chimeric
analysis, postnatal development of the PFN was critically depen-
The Purkinje fiber network (PFN) represents the most distal dent on the dose of Nkx2-5, which behaved in a cell-autonomous
aspect of the ventricular conduction system. In mammals, the manner.69
In summary, the genetic programs that dictate the cardiac developmental origins of the CCS has grown dramatically. Every
29
conduction system are fundamentally based on the intricate fundamental discovery, from the cellular origins of the CCS to the
balance between activating and repressing T-box transcription transcriptional networks that govern the properties of slow and
factors. Numerous regulatory feedback loops, involving Nkx2-5 fast conduction, has created a broader and richer understanding of
and other homeobox proteins, determine the gene dosage the how the heart keeps time. Incorporating newer modalities,
of Tbx5, Tbx2/3, and Tbx18 to specify whether a cardiac such as GWAS and ChIP-Seq, will further enhance understanding
progenitor cell becomes a conduction cell or a working cardio- of how these transcription factor networks regulate CCS specifica-
myocyte. Conduction cells are then imparted with fast and slow tion on a genome-wide scale. Ultimately, this understanding will
conduction properties based on the expression levels of Tbx5 and shape the future therapy of heart rhythm disorders.
Tbx3, respectively.
Acknowledgments
Conclusion
This work was supported by National Institutes of Health Grants
Since the first anatomic descriptions of the cardiac conduction R01HL82727 and R01HL105983 and New York State Grant
system in the early twentieth century, knowledge of the N08G-132.
15. Boogerd KJ, Wong LY, Christoffels VM, et al: 29. Hoogaars WM, Engel A, Brons JF, et al: Tbx3
References Msx1 and Msx2 are functional interacting part- controls the sinoatrial node gene program and
ners of T-box factors in the regulation of Con- imposes pacemaker function on the atria. Genes
1. Paff GH, Boucek RJ, Harrell TC: Observations nexin43. Cardiovasc Res 78:485–493, 2008. Dev 21:1098–1112, 2007.
on the development of the electrocardiogram. 16. Mommersteeg MT, Hoogaars WM, Prall OW, et 30. Habets PE, Moorman AF, Clout DE, et al: Coop-
Anat Rec 160:575–582, 1968. al: Molecular pathway for the localized formation erative action of Tbx2 and Nkx2.5 inhibits ANF
2. Kreuzberg MM, Willecke K, Bukauskas FF: of the sinoatrial node. Circ Res 100:354–362, 2007. expression in the atrioventricular canal: implica-
Connexin-mediated cardiac impulse propagation: 17. Kupershmidt S, Yang T, Anderson ME, et al: tions for cardiac chamber formation. Genes Dev
connexin 30.2 slows atrioventricular conduction Replacement by homologous recombination of 16:1234–1246, 2002.
in mouse heart. Trends Cardiovasc Med 16:266– the minK gene with lacZ reveals restriction of 31. Munshi NV, McAnally J, Bezprozvannaya S, et al:
272, 2006. minK expression to the mouse cardiac conduction Cx30.2 enhancer analysis identifies Gata4 as a
3. Viragh S, Challice CE: The development of the system. Circ Res 84:146–152, 1999. novel regulator of atrioventricular delay. Devel-
conduction system in the mouse embryo heart. II. 18. Rentschler S, Vaidya DM, Tamaddon H, et al: opment 136:2665–2674, 2009.
Histogenesis of the atrioventricular node and Visualization and functional characterization of 32. Moskowitz IP, Kim JB, Moore ML, et al: A
bundle. Dev Biol 56:397–411, 1977. the developing murine cardiac conduction system. molecular pathway including Id2, Tbx5, and
4. Viragh S, Challice CE: The development of the Development 128:1785–1792, 2001. Nkx2-5 required for cardiac conduction system
conduction system in the mouse embryo heart. I. 19. Kondo RP, Anderson RH, Kupershmidt S, et al: development. Cell 129:1365–1376, 2007.
The first embryonic A-V conduction pathway. Development of the cardiac conduction system as 33. Moskowitz IP, Pizard A, Patel VV, et al: The
Dev Biol 56:382–396, 1977. delineated by minK-lacZ. J Cardiovasc Electro- T-Box transcription factor Tbx5 is required for
5. Viragh S, Challice CE: The development of the physiol 14:383–391, 2003. the patterning and maturation of the murine
conduction system in the mouse embryo heart. 20. Miquerol L, Meysen S, Mangoni M, et al: Archi- cardiac conduction system. Development 131:
Dev Biol 80:28–45, 1980. tectural and functional asymmetry of the His- 4107–4116, 2004.
6. Wenink AC: Development of the human cardiac Purkinje system of the murine heart. Cardiovasc 34. Mori AD, Zhu Y, Vahora I, et al: Tbx5-dependent
conducting system. J Anat 121:617–631, 1976. Res 63:77–86, 2004. rheostatic control of cardiac gene expression and
7. Cheng G, Litchenberg WH, Cole GJ, et al: 21. Pallante BA, Giovannone S, Fang-Yu L, et al: morphogenesis. Dev Biol 297:566–586, 2006.
Development of the cardiac conduction system Contactin-2 expression in the cardiac Purkinje 35. Bruneau BG, Logan M, Davis N, et al: Chamber-
involves recruitment within a multipotent cardio- fiber network. Circ Arrhythm Electrophysiol specific cardiac expression of Tbx5 and heart
myogenic lineage. Development 126:5041–5049, 3:186–194, 2010. defects in Holt-Oram syndrome. Dev Biol 211:
1999. 22. Basson CT, Bachinsky DR, Lin RC, et al: Muta- 100–108, 1999.
8. Gourdie RG, Mima T, Thompson RP, et al: Ter- tions in human TBX5 [corrected] cause limb and 36. Mommersteeg MT, Dominguez JN, Wiese C, et
minal diversification of the myocyte lineage gen- cardiac malformation in Holt-Oram syndrome. al: The sinus venosus progenitors separate and
erates Purkinje fibers of the cardiac conduction Nat Genet 15:30–35, 1997. diversify from the first and second heart fields
system. Development 121:1423–1431, 1995. 23. Stennard FA, Harvey RP: T-box transcription early in development. Cardiovasc Res 87:92–101,
9. Miquerol L, Moreno-Rascon N, Beyer S, et al: factors and their roles in regulatory hierarchies in 2010.
Biphasic development of the mammalian ven- the developing heart. Development 132:4897– 37. Wiese C, Grieskamp T, Airik R, et al: Formation
tricular conduction system. Circ Res 107:153– 4910, 2005. of the sinus node head and differentiation of sinus
161, 2010. 24. Linden H, Williams R, King J, et al: Ulnar node myocardium are independently regulated by
10. Kitajima S, Miyagawa-Tomita S, Inoue T, et al: Mammary syndrome and TBX3: expanding the Tbx18 and Tbx3. Circ Res 104:388–397, 2009.
Mesp1-nonexpressing cells contribute to the ven- phenotype. Am J Med Genet A 149A:2809–2812, 38. Puskaric S, Schmitteckert S, Mori AD, et al:
tricular cardiac conduction system. Dev Dyn 2009. Shox2 mediates Tbx5 activity by regulating Bmp4
235:395–402, 2006. 25. Meneghini V, Odent S, Platonova N, et al: in the pacemaker region of the developing heart.
11. Poelmann RE, Jongbloed MR, Molin DG, et al: Novel TBX3 mutation data in families with Hum Mol Genet 19:4625–4633, 2010.
The neural crest is contiguous with the cardiac ulnar-mammary syndrome indicate a genotype- 39. Frank DU, Carter KL, Thomas KR, et al: Lethal
conduction system in the mouse embryo: a role in phenotype relationship: mutations that do not arrhythmias in Tbx3-deficient mice reveal
induction? Anat Embryol (Berl) 208:389–393, 2004. disrupt the T-domain are associated with less severe extreme dosage sensitivity of cardiac conduction
12. Hoogaars WM, Tessari A, Moorman AF, et al: limb defects. Eur J Med Genet 49:151–158, 2006. system function and homeostasis. Proc Natl Acad
The transcriptional repressor Tbx3 delineates the 26. Hiroi Y, Kudoh S, Monzen K, et al: Tbx5 associ- Sci U S A 109:E154–163, 2012.
developing central conduction system of the ates with Nkx2-5 and synergistically promotes 40. Blaschke RJ, Hahurij ND, Kuijper S, et al: Tar-
heart. Cardiovasc Res 62:489–499, 2004. cardiomyocyte differentiation. Nat Genet 28: geted mutation reveals essential functions of the
13. Meilhac SM, Kelly RG, Rocancourt D, et al: A 276–280, 2001. homeodomain transcription factor Shox2 in sino-
retrospective clonal analysis of the myocardium 27. Bruneau BG, Nemer G, Schmitt JP, et al: A murine atrial and pacemaking development. Circulation
reveals two phases of clonal growth in the devel- model of Holt-Oram syndrome defines roles of the 115:1830–1838, 2007.
oping mouse heart. Development 130:3877–3889, T-box transcription factor Tbx5 in cardiogenesis 41. Espinoza-Lewis RA, Yu L, He F, et al: Shox2 is
2003. and disease. Cell 106:709–721, 2001. essential for the differentiation of cardiac pace-
14. Christoffels VM, Hoogaars WM, Tessari A, et al: 28. Garg V, Kathiriya IS, Barnes R, et al: GATA4 maker cells by repressing Nkx2-5. Dev Biol
T-box transcription factor Tbx2 represses differ- mutations cause human congenital heart defects 327:376–385, 2009.
entiation and formation of the cardiac chambers. and reveal an interaction with TBX5. Nature 42. Wang J, Klysik E, Sood S, et al: Pitx2 prevents
Dev Dyn 229:763–770, 2004. 424:443–447, 2003. susceptibility to atrial arrhythmias by inhibiting
left-sided pacemaker specification. Proc Natl 61. Lalani SR, Thakuria JV, Cox GF, et al: 20p12.3 81. Pfeufer A, van Noord C, Marciante KD, et al:
Acad Sci U S A 107:9753–9758, 2010. microdeletion predisposes to Wolff-Parkinson- Genome-wide association study of PR interval.
43. Ammirabile G, Tessari A, Pignataro V, et al: Pitx2 White syndrome with variable neurocognitive Nat Genet 42:153–159, 2010.
confers left morphological, molecular, and func- deficits. J Med Genet 46:168–175, 2009. 82. Smith JG, Magnani JW, Palmer C, et al: Genome-
tional identity to the sinus venosus myocardium. 62. Chi NC, Shaw RM, De Val S, et al: Foxn4 directly wide association studies of the PR interval in
Cardiovasc Res 93:291–301, 2012. regulates tbx2b expression and atrioventricular African Americans. PLoS Genet 7:e1001304,
44. Gudbjartsson DF, Arnar DO, Helgadottir A, et al: canal formation. Genes Dev 22:734–739, 2008. 2011.
Variants conferring risk of atrial fibrillation on 63. Rutenberg JB, Fischer A, Jia H, et al: Develop- 83. van den Boogaard M, Wong LY, Tessadori F,
chromosome 4q25. Nature 448:353–357, 2007. mental patterning of the cardiac atrioventricular et al: Genetic variation in T-box binding element
45. Kapoor N, Galang G, Marban E, et al: Transcrip- canal by Notch and Hairy-related transcription functionally affects SCN5A/SCN10A enhancer.
tional suppression of connexin43 by TBX18 factors. Development 133:4381–4390, 2006. J Clin Invest 122:2519–2530, 2012.
undermines cell-cell electrical coupling in postna- 64. Milan DJ, Giokas AC, Serluca FC, et al: Notch1b 84. van Rijen HV, van Veen TA, van Kempen MJ,
tal cardiomyocytes. J Biol Chem 286:14073– and neuregulin are required for specification of et al: Impaired conduction in the bundle
14079, 2011. central cardiac conduction tissue. Development branches of mouse hearts lacking the gap junction
46. Ko YS, Yeh HI, Ko YL, et al: Three-dimensional 133:1125–1132, 2006. protein connexin40. Circulation 103:1591–1598,
reconstruction of the rabbit atrioventricular con- 65. Grego-Bessa J, Luna-Zurita L, del Monte G, 2001.
duction axis by combining histological, desmin, et al: Notch signaling is essential for ventricular 85. Tamaddon HS, Vaidya D, Simon AM, et al: High-
and connexin mapping data. Circulation 109: chamber development. Dev Cell 12:415–429, resolution optical mapping of the right bundle
1172–1179, 2004. 2007. branch in connexin40 knockout mice reveals slow
47. Efimov IR, Mazgalev TN: High-resolution, 66. Rentschler S, Harris BS, Kuznekoff L, et al: conduction in the specialized conduction system.
three-dimensional fluorescent imaging reveals Notch signaling regulates murine atrioventricular Circ Res 87:929–936, 2000.
multilayer conduction pattern in the atrioven- conduction and the formation of accessory path- 86. Sankova B, Benes J, Jr, Krejci E, et al: The effect
tricular node. Circulation 98:54–57, 1998. ways. J Clin Invest 121:525–533, 2011. of connexin40 deficiency on ventricular conduc-
48. Aanhaanen WT, Brons JF, Dominguez JN, et al: 67. Rentschler S, Yen AH, Lu J, et al: Myocardial tion system function during development. Car-
The Tbx2+ primary myocardium of the atrioven- Notch signaling reprograms cardiomyocytes to a diovasc Res 95:469–479, 2012.
tricular canal forms the atrioventricular node and Conduction-Like Phenotype. Circulation 2012. 87. Nguyen-Tran VT, Kubalak SW, Minamisawa S,
the base of the left ventricle. Circ Res 104:1267– 68. Pashmforoush M, Lu JT, Chen H, et al: Nkx2-5 et al: A novel genetic pathway for sudden cardiac
1274, 2009. pathways and congenital heart disease; loss of ven- death via defects in the transition between ven-
49. Horsthuis T, Buermans HP, Brons JF, et al: Gene tricular myocyte lineage specification leads to tricular and conduction system cell lineages. Cell
expression profiling of the forming atrioventricu- progressive cardiomyopathy and complete heart 102:671–682, 2000.
lar node using a novel tbx3-based node-specific block. Cell 117:373–386, 2004. 88. Hewett KW, Norman LW, Sedmera D, et al:
transgenic reporter. Circ Res 105:61–69, 2009. 69. Meysen S, Marger L, Hewett KW, et al: Nkx2.5 Knockout of the neural and heart expressed gene
50. Aanhaanen WT, Mommersteeg MT, Norden J, cell-autonomous gene function is required for the HF-1b results in apical deficits of ventricular
et al: Developmental origin, growth, and three- postnatal formation of the peripheral ventricular structure and activation. Cardiovasc Res 67:
dimensional architecture of the atrioventricular conduction system. Dev Biol 303:740–753, 548–560, 2005.
conduction axis of the mouse heart. Circ Res 2007. 89. Ismat FA, Zhang M, Kook H, et al: Homeobox
107:728–736, 2010. 70. Wakimoto H, Kasahara H, Maguire CT, et al: protein Hop functions in the adult cardiac con-
51. Ranganayakulu G, Elliott DA, Harvey RP, et al: Developmentally modulated cardiac conduction duction system. Circ Res 96:898–903, 2005.
Divergent roles for NK-2 class homeobox genes failure in transgenic mice with fetal or postnatal 90. Zhang SS, Kim KH, Rosen A, et al: Iroquois
in cardiogenesis in flies and mice. Development overexpression of DNA nonbinding mutant homeobox gene 3 establishes fast conduction in
125:3037–3048, 1998. Nkx2.5. J Cardiovasc Electrophysiol 13:682–688, the cardiac His-Purkinje network. Proc Natl Acad
52. Tanaka M, Chen Z, Bartunkova S, et al: The 2002. Sci U S A 108:13576–13581, 2011.
cardiac homeobox gene Csx/Nkx2.5 lies geneti- 71. Kasahara H, Wakimoto H, Liu M, et al: Progres- 91. Gourdie RG, Wei Y, Kim D, et al: Endothelin-
cally upstream of multiple genes essential for sive atrioventricular conduction defects and heart induced conversion of embryonic heart muscle
heart development. Development 126:1269– failure in mice expressing a mutant Csx/Nkx2.5 cells into impulse-conducting Purkinje fibers.
1280, 1999. homeoprotein. J Clin Invest 108:189–201, 2001. Proc Natl Acad Sci U S A 95:6815–6818, 1998.
53. Jay PY, Harris BS, Maguire CT, et al: Nkx2-5 72. Risebro CA, Petchey LK, Smart N, et al: Epistatic 92. Gassanov N, Er F, Zagidullin N, et al: Endothelin
mutation causes anatomic hypoplasia of the rescue of Nkx2.5 adult cardiac conduction induces differentiation of ANP-EGFP expressing
cardiac conduction system. J Clin Invest disease phenotypes by prospero-related homeo- embryonic stem cells towards a pacemaker phe-
113:1130–1137, 2004. box protein 1 and HDAC3. Circ Res 111:e19–31, notype. FASEB J 18:1710–1712, 2004.
54. Aanhaanen WT, Boukens BJ, Sizarov A, et al: 2012. 93. Takebayashi-Suzuki K, Yanagisawa M, Gourdie
Defective Tbx2-dependent patterning of the 73. Arnolds DE, Moskowitz IP: Inducible recombi- RG, et al: In vivo induction of cardiac Purkinje
atrioventricular canal myocardium causes acces- nation in the cardiac conduction system of minK: fiber differentiation by coexpression of
sory pathway formation in mice. J Clin Invest CreERT(2) BAC transgenic mice. Genesis preproendothelin-1 and endothelin converting
121:534–544, 2011. 49:878–884, 2011. enzyme-1. Development 127:3523–3532, 2000.
55. Bakker ML, Boukens BJ, Mommersteeg MT, et 74. Arnolds DE, Liu F, Fahrenbach JP, et al: TBX5 94. Sedmera D, Harris BS, Grant E, et al: Cardiac
al: Transcription factor Tbx3 is required for the drives Scn5a expression to regulate cardiac con- expression patterns of endothelin-converting
specification of the atrioventricular conduction duction system function. J Clin Invest 122:2509– enzyme (ECE): Implications for conduction system
system. Circ Res 102:1340–1349, 2008. 2518, 2012. development. Dev Dyn 237:1746–1753, 2008.
56. Harrelson Z, Kelly RG, Goldin SN, et al: Tbx2 75. Probst V, Kyndt F, Potet F, et al: Haploinsuffi- 95. Hall CE, Hurtado R, Hewett KW, et al:
is essential for patterning the atrioventricular ciency in combination with aging causes SCN5A- Hemodynamic-dependent patterning of endothe-
canal and for morphogenesis of the outflow tract linked hereditary Lenegre disease. J Am Coll lin converting enzyme 1 expression and differen-
during heart development. Development 131: Cardiol 41:643–652, 2003. tiation of impulse-conducting Purkinje fibers in
5041–5052, 2004. 76. Kyndt F, Probst V, Potet F, et al: Novel SCN5A the embryonic heart. Development 131:581–592,
57. Ma L, Lu MF, Schwartz RJ, et al: Bmp2 is essen- mutation leading either to isolated cardiac con- 2004.
tial for cardiac cushion epithelial-mesenchymal duction defect or Brugada syndrome in a large 96. Reckova M, Rosengarten C, deAlmeida A, et al:
transition and myocardial patterning. Develop- French family. Circulation 104:3081–3086, 2001. Hemodynamics is a key epigenetic factor in devel-
ment 132:5601–5611, 2005. 77. Schott JJ, Alshinawi C, Kyndt F, et al: Cardiac opment of the cardiac conduction system. Circ
58. Singh R, Horsthuis T, Farin HF, et al: Tbx20 conduction defects associate with mutations in Res 93:77–85, 2003.
interacts with smads to confine tbx2 expression to SCN5A. Nat Genet 23:20–21, 1999. 97. Rentschler S, Zander J, Meyers K, et al:
the atrioventricular canal. Circ Res 105:442–452, 78. Sotoodehnia N, Isaacs A, de Bakker PI, et al: Neuregulin-1 promotes formation of the murine
2009. Common variants in 22 loci are associated with cardiac conduction system. Proc Natl Acad Sci
59. Gaussin V, Morley GE, Cox L, et al: Alk3/ QRS duration and cardiac ventricular conduction. U S A 99:10464–10469, 2002.
Bmpr1a receptor is required for development of Nat Genet 42:1068–1076, 2010. 98. Patel R, Kos L: Endothelin-1 and Neuregulin-1
the atrioventricular canal into valves and annulus 79. Chambers JC, Zhao J, Terracciano CM, et al: convert embryonic cardiomyocytes into cells of
fibrosus. Circ Res 97:219–226, 2005. Genetic variation in SCN10A influences cardiac the conduction system in the mouse. Dev Dyn
60. Stroud DM, Gaussin V, Burch JB, et al: Abnormal conduction. Nat Genet 42:149–152, 2010. 233:20–28, 2005.
conduction and morphology in the atrioventricu- 80. Holm H, Gudbjartsson DF, Arnar DO, et al: 99. Harris BS, Spruill L, Edmonson AM, et al: Dif-
lar node of mice with atrioventricular canal tar- Several common variants modulate heart rate, PR ferentiation of cardiac Purkinje fibers requires
geted deletion of Alk3/Bmpr1a receptor. interval and QRS duration. Nat Genet 42:117– precise spatiotemporal regulation of Nkx2-5
Circulation 116:2535–2543, 2007. 122, 2010. expression. Dev Dyn 235:38–49, 2006.
Cardiac Fibroblasts and
Arrhythmogenesis 30
Nenad Bursac and Jong J. Kim
the heart wall and gradually attain a fibroblast or smooth muscle

CHAPTER OUTLINE
phenotype. This process is regulated by the spatially and tempo-
Cardiac Fibroblasts 297 rally coordinated expression of different growth factors, includ-
ing platelet-derived growth factor (PDGF), fibroblast growth
Fibroblast and Myofibroblast Electrophysiology 299
factors (FGFs) and transforming growth factor (TGF) β.6 During
Arrhythmogenic Effects of Myofibroblasts 301 prenatal growth, the number of fibroblasts in the heart steadily
increases. Shortly after birth, the proliferative capacity of cardio-
Cardiac Fibroblasts as Antiarrhythmic Targets 305
myocytes ceases, while the fibroblast number abruptly increases
Conclusions and Future Directions 306 (greater than twofold) for reasons that may be related to the
postnatal increase in blood pressure (and associated mechanical
strains in the heart wall), oxygen tension, or both.
Cardiac Fibroblasts
Phenotypic Diversity and Lack of Specific Markers
Cardiomyocytes in the heart occupy approximately 75% of the
myocardial volume, but the majority of cells by number are not Although there is significant knowledge of the role of cardiac
muscular, including fibroblasts, endothelial cells, pericytes, fibroblasts in the heart, the molecular properties of these cells
smooth muscle cells, macrophages, and mast and dendritic cells. remain poorly characterized, mainly because of the lack of specific
Cardiac fibroblasts, the most numerous of these non-muscular and ubiquitous markers of their phenotype.7 The markers cur-
cells, comprise 30% to 65% of all the cells in the healthy adult rently used are either expressed in only a fraction of fibroblasts or
heart, but only a minor fraction of the total heart volume.1 The also label other cells (Table 30-1). For example, immunostaining
number of fibroblasts in the heart is not constant, but dynami- for vimentin has been used widely to label cardiac fibroblasts;
cally changes during development and disease and with aging.2-4 however, this intracellular protein is expressed in other mesoderm-
Traditionally, cardiac fibroblasts have been considered as passive derived cells, including endothelial and smooth muscle cells.
cells that primarily maintain the structural and mechanical integ- Similarly, Thy-1 (CD90), a surface marker that labels cardiac
rity of the heart through the highly regulated synthesis and deg- fibroblasts and other mesenchymal cells also labels endothelial
radation of extracellular matrix (ECM) proteins including cells. Discoidin domain receptor 2 (DDR2), a membrane collagen-
collagens (mainly type I and III) and fibronectin. Recently, binding tyrosine kinase receptor, is expressed in a subset of cardiac
however, a large body of evidence has emerged suggesting the fibroblasts, but is also found in endothelial and smooth muscle
important roles of these connective tissue cells in cardiac devel- cells. Fibroblast-specific protein-1 (FSP1, S100A4) is another
opment, function, and pathology, including cardiac arrhythmias. commonly used fibroblast marker that is only sparsely expressed
Because cardiac fibroblasts are highly abundant and closely inter- in the normal heart, but is significantly upregulated in heart
laced with other cell types in the heart (virtually every cardio- disease, where it may also label smooth muscle cells.8 Recently,
myocyte touches one or more fibroblasts), they are in a position Acharya et al9 reported the generation of transgenic mice
to regulate actively and to modify heart function by direct contact in which transcription factor Tcf21 (Epicardin/Pod1/Capsulin)
with other cells and ECM, as well as through the secretion of was selectively expressed in a subset of epicardially derived cardiac
different cytokines, ECM proteins, and proteases. fibroblasts, but also in a multitude of other tissues.9 In general, the
best existing methods to identify and isolate cardiac fibroblasts
rely on the labeling with multiple positive and negative markers,
Fibroblast Origin such as CD31–/CD90+/DDR2+ for live cell isolation,4 or von Wil-
lebrand factor–/smooth muscle actin (SMA)–/vimentin+/DDR2+
Cardiac fibroblasts are the only cell type in the heart not associ- for immunolabeling.1,10 Although genetic fate mapping methods
ated with the basement membrane. They are recognized as for in situ labeling of cardiac fibroblasts in mice are extremely
single-nucleated, spindle-shaped cells with multiple processes valuable, they require a careful interpretation that accounts for
and prominent endoplasmic reticulum and Golgi apparatus. the potential pitfalls of this system.5
Fibroblasts reside in the self-secreted extracellular matrix Identifying specific molecular markers of cardiac fibroblasts is
arranged in sheets and strands that tightly envelop cardiomyocyte additionally complicated by the diversity of their phenotype as
fibers.1,2 Their spatial distribution in the healthy heart is mainly contributed by their specific developmental origin (e.g., epicar-
uniform, with the highest density found in the sinoatrial (SA) dial vs. endocardial), location in the heart (e.g., atria vs. ventricle,
node and annulus fibrosis, where these cells provide electrical left vs. right heart, valves, SA node, atrioventricular groove),
insulation to enable successful pacemaking function and coordi- pathologic state (e.g., infarction, pressure overload), and aging.
nated activation of atria and ventricles. Developmentally, cardiac For example, a subset of fibroblastic cells positive for PDGF
fibroblasts are derived from mesenchymal cells of a proepicardial receptor-α and stem cell antigen 1 (Sca-1) was recently identified
organ, which migrate over the heart surface to form the epicar- in perivascular interstitial regions of the mouse heart.11 These
dium.2,5 Through epithelial-to-mesenchymal transition, the epi- epicardially derived cells, termed cardiac colony-forming-unit fibro-
cardium gives rise to epicardium-derived cells that migrate into blasts, can undergo a long-term expansion for approximately 40
297
program during hypertrophic cardiomyopathy or heart failure

Table 30-1. Fibroblast and Myofibroblast Markers
may be contributed by analogous phenotypic changes in cardiac
Marker Cellular Overlap fibroblasts.
Detailed phenotypic and functional characterization of cardiac
Colla1 Various cells fibroblasts can be performed in vitro, but these studies must be
CD40 Various antigen presenting cells interpreted with caution because of various confounding factors,
including the possibility that enzymatic or outgrowth cell isola-
CD248 (TEM1) Pericytes, endothelial cells tion selects for a particular pool of cardiac fibroblasts and that
Cadherin-11 Carcinoma and retina epithelial cells fibroblast phenotype is significantly altered by higher stiffness of
the cell attachment substrate (several GPa for glass or plastic vs.
FSP1/S100A4 Smooth muscle cells, invasive
tens of KPa for intact tissue),13 increased oxygen tension (21%
carcinoma cells
for ambient air vs. 5% in intact tissue),14 and the lack of neuro-
Fibroblast surface antigen Monocytes/macrophages humoral and inflammatory factors, cyclic stretch, restricted extra-
(FSA) cellular space, and vasculature. Direct culture of freshly isolated
Discoidin domain receptor 2 Endothelial cells, smooth muscle fibroblasts within a biomimetic three-dimensional environment
(DDR2) cells (e.g., cyclically stretched soft hydrogel) is likely to better preserve
native cell phenotype compared with the use of standard two-
Fibroblast activation Activated melanocytes dimensional culture conditions.15
protein-1 (FAP1)
Prolyl-4-hydroxylase Endothelial cells, epithelial cells
PDGF receptor-β (PDGFRb) Smooth muscle cells, pericytes Cardiac Myofibroblasts
Heat shock protein-47 Monocytes/macrophages, various Although cardiac fibroblasts are the most dominant nonmyocyte
(HSP47) collagen-producing cells cell type in the healthy heart, cardiac disease or myocyte loss
Thymus cell antigen-1 Leukocytes, endothelial cells, various owing to myocardial infarction, hypertension, inflammation, and
(THY1/CD90) progenitor cells other stress signals is associated with the appearance and prolif-
eration of cardiac myofibroblasts to either replace dead myocytes
Vimentin Endothelial cells, smooth muscle with a collagenous scar (replacement fibrosis) or yield interstitial
cells, various cells or perivascular collagen accumulation (reactive fibrosis). Myofi-
Palladin 4Ig* Smooth muscle cells broblasts exhibit a contractile cell phenotype intermediate
between that of fibroblasts and smooth muscle cells.16,17 Com-
Periostin* Bone and carcinoma cells
pared with fibroblasts, myofibroblasts are larger cells with
Cofilin* Smooth muscle cells increased expression of stress fibers that exhibit enhanced prolif-
AngII type 1 receptor (AT1R)* Cardiomyocytes, smooth muscle eration, migration, and secretion of ECM proteins (e.g., collagen
cells I, collagen III, fibronectin), ECM degradation enzymes (matrix
metalloproteinases [MMPs]) and their inhibitors (tissue inhibi-
TGF-β receptor* Various cells tors of metalloproteinases [TIMPs]).18 In addition, microfibro-
Frizzled-2* Smooth muscle cells blasts show both the increased secretion of and responsiveness to
various cytokines and growth factors (e.g., AngII, TGF-β, PDGF,
α-Smooth muscle actin* Smooth muscle cells
ET-1, tumor necrosis factor [TNF] α, interleukin [IL] 1β).17,18
Integrins (αvβ3, α1β1, α2β1, Endothelial cells, various cells Although mainly absent from the normal heart (with the excep-
α11β1)* tion of the heart valves), myofibroblasts are found to participate
Collagen types I, III, IV, V, VI* Various cells actively in the cardiac wound healing response, where they
support rapid tissue remodeling and the formation of a fibrous
Lysyl oxidase* Smooth muscle cells scar. The efficient scar formation initially serves to prevent
Fibronectin ED-A* Smooth muscle cells harmful dilatation of the heart; however, the long-term persis-
tence and activity of myofibroblasts in the infarct scar or other
Tenascin C* Smooth muscle cells
regions of the myocardium can lead to excessive collagen accu-
* Myofibroblast markers only. mulation, stiffening of the heart, pathologic remodeling, and
eventually cardiac malfunction and failure. This persistence of
the activated myofibroblasts in the heart after the mature scar is
passages and differentiate into various cells of mesodermal formed contrasts wound-healing processes observed in other
lineage. The potential role of these highly proliferative cells in organs where myofibroblasts undergo apoptosis and disappear
cardiac repair, remodeling, and fibrotic disease remains to be upon scar formation.16
explored. Phenotypic diversity was also identified between canine Although myofibroblasts express all the known molecular
ventricular and atrial fibroblasts, with atrial fibroblasts being markers used to label cardiac fibroblasts (DDR2, vimentin, Thy1,
more proliferative in response to a variety of growth stimuli (e.g., FSP1, periostin), no specific myofibroblast markers have been
fetal bovine serum, PDGF, FGF-2, TGF-β1, angiotensin II identified (see Table 30-1).16 Compared with fibroblasts, de novo
[AngII], endothelin-1 [ET-1]).12 These differences were ampli- myofibroblasts also express or increase the expression of various
fied in congestive heart failure and eliminated using a PDGF cytoskeletal (α smooth muscle actin [α-SMA], SM22α, myosin
receptor blocker, AG1295. Different functional roles were also heavy chain-B, tropomyosin) and cell adhesion (paxillin, tensin,
reported for fetal and adult mouse cardiac fibroblasts. Ieda et al4 fibronectin ED-A) proteins.13,16 Furthermore, myofibroblasts and
have shown that adult fibroblasts promote cardiomyocyte hyper- fibroblasts do not express the intermediate filament marker
trophy, whereas fetal fibroblasts promote embryonic cardiomyo- desmin, and they lack the expression of mature smooth muscle
cyte proliferation by secreting fibronectin, collagen, and markers, such as smooth muscle myosin heavy chains. Regarding
heparin-binding endothelial growth factor that activate β1- their origin in fibrotic disease, myofibroblasts are traditionally
integrin signaling in cardiomyocytes.4 This finding suggests the thought to be derived from the resident (interstitial and adventi-
possibility that the observed reactivation of the cardiac fetal gene tial) fibroblasts that proliferate, migrate, and differentiate into a
Cardiac Fibroblasts and Arrhythmogenesis 299
HVF nRVF nMVF
30
A 50 µm B C
BV
BV
BV
D 50 µm E F
Figure 30-1. Specificity of myofibroblast labeling with anti–α-smooth muscle actin (α-SMA) antibodies. Cultured passage 2 fibroblasts from adult human (A), neonatal rat
(B), and neonatal mouse (C) ventricles are shown. D-F, Ventricular tissue sections from healthy 2-month-old mice. Samples were stained for α-SMA using a monoclonal
mouse antibody from Sigma-Aldrich (A2547, red) and polyclonal rabbit antibody from Abcam (ab5694, green) to label myofibroblasts and smooth muscle cells, filamentous
actin (gray in E) and sarcomeric α-actinin (gray in F) to label cardiomyocytes, and 4,6-diamino-2-phenylindole (DAPI; blue) to label nuclei. Yellow arrowheads in D and
E denote nonspecific, nonmyocyte labeling, which is absent in F. Anti–α-SMA images for the two antibodies were acquired using the same exposure time. BV, Blood vessel.
myofibroblast phenotype.5 This view is based on in vitro studies cell culture attributed to the high stiffness of the attachment sub-
showing increased fibroblast proliferation, migration, and con- strate or a switch to hyperoxic (ambient air) conditions.13,14 The
version to a myofibroblast phenotype in response to a variety of time course and extent of this process strongly depend on the
cytokines related to cardiac injury and disease (TGF-β, AngII, particulars of cell isolation and culture conditions, with some
TNF-α, IL-1β, IL-6, ET-1).18 However, the need to form a col- reports showing ubiquitous α-SMA expression as early as 1 to 2
lagenous scar rapidly to replace dead myocardium (thereby pre- days after fibroblast plating,24 whereas others describe little or no
venting wall rupture or dilatation after injury) would suggest that phenotype switch until passage 2 to 3.10 A contributing factor to
a significant portion of myofibroblasts at the injury site should this variability might be the use of different anti–α-SMA antibod-
be derived from nonresident or nonfibroblastic cells, rather than ies by different groups. Figure 30-1 shows passage-1 cultured
the activation and long-range migration of remote resident ventricular fibroblasts isolated from adult human, neonatal rat,
cardiac fibroblasts. In line with this reasoning, recent genetic fate and neonatal mouse tissues stained by two commonly used anti–
mapping studies in mice have demonstrated that during acute α-SMA antibodies, a monoclonal mouse antibody from Sigma-
cardiac injury, pressure overload, prolonged ischemia, or chronic Aldrich and polyclonal rabbit antibody from Abcam (MA, USA).
AngII treatment, myofibroblasts can originate from various non- The same antibodies were used to stain paraformaldehyde-fixed
fibroblastic sources such as (1) coronary endothelium in which adult mouse ventricular sections. The difference in the specificity
endothelial cells after cardiac damage undergo endothelial-to- of the two antibodies is obvious and suggests nonspecific staining
mesenchymal transition and migrate from the microvascular bed of fibroblasts by the antibody from Sigma-Aldrich. The extent to
into interstitium to become myofibroblasts,19 (2) epicardial epi- which cultured fibroblasts and myofibroblasts faithfully represent
thelium where epicardium-derived cells formed by an epithelial- the phenotype and function of their in vivo counterparts from
to-mesenchymal transition differentiate (potentially via Notch healthy or diseased hearts remains unknown and certainly war-
activation) into fibroblasts and myofibroblasts that remain to rants further study.
reside in the epicardium,20,21 and (3) circulating bone marrow-
derived cells, such as fibrocytes, that express both markers of
hematopoietic origin (CD45, CD13, CD34) and ECM proteins
(collagen I and III) or monocytes that upon recruitment to the
Fibroblast and Myofibroblast
site of injury or inflammation express both myofibroblast (FSP1, Electrophysiology
α-SMA) and monocytic (CD45, CD11b, CD14) markers.5 In
several studies, these non–fibroblast-derived myofibroblasts are Fibroblast Voltage-sensitive Channels
reported to comprise a significant fraction (20% to 75%) of all
myofibroblasts found in the fibrotic areas.7,19 Although some of Cardiac fibroblasts and myofibroblasts lack the required ion
these cells might not persist in the heart at later stages of the channels to initiate an action potential (AP) and are thus consid-
disease,22 their exact role in the initial and late adaptive and mal- ered unexcitable cells. They exhibit a relatively depolarized
adaptive fibrotic sequelae remains to be explored. resting membrane potential (RMP) of –50 to –20 mV, a cell
The mechanisms of fibroblast-to-myofibroblast conversion capacitance from approximately 6 pF for fibroblasts to 60 pF for
have been extensively studied in vitro despite the fact that, with myofibroblasts, and an input resistance at rest of 1 to 10 GΩ.25-27
time in culture, cardiac fibroblasts spontaneously attain a myofi- Figure 30-2, A1, shows typical current traces elicited in cultured
broblast phenotype and significantly upregulate α-SMA expres- human ventricular fibroblasts when membrane voltage in these
sion.23 This spontaneous phenotypic change does not occur in the cells was initially held at –40mV and increased by 10-mV incre-
homeostatic milieu of the healthy heart, but is rather an artifact of ments from –80 to 30 mV. Lowering the initial holding potential
Vh = −40 mV (BKCa) as well as a delayed outward rectifier K+ current (IK), which

is mediated by Kv1.6 channels in rat cells and by Kv1.5 and Kv1.6
in human cells. Similar slow and rapid delayed rectifier currents
(IKr and IKs) have been recorded in neonatal rat ventricular fibro-
blasts and are presumably conducted via Kv1.2, Kv1.4, Kv1.5,
and Kv2.1 channels.27 Cultured adult rat and mouse ventricular
A1 myofibroblasts express the gene encoding the Kir6.1 channel; in
mouse cells, the expression of Kir6.1 along with subunits of the
sulfonylurea receptor-2 (SUR2) channel generates a robust ATP-
100 pA
sensitive K+ current (IKATP) potentiated by pinacidil.28 Adult
Vh = −100 mV human and neonatal rat ventricular fibroblasts were also found
100 ms to express a transient outward K+ current (Ito) that in human cells
is conducted by Kv4.2 and Kv4.3 channels and in rat cells by
Kv4.2 and Kv1.4. Beside K+ currents, cultured human ventricular
fibroblasts express TTX-sensitive and TTX-resistant Na+ cur-
rents and swelling-induced Cl– current,29 whereas human atrial
fibroblasts express voltage-gated H+ currents.27 Interestingly, a
recent study has demonstrated that differentiation of human
atrial fibroblasts into myofibroblasts in cell culture is associated
A2 with the de novo expression of a fast voltage-gated Na+ current
predominantly carried by the α-subunit of the cardiac Na+
channel (Nav1.5).30 The potential relevance of this finding for
Steady state current (pA /pF)
atrial fibrotic disease and arrhythmogenesis is unknown. Endo
genous expression of a functional Na+-Ca2+ exchanger (NCX1 or
nRVF
4 NCX3 isoform) and L-type Ca2+ channel α-subunit (Cav1.2) has
nRVMF been shown to modulate the Ca2+ inflow in cultured fibroblasts
HVF and potentially contribute to regulation of myofibroblast prolif-
nHDF eration, migration, contraction, and collagen secretion.27,31,32
HEK293 2 Importantly, the majority of the ion currents described in the
NIH3T3 studies mentioned here were detected in only a fraction of all
the fibroblasts or myofibroblasts studied, thereby confirming the
−80 −60 −40 −20 20 large phenotypic and functional diversity of these cells.
Membrane
0 voltage
(mV) Fibroblast Mechanosensing and Transient
−2
Receptor Potential Channels
B Both fibroblasts and myofibroblasts have been considered mecha-

nosensitive cells, whereby their patterns of gene expression, pro-
Figure 30-2. Current-voltage characteristics of various fibroblasts. A1 and A2, Rep- liferation rate, contractile and electrical properties, and sensitivity
resentative whole-cell current traces from a human ventricular fibroblast when cell to and secretion of different soluble factors and ECM proteins
membrane voltage was held at –40 mV (A1) or –100 mV (A2) and stepped in are directly influenced by the mechanical state of their environ-
10-mV increments from −80 mV to 30 mV. B, Steady-state current-voltage (I-V)
relationship (mean ± SEM) recorded in neonatal rat ventricular fibroblasts (nRVF; n
ment.13 The phenomenon in which cell electrical properties are
= 5), neonatal rat ventricular myofibroblasts (nRVMF; adapted from Rohr et. al26), altered in response to a mechanical stimulus is called mechanoelec-
human ventricular fibroblasts (HVF; n = 4), neonatal human dermal fibroblasts tric feedback. In cardiac fibroblasts, mechanoelectric feedback is
(nHDF; n = 5), human embryonic kidney 293 fibroblasts (HEK 293; n = 3), and mouse likely mediated via Ca2+-permeable, stretch-sensitive channels of
NIH 3T3 fibroblasts (NIH3T3; n = 25). unknown molecular identity that likely belong to a family of
transient receptor potential (TRP) channels.27,33,34 TRP channels
are weakly sensitive to changes in membrane voltage and are
to –100 mV increased the amplitude of both the steady-state and regulated instead by stretch, oxidative stress, osmotic pressure,
small time-dependent current components (see Figure 30-2, A2). temperature, pH, or membrane receptor activation. The activity
Furthermore, the steady-state current-voltage (I-V) relationships of these channels sensitizes cardiac fibroblasts to physicochemical
of different types of fibroblasts and myofibroblasts exhibited changes in their environment. Transcripts of several TRP chan-
similar shapes, typical of unexcitable cells, with a moderate nels from the canonical (TRPC1,4,6), vanilloid (TRPV2,4), and
outward rectification present at higher membrane potentials (see melastatin (TRPM4,7) subfamilies were identified in human
Figure 30-2, B). In a recent study, cultured fibroblasts isolated atrial fibroblasts, and TRMP7 (but not TRPC6 or TRPV2,4)
from infarcted rat ventricles exhibited a hyperpolarized resting currents were also successfully recorded in these cells using
potential and increased outward current density compared with single-channel and whole-cell patch clamp. Similarly, adult rat
fibroblasts isolated from healthy ventricles.25 ventricular fibroblasts were reported to express transcripts of
Cardiac fibroblasts and myofibroblasts express a variety of TRPC2,3,5,6, TRPV2,4,6, and TRPM4,7 channels, and TRPV4
voltage-sensitive currents and related ion channel genes and and TRPC6 channels were shown to mediate Ca2+ entry into
proteins.27-29 For example, inward rectifier, Ba2+-sensitive K+ these cells. Furthermore, adult rat ventricular (but not human
current (IK1) controls RMP in freshly isolated adult rat ventricular atrial) fibroblasts were found to express nonselective cation cur-
fibroblasts (where it is likely mediated by Kir2.1 channels) and rents likely carried by TRPC3 and TRPC6 channels, or their
cultured commercially available (ScienCell Research Laboratory heteromers.34 In response to mechanical compression, cardiac
[CA, USA]) human ventricular fibroblasts (in which both Kir2.1 fibroblasts generate membrane potential depolarizations known
and Kir2.3 are expressed). Both rat and human adult ventricular as mechanically induced potentials that, through capacitive or poten-
fibroblasts express Ca2+-activated large conductance K+ channels tial electrotonic coupling with cardiomyocytes, can modulate
30
Impulse
Fibroblast Myofibroblast Collagen
conduction
A
B
O2
C D Blood vessel
Figure 30-3. Patterns of cardiac fibrosis. A, Healthy heart with normal impulse conduction. B, Interstitial fibrosis in which lateral deposition of collagen and myofibroblasts
mainly impedes transverse but not longitudinal impulse conduction. C, Patchy fibrosis where replacement of dead cardiomyocytes with fibrous tissue obstructs both
longitudinal and transverse impulse conduction. D, Perivascular fibrosis in which fibrous tissue around blood vessels impairs O2 supply and conduction in surrounding
myocardium.
cardiac electrical properties.35 In addition to voltage-gated and myofibroblasts respond to mechanical or oxidative stress or
TRP channels, cardiac fibroblasts and myofibroblasts express a various inflammatory and profibrotic cytokines (e.g., AngII,
variety of membrane-bound receptors and integrins that mediate TGF-β, TNF-α, IL-6, ET-1) by upregulating the secretion of
their sensitivity to different chemical and ECM-mediated stimuli ECM proteins (predominantly collagen III and I) and MMPs.16-18
and potentially support their long-range communication and The resulting degradation of old and excessive production of new
signal integration. ECM yields one or more of the following: (1) formation of a large
collagenous scar, (2) lateral separation of cardiomyocyte bundles
when collagen is deposited interstitially, (3) both longitudinal and
lateral separation of cardiomyocytes when lost myocytes are
Arrhythmogenic Effects of Myofibroblasts replaced with fibrous tissue, or (4) perivascular fibrosis (see
Figure 30-3). Based on histologic appearance, cardiac tissue
The described fibroblast-to-myofibroblast phenotype switch fibrosis can be classified as compact, interstitial, patchy, and
caused by different pathologic stimuli has a critical role in cardiac diffuse.36 Compact fibrosis is associated with healed infarct scars
remodeling, a process that involves changes in the size, structure, and can act as an anchor for macroscopic reentry. Interstitial
and function of the heart. Upon the onset of disease, cardiac fibrosis yields lateral separation of cardiomyocytes and, by selec-
remodeling initially allows the heart to adapt to changes in its tively impeding transverse conduction, can lead to the generation
environment in order to maintain normal cardiac output. of microreentrant circuits. Patchy fibrosis involves partial electri-
However, maladaptive cardiac remodeling (i.e., excessive cardiac cal insulation of groups of cardiomyocytes over relatively long
fibrosis) owing to persistent myofibroblast-related activity has distances (>1 mm) yielding the formation of tortuous conduction
been implicated as a major risk factor for cardiac death by increas- paths, increased vulnerability to block, and consequently, a highly
ing the susceptibility to cardiac mechanical dysfunction and arrhythmogenic substrate. Diffuse fibrosis is more disperse than
arrhythmias. In general, myofibroblasts can negatively affect patchy fibrosis and relatively nondisruptive to conduction;
cardiac electrical function through mechanisms related to their however, when excessive, it can also precipitate functional reentry.
altered turnover of ECM proteins (Figure 30-3), secretion of In fibrotic disease, nonconducting fibrous tissue typically creates
specific soluble factors (Figure 30-4), and potential for direct localized alterations in the source-load ratio of upstream (excited)
contact with adjacent cardiomyocytes (Figure 30-5). to downstream (unexcited) tissue areas, creating a substrate with
discontinuous or nonuniform conduction. Fibrotic insulation of
groups of cardiomyocytes can also facilitate the induction and
Extracellular Matrix–Mediated Effects on Cardiac propagation of early afterdepolarizations, such as those that arise
Electrical Activity during oxidative stress in aged myocardium.37 In this case, ectopic
activity is likely triggered in regions with an intermediate degree
The formation of a fibrotic tissue substrate because of of fibrosis, because too little fibrosis can prevent early afterdepo-
cardiac disease (e.g., chronic heart failure, hypertension, cardio- larization initiation, whereas too much fibrosis can block its
myopathy, myocardial infarction) or aging occurs when activated propagation.
In addition to the physical impediment of AP spread, fibrotic fibroblast proliferation, activation, and deposition of collagen, as
areas in the heart can also alter (via mechanoelectric feedback) well as by directly altering cardiomyocyte electrical activity (see
the electrical properties of bordering cardiomyocytes by stiffen- Figure 30-4) and survival.17,18,24 The most well-studied profibrotic
ing their attachment matrix or by altering the mechanical load cytokines in the heart are TGF-β1 and AngII. TGF-β signaling is
experienced by cardiomyocytes. In support of this concept, activated in response to cardiac injury, hypertrophy, or pressure
increased beta-myosine heavy chain (β-MHC) expression (char- overload, where it plays a central role in the generation of the
acteristic of pathologic cardiac hypertrophy in mice) was pre- profibrotic response.39,40 TGF-β in the heart is primarily expressed
dominantly detected in cardiomyocytes that bordered fibrous by fibroblasts but also by cardiomyocytes. Both cell types express
collagen-rich regions in pressure-overloaded and aged hearts.38 TGF-β receptors that represent heteromers of type I and type II
Furthermore, myofibroblast proliferation and collagen deposi- receptors. Upon TGF-β binding, the type I receptor (also known
tion can reduce the volume of cardiomyocyte extracellular space, as activin-linked kinase 5) phosphorylates Smad2 and Smad3,
increasing its resistance and propensity for rapid K+ accumulation which then bind to Smad4, and after translocation into the nucleus
during repetitive excitation, both of which are expected to activate downstream pathways, including Ras/MEK/ERK, p38,
increase the vulnerability to conduction slowing and block. and JNK. TGF-β signaling is inhibited by Smad7. Mice constitu-
Finally, excess myofibroblasts and collagen deposition in the peri- tively overexpressing TGF-β1 in cardiomyocytes show atrial but
vascular space can form a diffusion barrier to oxygen and nutri- not ventricular fibrosis, confirming the higher susceptibility of
ents and create local ischemic environments that can significantly atria to fibrotic disease. Treatment of cultured cardiac fibroblasts
alter cardiomyocyte electrophysiology (see Figure 30-3). with TGF-β induces their conversion to α-SMA+ myofibroblasts
and significantly enhances their ECM deposition through
increased expression of ECM genes, downregulation of MMPs,
Paracrine and Autocrine Effects on Cardiac and upregulation of plasminogen activator inhibitor (PAI)-1 and
Electrical Activity TIMPs.40 TGF-β can also directly alter myocyte electrical activity
by reducing inward rectifier K+ current, fast Na+ and L-type Ca2+
Various cytokines and growth factors are known to promote currents, and gap junctional coupling.24
arrhythmogenic changes in the cardiac substrate by increasing AngII is a potent vasoconstrictor and profibrotic factor abun-
dantly expressed in hearts undergoing myocardial remodeling.
Increased Ang II expression in transgenic mice by cardiac-
I Na
GJ
I to restricted upregulation of angiotensin-converting enzyme causes
atrial dilatation, focal fibrosis, and atrial fibrillation (AF).41 Ang
I Na I CaL I to II is produced in the injured heart by activated macrophages,
GJ cardiac myofibroblasts, and cardiomyocytes, where it acts through
I Kr I CaL I Ks
binding to its receptors AT1 and AT2, of which AT1 but not AT2
Ang II is expressed in cardiac fibroblasts. Binding of AngII to AT1 recep-
I CaL I K1 I KATP tor in fibroblasts stimulates their production of profibrotic factors
TGF-β Fibroblast TNF-α (TGF-β1, ET-1, FGF-2) and ECM proteins.17,18 TGF-β1 acts in
an autocrine fashion on fibroblasts (and in a paracrine fashion
Myofibroblast on surrounding cardiomyocytes) to stimulate AT1 expression
FGF-2 directly, thereby potentiating the fibrotic response. Thus, AngII
IL-1β
I VICa and TGF-β1 act synergistically as part of an integrated signaling
GJP ET-1 GJ network to drive fibrogenic responses in the heart by augmenting
I Na the effect that TGF-β has on fibroblast proliferation, differentia-
GJ tion, and ECM deposition.18,41
AngII can also directly affect cardiomyocyte electrophysiol-
ogy.24,42 In atrial myocytes, Ang II has been shown to upregulate
Figure 30-4. Soluble factors in fibrotic heart disease. Various growth factors and the expression of L-type Ca2+ channels via protein kinase C–
cytokines present in the fibrotic cardiac milieu mediate fibroblast-to-myofibroblast dependent and cyclic AMP response element-binding protein
conversion and directly alter the expression of membrane ion channels and gap (CREB)-dependent pathways, which can lead to sarcoplasmic
junctions (GJ) in cardiomyocytes. GJP, GJ phosphorylation. reticulum Ca2+ overload and triggered activity. AngII can also
Electrical coupling Mechanical coupling Fibroblast-to-myofibroblast

conversion
GJ SAC
Fibroblast Fibroblast Myofib-myocyte Myofib-myocyte
αSMA N-cad
electrical coupling mechanical coupling
αSMA
via gap junctions via N-cadherins
Myofib channel activation Myofib pulling of myocytes

& junctional current flow & myocyte SAC activation
RMP depolarization and
Myofibroblast Myofibroblast APD change in myocytes
Triggered Slowed
activity conduction
Cardiac arrhythmia
Figure 30-5. Effects of potential myofibroblast-cardiomyocyte coupling on cardiac electrophysiology. The existence of heterocellular electrical coupling via gap junctions
or mechanical coupling via adherens junctions could yield cardiomyocyte depolarization and APD alterations that could create arrhythmogenic conduction slowing, trig-
gered activity, or both.
enhance nuclear factor κB binding to the promoter for SCN5A TNF-α exerts multiple proarrhythmic changes in cardiomyocyte
30
(the gene encoding Nav1.5), which decreases the expression of electrical properties, including reduction of transient outward,
fast Na+ channels and reduces cardiomyocyte excitability. Binding delayed rectifier, and ATP-sensitive K+ currents, L-type Ca2+
of AngII to AT1 receptors in cardiomyocytes also leads to the current,24 and an increase in sarcoplasmic reticulum Ca2+ leak.49
formation and eventual internalization of a complex that includes Upon cardiac injury, interleukin-1β (IL-1β) is produced by
Kv4.3 channel, thus reducing the transient outward K+ current various cells including cardiac fibroblasts. IL-1β stimulates fibro-
(Ito) and potentially prolonging the cardiac action potential dura- blast migration and net ECM degradation through increased
tion (APD). Similarly, by binding to AT1 receptors and activating secretion of MMPs.18 IL-1β has been shown to downregulate
the protein kinase C–dependent pathway, AngII reduces the Cx43 expression in cardiomyocyte cultures and potentially in
delayed rectifier IKr/hERG current, which in turn can contribute canine cardiomyocytes, fibroblasts, and myofibroblasts within the
to APD prolongation and increased arrhythmogenesis in cardiac infarct border zone.50 Reduced Cx43 expression could in turn
hypertrophy and failure. On the other hand, AT1 stimulation by promote cardiac fibrosis via enhanced myofibroblast activity and
AngII enhances IKs in atrial cardiomyocytes, which promotes proliferation51,52 and result in the generation of an arrhythmo-
APD shortening and explains the basis for AngII-mediated genic substrate.
increase in AF vulnerability. In addition to its effects on cardiac In addition to studies describing the role of individual growth
voltage-gated ion channels, AngII can also stimulate Ca2+ influx factors or cytokines in cardiac hypertrophy, electrophysiology,
via activation of TRP3 and TRP6 channels leading to cardiac and fibrogenesis, several in vitro studies have examined how
hypertrophy. Lastly, AngII has been shown to mediate the reduc- fibroblast-conditioned media affect the function of neonatal rat
tion of connexin43 in ventricular myocytes via c-Src tyrosine ventricular myocytes (NRVMs). Although most of these studies
kinase upregulation, which could lead to increased risk of ven- show that fibroblast paracrine factors induce hypertrophy and
tricular fibrillation and death.43 enhance contractile activity of NRVMs, a study by Pedrotty
PDGF is another molecule that plays a significant role in et al53 also described the profound effects that neonatal rat ven-
cardiac fibrotic disease by stimulating fibroblast proliferation and tricular fibroblast conditioned media had on NRVM electrical
differentiation.17,44 Recently, upregulated expression of PDGF-A properties, including increased spontaneous activity, significantly
and -D, and PDGF receptors -α and -β by cardiac interstitial depolarized RMP, twofold conduction slowing, and twofold APD
cells including myofibroblasts was suggested as having a role in prolongation. These functional changes were accompanied by a
infarct scar formation and contributing interstitial fibrosis during significant reduction in the expression of Nav1.5, Kir2.1, and
the later stages of remodeling.45 Activation of PDGF receptors Kv4.3 genes, an increased β-MHC/α-MHC expression ratio, and
initiates signaling via mitogen-activated protein kinase, JAK/ no change in Cx43 or Cx45 gene expression. Interestingly,
STAT, and phospholipase C pathways, leading to fibroblast these deleterious effects were prevented if media conditioning
hyperresponsiveness, specifically in the atria. The regional differ- by fibroblasts occurred in noncontact cocultures with cardiomyo-
ence in PDGF receptor expression levels could also explain why cytes, but not in the presence of cardiac conditioned media.
the atria are more susceptible to fibrotic remodeling than the This result suggests the existence of local paracrine cross-talk
ventricles.12 PDGF can also act to promote fibrosis by elevating between cardiac fibroblasts and myocytes, which in physiologic
TGF-β levels.44 conditions, when numbers of the two cell types are balanced,
FGF-2 is primarily secreted by cardiac fibroblasts and cardio- prevents the adverse paracrine effects to occur. However, if
myocytes in response to adrenergic or AngII stimulation. FGF-2 the number of fibroblasts was locally increased or the number of
is translated into a high- and low-molecular-weight isoforms cardiomyocytes was locally decreased, such as in fibrotic disease,
(Hi-FGF-2 and Lo-FGF-2), of which Hi-FGF-2 is predomi- this protective balance would be compromised and the
nantly expressed by fibroblasts. In the extracellular environment, fibroblast paracrine action would negatively affect the electrical
FGF-2 acts in a paracrine fashion to induce fetal gene program function of the surrounding cardiomyocytes and potentially lead
and promote hypertrophy of cardiomyocytes and in an autocrine arrhythmic activity. Although impulse conduction was also
fashion to promote fibroblast proliferation and secretion of other slowed in the NRVM monolayers exposed to media conditioned
hypertrophic factors, such as cardiotrophin-1 (CT-1).17,46 In addi- by adult myofibroblasts from healthy or infarcted rat hearts, the
tion to its profibrotic and prohypertrophic actions, FGF-2 was APD of the treated NRVMs was reduced,25 suggesting that
reported to promote the opening of non–voltage-gated Ca2+- fibroblast-secreted factors and their functional roles may be also
permeable channels and stimulate Cx43 phosphorylation in age-dependent.
cardiomyocytes.24 Overall, paracrine factors from fibroblasts and myofibroblasts
ET-1 is secreted by a number of cells in the heart, including are expected to affect cardiac electrical function in vivo, but
endothelial cells, cardiomyocytes, and fibroblasts. TGF-β and this finding still awaits definite confirmation from studies that
Ang II promote ET-1 production via JNK and ERK/ROS path- would selectively manipulate fibroblast secretome in the intact
ways, respectively.44 ET-1 induces myocyte hypertrophy and heart. Whether the phenotypic diversity and specific spatial dis-
stimulates fibroblast ECM production and differentiation to tribution of fibroblasts or myofibroblasts in the heart can lead to
myofibroblasts.18 ET-1 also mediates electrical remodeling paracrine-mediated generation of a heterogeneous arrhythmo-
during heart failure by downregulating myocyte expression and genic substrate is currently unknown and also warrants further
phosphorylation of gap junctional proteins Cx40 and Cx43 and studies.
by reducing Nav1.5 protein expression and Na+ channel conduc-
tance.47 Similarly, exposure of cultured cardiomyocyte monolay-
ers to ET-1 yields reduced Cx43 expression and conduction Cell Contact–Mediated Effects on Cardiac
slowing.48 Electrical Activity
TNF-α in the healthy and injured heart is secreted by a
number of cells, including cardiac fibroblasts in which TNF-α The fact that cardiac fibroblasts and myofibroblasts express
secretion is enhanced by hypoxia and stimulation by AngII, sero- various voltage-sensitive and mechanosensitive ion channels sug-
tonin, or mechanical stretch.17,18 TNF-α binds to two cell gests the intriguing possibility that these cells could modulate
receptors—TNFRI and TNFRII—both of which are expressed cardiac electrical activity by directly coupling with cardiomyo-
by cardiac fibroblasts and myocytes. TNF-α binding to its recep- cytes through functional gap junctions (see Figure 30-5). Cur-
tors stimulates fibroblast proliferation and migration, as well as rently, however, there is no direct experimental evidence (except
the expression of MMPs and proinflammatory cytokines.18 for a dye transfer study in the SA node54) that cardiac fibroblasts
or myofibroblasts can functionally couple with cardiomyocytes in excitability by effectively decreasing the excitation threshold,
the intact heart or that this coupling could have a significant role whereas further RMP depolarization would inactivate Na+ chan-
in cardiac function or malfunction. In transmission electron nels and decrease the AP upstroke and conduction velocity. Even
micrographs from SA nodes or infarcted ventricles, cardiomyo- higher levels of cardiomyocyte RMP depolarization could induce
cytes, fibroblasts, and myofibroblasts formed close membrane triggered activity, whereas extreme RMP depolarization would
appositions but still remained separated by the basement mem- render cardiomyocytes unexcitable.24,62 If the cardiomyocyte IK1
brane.50,55 However, in the rabbit SA node,55 but not in the canine were reduced (e.g., in nodal tissues, infarction, failure, Andersen-
infarct border zone,50 small gap junction–like structures were Tawil syndrome), potential electrical coupling with fibroblasts
observed between abutting fibroblasts. The strongest evidence and myofibroblasts would be more likely to induce noticeable
for cardiomyocyte-fibroblast gap junctional coupling in situ changes in cardiac RMP. The resulting conduction slowing or
comes from Camelliti et al54 who showed examples of Cx45+ triggered activity could present arrhythmogenic risks that would
immunostaining between fibroblasts and cardiomyocytes in the be further exacerbated in conditions of high oxidative stress or
SA node. They also observed the infiltration of Cx45+/vimentin+ hypokalemia.63 However, in patients with AF and increased IK1,41
cells into sheep infarcts as early as a few hours after ligation, fol- cardiomyocyte RMP depolarization owing to electrical coupling
lowed by the appearance of Cx43+/vimentin+ cells within the with myofibroblasts could be antiarrhythmic.
subsequent 6 to 12 days. Most recently, the same group reported In addition to resistive (static) loading that would act to depo-
that approximately 3% of Cx43+ labeling in rabbit atria and ven- larize cardiomyocyte RMP, coupled fibroblasts would also impose
tricles resided in cardiomyocyte-fibroblast contacts, and up to a capacitive load on cardiomyocytes by contributing additional
9% of Cx40+ labeling in the atrioventricular node was membrane area to be charged and discharged by cardiomyocytes,
heterocellular.56 thus effectively diluting cardiomyocyte channel density. This in
In contrast to the scarce evidence for the structural and func- turn would affect any time-dependent (dynamic) change in the
tional coupling of fibroblasts and cardiomyocytes in the intact cardiomyocyte membrane potential, such as AP generation. In a
heart, there is a general agreement that these two cell types can recent study, McSpadden et al62 generated micropatterned pairs
functionally couple in vitro. The gap junction proteins poten- made of a neonatal rat cardiomyocyte coupled to a variable-sized
tially involved in this heterocellular coupling vary in different Cx43-overexpressing HEK293 cell (Cx43/HEK293), used as a
reports. Rohr24 has shown strong expression of Cx43 and Cx45 generic model of an unexcitable cell (see Figure 30-2, B) able to
but not Cx40 protein in neonatal rat ventricular myofibroblasts, strongly couple with cardiomyocytes.62 They showed that capaci-
but other groups have shown no Cx43 expression in these cells.57 tive loading by coupled fibroblasts acted to significantly slow the
Weak expression of Cx45 and no expression of Cx43 or Cx40 was AP upstroke velocity in cardiomyocytes, effectively reducing
found in neonatal rat ventricular fibroblasts10; however, cultured cardiac excitability. For the strong HEK293-cardiomyocyte cou-
adult mouse fibroblasts expressed Cx40 and Cx43, but not Cx45.58 pling studied, with an increase in HEK293 size (and capacitance),
Furthermore, in recent studies, cultured myofibroblasts isolated this capacitive effect on cardiomyocytes occurred before the
from infarcted hearts showed increased levels of Cx43 expression onset of significant RMP depolarization. Accordingly, coupling
compared to those isolated from healthy hearts.25,59 Although gap of Cx43/HEK293 cells to cardiomyocyte monolayers resulted in
junctions between cultured fibroblasts/myofibroblasts and car- a fivefold decrease in conduction velocity despite only a slight
diomyocytes can form without doubt, the frequency of this for- depolarization of cardiac RMP, signifying the capacitive loading
mation and the potential that selected cellular fractions are more of cardiomyocytes as a potentially powerful modulator of cardiac
amenable to coupling than others still remain to be explored. conduction.10,62
Using a well-controlled setting of micropatterned cell pairs, Besides passive (resistive and capacitive) loading of cardio-
Pedrotty et al60 found that only 9.6% of neonatal rat ventricular myocytes, coupled myofibroblasts could also actively modulate
myocyte-fibroblast pairs showed Cx43+ staining at the approxi- cardiac AP shape through the activation of their voltage-sensitive
mately 90-µm-long cell-cell border, and that instead of intercel- or mechanosensitive ion channels. For example, during cardiac
lular gap junction formation, heterocellular contacts with AP, fibroblasts coupled to an activating myocyte would also
fibroblasts predominantly yielded internalization of Cx43 in undergo depolarization and, when sufficiently depolarized, acti-
cardiomyocytes. vate their voltage-sensitive outward currents (see Figure 30-2),
Despite variable reports on the type of connexins involved, causing membrane repolarization. This in turn would tend to
most of the studies agree that a certain degree of functional generate a flow of gap junctional current from the cardiomyocyte
electrical coupling (from very weak to relatively strong) exists (source) into the fibroblast (sink), yielding shortening of the
between cultured fibroblasts/myofibroblasts and cardiomyocytes. cardiac APD. Depending on the fibroblast RMP, magnitude and
Because cardiac fibroblasts are unexcitable and have a depolarized type of its voltage-sensitive currents, strength of coupling with
RMP, their electrical coupling with cardiomyocytes will both cardiomyocytes, and the APD of unloaded cardiomyocyte, the
depolarize cardiomyocyte RMP and impose capacitive load described passive and active loading scenarios could result in dif-
during AP upstroke and repolarization. The specific shape of the ferent degrees of shortening or lengthening of the cardiac APD.
fibroblast I-V curve (see Figure 30-2, B) relative to that of a Furthermore, fibroblast mechanosensitive channels, dynamically
cardiomyocyte (mostly governed by its IK1) and the strength (con- activated during cardiac tissue contraction, could generate inward
ductance) of their gap junctional coupling will together deter- currents in fibroblasts and in turn depolarize coupled cardiomyo-
mine the level of cardiomyocyte RMP depolarization. Different cytes or accelerate their firing rate (e.g., in SA node). Coupled
studies report highly variable effects that fibroblast/myofibroblast myofibroblasts could also support passive (detrimental) impulse
coupling with cardiomyocytes has on cardiac RMP, starting from conduction over a distance of a few hundred micrometers and
a negligible change10,61 to strong depolarization.24 In the healthy electrotonically bridge remote cardiomyocytes.24 Whether this
adult myocardium with strong IK1 expression and large bridging effect would be proarrhythmic or antiarrhythmic would
cardiomyocyte size, one would generally expect that coupling to likely depend on the specific distribution and density of such
a fibroblast would only negligibly change the myocyte bridges in the cardiac tissue. At least in the NRVM-myofibroblast
resting potential. Thus, relatively strong coupling with multiple monolayers, increased myofibroblast density uniformly decreased
fibroblasts or a significant increase in fibroblast/myofibroblast conduction velocity and increased complexity of the reentrant
inward current without a change in resting potential would be activity while significantly reducing its dominant frequency.
necessary to induce significant cardiomyocyte depolarization. Interestingly, for a particular myofibroblast density studied, both
Slight depolarization of cardiomyocyte RMP can increase cardiac an increase and decrease in cardiomyocyte-myofibroblast
coupling strength increased the conduction velocity to a similar used in the AF ablation, can form only if cardiomyocytes are
30
degree.64 electrically coupled with myofibroblasts and, unlike in previous
In addition to electrical coupling, potential mechanical cou- computational studies,70 could not form because of collagen accu-
pling between myofibroblasts and cardiomyocytes has been pro- mulation alone. Furthermore, myofibroblast-myocyte coupling
posed recently as another mechanism that could contribute to the facilitated AF termination by complex fractionated atrial
increased arrhythmogenicity observed in cardiac fibrotic disease electrogram–targeted ablation, whereas the ablation in the
(see Figure 30-5). In this scenario, N-cadherin–mediated absence of this coupling only converted AF to sustained atrial
“tugging” of cardiomyocyte membranes by myofibroblasts would tachycardia.71 Although potentially important, these studies
activate mechanosensitive channels in myocytes, yielding RMP belong to a realm of mere academic exercise, pending proof that
depolarization and conduction slowing.57 Consistent with this the functional fibroblast-myocyte coupling in the heart is indeed
hypothesis, acute application of blebbistatin (to relax cellular a reality.
prestress and prevent tugging) as well as gadolinium and strep-
tomycin (to block stretch-sensitive channels nonspecifically) fully
reverted the observed conduction slowing. Silencing of Cx43 in
myofibroblasts did not alter these findings, suggesting the domi- Cardiac Fibroblasts as Antiarrhythmic Targets
nant involvement of mechanical rather than electrical coupling
in these observations, although potential roles of Cx45 coupling Recent clinical trials have identified myocardial fibrosis as an
were not addressed.57 Although direct proof for this mechanism independent predictor of cardiac arrhythmias, including ven-
by silencing N-cadherin expression in myofibroblasts was not tricular tachycardia and fibrillation.72 Given the important roles
provided, this study opened the possibility that, in the setting of of cardiac fibrosis in arrhythmogenic cardiac remodeling, the
fibrotic disease, mechanical contacts between myofibroblasts and attenuation or reversal of fibrogenic processes in the heart by
cardiomyocytes might significantly alter cardiomyocyte electro- selective pharmacologic or gene targeting could lead to effective
physiology. Naturally, the critical question remains to be deter- antiarrhythmic therapies (Table 30-2). Because cardiac fibroblasts
mined whether myofibroblasts in the heart couple to and myofibroblasts are believed to be the main mediators of
cardiomyocytes by N-cadherin junctions and whether the extracellular collagen accumulation in myocardial fibrosis as well
strength of this coupling and mechanical properties of the sur- as the generators of various profibrotic and inflammatory cyto-
rounding three-dimensional microenvironment, which is sub- kines, potential antifibrotic therapies for cardiac arrhythmias
stantially softer than the rigid substrate used for in vitro studies, would likely target fibroblast proliferation, differentiation, ECM
are conducive to the hypothesized tugging effect. Furthermore, accumulation, and response to or secretion of different soluble
studies of heterocellular contact-mediated interactions are factors. These therapies should be balanced and timely to allevi-
unavoidably confounded by the potential local paracrine or jux- ate pathologic cardiac remodeling but not interfere with the
tacrine signaling by which interacting cells might also affect each initial compensatory response to cardiac injury.73
other.
Other than potential functional coupling through gap or
adherens junctions, recent in vitro studies have suggested the
possibility that cardiomyocytes and fibroblasts can directly com-
municate through thin, long (up to ~20-30 micrometers) tubular Table 30-2. Examples of Antifibrotic Drugs
structures known as tunneling nanotubes.65 It is unknown whether
these structures are stable (or transiently formed secondary to cell Drug Action
motion) and how frequently they occur in vitro or in vivo. c-Src inhibitor Downregulation of RAAS by AngII
Regarding that the conductance of the tunneling nanotubes is in inhibition
the sub-nanosiemens (sub-nS) range66 and that velocity of Ca2+
propagation through these structures appears to be approxi- Pirfenidone Reduced TGF-β signaling by block of
mately 1 µm/s,65 their relevance for the cardiac electrophysiology Smad nuclear translocation
and arrhythmogenesis (assuming their presence in vivo) remains Relaxin Reduced TGF-β1 signaling and
to be established. Nonetheless, the possibility that heterocellular myofibroblast differentiation, enhanced
exchange (even if transient) of genetic material, organelles, and MMP-based collagen breakdown
other molecules can occur in the cardiac milieu could offer inter-
esting opportunities for the development of new cardiac Statins Downregulation of CTGF by RhoA
therapies.66 inhibition, ROS inhibition, reduced
Based on the in vitro evidence for functional electrical cou- MMP expression
pling between cardiomyocytes and myofibroblasts, a number of Omega-3 fatty acids Reduced TGF-β signaling by block of
recent computational studies have examined the potential effects Smad2/3 nuclear translocation via PKG
of fibroblast number, size, passive and active membrane proper- activation
ties, and spatial distribution on cardiomyocyte AP shape and
Geranylgeranylacetone Heat shock protein induction, TGF-β1
conduction. These studies suggest that heterogeneous spatial dis-
inhibition
tribution of fibroblast-cardiomyocyte electrical coupling in the
heart could yield proarrhythmic outcomes, including local con- Acetaminophen ROS inhibition
duction slowing and block, triggered activity, or spatially discor- β-blockers Reduced fibroblast proliferation and
dant Ca2+ alternans.67,68 Furthermore, in a realistic computational IL-6 secretion
model of infarcted rabbit ventricle, moderate densities of scar
myofibroblasts coupled to cardiomyocytes were found to enhance Losartan AngII receptor inhibition
arrhythmogenicity by augmenting APD dispersion in the periin- CTGF, Connective tissue growth factor; IL, interleukin; MMP, matrix
farct zone, while high myofibroblast densities significantly depo- metalloproteinase; PKG, cyclic guanosine monophosphate-dependent protein
larized myocytes, interrupted periinfarct zone conduction and kinase; RAAS, renin–angiotensin–aldosterone system; RhoA, Ras homolog gene
prevented arrhythmia induction.69 Recently, computational family, member A; ROS, reactive oxygen species; SAP, stress-activated protein
studies in a simplified model of fibrotic atrial tissue suggested that kinase; TGF, transforming growth factor.
complex fractionated atrial electrograms, often observed and
Antifibrotic Drug Therapies cardiac hypertrophy. Silencing miR-21 in failing hearts could
prove to be a potentially powerful antifibrotic therapy. Similarly,
As previously described, upregulation of cardiac or systemic miR-29 family members (miR-29a, miR-29b, and miR-29c) were
renin-angiotensin-aldosterone-system (RAAS) plays a critical found to be downregulated in the border zone of murine and
role in different aspects of cardiac fibrogenesis. Furthermore, human infarcts, and upregulating miR-29 expression decreased
several animal models and retrospective analyses of clinical data the collagen secretion of fibroblasts via suppression of TGF-β
have suggested that the downregulation of RAAS through the signaling. Recently, the expression of miR-101a and miR-101b
inhibition of AngII converting enzyme (e.g., lisinopril, imi- was found to be decreased in the periinfarct area of the rat heart,
daprilat, spironolactone), AngII receptors (e.g., eplerenone, and adenoviral overexpressing of miR-101a in a model of chronic
losartan), or aldosterone (e.g., telmisartan, candesartan) is a infarction decreased interstitial fibrosis via inhibition of c-Fos
promising strategy to reduce the incidence or persistence of atrial expression and TGF-β1 signaling.83
and ventricular arrhythmias. However, recent prospective ran- In addition to targeting the profibrotic function of fibroblasts,
domized clinical trials have been inconclusive; therefore, the recent studies in mice revealed the exciting possibility that viral
potential utility of RAAS inhibition for antiarrhythmic therapy delivery of cardiac transcription factors or miRs in the heart can
remains unproven.41 Because TGF-β signaling is also a major reprogram resident cardiac fibroblasts into functional cardiomyo-
mediator of cardiac fibrosis, its pharmacologic inhibition might cytes.8,84 Provided that the reprogramming efficiency is opti-
exert antifibrotic and potentially antiarrhythmic actions in the mized and that similar strategies can be applied to humans, this
heart. For example, pirfenidone, a blocker of TGF-β–induced approach could open the door to a new array of powerful gene
Smad nuclear translocation, reduced atrial and left ventricular therapies for heart disease. Similarly, successful genetic targeting
fibrosis and associated arrhythmias in models of canine heart of endogenous fibroblasts and myofibroblasts to alter their ion
failure and rat myocardial infarction.74 Similarly, relaxin, an currents and enhance or inhibit their coupling to surrounding
inhibitor of TGF-β effects, has been shown to reduce collagen cardiomyocytes could foster the development of new antiarrhyth-
content in murine models of fibrotic cardiomyopathy.75 mic therapies.85
Through their inhibition of fibroblast proliferation, migra-
tion, differentiation, and ECM turnover,18 cholesterol-lowering
statins (e.g., atorvastatin, pravastatin) exhibit strong antifibrotic
effects and could aid in the clinical prevention of AF.76 Similarly, Conclusions and Future Directions
the antiinflammatory effects of omega-3 fatty acids could also
translate into an antifibrotic and AF-suppressing therapy.77 Although cardiac fibroblasts have long been regarded as a uniform
Myeloperoxidase could represent yet another antiarrhythmic and static cell population, recent studies have revealed that these
target because it has been shown to stimulate profibrotic signal- cells are functionally and developmentally diverse and are actively
ing cascade via generation of reactive oxygen species, leading to involved in a number of processes that are essential to cardiac
atrial fibrosis and fibrillation.78 In addition, ET-1 receptor block- form and function. For this reason, an in-depth understanding of
ers (e.g., bosentan) aimed at the reduction of ECM production cardiac fibroblast origin, differentiation, and complex roles that
and myofibroblast differentiation could decrease the incidence of they play in cardiac physiology and pathology will be fundamen-
AF in patients with structural heart disease.44 Besides different tal to the ability to utilize these cells as a new therapeutic tool
profibrotic soluble factors, Ca2+ inflow through TRMP7 channels against arrhythmias and heart disease in general. However,
has been shown recently to regulate TGF-β–induced prolifera- current knowledge about cardiomyocyte-fibroblast interactions
tion and differentiation of human atrial fibroblasts from AF including complex bidirectional signaling through a shared extra-
patients33 and could potentially prove to be a novel and powerful cellular matrix, paracrine factors, or direct cell-cell contacts is
antiarrhythmic target. highly limited. There is an overwhelming number of important
questions remaining to be answered, such as:
1. Fibroblast origin and diversity in the heart—are there
Antifibrotic Gene Therapies multiple progenitors involved in the generation of cardiac
fibroblasts? How does the phenotype and function of
Genetic modifications of fibroblast function represent another fibroblasts changes with development and aging? Are there
promising approach for the treatment of fibrotic heart disease phenotypic and functional differences among fibroblasts
and related arrhythmias. Initial proof-of-concept studies have from different regions of the heart (left vs. right vs. septum,
successfully targeted the downregulation of TGF-β/AngII signal- base vs. apex, endocardial vs. epicardial)?
ing in infarcted or pressure-overloaded rodent hearts. Recently, 2. Fibroblast roles in cardiac pathology—do fibroblasts change
adeno-associated viral delivery of the small proteoglycan decorin their phenotype with cardiac disease? What are the
prevented cardiac fibrosis by reducing TGF-β/Smad2/3 activa- molecular differences between fibroblasts and myofibroblasts
tion and p38 MAPK signaling in hypertensive rat hearts.79 Simi- in situ, including the repertoire of their ion channels,
larly, in vitro knockdown of myocardia-related transcription coupling proteins, and paracrine factors? What is the role of
factor A and B in cardiac fibroblasts reduced their TGF-β1– resident fibroblasts versus myofibroblasts versus nonresident
induced transition to myofibroblasts, offering a potential gene (blood or endothelium-derived) myofibroblasts in cardiac
therapy approach for cardiac fibrotic disease.80 Furthermore, disease? Are myofibroblasts of different origins involved in
overexpression of Fc receptors in monocytes could represent distinct types of fibrosis (e.g., perivascular, patchy,
another strategy to alleviate fibrotic cardiomyopathy by prevent- interstitial)? When and why does the initially protective
ing serum amyloid P–mediated differentiation of monocytic cells action of myofibroblasts then become deleterious?
to fibroblasts.81 3. Fibroblast-cardiomyocyte relationships—how do multiple
Recently, small noncoding RNAs (microRNAs [miRs]) have paracrine factors interact in a bidirectional manner to
emerged as potentially important targets in the treatment of coordinate fibroblast and myocyte function in the heart? Do
cardiac fibrosis and arrhythmias.82 For example, miR-21 expres- potential differences in the fibroblast phenotype in different
sion in cardiac fibroblasts was found to be selectively upregulated regions of the heart correlate with regional differences in
in failing murine hearts where it enhanced ERK-MAP kinase cardiomyocyte electrical phenotype? Can the type and
activity and, by controlling fibroblast survival and cytokine secre- amount of secreted ECM directly change the electrical
tion, contributed to the formation of interstitial fibrosis and properties of the attached myocytes? Do fibroblasts and
myofibroblasts electrically or mechanically couple to understanding of how fibroblasts directly and indirectly interact
30
myocytes in the heart, and what are the consequences of this with cardiomyocytes and other cells in the heart is likely to result
coupling for cardiac electrical function? in the development of new non–cardiomyocyte-targeted drug,
The lack of suitable fibroblast markers and more efficient gene, and cell therapies for cardiac arrhythmias.
lineage tracing tools are probably the most important reasons
that attempts to understand the molecular and cellular events
involved in fibroblast function or malfunction continue to be
challenging for investigators. The inability to directly access or Acknowledgments
modulate fibroblast-cardiomyocyte interactions in situ is another
limiting factor. Nevertheless, rapid advances in genetic fate The authors thank Dr. Robert Kirkton for critical reading of the
mapping techniques (including the discovery of new markers manuscript and for generating Figure 30-1, and Hung Nguyen
of cardiac fibroblasts), in situ imaging technologies, and for generating Figure 30-2. This work is supported in part by
sophisticated tissue engineering and cell culture methods are National Institutes of Health–Nation Heart, Lung, and Blood
warranted to facilitate future progress in this field. An enhanced Institute grants HL106203, HL104326, and HL093711 to N.B.
18. Porter KE, Turner NA: Cardiac fibroblasts: At the current in acutely isolated rat cardiac fibroblasts
References heart of myocardial remodeling. Pharmacol Ther via natriuretic peptide c receptor-mediated signal-
123:255–278, 2009. ling. J Physiol 580:255–274, 2007.
1. Souders CA, Bowers SL, Baudino TA: Cardiac 19. Zeisberg EM, Tarnavski O, Zeisberg M, et al: 35. Kamkin A, Kirischuk S, Kiseleva I: Single
fibroblast: The renaissance cell. Circ Res 105:1164– Endothelial-to-mesenchymal transition contrib- mechano-gated channels activated by mechanical
1176, 2009. utes to cardiac fibrosis. Nat Med 13:952–961, deformation of acutely isolated cardiac fibroblasts
2. Snider P, Standley KN, Wang J, et al: Origin of 2007. from rats. Acta Physiol (Oxf) 199:277–292,
cardiac fibroblasts and the role of periostin. Circ 20. Zhou B, Honor LB, He H, et al: Adult mouse 2010.
Res 105:934–947, 2009. epicardium modulates myocardial injury by secret- 36. de Jong S, van Veen TA, van Rijen HV, et al: Fibro-
3. Biernacka A, Frangogiannis NG: Aging and cardiac ing paracrine factors. J Clin Invest 121:1894–1904, sis and cardiac arrhythmias. J Cardiovasc Pharma-
fibrosis. Aging and Disease 2:158–173, 2011. 2011. col 57:630–638, 2011.
4. Ieda M, Tsuchihashi T, Ivey KN, et al: Cardiac 21. Russell JL, Goetsch SC, Gaiano NR, et al: A 37. Karagueuzian HS: Targeting cardiac fibrosis: A
fibroblasts regulate myocardial proliferation dynamic notch injury response activates epicar- new frontier in antiarrhythmic therapy? Am J Car-
through beta1 integrin signaling. Dev Cell 16:233– dium and contributes to fibrosis repair. Circ Res diovasc Dis 1:101–109, 2011.
244, 2009. 108:51–59, 2011. 38. Pandya K, Kim H-S, Smithies O: Fibrosis, not cell
5. Zeisberg EM, Kalluri R: Origins of cardiac fibro- 22. van Amerongen MJ, Bou-Gharios G, Popa E, et al: size, delineates beta-myosin heavy chain reexpres-
blasts. Circ Res 107:1304–1312, 2010. Bone marrow-derived myofibroblasts contribute sion during cardiac hypertrophy and normal aging
6. Tian Y, Morrisey EE: Importance of myocyte- functionally to scar formation after myocardial in vivo. Proc Natl Acad Sci U S A 45:16864–16869,
nonmyocyte interactions in cardiac development infarction. J Pathol 214:377–386, 2008. 2006.
and disease. Circ Res 110:1023–1034, 2012. 23. Santiago JJ, Dangerfield AL, Rattan SG, et al: 39. Teekakirikul P, Eminaga S, Toka O, et al: Cardiac
7. Krenning G, Zeisberg EM, Kalluri R: The origin Cardiac fibroblast to myofibroblast differentiation fibrosis in mice with hypertrophic cardiomyopathy
of fibroblasts and mechanism of cardiac fibrosis. in vivo and in vitro: Expression of focal adhesion is mediated by non-myocyte proliferation and
J Cell Physiol 225:631–637, 2010. components in neonatal and adult rat ventricular requires tgf-beta. J Clin Invest 120:3520–3529,
8. Song K, Nam YJ, Luo X, et al: Heart repair by myofibroblasts. Dev Dyn 239:1573–1584, 2010. 2010.
reprogramming non-myocytes with cardiac tran- 24. Rohr S: Arrhythmogenic implications of fibroblast- 40. Bujak M, Frangogiannis NG: The role of tgf-beta
scription factors. Nature 485:599–604, 2012. myocyte interactions. Circ Arrhythm Electro- signaling in myocardial infarction and cardiac
9. Acharya A, Baek ST, Huang G, et al: The bhlh physiol 5:442–452, 2012. remodeling. Cardiovasc Res 74:184–195, 2007.
transcription factor tcf21 is required for lineage- 25. Vasquez C, Mohandas P, Louie KL, et al: Enhanced 41. Wakili R, Voigt N, Kaab S, et al: Recent advances
specific emt of cardiac fibroblast progenitors. fibroblast-myocyte interactions in response to in the molecular pathophysiology of atrial fibrilla-
Development 139:2139–2149, 2012. cardiac injury. Circ Res 107:1011–1020, 2010. tion. J Clin Invest 121:2955–2968, 2011.
10. McSpadden LC, Kirkton RD, Bursac N: Electro- 26. Rosker C, Salvarani N, Schmutz S, et al: Abolish- 42. Goette A, Lendeckel U: Electrophysiological
tonic loading of anisotropic cardiac monolayers by ing myofibroblast arrhythmogeneicity by pharma- effects of angiotensin ii. Part i: Signal transduction
unexcitable cells depends on connexin type and cological ablation of alpha-smooth muscle actin and basic electrophysiological mechanisms. Euro-
expression level. Am J Physiol Cell Physiol containing stress fibers. Circ Res 109:1120–1131, pace 10:238–241, 2008.
297:C339–351, 2009. 2011. 43. Sovari AA, Iravanian S, Dolmatova E, et al: Inhibi-
11. Chong JJ, Chandrakanthan V, Xaymardan M, et al: 27. Yue L, Xie J, Nattel S: Molecular determinants of tion of c-src tyrosine kinase prevents angiotensin
Adult cardiac-resident msc-like stem cells with a cardiac fibroblast electrical function and therapeu- ii-mediated connexin-43 remodeling and sudden
proepicardial origin. Cell Stem Cell 9:527–540, tic implications for atrial fibrillation. Cardiovasc cardiac death. J Am Coll Cardiol 58:2332–2339,
2011. Res 89:744–753, 2011. 2011.
12. Burstein B, Libby E, Calderone A, et al: Differen- 28. Vasquez C, Benamer N, Morley GE: The cardiac 44. Leask A: Potential therapeutic targets for cardiac
tial behaviors of atrial versus ventricular fibroblasts: fibroblast: Functional and electrophysiological fibrosis: Tgfbeta, angiotensin, endothelin, ccn2,
A potential role for platelet-derived growth factor considerations in healthy and diseased hearts. and pdgf, partners in fibroblast activation. Circ Res
in atrial-ventricular remodeling differences. Circu- J Cardiovasc Pharmacol 57:380–388, 2011. 106:1675–1680, 2010.
lation 117:1630–1641, 2008. 29. Li GR, Sun HY, Chen JB, et al: Characterization 45. Zhao W, Zhao T, Huang V, et al: Platelet-derived
13. Hinz B: The myofibroblast: Paradigm for a of multiple ion channels in cultured human cardiac growth factor involvement in myocardial remodel-
mechanically active cell. J Biomech 43:146–155, fibroblasts. PloS One 4:e7307, 2009. ing following infarction. J Mol Cell Cardiol
2010. 30. Chatelier A, Mercier A, Tremblier B, et al: A 51:830–838, 2011.
14. Roy S, Khanna S, Rink T, et al: P21waf1/cip1/sdi1 distinct de-novo expression of nav1.5 sodium 46. Kakkar R, Lee RT: Intramyocardial fibroblast
as a central regulator of inducible smooth muscle channels in human atrial fibroblasts differentiated myocyte communication. Circ Res 106:47–57,
actin expression and differentiation of cardiac into myofibroblasts. J Physiol 2012. 2010.
fibroblasts to myofibroblasts. Mol Biol Cell 31. Chen JB, Tao R, Sun HY, et al: Multiple ca2+ 47. Mueller EE, Momen A, Masse S, et al: Electrical
18:4837–4846, 2007. signaling pathways regulate intracellular ca2+ remodelling precedes heart failure in an endothelin-
15. Wang H, Haeger SM, Kloxin AM, et al: Redirect- activity in human cardiac fibroblasts. J Cell Physiol 1-induced model of cardiomyopathy. Cardiovasc
ing valvular myofibroblasts into dormant fibro- 223:68–75, 2010. Res 89:623–633, 2011.
blasts through light-mediated reduction in 32. Follonier Castella L, Gabbiani G, McCulloch CA, 48. Reisner Y, Meiry G, Zeevi-Levin N, et al: Impulse
substrate modulus. PloS One 7:e39969, 2012. et al: Regulation of myofibroblast activities: conduction and gap junctional remodelling by
16. van den Borne SW, Diez J, Blankesteijn WM, Calcium pulls some strings behind the scene. Exp endothelin-1 in cultured neonatal rat ventricular
et al: Myocardial remodeling after infarction: The Cell Res 316:2390–2401, 2010. myocytes. J Cell Mol Med 13:562–573, 2009.
role of myofibroblasts. Nat Rev Cardiol 7:30–37, 33. Du J, Xie J, Zhang Z, et al: Trpm7-mediated ca2+ 49. Duncan DJ, Yang Z, Hopkins PM, et al: Tnf-alpha
2010. signals confer fibrogenesis in human atrial fibrilla- and il-1beta increase ca2+ leak from the sarcoplas-
17. Takeda N, Manabe I: Cellular interplay between tion. Circ Res 106:992–1003, 2010. mic reticulum and susceptibility to arrhythmia in
cardiomyocytes and nonmyocytes in cardiac 34. Rose RA, Hatano N, Ohya S, et al: C-type natri- rat ventricular myocytes. Cell Calcium 47:378–
remodeling. Int J Inflam 2011:535241, 2011. uretic peptide activates a non-selective cation 386, 2010.
50. Baum JR, Long B, Cabo C, et al: Myofibroblasts rabbit ventricular myocytes and myofibroblasts. pressure-overloaded heart. Am J Physiol Heart
cause heterogeneous cx43 reduction and are J Physiol 583:225–236, 2007. Circ Physiol 298:H424–H432, 2010.
unlikely to be coupled to myocytes in the healing 62. McSpadden LC, Nguyen H, Bursac N: Size and 74. Nguyen DT, Ding C, Wilson E, et al: Pirfenidone
canine infarct. Am J Physiol Heart Circ Physiol ionic currents of unexcitable cells coupled to car- mitigates left ventricular fibrosis and dysfunction
302:H790–H800, 2012. diomyocytes distinctly modulate cardiac action after myocardial infarction and reduces arrhyth-
51. Jansen JA, van Veen TA, de Jong S, et al: Reduced potential shape and pacemaking activity in mias. Heart Rhythm 7:1438–1445, 2010.
cx43 expression triggers increased fibrosis due to micropatterned cell pairs. Circ Arrhythm Electro- 75. Du XJ, Xu Q, Lekgabe E, et al: Reversal of cardiac
enhanced fibroblast activity. Circ Arrhythm Elec- physiol 2012. fibrosis and related dysfunction by relaxin. Ann
trophysiol 5:380–390, 2012. 63. Nguyen TP, Xie Y, Garfinkel A, et al: Arrhythmo- N Y Acad Sci 1160:278–284, 2009.
52. Zhang Y, Kanter EM, Laing JG, et al: Connexin43 genic consequences of myofibroblast-myocyte cou- 76. Wang Z, Zhang Y, Gao M, et al: Statin therapy for
expression levels influence intercellular coupling pling. Cardiovasc Res 93:242–251, 2012. the prevention of atrial fibrillation: A meta-analysis
and cell proliferation of native murine cardiac 64. Zlochiver S, Munoz V, Vikstrom KL, et al: Elec- of randomized controlled trials. Pharmacotherapy
fibroblasts. Cell Commun Adhes 15:289–303, trotonic myofibroblast-to-myocyte coupling 31:1051–1062, 2011.
2008. increases propensity to reentrant arrhythmias in 77. Chen J, Shearer GC, Chen Q, et al: Omega-3 fatty
53. Pedrotty DM, Klinger RY, Kirkton RD, et al: two-dimensional cardiac monolayers. Biophys J acids prevent pressure overload-induced cardiac
Cardiac fibroblast paracrine factors alter impulse 95:4469–4480, 2008. fibrosis through activation of cyclic gmp/protein
conduction and ion channel expression of neonatal 65. He K, Shi X, Zhang X, et al: Long-distance inter- kinase g signaling in cardiac fibroblasts. Circula-
rat cardiomyocytes. Cardiovasc Res 83:688–697, cellular connectivity between cardiomyocytes and tion 123:584–593, 2011.
2009. cardiofibroblasts mediated by membrane nano- 78. Friedrichs K, Baldus S, Klinke A: Fibrosis in atrial
54. Camelliti P, Green CR, LeGrice I, et al: Fibroblast tubes. Cardiovasc Res 92:39–47, 2011. fibrillation—role of reactive species and mpo.
network in rabbit sinoatrial node: Structural and 66. Wang X, Gerdes HH: Long-distance electrical Frontiers in Physiology 3:214, 2012.
functional identification of homogeneous and het- coupling via tunneling nanotubes. Biochim 79. Yan W, Wang P, Zhao CX, et al: Decorin gene
erogeneous cell coupling. Circ Res 94:828–835, Biophys Acta 1818:2082–2086, 2012. delivery inhibits cardiac fibrosis in spontaneously
2004. 67. Xie Y, Garfinkel A, Camelliti P, et al: Effects of hypertensive rats by modulation of transforming
55. De Maziere AM, van Ginneken AC, Wilders R, et fibroblast-myocyte coupling on cardiac conduction growth factor-beta/smad and p38 mitogen-
al: Spatial and functional relationship between and vulnerability to reentry: A computational activated protein kinase signaling pathways. Hum
myocytes and fibroblasts in the rabbit sinoatrial study. Heart Rhythm 6:1641–1649, 2009. Gene Ther 20:1190–1200, 2009.
node. J Mol Cell Cardiol 24:567–578, 1992. 68. Xie Y, Garfinkel A, Weiss JN, et al: Cardiac alter- 80. Crider BJ, Risinger GM Jr, Haaksma CJ, et al:
56. Kohl P, Camelliti P: Fibroblast-myocyte connec- nans induced by fibroblast-myocyte coupling: Myocardin-related transcription factors a and b are
tions in the heart. Heart Rhythm 9:461–464, Mechanistic insights from computational models. key regulators of tgf-beta1-induced fibroblast to
2012. Am J Physiol Heart Circ Physiol 297:H775–H784, myofibroblast differentiation. J Invest Dermatol
57. Thompson SA, Copeland CR, Reich DH, et al: 2009. 131:2378–2385, 2011.
Mechanical coupling between myofibroblasts and 69. McDowell KS, Arevalo HJ, Maleckar MM, et al: 81. Haudek SB, Trial J, Xia Y, et al: Fc receptor
cardiomyocytes slows electric conduction in Susceptibility to arrhythmia in the infarcted heart engagement mediates differentiation of cardiac
fibrotic cell monolayers. Circulation 123:2083– depends on myofibroblast density. Biophys J fibroblast precursor cells. Proc Natl Acad Sci
2093, 2011. 101:1307–1315, 2011. U S A 105:10179–10184, 2008.
58. Louault C, Benamer N, Faivre JF, et al: Implication 70. Jacquemet V, Henriquez CS: Genesis of complex 82. Zhu H, Fan GC: Role of micrornas in the reper-
of connexins 40 and 43 in functional coupling fractionated atrial electrograms in zones of slow fused myocardium towards post-infarct remodel-
between mouse cardiac fibroblasts in primary conduction: A computer model of microfibrosis. ling. Cardiovasc Res 94:284–292, 2012.
culture. Biochim Biophys Acta 1778:2097–2104, Heart Rhythm 6:803–810, 2009. 83. Pan Z, Sun X, Shan H, et al: Microrna-101 inhib-
2008. 71. Ashihara T, Haraguchi R, Nakazawa K, et al: The ited postinfarct cardiac fibrosis and improved left
59. Zhang Y, Kanter EM, Yamada KA: Remodeling of role of fibroblasts in complex fractionated electro- ventricular compliance via the fbj osteosarcoma
cardiac fibroblasts following myocardial infarction grams during persistent/permanent atrial fibrilla- oncogene/transforming growth factor-beta1
results in increased gap junction intercellular tion: Implications for electrogram-based catheter pathway. Circulation 126:840–850, 2012.
communication. Cardiovasc Pathol 19:e233–240, ablation. Circ Res 110:275–284, 2012. 84. Qian L, Huang Y, Spencer CI, et al: In vivo repro-
2010. 72. O’Hanlon R, Grasso A, Roughton M, et al: Prog- gramming of murine cardiac fibroblasts into
60. Pedrotty DM, Klinger RY, Badie N, et al: Struc- nostic significance of myocardial fibrosis in hyper- induced cardiomyocytes. Nature 485:593–598,
tural coupling of cardiomyocytes and noncardio- trophic cardiomyopathy. J Am Coll Cardiol 2012.
myocytes: Quantitative comparisons using a novel 56:867–874, 2010. 85. Kirkton RD, Bursac N: Engineering biosynthetic
micropatterned cell pair assay. Am J Physiol Heart 73. Lucas JA, Zhang Y, Li P, et al: Inhibition of trans- excitable tissues from unexcitable cells for electro-
Circ Physiol 295:H390–H400, 2008. forming growth factor-beta signaling induces left physiological and cell therapy studies. Nat
61. Chilton L, Giles WR, Smith GL: Evidence of ventricular dilation and dysfunction in the Commun 2:300, 2011.
intercellular coupling between co-cultured adult
Models of Cardiac Excitation PART V
Ionic Mechanisms of Atrial

Action Potentials 31
Sandeep V. Pandit
CHAPTER OUTLINE human ventricular cell.26 By incorporating experimentally known

ionic differences between atrial and ventricular myocytes, the
Ionic Bases of Atrial Action Potentials in the Healthy model reconstructs essential aspects of atrial E-C coupling. It
Myocardium 309 allows a systematic comparison of the main ionic currents under-
lying the human atrial action potential (Figure 31-1, in red) and
Action Potential and Ionic Remodeling in Chronic
highlights key differences with those underlying the ventricular
Atrial Fibrillation 312 action potential (Figure 31-1, in black).
Alterations in Atrial Electrophysiology During
Ventricular Dysfunction 314
Action Potential Characteristics
Ionic Basis of Reentry (Spiral Waves) in the Atrium 314
The human atrial action potential typically demonstrates a trian-
gular morphology (compared with a spike-and-dome shape with
Our knowledge regarding the ionic bases of the atrial action a prominent plateau phase of the ventricular counterpart).1,27 The
potential has evolved continuously since the early transmem- human atrial action potential duration at 90% repolarization
brane recordings made in isolated human atrial tissue almost 50 (APD90) at 1 Hz shows large variations of between 150 and 500
years ago.1,2 The advent of the patch-clamp technique in the early msec, possibly influenced by recording conditions and ionic con-
1980s3 and the subsequent detailed characterization of the bio- centrations.2,4,27-31 The atrial resting membrane potential (Vrest)
physical properties of the various ionic currents in human atrial has been found to vary between −65 and −80 mV,2,4,27-31 and is
cells4-7 led to the development of the first detailed quantitative more depolarized than in the human ventricle26,27 (also see Figure
mathematical models of the human atrial action potential in 31-1). The depolarized Vrest in atrial cells compared with ven-
1998.8,9 This was followed by increased recognition that intracel- tricular cells is mainly attributed to differences in density of the
lular Ca2+ ([Ca2+]i) homeostasis plays a major role in influencing inward rectifier K+ current, IK117,32,33 (also see later). Maximum
atrial repolarization in both normal and pathophysiological con- upstroke velocities (Vmax) for atrial action potentials have been
ditions, particularly in atrial fibrillation (AF).10-16 It resulted in experimentally reported to vary between ≈150 and 300 V/s,2,27-31
formulations of newer quantitative models in 2011 that encapsu- in contrast to higher values of 300 to 400 V/s for human ven-
late the nonlinear interactions between ionic currents and (Ca2+)i tricular cells.26,33 Characteristic human atrioventricular action
to simulate human atrial excitation-contraction (E-C) cou- potential property/shape differences are replicated in most mam-
pling.17,18 Studies have also documented regional variations in malian species,19 except for the murine myocardium (rat/mouse),
atrial electrophysiology,19 the influences of age-related changes,20 where both atrial and ventricular cells display short, triangular
and the ionic bases of atrial remodeling that occur as the result action potentials.34,35
of AF21 and/or ventricular dysfunction.22 This chapter provides
an overview of the principal ionic determinants of atrial action
potentials at the cellular level in health and disease (mainly The Inward, Depolarizing Currents: (Na+, Ca2+)
humans), their regional heterogeneities and age-related changes,
and their behavior during functional reentry caused by one or Initiation of the atrial action potential is due to rapid activation
more spiral waves (rotors) that have now been shown to sustain of the inward Na+ current, INa, which is also the principal deter-
AF in animal models23 and more recently in humans.24,25 minant of Vmax. Functional differences in human ventricular and
atrial INa biophysical properties were reported to be minimal,36-39
but recent studies have shown that the molecular correlates of INa
in the human atrial and ventricular myocardium are different.
Ionic Bases of Atrial Action Potentials in the Although Nav1.5 was found to be the main α-subunit encoding
Healthy Myocardium for atrial/ventricular INa, the transcript for the β-subunit, Navβ1,
was more prominently expressed in the human atrium.40 In addi-
We recently developed a mathematical model of human atrial cell tion, recent reports in canine atrial cells suggest that atrial INa has
electrophysiology17 based on available experimental data. The higher current density and a more negative steady-state half-
atrial formulation was derived from an existing model of the inactivation voltage value, when compared with ventricular
309
310 MODELS OF CARDIAC EXCITATION
Ventricular myocyte Atrial myocyte human atrial and ventricular cells in current density and steady-
model model state inactivation properties.44 At the transcript level, greater
expression of the α-subunits Cav1.3 and Cav3.1 has been reported
0 in the atrium, compared with the ventricle.40 Moreover, a more
(mV)
Em “negative voltage” of the triangular plateau phase causes a larger

–81.3 mV –74.3 mV driving force (Vm − ECa; where Vm represents membrane voltage,
–100 and ECa represents reversal potential for ICaL), resulting in an ICaL
A current with a larger magnitude underlying the human atrial cell
500 action potential (compared with its ventricular counterpart), as
(nmol/L)
[Ca2+]I can be seen in Figure 31-1.17 Differences in regulation of human

atrial (but not ventricular) ICaL by second messengers such as
0 serotonin45 and phosphodiesterases46 (PDEs) have been reported
B and may further influence action potential morphology.47
0
INa
(A/F)
–200
–400
C The Outward, Repolarizing K+ Currents (K+)
0
The key current that sets the Vrest (i.e., the inward rectifier K+
(A/F)
–4 ICaL
–8 current IK1) has a much smaller density (≈5-6-fold less) in human
D atrial cells compared with its ventricular counterpart,17,32 and this
partially explains the depolarized Vrest in atrial as opposed to
0.2 ventricular cells (see Figure 31-1). The smaller density of IK1 is
(A/F)
0.1 IKr also responsible in part for the slower late phase of repolarization
0 in the atrium. The main molecular correlate of IK1, Kir2.1, is
E
0.01 more robustly expressed in the human ventricle than in the
atrium.40 The human atrium also expresses the fast, depolarization-
(A/F)
IKS
activated, 4-AP (aminopyridine)-sensitive, Ca2+-independent
0 transient outward K+ current, Ito1.4,27 This current is responsible
F for the initial repolarization phase of the atrial action poten-
10
tial4,17,27,48,49 (see Figure 31-1). It is interesting to note that despite
(A/F)
5 Fast Ito having almost a 2-fold higher density in the atrium than in the
0 Slow ventricle,27 the magnitude of Ito1 underlying the atrial and ven-
G tricular action potentials is similar, likely as the result of action
5
potential morphologies (see Figure 31-1). The main molecular
(A/F)
IKur correlates of human atrial Ito1 are Kv4.3/KChiP2 (α/β-subunits,

0 respectively).40 The slow component of Ito1,slow (mainly encoded
H
2 by Kv1.4) is absent in the human atrium, but present in the
ventricle.17,26 A Ca2+-dependent transient outward K+ current, Ito2,
(A/F)
1 IK1
has been reported in human atrial cells, but both its molecular
0 correlate and its contribution to the action potential remain
I
unclear.48 Human atrial cells also express three functional delayed
0.3 rectifier K+ currents, which contribute to repolarization: the
(A/F)
INaK
0.2 ultra-rapid delayed rectifier K+ current, IKur (molecular correlate
0.1 Kv1.5)6,50 and the rapid and slow delayed rectifier K+ currents,
J
viz, IKr, and IKs (molecular correlates HERG, KvLQT1).7,51,52 IKur
0 is present only in the atrium (not ventricle), is active during the
(A/F)
INCX
–0.5 plateau phase (see Figure 31-1), and is blocked by low concentra-
–1 tions of 4-AP (100 µM).5,6,50 Its atrial-selective nature has made
0 100 200 300 400 500 0 100 200 300 400 500
K it an attractive target for many antiarrhythmic drugs in develop-
Time (ms) Time (ms)
ment for terminating and/or preventing recurrence of AF.53
Figure 31-1. Key ionic currents underlying ventricular (black) and atrial (red) action Compared with IKur, the contribution of IKr and IKs to the human
potentials. Depicted in this figure are (A) action potentials, (B) Ca2+ transients atrial action potential is small (see Figure 31-1),17 in part because
([Ca2+]i), (C) sodium current, INa, (D) calcium current, ICaL, (E) rapid delayed rectifier of their small density, and in part because of their triangular shape
current, IKr, (F) slow delayed rectifier current, IKs, (G) transient outward current, Ito,
(H) ultra-rapid delayed rectifier current, IKur, (I) inward rectifier current, IK1, (J) Na+/
and plateau phase at relatively negative membrane voltages,
K+ pump current, INaK, and (K) Na+/Ca2+ exchanger current, INCX. which preclude both IKr and IKs from activating fully. However,
considerable variability in the shape of the human atrial action
potentials has been noted, and IKr/IKs are likely to contribute
cells.41 Similar differences were reported in guinea pig atrial and more in cells showing a prominent plateau phase and dome (the
ventricular cells.42 Further, because of the more depolarized Vrest, so-called type 1 action potentials).7
the availability of atrial INa is less, which results in a smaller INa
underlying the atrial action potential, compared with the ven-
tricular counterpart, as can be seen in Figure 31-1. The L-type [Ca2+]i, Electrogenic Pumps and Exchangers
Ca2+ current (ICaL) is mainly responsible for the plateau phase of
the atrial action potential, as well as the Ca2+-induced Ca2+ release The human atrial action potential is also modulated by [Ca2+]i,
in human atrial cells.43 Patch-clamp experiments suggested that which directly influences the inactivation of ICaL54 and the mag-
the T-type Ca2+ current (ICaT) was not present in human atrial nitude and temporal profile of the electrogenic Na+/Ca2+
cells.43 Experimental studies conducted in the Lederer lab exchanger current, INCX10,55 (molecular transcript NCX140), and
reported differences in the biophysical properties of ICaL between has an indirect influence on the Na+/K+ pump current, INaK, by
Ionic Mechanisms of Atrial Action Potentials 311
influencing intracellular Na+ ion accumulation17 (molecular tran- potentials in normal sinus rhythm.17 However, large variations in
31
scripts Na/K ATPase α1, α3, β140). The Na+/Ca2+ exchanger human atrial action potential shapes (mainly triangular, some
current (INCX) is the main Ca2+ extrusion and Na+ influx pathway dome shaped) have been reported in cells isolated from the right
in cardiac myocytes. It extrudes 1 Ca+ in exchange for 3 Na+, thus atrial appendage, and this has been attributed to differences in
generating an inward current that influences cardiac repolariza- current density ratios of Ito1/IKr in these cells.7 Recently, the
tion and arrhythmogenesis.56 The Na+/K+ pump (NKA) is the carbachol-activated K+ current (IKACh) was reported to be 70%
main route of Na+ efflux in cardiac cells, thus regulating intracel- larger in RA than in LA, in sinus rhythm patients, with corre-
lular [Na+]. By extruding 3 Na+ in exchange for 2 K+, it generates spondingly higher protein expressions of its molecular correlates
an outward current that is known to influence both resting mem- (Kir3.1/Kir3.4) in RA versus LA.79 To the best of our knowledge,
brane potential and repolarization.56 We recently simulated the no report has described the ionic mechanisms underlying human
[Ca2+]i homeostasis in human atrial cells, and our results showed PV cells. Thus available data regarding the regional variations in
that whereas INaK is primarily a repolarizing outward current, INCX human atrial electrophysiology/underlying ionic mechanisms
can be both outward and inward; it contributes to repolarization under healthy conditions are very limited, and further studies are
in the early phase of the action potential, and is a negative current needed in this regard. In the next paragraph, we briefly review
(depolarizing influence) during the later phase of the action the atrial action potential heterogeneity/ionic mechanisms in
potential, when it extrudes Ca2+ ions from the atrial cells (see other animal species.
Figure 31-1).17 INaK and INCX also influence the frequency depen- In healthy hearts of dogs and mice, the APD is shorter in LA
dence of the human atrial action potential duration, which is than in RA, primarily on account of differences in current densi-
discussed in later sections. ties of the delayed rectifier K+ currents, that is, IKr in dogs and
IKur and the steady-state K+ current Iss in mice display larger
current densities in LA than in RA.70,80,81 In sheep atrial cells, the
Other Ion Channels IKACh current has been shown to have a larger density in LA than
in RA, whereas the density of IK1 was similar.82 In contrast, IKACh
Besides the currents discussed earlier, other ion channels are current density was larger in RA than in LA in mouse,83 similar
activated under specific conditions and can influence the human to humans in sinus rhythm.79 The density of INa has been reported
atrial action potential. These include the acetylcholine-activated to be larger in LA than in RA in dogs.84 Besides LA and RA,
K+ current, IKACh28,57,58 (molecular correlates Kir3.1/3.440), the differences within RA have also been reported. For example, the
adenosine triphosphate (ATP)-sensitive K+ current, IKATP59,60 action potential properties of the pectinate muscles and the crista
(molecular correlate, Kir6.1/6.2/SuR2.X40), the Ca2+-dependent terminalis in the rabbit RA are different85; such differences have
nonselective cation current61,62 (molecular correlates unknown), been attributed to differences in various ion channels (Na+, Ca2+,
and the hyperpolarization-activated funny current, If63,78,93 K+) and have been integrated into a mathematical model.86 Fur-
(molecular correlates HCN1/HCN440). Some reports indicate thermore, differences in the electrophysiological properties of
that a Ca2+-activated K+ current IKCa (putative molecular correlate cells isolated from the LA and PV regions have been extensively
SK240) may modulate the human atrial action potential64,65; studied by the Nattel group in dogs73,74 and by the Chen group
however, its presence and its contribution to repolarization in rabbits.71,72 Canine PV cells exhibit a depolarized Vrest, a smaller
remain controversial, and species-specific variations have been Vmax, and shorter APD compared with LA cells.73 This has been
suggested.66 Stretch-activated ion channels (ISAC), which have attributed to a greater density of IKr, IKs, and a smaller density of
been reported in human atrial cells,67 influence both Vrest and ICaL, Ito1 in PV than in LA cells, whereas the densities of INCX and
APD; however, their molecular correlates remain unknown. ICaT were found to be similar between PV and LA cells.73 It is
Swelling-induced, outwardly rectifying chloride channels, such interesting to note that Ca2+-handling properties were not differ-
as ICl−68 (putative molecular correlates Cl3,Cl6,Cl740), have been ent between canine PV and LA cells.87 In rabbit PV cells, most
reported in human atrial cells. The transcripts of two-pore chan- cells demonstrated pacemaker activity.72 As reported in the ven-
nels (TWIK1, TASK1) have also been reported to be present in tricle, epi-endocardial differences also exist within the atrium. A
the human atrium.40 However, no information is available on the recent study in pigs reported a shorter refractory period in the
contribution of these currents or that of ICl− to the APD. atrial epicardium compared with the endocardium.88 This is in
line with a shorter atrial epicardial APD recorded in the canine
right atrial free wall compared with the endocardium.89 The
Ionic Bases of Regional Heterogeneity canine atrial endocardium was also more sensitive to APD short-
in Atrial Action Potentials ening after the addition of acetylcholine.89 However, the ionic
mechanisms underlying atrial epicardium-endocardium differ-
The atrium is a complicated three-dimensional (3D) structure, ences remain unexplored.
and the heterogeneities in electrical properties between different
regions, such as the left atrium (LA), the right atrium (RA), and
the pulmonary veins (PV) and their ionic basis have been exten- Ionic Bases of Atrial Action Potential
sively studied in many species, including mouse,69,70 rabbit,71,72 Variations With Age
and dog.73,74 Information regarding such heterogeneity in humans
is limited, but recent studies have begun to address the putative Age-related changes in atrial action potential differences in
ionic/molecular basis for these differences. The frequencies of humans have been most systematically investigated between
excitation in AF have been reported to show regional gradients infants (neonatal) and adults.29,30 Adult cells displayed a more
during paroxysmal AF and postoperative AF, with LA-PV regions prominent initial notch compared with neonatal atrial cells.29,30
displaying the highest frequencies.75-77 This is highly indicative The properties of Ito1 were significantly different as well; Ito1
of underlying differences in biophysical properties of some ionic displayed significantly higher density, faster inactivation, and a
currents, and these variations, which are more prominent in slower recovery from inactivation in neonatal compared with
disease conditions, are discussed later. In patients in normal sinus adult human atrial cells.29 Correspondingly, the protein density
rhythm, Caballero and colleagues reported a higher density of of Kv4.3 was higher and KChiP2 was lower in neonatal compared
IKur in the right than in the left atrial cardiomyocytes.52 Incorpo- with adult cells.29 The properties of ICaL were also different:
rating such data in human atrial mathematical models did not The basal density was smaller and the protein density of Giα3
result in appreciable differences between the LA/RA action was larger in neonatal compared with adult cells, resulting in a
different response to a lower dosage of β-adrenergic stimula- rate of 5 to 10 Hz.92 Several studies have investigated the ionic
tion.90 In addition, Ca2+ transients, which influence both adult mechanisms involved in the remodeling that occurs in the atria
and neonatal human atrial action potentials,14,17,91 may display of patients with long-standing chronic AF, and have shown that
inherently different biophysical properties but remain unex- electrophysiological remodeling contributes to the development
plored. It is well known that the propensity for AF increases with of a substrate that facilitates the tendency for persistence of
age,92 but very few studies have been conducted to explore poten- AF.102,103 Electrical remodeling induces changes in the biophysical
tial differences between adult and aged human atrial electro- properties of Na+, Ca2+, and K+ currents, which lead to shortening
physiology, and most of the knowledge in this area has been of the APD.104 Abnormal [Ca2+]i homeostasis also plays an impor-
derived from animal models (rat, rabbit, dog). This topic was tant role in AF-induced electrical remodeling.105,106 These changes
reviewed comprehensively in a recent article. We briefly describe in E-C coupling have been reviewed comprehensively in a recent
the salient points.20 Vrest was found to be more depolarized in rats/ article,16 and the important underlying ionic mechanisms are
dogs,94,95 whereas APD90 was more prolonged in rats/rabbits/dogs discussed in the following section.
in aged compared with adult atria.20,96,97 The density of INa was
not different,84 but the density of ICaL was reported to be smaller
in canine atrial cells.98 Further, the density of Ito1 was higher, its Effects of AF on AP and [Ca2+]i
steady-state half-inactivation value showed more positive values,
and recovery from inactivation was slower in canine aged atrial Atrial cells from patients in chronic AF (cAF) display shorter
cells compared with their adult counterparts.98 The density of the APD compared with healthy patients in normal sinus rhythm
sustained outward K+ current, Isus, was also reported to be larger (Figure 31-2, A).28,107-110 Furthermore, the normal human atrial
in aged canine atrial cells.98 In addition to the changes in ionic APD90 shortens when paced at progressively faster frequencies,
currents and action potentials, an increased level of interstitial but in cAF, this shortening is severely attenuated.17 The peak
fibrosis and modifications in intercellular coupling have been amplitude of [Ca2+]i is reduced in atrial myocytes from cAF
reported in aged compared with adult tissue,99,100 and this may patients compared with those of healthy patients (Figure 31-2, B),
further enhance the susceptibility to AF with age. It is likely that although the SR Ca2+ content is unaltered.17 [Ca2+]i decays more
Ca2+ handling is different between aged and adult atrial tissue, slowly in cAF compared with sinus rhythm.17,112 Our recently
but this issue remains unexplored, especially in humans. published mathematical model of the human atrial cell provided
novel insights into the ionic mechanisms underlying the altered
APD/[Ca2+]i in cAF (see Figure 31-2, A, B; black: sinus rhythm,
red: cAF); the salient points are discussed here.
Action Potential and Ionic Remodeling in
Chronic Atrial Fibrillation
Ionic Remodeling in cAF
AF is the most common cardiac arrhythmia seen in the clinic; it
affects approximately 2 million people in the United States Sarcolemmal Ion Channels
alone101 and is one of the main causes of embolic stroke.92 AF is INa Bosch et al reported that the density of INa and the voltage
characterized by rapid and irregular activation of the atria, at a dependence of steady-state activation were not altered in cAF in
AP CaT
50 mV
400 nM Sinus rhythm

0 mV
200 ms
Sinus rhythm 200 nM cAF

cAF
∆ current (A/F) 200 ms
Putative clone (A/F)
0 0
I 0 SCN5A INCX
~ Na–200 INaL
–0.02 –1
20 200
0 0.4
ICa α1c ~ INKA
–5
0
5 Sinus rhythm
Ito Kv4.3 B cAF
0
5
IKur Kv1.5
0
0.005
IKs KvLQT1/minK
0
~ IKr 0.1 ERG/miRP1

0
5
IK1 Kir2.1, 2.3
A 0
Figure 31-2. A, Comparison between ionic currents and action potentials, and (B) Ca2+ transients (CaT) and pumps (INaK) and exchanger (INCX) in atrial cells from normal
sinus rhythm (nsr) patients (black) versus chronic AF (cAF) patients (red).
humans, whereas steady-state inactivation was shifted to the right IK1 and IKACh In cAF, increases in both current density28,79,108,120,123
31
by ≈10 mV,110 and no changes were detected in mRNA levels of and mRNA levels of IK128,120 have been reported (see Figure
the Na+ channel gene SCN5A.114 In contrast, data from Sossalla 31-2, A, 8th row). Increased IK1 causes a more negative resting
et al show that both expression of Nav1.5 and peak INa density membrane potential in cAF versus sinus rhythm human atrial
are decreased (slightly) in the atrial myocardium of patients with myocytes.17,28,79 Patients with cAF exhibit agonist-independent
cAF.115 Additionally, the late Na+ current component, INaL (see constitutive IK,ACh activity that contributes to the enhanced basal
inset, Figure 31-2), was reported to be significantly increased in inward rectifier current and may result from abnormal channel
cAF patients.115 Sossalla et al proposed that this increase could be phosphorylation by PKC.28,79,128 Constitutively active IK,ACh is
due to the increase in neuronal Na+ channel isoforms (Nav1.1 considered to support the maintenance of AF, together
expression is increased),115 or it could be mediated by CaMKII, with increased IK1, by stabilizing reentrant activity sustained by
which is increased in AF111,116 and is known to regulate INaL,117 or rotors (faster activation, less meander129; also see later). Voigt et al
to be caused by oxidative stress.118,119 Simulations suggest that an found significant left-to-right gradients in IK1 and constitutively
increased INaL does not contribute significantly to repolarization active IKACh in patients with paroxysmal AF, which were dissipated
in cAF, where the overall APD90 is still shorter than that in in cAF; this may contribute to the left-to-right dominant
normal healthy cells.17 frequency gradients that are often more evident in paroxysmal
AF.75,79
ICaL A reduction in ICaL density by ≈50% in cAF compared
with sinus rhythm (see Figure 31-2, A, 3rd row) is one of the most IKATP Gene expression and electrophysiological studies in
consistent electrophysiological findings observed experimen- patients with AF demonstrated reduced mRNA levels of Kir6.2126
tally.17,107,108,112,120,121 Christ et al121 demonstrated that decreased and current activation,59 but increased current was also reported.60
ICaL density in cAF is not accompanied by altered expression of
the corresponding α1c- and β2a-channel subunits (although Ca2+ and Na+ Handling
other studies found different results122) and proposed that lower INCX Increased expression13,112,130 and abnormal function of
basal ICaL is due to decreased channel phosphorylation in cAF, INCX protein14,17,112 have been reported in human cAF. An increase
which results from an altered ratio of protein kinase/phosphatase in INCX may be an adaptive response to cellular Ca2+ loading and
activity in favor of increased phosphatase activity. It has been may contribute to diminish the Ca2+ overload induced by rapid
shown that blocking ICaL with nifedipine in normal human atrial pacing. The decay rate of caffeine-evoked [Ca2+]i (attribut-
atrial cells results in an AP characteristic typically seen in AF107 able to Ca2+ removal by INCX) is faster in human cAF versus sinus
with respect to morphology, duration, and impaired rate- rhythm myocytes.17,111,112 Note that simulated INCX during an
dependent adaptation. In other words, a reduction in ICaL seems action potential is smaller in AF than in sinus rhythm (see Figure
to be a critical component of the remodeled atrial action potential 31-2, B, 2nd row) as a result of the reduced amplitude of [Ca2+]i
in cAF. (see Figure 31-2, B, 1st row). Na+ overload–induced Ca2+ influx
via reverse-mode NCX has been implicated in Ca2+ overload and
If The hyperpolarization-activated pacemaker current, If, has related arrhythmogenesis, whereas increased Ca2+ extrusion via
been found to be increased in human AF compared with sinus forward mode has been linked to delayed afterdepolarizations.56,131
rhythm, at least at the mRNA level,63 and this could contribute Indeed, Na+ loading and Ca2+ loading are more favored at
to ectopic atrial pacemaker activity. However, no functional evi- increased atrial rates (with AF). However, additional studies are
dence suggests If involvement at the present time in cAF. needed to assess whether delayed afterdepolarizations (DADs)
are important in initiating arrhythmias in AF, and the underlying
Ito1 and IKur Human cAF is associated with strong reduction of role of INCX in mediating them, because an increased IK1 in cAF
Ito1 (see Figure 31-2, A, 4th row) density52,108,110,120,123-125 and down- will tend to oppose the occurrence of such DADs.
regulation of its α-subunit Kv4.3.114,126 IKur (see Figure 31-2, A,
5th row) was reported to be reduced in cAF,52,120,123,125,127 along INaK Workman et al found no difference in Na+/K+ pump
with diminished expression of Kv1.5 in some studies.114,123,126 current in myocytes from cAF patients compared with sinus
However, others have reported no changes in IKur density.108,110,124 rhythm, and concluded that INaK is not involved in AF-induced
Inconsistent results regarding IKur function have been commented electrophysiological remodeling in patients.132 Our simulations
on previously by Christ et al and were attributed to different show different INaK current underlying the AP in cAF versus sinus
strategies for identification of IKur (e.g., pharmacological or with rhythm (see Figure 31-2, B, 3rd row) caused by altered Na+ loading.
Ito1-inactivating prepulse) and to a fraction of IKur that is not Our simulations also indicate that the APD rate adaptation in
accounted for by Kv1.5.127 The reduction in Ito1 and IKur explains sinus rhythm atrial cells involves accumulation of intracellular
the slight prolongation in earlier phases of the action potential Na+ ([Na+]i) at high frequencies, and this causes outward shifts in
(see Figure 31-2, A, 1st row).17 Caballero et al recently looked at Na+/Ca2+ exchange and Na+/K+ pump currents. The model also
differences in current density and AF-induced alterations in the predicts that E-C coupling remodeling in cAF would reduce Na+
right versus the left human atrium. They found that cAF reduced accumulation, thus causing a blunted APD rate–dependent
the Ito1 amplitude and density more markedly in LA than in RA, response.17
thus creating a right-to-left gradient, whereas IKur was more
markedly reduced in RA than in LA, thus dissipating the left-to- Ryanodine Receptors (RyRs)
right gradient detected in sinus rhythm.52 Spontaneous Ca2+-release events (Ca2+ sparks) and Ca2+ waves
through leaky RyR channels have been reported in myocytes
IKs and IKr To date, no experimental evidence has been pro- from hearts in AF111-113,133,134 despite unaltered SR Ca2+ content.
vided regarding the involvement of IKr in AF-induced electrical One potential contributor to RyR hyperactivity may be oxidative
remodeling. Recently, Caballero et al provided the first demon- stress, which is known to play a critical role in AF pathophysiol-
stration that cAF significantly increased the amplitude of IKs ogy118 and to increase RyR open probability. Neef et al suggested
in both atria.52 They suggested that an increase in IKs could con- that the CaMKII-dependent increase in SR Ca2+ leak caused by
tribute to cAF-induced shortening of APD and could further RyR hyperphosphorylation in AF is a potential arrhythmogenic
promote fibrillatory conduction, especially with its accumulation mechanism,111 because elimination of Ca2+ via inward INCX could
at higher pacing frequencies, as has been shown in neonatal rat lead to cell depolarization and could cause DADs. Voigt et al
ventricular myocytes, where IKs was overexpressed.140 measured directly single RyRs isolated from cAF patients and
demonstrated a higher channel open probability in cAF that how they predispose to more frequent AF episodes culminating in
responded to CaMKII inhibition.14 cAF, remain poorly understood.
Changes in atrial electrical remodeling due to ventricular dys-
SR Ca2+ ATPase and PLN The SR Ca2+ ATPase (SERCA) is function (mainly tachypace-induced or myocardial infarction–
responsible for pumping Ca2+ back into the SR after Ca2+ release.56 induced heart failure) have also been studied in animal models,
The endogenous inhibitor PLN regulates SERCA and releases and the most systematic studies have been conducted in dogs by
its inhibition when phosphorylated by PKA or CaMKII.56,135 A the Nattel group.146,147 In atria of dogs with short-term (2 to 5
decrease in SERCA activity, associated with smaller SERCA weeks) tachypace-induced heart failure, the APD was not changed
protein expression, is evident in human cAF and explains the at slower pacing frequencies, but it increased in duration at rapid
slower [Ca2+]i decay compared with sinus rhythm.13,17,112 On the pacing frequencies.147 The densities of ICaL, Ito1, and IKs were
other hand, reduced inhibition of SERCA by hyperphosphory- reduced, but no change in their voltage dependencies or kinetics
lated PLN13 in cAF could help to maintain a normal SR Ca2+ load was noted. The densities of the K+ currents IK1, IKur, IKr, and ICaT
despite increased RyR activity. were not altered, but that of INCX was increased.147 Further, sub-
Findings of APD shortening and changes in Na+, Ca2+, K+ stantial dysregulation in [Ca2+]i and in its regulatory proteins was
currents, as well as electrogenic pumps and exchangers due to reported, which led to increased diastolic [Ca2+]i level, [Ca2+]i
cAF in humans, have been largely reproducible in canine/goat/ amplitude, SR load, and spontaneous release events.148 In con-
rabbit models of tachypace-induced AF, and suggest that electri- trast, for a long-term tachypaced canine model of ventricular
cal remodeling, including changes in APD and ionic mechanisms, heart failure (4 months), the APD was shortened significantly.149
occurs over a period of 1 to 2 weeks.136,137 In contrast, structural Ionic changes reported included an increased Ito1 and decreased
remodeling occurs over longer periods of atrial tachypacing IK1, IKur, and IKs. Upon study of ICaL under action potential clamp
(months) and is mostly irreversible.136,137 conditions, the heart failure atrial action potential reduced the
integral of ICaL in control, healthy cells, with a larger reduction
in myocytes from heart failure dogs.149 These data point toward
a complex electrical phenotype and underlying ionic mechanism
Alterations in Atrial Electrophysiology during changes in the atrium in the presence of heart failure, depending
Ventricular Dysfunction on the severity of the condition (short-term vs. long-term ven-
tricular pacing); this in part mirrors the complex and varied
Atrial remodeling that increases propensity for AF also occurs phenotype seen in patients with ventricular dysfunction.150
with cardiac disorders, such as coronary artery disease, congestive
heart failure, and left ventricular (LV) systolic dysfunction.138 In
atrial cells isolated from patients with LV dysfunction, the APD
was unaltered139,22 or prolonged.141 In sinus rhythm patients with Ionic Basis of Reentry (Spiral Waves)
reduced left ventricular ejection fraction (<45%), APD90 was in the Atrium
shorter than in patients with higher ejection fraction,22 and a sig-
nificant correlation was noted between cellular effective refrac- We conclude this chapter with a very brief summary about the
tory period (ERP) shortening and decreasing left ventricular ionic basis of spiral waves and rotors that have been shown to
ejection fraction.22 Furthermore, multivariate analysis adjusting sustain AF.151
for 10 relevant clinical covariates confirmed that LV dysfunction The initiation of atrial arrhythmias can occur at the cellular
was independently associated with atrial cellular ERP shortening, level, in the form of afterdepolarizations, early or late (EADs and
which may, therefore, be expected to contribute to a predisposi- DADs), or automaticity. The ionic mechanisms underlying these
tion to AF in these patients. Ionic remodeling due to heart failure/ arrhythmogenic action potentials are somewhat similar to their
LV dysfunction in human atrium remains poorly understood. ICaL ventricular counterpart. For example, our computer simulations
was decreased in patients with coronary artery disease, aortic valve have showed that, in the presence of a Kv1.5 mutation that
disease, or mitral valve disease142 or was unchanged in LV renders IKur inactive, β-adrenergic stress can give rise to EADs,
dysfunction/heart failure patients.22,143 Schreieck et al found primarily on account of the reactivation of ICaL.17 Similarly,
increased Ito1 density in human atrial myocytes of patients with abnormal Ca2+ homeostasis leading to enhanced Ca2+ leak, in
reduced LV function, with no change in its voltage dependence or combination with an increased density of INCX, was shown to
decay, but with faster recovery from activation.139 However, this underlie an increased propensity for DADs in cells isolated from
Ito1 increase may have been confounded by the lower proportion cAF patients compared with normal healthy ones.14 These
of patients treated with β-blockers in the reduced LV function arrhythmogenic action potentials can lead to vortex shedding and
group, because such treatment is associated with decreased Ito1 in wavebreaks in cardiac tissue, which can generate rotors.152
human atrium.144 In contrast, Workman et al found that LV dys- We129 and others153-155 have studied the ionic basis of these
function was associated with decreased Ito1, a positive shift in its rotors. In a simple two-dimensional (2D) homogeneous and iso-
activation voltage, and no change in its decay kinetics.22 Koumi tropic sheet (5 cm × 5 cm) that incorporated mathematical
et al reported depolarized Vrest in atrial myocytes from heart models of human atrial cells (500 × 500 cells), rotors were simu-
failure patients, possibly due to reduced density of IK1 and IKACh.141 lated by cross-field stimulation.129 The ionic conditions seen
Workman et al reported unchanged IK1 in LV dysfunction,22 experimentally in chronic AF were mimicked in two cases using
although Ba2+-sensitive IK1 or IKACh was not measured. Unchanged the Courtemanche atrial model9 (1) in which downregulation in
atrial IKur has also been reported in human LV dysfunction.22,139 ICaL, Ito1, and IKur was implemented (CAF1),156 and (2) where, in
Cardiac dilatation is known to develop frequently during the addition to reduced densities of ICaL, Ito1, and IKur, the density of
course of cardiac failure.145 In trabeculae and myocytes taken from IK1 was increased twofold (CAF2), as was found in experiments.123
dilated atria, the APD was shorter and the plateau was markedly The resulting APD shortening and restitution are shown for
depressed compared with those in trabeculae and myocytes from CAF1 and CAF2 (Figure 31-3, A, B, respectively; the DI in
nondilated atria.145 However, it must be noted that ventricular Figure 31-3, B represents the diastolic interval for an S1-S2
dysfunction was not quantified in these patients. AP changes were stimulus protocol). The increase in IK1, in addition to shortening
explained with more severely depressed ICaL compared with the the APD further, caused a hyperpolarization of Vrest (see Figure
reduction in total outward current.145 Overall, the ionic bases of 31-3, A). Both conditions allowed for the maintenance of stable
altered atrial APD in patients with ventricular dysfunction, and rotors in 2D sheets, as illustrated in Figure 31-3, C. However,
40
400
31
Control
0 APD–70 (msec)
Control
CAF1
mV
200
–40 CAF1
–80
CAF2
CAF2 0
0 250 500 0 200 400
A ms B DI (ms)
CAF1 CAF2
Spiral
2.4 s 4.8 s snapshots 2.4 s 4.8 s
7.2 s 9.6 s 7.2 s 9.6 s
5.0 5.0
Tip
cm
cm
2.5 meander 2.5
0.0 0.0
0.0 2.5 5.0 0.0 2.5 5.0
cm cm
5.7 Hz 8.4 Hz
Power
spectrum
0 5 10 15 0 5 10 15
Freq (Hz) Freq (Hz)
C
Figure 31-3. A, Action potentials in control and chronic AF (CAF1, CAF2). B, Electrical restitution plotted as APD−70 versus the diastolic interval (DI) in control and chronic
AF. C, Spiral waves (phase movie), tip meander, and power spectral densities in chronic AF conditions CAF1 and CAF2. Phase movies are shown at separate times (2.4, 4.8,
7.2, and 9.6 s). The tip meander is plotted in a 5 × 5-cm2, two-dimensional (2D) atrial sheet.
the rotors had different properties: The rotor in the CAF2 condi- faster repolarization and hyperpolarized Vrest.129 Because INa is the
tion had a higher frequency of rotation (8.4 Hz) compared with key determinant of excitation, this allowed for faster rotor activ-
that in CAF1 (5.7 Hz).129 Further, the meander of the spiral wave ity. It is interesting to note that when ICaL in the CAF1 condition
tip was smaller in the CAF2 condition (see Figure 31-3, C). Thus, was further reduced, it shortened APD but did not hyperpolarize
an increase in IK1 caused the rotors to become more stable (faster Vrest, and the rotor was able to accelerate only a little (6.3 Hz),
frequency, less meander). Further analyses (not shown) revealed thus providing further support to the key role of IK1 in accelerat-
that the key mechanism underlying rotor acceleration was the ing reentry. These simulations have been supported by experi-
removal of INa inactivation (increased availability) on account of mental results in mice and dogs. Transgenic mice overexpressing
IK1 were able to sustain much faster and more sustained rotors of adenosine, which was shown to increase AF frequency of
(frequency ≈44 Hz) compared with wild type mice (≈26 Hz).157 excitation.160 The role of other K+ currents in sustaining rotors
Further, the Nattel lab compared AF frequencies in dogs with in AF is less clear. When the maximum conductance of Ito1 and
matched effective refractory periods in two models of AF: one in IKur was reduced by 90%, the rotor was terminated, but when the
which APD was shortened as the result of cholinergic activation same was done for IKr and IKS, the rotor activity remained sus-
of IKACh, an inward rectifier K+ current, and one in which the APD tained.129 Our simulations were devoid of any ionic and structural
was shortened through atrial tachypacing (thus the ionic mecha- heterogeneities, which play an important role in sustaining AF;
nism of shortening consisted of both a reduction in ICaL and an thus more work is needed to allow workers to define the ionic
increase in IK1).158 It was found that the frequency of AF was bases of reentry in detailed 3D structural models of the atria161
higher in the case where the APD was shortened through IKACh that incorporate ionic heterogeneities,19 as well as interstitial
alone, compared with shortening induced by remodeling of both fibrosis.162 Last, simulations that incorporate more recent ionic
ICaL and IK1 in tachypaced hearts.158 These experiments provide models with detailed Ca2+ formulations,17 in combination with
support for the notion that inward rectifier currents are critical further experiments, are needed to obtain more detailed quantita-
determinants of reentry and its activation frequency.159 This tive insights into whether Ca2+ homeostasis is important in main-
effect has been verified in the human atria by infusion taining rotors in AF.
Mechanistic insights and therapeutic opportuni- 30. Escande D, Loisance D, Planche C, et al: Age-
References ties. Eur Heart J 33:1870–1877, 2012. related changes of action potential plateau shape
16. Grandi E, Workman AJ, Pandit SV: Remodeling in isolated human atrial fibers. Am J Physiol 249(4
1. Trautwein W, Kassebaum DG, Nelson RM, et al: of E-C coupling in atrial fibrillation: Invited Pt 2):H843–H850, 1985.
Electrophysiological study of human heart review. J Atrial Fibrillation 2:1–17, 2012. 31. Wang ZG, Pelletier LC, Talajic M, et al: Effects
muscle. Circ Res 10:306–312, 1962. 17. Grandi E, Pandit SV, Voigt N, et al: Human atrial of flecainide and quinidine on human atrial action
2. Gelband H, Bush HL, Rosen MR, et al: Electro- action potential and Ca2+ model: Sinus rhythm potentials: Role of rate-dependence and compari-
physiologic properties of isolated preparations of and chronic atrial fibrillation. Circ Res 109:1055– son with guinea pig, rabbit, and dog tissues. Cir-
human atrial myocardium. Circ Res 30:293–300, 1066, 2011. culation 82:274–283, 1990.
1972. 18. Koivumäki JT, Korhonen T, Tavi P: Impact of 32. Varró A, Nánási PP, Lathrop DA: Potassium cur-
3. Hamill OP, Marty A, Neher E, et al: Improved sarcoplasmic reticulum calcium release on calcium rents in isolated human atrial and ventricular car-
patch-clamp techniques for high-resolution dynamics and action potential morphology in diocytes. Acta Physiol Scand 149:133–142, 1993.
current recording from cells and cell-free mem- human atrial myocytes: A computational study. 33. O’Hara T, Virág L, Varró A, et al: Simulation of
brane patches. Pflugers Arch 391:85–100, PLoS Comput Biol 7:e1001067, 2011. the undiseased human cardiac ventricular action
1981. 19. Schram G, Pourrier M, Melnyk P, et al: Differen- potential: Model formulation and experimental
4. Shibata EF, Drury T, Refsum H, et al: Contribu- tial distribution of cardiac ion channel expression validation. PLoS Comput Biol 7:e1002061, 2011.
tions of a transient outward current to repolariza- as a basis for regional specialization in electrical 34. Xu H, Barry DM, Li H, et al: Attenuation of the
tion in human atrium. Am J Physiol 257(6 Pt function. Circ Res 90:939–950, 2002. Review. slow component of delayed rectification, action
2):H1773–H1781, 1989. 20. Dun W, Boyden PA: Aged atria: Electrical remod- potential prolongation, and triggered activity in
5. Firek L, Giles WR: Outward currents underlying eling conducive to atrial fibrillation. J Interv Card mice expressing a dominant-negative Kv2 alpha
repolarization in human atrial myocytes. Cardio- Electrophysiol 25:9–18, 2009. subunit. Circ Res 85:623–633, 1999.
vasc Res 30:31–38, 1995. 21. Van Wagoner DR, Nerbonne JM: Molecular basis 35. Xu H, Li H, Nerbonne JM: Elimination of the
6. Fedida D, Wible B, Wang Z, et al: Identity of a of electrical remodeling in atrial fibrillation. J Mol transient outward current and action potential
novel delayed rectifier current from human heart Cell Cardiol 32:1101–1117, 2000. Review. prolongation in mouse atrial myocytes expressing
with a cloned K+ channel current. Circ Res 22. Workman AJ, Pau D, Redpath CJ, et al: Atrial a dominant negative Kv4 alpha subunit. J Physiol
73:210–216, 1993. cellular electrophysiological changes in patients 519(Pt 1):11–21, 1999.
7. Wang Z, Fermini B, Nattel S: Delayed rectifier with ventricular dysfunction may predispose to 36. Furukawa T, Koumi S, Sakakibara Y, et al: An
outward current and repolarization in human AF. Heart Rhythm 6:445–451, 2009. analysis of lidocaine block of sodium current in
atrial myocytes. Circ Res 73:276–285, 1993. 23. Skanes AC, Mandapati R, Berenfeld O, et al: Spa- isolated human atrial and ventricular myocytes.
8. Nygren A, Fiset C, Firek L, et al: Mathematical tiotemporal periodicity during atrial fibrillation in J Mol Cell Cardiol 27:831–846, 1995.
model of an adult human atrial cell: The role of the isolated sheep heart. Circulation 98:1236– 37. Sakakibara Y, Furukawa T, Singer DH, et al:
K+ currents in repolarization. Circ Res 82:63–81, 1248, 1998. Sodium current in isolated human ventricular
1998. 24. Narayan SM, Krummen DE, Rappel WJ: Clinical myocytes. Am J Physiol 265(4 Pt 2):H1301–
9. Courtemanche M, Ramirez RJ, Nattel S: Ionic mapping approach to diagnose electrical rotors H1309, 1993.
mechanisms underlying human atrial action and focal impulse sources for human atrial fibril- 38. Jia H, Furukawa T, Singer DH, et al: Character-
potential properties: Insights from a mathematical lation. J Cardiovasc Electrophysiol 23:447–454, istics of lidocaine block of sodium channels in
model. Am J Physiol 275(1 Pt 2):H301–H321, 2012. single human atrial cells. J Pharmacol Exp Ther
1998. 25. Narayan SM, Patel J, Mulpuru S, et al: Focal 264:1275–1284, 1993.
10. Bénardeau A, Hatem SN, Rücker-Martin C, et al: impulse and rotor modulation ablation of sustain- 39. Sakakibara Y, Wasserstrom JA, Furukawa T, et al:
Contribution of Na+/Ca2+ exchange to action ing rotors abruptly terminates persistent atrial Characterization of the sodium current in single
potential of human atrial myocytes. Am J Physiol fibrillation to sinus rhythm with elimination on human atrial myocytes. Circ Res 71:535–546,
271(3 Pt 2):H1151–H1161, 1996. follow-up: A video case study. Heart Rhythm 1992.
11. Hatem SN, Bénardeau A, Rücker-Martin C, et al: 2012 Mar 28. [Epub ahead of print] 40. Gaborit N, Le Bouter S, Szuts V, et al: Regional
Different compartments of sarcoplasmic reticu- 26. Grandi E, Pasqualini FS, Bers DM: A novel com- and tissue specific transcript signatures of ion
lum participate in the excitation-contraction cou- putational model of the human ventricular action channel genes in the non-diseased human heart.
pling process in human atrial myocytes. Circ Res potential and Ca transient. J Mol Cell Cardiol J Physiol 582(Pt 2):675–693, 2007.
80:345–353, 1997. 48:112–121, 2010. 41. Burashnikov A, Di Diego JM, Zygmunt AC, et al:
12. Hove-Madsen L, Llach A, Bayes-Genís A, et al: 27. Amos GJ, Wettwer E, Metzger F, et al: Differ- Atrium-selective sodium channel block as a strat-
Atrial fibrillation is associated with increased ences between outward currents of human atrial egy for suppression of atrial fibrillation: Differ-
spontaneous calcium release from the sarcoplas- and subepicardial ventricular myocytes. J Physiol ences in sodium channel inactivation between
mic reticulum in human atrial myocytes. Circula- 491(Pt 1):31–50, 1996. atria and ventricles and the role of ranolazine.
tion 110:1358–1363, 2004. 28. Dobrev D, Graf E, Wettwer E, et al: Molecular Circulation 116:1449–1457, 2007.
13. El-Armouche A, Boknik P, Eschenhagen T, et al: basis of downregulation of G-protein-coupled 42. Li GR, Lau CP, Shrier A: Heterogeneity of
Molecular determinants of altered Ca2+ handling inward rectifying K(+) current I(K,ACh) in sodium current in atrial vs epicardial ventricular
in human chronic atrial fibrillation. Circulation chronic human atrial fibrillation: Decrease in myocytes of adult guinea pig hearts. J Mol Cell
114:670–680, 2006. GIRK4 mRNA correlates with reduced I(K,ACh) Cardiol 34:1185–1194, 2002.
14. Voigt N, Li N, Wang Q, et al: Enhanced sarco- and muscarinic receptor-mediated shortening of 43. Li GR, Nattel S: Properties of human atrial ICa
plasmic reticulum Ca2+ leak and increased Na+- action potentials. Circulation 104:2551–2557, at physiological temperatures and relevance to
Ca2+ exchanger function underlie delayed 2001. action potential. Am J Physiol 272(1 Pt 2):H227–
afterdepolarizations in patients with chronic atrial 29. Wang Y, Xu H, Kumar R, et al: Differences in H235, 1997.
fibrillation. Circulation 125:2059–2070, 2012. transient outward current properties between 44. Cohen NM, Lederer WJ: Calcium current in
15. Nattel S, Dobrev D: The multidimensional role neonatal and adult human atrial myocytes. J Mol single human cardiac myocytes. J Cardiovasc
of calcium in atrial fibrillation pathophysiology: Cell Cardiol 35:1083–1092, 2003. Electrophysiol 4:422–437, 1993.
45. Ouadid H, Seguin J, Dumuis A, et al: Serotonin 64. Xu Y, Tuteja D, Zhang Z, et al: Molecular iden- unequal left/right atrial frequencies and rotor
increases calcium current in human atrial myo- tification and functional roles of a Ca(2+)- dynamics. Cardiovasc Res 59:863–873, 2003.
cytes via the newly described 5-hydroxytryptamine4
receptors. Mol Pharmacol 41:346–351, 1992.
46. Rivet-Bastide M, Vandecasteele G, Hatem S, et
activated K+ channel in human and mouse hearts.
J Biol Chem 278:49085–49094, 2003.
65. Skibsbye L, Poulet C, Diness JG, et al: Inhibition
83. Lomax AE, Rose RA, Giles WR: Electrophysio-
logical evidence for a gradient of G protein-gated
K+ current in adult mouse atria. Br J Pharmacol
31
al: cGMP-stimulated cyclic nucleotide phospho- of Ca-activated K channels in human atrial tissue, 140:576–584, 2003.
diesterase regulates the basal calcium current in and its role in atrial fibrillation: S56. Heart 84. Baba S, Dun W, Hirose M, et al: Sodium current
human atrial myocytes. J Clin Invest 99:2710– Rhythm 9(Suppl), 2012. Abstract. function in adult and aged canine atrial cells. Am
2718, 1997. 66. Nagy N, Szuts V, Horváth Z, et al: Does small- J Physiol Heart Circ Physiol 291:H756–H761,
47. Hatem SN, Coulombe A, Balse E: Specificities of conductance calcium-activated potassium channel 2006.
atrial electrophysiology: Clues to a better under- contribute to cardiac repolarization? J Mol Cell 85. Yamashita T, Nakajima T, Hazama H, et al:
standing of cardiac function and the mechanisms Cardiol 47:656–663, 2009. Regional differences in transient outward current
of arrhythmias. J Mol Cell Cardiol 48:90–95, 67. Kamkin A, Kiseleva I, Wagner KD, et al: Charac- density and inhomogeneities of repolarization in
2010. Review. terization of stretch-activated ion currents in iso- rabbit right atrium. Circulation 92:3061–3069,
48. Escande D, Coulombe A, Faivre JF, et al: Two lated atrial myocytes from human hearts. Pflugers 1995.
types of transient outward currents in adult Arch 446:339–346, 2003. 86. Aslanidi OV, Boyett MR, Dobrzynski H, et al:
human atrial cells. Am J Physiol 252(1 Pt 68. Demion M, Guinamard R, El Chemaly A, et al: Mechanisms of transition from normal to reen-
2):H142–H148, 1987. An outwardly rectifying chloride channel in trant electrical activity in a model of rabbit atrial
49. Fermini B, Wang Z, Duan D, et al: Differences human atrial cardiomyocytes. J Cardiovasc Elec- tissue: interaction of tissue heterogeneity and
in rate dependence of transient outward current trophysiol 17:60–68, 2006. anisotropy. Biophys J 96:798–817, 2009.
in rabbit and human atrium. Am J Physiol 263(6 69. Lomax AE, Rose RA, Giles WR: Electrophysio- 87. Coutu P, Chartier D, Nattel S: Comparison of
Pt 2):H1747–H1754, 1992. logical evidence for a gradient of G protein-gated Ca2+-handling properties of canine pulmonary
50. Wang Z, Fermini B, Nattel S: Sustained K+ current in adult mouse atria. Br J Pharmacol vein and left atrial cardiomyocytes. Am J Physiol
depolarization-induced outward current in 140:576–584, 2003. Heart Circ Physiol 291:H2290–H2300, 2006.
human atrial myocytes: Evidence for a novel 70. Lomax AE, Kondo CS, Giles WR: Comparison 88. Michowitz Y, Nakahara S, Bourke T, et al: Elec-
delayed rectifier K+ current similar to Kv1.5 of time- and voltage-dependent K+ currents in trophysiological differences between the epicar-
cloned channel currents. Circ Res 73:1061–1076, myocytes from left and right atria of adult mice. dium and the endocardium of the left atrium.
1993. Am J Physiol Heart Circ Physiol 285:H1837– Pacing Clin Electrophysiol 34:37–46, 2011.
51. Wang Z, Fermini B, Nattel S: Rapid and slow H1848, 2003. 89. Anyukhovsky EP, Rosenshtraukh LV:
components of delayed rectifier current in human 71. Tsai WC, Chen YC, Lin YK, et al: Sex differences Electrophysiological responses of canine atrial
atrial myocytes. Cardiovasc Res 28:1540–1546, in the electrophysiological characteristics of pul- endocardium and epicardium to acetylcholine and
1994. monary veins and left atrium and their clinical 4-aminopyridine. Cardiovasc Res 43:364–370,
52. Caballero R, de la Fuente MG, Gómez R, et al: implication in atrial fibrillation. Circ Arrhythm 1999.
In humans, chronic atrial fibrillation decreases the Electrophysiol 4:550–559, 2011. 90. Tipparaju SM, Kumar R, Wang Y, et al: Develop-
transient outward current and ultrarapid compo- 72. Chen YJ, Chen SA, Chen YC, et al: Electrophysi- mental differences in L-type calcium current of
nent of the delayed rectifier current differentially ology of single cardiomyocytes isolated from human atrial myocytes. Am J Physiol Heart Circ
on each atria and increases the slow component rabbit pulmonary veins: Implication in initiation Physiol 286:H1963–H1969, 2004.
of the delayed rectifier current in both. J Am Coll of focal atrial fibrillation. Basic Res Cardiol 91. Wagner MB, Wang Y, Kumar R, et al: Calcium
Cardiol 55:2346–2354, 2010. 97:26–34, 2002. transients in infant human atrial myocytes.
53. Dobrev D, Carlsson L, Nattel S: Novel molecular 73. Ehrlich JR, Cha TJ, Zhang L, et al: Cellular elec- Pediatr Res 57:28–34, 2005.
targets for atrial fibrillation therapy. Nat Rev trophysiology of canine pulmonary vein cardio- 92. Benjamin EJ, Chen PS, Bild DE, et al: Prevention
Drug Discov 11:275–291, 2012. myocytes: Action potential and ionic current of atrial fibrillation: report from a national heart,
54. Sun H, Leblanc N, Nattel S: Mechanisms of inac- properties. J Physiol 551(Pt 3):801–813, 2003. lung, and blood institute workshop. Circulation
tivation of L-type calcium channels in human 74. Ehrlich JR, Cha TJ, Zhang L, et al: Characteriza- 119:606–618, 2009.
atrial myocytes. Am J Physiol 272(4 Pt 2):H1625– tion of a hyperpolarization-activated time- 93. Pino R, Cerbai E, Calamai G, et al: Effect of
H1635, 1997. dependent potassium current in canine 5-HT4 receptor stimulation on the pacemaker
55. Li GR, Nattel S: Demonstration of an inward cardiomyocytes from pulmonary vein myocardial current I(f) in human isolated atrial myocytes.
Na(+)-Ca2+ exchange current in adult human sleeves and left atrium. J Physiol 557(Pt 2):583– Cardiovasc Res 40:516–522, 1998.
atrial myocytes. Ann N Y Acad Sci 779:525–528, 597, 2004. 94. Su N, Duan J, Moffat MP, et al: Age-related
1996. 75. Sanders P, Berenfeld O, Hocini M, et al: Spectral changes in electrophysiological responses to
56. Bers DM: Excitation-Contraction Coupling and analysis identifies sites of high-frequency activity muscarinic receptor stimulation in rat myocar-
Cardiac Contractile Force, Dordrecht, The maintaining atrial fibrillation in humans. Circula- dium. Can J Physiol Pharmacol 73:1430–1436,
Netherlands, 2001, Kluwer Academic tion 112:789–797, 2005. 1995.
Publishers. 76. Atienza F, Almendral J, Jalife J, et al: Real-time 95. Anyukhovsky EP, Sosunov EA, Chandra P, et al:
57. Koumi S, Arentzen CE, Backer CL, et al: Altera- dominant frequency mapping and ablation of Age-associated changes in electrophysiologic
tions in muscarinic K+ channel response to ace- dominant frequency sites in atrial fibrillation with remodeling: A potential contributor to initiation
tylcholine and to G protein-mediated activation left-to-right frequency gradients predicts long- of atrial fibrillation. Cardiovasc Res 66:353–363,
in atrial myocytes isolated from failing human term maintenance of sinus rhythm. Heart Rhythm 2005.
hearts. Circulation 90:2213–2224, 1994. 6:33–40, 2009. 96. Toda N: Age-related changes in the transmem-
58. Heidbüchel H, Vereecke J, Carmeliet E: Three 77. Swartz MF, Fink GW, Lutz CJ, et al: Left versus brane potential of isolated rabbit sino-atrial nodes
different potassium channels in human atrium: right atrial difference in dominant frequency, K(+) and atria. Cardiovasc Res 14:58–63, 1980.
Contribution to the basal potassium conductance. channel transcripts, and fibrosis in patients devel- 97. Huang C, Ding W, Li L, et al: Differences in the
Circ Res 66:1277–1286, 1990. oping atrial fibrillation after cardiac surgery. aging-associated trends of the monophasic action
59. Balana B, Dobrev D, Wettwer E, et al: Decreased Heart Rhythm 6:1415–1422, 2009. potential duration and effective refractory period
ATP-sensitive K(+) current density during 78. Porciatti F, Pelzmann B, Cerbai E, et al: The of the right and left atria of the rat. Circ J 70:352–
chronic human atrial fibrillation. J Mol Cell pacemaker current I(f) in single human atrial 357, 2006.
Cardiol 35:1399–1405, 2003. myocytes and the effect of beta-adrenoceptor and 98. Dun W, Yagi T, Rosen MR, et al: Calcium and
60. Wu G, Huang CX, Tang YH, et al: Changes of A1-adenosine receptor stimulation. Br J Pharma- potassium currents in cells from adult and aged
IK,ATP current density and allosteric modulation col 122:963–969, 1997. canine right atria. Cardiovasc Res 58:526–534,
during chronic atrial fibrillation. Chin Med J 79. Voigt N, Trausch A, Knaut M, et al: Left-to-right 2003.
(Engl) 118:1161–1166, 2005. atrial inward rectifier potassium current gradients 99. Spach MS, Heidlage JF, Dolber PC, et al: Mecha-
61. Guinamard R, Chatelier A, Demion M, et al: in patients with paroxysmal versus chronic atrial nism of origin of conduction disturbances in aging
Functional characterization of a Ca(2+)-activated fibrillation. Circ Arrhythm Electrophysiol 3:472– human atrial bundles: Experimental and model
non-selective cation channel in human atrial car- 480, 2010. study. Heart Rhythm 4:175–185, 2007.
diomyocytes. J Physiol 558(Pt 1):75–83, 2004. 80. Li D, Zhang L, Kneller J, et al: Potential ionic 100. Burstein B, Nattel S: Atrial fibrosis: Mechanisms
62. Koster OF, Szigeti GP, Beuckelmann DJ: Charac- mechanism for repolarization differences between and clinical relevance in atrial fibrillation. J Am
terization of a [Ca2+]i-dependent current in human canine right and left atrium. Circ Res 88:1168– Coll Cardiol 51:802–809, 2008. Review.
atrial and ventricular cardiomyocytes in the 1175, 2001. 101. Damani SB, Topol EJ: Molecular genetics of atrial
absence of Na+ and K+. Cardiovasc Res 41:175, 81. Hu Y, Jones SV, Dillmann WH: Effects of hyper- fibrillation. Genome Med 1:54, 2009.
1999. thyroidism on delayed rectifier K+ currents in left 102. Nattel S, Burstein B, Dobrev D: Atrial remodel-
63. Lai LP, Su MJ, Lin JL, et al: Measurement of and right murine atria. Am J Physiol Heart Circ ing and atrial fibrillation. Circ Arrhythm Electro-
funny current (If) channel mRNA in human atrial Physiol 289:H1448–H1455, 2005. Erratum in: physiol 1:62–73, 2008.
tissue: Correlation with left atrial filling pressure Am J Physiol Heart Circ Physiol 290:H489, 2006. 103. Wakili R, Voigt N, Kaab S, et al: Recent advances
and atrial fibrillation. J Cardiovasc Electrophysiol 82. Sarmast F, Kolli A, Zaitsev A, et al: Cholinergic in the molecular pathophysiology of atrial fibril-
10:947–953, 1999. atrial fibrillation: I(K,ACh) gradients determine lation. J Clin Invest 121:2955–2968, 2011.
104. Workman AJ, Kane KA, Rankin AC: Cellular expression are reduced in chronic human atrial a high risk of atrial fibrillation. Eur Heart J
bases for human atrial fibrillation. Heart Rhythm fibrillation. Circ Res 80:772–781, 1997. 29:1190–1197, 2008.
5:S1–S6, 2008. 124. Grammer JB, Bosch RF, Kuhlkamp V, et al: 143. Cheng TH, Lee FY, Wei J, et al: Comparison of
105. Dobrev D, Nattel S: Calcium handling abnor- Molecular remodeling of Kv4.3 potassium chan- calcium-current in isolated atrial myocytes from
malities in atrial fibrillation as a target for innova- nels in human atrial fibrillation. J Cardiovasc failing and nonfailing human hearts. Mol Cell
tive therapeutics. J Cardiovasc Pharmacol Electrophysiol 11:626–633, 2000. Biochem 157:157–162, 1996.
52:293–299, 2008. 125. Brandt MC, Priebe L, Bohle T, et al: The ultrar- 144. Marshall GE, Russell JA, Tellez JO, et al: Remod-
106. Dobrev D, Voigt N, Wehrens XHT: The ryano- apid and the transient outward K+ current in elling of human atrial K+ currents but not ion
dine receptor channel as a molecular motif in human atrial fibrillation: Their possible role in channel expression by chronic β-blockade.
atrial fibrillation: Pathophysiological and thera- postoperative atrial fibrillation. J Mol Cell Cardiol Pflugers Arch Eur J Physiol 463:537–548,
peutic implications. Cardiovasc Res 89:734–743, 32:1885–1896, 2000. 2012.
2011. 126. Brundel BJ, Van Gelder IC, Henning RH, et al: 145. Le Grand BL, Hatem S, Deroubaix E, et al:
107. Van Wagoner DR, Pond AL, Lamorgese M, et al: Alterations in potassium channel gene expression Depressed transient outward and calcium cur-
Atrial L-type Ca2+ currents and human atrial in atria of patients with persistent and paroxysmal rents in dilated human atria. Cardiovasc Res
fibrillation. Circ Res 85:428–436, 1999. atrial fibrillation: Differential regulation of 28:548–556, 1994.
108. Workman AJ, Kane KA, Rankin AC: The contri- protein and mRNA levels for K+ channels. J Am 146. Li D, Fareh S, Leung TK, et al: Promotion of
bution of ionic currents to changes in refractori- Coll Cardiol 37:926–932, 2001. atrial fibrillation by heart failure in dogs: Atrial
ness of human atrial myocytes associated with 127. Christ T, Wettwer E, Voigt N, et al: Pathology- remodeling of a different sort. Circulation
chronic atrial fibrillation. Cardiovasc Res 52:226– specific effects of the IKur/Ito/IK,ACh blocker 100:87–95, 1999.
235, 2001. AVE0118 on ion channels in human chronic atrial 147. Li D, Melnyk P, Feng J, et al: Effects of experi-
109. Boutjdir M, Le Heuzey J, Lavergne T, et al: Inho- fibrillation. Br J Pharmacol 154:1619–1630, 2008. mental heart failure on atrial cellular and ionic
mogeneity of cellular refractoriness in human 128. Dobrev D, Friedrich A, Voigt N, et al: The G electrophysiology. Circulation 101:2631–2638,
atrium: Factor of arrhythmia? Pacing Clin Elec- protein-gated potassium current IK,ACh is constitu- 2000.
trophysiol 9:1095–1100, 1986. tively active in patients with chronic atrial fibril- 148. Yeh YH, Wakili R, Qi XY, et al: Calcium-handling
110. Bosch RF, Zeng X, Grammer JB, et al: Ionic lation. Circulation 112:3697–3706, 2005. abnormalities underlying atrial arrhythmogenesis
mechanisms of electrical remodeling in human 129. Pandit SV, Berenfeld O, Anumonwo JM, et al: and contractile dysfunction in dogs with conges-
atrial fibrillation. Cardiovasc Res 44:121–131, Ionic determinants of functional reentry in a 2-D tive heart failure. Circ Arrhythm Electrophysiol
1999. model of human atrial cells during simulated 1:93–102, 2008.
111. Neef S, Dybkova N, Sossalla S, et al: CaMKII- chronic atrial fibrillation. Biophys J 88:3806– 149. Sridhar A, Nishijima Y, Terentyev D, et al:
dependent diastolic SR Ca2+ leak and elevated 3821, 2005. Chronic heart failure and the substrate for atrial
diastolic Ca2+ levels in right atrial myocardium of 130. Schotten U, Greiser M, Benke D, et al: Atrial fibrillation. Cardiovasc Res 84:227–236, 2009.
patients with atrial fibrillation. Circ Res fibrillation-induced atrial contractile dysfunction: 150. Darby AE, Dimarco JP: Management of atrial
106:1134–1144, 2010. A tachycardiomyopathy of a different sort. Car- fibrillation in patients with structural heart
112. Voigt N, Trafford AW, Ravens U, et al: Cellular diovasc Res 53:192–201, 2002. disease. Circulation 125:945–957, 2012.
and molecular determinants of altered atrial Ca2+ 131. Priori SG, Corr PB: Mechanisms underlying Review.
signaling in patients with chronic atrial fibrilla- early and delayed afterdepolarizations induced by 151. Jalife J: Déjà vu in the theories of atrial fibrillation
tion. Circulation 120:S667–S668, 2009. Abstract catecholamines. Am J Physiol 258:H1796–H1805, dynamics. Cardiovasc Res 89:766–775, 2011.
2630. 1990. Review.
113. Liang X, Xie H, Zhu PH, et al: Ryanodine 132. Workman AJ, Kane KA, Rankin AC: Characteri- 152. Cabo C, Pertsov AM, Davidenko JM, et al: Vortex
receptor-mediated Ca2+ events in atrial myocytes sation of the Na, K pump current in atrial shedding as a precursor of turbulent electrical
of patients with atrial fibrillation. Cardiology cells from patients with and without chronic activity in cardiac muscle. Biophys J 70:1105–
111:102–110, 2008. atrial fibrillation. Cardiovasc Res 59:593–602, 1111, 1996.
114. Brundel BJ, Van Gelder IC, Henning RH, et al: 2003. 153. Kneller J, Zou R, Vigmond EJ, et al: Cholinergic
Ion channel remodeling is related to intraopera- 133. Chelu MG, Sarma S, Sood S, et al: Calmodulin atrial fibrillation in a computer model of a two-
tive atrial effective refractory periods in patients kinase II-mediated sarcoplasmic reticulum Ca2+ dimensional sheet of canine atrial cells with real-
with paroxysmal and persistent atrial fibrillation. leak promotes atrial fibrillation in mice. J Clin istic ionic properties. Circ Res 90:E73–E87,
Circulation 103:684–690, 2001. Invest 119:1940–1951, 2009. 2002.
115. Sossalla S, Kallmeyer B, Wagner S, et al: Altered 134. Vest JA, Wehrens XHT, Reiken SR, et al: Defec- 154. Kharche S, Adeniran I, Stott J, et al: Pro-
Na+ currents in atrial fibrillation effects of rano- tive cardiac ryanodine receptor regulation during arrhythmogenic effects of the S140G KCNQ1
lazine on arrhythmias and contractility in human atrial fibrillation. Circulation 111:2025–2032, mutation in human atrial fibrillation: Insights
atrial myocardium. J Am Coll Cardiol 55:2330– 2005. from modelling. J Physiol 2012 Jun 11 [Epub
2342, 2010. 135. Maier LS, Bers DM: Role of Ca2+/calmodulin- ahead of print]
116. Tessier S, Karczewski P, Krause EG, et al: Regula- dependent protein kinase (CaMK) in excitation- 155. Krogh-Madsen T, Abbott GW, Christini DJ:
tion of the transient outward K+ current by Ca2+/ contraction coupling in the heart. Cardiovasc Res Effects of electrical and structural remodeling on
calmodulin-dependent protein kinases II in 73:631–640, 2007. atrial fibrillation maintenance: A simulation study.
human atrial myocytes. Circ Res 85:810–819, 136. Allessie M, Ausma J, Schotten U: Electrical, con- PLoS Comput Biol 8:e1002390, 2012.
1999. tractile and structural remodeling during atrial 156. Courtemanche M, Ramirez RJ, Nattel S: Ionic
117. Wagner S, Dybkova N, Rasenack EC, et al: Ca2+/ fibrillation. Cardiovasc Res 54:230–246, 2002. targets for drug therapy and atrial fibrillation-
calmodulin-dependent protein kinase II regulates Review. induced electrical remodeling: insights from a
cardiac Na+ channels. J Clin Invest 116:3127– 137. Schotten U, Verheule S, Kirchhof P, et al: Patho- mathematical model. Cardiovasc Res 42:477–489,
3138, 2006. physiological mechanisms of atrial fibrillation: A 1999.
118. Mihm MJ, Yu F, Carnes CA, et al: Impaired myo- translational appraisal. Physiol Rev 91:265–325, 157. Noujaim SF, Pandit SV, Berenfeld O, et al:
fibrillar energetics and oxidative injury during 2011. Review. Up-regulation of the inward rectifier K+ current
human atrial fibrillation. Circulation 104:174– 138. Neuberger HR, Mewis C, van Veldhuisen DJ, et (I K1) in the mouse heart accelerates and stabilizes
180, 2001. al: Management of atrial fibrillation in patients rotors. J Physiol 578(Pt 1):315–326, 2007.
119. Wagner S, Ruff HM, Weber SL, et al: Reactive with heart failure. Eur Heart J 28:2568–2577, 158. Katsouras G, Sakabe M, Comtois P, et al: Differ-
oxygen species-activated Ca/calmodulin kinase 2007. ences in atrial fibrillation properties under vagal
IIδ is required for late INa augmentation leading 139. Schreieck J, Wang Y, Overbeck M, et al: Altered nerve stimulation versus atrial tachycardia remod-
to cellular Na and Ca overload. Circ Res 108:555– transient outward current in human atrial myo- eling. Heart Rhythm 6:1465–1472, 2009.
565, 2011. cytes of patients with reduced left ventricular 159. Jalife J: Inward rectifier potassium channels
120. Dobrev D, Ravens U: Remodeling of cardiomyo- function. J Cardiovasc Electrophysiol 11:180– control rotor frequency in ventricular fibrillation.
cyte ion channels in human atrial fibrillation. 192, 2000. Heart Rhythm 6(11 Suppl):S44–S48, 2009.
Basic Res Cardiol 98:137–148, 2003. 140. Muñoz V, Grzeda KR, Desplantez T, et al: Ade- Review.
121. Christ T, Boknik P, Wohrl S, et al: L-type Ca2+ noviral expression of IKs contributes to wave- 160. Atienza F, Almendral J, Moreno J, et al: Activation
current downregulation in chronic human atrial break and fibrillatory conduction in neonatal rat of inward rectifier potassium channels accelerates
fibrillation is associated with increased activity of ventricular cardiomyocyte monolayers. Circ Res atrial fibrillation in humans: evidence for a reen-
protein phosphatases. Circulation 110:2651– 101:475–483, 2007. trant mechanism. Circulation 114:2434–2442,
2657, 2004. 141. Koumi S, Arentzen CE, Backer CL, et al: Altera- 2006.
122. Brundel BJ, van Gelder IC, Henning RH, et al: tions in muscarinic K+ channel response to ace- 161. Harrild D, Henriquez C: A computer model of
Gene expression of proteins influencing the tylcholine and to G protein-mediated activation normal conduction in the human atria. Circ Res
calcium homeostasis in patients with persistent in atrial myocytes isolated from failing human 87:E25–E36, 2000.
and paroxysmal atrial fibrillation. Cardiovasc Res hearts. Circulation 90:2213–2224, 1994. 162. Tanaka K, Zlochiver S, Vikstrom KL, et al: Spatial
42:443–454, 1999. 142. Dinanian S, Boixel C, Juin C, et al: Downregula- distribution of fibrosis governs fibrillation wave
123. Van Wagoner DR, Pond AL, McCarthy PM, et tion of the calcium current in human right atrial dynamics in the posterior left atrium during heart
al: Outward K+ current densities and Kv1.5 myocytes from patients in sinus rhythm but with failure. Circ Res 101:839–847, 2007.
Models of the Ventricular Action
Potential in Health and Disease 32
Eleonora Grandi and Donald M. Bers
rectifier current (IK1) maintains and stabilizes the resting poten-

CHAPTER OUTLINE
tial. Other important players in shaping properties of the cardiac
Computational Models of the Ventricular Myocyte 319 AP are the Na+/K+ pump (NKA), which generates an outward
2+ 2+ current by extruding 3 Na+ ions and importing 2 K+ ions on each
Ca -Induced Ca Release 322
cycle, and the Na+/Ca2+ exchanger (NCX), which mostly operates
Force 323 in the Ca2+-extrusion mode and yields a net inward charge move-
ment when 1 Ca2+ ion is exchanged for 3 Na+ ions (Figure 32-2, D).
Energetics 323
The plasma membrane Ca2+ ATPase (PMCA) also extrudes Ca2+
Cell Signaling 324 from the cell and finely regulates cytosolic Ca2+ levels.
Computational models of cardiac electrophysiology and the
Models of Cardiac Disease 326
close interplay between model development and experimentation
enabled us to achieve a quantitative understanding of the integra-
tive behavior of the cardiac myocyte in health and disease (see
Section VI). We will present the evolution of computational
electrophysiological models from the early formulations to the
Computational Models most recent comprehensive descriptions integrating Ca2+ and
of the Ventricular Myocyte other signaling pathways, force, and metabolism.
The Ventricular Action Potential

Early Canonical Models of the Cardiac AP
The action potential (AP) is a transient depolarization of the cell
membrane that emerges from the dynamic behavior of a diverse Hodgkin and Huxley laid the foundation for the use of integrative
population of membrane ion channels (depicted in Figure 32-1). models in biology describing both voltage clamp measurements
A prototypical ventricular myocyte AP is shown in Figure 32-2, A of membrane currents and an integrative AP model of the squid
(black). The AP exhibits a steep upstroke, followed by a sustained giant axon.3,4 Their modeling approach, which postulated that
slowly decaying plateau phase, which eventually gives way to gating mechanisms regulated membrane permeability, whereby
repolarization. Below the AP are shown the associated depolar- distinct entities (i.e., gates) controlled the flux of both Na+ and
izing currents, which are carried by inward Na+ and Ca2+ cur- K+ ions, has served to this day as a paradigm for a quantitative
rents. Under physiological conditions, the Na+ current (INa) description of cell membrane excitability. In the early 1960s,
activates rapidly, producing the AP upstroke, and then inactivates Denis Noble presented the first computational models of a
completely (Figure 32-2, B, red). The L-type Ca2+ current (ICaL) cardiac myocyte5,6 and addressed the issue of whether Na+ and
inactivates more slowly (Figure 32-2, B, blue), and less completely, K+ current descriptions similar to those of Hodgkin and Huxley
allowing for the inward Ca2+ current to maintain the plateau could be employed to account for the long AP plateau in Purkinje
phase of the AP. The influx of Ca2+ via ICaL triggers the release fibers. Modifications to Na+ and K+ current kinetics were able to
of Ca2+ from the sarcoplasmic reticulum (SR), the subcellular generate a plateau, which was supported by the inward Na+
organelle that stores and releases the majority of Ca2+ during each current. It is now known that Ca2+ currents, which had not yet
heartbeat.1 This event is known as Ca2+-induced Ca2+ release been discovered and were later introduced in the McAllister-
(CICR) and will be discussed in Section II. The SR then actively Noble-Tsien model,7 mainly support the AP plateau. Beeler and
re-sequesters Ca2+ via the SR Ca2+ adenosine triphosphatase Reuter extended the latter model to develop the first computa-
(ATPase) (SERCA), which is the primary mechanism removing tional ventricular myocyte model.8 Their work introduced a
Ca2+ from the cytosol to allow relaxation in between heartbeats.2 time-varying intracellular Ca2+ concentration and reinforced the
Throughout this process, Ca2+ is buffered by Ca2+-binding pro- idea that a balance between K+ and Ca2+ currents maintains the
teins such as calmodulin (CaM) and troponin. The rise in cyto- AP plateau, and repolarization is regulated by activation of K+
solic Ca2+ (Figure 32-2, A, blue) ultimately leads to cell contraction and inactivation of Ca2+ currents. Aspects of intracellular Ca2+
(see Section III). The intracellular Ca2+ signal also feeds back on handling were introduced in the DiFrancesco-Noble Purkinje
the L-type Ca2+ channel (LTCC), mediating inactivation of the cell,9 which constituted the basis of all subsequent models of the
current, and therefore plays a role in influencing the AP shape. cardiac myocyte. One of the landmarks in the development of
Furthermore, intracellular Ca2+ regulates a variety of processes models of ventricular electrophysiology was the formulation of
including mitochondrial adenosine triphosphate (ATP) produc- the dynamic Luo-Rudy ventricular cell model.10,11 Important fea-
tion (see Section IV), intracellular signaling (presented in Section tures of this model were the inclusion of the intracellular SR
V), and gene expression. compartment, time-varying intracellular ion concentrations, and
Various types of K+ channels (Figure 32-2, C) drive cell mem- ion pumps and exchangers. The continued interactive iteration
brane repolarization. The transient outward currents (Ito and between experiments and simulations has led to improved mech-
IClCa), carried by K+ and Cl− respectively, determine the notch that anistic insights into cardiac myocyte electrophysiology, as was
follows the upstroke, and components of the delayed rectifier K+ recently reviewed.12 Table 32-1 summarizes the existing models
current (IKr and IKs) contribute to AP repolarization. The inward of the ventricular myocyte, based on previous classifications.12-14
319
Ca 2K INa, INaL
PLM
Sarcolemma ATP NCX ATP
Sub-Sarcolemma
Na
Ca 3Na
Cleft
Ca
RyR
Ca
-AR
ICa Ca CaMKII
ATP
PLB
SR
G
CaM
AC
cAMP
Ca
Ca
PKA
Myofilaments
Bulk Cytosol
Cl K K K K K
IClCa IKs IKr Ito IKur IK1

Figure 32-1. Schematic representation of the processes described in ECC models. Em homeostasis is regulated by a diverse population of ionic currents: INa: fast Na+ current;
INaL: late Na+ current; ICaL: L-type Ca2+ current; IK1: inward rectifier K+ current; IKur: ultra-rapid delayed rectifier K+ current; Ito: transient outward K+ current; IKr: rapidly activating
delayed rectifier K+ current; IKs: slowly activating delayed rectifier K+ current; and IClCa: Ca2+-activated Cl− current. The depicted compartmentation for Ca2+ (and Na+) signaling,
including a cleft subspace for SR Ca2+ release and a submembrane compartment where Ca2+ raises higher versus bulk [Ca2+], was proposed by Shannon et al20 and was
applied by Mahajan et al,22 Grandi et al,36 and Li et al31 in rabbit, human, and mouse models, respectively. Ca2+ signaling interacts with CaMKII and PKA signaling pathways,
as illustrated (modeled) in Figure 32-4.
500 0 0
A B
30
-1
400
-100
0 ICa -2
ICa (A/F)
[Ca]i (nM)
INa (A/F)
Em (mV)
300 5 ms
-200 -3
–30
200 -4
–60 -300
-5
100 INa
–90 -400 -6
0 100 200 300 400
Time (ms)
C 8 D 0.4 INa/K
1.5
IK1
0.2
INa/K or INaCax (A/F)
IK1, IK1, IKs, (A/F)
Ito (A/F)
1.0 -0.0
4
-0.2
IKs 0.5
Ito INaCax
-0.4
IKr
0 0.0 -0.6
Figure 32-2. (A) Human ventricular AP and CaT simulated with the Grandi-Pasqualini-Bers model.36 (B) Na+ and Ca2+ currents (inset shows the different activation and decay
times), (C) K+ currents, and (D) Na+/K+ pump and Na+/Ca2+ exchange currents during an AP.
Models of the Ventricular Action Potential in Health and Disease 321
Table 32-1 Models of Ventricular Myocyte Electrophysiology
Model
8
Species Parent Model Comments 32
Beeler and Reuter 1977 Mammalian First ventricular model
Matsuoka et al 200356 Mammalian Integration of a contraction model
68 56
Matsuoka et al 2004 Mammalian Matsuoka et al 2003 Description of ATP metabolism
Luo and Rudy 199115 Guinea pig Beeler and Reuter 19778 First guinea pig ventricular model17
Noble et al 199117 Guinea pig Hilgemann and Noble 1987104 First guinea pig ventricular model15 from parent rabbit atrial
model
Nordin 199316 Guinea pig DiFrancesco and Noble 19859 Updated Ca2+ cycling and compartmentation
11 15
Luo and Rudy 1994 Guinea pig Luo and Rudy 1991 Inclusion of Ca2+ dynamics and updated currents
105 11
Zeng et al 1995 Guinea pig Luo and Rudy 1994 Introduction of slowly and rapidly activating IK components
Jafri et al 199843 Guinea pig Luo and Rudy 199411 Mechanistic Ca2+ dynamics, dyadic space, LTCC mode switching
55 17
Noble et al 1998 Guinea pig Noble et al 1991 Introduction of the dyadic space for Ca2+ and stretch-dependent
processes
Rice et al 200061 Guinea pig Jafri et al 199843 Inclusion of a contraction model
Faber and Rudy 200076 Guinea pig Luo and Rudy 199411 Incomplete deactivation of IKs and Na+ accumulation at fast
pacing rates
Crampin and Smith 200698 Guinea pig Luo and Rudy 199411 Inclusion of pH regulation
71 43
Cortassa et al 2006 Guinea pig Jafri et al 1998 Inclusion of contraction and mitochondrial bioenergetics
Pandit et al 200118 Rat Demir et al 1994106 From parent sinoatrial node model
19 18
Pandit et al 2003 Rat Pandit et al 2001 With diabetes
Saucerman et al 200374 Rat Puglisi and Bers 200121 First β-adrenergic signaling formulation
Pasek et al 200651 Rat Pandit et al 200118 Inclusion of T-tubule system
60 18
Niederer and Smith 2007 Rat Pandit et al 2001 Inclusion of contraction model
Bondarenko et al 200428 Mouse First mouse ventricular model
30 28
Li et al 2010 Mouse Bondarenko et al 2004 Complete refit of mouse model from mouse data
Li et al 201231 Mouse Li et al 201030 Study of SERCA-KO and ATP consumption by ion transport
Shannon et al 200420
Yang and Saucerman 201279 Mouse Bondarenko et al 200428 Simplified CICR and embedded β-adrenergic signaling to study
Greenstein et al 200647 the impact of PLM phosphorylation on CaT
Puglisi and Bers 200121 Rabbit Luo and Rudy 199411 First rabbit ventricular model with user-friendly interface
20 21
Shannon et al 2004 Rabbit Puglisi and Bers 2001 New Ca2+ dynamics formulation, submembrane Ca2+ space
86 20
Saucerman and Bers 2008 Rabbit Shannon et al 2004 Incorporation of CaM, CaMKII, CaN modules
22 20
Mahajan et al 2008 Rabbit Shannon et al 2004 Markov formulation for ICaL, and Ca2+ cycling model for the study
of APD and calcium alternans at rapid heart rates
Soltis and Saucerman 201078 Rabbit Shannon et al 200420 CaMKII and PKA pathways merged
Saucerman and Bers 200886
Morotti et al 201296 Rabbit Shannon et al 200420 Reparameterization of Mahajan et al22 model of ICaL to account for
Ba2+-dependent inactivation
Winslow et al 199923 Canine Jafri et al 199843 First canine ventricular model
25
Greenstein et al 2000 Canine Winslow et al 199923 Markov formulation for transient outward currents
25
Greenstein and Winslow Canine Greenstein et al 2000 Stochastic local control of SR Ca2+ release
200245
Cabo and Boyden 200324 Canine Luo and Rudy 199411 Normal and infarcted hearts
26
Hund and Rudy 2004 Canine Luo and Rudy 199411 Inclusion of CaMKII regulation of ECC
66 23
Michailova et al 2004 Canine Winslow et al 1999 Modeling of metabolism (i.e., Ca2+/Mg2+ buffering, ATP, ADP,
MgATP regulation of NKA and PMCA)
Greenstein et al 200647 Canine Greenstein et al 200025 Simplified CICR from Hinch et al 200446
82 11
Livshitz and Rudy 2007 Canine Luo and Rudy 1994 New mechanistic SR Ca2+ release current
Hund and Rudy 200426
Decker et al 200927 Canine Hund and Rudy 200426 Simulation of rate-dependent phenomena
Continued
Table 32-1 Models of Ventricular Myocyte Electrophysiology—cont’d
Model Species Parent Model Comments

87
Hashambhoy et al 2009 Canine Greenstein and Winslow CaMKII regulation and LTCC phosphorylation
200245
Hashambhoy et al 201088 Canine Greenstein and Winslow RyR phosphorylation by CaMKII
200245
Hashambhoy et al 200987
Heijman et al 201180 Canine Luo and Rudy 199411 New β-adrenergic signaling
11
Priebe and Beuckelmann Human Luo and Rudy 1994 First human ventricular model. Introduced formulations for
199832 normal and failing hearts
Iyer et al 200435 Human Winslow et al 199923 Markov formulations for INa, Ito, IKr, and ICaL
33
tenTusscher et al 2004 Human Priebe and Beuckelmann Updated currents, transmural variation
199832
tenTusscher & Panfilov 200634 Human ten Tusscher et al 200433 Updated currents, cleft subspace
Grandi et al 2010 36
Human Shannon et al 2004 20
Updated K+ currents and Ca2+ handling, Na+ accumulation
107 36
Carro et al 2011 Human Grandi et al 2010 Improved restitution properties
37 27
O’Hara & Rudy 2011 Human Decker et al 2009 Increased human-specific model accuracy from human data
Species-Specific and Human Models K+, and NCX currents from undiseased human ventricle, and
included the effects of Ca2+/CaM-dependent protein kinase II
Although most of the early models generically integrated experi- (CaMKII) on ionic currents and Ca2+ cycling. This model also
mental data from mammalian hearts (mostly guinea pig11,15-17), by reproduced early afterdepolarizations (EADs) and alternans.
the mid-1990s, electrophysiological studies had shown species Species-specific models provide a useful platform for investi-
differences in AP waveforms and ionic currents (e.g., mouse and gating species differences (e.g., with respect to arrhythmia
rat APs lack a plateau phase), but exhibited rapid repolarization and drug response). Comparison of AP repolarization, rate-
and very short AP duration (APD) if compared with the promi- dependent behavior, and drug response in human, dog, and
nent plateau phase and long AP seen in humans, rabbit, guinea guinea pig demonstrated major species differences and suggested
pig, and dog. Thus, the emphasis shifted to more detailed models that great caution should be taken when attempts are made to
on the basis of data obtained from isolated cells from one particu- extrapolate results from nonhuman species to human cellular
lar species (see Table 32-1). Rat ventricular cell models have been electrophysiology.38
developed,18,19 as well as rabbit20-22 and dog.23-27 Mouse models28-31
are becoming increasingly important, as genetic manipulation in
mice has proven to be a powerful tool for studying the physio-
logical effects of gene mutations, knockouts, and overexpression. Ca2+-Induced Ca2+ Release
The wealth of experimental data that can be measured in these
mouse models cannot be obtained in humans, but integrative The opening of LTCCs upon membrane potential (Em) depolar-
models provide a systematic framework to make inferences con- ization allows Ca2+ to flow down its electrochemical gradient into
cerning the effects expected in other species and in the clinical the dyadic subspace and trigger the opening of the ryanodine
setting (see caveat later). receptors (RyRs) that release Ca2+ from the SR via CICR. At
The first human ventricular cell model was published by more negative Em, fewer Ca2+ channels open, but the unitary flux
Priebe and Beuckelmann32 and was largely based on the dynamic is larger (because of the larger driving force), causing more effi-
Luo-Rudy model,11 in which formulations for the major ionic cient Ca2+-induced Ca2+ release and higher ECC gain (ratio of
currents were adjusted to the data (limited at that time) available Ca2+ released from the SR to the Ca2+ trigger via LTCCs). Gain
for human ventricular cells. The tenTusscher et al33 model and at negative Em is high, in part because a single Ca2+ channel
an updated version34 included many reformulated currents and opening may trigger junctional release, but as more channels are
recapitulated several electrophysiological phenomena. The Iyer recruited (e.g., approaching 0 mV), greater trigger redundancy
et al35 model addressed whole-cell Ca2+ homeostasis carefully. of LTCC opening (and a consequently large denominator in the
The most relevant ionic currents were based almost entirely on gain equation) is seen per release unit.39 At increasingly positive
data from human channels expressed in non-myocytes and for- Em (where unitary currents are small), multiple openings are
mulated with Markovian chains, which made this model much needed to ensure RyR opening, which results in lower gain. The
more complex than those previously described. The Grandi- Ca2+ release flux is a smooth, continuous function of trigger
Pasqualini-Bers model36 included K+ current reformulations influx—a behavior observed originally by Fabiato40 and known as
using undiseased human data, provided an accurate description graded Ca2+ release.
of Ca2+ and Na+ handling in the human ventricular myocyte, and Whereas the dynamic Luo-Rudy model11 generated APs using
accurately reproduced diverse aspects of excitation-contraction detailed kinetic descriptions of membrane currents, the Ca2+ sub-
coupling (ECC), including the contribution of various K+ cur- system was represented by a phenomenological model mimicking
rents to repolarization reserve, the phenomenon of reverse-rate the process of CICR, but unable to capture the biophysical details
dependence of APD upon K+ current block, and feedback on the involved. Many advances have resulted from improved descrip-
effects of changes in Na+ levels on APD and Ca2+ handling. The tions of intracellular Ca2+ handling (see recent reviews41,42). In
O’Hara and Rudy model37 included new measurements for ICaL, 1998, Jafri et al incorporated in a model of the guinea pig
ventricular myocyte43 mechanistic Markov models for both with models of myofilament activation and crossbridge interac-
32
LTCCs and RyRs and a restricted subspace (a single compart- tions. However, only a few groups have developed detailed
ment representing the total volume of all dyads) into which all systems models of cooperative myofilament activation, in part
Ca2+ fluxes through these channels are directed and [Ca2+] because of the complexity of spatially explicit models and the
increases faster and higher than measured in the bulk cytosolic theoretical shortcomings of spatially compressed models.52
compartment. Winslow et al23 adapted this Ca2+ subsystem in a The first ventricular myocyte model to couple myocyte ionic
canine model. These “common pool” models, whereby sarcolem- currents and Ca2+ handling to myofilament interactions was pub-
mal Ca2+ influx enters the same Ca2+ pool into which SR Ca2+ is lished by Michailova et al in 199253 and was extended in 1997.54
released and by which it is regulated, cannot reproduce both high Noble et al55 studied length- and tension-dependent changes in
gain and graded release.44 Most models have circumvented this mechanical and electrophysiological processes by incorporating
problem by introducing a dependence of the SR Ca2+ flux on in the earlier Noble et al model17 a description of force genera-
LTCC influx or Em, which removes the positive feedback effect tion. Matsuoka et al56 included the Negroni-Lascano myofila-
inherent to common pool models, resulting in nonphysiological ment model57 and recapitulated several aspects of ECC. Niederer
regenerative, all-or-none rather than graded, Ca2+ release. and Smith58 integrated the Hunter-McCulloch-ter Keurs myo-
Graded release arises from local stochastic interactions filament model59 into the Pandit et al rat ventricular cell model,18
between LTCCs and RyRs in thousands of Ca2+-release units; this and this framework was subsequently used to evaluate proposed
process is called local control of Ca2+ release. Several computa- mechanisms underlying the slow force response to stretch.60
tional models have been developed to investigate properties of The Rice-Jafri-Winslow model61 incorporated into the Jafri
local Ca2+ release, but only in 2002, Greenstein and Winslow45 et al43 guinea pig model a formulation of contraction62 that
developed a comprehensive model of the ventricular myocyte included a phenomenological representation of cooperative inter-
based on the theory of local control of SR Ca2+ release, which action between neighboring troponin/tropomyosin units. The
recapitulated several behaviors from the single channel to the model was used to examine aspects of short-term interval–force
whole-cell level. A simplified version of local control of CICR relationships in cardiac muscle, and showed that cooperative prop-
describing the (deterministic) ensemble behavior of release units erties of the myofilaments profoundly affect the developed force.
was developed by Hinch et al46 and was incorporated in a less Rice et al have expanded this formulation in a new approximate
computationally expensive ECC model.47 model of myofilament activation and crossbridge cycling (Figure
Shannon et al20 introduced a subsarcolemmal Ca2+ compart- 32-3, A) that captured many experimentally observed features of
ment (see Figure 32-1) that was based on experiments suggesting cooperative length-dependent thin-filament activation.63 The
that the Na+/Ca2+ exchanger senses elevated Ca2+ levels (com- contraction model was then integrated into the Shannon et al20
pared with the bulk [Ca2+]i), and used a model of RyR that rabbit ECC framework, showing its suitability for coupling with
included regulation by both cytosolic and luminal [Ca2+]. The existing models of electrophysiology and Ca2+ handling, and the
model reproduced the steep relationship between SR Ca2+ load ability to recapitulate common experimental characterization such
and release, and conferred importance to the (partial) contribu- as cell shortening (Figure 32-3, B). The McCulloch group coupled
tion of SR depletion in the regulation of SR Ca2+-release dynam- this model to the canine ECC model64 and found that heterogene-
ics and termination. This Ca2+ subsystem has been adapted in the ities in ion channel and Ca2+ handling protein expression between
Grandi-Pasqualini-Bers36 human ventricular myocyte model. epicardial, M-cell, and endocardial myocytes in the dog were suf-
Localization and compartmentation of ion channels and trans- ficient to explain most but not all differences between unloaded
porters have proven crucial for proper ECC (especially in the shortening twitches measured in canine myocytes, and indicated
dyadic space). It has been suggested that NCX and NKA consti- that variations in crossbridge function may be present.
tute a molecular complex with ankyrin-B and the inositol
1,4,5-phosphate-operated Ca2+-releasing channel in the T-tubules
(distinct from the LCC-RyR complex) that may play an important
role in the regulation of local Na+ and Ca2+ concentrations and may Energetics
possibly modulate CICR. In fact, alterations in ankyrin-B levels
can affect local [Na+]i and cellular and SR Ca2+ cycling48,49 and are Oxidative phosphorylation is the main source of energy for
associated with a broad spectrum of cardiac consequences.50 metabolic/contractile works in cardiac myocytes. Mitochondria
Other local Ca2+ (and Na+) signals may be relevant to ECC. provide the ATP needed for contractile function and sarcolem-
For example, Pasek et al51 proposed a model of the guinea pig mal and sarcoplasmic ion transport, which is responsible for
ventricular myocyte that includes a diffusive T-tubule system myocyte electrical activity. Energetics drives ion transport pro-
with a heterogeneous distribution of ion channels between cesses by means of their dependence on the proton motive force
tubular and surface membranes, and predicted that Ca2+ deple- and phosphorylation potential, as well as by direct transport
tion and K+ accumulation in the T-tubule during APs impact across the mitochondrial inner membrane (it has been shown that
ECC. Also, the hypothesis that juxtaposition of mitochondria and mitochondrial Ca2+ transport can influence Ca2+ signals in the
dyads may give rise to an additional Ca2+-signaling microdomain cytoplasm65). On the other hand, energy demands change in
regulating mitochondrial ATP production and modulating dyadic response to activation of myofilaments and, to a lesser extent,
Ca2+ concentrations may guide future modeling efforts. NKA, SERCA, and PMCA.31 Coupling of mitochondrial ener-
getics to ECC models is therefore needed to investigate the key
role of energetics in modulating myocyte mechanical activity and
ion concentration gradients, especially during impaired meta-
Force bolic states in pathologic conditions, such as ischemia and heart
failure (HF).
Ca2+ released in the dyadic subspace during CICR is then free to To link metabolism and Ca2+ handling in (canine) ventricu-
diffuse out into the myoplasm and throughout the sarcomere, lar myocytes, Michailova et al66,67 extended the Winslow
where it binds to the Ca2+-binding protein troponin in the myo- et al model23 by incorporating descriptions of Ca2+ and Mg2+
filaments and initiates cell contraction. Not only does Ca2+ acti- buffering and transport by ATP and ADP, and MgATP regulation
vate the myofilaments, but also the Ca2+ transient (CaT) is of ion transporters (NKA, SERCA, and PMCA). Matsuoka
influenced when developed force is changed. Thus it is important et al68 coupled a model of mitochondrial metabolism69 to their
to couple models of cardiac electrophysiology and Ca2+ cycling electrophysiological model56 and simulated nicotinamide adenine
Simulation - rabbit
50
Regulatory Ca binding
affects activation only
(ratio set by SL alone) 0
Em(mV)
T Txb Nxb
–50
100 nM [Ca]
2% length
Kd:> K’d:
Knp(TCaTot )7.5 Knp(TCaTot)–7.5 Ca
TCa TxbCa –100
0 200 400 600 800
Pxb
time (ms)
Apparent Ca binding affects the Experimental data - rabbit
cytosolic Ca transient gxbT fappT
(ratio set by SL and XBPreR/XBPostR) gappT [Ca]i
40
hft
T Txb XBPostR XBPreR
Em(mV)
0 Contraction
Ca hbt Ca
Kd:> K’d:
100 nM [Ca]
2% length
–40
AP
TCa TxbCa
–80
0 200 400 600 800
A B time (ms)
Figure 32-3. A, The Rice et al model63 combines myofilament activation and crossbridge cycling in a coupled system. To implement Ca2+-based activation, it is assumed
that troponin and tropomyosin act as regulatory units that exist in a nonpermissive (N, preventing the formation of strongly bound crossbridges) or a permissive (P, allowing
transitions to strongly bound crossbridges state). The next transition is to the prerotated state, which is strongly bound with the head extended. The transition to the post-
rotated force-generating state represents isomerization to induce strain in the extensible neck region. B, The myofilament model developed by Rice et al63 is coupled to
the Shannon et al20 model of the rabbit ventricular myocyte. Results are shown for the combined model (top) along with experimental data in the rabbit at 37° C (bottom).
The labeled responses show APs, bulk myoplasmic CaTs, and cell shortening signals.
(Reproduced from Rice JJ, Wang F, Bers DM, et al: Approximate model of cooperative activation and crossbridge cycling in cardiac muscle using ordinary differential equations.
Biophys J 95:2368–2390, 2008.)
dinucleotide (NADH) and mithocondrial CaTs following an PKA

increase in workload. The O’Rourke group has extensively
studied cardiac energetics both experimentally and theoretically. β-Adrenergic control of cardiac function is initiated by
Cortassa et al70 developed the first integrative kinetic and ther- β-adrenergic receptor (βAR) coupling to stimulatory G protein
modynamic mathematical model of cardiac energy metabolism and subsequent stimulation of adenylyl cyclase, which synthesizes
that takes into account mitochondrial matrix- and membrane- the classical second messenger cyclic AMP (cAMP). Cyclic AMP
based processes such as tricarboxylic acid cycle, oxidative phos- in turn activates protein kinase A (PKA), which phosphorylates a
phorylation, and Ca2+ dynamics. This model was subsequently wide spectrum of target proteins (Figure 32-4, A). These include
integrated with models of the electrophysiological, Ca2+-han- LTCC, RyR, and phospholamban (PLB), which are important
dling,43 and force-generation61 subsystems of the cardiac myocyte players in the regulation of Ca2+ dynamics and transport; the
to study the complex dynamics of the response of mitochondrial myofilament protein troponin I (TnI); and phospholemman
bioenergetics to alterations in myocyte contractile and electrical (PLM), which regulates NKA. Sarcolemmal ion channels respon-
activity.71 The formulation explicitly incorporates cytoplasmic sible for IKs and IKur and the cystic fibrosis transmembrane
ATP-consuming processes associated with force generation and conductance regulator (CFTR) are also targets of PKA phos-
ion transport, as well as the creatine kinase reaction. Changes in phorylation. To probe the functional interactions of these altera-
the electrical and contractile activity of the myocyte are coupled tions in the integrative cellular environment, Saucerman et al74
to mitochondrial energetics through the ATP, Ca2+, and Na+ were the first to develop and validate a functionally integrated
concentrations in the myoplasmic and mitochondrial matrix systems coupling of a model of β-adrenergic signaling with
compartments. Extensions of this model have been used to inves- models of Ca2+ handling and electrophysiology (study21 adapted
tigate the mechanisms of oxidative stress.72,73 to simulate the rat ventricular myocyte with experimental Ca2+
handling and K+ current data). The model was also used for a
systematic sensitivity analysis,75 which supported the conclusion
that PLB phosphoregulation is the primary mechanism for
Cell Signaling increased SR load and Ca2+ relaxation rate during β-adrenergic
stimulation, while both PLB and LTCC phosphorylation con-
Cardiac myocyte electrical activity, CICR, contraction, metabo- tribute to increased systolic Ca2+. The Saucerman et al74
lism, and gene regulation are subject to modulation by a number β-adrenergic signaling model was recently embedded in the
of signaling pathways, which involve cascades of signaling mol- Faber and Rudy guinea pig model76 to study the role of βAR
ecules resulting in posttranslational modifications (e.g., phos- agonists in LQTS,77 in the Shannon et al rabbit model20 to study
phorylation) of target proteins. The most widely studied signaling the synergy between PKA and CaMKII pathways,78 and in the
pathways in cardiac myocytes are the β-adrenergic and CaMKII Yang and Saucerman mouse model79 to study the role of PLM
pathways. phosphorylation in the fight-or-flight response. Heijman et al80
32
DADs? EADs? INa Ito IK1 IKs ICFTR
ICa
3Na
PLM
ATP
sarcolemma GPCR
NCX
PKA
RyR CaMKII C
Ca
CaM
SR
[Ca]i
Ca Ca
PLB
Heart Failure
ATP
Ca Na LTCCp
Ca Ca
100
dyadic cleft LCC LCC NCX
PP2A PKA
PKA PP1
80 RyRp
% Activity/Phosphorylation
CAMKII
Ca Na
CAMKII CaM Ca Ca CaM
PP2A PP1 Na
CamKII OE
60
PKA INa)f,L)
RyR
K Dyadic CaMKII
Ito 40
Sub-sarcolemma
Ca CaM
sarcoplasmic Ca
cytosol
reticulum SERCA K
CAMKII IKS 20
PLB PKA PP1 PLBp
PKA Tnl
PP1 CI
ICFTR
0
I1 PP2A 0 10 20 30
B C Time (s)
Figure 32-4. A, PKA and CaMKII influence ECC by affecting several electrical and Ca2+-handling proteins, including PLB, RyR, and LTCC. In addition, INa and K+ currents (e.g.,
Ito and IK1) are regulated by CaMKII. PKA also targets sarcolemmal ion channels, such as IKs and ICFTR, and the myofilament protein TnI. By exerting multiple effects on these
numerous targets, CaMKII can simultaneously favor heart failure and arrhythmias. PKA can exacerbate this via positive Ca2+ feedback. B, Soltis and Saucerman78 model
schematic. CaMKII is activated by Ca2+/CaM binding in the dyadic cleft, subsarcolemma, and cytosolic compartments. Active CaMKII phosphorylates LCCs, RyRs, and PLB.
CaMKII-dependent alterations to INa and Ito were included. PKA phosphorylates LCCs, RyRs, PLB, Inhibitor-1, TnI, IKs, and ICFTR, and phosphatases 1 and 2A oppose phosphoryla-
tion by either kinase. C, Time course and frequency dependence of CaMKII phosphorylation: dyadic CaMKII activity and phosphorylation profiles during CaMKII overexpres-
sion (6×) simulation at low (0.5 Hz) and higher (2 Hz) pacing frequencies.78
(From Soltis AR, Saucerman JJ: Synergy between CaMKII substrates and beta-adrenergic signaling in regulation of cardiac myocyte Ca2+ handling. Biophys J 99:2038–2047, 2010.)
presented a computational compartmental model of βAR signal- CaMKII

ing and its effects on canine ventricular myocyte electrophysiol-
ogy. The model included localized signaling domains, β1 and β2 CaMKII is a multifunctional protein kinase expressed abundantly
receptor isoforms, and PKA (and CaMKII) effects on a wide in the heart, which activates in response to increasing [Ca2+] (e.g.,
range of substrates affecting whole-cell electrophysiology and during systole). In this event, intracellular CaM binds up to four
CaT. The model also showed how activation of specific cAMP Ca2+ ions and the Ca2+/CaM complex binds to the regulatory
domains by different receptor isoforms allows for specific control domain of CaMKII and displaces the autoinhibitory domain on
of AP and CaT properties, and confirmed the synergistic nature CaMKII, thereby activating the enzyme with half maximal activa-
of CaMKII and PKA crosstalk. tion at [Ca2+]i of 0.1 to 10 µM. After this Ca2+/CaM-dependent
activation, CaMKII can lock itself into an activated state upon guanine nucleotide exchange factor Epac (independent of PKA)
autophosphorylation of Thr287 on the autoinhibitory segment. to activate a CaMKII- and RyR-dependent SR Ca2+ leak.89 This
Autophosphorylation can maintain CaMKII active even after interaction has been reported to mediate βAR-induced cardiac
[Ca2+]i has declined (e.g., during diastole), when Ca2+/CaM has hypertrophy.90 Curran et al91 showed that βAR activation of
dissociated from its binding region (the so-called autonomous CaMKII, which is Ca2+- and PKA-independent, is also cAMP-
state). CaMKII phosphorylates numerous target proteins involved independent.92 A systems model incorporating these pathways
in Ca2+ influx, release from and uptake into the SR, and sarcolem- could provide a quantitative framework for discriminating canon-
mal Na+ and K+ channels (see Figure 32-4, A). This in turn can ical and novel mechanisms of crosstalk between βAR and CaMKII
influence myocyte Ca2+ regulation and confers further Ca2+ signaling in myocytes.
dependence to electrophysiological effects.81 Recent models have
incorporated elements of the CaMKII signaling cascade to study
the role of CaMKII in regulating cardiomyocyte contractility and
excitability. Hund and Rudy26 for the first time integrated Models of Cardiac Disease
CaMKII signaling in a model of canine ECC, including its effects
on Ca2+ handling, and postulated its role in the rate dependence Heart Failure
of CaT. Their model was extended by Livshitz and Rudy,82 who
reformulated SR Ca2+ release and linked CaMKII activity to an HF is a rapidly growing health problem in the United States.
increased propensity for T wave alternans, and was further About half of HF deaths are due to arrhythmias (and half to pump
updated by Christensen et al83 with the addition of an oxidized failure). Alterations in myocyte ion currents, Ca2+ handling, con-
(active) state of CaMKII accounting for the oxidation-induced tractile function, and their neurohormonal regulation, accompa-
activation of CaMKII that occurs during myocardial infarction. nied by ventricular hypertrophy and structural remodeling, all
Grandi et al84 have investigated in silico the effects of CaMKII contribute to the HF phenotype. Quantitative systems models
overexpression on cardiac myocyte ionic currents and APs in the that integrate across interacting biochemical and biophysical
comprehensive Shannon et al20 ECC model. By incorporating the functions (and multiscale models) have proved essential for
modulatory effects of CaMKII on INa, ICaL, and Ito, it was shown a mechanistic understanding of this complex clinical
that APs from rabbit myocytes overexpressing CaMKII are syndrome.23,32,93,94
shorter compared with control, as measured experimentally. CaMKII is more expressed and more active in HF, and has
CaMKII-induced Na+ channel loss of function was predicted to been causally linked to the initiation of triggered arrhythmias via
reduce AP rate of rise and conduction velocity, especially at fast both EADs and delayed afterdepolarizations (DADs95). For simu-
heart rates (like in the Brugada syndrome), whereas increases in lating the effects of CaMKII overexpression on sarcolemmal ion
the late component could potentially resemble a long QT-3 phe- channels (see earlier), we have shown that when Ito expression is
notype at slow rates. Wagner et al85 incorporated the same low (e.g., in HF or endocardial myocytes), the effects on INa and
CaMKII overexpression effects into a mouse model and showed ICaL would be predominant, thus potentially leading to a LQT3-
species specificity and differences between acute and chronic like phenotype (Figure 32-5, A, middle and right) and possibly
CaMKII effects. The Saucerman and Bers model86 was the first EAD-induced triggered arrhythmias. In fact, prolongation of the
to incorporate and validate compartmental models of CaM, AP plateau may slow deactivation of ICaL, which depolarizes the
CaMKII, and phosphatases (including calcineurin [CaN]), based membrane and generates EADs.96 In addition, experimental and
on measured concentrations, binding kinetics, and target phos- modeling data indicate that phosphorylation events promoting
phorylation, into the Shannon et al20 rabbit model of ECC. mode 2 gating of ICaL contribute to EADs and ventricular arrhyth-
Model simulations showed that different affinities of CaM for mias.97 Heterogeneity of transmural ventricular repolarization has
CaMKII and CaN are important for determining their sensitivity been linked to a variety of arrhythmic manifestations. A certain
to local Ca2+ signals. CaM is highly activated in the dyadic cleft degree of transmural dispersion of repolarization is normal
from beat to beat with no significant “memory” of previous beats, (Figure 32-5, B, left) and may be attributed in part to differential
but less than 1 in 105 free bulk cytosolic CaM molecules are expression of Ito (where there is higher Ito in myocytes at the epi-
activated during that same beat. They predicted that CaMKIIδC, cardial side). If CaMKII prolongs APDs in the endocardium (see
often described as a cytosolic isoform, is relatively insensitive to Figure 32-5, A, middle and right) and shortens APDs in the epicar-
cytosolic Ca2+ because of its relatively low CaM affinity, whereas dium (Figure 32-5, A, left), this could amplify transmural disper-
the higher affinity of CaN allows for its gradual activation. When sion of repolarization (Figure 32-5, B, right), thus predisposing to
targeted to the cleft, CaMKII exhibits dynamic, frequency- the development of reentrant arrhythmias. When incorporating
dependent responses to Ca2+ (Figure 32-4, C). Hashambhoy et into the ECC model the full spectrum of HF-induced changes,
al87 modeled dynamic CaMKII phosphorylation of LTCCs within such as CaMKII-mediated RyR-Ca2+ sensitization, enhanced SR
the local control model45 and postulated that a CaMKII- Ca2+ leak, and decreased IK1 and Ito (besides the changes in INa, ICaL,
dependent shift between gating modes underlies ICa facilitation. and Ito gating), coupled with other cellular changes, such as
Subsequently, RyR regulation by CaMKII was included in this increased INCX, we observed DAD-induced triggered arrhythmias
framework.88 Soltis and Saucerman78 integrated dynamic upon βAR stimulation (Figure 32-5, C). DADs are a consequence
CaMKII-dependent regulation of LTCC, PLB, and RyR (Figure of a βAR-induced increase in [Ca2+]SR that activates SR Ca2+
32-4, B, C) and identified CaMKII-dependent RyR hyperphos- release via hypersensitive RyR and causes inward INCX, which
phorylation as a proarrhythmogenic trigger. By linking the depolarizes the membrane toward threshold for a triggered AP.
CaMKII and PKA pathways to ECC, they showed that CaMKII Notably, this mechanism may also explain EADs.96
and PKA activities during β-adrenergic stimulation may syner-
gistically facilitate inotropic responses, as βAR stimulation
can activate CaMKII via PKA-dependent enhancement of ICa Ischemia
and SR Ca2+ uptake and consequent increase in SR Ca2+
release and CaT amplitude (also shown by Heijman et al80), Regulation of metabolic and electrophysiological processes may
which served as the basis for a systematic analysis of the interac- become dysfunctional during ischemia-reperfusion, often result-
tion between βAR-PKA and CaM-CaMKII. Furthermore, there ing in alterations in ionic balance that may lead to lethal cardiac
is evidence for a Ca2+- and PKA-independent pathway by which arrhythmias and/or contractile failure. Increased concentrations
βAR cause CaMKII activation. Cyclic AMP can activate the of intracellular Na+ and Ca2+, increased extracellular K+, and
Ito 100% Ito 25% Ito 10%
32
(control/epi) (HF/endo) (HF/endo)
(mV)
0
-100
200 ms [1 Hz]
A Control
CaMKII
endo endo
epi epi
200 ms
B 78 ms 183 ms
HF and CaMKII-overexpression
Control Isoproterenol
(mV)
−100
Stimulation (1 Hz) Stimulation (1 Hz)
0.6 DADs
2
0.4
[Ca] i (µM)
1
0.2
0 0
C 1s
DADs
Figure 32-5. A, Simulated APs at 1 Hz in control and CaMKIIδC-overexpressing cardiac myocytes in the presence of 100% (left), 25% (middle), and 10% (right) Ito. When Ito
is fully expressed (epi or control), CaMKII shortens the AP, whereas the AP is prolonged by Ito downregulation (e.g., endo or HF). B, This could amplify transmural dispersion
of repolarization, which may facilitate reentry phenomena. C, Stimulation of steady-state AP and Ca2+ transients followed by a period of rest in digital HF cell (left) and with
isoproterenol (right) at 1 Hz. In isoproterenol, cessation of 1-Hz stimulation, leads to spontaneous SR Ca2+ release (right, lower panel), activating inward INCX, which depolarizes
the membrane-generating DADs (right, upper panel). DADs were not seen in HF (left). (B, Redrawn from Bers DM, Grandi E: Calcium/calmodulin-dependent kinase II regula-
tion of cardiac ion channels. J Cardiovasc Pharmacol 54:180–187, 2009. C, Redrawn from Bers DM, Grandi E: Calcium/calmodulin-dependent kinase II regulation of cardiac
ion channels. J Cardiovasc Pharmacol 54:180–187, 2009.)
decreased intracellular ATP and pH have been measured experi- interaction of protons with the contractile machinery, and recon-
mentally in ischemia. ciled these various contributions to understand the overall effects
Acidosis in cardiac myocytes is a major factor in the reduced of acidosis in the beating heart.
inotropy that occurs in the ischemic heart as a net result of Metabolic blockade causes activation of the ATP-sensitive K+
complex interactions between protons and a variety of intracel- current (IKATP), which can markedly shorten the cardiac AP and
lular processes. Crampin and Smith98 developed a dynamical was quantitatively investigated.99,100 Ch’en et al101 simulated the
model of pH regulation and ECC to predict the time courses of biochemical changes in cytosolic ATP, pH, and Ca2+ that occur
key ionic species during acidosis, in particular intracellular pH, during ischemia-reperfusion. The model produced NCX-
Na+, Ca2+, and contraction. They suggested that the most signifi- mediated Ca2+-overload arrhythmias and identified the electro-
cant effects are elevated Na+, inhibition of NCX, and the direct physiological effects of therapeutics such as Na+/H+ exchange
block. Michailova et al66,67 explored the role of free Mg2+, MgATP,

and MgADP in IKATP, ICaL, and [Ca2+]i in a canine cell model, and Summary
showed that either increases in free cytosolic Mg2+ (0.2 to 5 mM)
with fixed Mg-nucleotide concentrations, or decreases in the We have discussed how computational modeling has evolved to
ATP/ADP ratio with fixed total Mg2+, could activate IKATP and incorporate descriptions of ion channels and membrane trans-
systematically change APD, ICaL, and CaTs.66 More recently, porters and to integrate mechanistic models of the CICR process
Zhou et al102 expanded the comprehensive ECC-energetics (common pool vs. local control formulations) and models of force
model of Cortassa et al70 (see earlier) that includes realistic math- generation, mitochondrial ATP production and its regulation by
ematical representations of mitochondrial energetics to drive the Ca2+, and the coupling of electrophysiological models with sig-
changes in ATP/ADP ratio and integrates reactive oxygen species naling pathways. Finally, we have shown how cellular models of
(ROS)-induced ROS release processes to simulate the phenom- electrophysiology, cell signaling, and metabolism have been used
enon of oxidative stress–induced mitochondrial oscillations and to investigate the mechanisms underlying cardiac diseases includ-
their effects on whole cardiomyocyte function. This allowed ing heart failure and ischemia, with the ultimate goal of improv-
examination of the sequence of events that activate IKATP during ing treatment.103
oxidative stress and provided a new tool for examining how alter-
ations in mitochondrial energetic state will impact the electro-
physiology and electrical activities of the cardiac cell in both
health and disease. Notably, oxidative stress (e.g., due to Acknowledgments
mitochondria-derived ROS) causes a number of functional alter-
ations in key ECC targets, including kinases (e.g., CaMKII). We wish to thank Drs. Yuanfang Xie and Stefano Morotti for
Thus, future integrative modeling that incorporates those their critical reading of the manuscript. Sources of support:
changes can be a powerful tool in the study of synergy and cross- NHLBI Grants P01-HL080101 and R37-HL30077, and the
talk among these various pathways. Fondation Leducq (to DMB).
15. Luo CH, Rudy Y: A model of the ventricular 28. Bondarenko VE, Szigeti GP, Bett GC, et al: Com-
References cardiac action potential: Depolarization, repolar- puter model of action potential of mouse ven-
ization, and their interaction. Circ Res 68:1501– tricular myocytes. Am J Physiol Heart Circ
1. Bers DM: Excitation-Contraction Coupling and 1526, 1991. Physiol 287:H1378–H1403, 2004.
Cardiac Contractile Force, ed 2, Dordrecht, The 16. Nordin C: Computer model of membrane current 29. Li L, Louch WE, Niederer SA, et al: Calcium
Netherlands, 2001, Kluwer Academic Press. and intracellular Ca2+ flux in the isolated guinea dynamics in the ventricular myocytes of SERCA2
2. Bassani JW, Bassani RA, Bers DM: Relaxation in pig ventricular myocyte. Am J Physiol 265(6 Pt knockout mice: A modeling study. Biophys J
rabbit and rat cardiac cells: Species-dependent 2):H2117–H2136, 1993. 100:322–331, 2011.
differences in cellular mechanisms. J Physiol 17. Noble D, Noble SJ, Bett GC, et al: The role of 30. Li L, Niederer SA, Idigo W, et al: A mathematical
476:279–293, 1994. sodium-calcium exchange during the cardiac model of the murine ventricular myocyte: A data-
3. Hodgkin AL, Huxley AF: A quantitative descrip- action potential. Ann N Y Acad Sci 639:334–353, driven biophysically based approach applied to
tion of membrane current and its application to 1991. mice overexpressing the canine NCX isoform.
conduction and excitation in nerve. J Physiol 18. Pandit SV, Clark RB, Giles WR, et al: A mathe- Am J Physiol Heart Circ Physiol 299:H1045–
117:500–544, 1952. matical model of action potential heterogeneity in H1063, 2010.
4. Hodgkin AL, Huxley AF, Katz B: Measurement adult rat left ventricular myocytes. Biophys J 31. Li L, Louch WE, Niederer SA, et al: Sodium
of current-voltage relations in the membrane of 81:3029–3051, 2001. accumulation in SERCA knockout-induced heart
the giant axon of Loligo. J Physiol 116:424–448, 19. Pandit SV, Giles WR, Demir SS: A mathematical failure. Biophys J 102:2039–2048, 2012.
1952. model of the electrophysiological alterations in 32. Priebe L, Beuckelmann DJ: Simulation study of
5. Noble D: Cardiac action and pacemaker poten- rat ventricular myocytes in type-I diabetes. cellular electric properties in heart failure. Circ
tials based on the Hodgkin-Huxley equations. Biophys J 84(2 Pt 1):832–841, 2003. Res 82:1206–1223, 1998.
Nature 188:495–497, 1960. 20. Shannon TR, Wang F, Puglisi J, et al: A mathe- 33. ten Tusscher KH, Noble D, Noble PJ, et al: A
6. Noble D: A modification of the Hodgkin-Huxley matical treatment of integrated Ca dynamics model for human ventricular tissue. Am J Physiol
equations applicable to Purkinje fibre action and within the ventricular myocyte. Biophys J Heart Circ Physiol 286:H1573–H1589, 2004.
pace-maker potentials. J Physiol 160:317–352, 87:3351–3371, 2004. 34. ten Tusscher KH, Panfilov AV: Alternans and
1962. 21. Puglisi JL, Bers DM: LabHEART: an interactive spiral breakup in a human ventricular tissue
7. McAllister RE, Noble D, Tsien RW: Reconstruc- computer model of rabbit ventricular myocyte ion model. Am J Physiol Heart Circ Physiol
tion of the electrical activity of cardiac Purkinje channels and Ca transport. Am J Physiol Cell 291:H1088–H1100, 2006.
fibres. J Physiol 251:1–59, 1975. Physiol 281:C2049–C2060, 2001. 35. Iyer V, Mazhari R, Winslow RL: A computational
8. Beeler GW, Reuter H: Reconstruction of the 22. Mahajan A, Shiferaw Y, Sato D, et al: A rabbit model of the human left-ventricular epicardial
action potential of ventricular myocardial fibres. ventricular action potential model replicating myocyte. Biophys J 87:1507–1525, 2004.
J Physiol 268:177–210, 1977. cardiac dynamics at rapid heart rates. Biophys J 36. Grandi E, Pasqualini FS, Bers DM: A novel com-
9. DiFrancesco D, Noble D: A model of cardiac 94:392–410, 2008. putational model of the human ventricular action
electrical activity incorporating ionic pumps and 23. Winslow RL, Rice J, Jafri S, et al: Mechanisms of potential and Ca transient. J Mol Cell Cardiol
concentration changes. Philos Trans R Soc Lond altered excitation-contraction coupling in canine 48:112–121, 2010.
B Biol Sci 307:353–398, 1985. tachycardia-induced heart failure, II: Model 37. O’Hara T, Virag L, Varro A, et al: Simulation of
10. Luo CH, Rudy Y: A dynamic model of the cardiac studies. Circ Res 84:571–586, 1999. the undiseased human cardiac ventricular action
ventricular action potential. II. Afterdepolariza- 24. Cabo C, Boyden PA: Electrical remodeling of the potential: Model formulation and experimental
tions, triggered activity, and potentiation. Circ epicardial border zone in the canine infarcted validation. PLoS Comput Biol 7:1002061,
Res 74:1097–1113, 1994. heart: a computational analysis. Am J Physiol 2011.
11. Luo CH, Rudy Y: A dynamic model of the cardiac Heart Circ Physiol 284:H372–H384, 2003. 38. O’Hara T, Rudy Y: Quantitative comparison of
ventricular action potential. I. Simulations of 25. Greenstein JL, Wu R, Po S, et al: Role of the cardiac ventricular myocyte electrophysiology
ionic currents and concentration changes. Circ calcium-independent transient outward current I and response to drugs in human and nonhuman
Res 74:1071–1096, 1994. (to1) in shaping action potential morphology and species. Am J Physiol Heart Circ Physiol
12. Winslow RL, Cortassa S, O’Rourke B, et al: Inte- duration. Circ Res 87:1026–1033, 2000. 302:H1023–H1030, 2012.
grative modeling of the cardiac ventricular 26. Hund TJ, Rudy Y: Rate dependence and regula- 39. Altamirano J, Bers DM: Voltage dependence of
myocyte. Wiley Interdiscip Rev Syst Biol Med tion of action potential and calcium transient in a cardiac excitation-contraction coupling: Unitary
3:392–413, 2011. canine cardiac ventricular cell model. Circulation Ca2+ current amplitude and open channel prob-
13. Noble D, Garny A, Noble PJ: How the Hodgkin- 110:3168–3174, 2004. ability. Circ Res 101:590–597, 2007.
Huxley equations inspired the cardiac Physiome 27. Decker KF, Heijman J, Silva JR, et al: Properties 40. Fabiato A: Time and calcium dependence of acti-
Project. J Physiol 590:2613–2628, 2012. and ionic mechanisms of action potential adapta- vation and inactivation of calcium-induced release
14. Wilders R: Computer modelling of the sinoatrial tion, restitution, and accommodation in canine of calcium from the sarcoplasmic reticulum of a
node. Med Biol Eng Comput 45:189–207, epicardium. Am J Physiol Heart Circ Physiol skinned canine cardiac Purkinje cell. J Gen
2007. 296:H1017–H1026, 2009. Physiol 85:247–289, 1985.
41. Greenstein JL, Winslow RL: Integrative systems responses. Am J Physiol 276(5 Pt 2):H1734– 84. Grandi E, Puglisi JL, Wagner S, et al: Simulation
models of cardiac excitation-contraction cou- H1754, 1999. of Ca-calmodulin-dependent protein kinase II on
pling. Circ Res 108:70–84, 2011.
42. Williams GS, Smith GD, Sobie EA, et al: Models
of cardiac excitation-contraction coupling in ven-
63. Rice JJ, Wang F, Bers DM, et al: Approximate
model of cooperative activation and crossbridge
cycling in cardiac muscle using ordinary differen-
rabbit ventricular myocyte ion currents and action
potentials. Biophys J 93:3835–3847, 2007.
85. Wagner S, Hacker E, Grandi E, et al: Ca/
32
tricular myocytes. Math Biosci 226:1–15, 2010. tial equations. Biophys J 95:2368–2390, 2008. calmodulin kinase II differentially modulates
43. Jafri MS, Rice JJ, Winslow RL: Cardiac Ca2+ 64. Campbell SG, Flaim SN, Leem CH, et al: Mech- potassium currents. Circ Arrhythm Electro-
dynamics: The roles of ryanodine receptor adap- anisms of transmurally varying myocyte electrophysiol 2:285–294, 2009.
tation and sarcoplasmic reticulum load. Biophys J mechanics in an integrated computational model. 86. Saucerman JJ, Bers DM: Calmodulin mediates
74:1149–1168, 1998. Philos Transact A Math Phys Eng Sci 366:3361– differential sensitivity of CaMKII and calcineurin
44. Stern MD: Theory of excitation-contraction cou- 3380, 2008. to local Ca2+ in cardiac myocytes. Biophys J
pling in cardiac muscle. Biophys J 63:497–517, 65. Maack C, Cortassa S, Aon MA, et al: Elevated 95:4597–4612, 2008.
1992. cytosolic Na+ decreases mitochondrial Ca2+ 87. Hashambhoy YL, Winslow RL, Greenstein JL:
45. Greenstein JL, Winslow RL: An integrative uptake during excitation-contraction coupling CaMKII-induced shift in modal gating explains
model of the cardiac ventricular myocyte incor- and impairs energetic adaptation in cardiac myo- L-type Ca2+ current facilitation: a modeling study.
porating local control of Ca2+ release. Biophys J cytes. Circ Res 99:172–182, 2006. Biophys J 96:1770–1785, 2009.
83:2918–2945, 2002. 66. Michailova AP, Belik ME, McCulloch AD: Effects 88. Hashambhoy YL, Greenstein JL, Winslow RL:
46. Hinch R, Greenstein JL, Tanskanen AJ, et al: A of magnesium on cardiac excitation-contraction Role of CaMKII in RyR leak, EC coupling and
simplified local control model of calcium-induced coupling. J Am Coll Nutr 23:514S–517S, 2004. action potential duration: A computational model.
calcium release in cardiac ventricular myocytes. 67. Michailova A, McCulloch AD: Effects of Mg2+, J Mol Cell Cardiol 49:617–624, 2010.
Biophys J 87:3723–3736, 2004. pH and PCr on cardiac excitation-metabolic cou- 89. Pereira L, Metrich M, Fernandez-Velasco M, et
47. Greenstein JL, Hinch R, Winslow RL: Mecha- pling. Magnes Res 21:16–28, 2008. al: The cAMP binding protein Epac modulates
nisms of excitation-contraction coupling in an 68. Matsuoka S, Sarai N, Jo H, et al: Simulation of Ca2+ sparks by a Ca2+/calmodulin kinase signalling
integrative model of the cardiac ventricular ATP metabolism in cardiac excitation-contraction pathway in rat cardiac myocytes. J Physiol 583(Pt
myocyte. Biophys J 90:77–91, 2006. coupling. Prog Biophys Mol Biol 85:279–299, 2):685–694, 2007.
48. Camors E, Mohler PJ, Bers DM, et al: Ankyrin-B 2004. 90. Metrich M, Lucas A, Gastineau M, et al: Epac
reduction enhances Ca spark-mediated SR Ca 69. Korzeniewski B, Zoladz JA: A model of oxidative mediates beta-adrenergic receptor-induced car-
release promoting cardiac myocyte arrhythmic phosphorylation in mammalian skeletal muscle. diomyocyte hypertrophy. Circ Res 102:959–965,
activity. J Mol Cell Cardiol 52:1240–1248, 2012. Biophys Chem 92:17–34, 2001. 2008.
49. Wolf RM, Mitchell CC, Christensen MD, et al: 70. Cortassa S, Aon MA, Marban E, et al: An inte- 91. Curran J, Hinton MJ, Rios E, et al: Beta-
Defining new insight into atypical arrhythmia: A grated model of cardiac mitochondrial energy adrenergic enhancement of sarcoplasmic reticu-
computational model of ankyrin-B syndrome. Am metabolism and calcium dynamics. Biophys J lum calcium leak in cardiac myocytes is mediated
J Physiol Heart Circ Physiol 299:H1505–H1514, 84:2734–2755, 2003. by calcium/calmodulin-dependent protein kinase.
2010. 71. Cortassa S, Aon MA, O’Rourke B, et al: A com- Circ Res 100:391–398, 2007.
50. Mohler PJ, Schott JJ, Gramolini AO, et al: putational model integrating electrophysiology, 92. Grimm M, Brown JH: Beta-adrenergic receptor
Ankyrin-B mutation causes type 4 long-QT contraction, and mitochondrial bioenergetics in signaling in the heart: Role of CaMKII. J Mol
cardiac arrhythmia and sudden cardiac death. the ventricular myocyte. Biophys J 91:1564–1589, Cell Cardiol 48:322–330, 2010.
Nature 421:634–639, 2003. 2006. 93. Trenor B, Cardona K, Gomez JF, et al: Simulation
51. Pasek M, Simurda J, Christe G: The functional 72. Cortassa S, Aon MA, Winslow RL, et al: A mito- and mechanistic investigation of the arrhythmo-
role of cardiac T-tubules explored in a model of chondrial oscillator dependent on reactive oxygen genic role of the late sodium current in human
rat ventricular myocytes. Philos Transact A Math species. Biophys J 87:2060–2073, 2004. heart failure. PLoS One 7:e32659, 2012.
Phys Eng Sci 364:1187–1206, 2006. 73. Zhou L, Aon MA, Almas T, et al: A reaction- 94. Shannon TR, Wang F, Bers DM: Regulation of
52. Rice JJ, Tu Y, Poggesi C, et al: Spatially- diffusion model of ROS-induced ROS release in cardiac sarcoplasmic reticulum Ca release by
compressed cardiac myofilament models generate a mitochondrial network. PLoS Comput Biol luminal [Ca] and altered gating assessed with a
hysteresis that is not found in real muscle. Pac 6:e1000657, 2010. mathematical model. Biophys J 89:4096–4110,
Symp Biocomput 366–377, 2008. 74. Saucerman JJ, Brunton LL, Michailova AP, et al: 2005.
53. Michailova AP, Spassov VZ: Theoretical model Modeling beta-adrenergic control of cardiac 95. Anderson ME: Multiple downstream proarrhyth-
and computer simulation of excitation-contraction myocyte contractility in silico. J Biol Chem mic targets for calmodulin kinase II: Moving
coupling of mammalian cardiac muscle. J Mol 278:47997–48003, 2003. beyond an ion channel-centric focus. Cardiovasc
Cell Cardiol 24:97–104, 1992. 75. Saucerman JJ, McCulloch AD: Mechanistic Res 73:657–666, 2007.
54. Michailova A, Spassov V: Computer simulation of systems models of cell signaling networks: A case 96. Morotti S, Grandi E, Summa A, et al: Theoretical
excitation-contraction coupling in cardiac muscle: study of myocyte adrenergic regulation. Prog study of L-type Ca2+ current inactivation kinetics
A study of the regulatory role of calcium binding Biophys Mol Biol 85:261–278, 2004. during action potential repolarization and early
to troponin C. Gen Physiol Biophys 16:29–38, 76. Faber GM, Rudy Y: Action potential and contrac- afterdepolarizations. J Physiol 590(Pt 18):4465–
1997. tility changes in [Na+]i overloaded cardiac myo- 4481, 2012.
55. Noble D, Varghese A, Kohl P, et al: Improved cytes: A simulation study. Biophys J 78:2392–2404, 97. Tanskanen AJ, Greenstein JL, O’Rourke B,
guinea-pig ventricular cell model incorporating a 2000. et al: The role of stochastic and modal gating
diadic space, IKr and IKs, and length- and 77. Ahrens-Nicklas RC, Clancy CE, Christini DJ: of cardiac L-type Ca2+ channels on early after-
tension-dependent processes. Can J Cardiol Re-evaluating the efficacy of beta-adrenergic ago- depolarizations. Biophys J 88:85–95, 2005.
14:123–134, 1998. nists and antagonists in long QT-3 syndrome 98. Crampin EJ, Smith NP: A dynamic model of
56. Matsuoka S, Sarai N, Kuratomi S, et al: Role of through computational modelling. Cardiovasc excitation-contraction coupling during acidosis in
individual ionic current systems in ventricular Res 82:439–447, 2009. cardiac ventricular myocytes. Biophys J 90:3074–
cells hypothesized by a model study. Jpn J Physiol 78. Soltis AR, Saucerman JJ: Synergy between 3090, 2006.
53:105–123, 2003. CaMKII substrates and beta-adrenergic signaling 99. Shaw RM, Rudy Y: Electrophysiologic effects of
57. Negroni JA, Lascano EC: A cardiac muscle model in regulation of cardiac myocyte Ca2+ handling. acute myocardial ischemia: A mechanistic investi-
relating sarcomere dynamics to calcium kinetics. Biophys J 99:2038–2047, 2010. gation of action potential conduction and conduc-
J Mol Cell Cardiol 28:915–929, 1996. 79. Yang JH, Saucerman JJ: Phospholemman is a tion failure. Circ Res 80:124–138, 1997.
58. Niederer SA, Hunter PJ, Smith NP: A quantita- negative feed-forward regulator of Ca2+ in beta- 100. Ferrero JM Jr, Saiz J, Ferrero JM, et al: Simula-
tive analysis of cardiac myocyte relaxation: A adrenergic signaling, accelerating beta-adrenergic tion of action potentials from metabolically
simulation study. Biophys J 90:1697–1722, inotropy. J Mol Cell Cardiol 52:1048–1055, 2012. impaired cardiac myocytes: Role of ATP-sensitive
2006. 80. Heijman J, Volders PG, Westra RL, et al: Local K+ current. Circ Res 79:208–221, 1996.
59. Hunter PJ, McCulloch AD, ter Keurs HE: Mod- control of beta-adrenergic stimulation: Effects on 101. Ch’en FF, Vaughan-Jones RD, Clarke K, et al:
elling the mechanical properties of cardiac ventricular myocyte electrophysiology and Ca2+ Modelling myocardial ischaemia and reperfusion.
muscle. Prog Biophys Mol Biol 69:289–331, transient. J Mol Cell Cardiol 50:863–871, 2011. Prog Biophys Mol Biol 69:515–538, 1998.
1998. 81. Bers DM, Grandi E: Calcium/calmodulin-depen- 102. Zhou L, Cortassa S, Wei AC, et al: Modeling
60. Niederer SA, Smith NP: A mathematical model dent kinase II regulation of cardiac ion channels. cardiac action potential shortening driven by oxi-
of the slow force response to stretch in rat ven- J Cardiovasc Pharmacol 54:180–187, 2009. dative stress-induced mitochondrial oscillations in
tricular myocytes. Biophys J 92:4030–4044, 2007. 82. Livshitz LM, Rudy Y: Regulation of Ca2+ and guinea pig cardiomyocytes. Biophys J 97:1843–
61. Rice JJ, Jafri MS, Winslow RL: Modeling short- electrical alternans in cardiac myocytes: Role of 1852, 2009.
term interval-force relations in cardiac muscle. Am CAMKII and repolarizing currents. Am J Physiol 103. Amanfu RK, Saucerman JJ: Cardiac models in
J Physiol Heart Circ Physiol 278:H913–H931, Heart Circ Physiol 292:H2854–H2866, 2007. drug discovery and development: A review. Crit
2000. 83. Christensen MD, Dun W, Boyden PA, et al: Oxi- Rev Biomed Eng 39:379–395, 2011.
62. Rice JJ, Winslow RL, Hunter WC: Comparison dized calmodulin kinase II regulates conduction 104. Hilgemann DW, Noble D: Excitation-contraction
of putative cooperative mechanisms in cardiac following myocardial infarction: A computational coupling and extracellular calcium transients in
muscle: Length dependence and dynamic analysis. PLoS Comput Biol 5:e1000583, 2009. rabbit atrium: Reconstruction of basic cellular
mechanisms. Proc R Soc Lond B Biol Sci oretical formulation and their role in repolariza- 107. Carro J, Rodriguez JF, Laguna P, et al: A human
230:163–205, 1987. tion. Circ Res 77:140–152, 1995. ventricular cell model for investigation of cardiac
105. Zeng J, Laurita KR, Rosenbaum DS, et al: Two 106. Demir SS, Clark JW, Murphey CR, et al: A math- arrhythmias under hyperkalaemic conditions.
components of the delayed rectifier K+ current in ematical model of a rabbit sinoatrial node cell. Am Philos Transact A Math Phys Eng Sci 369:4205–
ventricular myocytes of the guinea pig type: The- J Physiol 266(3 Pt 1):C832–C852, 1994. 4232, 2011.
Calcium Signaling in Cardiomyocyte
Models With Realistic Geometries 33
Anushka P. Michailova, Andrew G. Edwards, Johan Hake,
Masahiko Hoshijima, and Andrew D. McCulloch*
CHAPTER OUTLINE transport mechanisms. Second, it provides a platform for region-

alization of specialized signaling structures. Here we focus on
Electrophysiological Structure of the Ventricular structures that are directly involved in cardiac Ca2+ signaling; the
Myocyte 331 best described are the calcium release unit (CRU; Figure 33-1,
Structure-Function Relationships at the Nanometer right panel) and the cardiac couplon.
Scale (CRU and Couplon) 332
Structure-Function Relationships at the Micron Scale A Brief Ultrastructural History of Cardiac
(T-Tubule and T-System) 335 Excitation-Contraction (E-C) Coupling
Initial impressions of the shape and structure of the CRU first
The microarchitecture of cellular substructures involved in appeared roughly half a century ago—well before the function of
calcium signaling is highly organized in all forms of mammalian the couplon was understood, or the term itself had been coined.7
striated muscle, and the cardiac ventricular myocyte (VM) is no At that time, transmission electron microscopy (EM) had shown
exception. The ultrastructural characteristics of the T-tubular junctions between the SR and T-tubular membranes, which, in
system and the sarcoplasmic reticulum (SR) play an important cardiac muscle, were termed dyads for their characteristic two-
role in normal cardiac electrophysiology, and their degradation component appearance in longitudinal tissue sections. A few
has dire consequences in a number of pathologic contexts. Below years later, Constantin et al8 identified these SR structures as the
we discuss several electrophysiologically important aspects of site of intracellular calcium storage; Winegrad’s landmark study9
VM ultrastructure, and place emphasis on describing structure- shortly thereafter showed that they were also the site of intracel-
function relationships in the healthy and diseased myocardium. lular Ca2+ release. Facett and McNutt5 accordingly refined Porter
Because interaction within and between these structures often and Palade’s well-known sketch of the cardiac sarcomere,10 and
occurs at or below the limit of resolution for traditional live- their rendition remains popular today (Figure 33-1, left panel). In
imaging techniques, quantitative computational modeling has this schematic, Ca2+ is released from specialized projections of
made an essential contribution to our understanding of how the SR, which closely juxtaposes the T-tubular membrane. These
structure determines function. As such, we highlight ongoing projections were originally named for their anatomical appear-
computational efforts that make use of reconstructed subcellular ance in transmission EM (i.e., junctional—jSR; terminal—
geometries to describe VM physiology at the optical diffraction cisternal SR) or for characteristics of physical separation (i.e.,
limit and below. heavy SR). Over the next several decades, classic studies were
published by a number of groups to define Ca2+-induced Ca2+
release (CICR) as the essential mechanism of E-C coupling
occurring at cardiac dyads, and responsible for activating contrac-
Electrophysiological Structure of the tion of the heart.11,12 Key aspects of CICR are detailed in the next
Ventricular Myocyte sections, but to review briefly, voltage-dependent opening of
local L-type Ca2+ channels (LCC) permits influx of Ca2+, which
Anatomy of the Myocyte Sarcolemma binds to and activates nearby ryanodine receptors (RyRs) in the
jSR membrane. This results in release, from the jSR, of the bulk
The plasma membrane of ventricular myocytes, and of most of Ca2+, which goes on to participate in contraction.12
mammalian striated muscle, exhibits regular invaginations that Over the past two decades, it has become clear that a number
project into the cell perpendicular to its surface (Figure 33-1). of important properties of cardiac E-C coupling are critically
These structures align with the sarcomeric Z-disc and were origi- dependent upon the biophysical characteristics of Ca2+-mediated
nally termed transverse tubules (T-tubules) for their dominant interaction between the LCC and RyR. These processes
orientation with respect to the long axis of the cell. However, in and phenomena will be discussed in detail below, but at this
most species, a large number of longitudinal branches have also point it behooves us to introduce several terms that have
been observed,1,2 and this has caused different authors to describe been adopted to reflect fundamental structure-function relation-
the lattice architecture variously as the transverse-axial-tubular ships in the current paradigm of E-C coupling, which is
system (TATS),3 the sarcolemmal tubule system,4 the T-system,5 known as “local control.”13 First, the term CRU refers to an
and the sarcolemmal Z-rete.6 At a macroscopic level, the T-tubular individual jSR terminal and its associated RyR ensemble14 and
lattice affects myocyte function through at least two important helps to indicate the discrete functional nature of these struc-
mechanisms. First, it expands VM surface area, and thereby tures, that is, individual CRUs are generally activated in an all-
increases the density (per unit cell volume) of all sarcolemmal or-nothing fashion but are otherwise functionally isolated from
one another by a sufficiently large inter-CRU diffusion distance.
The term couplon was introduced to explicitly define the combina-
*Corresponding authors A.G. Edwards, J. Hake, and A.P. Michailova also contributed to this chapter. tion of an LCC cluster and juxtaposed CRU, which together are
331
Sarcolemma T-tubule
Ca L-type Ca Ryanodine Na/Ca
ions channel receptors exchanger
Mitochondrion SERCA
pump
Ca release Junctional
unit sarcoplasmic
Sarcoplasmic reticulum
reticulum
ar
cellu
Intra
luar
acel
Extr T-tubule
Myofilaments
Dyadic cleft
Z-line Z-line
Figure 33-1. Overview of structures involved in cardiac Ca2+ handling and excitation-contraction (E-C) coupling. T-tubules are invaginations of the sarcolemma that facilitate
spatiotemporal synchronization of intracellular Ca2+ release by bringing L-type Ca2+ channels (LCCs) into close proximity with most of the cell’s volume. Ca2+ release units
(CRUs) juxtapose the T-tubules along their length, and the narrow space separating the jSR and the T-tubules is named the dyadic cleft. LCCs reside in the sarcolemma,
predominantly at dyadic clefts, and the surface of the release unit that faces the T-tubule (the dyadic surface) contains ryanodine receptors (RyRs). In a ventricular myocyte,
typically ≈20 000 CRUs are present, and these are synchronously activated during an action potential (AP). Upon arrival of the AP, LCCs open and permit Ca2+ influx, which
then triggers further Ca2+ release from the sarcoplasmic reticulum (SR) via binding to local RyRs. Ca2+ release from a single CRU is called a Ca2+ spark. After a spark, the SERCA
pump and the Na/Ca exchanger either resequester the released Ca2+ into the SR, or extrude it from the cell, respectively.
(Left panel is retouched from Fawcett DW, McNutt NS: The ultrastructure of the cat myocardium. I. Ventricular papillary muscle. J Cell Biol 42:1–45, 1969.)
capable of functionally interacting to contribute to E-C cou- to 15 nm, and appears to be interspersed among the RyR mole-
pling.7 For this reason, the couplon is, by definition, the elemen- cules within the dyadic surface of the jSR membrane.19
tary site of E-C coupling in the heart (see Figure 33-1) and along The exact numbers of LCCs and RyRs inside the couplon are
with the CRU will be a feature of much further discussion in this still debated. Early studies employed transmission EM to measure
chapter. the cross-sectional diameter of individual CRU terminals, and to
show that, in mammalian cardiac muscle, RyRs form dense clus-
ters at these terminals.22 Assuming that the CRU terminals are
Role of Imaging in Defining Cardiac Ultrastructure circular (i.e., that the size corresponds to the diameter) and are
and Subcellular Modeling densely packed with receptors, these authors were able to approx-
imate the number of RyRs per couplon. This estimate suggested
Much of our knowledge of cardiac ultrastructure has been per- that 130 to 150 receptors are present in mouse ventricular cou-
mitted by steadily improving approaches to both EM and light plons.22 Given that the RyR tetramer is 30 nm at each side, this
microscopy (LM). In combination with tomographic reconstruc- would require a dyad containing 100 RyRs to be 440 nm in
tion algorithms, high-voltage EM (HVEM) has both enhanced diameter and 1.8 × 10−12 µℓ in volume. Recent studies have chal-
resolution and improved penetration of thick-section prepara- lenged these initial estimates.15,18,23 Using 3D EM tomography,
tions, which has now permitted 3-dimensional (3D) reconstruc- Hayashi et al15 reported that the size of each dyad is almost an
tion of the T-tubule system and associated E-C coupling order of magnitude smaller than was previously reported (mean
structures.15,16 Improvements in acquisition and analysis of LM 0.44 × 10−12 µℓ) and that each dyad includes a large fraction of
images have now permitted imaging at or below the optical dif- tiny dyads (median 0.28 × 10−12 µℓ). They also showed that the
fraction limit. This progress has brought the living T-tubule,6,17 density of RyRs within each CRU was sparse. By segmenting
and even the detailed morphology of individual RyR clusters,18,19 features, such as the T-tubule (green), jSR (yellow), dyadic space
into view. Together, these techniques have been central to defin- (white), and RyR occupancy (blue), they observed spaces within
ing the structural characteristics of VM microdomains, describ- the dyad that did not contain any RyRs (Figure 33-2, E). It is
ing protein localization in and around those domains, and interesting to note that Hayashi et al also found that 80% of all
permitting the geometrically detailed quantitative approaches we dyads have a neighboring dyad within 25 nm, which is dramati-
describe here. cally closer than was previously thought. Using super-resolution
light microscopy, Baddeley et al18 found support for most of these
surprising results. They similarly observed RyR clusters to be
much smaller, more closely arranged, and more irregularly
Structure-Function Relationships at the shaped than had been reported or assumed in earlier work. It is
Nanometer Scale (CRU and Couplon) not completely clear why recent studies have retrieved such dif-
ferent estimates of couplon size compared with those of Franzini-
Couplon and CRU Microarchitecture Armstrong et al,22 although it is at least clear that the early
assumption of a circular dyadic geometry was oversimplified, and
The small space between the jSR and T-tubular membranes pro- this fact may contribute to the discrepancy. Based on agreement
vides a confined volume (“dyadic cleft,” “junctional cleft,” or among recent structural studies and even earlier functional
“fuzzy space”) in which large and rapid changes in local [Ca2+] can work,24 it is probably safe to conclude that the number of RyRs
be generated by local transporters, particularly LCCs and RyRs. in the average CRU is on the order of tens rather than greater
This constrained architecture is fundamental to high-fidelity cou- than 100.
pling between those transporters, and is maintained by the spe- The observation by both Baddeley et al18 and Hayashi
cialized anchoring protein, junctophilin, which tethers the jSR to et al15—that many small RyR clusters exist very close to each
the T-tubular membrane.19-21 In healthy myocytes, junctophilin other—may have important functional implications. Baddeley
keeps the distance between the jSR and the T-tubule to within 12 et al defined a class of RyR super clusters, which combined small
Calcium Signaling in Cardiomyocyte Models With Realistic Geometries 333
33
200 nm
D
A 200 nm
E
B 200 nm C 200 nm F 200 nm
Figure 33-2. Fine anatomy of dyadic clefts in the mouse myocardium. A, High- resolution mesh models of a T-tubule (green) and jSRs (yellow) shown with a slice image;
these were constructed by dual-axis electron microscopy (EM) tomography. B and C, The ultrathin serial slice images of this structure revealed inhomogeneous distribution
of ryanodine receptor (RyR) feet in dyadic cleft spaces (B), and their RyR foot–rich subdomains were segmented (light blue lines in C). D through F, The intra-anatomy of
three closely assembled dyadic clefts. From the complete mesh model (D), jSR membranes are removed to expose eight RyR foot–rich subdomains (the surface meshes of
these subdomains are shown in light blue in E) that partially occupy dyadic cleft spaces (the whole dyadic cleft spaces are indicated as the junctional regions of T-tubular mem-
branes in white). Both jSR meshes and meshes that identify the RyR-rich subdomains are removed in F. Scale bars: 200 nm.
(From Hayashi T, et al: Three-dimensional electron microscopy reveals new details of membrane systems for Ca2+ signaling in the heart. J Cell Sci 122[Pt 7]:1005–1013, 2009.)
clusters within 100 nm of each other, because conventional myocardial Na+/Ca2+ exchanger (NCX) is also worth mentioning
understanding suggests that these clusters would function as a at this point, as its potential involvement in E-C coupling is a
single CRU. However, at some combination of short intercluster long-debated topic that has recently been revisited.32 Although
distance and low SR release flux, these clusters might be expected NCX does exhibit a punctate distribution in the T-tubules, precise
to exhibit a variety of spark amplitudes as the result of partial colocalization of NCX with RyR occurs for only ∼10 % of the
activation of the super cluster. Such subspark amplitude CRU total NCX signal.27,33 With this in mind, the latter authors also
activations have been observed experimentally and have been note that ∼40% of NCX puncta reside within 150 nm of the
called quarks or sparklets.25 It is conceivable that the sparse distri- nearest RyR cluster. Thus, even though NCX is probably not
bution of RyRs within a single dyad (observed by Hayashi et al15) selectively concentrated within the dyadic portion of the T-tubular
could yield such behavior, and this sparse arrangement has since membrane, it is likely to be nearby.
been observed by other investigators.23 The protein exhibiting strongest colocalization with RyR is
Immunofluorescence-based colocalization studies have defini- calsequestrin (CSQN). A total of 95% of all CSQN labeling is
tively shown that the LCCs present in ventricular myocytes coincident with RyR, and only ∼10% of RyR occur in the absence
exhibit a punctate pattern of distribution, and that ∼90% of these of CSQN. These observations are consistent with our under-
puncta are coincident with RyR.26,27 Thus, a vast majority of LCCs standing of the function of CSQN, which, through its ability to
are probably components of couplon structures. The number of rapidly buffer Ca2+, acts to both limit the thermodynamic gradi-
LCCs within each couplon is less certain, as is the number of ent that opposes SR Ca2+ reuptake and provide a large local
active LCCs required to trigger local Ca2+ release during physi- supply of Ca2+ for RyR-mediated Ca2+ release. Finally, localiza-
ological E-C coupling. One range of estimates suggests that 17 to tion of the sarco-endoplasmic reticulum Ca2+ ATPase (SERCA2)
53 LCCs are likely to be involved in couplon activation28-30; others is less clear than that of RyR or CSQN. It is generally agreed
have observed higher coupling fidelities and therefore suggest that both major (SERCA2a) and minor (SERCA2b) cardiac splice
that fewer LCCs are required to trigger local SR Ca2+ release.31 variants are present at the Z-disc, and therefore in the vicinity of
Part of the uncertainty here is due to discrepancy in the estimates the jSR.34-36 However, fluorescent labeling also appears to deco-
of RyR cluster size, which are required to extrapolate the LCC rate the M-line SR,34 and no investigation to date has definitively
number in some studies.29 Otherwise, differences in the details of demonstrated that SERCA2 exists within the same functional
approach across studies, such as the potential ranges for LCC domain as RyR, and CSQN. As mentioned below, our quantita-
activation and partial pharmacological inhibition of the LCC tive approaches suggest that uncertainty in SERCA2 localization
pool, have made it difficult to determine truly physiologically at this level may be functionally important, particularly with
representative constraints for E-C coupling. Localization of the respect to Ca2+ spark dynamics.
Subcellular Modeling at the Nanometer Scale been estimated from indirect measures, such as Ca2+ spark fluo-
rescence.42 For this uncertainty, modelers have turned to func-
Functional data gained from LM and patch-clamp electrophysi- tionalistic models of the CRU43 or have constrained their
ology have formed the basis for multiple model studies of whole- approaches to lipid bilayer recordings of RyRs extracted from
cell E-C coupling and single Ca2+ sparks over the past 20 years.37,38 their native environment.44 The former have been most popular
A set of four specific properties of macroscopic E-C coupling, in integrative models.45,46 However, we still lack a proper under-
and isolated RyR function, have been used to drive and constrain standing of how models based on bilayer experiments can be
the development of these computational models: (1) RyR activa- incorporated into integrative models, and whether they fulfill the
tion should exhibit a steep and nonlinear dependence upon sub- four E-C coupling properties previously described. One of the
membrane cytosolic [Ca2+]; (2) reliable termination of release reasons for this is that to be able to reliably terminate a Ca2+ spark
should be achieved by experimentally defined RyR gating proper- (property (2) mentioned earlier), the functionalistic models have
ties; (3) the total Ca2+ released from the SR should be about an employed Ca2+-dependent inactivation as the main termination
order of magnitude larger than that entering through activated mechanism. However, lipid bilayer experiments do not support
LCCs, that is, E-C–coupling gain should be high; and (4) the such an inactivation mechanism at physiologically relevant cyto-
total Ca2+ released from the SR should be proportional to the solic [Ca2+].47 Alternatively, experimental measurements have
macroscopic LCC current, that is, SR Ca2+ release should be demonstrated that RyR channel activity is also regulated by
graded by L-type Ca2+ current (ICaL). luminal SR [Ca2+], and it has been thought for some time that
The multiscale aspects of modeling E-C coupling became this mechanism may contribute to spark termination via local jSR
obvious quite early, as (3) and (4) are whole-cell measures, and Ca2+ depletion.48 When luminal Ca2+ regulation was included in
E-C coupling is controlled locally within each of the ∼20 000 the model, together with a limited supply of Ca2+ in the jSR,
couplons in a myocyte. The first attempts to model E-C coupling reliable spark termination was accomplished without inactivation
used the mean cytosolic [Ca2+] as a trigger for RyR release. None by cytosolic [Ca2+].49 A requirement of this model, however, is
of these early models, however, were able to reconcile (1) to (3) that luminal [Ca2+] is depleted to a level not supported by direct
with (4).13 The remedy was to let local [Ca2+] within each CRU measurements during a spark.50 Sobie and Lederer recently sug-
control the gating of the local RyRs,13,39,40 and thus “local control” gested that this discrepancy may be resolved if the luminal Ca2+
models were born. These models kept a steep all-or-none control signal is averaged over neighboring SR compartments, which is
of RyR release within each couplon, and graded release was probably the most realistic reference to experimental measure-
accomplished by a probabilistic recruitment of the CRUs, where ments, where recording volumes are much larger than a single
CRUs were activated in proportion with ICaL. The problem of CRU.51 To determine whether physiological spark termination
combining local control and whole-cell aspects of Ca2+ signaling can be achieved with this modification, a more detailed computer
is an inherently multiscale problem that has been addressed by model of spark generation, including Ca2+ diffusion within local
several approaches of varying complexity.41 SR compartments, was needed.
At the level of a single CRU, modelers have been challenged To that end, Hake et al managed to combine recent develop-
by the lack of direct measurements of local SR Ca2+ release ments in 3D EM tomography with new techniques in computa-
current. Because RyRs are not members of the sarcolemma, tional mesh generation to develop a 3D model of the Ca2+ spark.52
patch-clamping techniques are not available to assay channel The CRU geometry (Figure 33-3, A) was manually segmented
function in the intact cell. Instead local SR release current has from 3D EM tomography data for a mouse ventricular myocyte.15
Mit
nSRZ
SR TT
TT
Mit
jSR
A 60 nm B C D Cytosolic domain
SRZ1 nSR1
nSR2
Mit
nSR3 nSR4
TT Mit
SRZ2 jSR
nSR nSR6
nSR5
nSRZ
Cytosolic domain
E F G
Figure 33-3. Generation of reconstructed geometries is a multistage process. A, First, features are segmented from electron tomography data. B, From the segmented
data, an initial low-quality surface mesh is generated. C, The quality of the surface mesh is improved. D, The surface mesh is annotated, for easy application of boundary
fluxes. Blue is the T-tubule; red, orange, and yellow show the sarcoplasmic reticulum; and purple is a mitochondrion. E, The full, annotated surface mesh. The dimensions of
the mesh are 1430 × 940 × 406 nm. F, The surface mesh is converted to an annotated volumetric mesh using TetGen. G, The geometry included a full representation of
the local SR.
(Panels A, D, E, and G from Hake J, et al: Modeling cardiac calcium sparks in a three-dimensional reconstruction of a calcium release unit. J Physiol 590[Pt 18]:4403–4422, 2012.)
An annotated surface mesh then was generated from the seg- simulated E-C coupling.59 As described in next section, this lack of
33
mented features (Figure 33-3, B-E), and eventually was trans- synchrony has become a hallmark characteristic of E-C coupling
formed into a volumetric tetrahedral mesh, seen in Figure 33-3, F. dysfunction in a variety of cardiac pathologies. To our knowledge,
The SR was compartmentalized by dividing the SR geometry these studies have provided the only available theoretical analyses
into individual parts (Figure 33-3, G). of disease-associated couplon remodeling. More broadly, the
potential for using reconstructed geometries of the CRU to model
Ca2+ sparks and other microdomain-defined processes is huge. As
Modeling of Ca2+ Sparks in a Realistic 3D EM tomography and super-resolution microscopy techniques
CRU Geometry improve, large volumes of ultrastructural imaging data will become
available. 3D computer models enable us to investigate the
In Hake’s model, Ca2+ release was simulated by coupling Ca2+ structure-function relationships of localized proteins in geometri-
diffusion within the local SR to Ca2+ diffusion in the cytosol. By cally confined environments. Therefore, many questions remain
including the SR Ca2+ dye, Fluo5, these authors also modeled the to be investigated with respect to how E-C is controlled locally,
impact of dye dynamics on experimental measurements of luminal particularly those associated with the structural basis for subspark
[Ca2+]. Specific RyR dynamics were not used to define release release events (quarks and sparklets) and the role of protein local-
termination in this model. Instead release was terminated such ization in shaping spark dynamics. It is exciting to consider what
that the modeled Fluo5 signal decayed to a level corresponding these detailed analyses can uncover if they can be scaled up to
with experimentally measured values.50 In this way, Hake et al reference the volumes of data available to describe macroscopic
applied a functionalistic termination criterion to directly relate properties of E-C coupling.
experimentally measured Fluo5 fluorescence with jSR [Ca2+] at
the end of a spark. It is important to note that these dynamics
included local competition between Fluo5 and CSQN, and a Structure-Function Relationships at the
maximal RyR release flux that was within the range of the best
available experimental estimates. These analyses suggest that the
Micron Scale (T-Tubule and T-System)
jSR [Ca2+] at spark termination is indeed low (∼10% of resting
value, see Figure 33-4, C), even though Fluo5 fluorescence decays T-Tubule Microanatomy and Its Role in Regulating
to only 60% of the resting value.52 The available Ca2+ reserve E-C Coupling
within the neighboring SR compartments (44 %) was also com-
parable with experimentally measured values.53 As is true for the CRU, advancements in imaging modalities have
Another advantage of the realistic geometry used in this model driven progress in defining the architecture of cardiac T-tubules,
is that it inherently provides the correct surface area-to-volume and a consistent finding among these studies is the strong species
relationship for local structures. This is crucial to finding correct specificity of T-tubular structure. In the rat, the average ventricu-
quantitative relationships between the local surface fluxes repre- lar tubule is ∼250 nm in diameter, and exhibits longitudinal (axial)
sented in the model by the RyRs, SERCA, and NCX. SERCA branches every 6.87 µm.6 Rabbit and human myocytes generally
localized close to the CRU will experience a much larger cyto- have larger T-tubular diameters (∼450 nm), and exhibit less pro-
solic [Ca2+] during the spark than peripherally located pumps, as nounced branching.1,17,60,61 This frequency of T-tubular branch-
is illustrated by the cytosolic [Ca2+] gradients shown in Figure ing indicates a key aspect of species differences in the relationship
33-4, B. Hake et al identified a new potential role for the SERCA between T-tubular structure and function: The T-tubular rete of
pump during the Ca2+ spark. By pumping Ca2+ back into the SR rodents is generally more intricate than that of larger mammals,
during the spark, local SERCA activity was able to slow down particularly humans. The most important outcome of this
the depletion of luminal Ca2+, and hence increase spark duration. complex branching is a decreased distance from any part of the
This was possible because at the late phase of a Ca2+ spark, when cytoplasm to the sarcolemmal membrane, within the rodent cell.
luminal Ca2+ in jSR is low and cytosolic Ca2+ close to the CRU Two-dimensional analyses of transverse sections indicate that, in
is high, the SERCA pump operates under thermodynamically the human, ∼64% of the cytoplasm resides within 250 nm of a
optimal conditions. This modulatory role needs to be investi- T-tubular membrane, compared with ∼94% in the rat.1 It is
gated further, but it is a clear prediction of the model, which was generally held that this intricate lattice-like structure facilitates a
made possible by the correct surface area-to-volume relationship more synchronous Ca2+ release in the roden (e.g., mice, rats,
for local SR components. squirrels, porcupine, beavers, guinea, hamsters), thereby reduc-
ing the intracellular heterogeneity of E-C coupling, which may
be necessary to support efficient contraction at the very high
Future Work heart rates of these animals. As we describe at the end of this
chapter, this principle probably also applies in the converse under
Using models to quantitatively demonstrate mechanisms of circumstances of pathologic remodeling, which degrades the
disease associated with ultrastructural remodeling is a clear and lattice architecture, resulting in a sparse T-tubular architecture.
important application of these techniques. To date, a couple of Although it has not been directly tested, a commonly held
studies have looked at the effects of altered phosphorylation of hypothesis is that these differences in T-tubular density influence
RyR, LCC, and SERCA/phospholamban in failing cells.54-56 E-C coupling by altering the density of functional couplons. As
However, because of the lack of published structural data on described earlier, CRU nearest neighbor distances (from the edge
disease-induced remodeling of CRU geometry, very few modeling of one CRU to the edge of its closest adjacent CRU) appear to be
studies have addressed this problem. Koh et al showed that spark somewhere on the order of tens of nanometers in the rat.15,18
dynamics were highly sensitive to changes in dyadic cleft height— Center-to-center distances are larger on average, typically 300 to
a hypothetical characteristic of heart failure.57 For the cleft height 800 nm.18,22,26 These distances are ∼15% to 50% greater in the
to be changed, expression or distribution of junctophilin most human, resulting in an RyR cluster density (per unit cell volume)
probably has to be altered.58 In a model based on experimental that is roughly halved compared with that of the rat.62 Remodeling
junctophilin knockdown in the mouse, van Oort et al showed that accompanying heart failure, also in humans, reduces RyR cluster
the E-C coupling gain was lowered as a result of decreased size of density by a further 20%,61 and, as is described below, the density
the CRU.21 Similarly, Cannell et al demonstrated that increasing of functional couplons is probably even more severely degraded by
the height of the dyadic cleft may reduce spark synchrony during pathologic remodeling.
Ca[µM]
5
0
A
6 jSRBack boundry 1200 nSR6 compartment
5 Cytosol nSZ1 compartment
nSR6 boundry 900 jSR compartment
4
[Ca][µM]
[Ca][µM]
3 600
2
300
1
0 0
0 20 40 60 80 100 0 20 40 60 80 100
B Time [ms]
Ca[µM]
C Time [ms]
2.0
2.5 2.5
1.5
1.0
Distance [µM]
–0.5 2.0 2.0

F/F0
–0.0
–0.5
1.5 1.5
–1.0
–1.5
–2.0 1.0
1.0
0 20 40 60 80 100 0 20 40 60 80 100
D Time [ms] E Time [ms]
Figure 33-4. Large Ca2+ gradients within a single Ca2+ release unit (CRU). A, Volumetric representation of the [Ca2+] in the cytosolic domain after 5 ms. B, Average [Ca2+] at
three different positions in the CRU: backside boundary of jSR (continuous line), the whole cytosolic domain (dashed line), and at the boundary of the 6th nSR compartment
(dash-dotted line). C, Free Ca2+ content in three SR domains during a spark: the 6th nSR compartment (continuous line), the 1st Z-line SR compartment (dashed line), and the
jSR compartment (dash-dotted line). D, A generated line-scan image from the cytosolic Fluo4 signal with added noise. The width and duration of the spark are 1.0 µm and
28.5 ms, respectively. E, A Fluo4 trace from the red marker at the right of the line-scan image; SR, Junctional sarcoplasmic reticulum; nSR, network sarcoplasmic
reticulum.
(From Hake J, et al: Modeling cardiac calcium sparks in a three-dimensional reconstruction of a calcium release unit. J Physiol 590[Pt 18]:4403–4422, 2012.)
As is true within the couplon, protein localization is a key does the sustained K+ current (ISS).60 Other K+ currents appear to
component of T-tubular structure-function relationships, and a be homogeneously distributed. Many of these current carriers are
number of sarcolemmal transporters are selectively enriched sensitive to local ionic concentrations, particularly Na+ and Ca2+,
within the T-tubules. Functionalistic approaches involving rapid and this is just one reason why spatially realistic quantitative
changes in permeant ion concentration, or experimental detubu- models will be fundamental in developing a complete under-
lation, indicate that currents carried by LCC,63,64 NCX,65 and standing of their function.
Na+/K+-ATPase65,66 exhibit a dominant T-tubular distribution.
The fold enrichment (ICaL : INaCa:INaK) of each of these currents,
with respect to the surface sarcolemma, and after accounting for dif- Subcellular Modeling at the Micron Scale
ferences in corresponding membrane areas, is approximately
6 : 3 : 3.63 It is interesting to note that the neuronal Na+ current Whole-cell computational models have proved to be powerful
(INaN), which has recently been proposed to be a driver of reverse tools for predicting and analyzing interactions among sarcolem-
mode INaCa,32 exhibits strong T-tubular localization (∼9-fold), as mal ion fluxes, the action potential, and intracellular
Ca2+ handling under normal and pathologic conditions. For a
33
complete view of successes and failures in modeling pursuits of
this type, we refer the reader to excellent recent reviews from
Noble and colleagues.67,68 Using whole-cell modeling approaches,
Orchard and collaborators also investigated how changes in Nuc
T-system volume and distribution of ion fluxes between the
surface and the T-tubular membrane affect whole-cell electro-
physiology.69,70 However, as was the case for modeling of the
couplon, an important limitation of these whole-cell models is
that they treat subcellular spaces as lumped compartments, and
are thus unable to dissect the impact of structural changes in the
T-tubule system and other organelles. To overcome this limita-
tion, recent modeling approaches have introduced simplified
representations of T-system geometry to enable introduction of
spatial control of Ca2+ handling. In rodents, Lu et al and Yao and
Yu assumed a cylindrical T-tubule geometry.71,72 These studies
suggest that the Ca2+ transient is tightly regulated by the localiza-
tion of sarcolemmal Ca2+ transporters and is strongly reliant upon
the presence of Ca2+ buffers when SR Ca2+ fluxes are pharmaco-
Figure 33-5. The rodent ventricular T-tubular network and a single-tubule geomet-
logically inhibited. Hatano and collaborators73,74 extended the ric model. Upper left panel: Cardiac sarcolemma including T-tubules is visualized by
approach of Lu et al by developing a detailed 3D geometric 2-photon microscopy: external membrane (blue arrows); T-tubules (white arrows);
model of a guinea pig cardiomyocyte in which the subcellular nucleus (Nuc); bar 2 µm. Lower left panel: Geometric model of the T-tubule subsys-
structures (T-tubules, myofibrils, SR, mitochondria) were tem extracted from the 2-photon image. Middle and right panels: Expanded view
modeled using simplified geometries. Their analyses of sarco- of single T-tubule geometry and its surrounding half-sarcomeres.
mere dynamics revealed that the dyssynchronous contraction
(Adapted from Cheng Y, et al: Numerical analysis of Ca2+ signaling in rat ventricular
caused by detubulation leads to impairment of contractile
myocytes with realistic transverse-axial tubular geometry and inhibited sarcoplasmic
efficiency.
reticulum. PLoS Comput Biol 6:e1000972, 2010.)
The use of idealistic shapes, however, may alter diffusion
distances in longitudinal and axial directions and, consequently,
the predicted local and global Ca2+ signals. Using published
images of T-tubule ultrastructure from rodents and rabbits, considered because the entire T-tubular system in ventricular
investigators have examined Ca2+ dynamics in 3D reconstructions myocytes is roughly periodic.15,79 The surrounding half-
of single-tubule and multi-tubule geometric domains.75-77 Before sarcomeres were modeled as a rectangular box of 2 µm × 2 µm
discussing work that has made use of this relatively new paradigm in the plane of external sarcolemma and 5.96 µm in depth (Figure
in subcellular modeling, it is best to briefly describe the process 33-5, middle). Because Yu’s T-tubule model did not include the
by which these geometries are generated. realistic cell surface, one of the box faces (top red surfaces in Figure
33-5) was assumed to be the external sarcolemma. The T-tubule
inside this compartment was extracted from the T-PM data cor-
3D Imaging and Geometric Modeling of a Native responding to the region indicated in Figure 33-5 (lower left).
Ventricular T-Tubular System T-tubule diameter varied from 0.19 µm to 0.469 µm, and
T-tubule depth was 5.645 µm. The volume of the model com-
As has been described, Hake et al made use of 3D EM tomograms partment was ∼23.31 µm3. The compartment membrane area
to generate a computational geometry of a single CRU, but it is measured ∼9.00 µm2 where the percentage of cell membrane
not only EM data that can be used for these purposes. The gap within the T-tubule was 64% (∼5.75 µm2), and that within the
between imaging and simulation involves two major steps: (1) external membrane was 36% (∼3.25 µm2). Four exogenous and
extracting features (boundary or skeleton) from imaging data; and endogenous Ca2+ buffers (Fluo3, ATP, calmodulin, and troponin
(2) constructing geometric models (represented by meshes) from C) were modeled within the cytosolic domain (Figure 33-6, left
the detected features. Thus, the simple requirement of this panel). When 100 µM Fluo3 was included, along with ∼1.7-fold
approach is that the imaging modality provides sufficient resolu- enrichment of LCC and 3-fold enrichment of NCX in the
tion and dimension to both (1) capture the volume of the desired T-tubular membrane, the model-predicted [Ca2+] dynamics
microarchitectural feature; and (2) permit sufficiently accurate closely resembled experimental data collected in rat ventricular
segmentation of that feature. To this end, image preprocessing is myocytes with blocked SR Ca2+ fluxes80 (Figure 33-6, right panels).
usually necessary for better feature extraction, particularly when Counterintuitively, when LCC density was heterogeneously dis-
the original image is noisy, or the contrast between features and tributed within the T-tubule model, the spatial heterogeneity of
background is low. With 3D two-photon microscopy (T-PM) cytosolic [Ca2+] was reduced relative to a homogeneous sarcolem-
images, Yu and collaborators developed a set of image processing mal LCC distribution. Strongly nonuniform spatial Ca2+ gradi-
and analysis tools, which, combined with the mesh generator, ents, not observed during experiments, were found when LCC
GAMer, were able to generate high-quality meshes for 3D and NCX fluxes were uniformly distributed along the sarco-
T-tubular systems in mice (Figure 33-5, lower left).16,78 lemma. This unexpected result may be due to the high curvature
of the membrane near the mouth of the T-tubule, which increases
the LCC flux per unit cytosolic volume in that region. When the
Modeling the Effects of Normal T-Tubule LCC distribution is homogeneous, this effect is exaggerated at
Ultrastructure on Subcellular Ca2+ Signals these points of high curvature. A second important finding of this
study, which may contribute to the effects regarding LCC distri-
Cheng and collaborators75 used the Yu et al16 geometry to inves- bution, is that including the mobile Ca2+ buffering effect of
tigate the role of individual T-tubular architecture and Ca2+ 100 µM Fluo3, allowed masking of spatial nonuniformities in
buffer activity in determining the characteristics of Ca2+ signaling cytosolic [Ca2+] that occurred in the absence of dye (even when
in rat ventricular myocytes. A small compartment containing a Ca2+ transporters were heterogeneously distributed). Thus,
single T-tubule and its surrounding half-sarcomeres was during physiological Ca2+ influx, (i.e., no Fluo3), large and steep
0 0.25
surface membrane [µm]

1
Ca2+ concentration [µM]

Distance to
2 0.2
3
Ca2+ Ca2+ Ca2+
4 0.15
LCC
NCX
Leak
5
Ca2+
Ca2+ 6
0 100 200 300 400 0.1
NCX
Time [ms]
SR
LCC
0.5
Leak
Ca2+ 0.45
Intensity
Fluo 3
0.4
Ca2+ ATP
CAL 0.35
TN
0 100 200 300 400 500

Time [ms]
Figure 33-6. Left panel: Diagram illustrating Ca2+ entry and extrusion via the sarcolemma, and Ca2+ buffering and diffusion inside the cytosol with sarcoplasmic reticulum
(SR) Ca2+ fluxes pharmacologically blocked: L-type Ca2+ channel (LCC); Na+/Ca2+ exchanger (NCX); membrane Ca2+ leak (Leak); sarcoplasmic reticulum (SR); troponin C (TN),
adenosine triphosphate (ATP); calmodulin (CAL); Ca2+ fluorophore (Fluo3). Upper right panel: Calcium concentrations visualized as line-scan images in the transverse cell
direction. LCC current density heterogeneously distributed along the length of the T-tubule. Na+/Ca2+ flux density was three times higher in the T-tubule, and the Ca2+ leak
current was homogeneously distributed. Line scan was positioned 200 nm from the T-tubule membrane at an angle of 120 degrees. Lower right panel: Local Ca2+ time
courses with re-plot from experimental data.80 The re-plots are taken along the scan line at 0 µm (blue), 3.96 µm (green), and 5.65 µm (red) from the near-surface location.
The scan line in the Cheng et al experiments80 was located 200 nm from the surface of the T-tubule.
(Adapted from Cheng Y, et al: Numerical analysis of Ca2+ signaling in rat ventricular myocytes with realistic transverse-axial tubular geometry and inhibited sarcoplasmic reticulum.
PLoS Comput Biol 6:e1000972, 2010.)
Ca2+ gradients might be expected in the narrow subsarcolemmal spatial irregularities of RyR cluster distribution may significantly
space (∼40 to 50 nm in depth). It is interesting to note that in influence Ca2+ dynamics, although the role of T-tubular micro-
examining rabbit cells, Kekenes-Huskey et al revealed qualita- anatomy was not directly assessed in that study.
tively similar results to those of Cheng et al. Local [Ca2+] gradi-
ents within the cytosol and subsarcolemmal regions were highly
sensitive to details of T-tubule ultrastructure and membrane Ca2+ Toward a Quantitative Understanding
flux distribution, when SR Ca2+ fluxes were blocked.76 For this of Pathologic T-Tubule Remodeling
reason, it seems likely that several of the key effects of single
T-tubule microarchitecture on local Ca2+ dynamics are conserved Over the past 5 to 15 years, remodeling of the T-tubular lattice
across species. has emerged as a consistent characteristic of a number of cardiac
These two models are the first to have combined sophisticated pathologies82 and across a range of species, including humans.61,83
computational methods and detailed structural information to The characteristics and pathology specificity of these processes
understand the role of T-tubular microarchitecture in defining are not yet firmly established. However, various investigators
cardiac Ca2+ microdomains. Important limitations of these studies have observed a reduction in the number or density of T-tubules,83
are (1) the relatively small size of the modeled compartments, expansion of T-tubular diameter,61 and/or an increase in the
which contained only a single realistic T-tubule; and (2) as prevalence of axial/longitudinal tubule elements.84 A consistent
described previously, significant branching of the rat T-tubular functional outcome associated with these remodeling processes
lattice, which suggests that our assumption that the modeled is reduced synchrony of CICR,82,85 which is generally thought to
compartment is a repeating unit inside the cell is, to some extent, contribute to contractile dysfunction.
unrealistic. The precise mechanisms by which remodeling leads to CICR
In attempting to reach beyond these limitations, Yu et al dyssynchrony are not completely clear, although decoupling of
extended the approach of Cheng et al by reconstructing a 3D RyR clusters appears to be the dominant end point.86,87 This
geometry of several T-tubules in the mouse.77 Briefly, this study decoupling may occur through orphaning of entire T-tubules
affirmed that local Ca2+ dynamics are sharply affected by T-system from the surface membrane, such that they can no longer conduct
ultrastructure and Ca2+ fluxes at the surface membrane. Finally, the AP to their associated CRUs, or as the result of local changes
Soeller et al constructed a model of stochastic Ca2+ dynamics in in couplon geometry that orphan individual receptor clusters. As
rat myocytes using measured 3D distributions of RyR clusters, was mentioned earlier, it is also possible that reduced overall
which are adjacent to the T-system.81 These authors found that CRU density causes heterogeneity in Ca2+ release accompanying
T-tubule loss.61 The approach presented by Heinzel et al85 offers To date, no modeling study has attempted to explicitly simu-
33
some promise for describing these structure-function relation- late the effects of T-tubular remodeling with realistic geometries.
ships experimentally. These authors measured regional CICR As mentioned earlier, some researchers have taken idealized
kinetics and T-tubular density in the same patch-clamped myo- approaches to understanding how increases in junctional cleft
cytes. Although they did not quantitatively correlate delayed width promote heterogeneity in CICR latency59 or E-C coupling
CICR with distance from the local T-tubule, they were able to gain,21 but more complex structural changes have not been inves-
characterize spark kinetics at sites that were designated as decou- tigated. These questions remain an obvious target of the
pled based on the latency of their activation. approaches we have described in this chapter.
18. Baddeley D, et al: Optical single-channel resolu- same sites in the murine heart. Cell Calcium
References tion imaging of the ryanodine receptor distribution 34:457–464, 2003.
in rat cardiac myocytes. Proc Natl Acad Sci U S A 37. Soeller C, Cannell MB: Analysing cardiac
1. Jayasinghe I, et al: Comparison of the organization 106:22275–22280, 2009. excitation-contraction coupling with mathematical
of t-tubules, sarcoplasmic reticulum and ryanodine 19. Jayasinghe ID, et al: Nanoscale organization of models of local control. Prog Biophys Mol Biol
receptors in rat and human ventricular myocar- junctophilin-2 and ryanodine receptors within 85:141–162, 2004.
dium. Clin Exp Pharmacol Physiol 39:469–476, peripheral couplings of rat ventricular cardiomyo- 38. Williams GS, et al: Models of cardiac excitation-
2012. cytes. Biophys J 102:L19–L21, 2012. contraction coupling in ventricular myocytes.
2. Sommer JR, Waugh RA: The ultrastructure of the 20. Takeshima H, et al: Junctophilins: A novel family Math Biosci 226:1–15, 2010.
mammalian cardiac muscle cell—with special of junctional membrane complex proteins. Mol 39. Greenstein JL, Winslow RL: An integrative model
emphasis on the tubular membrane systems: A Cell 6:11–22, 2000. of the cardiac ventricular myocyte incorporating
review. Am J Pathol 82:192–232, 1976. 21. van Oort RJ, et al: Disrupted junctional membrane local control of Ca2+ release. Biophys J 83:2918–
3. Forbes MS, Hawkey LA, Sperelakis N: The complexes and hyperactive ryanodine receptors 2945, 2002.
transverse-axial tubular system (TATS) of mouse after acute junctophilin knockdown in mice. Cir- 40. Hinch R, et al: A simplified local control model of
myocardium: Its morphology in the developing culation 123:979–988, 2011. calcium-induced calcium release in cardiac ven-
and adult animal. Am J Anat 170:143–162, 1984. 22. Franzini-Armstrong C, Protasi F, Ramesh V: tricular myocytes. Biophys J 87:3723–3736, 2004.
4. Bers D: Excitation Contraction Coupling and Shape, size, and distribution of Ca(2+) release units 41. Williams GS, et al: Models of cardiac excitation-
Cardiac Contractile Force, Boston, 2001, Kluwer and couplons in skeletal and cardiac muscles. contraction coupling in ventricular myocytes.
Academic Publishers. Biophys J 77:1528–1539, 1999. Math Biosci 226:1–15, 2010.
5. Fawcett DW, McNutt NS: The ultrastructure of 23. Asghari P, et al: The structure and functioning of 42. Cheng H, Lederer WJ, Cannell MB: Calcium
the cat myocardium. I. Ventricular papillary the couplon in the mammalian cardiomyocyte. sparks: Elementary events underlying excitation-
muscle. J Cell Biol 42:1–45, 1969. Protoplasma 249(Suppl 1):S31–S38, 2012. contraction coupling in heart muscle. Science
6. Soeller C, Cannell MB: Examination of the trans- 24. Bridge JH, Ershler PR, Cannell MB: Properties of 262:740–744, 1993.
verse tubular system in living cardiac rat myocytes Ca2+ sparks evoked by action potentials in mouse 43. Stern MD, et al: Local control models of cardiac
by 2-photon microscopy and digital image- ventricular myocytes. J Physiol 518(Pt 2):469–478, excitation-contraction coupling: A possible role for
processing techniques. Circ Res 84:266–275, 1999. allosteric interactions between ryanodine recep-
1999. 25. Brochet DX, et al: Quarky calcium release in the tors. J Gen Physiol 113:469–489, 1999.
7. Stern MD, Pizarro G, Rios E: Local control model heart. Circ Res 108:210–218, 2011. 44. Zahradnik I, Gyorke S, Zahradnikova A: Calcium
of excitation-contraction coupling in skeletal 26. Scriven DR, et al: Analysis of Cav1.2 and ryanodine activation of ryanodine receptor channels—
muscle. J Gen Physiol 110:415–440, 1997. receptor clusters in rat ventricular myocytes. reconciling RyR gating models with tetrameric
8. Constantin LL, Franzini-Armstrong C, Podolsky Biophys J 99:3923–3929, 2010. channel structure. J Gen Physiol 126:515–527,
RJ: Localization of calcium-accumulating struc- 27. Scriven DR, Dan P, Moore ED: Distribution of 2005.
tures in striated muscle fibers. Science 147:158– proteins implicated in excitation-contraction cou- 45. Tanskanen AJ, et al: Protein geometry and place-
160, 1965. pling in rat ventricular myocytes. Biophys J ment in the cardiac dyad influence macroscopic
9. Winegrad S: Autoradiographic studies of intracel- 79:2682–2691, 2000. properties of calcium-induced calcium release.
lular calcium in frog skeletal muscle. J Gen Physiol 28. Louch WE, et al: Control of Ca2+ release by action Biophys J 92:3379–3396, 2007.
48:455–479, 1965. potential configuration in normal and failing 46. Winslow RL, et al: Integrative modeling of the
10. Porter KR, Palade GE: Studies on the endoplasmic murine cardiomyocytes. Biophys J 99:1377–1386, cardiac ventricular myocyte. Wiley Interdiscip Rev
reticulum. III. Its form and distribution in striated 2010. Syst Biol Med 3:392–413, 2011.
muscle cells. J Biophys Biochem Cytol 3:269–300, 29. Polakova E, et al: Local calcium release activation 47. Laver DR, Honen BN: Luminal Mg2+, a key factor
1957. by DHPR calcium channel openings in rat cardiac controlling RYR2-mediated Ca2+ release: Cyto-
11. Endo M, Tanaka M, Ogawa Y: Calcium induced myocytes. J Physiol 586:3839–3854, 2008. plasmic and luminal regulation modeled in a tetra-
release of calcium from the sarcoplasmic reticulum 30. Sobie EA, Ramay HR: Excitation-contraction cou- meric channel. J Gen Physiol 132:429–446, 2008.
of skinned skeletal muscle fibres. Nature 228:34– pling gain in ventricular myocytes: Insights from a 48. Gyorke S, et al: Regulation of sarcoplasmic reticu-
36, 1970. parsimonious model. J Physiol 587(Pt 6):1293– lum calcium release by luminal calcium in cardiac
12. Fabiato A, Fabiato F: Contractions induced by a 1299, 2009. muscle. Front Biosci 7:1454–1463, 2002.
calcium-triggered release of calcium from the sar- 31. Altamirano J, Bers DM: Voltage dependence of 49. Sobie EA, et al: Termination of cardiac Ca(2+)
coplasmic reticulum of single skinned cardiac cells. cardiac excitation-contraction coupling: Unitary sparks: An investigative mathematical model of
J Physiol 249:469–495, 1975. Ca2+ current amplitude and open channel proba- calcium-induced calcium release. Biophys J 83:59–
13. Stern MD: Theory of excitation-contraction cou- bility. Circ Res 101:590–597, 2007. 78, 2002.
pling in cardiac muscle. Biophys J 63:497–517, 32. Torres NS, et al: Na+ currents are required for 50. Zima AV, et al: Termination of cardiac Ca2+ sparks:
1992. efficient excitation-contraction coupling in rabbit Role of intra-SR [Ca2+], release flux, and intra-SR
14. Isenberg G, Han S: Gradation of Ca(2+)-induced ventricular myocytes: A possible contribution of Ca2+ diffusion. Circ Res 103:e105–e115, 2008.
Ca2+ release by voltage-clamp pulse duration in neuronal Na+ channels. J Physiol 588(Pt 21):4249– 51. Sobie EA, Lederer WJ: Dynamic local changes in
potentiated guinea-pig ventricular myocytes. 4260, 2010. sarcoplasmic reticulum calcium: Physiological and
J Physiol 480(Pt 3):423–438, 1994. 33. Jayasinghe ID, Cannell MB, Soeller C: Organiza- pathophysiological roles. J Mol Cell Cardiol
15. Hayashi T, et al: Three-dimensional electron tion of ryanodine receptors, transverse tubules, and 52:304–311, 2012.
microscopy reveals new details of membrane sodium-calcium exchanger in rat myocytes. 52. Hake J, et al: Modeling cardiac calcium sparks in a
systems for Ca2+ signaling in the heart. J Cell Sci Biophys J 97:2664–2673, 2009. three-dimensional reconstruction of a calcium
122(Pt 7):1005–1013, 2009. 34. Dally S, et al: Multiple and diverse coexpression, release unit. J Physiol 590(Pt 18):4403–4422, 2012.
16. Yu Z, et al: Three-dimensional geometric model- location, and regulation of additional SERCA2 and 53. Antoons G, et al: Mechanisms underlying the fre-
ing of membrane-bound organelles in ventricular SERCA3 isoforms in nonfailing and failing human quency dependence of contraction and [Ca(2+)](i)
myocytes: Bridging the gap between microscopic heart. J Mol Cell Cardiol 48:633–644, 2010. transients in mouse ventricular myocytes. J Physiol
imaging and mathematical simulation. J Struct Biol 35. Greene AL, et al: Overexpression of SERCA2b in 543(Pt 3):889–898, 2002.
164:304–313, 2008. the heart leads to an increase in sarcoplasmic retic- 54. Hashambhoy YL, Greenstein JL, Winslow RL:
17. Savio-Galimberti E, et al: Novel features of the ulum calcium transport function and increased Role of CaMKII in RyR leak, EC coupling and
rabbit transverse tubular system revealed by quan- cardiac contractility. J Biol Chem 275:24722– action potential duration: A computational model.
titative analysis of three-dimensional reconstruc- 24727, 2010. J Mol Cell Cardiol 49:617–624, 2010.
tions from confocal images. Biophys J 36. Vangheluwe P, et al: Ca2+ transport ATPase iso- 55. Saucerman JJ, Bers DM: Calmodulin binding pro-
95:2053–2062, 2008. forms SERCA2a and SERCA2b are targeted to the teins provide domains of local Ca2+ signaling in
cardiac myocytes. J Mol Cell Cardiol 52:312–316, 67. Noble D: Successes and failures in modeling heart 77. Yu Z, et al: Multiscale modeling of calcium dynam-
2008. cell electrophysiology. Heart Rhythm 8:1798– ics in ventricular myocytes with realistic transverse
56. Stokke MK, et al: Ca(2+) wave probability is deter- 1803, 2011. tubules. IEEE Trans Biomed Eng 58:2947–2951,
mined by the balance between SERCA2-dependent 68. Noble D, Garny A, Noble PJ: How the Hodgkin- 2011.
Ca(2+) reuptake and threshold SR Ca(2+) content. Huxley equations inspired the cardiac Physiome 78. Yu Z, et al: Feature-preserving adaptive mesh gen-
Cardiovasc Res 90:503–512, 2011. Project J Physiol 590(Pt 11):2613–2628, 2012. eration for molecular shape modeling and simula-
57. Koh X, et al: A 3D Monte Carlo analysis of the role 69. Pasek M, Simurda J, Christe G: The functional tion. J Mol Graph Model 26:1370–1380, 2008.
of dyadic space geometry in spark generation. role of cardiac T-tubules explored in a model of rat 79. Bers DM, Despa S: Na+ transport in cardiac myo-
Biophys J 90:1999–2014, 2006. ventricular myocytes. Philos Transact A Math Phys cytes: Implications for excitation-contraction cou-
58. Wei S, et al: T-tubule remodeling during transition Eng Sci 364:1187–1206, 2006. pling. IUBMB Life 61:215–221, 2009.
from hypertrophy to heart failure. Circ Res 70. Pasek M, Simurda J, Orchard CH: Role of t-tubules 80. Cheng H, Cannell MB, Lederer WJ: Propagation
107:520–531, 2010. in the control of trans-sarcolemmal ion flux and of excitation-contraction coupling into ventricular
59. Cannell MB, Crossman DJ, Soeller C: Effect of intracellular Ca(2+) in a model of the rat cardiac myocytes. Pflugers Arch 428:415–417, 1994.
changes in action potential spike configuration, ventricular myocyte. Eur Biophys J 41:491–503, 81. Soeller C, et al: Three-dimensional high-resolution
junctional sarcoplasmic reticulum micro- 2012. imaging of cardiac proteins to construct models of
architecture and altered t-tubule structure in 71. Lu S, et al: Multiscale modeling in rodent ventricu- intracellular Ca2+ signalling in rat ventricular
human heart failure. J Muscle Res Cell Motil lar myocytes. IEEE Eng Med Biol Mag 28:46–57, myocytes. Exp Physiol 94:496–508, 2009.
27:297–306, 2006. 2009. 82. Louch WE, Sejersted OM, Swift F: There goes the
60. Brette F, Orchard C: T-tubule function in mam- 72. Yao G, Yu Z: A localized meshless approach for neighborhood: Pathological alterations in T-tubule
malian cardiac myocytes. Circ Res 92:1182–1192, modeling spatial-temporal calcium dynamics in morphology and consequences for cardiomyocyte
2003. ventricular myocytes. Int J Numer Method Biomed Ca2+ handling. J Biomed Biotechnol 2010:503906,
61. Crossman DJ, et al: Changes in the organization of Eng 28:187–204, 2012. 2010.
excitation-contraction coupling structures in 73. Hatano A, et al: Critical role of cardiac t-tubule 83. Lyon AR, et al: Loss of T-tubules and other
failing human heart. PLoS One 6:e17901, 2011. system for the maintenance of contractile function changes to surface topography in ventricular myo-
62. Soeller C, et al: Analysis of ryanodine receptor revealed by a 3D integrated model of cardiomyocytes from failing human and rat heart. Proc Natl
clusters in rat and human cardiac myocytes. Proc cytes. J Biomech 45:815–823, 2012. Acad Sci U S A 106:6854–6859, 2009.
Natl Acad Sci U S A 104:14958–14963, 2007. 74. Hatano A, et al: A three-dimensional simulation 84. Louch WE, et al: T-tubule disorganization and
63. Brette F, Orchard C: Resurgence of cardiac model of cardiomyocyte integrating excitation- reduced synchrony of Ca2+ release in murine car-
t-tubule research. Physiology (Bethesda) 22:167– contraction coupling and metabolism. Biophys J diomyocytes following myocardial infarction.
173, 2007. 101:2601–2610, 2011. J Physiol 574(Pt 2):519–533, 2006.
64. Pasek M, et al: Quantification of t-tubule area and 75. Cheng Y, et al: Numerical analysis of Ca2+ signal- 85. Heinzel FR, et al: Remodeling of T-tubules and
protein distribution in rat cardiac ventricular myo- ing in rat ventricular myocytes with realistic reduced synchrony of Ca2+ release in myocytes
cytes. Prog Biophys Mol Biol 96:244–257, 2008. transverse-axial tubular geometry and inhibited from chronically ischemic myocardium. Circ Res
65. Despa S, et al: Na/Ca exchange and Na/K-ATPase sarcoplasmic reticulum. PLoS Comput Biol 102:338–346, 2008.
function are equally concentrated in transverse 6:e1000972, 2010. 86. Biesmans L, et al: Subcellular heterogeneity of
tubules of rat ventricular myocytes. Biophys J 76. Kekenes-Huskey PM, Cheng Y, Hake J, et al: ryanodine receptor properties in ventricular myo-
85:3388–3396, 2003. Modeling effects of L-type Ca2+ current and Na2+- cytes with low T-tubule density. PLoS One
66. Despa S, Bers DM: Functional analysis of Na+/ Ca2+ exchanger on Ca2+ trigger flux in rabbit myo- 6:e25100, 2011.
K+-ATPase isoform distribution in rat ventricular cytes with realistic T-tubule geometries. Front 87. Song LS, et al: Orphaned ryanodine receptors in
myocytes. Am J Physiol Cell Physiol 293:C321– Comp Physiol Med: Calcium Signaling in Cardiac the failing heart. Proc Natl Acad Sci U S A
C327, 2007. Myocyte 3:351, 2012. 103:4305–4310, 2006.
Theory of Rotors and Arrhythmias
Richard A. Gray
34
CHAPTER OUTLINE Cellular Phenomena (0D)
Cellular Phenomena (0D) 341
A variety of cellular phenomena contribute to irregular cardiac
Cable Phenomena (1D) 343 rhythms. Because of the large number and variety of known
cardiac ion channels, a plethora of studies have been undertaken
Sheet Phenomena (2D) 344
regarding the relationships of genetic and membrane kinetic prop-
Slab Phenomena (3D) 347 erties to arrhythmias. In addition, intracellular signaling as well as
complex subcellular structures (e.g., dyadic clefts, sarcoplasmic
Influence of Whole-Heart Geometry 348
reticulum, mitochondria) can have important physiological effects.
Summary 348 Here, I focus on generic cellular phenomena that “emerge” from
the underlying physiology and discuss their dynamics with regard
to arrhythmias (specifically, irregular rhythms and block).
Sudden cardiac death is the leading cause of fatalities in the The two most fundamental nonlinear properties of cardiac
industrialized world and is most often the result of result of cells are excitability and refractoriness, which are the two necessary
ventricular fibrillation (VF). VF, the most lethal cardiac arrhyth- attributes for a system to exhibit nonlinear wave propagation.
mia, usually occurs through a transition from ventricular tachy- The all-or-none behavior of the cardiac action potential upstroke
cardia (VT). Although the mechanisms underlying the initiation is characterized by the cell’s excitability. Excitability is deter-
of VF are no doubt multifactorial and varied in the diverse patient mined primarily by the diastolic membrane potential and the
population, it is generally believed that most episodes of VF and “threshold” for activation. In normal cells, the dynamics of sub-
VT are maintained by self-sustaining, reentrant waves of electri- threshold responses of the cardiac membrane are determined
cal impulse propagation that are commonly called spiral waves or primarily by the potassium-rectifying current (IK1), and under
rotors. normal conditions, the rapid all-or-none upstroke of the action
Since the landmark publication of Gordon Moe’s multiple potential is a result of the rapid sodium current INa.2 Regenerative
wavelet hypothesis,1 it has been assumed that heterogeneity of depolarization results in threshold-like behavior and occurs when
the refractory period is an essential element of VF. Only in the the magnitude of the inward ionic current (Iion, primarily INa)
past two decades (much longer in Russia) has it become clear becomes greater than that of the outward capacitive current (Ic).
through theory and mathematics, as well as by counter-examples In normal cells, refractoriness coincides with complete recovery
in numerical experiments, that reentry can occur in tissue with of the action potential, and hence the action potential duration
homogeneous refractoriness. Rotors form as a result of regional (APD) provides an excellent surrogate for refractoriness in
conduction block of an action potential wave propagating within healthy tissue. APD is the result of a complex interplay of numer-
the heart; this block can arise via static or functional heteroge- ous voltage- and time-dependent ion currents (primarily potas-
neities (or a combination of both). I restrict myself to the discus- sium, sodium, and calcium) that make up Iion as well as sarcolemma
sion of functional spatial heterogeneities in continuous pumps and intracellular ions. In summary, the cardiac cell trans-
homogeneous cardiac tissue, not because I view static inhomoge- membrane response (ΔVm) to intracellular current injection is a
neities as unimportant, but rather to allow (it is hoped) a clear nonlinear function of the timing, amplitude, and polarity of the
and focused presentation of theoretical concepts. In many “real- stimulus (Figure 34-1).
life” situations, I believe that the underlying mechanisms are a Normally, during rest, the APD accounts for a large portion
variation on the themes presented in this chapter, although I have of the interval between beats; however, during exercise, APD
no doubt that certain phenomena not discussed here (e.g., dis- must shorten to accommodate the increased heart rate. There-
crete effects such as gap junction uncoupling and transmural fore, nearly all of the dynamics of cellular processes are rate
heterogeneities) are important in certain situations. dependent and depend on a plethora of ionic mechanisms. These
Any text on the theory of rotors and arrhythmias will neces- rate dependencies have been extensively studied, usually in
sarily be incomplete and biased; every topic in this chapter could response to electrical stimulation pacing protocols. At low heart
serve as the basis for an extended treatise. First, an important rates, cardiac cells respond to stimuli in a one-to-one manner
disclaimer: I am not a mathematician; my understanding of the (i.e., 1 : 1, or stable behavior). Here I concentrate on the generic
theory and pertinent literature that I present is terribly incom- cellular phenomena that result in the unstable responses (not 1 : 1)
plete. I take a “dimensional” approach and discuss theoretical that can occur when heart rate increases.
concepts relevant to arrhythmias, starting with purely cellular The theoretical study of the response of APD as a function of
characteristics and followed by phenomena in a cable, sheet, slab, heart rate heart began with the seminal study of Nolasco and
and whole heart. Although rotors can form only in two- Dahlen3; “APD restitution” describes the relationship APDi =
dimensional (2D) or three-dimensional (3D) tissue, cellular (0D) f(DIi), where DI is the previous diastolic interval and i is the beat
dynamics play an important role in the initiation and mainte- number. These variables are shown schematically in Figure
nance of cardiac arrhythmias. 34-2, A, where the inter-beat interval is called the basic cycle
341
Stim Stim Stim Stim
BCL
APDi APD i +1
A DIi DIi +1 Time
π
APDi = f (DIi )
APD
∆θ 0
BCL = DIi +1 + APD i

–π
2 Start
–80 1
–60 0
B DI
–40
– 20 –1 S2amp
pre
Vm 0
20 – 2
APD
Figure 34-1. Response of isolated rabbit myocytes to intracellular current stimuli.
Vm was recorded from 10 cells, and S2 stimuli (nA) of either polarity were applied
at various Vm values throughout the action potential (Vmpre ). The response was quan-
X !
tified with a cardiac phase variable, θ = arctan[Vm (t + td) − Vm*, Vm (t) − Vm*], where
t is time, td = 3 ms, and Vm* = −55 mV (threshold). Regenerative depolarization is Start
characterized by Δθ ≈ +π, and regenerative repolarization by Δθ ≈ −π.81
DI min DI DI i + 1
C (2BCL – APDi)
length (BCL). Nolasco and Dahlen3 presented a seminal graphi-
cal method to describe the dynamic response of APD to BCL
(see Figure 34-2, B), which was later formalized in equation form
by Guevara et al.4 If BCL is constant, the solution to the restitu- Long
tion function is APD* = f(DI*) = f(BCL − APD*), where the
asterisks denote the values corresponding to the steady state,
which is also called a fixed point. Through this simplified approach,
APD
APD is considered to be determined only by the preceding DI,

and this fixed point is stable if the slope of the restitution curve Short
at the fixed point is less than one, that is, f′(DI*) < 1, where f′ is
df/dDI. As Nolasco and Dahlen reported, any change in cycle
length generates oscillations as APD “settles” into equilibrium,
as is shown in Figure 34-2, B. Start
Many of the complex unstable (non-1 : 1) responses result D DI
from the fact that a beat is skipped if the stimulus strength and
duration are not sufficient to generate a new action potential. Figure 34-2. Action potential duration (APD) restitution. A, A sequence of identical
There is a minimum DI for which an all-or-none action potential action potentials recorded during constant pacing (i.e., stable response). Basic cycle
can be generated, which leads to a skipped beat when the cell is length (BCL) is the interstimulus interval; DI is the diastolic interval; and APD is the
action potential duration. B, Graphical iteration illustrating the response of APD to
still refractory and recovery of excitability is not sufficient to shortening of BCL to (another) stable fixed point (see text for additional details).
produce regenerative depolarization. A skipped beat allows extra The restitution relationship, f(DI), is shown as a thick black curve. The dashed line
time (essentially an interval equal to 2 × BCL) for recovery of the represents the relationship between APD and DI for the new BCL. The APD response
cell, and hence the APD after a skipped beat tends to be longer is characterized by alternating short and long values of decreasing amplitude.
than that corresponding to the last captured beat (Figure 34-2, C). C, Graphical iteration illustrating the response of APD to shortening of BCL to a
This important cellular dynamic of a stimulus failing to generate fixed point corresponding to a DI shorter than minimal value for regenerative
an all-or-none action potential is the key to localized conduction depolarization (DImin), shown as a vertical dashed red line. This leads to a skipped beat
block as described later in this chapter. (X), and hence the following DI is prolonged (see text for additional details).
If the slope of the APD restitution curve at the fixed point is D, Graphical iteration illustrating the response of APD corresponding to BCL near
an unstable fixed point. The APD response is characterized by alternating short and
greater than one ( f′ [DI*] > 1), the fixed point is unstable and 1 : 1 long values of increasing amplitude until a persistent period of 2 long-short APD
responses are not possible. For monotonic restitution curves, sequence is maintained.
Theory of Rotors and Arrhythmias 343
oscillations in APD will grow until a beat is skipped or a stable current-voltage relationship. fhyp must have an asymmetrical
34
2 : 2 rhythm called alternans is established, as shown in beat Figure “N-shape” (i.e., polynomial with degree ⩾ 3) to sustain nonlinear
34-2, D. If the APD restitution curve is not monotonic, many propagation.15
more types of behavior are possible, including chaotic, 4 : 4, and Of course, the membrane response during propagation is not
3 : 3, even without a skipped beat!5 Discussion of the rich dynam- instantaneous, as is assumed in Equation 2, and numerous inves-
ics of non-monotonic APD restitution relationships6-8 and thor- tigators have studied the effect of a second variable on 1D propa-
ough coverage of alternans are beyond the scope of this chapter. gation theoretically. The effects of the INa activation and
It should be appreciated that the theory and dynamical analysis inactivation gates as well as of maximal conductance on propaga-
of APD restitution can be applied to other cellular properties tion have been studied in depth.16-19 The FitzHugh model
such as excitability,9 latency,5 and intracellular calcium.10 includes a second “recovery” variable (U) with much slower
kinetics compared with V. Thus, combining Equation 2 with
∂w
fhyp(Vm,U) and coupling it with τ w = f w (V , U ) provides the
Cable Phenomena (1D) ∂t
classic FitzHugh-Nagumo equation.12,20 The addition of a slow
The theory of transmembrane potential and current flow in 1D variable (U) allows a separation of the problem into different time
is well established in the form of “the cable equation”: scales, with the result that c = c(U), which allows a quantitative
∂ 2V m ∂V m analysis of conduction velocity (CV) restitution. Just as with
2 − τm − rm I ion = 0 [Eqn 1] APD, CV in cardiac tissue is a function of the previous DI and
∂x 2 ∂t tends to decrease monotonically as DI decreases (and recovery,
rm U, increases), and there is a minimum DI for propagation at a
where x is the direction of propagation, = is the length finite CV. CV is highest for fully recovered (resting or quiescent)
ri + re tissue (Urest) and decreases as U increases until it reaches a critical
constant, τm = rm cm is the membrane time constant, ri is intracel- stall value (Ucrit), that is, propagation occurs for recovery values
lular axial resistance, and re is extracellular axial resistance (low- U < Ucrit, where Ucrit is defined as
ercase parameters represent quantities per unit length).
The conditions for generating a nonlinear propagating wave ∫ f (V , U ) dV = 0
hyp crit
in 1D are considerably more complicated than those correspond- APD alternans can occur uniformly in a cable via the 0D
ing to the elicitation of an all-or-none action potential in 0D. Just mechanism previously described, but the influence of CV restitu-
as for single cells, INa generates the source to sustain propagation tion allows for the uniquely 1D phenomenon of spatially discor-
in 1D, but the load imposed by downstream tissue is significantly dant alternans.21 It has been shown that spatially discordant
greater compared with 0D. To initiate a propagating wave in a alternans can occur in a homogeneous cable exhibiting both
cable, it is not enough to bring a single cell to threshold because APD and CV restitution because repolarization is affected by
a single cell does not provide enough source current to bring spatial coupling.22 Spatially discordant alternans occurs as the
neighboring cells to threshold. A certain “liminal length” is result of a pattern-forming linear instability, and the out-of-
required to generate sufficient inward current to overcome the phase APD spatial patterns can be stationary (resulting from
downstream load (sink) and to initiate a propagating wave in a amplification of a unique finite wavelength mode) or nonstation-
fully excitable cable.11 In addition, there exists a Vm spatial profile ary (resulting from the amplification of a discrete set of complex
shape called a critical nucleus, which can be computed analytically modes); the distance between nodes is independent of cable
for the FitzHugh model,12 which acts as a “threshold” for propa- length.22 It has been shown that CV restitution and the shape
gating wave fronts (profiles above this critical nucleus propagate of action potential recovery can suppress alternans even for
while those lying below it do not).13 f′ > 1.23,24
When an action potential is propagating into quiescent tissue, “Wavelength” (λ) in cardiac electrophysiology is often defined
the subsequent recovery Vm profile is nonuniform; hence if a as the length of excited tissue represented as λ = APD * CV. λ
region of tissue greater than the liminal length is depolarized in tends to decrease with increasing rate (decreasing DI); this fact
the recovering wave tail, the acute response will be one of three is of paramount importance because it allows reentrant waves to
outcomes: (1) no new wave front generation if the entire region form in tissues sizes smaller than the resting wavelength value.
is refractory; (2) “unidirectional block” if part of the region is Reentry is possible in 1D as unidirectional wave propagation
excitable with propagation in the retrograde direction; or (3) within a ring if it is of sufficient length (i.e., >λ). When the
propagation away from the excited region in both directions. ring length is large, the dynamics of the wave front and tail are
Unidirectional block is the hallmark of rotor formation (details stable and an excited region of size λ propagates continuously
below), so details of the stimulation, specifically, area affected, around the ring (CV, APD, and DI are constant along the
amplitude, and timing—called the vulnerable window (or period)— ring at values determined by the restitution curves). When the
are of particular importance. ring length is progressively shortened, the propagation may
Theoretical results regarding the nonlinear cable equation stop (i.e., conduction block) or may transition to irregular behav-
(Equation 1) are too numerous to address here (see References ior, including alternans and quasi-periodicity.25 Courtemanche
14 through 16 for excellent reviews). It is convenient to study et al derived an integro-delay equation to represent this phenom-
stable propagation using a moving coordinate system ξ = X + cT, enon and predicted that loss of stability occurred at the
where c is the conduction speed in normalized units (X = x/ℓ, length where the APD restitution slope was greater than 1 (just
T = t/τm), because the partial differential equation (PDE) in Equa- as in 0D).26 However, their approach assumed repolarization
tion 1 can be converted into an ordinary differential equation was an intrinsic cellular property and therefore that the
(ODE): predicted bifurcation is degenerate (an infinite dimensional
d 2V dV Hopf bifurcation),26 although it was shown that including
−c − f hyp = 0 [Eqn 2] spatial coupling in repolarization removed this degeneracy.27
dξ 2 dξ Cytrynbaum and Kenner have extended the analytical results of
where V represents Vm normalized from 0 to 1, and fhyp(Vm) Courtemanche et al to include the phenomenon of “triggered”
represents a hypothetical, nonlinear, time-independent ion repolarization.23
studied; Rcrit was estimated to be ≈1 mm.30 Thus the 2D equiva-

Sheet Phenomena (2D) lent of liminal length is that a circle greater than radius Rcrit must
be excited to generate a propagating wave because of the effect
The 1D concept of liminal length does not extend directly to 2D of wave front curvature; thus the “liminal area” for 2D propaga-
because a new 2D characteristic comes into play, namely, the tion is π Rcrit
2
.
shape of the wave front. One might think that a circle with diam- The relationships of Rcrit to other length scales in cardiac
eter equal to the liminal length would generate a propagating tissue, most important, the wavelength and the width of the wave
wave in 2D, but it does not; this fact results from the 2D (and front, are relevant to wave propagation in 2D. The width of the
3D) effects of wave front curvature (κ).28 Because a convex wave wave front (wF) is equal to the duration of the upstroke of the
front sees an increased load compared with a planar wave front, action potential multiplied by CV. The important relationship
it propagates more slowly (just based on geometrical factors, i.e., between wF and Rcrit can be understood through the phenomenon
the current density at the wave front); the reverse is true for a of wave front detachment as illustrated in Figure 34-4.31 Imagine
concave front (Figure 34-3). The effect of wave front curvature an action potential propagating horizontally in the lower half of
on propagation speed is well known for small values of κ, and a 2D sheet containing a thin “insulating” nonconductive barrier
this effect is linear.29 that extends part of the way across the sheet. The wave will
propagate as a plane wave exactly analogous to 1D initially, but
CV = CV plane − Dκ [Eqn 3]
where CVplane is the CV at zero curvature (plane wave) and D is
the diffusion coefficient (discussed in detail later). Equation 3 led
investigators to extrapolate this relationship to estimate the “crit-
ical curvature for propagation” (κcrit, corresponding to CV = 0)
and its inverse, the “critical radius for propagation” (Rcrit), as
1 D
κ crit = = . These equations have been confirmed in
Rcrit CV plane w F ~ RC R
clever experiments by Cabo et al, in which conduction speed and
block through narrow isthmuses in sheets of cardiac tissue were
w F >> R C
Concave Planar Convex

κ<0 κ=0 κ>0
CV
κ crit?
w F << R C
0
0
κ
C
Figure 34-3. Effect of wave front curvature (κ) on conduction velocity (CV). The
current density at the wave front is dependent on geometrical factors via the spatial Figure 34-4. Wave front detachment. The behavior of a wave front propagating
Laplacian; this fact is manifested as a dependence of CV on curvature. A concave around a linear (horizontal) obstacle depends on the relationship between the
wave front (κ < 0) propagates more quickly than a planar one (κ = 0), and a convex wave front width (wF) and the critical curvature for propagation (Rcrit). See text for
wave front (κ > 0) propagates more slowly. For low curvature, CV is a linear function details. Light grey represents resting tissue, dark grey the wave front, and white the
of κ with negative slope equal to the diffusion coefficient (see Equation 3). plateau. Red semicircles represent Rcrit.
what will happen when the wave front reaches the end of the around a circular region of tissue with “centripetal wavelets”
barrier? If wF ≫ Rcrit, the wave front “tip” will pivot around the
34
continuously propagating inward and blocks, while the main
end of the barrier, maintaining very close contact with the barrier wave emanates a curved waved (whose shape will be an involute
(Figure 34-4, B). If wF ≪ Rcrit, the wavefront has enough source of a circle based on Huygens’ principle) and emanates away from
“strength” to activate tissue ahead of the wave front (horizontal the circle, as shown in Figure 34-5 (left). Because this model
direction) but is not strong enough to cause regenerative depo- contains no excitable gap, a leading circle reentry circuit cannot
larization in the vertical direction, hence the wave front detaches move and cannot be influenced by external waves (unless the
from the barrier and shrinks, and dies out as it collides with a incoming waves provide greater strength than is provided by the
boundary or the tip propagates around the barrier at a large leading circle wave front).
distance, as shown in Figure 34-4, C. If wF ≈ Rcrit, the wave front A qualitative theory of “spiral wave reentry” extends these
detaches from the barrier but curls back with a circular trajectory ideas by including the effects of wave front curvature (Equation
of radius ≈ Rcrit (Figure 34-4, A; note the geometric relationship 3) and Rcrit. A spiral wave has increasing curvature as one traces
between wF/2 and Rcrit). It is important to note that this phenom- the front toward the center of rotation, assuming polar coordi-
enon of wave detachment can give rise to the development of new nates with r = 0 are defined as the center of rotation, propagation
reentrant waves; many factors (e.g., partial block of INa via drugs fails for r < Rcrit, and there is an excitable “core” region for 0 < r
or fast heart rates decrease CV and hence wF) increase this pos- < Rcrit. Spiral waves can be stable or unstable with a wide variety
sibility and hence are thought to be proarrhythmic.31 In normal of wave tip motion. Stable spirals can rotate around a circular
tissue, the upstroke duration is ≈1 to 2 ms and CV is ≈20 to core or along a line, as is shown in Figure 34-5 (right). Circular
60 cm/s, so wF is ≈0.4 to 1 mm, although this value is dependent cores of radius Rcrit result when Rcrit ≫ λ, and linear cores result
on alterations in cellular excitability that occur with heart rate, when Rcrit ≪ λ, as is shown in Figure 34-5 (right).36 For circular
ischemia, and so forth. This brings us to a very important gen- cores, the wave front is not influenced by the wave tail, but for
eralization, that is, any spatial discontinuity of size much less than linear cores, the wave front impinges on the tail, which is similar
that of wF will generally not influence macroscopic propagation to the leading circle concept of reentry. Spiral waves exhibit a
patterns.32 wide variety of wave tip patterns, in addition to circular and
The simplest form of reentry in 2D is anatomical reentry, which linear, usually referred to as “meander”—a term coined by Art
is similar to propagation around a 1D ring. In anatomical reentry, Winfree.37 Numerical simulations of spiral waves in homoge-
a wave is anchored to an obstacle and can be self-sustaining if the neous sheets have demonstrated that model parameters greatly
perimeter of the obstacle is greater than λ. The dynamics can be influence tip trajectories patterns.38,39 The most common meander
stable or unstable as has been mentioned, but in either case, an patterns are “flower-like” with inward or outward “petals.” A
“excitable gap” exists, that is, a region of the reentrant circuit is stable spiral has an unambiguous period Ts=2π/ωs, whose tip
excitable and can be stimulated by an electrode, or more realisti- circumvents a circle (core) but can transition to a meandering
cally by an incoming wave. As is well known clinically, the flower pattern via a supercritical Hopf bifurcation that introduces
dynamics of anatomical reentry can be easily disrupted (e.g., a second frequency (ω2).40,41 Spiral wave tip trajectories can be
terminated or reset) by pacing from a nearby location at a BCL more complex than rosette patterns,39,42 and we use Winfree’s
shorter than the period of reentry. term “hyper-meander” to describe these patterns. Hyper-meander
In 1946, Wiener and Rosenbleuth published a quantitative
description of wave propagation and reentry around obstacles in
2D and related their findings to cardiac arrhythmias.33 They
presented certain “postulates” from which they derived predic- Leading circle Spiral wave
tions for possible wave dynamics in continuous tissue with
boundaries and holes. These postulates have been described as
“a simplified form of the laws of excitation”34 and are as follows:
(1) an impulse, once started, will spread with constant velocity in
all directions within the tissue; (2) the amplitude of the wave
remains constant and exceeds the threshold of adjacent regions
when these regions are in the resting state; (3) there are three –
states in which the tissue can exist—active (only at the infinitesi- X
mally small wave front), refractory (with constant duration), and
resting. Wiener and Rosenbleuth write, “The law of the propaga-
tion of impulses in a homogeneous 2-D system is Huygens’
principle in its simplest form.” Wiener and Rosenbleuth con-
cluded that reentry in the heart could occur only around holes λ << R C
and not in continuous tissue.33
In contrast to anatomical reentry around obstacles, 2D “func-
Strength
tional” reentry is a self-sustaining wave in a sheet with no obsta-

cles and can occur in homogeneous tissue. In 1948, Selfridge
relaxed the “theorem” of Wiener and Rosenbleuth, which claimed
that a wave front must not have any free ends and must end on
a boundary or obstacle.34 Selfridge claimed that the free ends of λ >> R C
wave fronts could be present in 2D homogenous tissue with no Interval
obstacles, and thus functional reentry was theoretically possible
around a line of distance λ/2, although this situation is unstable Figure 34-5. Functional reentry. Self-sustaining reentry can exist in a homogenous
medium with no obstacles. The leading circle (top right) concept involves no excit-
because there is no distance between the wave tail and the wave
able gap, with the wave front continuously impinging on the wave tail according
front. A similar concept of 2D functional reentry, called the to the strength-interval relationship (bottom left), and with the central core area
“leading circle” theory, was introduced by Allessie in 1977.35 In being refractory and nonexcitable. The spiral wave concept (right) is based on the
this form of reentry, the wave front is located precisely at the idea that wave fronts cannot propagate with extremely high curvature, so the wave
wave tail (i.e., refractoriness is determined by the strength- tip curls around a circle of radius Rcrit if the wavelength (λ) is short (top right) but
interval relationship). In this model, the main wave propagates around a line if is λ long (bottom right).
is characterized by rapid tip motion over large areas and thus is with long DI) above, below, and to the right, at a propagation
particularly important to spiral wave breakup, which is discussed velocity that is dependent on curvature (see Equation 3), so the
later. In addition to intrinsic motion of spiral waves in homoge- speed will increase as it expands (Figure 34-6, B). However,
nous tissue, a variety of inhomogeneities such as fiber curvature43 propagation to the left occurs in tissue that is not fully excitable,
and gradients in cellular properties44 have been shown to elicit has short DI, and hence is slower; in addition, because of the
spiral wave drift. geometry of the wave front and tail, the wave front propagating
The initiation of reentrant waves within the heart occurs via to the left is less curved (more planar). These local effects to the
three basic mechanisms: (1) an unphysiologically large and brief left of the S2 site (slow propagation, increasing DI in front of the
electrical or mechanical disturbance that directly results in the wave, and decreased curvature) act to speed up propagation to
transmembrane potential patterns necessary to generate reentry; the left (see asterisk in Figure 34-6, B). Therefore, propagation
(2) “phase 2” reentry, resulting from spatial differences in “spike- to the left of the S2 site first is slow but then speeds up, and the
and-dome” morphology (mediated by the transient outward wave front resulting from the S2 stimulus becomes nearly circu-
current, Ito) during repolarization45; and (3) a premature wave lar. If the recovery pattern of the S1 beat is not circular (Figure
propagating into a region that encompasses both excitable and 34-6, C), as could occur with spatially discordant alternans in
refractory tissue, resulting from intrinsic and/or functional tissue homogeneous tissue, the initial wave front generated by S2 will
inhomogeneities. This last mechanism most likely results from a be different compared with the uniform APD case. The wave
generic asymmetrical wave front–tail interaction and is thought resulting from the S2 will propagate into resting (i.e., fully excit-
to underlie nearly all spontaneous initiation of reentrant waves. able) tissue above, below, and to the right, just as before, but
Two phenomena play a role in this interaction: a premature propagation directly to the left will not occur. The wave front in
ectopic heartbeat (e.g., resulting from an afterdepolarization) that this case is not a closed curve but is broken, and the two ends of
will propagate at speeds according to its curvature and CV resti- the wave are called wavebreaks and are labeled (“−” and “+”) in
tution, and a nonuniform spatial distribution of refractoriness Figure 34-6, D. The wave front will curl around both wavebreaks
(influenced by APD restitution). Both early and delayed afterde- with opposite chirality; the wave front near the breaks will be
polarizations that result from intracellular calcium overload are affected by local refractoriness directly and via Rcrit (which is a
important cellular mechanisms responsible for the initiation of function of excitability). According to spiral wave theory, the
premature beats, which can trigger arrhythmias. wave front cannot propagate with curvature greater than κcrit, so
It is well known that a wave propagating across a sheet of there will be a region between, and to the left of, the wavebreaks
tissue with a central region of increased refractoriness can result that is not activated. The two ends of the broken wave front will
in the formation of counter-rotating reentrant waves called curl around and merge, resulting in a curved wave propagating
figure-of-eight reentry. Let us examine this phenomenon in detail to the left (see black arrow pointing left in Figure 34-6, D) and
by considering the simple case (Figure 34-6) of pacing the (iso- a nearly planar wave between the wavebreaks moving to the right;
tropic) heart from two different locations at different times. The this occurs at the same time that the tissue to the right of the line
first stimulus generates a circular wave that propagates rapidly connecting wavebreaks is recovering. The balance of source and
into quiescent tissue, and if APD is homogeneous, the wave tail sink at the wave front (as described for 1D) connecting the wave-
will also trace out a circle centered at site 1 (dashed circle), as breaks will determine whether it propagates to the right (see
shown in Figure 34-6, A. A second stimulus applied at the site purple arrow pointing right) or blocks (purple “X”).
indicated by the black circle (site S2) will generate a circular wave In my opinion, the disparity of source and sink strengths of
that will propagate into resting tissue (i.e., fully excitable tissue this wave moving to the right between the wavebreaks is the funda-
mental principle underlying rotor formation. If this wave does
not propagate (sink > source), then all (or nearly all) sites are
activated twice (and only twice) for the S1-S2 protocol described
here, even though two wavebreaks were formed. This is why I’ve
stated that wavebreaks are a necessary but not sufficient condition for
rotor formation.46 If this isthmus wave does propagate to the right
Short
t=0 t=0 (source > sink), a “self-sustaining” rotor is formed. In this case,
? the rotor contains two wavebreaks and one wave front, although
rotors can contain only one wavebreak (if there is a boundary).47
Long
Self-sustaining should be taken with caution, however, because

Short
? the source-sink balance of the wave as it moves through the line

joining the wavebreaks (common isthmus) determines whether it
“lives” (propagates) or “dies” (blocks) during each rotation.46
A C It is generally accepted that VF is the result of multiple unsta-
ble reentrant waves, and because most episodes of VF begin as
tachycardia, many of us think that the transition of a single spiral
wave into many (i.e., spiral wave breakup) is of paramount impor-
2 tance clinically. Here I present a few fundamental mechanisms
2 1
1 underlying spiral wave breakup.
0 0 – 0
As has been described, non-monotonic restitution functions
? or
+ X
(APD, excitability, CV, etc.) can give rise to irregular behavior in
0D and will translate into conduction block and reentry in 2D
with the proper initial conditions. These are described in an
extensive review by Fenton.8 As has been described, monotonic
B D APD and CV restitution can lead to alternans, most notably
Figure 34-6. Rotor formation. During S1-S2 stimulation from different sites, the S2
spatial discordant ones, such that a wave front may encounter
wave front interacts with the S1 wave tail. As a result of the asymmetry of pacing nearby regions of long and short APD. It is well known that this
sites, the circular S2 wave front encounters an asymmetrical recovery gradient. This situation can lead to regional conduction block and rotor forma-
can slow down propagation to the left (A and B) but can also lead to wavebreak tion (as is shown in Figure 34-6).48 Therefore, spatially discordant
and rotor formation (C and D). See text for details. alternans resulting from restitution properties is a mechanism of
spiral wave breakup worthy of considerable study (which it has
34
undergone). However, despite many studies showing the corre-
spondence of APD restitution slope and spiral wave breakup too
numerous to cite, direct evidence of the spatio-temporal patterns
leading to rotor formation consistent with alternans and restitu-
tion theory is very sparse (see Figure 3 in Reference 49 and
Figure 8 in Reference 50).
Since the first experimental evidence of spiral waves in 2D
slices of cardiac tissue was gathered,51 it has been noted that they
may drift, leading to Doppler-induced differences in local activa-
tion periods along the direction of drift.44 In general, the drift
speed of spiral waves tends to be slow; however, the relationship
between changes in period and drift speed is nonlinear, and Gray
et al52 showed examples of Doppler-induced changes in period A
an order of magnitude larger than had previously been demon-
strated in cardiac experiments. Doppler-induced short periods
directly in front of the rapidly moving spiral wave (no period
changes are expected in the perpendicular direction) can lead to
regional conduction block. Thus rapidly moving spirals exhibit-
ing hyper-meander can lead to breakup via the Doppler effect,53
although once again, experimental evidence of this mechanism is
paltry (see Figure 4 in Reference 46).
Of course the same mechanisms that lead to initiation of
rotors (i.e., premature ectopic beats) can disrupt stable rotors, so
afterdepolarizations can lead to destabilization and breakup of
rotors.54,55 In addition, oscillatory membrane properties (e.g., ICaL)
can destabilize spiral waves, giving rise to VF-like patterns, pro-
vided that the characteristic period of the membrane oscillations
(i.e., early after depolarizations) is similar to the spiral wave B
period (Ts).56
Slab Phenomena (3D)

All the 2D phenomena already described are relevant in 3D as
well, and cardiac tissue thinner than one space constant (≈1 mm,
see Reference 57) is effectively 2D. In 2D, a rotor is a spiral wave
rotating about a 0D phase singularity; in 3D, a rotor is a 3D spiral
(i.e., “scroll”) wave that rotates around a 1D phase singularity line
called a filament. These filaments and the resulting scroll waves
can be of a variety of shapes. The simplest is a 3D scroll wave
rotating around a straight filament connecting the epicardial and
endocardial surfaces, in which a spiral wave is evident on both
surfaces (Figure 34-7, A). These filaments can also be nonlinear: C
a “U-shaped” filament gives rise to a figure-of-eight pattern on Figure 34-7. Scroll waves. In 3D, rotors are 3D “scroll” waves that rotate around a
one surface, and a target or “breakthrough” pattern on the oppo- 1D phase singularity line called a filament. Filaments can be linear (A), can be
site surface (Figure 34-7, B). Filaments can even be closed curves, curved such as U-shaped (B), and can form closed rings (C).
giving rise to scroll rings in which no evidence of reentry is found
on any surface (Figure 34-7, C); filaments can even be twisted
and linked.
Considerable mathematical and numerical analysis of the especially in the presence of boundaries (see Reference 64 for an
dynamics of filaments in excitable media has been performed, interesting numerical finding of persistent tangled scroll rings).
once again with a strong Russian influence still being discovered In addition to intrinsic motion of scroll waves in homogeneous
in the West.16,58-61 Some of this work involves diffusion of mul- tissue, a variety of inhomogeneities such as surface curvature,65
tiple variables, although in cardiac tissue, only one voltage dif- wall thickness,66 and transmural fiber rotation67 have been shown
fuses, and mathematically these situations can be quite different. to elicit rotor drift.
It has been suggested that scroll rings shrink at a rate directly Vladimir Biktashev published a landmark paper in 1994 in
related to the diffusion coefficient and inversely related to its which he demonstrated that a filament will shrink or expand
radius,58 and that filament twist shortens the rotation period.59 based on a single coefficient (filament “tension”), which depends
The eikonal formulation provides predictions based on the geo- on the parameters of the medium.68 It was thought that in normal
metrical properties of wave fronts62 and has provided insight into cardiac tissue, which is highly excitable, filaments exhibit positive
scroll wave dynamics in the limit of small curvature and twist. tension and tend to shrink, although during ischemia, in which
For example, scroll waves tend to resist twist (unless locked in, excitability is significantly decreased, filaments may exhibit nega-
as in a ring); the stability of twisted rings depends on medium tive tension and thus may be unstable and may expand (even
parameters; and knotted filaments can exist.63 The eikonal without twist or meander).68 However, more recently, it has been
approach has provided insight into the topological features of shown that negative filament tension can occur during normal
scroll waves but not into quantitative behavior nor persistence, excitability.69 In my opinion, this mechanism of negative filament
tension is a strong candidate to play a primary role in VF under condi- estimated the rotor surface density to be one per 12 cm2, which
tions in which excitability is low and wall thickness is sufficient for the translates to approximately 1 to 2 for rabbits, 5 for sheep, and 15
phenomenon to manifest.47 for healthy humans).46
Another attractive purely 3D mechanism of rotor instability Whether VF is essentially a 2D or a 3D phenomenon is a
involves the high scroll wave front curvature resulting from the question of considerable interest and significance. Exclusively 3D
twisting of fiber rotation across the ventricular wall. Fenton and mechanisms of rotor instability are few (see earlier 3D section).
Karma described how phase shifts of spiral wave rotation across Winfree has argued that rotors might be stable in 2D slices of
the wall can lead to significant transmural gradients of Vm, result- cardiac tissue but not in 3D hearts, whose walls are above a
ing in “twistons” that propagate along the filament and sometimes certain critical thickness that allow unstable 3D filaments77; thus
break off, forming new filaments.70 the dimensionless ratio of wall thickness to rotor diameter seems
of great theoretical importance.
Influence of Whole-Heart Geometry (3D)

Summary
Most life-threatening arrhythmias are thought to be reentrant;
therefore the theoretical study of the existence and stability of The theory of rotors and arrhythmias is very rich and broad. I
rotors and phase singularities in the whole heart is of paramount hope that I have presented the tip of multiple theoretical icebergs
importance. It is intuitive that a single stable rotor or multiple to the cardiac electrophysiological community and expect
stable rotors will give rise to monomorphic ventricular tachycar- continued, fruitful cross-fertilization of experimental and theo-
dia, and a moving rotor will give rise to polymorphic tachycardia. retical approaches in our field. In the last edition of this book, I
It is also intuitive that the faster the speed and the larger the area reviewed experimental studies on VF74; here I concentrate on
corresponding to spiral wave movement,71,72 the more irregular theoretical topics almost exclusively; neither chapter has reviewed
is the corresponding electrocardiogram (ECG), and it has been the vast related numerical studies. One take-home message for
shown that even a single spiral wave gives rise to ECG patterns the reader of this chapter is that there are important temporal
that resemble ventricular fibrillation (VF).52 and spatial scales of significant importance to cardiac wave propa-
The geometry of the whole heart is very complicated. In the gation and arrhythmias. Some important electrophysiological
healthy heart, cardiac myocytes are well connected electrically via time scales include the duration of the action potential upstroke,
gap junctions and are aligned anisotropically in fibers and sheets, the action potential duration (APD), and the period of reentry
which (in the ventricle) rotate across the ventricular wall and are (Ts). Some important spatial scales are liminal length; wave front
arranged into laminar sheets and cleavage planes. The structures width (wF), wavelength (λ); critical curvature (κcrit) and radius
of the atria and ventricles are very different; therefore, the effects (Rcrit) for propagation; and surface area and wall thicknesses of
of geometrical factors on atrial (with its many “holes” capable of the heart. Unfortunately, because of extreme nonlinearities, very
supporting anatomical reentry) and ventricular fibrillation are few analytical closed-form solutions are known; nevertheless,
expected to be quite different.47 theory has provided invaluable quantitative and qualitative insight
Rotors have a characteristic size, so the maximum number of into the extremely complex and clinical significant phenomena of
rotors can be estimated for a given heart size,73 although one can wave propagation and arrhythmias in the heart.
expect the rotors to be much less densely packed than this theo- Nonlinear wave and rotor theory provides a solid foundation
retical maximum value. In addition to naturally (and unnaturally) for the study of reentrant cardiac arrhythmias and has inspired
occurring obstacles (and long-range connections such as trabecu- numerous experimental and numerical studies. Most experimen-
lae), which provide the substrate for anatomical reentry,47 I tal studies have been conducted in mammalian hearts of various
believe that two main geometrical factors are related to the theo- sizes, so differences among species, including heart size, are
retical maximum number of rotors in the heart: surface area and important to remember.73 Most animal studies have been carried
wall thickness. The concept of a “critical mass” required for out in healthy hearts, and we have learned a lot regarding the
sustained fibrillation is well known and intuitive if we consider fundamental mechanisms of wave propagation and reentry in
the underlying cause to be multiple unstable rotors. These rotors cardiac tissue. However, clinical arrhythmias occur in sick indi-
need a certain amount of space to exist, move, and break up in viduals (mostly) as the result of a wide variety of causes. Much
perpetuity. By estimating the CV and the period of rotors during more work is required to understand these highly complex situ-
VF to be approximately 30 cm/s and 100 ms, we can estimate the ations, including developing relevant animal models of disease.
minimum perimeter for the rotating wave tip to be its product In many cardiac electrophysiology basic research labs, experi-
(3 cm); therefore, the minimum rotor diameter is approximately ments and theory have been tightly integrated with the develop-
1 cm (3 cm ÷ π). It should be noted that the rotor period (and ment of experimental methods and design of protocols based on
the APD) in mice and guinea pig can be much shorter than for theoretical ideas. Experimental methods78,79 and signal process-
other mammals,74 and the period of rotors in humans is closer to ing80 are starting to illuminate the rotor filaments within the
200 ms.75,76 In my opinion, this longer rotor period in human VF ventricular wall! I expect future experiments to delineate among
is one of the effects of disease, which paradoxically may increase the various candidate mechanisms of spiral wave breakup and
rotor size and decrease irregularity! introduce new ones.
The surface area necessary to support a 2D rotor therefore
must be larger than 1 cm × 1 cm, and the first high–spatial reso-
lution (video images of Vm) experimental evidence of 2D spiral
waves in cardiac tissue demonstrated stationary and drifting pat-
terns in tissue slices 2 cm × 2 cm.51 Subsequently, video movies Acknowledgments
of spiral waves of Vm were recorded on the heart surface of rabbit
and sheep hearts, in which multiple windings were conspicuously I would like to sincerely thank Pras Pathmanathan for many
absent.71 In fact, although the spatial density of phase singularities useful discussions during the preparation of this chapter. I would
on the surface was high, only 20% of these lasted longer than one especially like to thank my inspirational mentors, Pepe Jalife and
rotation, so the rotor density on the surface was relatively low (we Arkady Pertsov.
27. Comtois P, Vinet A: Stability and bifurcation in an 51. Davidenko JM, et al: Stationary and drifting spiral
References integral-delay model of cardiac reentry including waves of excitation in isolated cardiac muscle.
1. Moe GK, Rheinboldt WC, Abildskov JA: A com-

puter model of atrial fibrillation. Am Heart J
spatial coupling in repolarization. Phys Rev E
Stat Nonlin Soft Matter Phys 68(5 Pt 1):51903,
2003.
Nature 355:349, 1992.
52. Gray RA, et al: Mechanisms of cardiac fibrillation.
Science 270:1222, 1995.
34
67:200, 1964. 28. Fast VG, Kleber AG: Role of wavefront curvature 53. Bar M, Eiswirth M: Turbulence due to spiral
2. Ebihara L, Johnson EA: Fast sodium current in in propagation of cardiac impulse. Cardiovasc Res breakup in a continuous excitable medium. Phys
cardiac muscle: A quantitative description. Biophys 33:258, 1997. Rev E Stat Phys Plasmas Fluids Relat Interdiscip
J 32:779, 1980. 29. Zykov VS: Analytic evaluation of the relationship Topics 48:R1635, 1993.
3. Nolasco JB, Dahlen RW: A graphic method for the between the speed of a wave of excitation in a two- 54. Chudin E, et al: Intracellular Ca(2+) dynamics and
study of alternation in cardiac action potentials. dimensional excitable medium and the curvature of the stability of ventricular tachycardia. Biophys J
J Appl Physiol 25:191, 1968. its front. Biofizika 25:888, 1980. 77:2930, 1999.
4. Guevara MR, Ward G, Shrier A: Electrical alter- 30. Cabo C, et al: Wave-front curvature as a cause of 55. Pollard AE, Lit LM, Rogers JM: Functional reen-
nans and period doubling bifurcations. Biophys J slow conduction and block in isolated cardiac try’s influence on intracellular calcium in the LRd
167, 1984. muscle. Circ Res 75:1014, 1994. membrane equations. IEEE Trans Biomed Eng
5. Chialvo DR, Gilmour RF Jr, Jalife J: Low dimen- 31. Cabo C, et al: Vortex shedding as a precursor of 47:1228, 2000.
sional chaos in cardiac tissue. Nature 343:653, turbulent electrical activity in cardiac muscle. 56. Gray RA, Huelsing DJ: Excito-oscillatory dynam-
1990. Biophys J 70:1105, 1996. ics as a mechanism of ventricular fibrillation. Heart
6. Watanabe M, Otani NF, Gilmour RJ Jr: Biphasic 32. Pertsov AM: Scale of geometric structures respon- Rhythm 5:575, 2008.
restitution of action potential duration and complex sible for discontiuous propagation in myocardial 57. Akar FJ, Roth BJ, Rosenbaum DS: Optical
dynamics in ventricular myocardium. Circ Res tissue. In Spooner PM, Joyner RW, Jalife J, editors: measurement of cell-to-cell coupling in intact
76:915, 1995. Discontinuous Propagation in the Heart, Armonk, heart using subthreshold electrical stimulation.
7. Franz MR, et al: Cycle length dependence of NY, Futura, 1997. Am J Physiol Heart Circ Physiol 281:H533,
human action potential duration in vivo. J Clin 33. Wiener N, Rosenbleuth A: The mathematical for- 2001.
Invest 82:972, 1988. mulation of the problem of conduction of impulses 58. Panfilov AV, Pertsov AM: [Vortex ring in a
8. Fenton FH, et al: Multiple mechanisms of spiral in a network of connected excitable elements, spe- 3-dimensional active medium described by
wave breakup in a model of cardiac electrical activ- cifically in cardiac muscle. Arch Inst Cardiol Mex reaction-diffusion equations]. Dokl Akad Nauk
ity. Chaos 12:852, 2002. 16:205, 1946. SSSR 274:1500, 1984.
9. Delmar M, et al: Ionic basis and analytical solution 34. Selfridge O: Studies on flutter and fibrillation: 59. Mikhailov AS, Panfilov AV, Rudenko AN: Twisted
of the wenckebach phenomenon in guinea pig ven- Some notes on the theory of flutter. Arch Inst scroll waves in active three-dimensional medium.
tricular myocytes. Circ Res 65:775, 1989. Cardiol Mex 18:177, 1948. Phys Lett 109:246, 1985.
10. Shiferaw Y, et al: Model of intracellular calcium 35. Allessie MA, Bonke FI, Schopman FJ: Circus 60. Tyson JJ, Strogatz SH: The differential geometry
cycling in ventricular myocytes. Biophys J 85:3666, movement in rabbit atrial muscle as a mechanism of scroll waves. Int J Bifurc Chaos 1:723, 1991.
2003. of tachycardia. III. The “leading circle” concept: A 61. Pertsov AM, Jalife J: Scroll waves in three-
11. Rushton WAH: Initiation of the propagated distur- new model of circus movement in cardiac tissue dimensional cardiac muscle. In Zipes DP, Jalife J,
bance. Proc R Soc London B 124:210, 1937. without the involvement of an anatomical obstacle. editors: Cardiac Electrophysiology: From Cell to
12. Fitzhugh R: Impulses and physiological states in Circ Res 41:9, 1977. Bedside, ed 2, Philadelphia, 2000, WB Saunders,
theoretical models of nerve membrane. Biophys J 36. Krinsky VI, Efimov IR, Jalife J: Vortices with linear p 336.
1:445, 1961. cores in excitable media. Proc R Soc London Ser 62. Keener JP: A geometrical theory for spiral waves
13. Neu JC, Preissig RS, Krassowska W: Initiation A 437:645, 1992. in excitable media. SIAM J Appl Math 46:1039,
of propagation in a one-dimensional excitable 37. Winfree AT: Spiral waves of chemical activity. 1986.
medium. Phys D 102:285, 1997. Science 175:634, 1972. 63. Keener JP, Tyson JJ: The dynamics of scroll waves
14. Jack JJB, Noble D, Tsien RW: Electric Current 38. Zykov VS: [Cycloid circulation of spiral waves in in excitable media. SIAM Rev 34:1, 1992.
Flow in Excitable Cells, Oxford, 1975, Clarendon an excitable medium]. Biofizika 25:906, 1986. 64. Winfree AT: Persistent tangled vortex rings
Press. 39. Winfree AT: Varieties of spiral wave behavior: An in generic excitable media. Nature 371:233,
15. Scott AC: The electrophyscis of a nerve fiber. Rev experimentalist’s approach to the theory of excit- 1994.
Mod Phys 47:487, 1975. able media. Chaos 1:303, 1991. 65. Davydov VA, et al: Excitation fronts on a periodi-
16. Zykov SV: Simulation of Wave Processes in Excit- 40. Karma A: Meandering transition in two- cally modulated curved surface. Phys Rev Lett
able Media, New York, 1987, Manchester Univer- dimensional excitable media. Phys Rev Lett 85:868, 2000.
sity Press. 65:2824, 1990. 66. Panfilov AV, Keener JP: Generation of reentry in
17. Hunter PJ, McNaughton PA, Noble D: Analytical 41. Barkley D: Euclidean symmetry and the dynamics anisotropic myocardium. J Cardiovasc Electro-
models of propagation in excitable cells. Prog of rotating spiral waves. Phys Rev Lett 72:164, physiol 4:412, 1993.
Biophys Mol Biol 30:99, 1975. 1994. 67. Wellner M, Berenfeld O, Pertsov AM: Predicting
18. Hinch R: An analytical study of the physiology 42. Courtemanche M, Winfree AT: Reentrant rotating filament drift in twisted anisotropy. Phys Rev E
and pathology of the propagation of cardiac waves in a Beeler-Reuter based model of two- Stat Phys Plasmas Fluids Relat Interdiscip Topics
action potentials. Prog Biophys Mol Biol 78:45, dimensional cardiac conduction. Int J Bifurc Chaos 61:1845, 2000.
2002. 1:431, 1991. 68. Biktashev V, Holden A, Zhang H: Tension of orga-
19. Chernyak YB, Starobin JM: Characteristic and 43. Rogers JM, McCulloch AD: Nonuniform muscle nizing filaments of scroll waves. Philos Trans R Soc
critical excitation length scales in 1-D and 2-D fiber orientation causes spiral wave drift in a London Ser A 347:611, 1994.
simulations of reentrant cardiac arrhythmias using finite element model of cardiac action potential 69. Alonso S, Panfilov AV: Negative filament tension
simple two-variable models. Crit Rev Biomed Eng propagation. J Cardiovasc Electrophysiol 5:496, in the Luo-Rudy model of cardiac tissue. Chaos
27:359, 1999. 1994. 17:15102, 2007.
20. Nagumo J, Arimoto S, Yoshizawa H: An active 44. Pertsov AM, et al: Spiral waves of excitation under- 70. Fenton F, Karma A: Vortex dynamics in three-
pulse transmission line simulating nerve axon. Proc lie reentrant activity in isolated cardiac muscle. dimensional continuous myocardium with fiber
Inst Radio Engineers 50:2061, 1962. Circ Res 72:631, 1993. rotation: Filament instability and fibrillation.
21. Watanabe MA, et al: Mechanisms for discordant 45. Lukas A, Antzelevitch C: Phase 2 reentry as a Chaos 8:20, 1998.
alternans. J Cardiovasc Electrophysiol 12:196, mechanism of initiation of circus movement 71. Gray RA, et al: Nonstationary vortexlike reentrant
2001. reentry in canine epicardium exposed to simulated activity as a mechanism of polymorphic ventricular
22. Echebarria B, Karma A: Instability and spatiotem- ischemia. Cardiovasc Res 32:593, 1996. tachycardia in the isolated rabbit heart. Circulation
poral dynamics of alternans in paced cardiac tissue. 46. Gray RA, Pertsov AM, Jalife J: Spatial and tempo- 91:2454, 1995.
Phys Rev Lett 88:208101, 2002. ral organization during cardiac fibrillation. Nature 72. Starmer CF, Starobin J: Spiral tip movement: the
23. Cytrynbaum E, Keener JP: Stability conditions for 392:75, 1998. role of the action potential wavelength in polymor-
the traveling pulse: Modifying the restitution 47. Gray RA, Jalife J: Ventricular fibrillation and atrial phic cardiac arrhythmias. Int J Bifurc Chaos
hypothesis. Chaos 12:788, 2002. fibrillation are two different beasts. Chaos 8:65, 6:1909, 1996.
24. Cherry EM, Fenton FH: Suppression of alternans 1998. 73. Panfilov AV: Is heart size a factor in ventricular
and conduction blocks despite steep APD restitu- 48. Karma A: Spiral breakup in model equations of fibrillation? Or how close are rabbit and human
tion: Electrotonic, memory, and conduction veloc- action potential propagation in cardiac tissue. Phys hearts? Heart Rhythm 3:862, 2006.
ity restitution effects. Am J Physiol Heart Circ Rev Lett 71:1103, 1993. 74. Gray RA: Rotors and spiral waves in the heart. In
Physiol 286:H2332, 2004. 49. Lee MH, et al: Patterns of wave break during ven- Zipes DP, Jalife J, editors: Cardiac Electrophysiol-
25. Vinet A, Roberge FA: The dynamics of sustained tricular fibrillation in isolated swine right ventricle. ogy: From Cell to Bedside, ed 5, Philadelphia,
reentry in a ring model of cardiac tissue. Ann Am J Physiol Heart Circ Physiol 281:H253, 2009, WB Saunders, p 417.
Biomed Eng 22:568, 1994. 2001. 75. Clayton RH, Murray A, Campbell RW: Analysis of
26. Courtemanche M, Glass L, Keener JP: Instabilities 50. Choi BR, Jang W, Salama G: Spatially discordant the body surface ECG measured in independent
of a propagating pulse in a ring of excitable media. voltage alternans cause wavebreaks in ventricular leads during ventricular fibrillation in humans.
Phys Rev Lett 70:2182, 1993. fibrillation. Heart Rhythm 4:1057, 2007. Pacing Clin Electrophysiol 18:1876, 1995.
76. Nanthakumar K, et al: Epicardial organization of 79. Matiukas A, et al: Near-infrared voltage-sensitive isolated rabbit heart. J Cardiovasc Electrophysiol
human ventricular fibrillation. Heart Rhythm 1:14, fluorescent dyes optimized for optical mapping in 10:1452, 1999.
2004. blood-perfused myocardium. Heart Rhythm 81. Gray RA, et al: Effect of strength and timing of
77. Winfree AT: Electrical turbulence in three- 4:1441, 2007. transmembrane current pulses on isolated ven-
dimensional heart muscle. Science 266:1003, 1994. 80. Efimov IR, et al: Evidence of three-dimensional tricular myocytes. J Cardiovasc Electrophysiol
78. Baxter WT, et al: Visualizing excitation waves scroll waves with ribbon-shaped filament as a 12:1129, 2001.
inside cardiac muscle using transillumination. mechanism of ventricular tachycardia in the
Biophys J 80:516, 2001.
Supraventricular Arrhythmias in a
Realistic 3D Model of the Human Atria 35
Javier Saiz and Catalina Tobón
Despite important experimental and clinical evidence

CHAPTER OUTLINE
gained over past decades, the relationships between different
Brief Summary of Atrial Computer Models 351 characteristics of EGM and the propagation patterns that under-
lie them are still far from our complete understanding. Inherent
Building a 3D Model of the Human Atria 352
limitations of the experimental mapping studies because they are
Atrial Activation During Sinus Rhythm 354 performed sequentially, assuming temporal and spatial stability,
together with limited spatial resolution, have contributed to this
Atrial Arrhythmias in Electrically Remodeled Atria 354
incomplete knowledge.
Atrial Fibrillation Ablation 356 In recent years, computational modeling has provided a
framework of multiscale integrated models for the study of
Conclusions 358
cardiac arrhythmias.17,18 Computational cardiac models that sim-
ulate atrial activity have proved to be an important tool in facili-
tating understanding of the complex mechanisms underlying
atrial arrhythmias. Computer simulations of atrial tissue have
Atrial arrhythmias are the most common sustained cardiac provided hypotheses that have been tested experimentally and,
arrhythmias in humans. Atrial arrhythmias, mainly atrial fibrilla- additionally, have been used to investigate and to explain experi-
tion (AF), which often provoke disabling symptoms and severe mental and clinical observations. This chapter reviews the insights
complications,1 are considered major causes of morbidity and provided by these atrial models, with emphasis on the contribu-
mortality. tions of three-dimensional (3D) atrial models, and shows various
Additionally, it has been observed that AF induces changes in examples of atrial arrhythmias simulated using a realistic 3D
the atrial myocardium that help to perpetuate the arrhythmia model of human atria developed by our group.
“AF begets AF.” These changes, called atrial remodeling, include
alterations in the expression of various ion channels affecting
electrical activity of the atrial cells and changes in anatomical
structure. They have been described in animals2 and in humans.3,4 Brief Summary of Atrial Computer Models
Changes in the electrical activity cause a decreased effective
refractory period (ERP) and slowed conduction and reduction in Several computer atrial models have been developed and used to
rate adaptation of ERP,2,3 which may help the initiation and per- study atrial arrhythmias and to evaluate the efficacy of different
sistence of AF, as suggested by experimental studies.2,3 therapeutic approaches. The first was developed by Moe et al19
Despite significant advances in our knowledge about the Using a cellular automata model, they suggested that AF can
mechanisms that lead to the onset and maintenance of atrial result from the activity of multiple independent wave fronts
arrhythmias, they remain incompletely elucidated. It is thought propagating simultaneously throughout the entire atria. The
that atrial arrhythmias can be caused by focal ectopic activity, multiple wavelet hypothesis, as a mechanism underlying AF, was
localized reentry, or multiple propagating wavelets.4-9 It is impor- confirmed experimentally after several years.4 The development
tant to understand the mechanisms of initiation and perpetuation of comprehensive mathematical models of the electrophysiologi-
of atrial arrhythmias because they have a strong influence in the cal activity of human atrial cells20-23 has provided a useful tool for
design of antiarrhythmic therapies. Several experimental and investigating the contribution of different ionic currents to atrial
clinical studies have shown that different mechanisms lead to arrhythmias. Dr. Jalife’s group have been pioneers in combining
differences in the characteristics of spatiotemporal organization experiments with simulation studies to understand the origin and
of atrial arrhythmias. maintenance of AF. Using a simplified 2D model of human atrial
Studies of the spatiotemporal organization of atrial arrhyth- tissue, they have shown the important role of the IK1 current in
mias are currently being performed by analyzing the electro- stabilizing rotors during chronic AF, and how the blockade of IKur
grams (EGM) recorded at different points on the atrial surface or Ito can terminate rotor activity.24 On the other hand, a recent
using different signal analysis techniques, including analysis of study using a 2D model of atrial tissue suggested that electrotonic
EGM morphology,10,11 dominant frequency (DF),7,12,13 and regu- interaction between myocytes and fibroblasts plays an important
larity12 or organization indexes (OI).14 role in the genesis of CFAE.25
In recent high-density mapping studies, areas of complex frac- The first 3D model of the human atria, which combined a
tionated atrial electrograms (CFAE) and high DF have been detailed cellular ionic model (Nygren cell model20) with realistic
proposed as critical regions for the maintenance of AF6,8,15 and geometry, was presented by Harrild and Henriquez.26 They
have become target sites for AF ablation. However, ablation of showed how bundles of atrial muscle including the crista termi-
high-DF areas has proven to be incompletely effective in patients nalis (CT), the pectinate muscles (PM), Bachmannn’s bundle
with persistent AF.12 In addition, it is not universally accepted (BB), and the limbus of the fossa ovalis play an important role in
that improvement in AF ablation occurs after CFAE ablation in determining spread of the propagation wave front. Vigmond
patients with persistent AF.16 et al27 presented a less geometrically realistic model but included
351
morphologic details such as electrical interatrial communica-

tions, anisotropic conduction, and muscular structures. They Building a 3D Model of the Human Atria
observed how specific structures in the atria—coronary sinus
(CS), CT, PM, and orifices such as inferior caval vein (ICV) and Anatomical Characteristics
superior caval vein (SCV)—play an important role in reentrant
activity. Lausanne’s group introduced a simplified fiber structure A realistic 3D model of human atria that includes fiber orienta-
in a 3D model for studying the impact of anisotropy on the tion was previously developed by our group.35-37 The original set
morphology of electrograms10 and later evaluated the effects of of surfaces of the model was based on the work of Harrild and
different ablation patterns in the treatment of AF,28 showing that Henriquez.26 These surfaces were modified in accordance with
the prediction of converting AF to sinus rhythm observed experi- data from the literature and histologic observations.42,43 The
mentally was very similar to what was predicted by the computer model comprises the main structures (Figure 35-1): left and right
model.29 This model, in combination with clinical data, has atrial chambers (LA and RA), 20 pectinate muscles (PM), the
also been used to show that atrial fibrillatory cycle length is an fossa ovalis (FO), Bachmann’s bundle (BB), the crista terminalis
important predictor of AF duration and of the rate of AF (CT), left and right appendages (LAPG and RAPG), left and
termination by ablation.30 In 2006, Seeman et al used data right pulmonary veins (LPV and RPV), superior and inferior
extracted from a visible female dataset to develop a 3D model of caval veins (SCV and ICV), the isthmus of RA, atrioventricular
the atria with great anatomical detail. They included anisotropic rings (AVR) and the coronary sinus (CS). The sinoatrial node
properties of the tissue and heterogeneous electrophysiological (SAN) is situated near the ostium of the SCV.
properties to study the contributions of different anatomical Several experimental observations have been made about the
structures in normal atrial conduction.31 Aslanidi et al presented role of anatomical structure and electrophysiological heterogene-
a multiscale computational model based on this 3D model that ity in atrial electrical activity in both physiological and pathologic
included a human torso model to study the mechanisms underly- conditions.42,44 Notably, this atrial model includes a realistic fiber
ing atrial arrhythmias32 and to test the treatment efficacy of anti- orientation based on histologic observations.42,45 The model was
arrhythmic drugs.33 The effect of electrical versus structural divided into 42 areas (see the colored zones in Figure 35-1), and a
remodeling on AF perpetuation has been studied using a 3D realistic fiber direction was assigned to each region. In Figure
model based on Harrild and Henriquez’s work.26 It has been 35-1, the main areas of the model and their fiber orientation
observed that both electrical remodeling and structural remodel- (indicated by arrows) are shown. It is noted that circulating muscle
ing contribute to APD shortenings, whereas structural remodel- bundles42,46 around the CS, the LPV and RPV, the SCV and ICV,
ing is the main contributor to reduced conduction velocity.34 the AVR, and both appendages (RAPG and LAPG) are present,
All of the described 3D atrial models used a simplistic fiber whereas BB, CT, and PM show aligned fibers along their longi-
structure lacking detailed description of fiber direction in the tudinal axes. The posterior walls of both atria (RA and LA)
atria. In 2009, our group developed a 3D atrial model that inte- consist of mainly vertical fibers, whereas the fibers between the
grated realistic geometry and structure, as well as heterogeneous SCV and the ICV have a horizontal direction, provoking a
electrical properties, with detailed fiber orientation in the whole complex arrangement of vertical, horizontal, and circular fibers
atria.35-39 In recent reports,40,41 semiautomatic methods were used in the ostium of the pulmonary veins.45
to incorporate atrial anisotropy (including fiber orientation) and
heterogeneities into patient-specific 3D models based on geom- Electrophysiological Models
etries obtained from magnetic resonance imaging (MRI) data,
gaining an important step toward the use of specific atrial models Nygren’s model20 of human atrial action potential (AP) was used
in the clinic. to simulate cellular electrical activity. Electrophysiological
SCV LA
BB
RA
RAPG
FO SAN
RPV
LPV
CT
PM
AVR Isthmus LAPG
CS ICV
Figure 35-1. Frontal and dorsal views of the human atrial model. Colors represent areas with different fiber orientation. Arrows indicate the orientation of the fibers in the
main areas of the atria. AVR, Atrioventricular rings; BB, Bachmann’s bundle; CS, coronary sinus; CT, crista terminalis; FO, fossa ovalis; ICV and SCV, inferior and superior caval
veins; LA and RA, left and right atria; LAPG and RAPG, left and right appendages; PM, pectinate muscles; RPV and LPV, right and left pulmonary veins; SAN, sinoatrial node.
(From Tobón, C, Ruiz-Villa C, Heidenreich E. et al: Three dimensional human atrial model with fiber orientation. Electrograms and arrhythmic activation patterns relationship.
Plos One 8(2):1–13, 2013, Fig. 1.)
Supraventricular Arrhythmias in a Realistic 3D Model of the Human Atria 353
heterogeneity was included to reproduce AP in different zones 1 ∂V m

∇ ⋅ ( D∇ V m ) = C m + I ion − I stim [Eqn 1]
35
of the atria47: PM, CT, AVR, left and right appendages (APG),
Sv ∂t
and atrial working myocardium (AWM), which includes the
remaining atrial structures. To obtain these AP models, maximum
conductance of It, IKr, and ICaL was modified.35 To reproduce where Sv corresponds to the surface-to-volume ratio, D is the
electrical remodeling conditions, changes in conductance and conductivity tensor, Cm is the specific membrane capacitance
kinetics of different ionic channels observed in experimental (50 pF), Iion is the total ionic current that crosses the membrane
studies of permanent AF3,48 have been incorporated into these AP cells, Vm is the membrane potential, and Istim is the stimulus
models. Maximum conductance for IK1 was increased by 250% current. The monodomain equation was solved using a finite
and for ICaL and It was decreased by 74% and 85%, respectively; element method.
kinetics of the fast inactivation of ICaL was increased by 62%, the Conductivity values were assigned to obtain realistic conduc-
activation curve of It was shifted by +16 mV; and the inactivation tion velocities observed in the different atrial zones: 25 cm/s in
curve of INa was shifted by +1.6 mV.49 very slow regions (SAN), 54 cm/s in slow regions (PV), 120 cm/s
Figure 35-2 shows the last AP obtained when a train of 10 in fast regions (BB, limbus of the FO and PM), and 143 cm/s in
stimuli at a basic cycle length of 1000 ms was applied for the very fast regions (CT bundle), whereas the remaining atria had a
different atrial cellular models considered (AWM, PM, CT, APG, conduction velocity of 69 cm/s.50,51 An anisotropic ratio of conduc-
and AVR) under both physiological (control) (Figure 35-2, A) and tivity was also introduced in agreement with experimental data.51,52
remodeling conditions (Figure 35-2, B). The APD90 values (see The isthmus of the RA and the SAN were set isotropic, whereas
table in Figure 35-2, C) show that under control conditions, an anisotropic ratio of 1 : 2 was used for BB, limbus of FO, PV, and
APD90 ranged from 180 ms to 307 ms. It is important to note, AWM.52 Finally, an anisotropy ratio of 1 : 9 was used for CT.51
however, that atrial remodeling decreased the APD90 values in Unipolar pseudo-electrograms (EGM) were simulated in
the whole atria (ranging from 56 ms to 92 ms) as well as the APD more than 43,000 points along the atrial surface. The extracel-
dispersion. lular potential (Φe) was calculated by the following equation53:
Figure 35-2, D illustrates restitution curves for AWM, for
both physiological (control) and electrically remodeled cells. It is
 1 
possible to observe how the remodeling condition provokes a Φ e ( r ) = − K ∫∫∫ ∇′V m ( r ′ ) ⋅ ∇′  dv [Eqn 2]
reduction in APD and in ADP rate adaptation, inducing a  r ′ − r 
decrease in ERP and a reduction in rate adaptation of ERP, which
will facilitate ectopic beat propagation. where ∇′Vm is the spatial gradient of transmembrane potential
Vm, K is a constant that includes the ratio of intracellular and
extracellular conductivities, r is the distance from the source
Propagation and Atrial Electrograms point (x, y, z) to the measuring point (x′, y′, z′), and dv is the dif-
ferential volume.
To simulate the electrical propagation of AP, we used the EGM on the entire atrial surface were processed with a 40- to
monodomain model described by the following reaction-diffusion 250-Hz band-pass filter, rectified, and further low-pass filtered at
equation: 20 Hz.54 Spectral analysis of the signals was then performed with
20 Control Remodeled
AWM AWM
PM PM
0 CT CT
APG APG
AVR AVR
–20
mV
–40
–60
–80
50 100 150 200 250 300 50 100 150 200 250 300
A t (ms) B t (ms)
AWM APD 90 (ms)
300
CT PM APG AVR AWM
200 Control
APD90 Control (ms) 307 237 245 180 282 Remodeled
100
APD90 Remodeled (ms) 92 73 78 56 80
0
200 400 600 800 1000
C D Cl (ms)
Figure 35-2. AP for the different types of atrial cells: crista terminalis (CT), pectinate muscles (PM), left and right appendages (APG), atrioventricular rings (AVR), and atria
working myocardium (AWM), which includes the rest of the tissue; under (A) physiological conditions (control) and (B) remodeling conditions (remodeled), at a
BCL = 1000 ms. C, APD90 values. D, Restitution curves for physiological and remodeling conditions for cells of the AWM.
(From Tobón, C, Ruiz-Villa C, Heidenreich E. et al: Three dimensional human atrial model with fiber orientation. Electrograms and arrhythmic activation patterns relationship.
Plos One 8(2):1–13, 2013, Fig. 3.)
fast Fourier transform (FFT), obtaining a spectral resolution of atrial depolarization occurred at 120 ms in normal conditions and
0.12 Hz. The DF, defined as the frequency corresponding to the at 138 ms in remodeled atria, ending in the distal LAPG. These
highest peak of the power spectrum, was determined. To measure values are in agreement with the values observed experimentally.
frequency variability in the spectrum, the OI was also calculated. Canavan et al55 showed that under physiological conditions, the
Spectral power of the DF and its three harmonic peaks were last activation of atrial tissue occurred just before 120 ms. Another
calculated by computing the area under the peaks. The OI was study56 observed the last activation of LA tissue at approximately
obtained as the ratio of this spectral power to the total power of 116 ms. It is interesting to note that the entire atria repolarized
the spectrum.14,54 False color DF and OI maps were constructed faster in remodeled conditions than under physiological condi-
by assigning a color between blue and red to each point for lowest tions (see snapshots at 135 ms in Figure 35-3).
and highest values, respectively.
Simulation Protocols Atrial Arrhythmias in Electrically

Remodeled Atria
Different atrial arrhythmias were generated using an S1-S2 pro-
tocol as follows: A train of stimuli (rectangular current pulses of After the 10th pulse was applied in the SAN area of the remod-
6 ms of duration and 30 µA of amplitude, stimulation area of eled atria (see Figure 35-3, B), an ectopic focus (S2) was applied
approximately 10 mm2) with a basic cycle length of 1000 ms was at different points of the atria. After the application of the bursts
applied during 10 seconds in the SAN area to simulate sinus of six beats (S2) in the center of the posterior wall of the LA, the
rhythm (S1). Two different ectopic stimuli (S2) were simulated. wave front induced by the ectopic focus fragmented, generating
After the last beat of the S1 stimulation, a burst of six ectopic a reentrant activity (Figure 35-4). Different reentries in the RPV,
beats or a continuous ectopic focus was delivered at a cycle length SVC, and CT and wave fragmentations were observed (see snap-
(CL) of 130 ms (S2). Once an ectopic focus was applied, the shots at 3140 ms and 4824 ms in Figure 35-4, A). Thereafter, this
simulation was maintained for 10 seconds. complex activity became a more stable reentrant tachycardia
rounded to RPV (see snapshot at 9317 ms in Figure 35-4, A),
showing fragmentation of the wave front only in the CT arch
area. The EGM calculated in three points of the LA—center of
Atrial Activation during Sinus Rhythm the posterior wall, center of the superior wall, and LAPG—and
in three points of the RA—center of the posterior wall, CT, and
Normal atrial propagation patterns in sinus rhythm were simu- RAPG—are shown in Figure 35-4, B. Figure 35-4, B also depicts
lated by applying a periodic stimulation of 10 beats, at a basic the DF and OI values calculated from the spectral analysis of
cycle length of 1000 ms, in the SAN area. Figures 35-3 depicts these EGM, using the method described in previous sections. In
various snapshots of the propagation of the last beat applied for the entire atria, the EGM showed a stable and regular atrial
both physiological (Figure 35-3, A) and remodeling (Figure activity (typical of a macro-reentrant atrial tachycardia) with
35-3, B) conditions. As shown in the figures, the stimulus applied single potentials, with the exception of the CT, where double
in the SAN region caused initiation of an almost triangular wave potentials were observed as the result of fragmentation of the
front that quickly spread to the ICV favored by high conductivity wave fronts that occurred in this region. Spectral analysis of the
and anisotropy of the CT (see snapshots for 25 ms in Figure EGM also indicated organized and periodic overall activity, with
35-3, A, B). The depolarizing wave propagated to the anterior DF values near 6 Hz and high OI values (near to 1), except in
septal portion of the LA through the interatrial BB, inducing the the CT with an OI of 0.84. It is noteworthy that the model
first activation of the LA after 46 ms (in normal atria) and 54 ms reproduced stable and regular activation during reentrant atrial
(in remodeled atria) from the SAN activation. arrhythmias experimentally shown,57,58 and that the high OI
Figure 35-3 also illustrated (see snapshots at 94 ms) the delay values corresponded with this high regularity. Moreover, the
in propagation induced by remodeling conditions. The entire model also reproduced the double potentials in the EGM
Physiological
mV
25 ms 94 ms 135 ms
20
0
–20
–40
–60
A –80
Remodeled
25 ms 94 ms 135 ms mV
20
0
–20
–40
–60
B –80
Figure 35-3. Snapshots of the propagation of the last stimulus applied in the SAN area (the 10th beat) for both physiological (A) and electrically remodeled atria (B). Colors
represent the distribution of membrane voltage (AP values in mV) in the entire atria. Depolarized zones are illustrated by a red color, whereas repolarized zones are repre-
sented by a blue color.
3140 ms 4824 ms 9317 ms

mV
20
–5
–30
35
–55
–80
mV
20
–5
–30
–55
–80
A
Center of the LA posterior wall. DF = 5.9 Hz OI = 0.99 Center of the RA posterior wall. DF = 5.9 Hz OI = 0.96
Center of the LA superior wall. DF = 5.9 Hz OI = 0.98 Crista terminalis (CT). DF = 5.9 Hz OI = 0.84
Left appendage (LAPG). DF = 5.9 Hz OI = 0.95 Right appendage (RAPG). DF = 6.0 Hz OI = 0.92
2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 10
B t (s) t (s)
Figure 35-4. A, Snapshots of the reentrant activity observed in electrically remodeled atria when a burst of six beats (CL = 130 ms) was applied in the center of the LA
posterior wall. Colors represent the distribution of membrane voltage (in mV). Depolarized zones are illustrated by a red color, whereas repolarized zones are represented
by a blue color. Different••reentries and wave fragmentations are observed before a reentrant activity is established around the RPV. Arrows indicate the direction of wave
front propagation. The “ ” indicates conduction block. B, EGM at different points of the atria. Values of DF and OI are illustrated. See text for details.
recorded at points along the blocking line that become the CT, rotor was triggered in the posterior wall of the LA, while frag-
in accordance with experimental observations.11 mentations and collisions of wave fronts, mainly at the level of
Upon application of the burst of six ectopic beats (S2) in the CT, were observed (see snapshot at 3701 ms in Figure 35-6, A).
SCV, different reentrant mechanisms were observed (Figure AF was then maintained by figure-of-eight reentries and macro-
35-5). During the first few seconds, a reentrant tachycardia was reentries in the LA, whereas fragmentation of wave fronts and
maintained by a reentry wave front anchored in the CT area (see reentries around the CT continued in the RA (see snapshot at
snapshot at 2549 ms in Figure 35-5, A). Very quickly (at approxi- 9303 ms in Figure 35-6, A); at the end of simulation, it was pos-
mately 5.5 s after initiation of reentrant tachycardia), the reentry sible to observe a rotor in the posterior wall of the LA (see
that maintained the tachycardia collided and became a macro- snapshot at 9663 ms in Figure 35-6, A).
reentry, turning the tricuspid annulus, starting a typical atrial The EGM calculated at various points in the atria (see Figure
flutter (AFL) (see snapshots at 5789 ms and 9965 ms in Figure 35-6, B) present mainly polymorphic single potentials. Double
35-5, A). The EGM calculated at the six different points of the potentials and fragmented potentials (CFAE) are observed at
atria showed a change in morphology after the second 5.5, caused pivot points, conduction blocks, and wave fragmentations. CL
by the evolution from atrial tachycardia to typical AFL (Figure variations are also shown. Spectral analysis showed similar DF
35-5, B). During the first 5.5 seconds, the potentials are mainly values and low OI values for the entire atria, accordingly with its
simple in the entire atria, except in the CT, where double poten- irregular activity. To show distribution of DF and OI values
tials are provoked by the wave front turning around the blocking throughout the whole atria, we developed DF and OI maps by
line in this area. Thereafter, the model reproduced a typical AFL processing more than 43,000 unipolar EGM distributed along
episode with the reentrant excitation traveling around the tricus- the atrial surface. DF maps (see Figure 35-6, C) depict atrial
pid valve ring, similar to that experimentally observed.59 This zones with slight DF differences around a value of 5.8 Hz,
macro-reentry propagated to the rest of the atria with a 1 : 1 whereas OI maps (see Figure 35-6, D) show how the overall atrial
pattern of activation, according to experimental observations,60 activity presents a low degree of organization (lower than 0.7).
provoking uniform and regular EGM in the entire atria with DF The application of an ectopic focus S2 of continuous activity
values near 5 Hz, in agreement with the range observed in at the base of the RPV induced multiple fronts of reentry with
humans (4.2 to 5.8).61 The OI values near to the unit in the whole irregular trajectories (see Figure 35-7). Zones of LA are activated
atria also indicated high regularity in overall atrial activation. It periodically by the ectopic focus, and wave fragmentation and
is important to note that spectral analysis of the EGM was per- collisions are mainly observed throughout the rest of the atria
formed on the AFL pattern (the last 4.5 seconds of simulation). (see snapshots at 3825 ms and 5928 ms in Figure 35-7, A). Rotors
When the ectopic focus S2 (burst of six beats) was applied at that collided and fragmented creating new wavefronts are mainly
the base of the RPV, an AF maintained by different reentrant observed in RA (see snapshot at 7906 ms in Figure 35-7, A). In
patterns in the LA was observed (see Figure 35-6). Initially, a points near the ectopic focus, the EGM showed rapid and regular
2549 ms 5789 ms 9965 ms

mV
20
–5
–30
–55
–80
mV
20
–5
–30
–55
–80
A
Center of the LA posterior wall. DF = 5.0 Hz OI = 0.99* Center of the RA posterior wall. DF = 5.0 Hz OI = 0.95*
Center of the LA superior wall. DF = 5.0 Hz OI = 0.66* Crista terminalis (CT). DF = 5.0 Hz OI = 0.98*
Left appendage (LAPG). DF = 5.0 Hz OI = 0.97* Right appendage (RAPG). DF = 5.0 Hz OI = 0.93*
2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 10
B t (s) t (s)
* AFL values (from 5.5 to 10 s)

Figure 35-5. A, Snapshots of the reentrant activity observed in the electrically remodeled atria when a burst of six beats (CL = 130 ms) was applied in the SCV. Colors
represent the distribution of membrane voltage (in mV). Depolarized zones are illustrated by a red color, whereas repolarized zones are represented by a blue color. In the
first seconds, reentrant activity was maintained by reentry anchors in the CT. At 5.5 s, this reentrant activity became a typical AFL. Arrows indicate the direction of wave
front propagation. B, EGM at different points of the atria. Values of DF and OI for these different points, calculated taking into account only the flutter activity (AFL), are
illustrated. See text for details.
atrial activation with the highest DF and OI values, 7.7 Hz and activity of the atria during sinus rhythm.63 Since the introduction
0.83, respectively, in the EGM of the center of the LA posterior of Maze III, various ablation strategies have been tested to limit
wall (see Figure 35-7, B). However, in the posterior wall of the the number and length of ablation lines and, as a result, reduce
RA and in the CT, the EGM presented polymorphic potentials undesirable damage to the atria.65,66 Evaluation of the effective-
and CL variations. For this case, the DF map (see Figure 35-7, C) ness of different ablation patterns is usually performed in experi-
indicated that the entire LA was activated homogeneously at high mental and clinical studies. Recently, computer models have also
frequency (around 7.7 Hz), whereas RA was activated at a lower been used to systematically evaluate the efficacy of different com-
frequency (around 5 Hz) with slightly higher DF values in small binations of the ablation lines.29
areas of SCV, tricuspid ring, and posterior wall. The OI map (see Using the 3D model of human atria, we have simulated the
Figure 35-7, D) showed a high degree of organization in various effects of application of two ablation procedures on the AF
zones of the LA, such as in part of the posterior wall and in areas pattern induced by application of a burst of six beats at the base
of the inferior wall, with the highest OI values (close to 1) at the of the RPV (shown in Figure 35-6). After AF was initiated, 10
site of the application of ectopic focus and in neighboring areas. seconds from starting the application of the burst, ablation lines
In the RA, a low degree of organization can be observed with OI were introduced in the model following two patterns: Maze III
values of ≈0.3 in some areas. These results are in accordance with (Figure 35-8, A) and left-sided partial Maze (Figure 35-8, B). The
clinical observations in which low OI values were observed in ablation lines were of two to three elements of thickness and zero
areas where multiple wave fronts interacted, whereas high OI conductivity, simulating perfect transmural obstacles to the wave
values were associated with the source that maintained the AF.54 front propagation. Maze III consisted of ten lines localized in
both RA and LA,62 whereas the left-sided partial Maze consisted
of only four lines in LA.66 Figure 35-8, A depicts AF progression
after the application of Maze III. It is possible to observe that
Atrial Fibrillation Ablation Maze III finished AF in less than 1 second, and the entire atria
reached a resting state awaiting the next SAN beat (see snapshot
Surgical ablation is one of the most common treatments for AF. at 314 ms in Figure 35-8, A). The ablation lines in LA blocked
The surgical Maze III procedure, introduced by Cox et al,62 con- the rotors in this area, which maintained FA (see previous section).
sists of creating lines of blockage to prevent all possible reentrant Simultaneously, lines in the RA blocked reentrant wave fronts
circuits in both atria. Although Maze III has proven its efficacy and the fragmentations observed in this area. However, when
in terms of AF termination,63,64 it affects the correct mechanical left-sided partial Maze was applied, the AF did not stop but
3701 ms 9303 ms 9663 ms
mV
20
–5
–30
–55
–80
mV
20
–5
–30
–55
–80
A
Center of the LA posterior wall. DF = 6.0 Hz OI = 0.65
Center of the RA posterior wall. DF = 5.9 Hz OI = 0.42
Hz
C 4.5 5.5 6.5
Crista terminalis (CT). DF = 6.0 Hz OI = 0.49
OI
2 3 4 5 6 7 8 9 10
0.0 0.5 1.0
B t (s) D
Figure 35-6. A, Snapshots of the AF induced in the electrically remodeled atria when a burst of six beats (CL = 130 ms) was applied at the base of the RPV. Colors represent
the distribution of membrane voltage (in mV). Depolarized zones are illustrated by a red color, whereas repolarized zones are represented by a blue color. Different reentries
are observed in LA, whereas wave fragmentations are shown in RA. Arrows indicate the direction of wave front propagation. B, EGM at different points of the atria. Values
of DF and OI are illustrated. C and D, DF and OI maps calculated by processing more than 43,000 EGM; red color indicates the highest values and blue color the lowest
values, according to the color scale. See text for details.
3825 ms 5928 ms 7906 ms

mV
20
–5
–30
–55
–80
mV
20
–5
–30
–55
–80
A
Center of the LA posterior wall. DF = 7.7 Hz OI = 0.83
Center of the RA posterior wall. DF = 5.2 Hz OI = 0.25

Hz
C 2.0 5.0 8.0
Crista terminalis (CT). DF = 5.2 Hz OI = 0.32
OI
2 3 4 5 6 7 8 9 10
0.0 0.5 1.0
B t (s) D
Figure 35-7. A, Snapshots of the AF induced in the electrically remodeled atria when a continuous ectopic focus was applied at the base of the RPV. Colors represent the
distribution of membrane voltage (in mV). Depolarized zones are illustrated by a red color, whereas repolarized zones are represented by a blue color. Regular activation
is observed in points near the focus area, whereas rotors, fragmentation, and collisions are observed in the rest of the atria. Arrows indicate the direction of wave front
propagation. B, EGM at different points of the atria. Values of DF and OI are illustrated. C and D, DF and OI maps calculated by processing more than 43,000 EGM; red color
indicates the highest values and blue color the lowest values, according to the color scale. See text for details.
Maze III
46 ms 192 ms 314 ms mV
20
0
–20
–40
–60
–80
mV
20
0
–20
–40
–60
A –80
Left-sided partial Maze
46 ms 192 ms 9660 ms mV
20
0
–20
–40
–60
–80
mV
20
0
–20
–40
–60
B –80
Figure 35-8. Snapshots illustrating the evolution of the AF obtained by the application of a burst of six beats at the base of the RPV (see Figure 35-3) when two different
ablation patterns were applied. Ablation lines are indicated in the left figures and consist of applying zero conductivity in elements along the line, 10 seconds after applica-
tion of the burst. A, Maze III ablation patterns finished AF very quickly. B, Left-sided partial Maze did not stop AF, but it converts AF into reentrant tachycardia. Arrow
indicates the direction of the reentry. See text for details.
became a reentrant tachycardia around the CT (see Figure

35-8, B). In this case, the ablation lines blocked the rotors in the Conclusions
LA, but RA could maintain the reentrant activity around a block-
ing line in the CT (see snapshot at 9660 ms in Figure 35-8, B). The treatment of atrial arrhythmias, and particularly of atrial
Fragmentations and collisions were not observed in the RA, fibrillation, is far from completely satisfactory. To the complexity
thereby a stable reentrant tachycardia was established. It has of the responsible mechanisms, we have to add the inherent limi-
been reported that atrial tachycardia follows AF ablation in a tations of experimental studies. This chapter described some
high percentage of patients, ranging from 1% to 50%.16,67 examples of how a realistic human atrial model can facilitate
Although surgical ablation procedures are common interven- understanding of the relationship between different characteris-
tions, and a great number of investigators have attempted tics of signals recorded in the atrial surface and the propagation
to optimize their efficacy, no universally accepted strategies for patterns that induced them. 3D realistic human atrial computer
ablation are known, and only international recommendations can models have proved to be a complementary tool for the study of
be reached, such as the recent Heart Rhythm Society (HRs)/ atrial arrhythmias and for the advancement of improved therapies
European Heart Rhythm Association (EHRA)/European by offering the possibility of testing the effects of different antiar-
Carbon Arrhythmia Society (ECAS) expert consensus statement rhythmic strategies in a human model “without irreversible
on catheter and surgical ablation of atrial fibrillation.16 damage to the patients.”
originating in the pulmonary veins. N Engl J Med 11. Konings KT, Smeets JL, Penn OC, et al: Configu-
References 339:659–666, 1998. ration of unipolar atrial electrograms during elec-
6. Mandapati R, Skanes A, Chen J, et al: Stable micro- trically induced atrial fibrillation in humans.
1. Wolf P, Benjamin E, Belanger A, et al: Secular reentrant sources as a mechanism of atrial fibrilla- Circulation 95:1231–1241, 1997.
trends in the prevalence of atrial fibrillation: The tion in the isolated sheep heart. Circulation 12. Sanders P, Berenfeld O, Hocini M, et al: Spectral
Framingham study. Am Heart J 113:790–796, 101:194–199, 2000. analysis identifies sites of high-frequency activity
1996. 7. Berenfeld O, Mandapati R, Dixit S, et al: Spatially maintaining atrial fibrillation in humans. Circula-
2. Nattel S: New ideas about atrial fibrillation 50 distributed dominant excitation frequencies reveal tion 112:789–797, 2005.
years on. Nature 415:219–226, 2002. hidden organization in atrial fibrillation in the 13. Lin YJ, Tai CT, Chen SA: Can mapping and abla-
3. Workman AJ, Kane AK, Rankin AC: The contribu- Langendorff-perfused sheep heart. J Cardiovasc tion of atrial fibrillation be guided by frequency
tion of ionic currents to changes in refractoriness Electrophysiol 11:869–879, 2000. analysis of fibrillatory waves? J Cardiovasc Electro-
of human atrial myocytes associated with chronic 8. Mansour M, Mandapati R, Berenfeld O, et al: Left- physiol 17:S44–S49, 2006.
atrial fibrillation. Cardiovasc Res 52:226–235, to-right gradient of atrial frequencies during acute 14. Everett TH, Wilson EE, Verheule S, et al:
2001. atrial fibrillation in the isolated sheep heart. Circu- Structural atrial remodeling alters the substrate
4. Allessie MA, Bonke FI, Schopman FJ: Circus lation 103:2631–2636, 2001. and spatiotemporal organization of atrial
movement in rabbit atrial muscle as a mechanism 9. Jalife J: Rotors and spiral waves in atrial fibrillation. fibrillation: A comparison in canine models of
of tachycardia. III. The “leading circle” concept: A J Cardiovasc Electrophysiol 14:776–780, 2003. structural and electrical atrial remodeling. Am J
new model of circus movement in cardiac tissue 10. Jacquemet V: A Biophysical Model of Atrial Physiol Heart Circ Physiol 291:H2911–H2923,
without the involvement of an anatomical obstacle. Fibrillation and Electrograms: Formulation, Vali- 2006.
Circ Res 41:9–18, 1977. dation and Applications [PhD thesis], Lausanne, 15. Nademanee K, Lockwood E, Oketani N, et al:
5. Haissaguerre M, Jais P, Shah DC, et al: Spontane- École Polytechnique Fédérale de Lausanne, Catheter ablation of atrial fibrillation guided by
ous initiation of atrial fibrillation by ectopic beats 2004. complex fractionated atrial electrogram mapping
of atrial fibrillation substrate. J Cardiol 55:1–12, 34. Krogh-Madsen T, Abbott GW, Christini DJ: studied during open heart surgery. Eur Heart J
2010. Effects of electrical and structural remodeling on 19:293–300, 1998.
16. Calkins H, Kuck KH, Cappato R, et al: HRS/
EHRA/ECAS Expert consensus statement on
catheter and surgical ablation of atrial fibrillation:
atrial fibrillation maintenance: A simulation study.
PLoS Comput Biol 8:e1002390, 2012.
35. Tobón C, Ruiz-Villa CA, Heidenreich E, et al:
52. Kleber AG, Rudy Y: Basic mechanisms of cardiac
impulse propagation and associated arrhythmias.
Physiol Rev 84:431–488, 2004.
35
Recommendations for patient selection, proce- Effect of the ectopic beats location on vulnerability 53. Clayton RH, Holden AV: Propagation of normal
dural techniques, patient management and to reentries in a three-dimensional realistic model beats and re-entry in a computational model of
follow-up, definitions, endpoints, and research trial of human atrium. In Programs and Abstracts of the ventricular cardiac tissue with regional differences
design. Europace 14:528–590, 2012. 36th Annual International Conference of Comput- in action potential shape and duration. Prog
17. Noble D: Modeling the heart—from genes to cells ers in Cardiology, 2009, Park City, Utah, pp Biophys Mol Biol 85:473–499, 2004.
to the whole organ. Science 295:1678–1682, 2002. 449–452. 54. Takahashi Y, Sanders P, Jais P, et al: Organization
18. Rudy Y: From genes and molecules to organs and 36. Ruiz-Villa CA: Estudio de la vulnerabilidad a reen- of frequency spectra of atrial fibrillation: Relevance
organisms: Heart. In Egelman EH, editor: Com- tradas a través de modelos matemáticos y simu- to radiofrequency catheter ablation. J Cardiovasc
prehensive Biophysics: Textbook of Simulation and lación de la aurícula humana [PhD thesis], Valencia, Electrophysiol 17:382–388, 2006.
Modeling, vol 9, Oxford, 2012, Academic Press, pp 2010, Universitat Politècnica de València. 55. Canavan TE, Schuessler RB, Cain ME, et al: Com-
268–327. 37. Tobón C: Modelización y evaluación de factores puterized global electrophysiological mapping of
19. Moe GK, Rheinboldt WC, Abildskov JA: A com- que favorecen las arritmias auriculares y su trata- the atrium in a patient with multiple supraventricu-
puter model of atrial fibrillation. Am Heart J miento mediante técnicas quirúrgicas [PhD thesis], lar tachyarrhythmias. Ann Thorac Surg 46:232–
67:200–220, 1964. Valencia, 2010, Universitat Politécnica de 235, 1988.
20. Nygren A, Fiset C, Firek L, et al: Mathematical València. 56. Lemery R, Guiraudon G, Veinot JP: Anatomic
model of an adult human atrial cell: The role of K+ 38. Tobón C, Rodríguez JF, Ferrero JM, et al: Domi- description of Bachmann’s bundle and its relation
currents in repolarization. Circ Res 82:63–81, nant frequency and organization index maps in a to the atrial septum. Am J Cardiol 91:1482–1485,
1998. realistic 3D computational model of atrial fibrilla- 2003.
21. Courtemanche M, Ramirez RJ, Nattel S: Ionic tion. Europace 14:v25–v32, 2012. 57. Kalman JM, Olgin JE, Karch MR, et al: “Cristal
mechanisms underlying human atrial action poten- 39. Tobón C, Ruiz-Villa CA, Heidenreich E, et al: A tachycardias”: Origin of right atrial tachycardias
tial properties: Insights from a mathematical three-dimensional human atrial model with fiber from the crista terminalis identified by intracardiac
model. Am J Physiol 275:H301–H321, 1998. orientation: Electrograms and arrhythmic activa- echocardiography. J Am Coll Cardiol 31:451–459,
22. Maleckar MM, Greenstein JL, Giles WR, et al: tion patterns relationship. PloS ONE accepted. 1998.
Electrotonic coupling between human atrial myo- 40. Krueger MW, Schmidt V, Tobón C, et al: Model- 58. Ryu K, Sahadevan J, Khrestian CM, et al: Use of
cytes and fibroblasts alters myocyte excitability and ing atrial fiber orientation in patient-specific fast Fourier transform analysis of atrial electro-
repolarization. Biophys J 97:2179–2190, 2009. geometries: A semi-automatic rule-based approach. grams for rapid characterization of atrial
23. Grandi E, Pandit SV, Voigt N, et al: Human atrial In Programs and Abstracts of the 6th International activation—Implications for delineating possible
action potential and Ca2+ model: Sinus rhythm Conference on Functional Imaging and Modeling mechanisms of atrial tachyarrhythmias. J Cardio-
and chronic atrial fibrillation. Circ Res 109:1055– of the Heart, 2011, New York, pp 223–232. vasc Electrophysiol 17:198–206, 2006.
1066, 2009. 41. McDowell KS, Vadakkumpadan F, Blake R, et al: 59. Saoudi N, Cosío F, Waldo A, et al: A classification
24. Pandit SV, Berenfeld O, Anumonwo JM, et al: Methodology for patient-specific modeling of of atrial flutter and regular atrial tachycardia
Ionic determinants of functional reentry in a 2-D atrial fibrosis as a substrate for atrial fibrillation. according to electrophysiological mechanisms and
model of human atrial cells during simulated J Electrocardiol 45:640–645, 2012. anatomical bases; a Statement from a Joint Expert
chronic atrial fibrillation. Biophys J 88:3806–3821, 42. Ho SY, Anderson RH, Sanchez-Quintana D: Atrial Group from The Working Group of Arrhythmias
2005. structure and fibres: Morphologic bases of atrial of the European Society of Cardiology and the
25. Ashihara T, Haraguchi R, Nakazawa K, et al: The conduction. Cardiovasc Res 54:325–336, 2002. North American Society of Pacing and Electro-
role of fibroblasts in complex fractionated electro- 43. Cohen GI, White M, Sochowski RA, et al: physiology. Eur Heart J 22:1161–1182, 2001.
grams during persistent/permanent atrial fibrilla- Reference values for normal adult transesophageal 60. Friedman PA, Luria D, Fenton AM, et al: Global
tion: Implications for electrogram-based catheter echocardiographic measurements. J Am Soc Echo- right atrial mapping of human atrial flutter: The
ablation. Circ Res 110:275–284, 2012. cardiogr 8:221–230, 1995. presence of posteromedial (sinus venosa region)
26. Harrild D, Henriquez C: A computer model of 44. Wilders R, Wagner MB, Golod DA, et al: Effects functional block and double potentials: A study in
normal conduction in the human atria. Circ Res of anisotropy on the development of cardiac biplane fluoroscopy and intracardiac echocardiog-
87:E25–E36, 2000. arrhythmias associated with focal activity. Pflugers raphy. Circulation 101:1568–1577, 2000.
27. Vigmond EJ, Ruckdeschel R, Trayanova N: Arch 441:301–312, 2000. 61. Olshansky B, Wilber DJ, Hariman RJ:
Reentry in a morphologically realistic atrial model. 45. Nathan H, Eliakim M: The junction between the Atrial flutter—Update on the mechanism and
J Cardiovasc Electrophysiol 12:1046–1054, 2001. left atrium and the pulmonary veins: An anatomic treatment. Pacing Clin Electrophysiol 15:2308–
28. Ruchat P, Dang L, Virag N, et al: A biophysical study of human hearts. Circulation 34:412–422, 2335, 1992.
model of atrial fibrillation to define the appropriate 1966. 62. Cox JL: The surgical treatment of atrial fibrilla-
ablation pattern in modified maze. Eur J Cardio- 46. Ho S, Sanchez-Quintana D: The importance of tion. IV. Surgical technique. J Thorac Cardiovasc
thorac Surg 31:65–69, 2001. atrial structure and fibers. Clin Anat 22:52–63, Surg 101:584–592, 1991.
29. Dang L, Virag N, Ihara Z, et al: Evaluation of 2009. 63. Cox JL, Schuessler RB, Lappas DG, et al: An
ablation patterns using a biophysical model of atrial 47. Feng J, Yue L, Wang Z, et al: Ionic mechanisms of 8½-year clinical experience with surgery for atrial
fibrillation. Ann Biomed Eng 33:465–474, 2005. regional action potential heterogeneity in the fibrillation. Ann Surg 224:267–273, 1996.
30. Haïssaguerre M, Lim KT, Jacquemet V, et al: Atrial canine right atrium. Circ Res 83:541–551, 1998. 64. Khargi K, Lemke B, Deneke T: Concomitant anti-
fibrillatory cycle length: Computer simulation and 48. Bosch RF, Zeng X, Grammer JB, et al: Ionic mech- arrhythmic procedures to treat permanent atrial
potential clinical importance. Europace 9:v64–v70, anisms of electrical remodeling in human atrial fibrillation in CABG and AVR patients are as effec-
2007. fibrillation. Cardiovasc Res 44:121–131, 1999. tive as in mitral valve patients. Eur J Cardiothorac
31. Seemann G, Hoper C, Sachse FB, et al: Heteroge- 49. Zhang H, Liu J-H, Garratt CJ, et al: Cellular mod- Surg 27:841–846, 2005.
neous three-dimensional anatomical and electro- elling of electrical remodelling in two different 65. Sie HT, Beukema WP, Elvan A, et al: New strate-
physiological model of human atria. Philos models of human atrial myocytes. Presented at the gies in the surgical treatment of atrial fibrillation.
Transact A Math Phys Eng Sci 364:1465–1481, 30th Annual International Conference of Comput- Cardiovasc Res 58:501–509, 2003.
2006. ers in Cardiology, 2003, Thessaloniki Chalkidiki, 66. Sueda T, Nagata H, Orihashi K, et al: Efficacy of
32. Aslanidi OV, Colman MA, Stott J, et al: 3D virtual pp 777–780. a simple left atrial procedure for chronic atrial
human atria: A computational platform for study- 50. Boineau JP, Canavan TE, Schuessler RB, et al: fibrillation in mitral valve operations. Ann Thorac
ing clinical atrial fibrillation. Prog Biophys Mol Demonstration of a widely distributed atrial pace- Surg 63:1070–1075, 1997.
Biol 107:156–168, 2011. maker complex in the human heart. Circulation 67. Kobza R, Kottkamp H, Dorszewski A, et al: Stable
33. Aslanidi OV, Al-Owais M, Benson AP, et al: Virtual 77:1221–1237, 1988. secondary arrhythmias late after intraoperative
tissue engineering of the human atrium: Modeling 51. Hansson A, Holm M, Blomstrom P, et al: Right radiofrequency ablation of atrial fibrillation: Inci-
pharmacological actions on atrial arrhythmogene- atrial free wall conduction velocity and degree of dence, mechanism, and treatment. J Cardiovasc
sis. Eur J Pharm Sci 46:209–221, 2012. anisotropy in patients with stable sinus rhythm Electrophysiol 15:1246–1249, 2004.
Cardiac Electromechanical Models
Natalia A. Trayanova, Jason Constantino, and Yuxuan Hu
36
CHAPTER OUTLINE General Approach to Multi-scale

General Approach to Multi-scale Electromechanical Electromechanical Modeling
Modeling of the Heart and Its Validation 361 of the Heart and Its Validation
Applications of Electromechanical Modeling 363
Initial Efforts Toward Patient-Specific
Computational Representation of the
Electromechanical Modeling of the Heart 367 Electromechanical Processes in the Heart
Concluding Remarks 368 A schematic of the general approach to modeling multi-scale
cardiac electromechanical function at the level of the organ is
shown in Figure 36-1, A. The general approach consists of two
The cyclic pumping of the heart arises from the synergy of its coupled problems, simulating the electrical and mechanical func-
electrical and mechanical functions. Understanding the individ- tions of the heart. A flowchart of the simulation process with
ual functions has been the subject of intense research in basic connections within the model representation of each function as
science and clinical cardiology. Over the years, experimental and well as inter-relations between the two parts of the electrome-
clinical studies have provided significant insight into the electrical chanical model is presented in Figure 36-1, B.
and mechanical activity of the beating heart from the molecular The electrical problem of the model simulates the propaga-
to the organ level. However, detailed information regarding the tion of a wave of transmembrane potential by solving the
intricate electrical or mechanical processes at each level of this monodomain reaction-diffusion partial differential equation
hierarchy might not be sufficient to elucidate the causes of emer- (PDE; or a system of coupled PDEs if the extracellular current
gent phenomena at the level of the entire organ arising from the flow is explicitly accounted for, i.e., the bidomain problem) over
interactions between electrical and mechanical processes. With the volume of the heart.1 The reaction-diffusion PDE describes
current experimental methods limited in their inability to explore current flow through myocytes that are electrically connected via
the three-dimensional coupled electrical and mechanical activity low-resistance gap junctions. Cardiac tissue has orthotopic
in the heart simultaneously and with sufficient spatiotemporal passive electrical conductivities that arise from the cellular orga-
resolution, computer modeling of whole-heart electromechanical nization of the heart into fibers and laminar sheets. Global con-
function is rapidly becoming an important investigative tool in ductivity values are obtained by combining fiber and sheet
its own right. Today, owing to advancements in computational organization with myocyte-specific local conductivity values.
techniques and tools as well as in image processing, electrome- Current flow in the tissue is driven by the active processes of ionic
chanical modeling of the heart has become a comprehensive exchanges across myocyte membranes. These processes are rep-
methodology that combines detailed information regarding the resented by the cellular ionic model (see Figure 36-1, A, B),
electrophysiological and mechanical processes across the spatial where current flow through ion channels, pumps, and exchangers
scales in the heart, and serves to provide a higher level of under- and subcellular calcium cycling are governed by a set of ordinary
standing of the complex electromechanical interactions in the differential (ODE) and algebraic equations; ionic models of dif-
heart. ferent complexity are currently in use.2 Simultaneous solution of
In this chapter, we present an overview of the current state- the PDE(s) with the set of ionic model equations represents
of-the-art in whole-heart electromechanical modeling, focusing simulation of electrical wave propagation in the heart.
on realistic-geometry biophysically detailed model develop- The intracellular calcium released during electrical activation
ments. We first present the general framework in modeling couples the electrical and mechanical components of the model
the electromechanical behavior of the heart. We then showcase by providing a bi-directional link between the cellular ionic and
the powerful utility of such realistic electromechanical models myofilament models (see Figure 36-1, A, B). The cellular myo-
in revealing mechanisms at play in the normal and diseased filament model consists of another set of ODEs that represent
heart by reviewing the latest insights obtained with such models. the biophysical processes of calcium binding to troponin and
We conclude this chapter with a discussion of the developments cross-bridge cycling, as well as the mechanisms of cooperativity.
in patient-specific electromechanical modeling, emphasizing Compared with the evolution of cellular ionic models, the devel-
translational efforts toward bringing computer modeling of heart opment of myofilament models has been slower and more diffi-
electromechanics from the realm of the basic science into the cult, as no clear consensus has been reached regarding the
clinic. mathematical approach to model myofilament dynamics. An
361
Circulatory system
SAC
C
Monodomain/Bidomain Ca Continuum mechanics
Cellular ionic model Cellular myofilament model

A Electrical problem Mechanical problem
Compute Calculate D
Mechanical problem
Electrical problem
propagation of state variables of Compute volumes

electrical activity active tension model of the pulmonic and
circulatory system
Iion
Ca++ Ta
VLV
Vm VRV
PLV
Compute Calculate PRV
membrane Stretch, balance of forces
kinetics stretch rate
B E
Figure 36-1. A, Schematic of the general approach to modeling cardiac electromechanical function. B, Flowchart of the simulation process. C, Geometric models of the
heart (rabbit and canine). D, Computational meshes of the canine heart for electrical and mechanical problems. E, Fiber and sheet orientations obtained from diffusion
tensor (DT)-magnetic resonance imaging (MRI) of the canine heart.
(Images modified with permission from Vadakkumpadan F, Arevalo H, Prassl AJ, et al: Image-based models of cardiac structure in health and disease. Wiley Interdiscip Rev Syst
Biol Med 2:489–506, 2010; Gurev V, Lee T, Constantino J, et al: Models of cardiac electromechanics based on individual hearts imaging data: Image-based electromechanical
models of the heart. Biomech Model Mechanobiol 10:295–306, 2011.)
up-to-date review of cellular myofilament models can be found as cylindrical and elliptical shapes to represent the ventricles) or
in a recent publication.3 anatomically accurate, the latter representing ventricular aver-
The contraction of the heart arises from the active tension aged geometries obtained from histologic sectioning6-8 or the
generated by the myofilaments within the cardiac cell. In the geometry and structure of individual hearts,9-11 as obtained from
mechanics part of the model, deformation of the organ is magnetic resonance imaging (MRI).12 Figure 36-1, C presents
described by the equations of continuum mechanics,4 with the some of the ventricular geometries used in electromechanical
passive properties of the myocardium described by a constitutive modeling: The University of California San Diego (UCSD)
law. The most comprehensive formulation of cardiac tissue con- rabbit ventricular geometry6 is an example of averaged geometry
stitutive relation can be found in a recent article by Holzapfel obtained through histologic sectioning, while another image
and Ogden.5 Simultaneous solution of the myofilament model exemplifies an MRI-based individual heart geometry.9 An example
equations and of those representing passive cardiac mechanics of an MRI-based atrial geometry only can be found in a recent
over the volume of the heart (see Figure 36-1, A, B) constitutes publication.13
simulation of cardiac contraction. Finally, to simulate the cardiac The aforesaid description of electromechanical modeling
cycle and the corresponding pressure-volume loops, conditions refers to models of strong coupling, where the electrical and
on chamber volume and pressure are imposed, arising typically mechanical problems are solved simultaneously.14 In cases that do
from lumped-parameter models of the systemic and pulmonic not necessitate strong coupling, as for instance during examina-
circulatory systems (see Figure 36-1, A, B). tion of how the mechanical activation of the heart follows the
In addition to the bi-directional relationship between electri- electrical activation, weakly coupled schemes are employed. In a
cal and mechanical components, provided by intracellular calcium weakly coupled model, the electrical activation times (calculated
cycling, a key feedback mechanism in the electromechanical from the electrical problem) are inputted into the mechanical
model (acting within the mechanics component) is the length and part as the instances when the myofilament model is activated.
velocity dependence of tension (see Figure 36-1, B): The stretch In a more sophisticated approach, the ionic and cardiac myofila-
and stretch rate, as determined by the deformation of the heart, ment models can be coupled within the mechanics component,
affect tension development in the cell. Mechanical deformation with the electrical activation times determining the instant at
could further affect the electrical activity of the heart via the which this combined ionic-myofilament model is activated10; in
opening of stretch-activated channels (see Figure 36-1, A, B). To this case, cooperativity mechanisms, such as calcium binding to
simulate this feedback mechanism, the stretch and stretch rate troponin C, are represented in the model.
calculated from the mechanics component serve as an input into The governing equations describing cardiac electromechani-
the electrical component: They determine the conductance of cal behavior are solved on a spatially discretized version of
stretch-activated channels, the latter represented within the cel- the heart volume (i.e., on the computational mesh). The
lular ionic model. electrical and mechanical parts of the model have different
Solutions to organ-level electromechanical problems entail requirements regarding the degree of discretization (i.e., element
the use of organ-level geometries, which could be idealized (such size) and the element type; thus the two parts of the model
Cardiac Electromechanical Models 363
require two different computational meshes. The electrical mesh maps. The mechanical component is validated with local strain
36
requirements are based on spatiotemporal characteristics of wave measurements calculated from MRI or ultrasound images, or
propagation; a spatial resolution of about 250 to 300 micrometers with hemodynamical metrics such as left ventricular (LV) pres-
is appropriate for electrophysiological finite element models. A sure and volume, ejection fraction, or maximal rate of pressure
novel approach was recently published for electrical mesh gen- change.
eration directly from segmented MRI15 (Figure 36-1, D). The A study by Provost et al21 provides a unique example of elec-
mechanical mesh, on the other hand, typically consists of hexa- tromechanical (rather than separate electrical or mechanical)
hedral elements with a Hermite basis. This choice of finite ele- heart model validation. It is based on the use of electromechanical
ments increases the degree of strain continuity and is appropriate wave imaging (EWI),22 a novel noninvasive ultrasound-based
for maintaining incompressibility constraints. The mechanical imaging technique capable of mapping the propagation of the
mesh of the heart (see Figure 36-1, D) can also be generated electromechanical wave along echocardiographic planes; this is
directly from segmented MRI.10,16 achieved by mapping the interframe axial strains. In the Provost
Fiber and laminar sheet organization underlies the ortho- et al study,21 data from EWI were used to validate the MRI-based
tropic electrical conductivities of the tissue and its mechanical electromechanical model of the normal canine ventricles for dif-
properties. In the electrical mesh, local fiber and sheet directions ferent pacing protocols. Figure 36-2, A, B presents experimental
are typically mapped at the centroids of the finite elements, and and simulated EWI maps for pacing from the LV base; the cor-
in the mechanics mesh, fiber and sheet orientations and their responding isochronal maps of electromechanical activation are
derivatives are defined at mesh nodes and then are interpolated shown in Figure 36-2, C. In both experiments and simulations,
over the elements. This is typically done using histologic section- the EW emerged from the basal region of the lateral wall and
ing information6 or diffusion tensor (DT) MRI data.9,10 Use of propagated toward the apex, the septum, and the right ventricular
DT-MRI data is based on the fact that the primary, secondary, (RV) wall. Representative curves of the interframe strains over
and tertiary eigenvectors of the water DTs are aligned with fiber time in the lateral and septal wall are shown in Figure 36-2, D.
direction, with the direction transverse to the fiber direction and This study provides an illustrative example of how emerging
in the plane of the laminar sheet, and with that normal to the experimental techniques like EWI could be used for validation
laminar sheet, respectively.12 Figure 36-1, E presents fiber and of cardiac electromechanical models.
sheet orientation, as reconstructed from DT-MRI of the canine
heart. In cases where neither histologic nor DT-MRI informa-
tion is available, rule-based approaches17 or image transformation
algorithms18 have been used to assign fiber and sheet orientation Applications of Electromechanical Modeling
consistent with measurements.
Simulations of heart electromechanical function are typically Electromechanical Interactions in the Heart
executed on parallel high-performance computing hardware.
Reviews of numeric approaches to simulating the electrome- Understanding the mechanical consequences of an altered cardiac
chanical activity of the heart can be found1,10,19,20; these articles activation sequence is of great importance because dyssynchro-
also address the challenges involved in developing multi-scale nous electrical activation can cause abnormalities in perfusion and
models at the organ level. pump function. Early computational studies of electromechanics
employed truncated ellipsoids as LV representations in an attempt
to provide insight into the relationship between the spatial pattern
Approaches to Experimental Validation of of electrical activation and the resultant contraction.23 The study
Electromechanical Models by Usyk and McCulloch24 was the first to examine the distribution
of the time interval between myocyte depolarization and onset of
Validating the ventricular (or atrial) model of cardiac electrome- myofiber shortening, termed the electromechanical delay, in a
chanics with experimental data is a pivotal component of model realistic-geometry ventricular canine model. The study by Gurev
development. It constrains the model parameter space and et al25 further advanced understanding of the three-dimensional
enhances the physiological relevance of the model. To validate (3D) electromechanical delay distribution in the intact ventricles
the electrical component of the model, simulation results are for sinus rhythm and epicardial pacing. The authors employed an
often compared with electrical activation maps obtained from electromechanical model of the rabbit ventricles and dissected the
epicardial sock or plunge electrodes, from optical mapping, or role of loading conditions in altering the 3D distribution of elec-
from electrocardiograms (ECGs) and body surface potential tromechanical delay. Figure 36-3 presents the epicardial and
50 ms 60 ms 70 ms 90 ms 110 ms
0.25% 30 ms
0.3
Inter-frame strain (%)
0.2
0.1
0
A –0.1
–0.2
–0.3
–0.4
–0.25% 110 ms 0 50 100 150 200
B C D Time (ms)
Figure 36-2. Validation of the electromechanical model of the canine heart with electromechanical wave imaging. Experimental (top) and simulated (bottom) interframe
strain distribution associated with an electromechanical wave (A) and the corresponding isochronal maps of electromechanical activation (B) for LV base pacing. C, Experi-
mental and simulated interframe strain traces at the septum (black) and the lateral wall (blue).
(Images modified with permission from Provost J, Gurev V, Trayanova N, et al: Mapping of cardiac electrical activation with electromechanical wave imaging: An in silico-in vivo
reciprocity study. Heart Rhythm 8:752–759, 2011.)
Sinus rhythm Epicardial pacing activated by mechanical stimuli. Of these, stretch-activated chan-
nels (SACs) have long been implicated as important contributors
to the proarrhythmic substrate in the heart. The nonuniform
distribution of positive myofiber strain (stretching) during
mechanical contraction under a variety of pathologic conditions
could produce, via SACs, proarrhythmic dispersion in electro-
physiological properties. SACs have been shown to shorten or
lengthen the action potential duration (APD) of a single myocyte
or to produce ectopic beats, depending on the timing of the
Electromechanical delay mechanical stimulus application relative to the phase of the action
25 ms 40 ms potential. However, uncovering the mechanisms by which SACs
Figure 36-3. Electromechanical delay during sinus rhythm and epicardial pacing contribute to ventricular arrhythmogenesis under a variety of
in a model of rabbit ventricular electromechanics. Each panel presents the left pathologic conditions has been hampered by the lack of experi-
ventricular (LV) lateral view of epicardium (left) and endocardium (right). The lines mental methods that can record the 3D electrical and mechanical
represent fiber directions. activity simultaneously and with high spatiotemporal resolution.
Thus, computer simulations have emerged as a valuable tool
(Images modified with permission from Gurev V, Lee T, Constantino J, et al: Models of
in dissecting the mechanisms by which SACs contribute to
cardiac electromechanics based on individual hearts imaging data: Image-based elec-
arrhythmogenesis.
tromechanical models of the heart. Biomech Model Mechanobiol 10:295–306, 2011.)
Role of Mechanoelectrical Feedback in the Stability

endocardial electromechanical delay distributions for sinus of Ventricular Tachycardia/Fibrillation
rhythm and epicardial pacing obtained in the study. Results A study by Keldermann et al28 employed an electromechanical
revealed that during normal sinus rhythm, the electromechanical model of human LV to study the effect of mechanoelectrical feed-
delay was longer on the epicardium than on the endocardium and back via SAC on reentrant wave dynamics. SAC conductance was
at the base than at the apex. After epicardial pacing, electrome- dependent on the stretch ratio in the fiber direction. The authors
chanical delay distribution was markedly different. For both elec- found that nonuniform activation of SACs can result in the degen-
trical activation sequences, the late-depolarized regions were eration of a stable scroll wave (representing ventricular tachycar-
characterized by significant myofiber prestretch caused by con- dia [VT]) into turbulent patterns characteristic of ventricular
traction of the early-depolarized regions. This prestretch delayed fibrillation (VF). Simulation results revealed that regions of wave-
the onset of myofiber shortening, thus resulting in a longer elec- break were those undergoing stretch. At these regions, depolar-
tromechanical delay, giving rise to heterogeneities in 3D electro- ization due to SAC opening blocked propagation at the stretched
mechanical delay distribution. This study underscored the central region. This simulation study suggested an explanation of the
role that the electrical activation sequence and thus the loading degeneration of ventricular tachycardia into ventricular fibrilla-
conditions play in modulating the relationship between electrical tion that offered an alternative to the restitution hypothesis.
activation and mechanical contraction. However, Keldermann et al analyzed only how opening of SACs
The heart achieves an efficient coordinated contraction via a with a reversal potential close to zero and large conductance
complex network of feedback mechanisms. Organ-level electro- affected the stability of VT/VF. Because experimental studies
mechanical modeling by Niederer and Smith26 explored how have found that SACs exhibit a variety of reversal potentials and
cellular-level behaviors affect work transduction, stress and strain conductances,29 a more comprehensive study of the effects of
homogeneity at the whole-ventricle level, and the feedback loops SACs on the stability of VT/VF was needed.
that regulate normal contraction in an electromechanical model Hu et al30 aimed to analyze spiral wave stability using an MRI-
of the rat LV. The simulation research demonstrated that length- based electromechanical model of the human heart that included
dependent changes in Ca sensitivity and the filament overlap, a variety of SAC reversal potentials and channel conductances.
which is believed to constitute the Frank-Starling law, were the Investigators found that recruitment of SACs can also have the
two dominant regulators of the efficient transduction of work. opposite effect: Opening of SACs with large negative reversal
The absence of either mechanism not only altered the spatial potentials or of those with reversal potentials close to zero and
distribution of stress and strain, but also determined the trans- low conductance led to suppression of scroll wave breakup. As
mural variation in work. These results showed that feedback from shown in Figure 36-4, A, in a model without SAC representation,
muscle length to tension generation at the cellular level is an scroll waves break up continuously, sustaining VF. Representing
important control mechanism of the pumping efficiency of the SAC with negative reversal potentials (e.g., −60 mV) decreased
heart. In a recent study, Land et al27 developed a murine model the average number of scroll wave filaments by 46% to 62%.
of electromechanics to explore how length-dependent and Mechanistic analysis revealed that recruitment of SACs in this
velocity-dependent feedback mechanisms alter the pressure case inhibited scroll wave breakup through flattening of the APD
developed by the ventricles. Simulation results revealed that the restitution relationship, although to a different degree in differ-
length-dependent changes in Ca sensitivity and filament overlap ent regions. Opening of SACs with less negative reversal poten-
are the principal regulators of ejection and isovolumetric relax- tials (e.g., −10 mV) and with low channel conductances also
ation. The model also revealed that including velocity depen- suppressed scroll wave breakup (Figure 36-4, B) but by a different
dence of tension in the model extended the LV pressure plateau, mechanism, rendering the conduction velocity restitution curve
resulting in a better match between experiment and simulation. shallower. This study revealed that recruitment of SACs affects
These two studies illustrate the importance of the cellular feed- scroll wave stability via different mechanisms, depending on the
back mechanisms in regulating the emergent electromechanical SAC population characteristics.
activity of the whole heart.
Role of Mechanoelectrical Feedback in Initiation
and Perpetuation of Atrial Fibrillation
Mechanoelectrical Coupling Because atrial dilatation increases vulnerability to atrial fibrilla-
tion (AF), understanding the mechanism for initiation and per-
One of the most important mechanisms of mechanoelectrical petuation of AF during acute stretch is of paramount importance.
coupling in the heart is seen in the sarcolemmal channels that are The study by Kuijpers et al31 addressed this issue by developing
LV LV
30 mV 30 mV 36
Vm
Vm
RV RV
–80 mV –80 mV
LV LV
RV RV
A B
Figure 36-4. Recruitment of stretch-activated channel (SAC) and scroll wave stability. Epicardial transmembrane potential distribution maps on the posterior wall (top) and
corresponding semitransparent view of the ventricles (bottom) at two different time points (1 s apart) from a simulation of ventricular fibrillation without SAC representation
(A) and a simulation of spiral wavebreak suppression resulting from SAC opening (B). Pink dots in transmembrane potential maps indicate the locations of epicardial phase
singularities. Blue denotes filament distribution in the semitransparent view of the ventricles; activation wavefronts are shown in red.
an image-based model of atrial electromechanics. In this model, 240

the mechanical problem was represented by interconnected seg- 190 (ms) –85 +35 (mV)
ments, each consisting of elastic and contractile components,
rather than continuum mechanics equations. The model incor- 191 193 195
porated mechanoelectrical feedback via SAC opening. Simula- LV
RV
tion results revealed that regional differences in the stretch ratio 360
resulted in heterogeneities in membrane excitability, impulse
propagation, APD, and the effective refractory period (ERP).
Dispersions of APD and ERP were further enhanced by contrac-
tion of active parts of the atria and passive stretch of the rest. The
study by Kuijpers et al provided evidence that mechanoelectrical Figure 36-5. Evolution of a mechanically induced ventricular premature beat (VPB)
feedback during atrial dilatation resulted in regional changes in and reentry in the regionally ischemic heart. Insets of 191 to 195 ms present short-
electrophysiological properties of the atria, contributing to the and long-axis views of the apical region for better visualization of the spontaneous
inducibility and perpetuation of AF. generation of a propagating wave. 240- to 360-ms panels present a titled anterior
view of the ventricles to allow visualization of reentry formation. Arrow in 191-ms
Spontaneous Induction of Arrhythmias in the Regionally inset indicates the location of earliest spontaneous firing.
Ischemic Heart (Images modified with permission from Jie X, Gurev V, Trayanova N: Mechanisms of
In the normal heart, tissue stretch can result in spontaneous firing mechanically induced spontaneous arrhythmias in acute regional ischemia. Circ Res
of myocytes and ventricular premature beats (VPBs). In cases of 106:185–192, 2010.)
significant mechanical stimuli, such as impact to the precordial
region of the chest, as often occurs in normal young athletes,
tissue stretch can lead to arrhythmia induction, a phenomenon
known as commotio cordis. explored using electromechanical modeling. Wall et al33 devel-
In the acutely ischemic heart, occurrence of VPBs has also oped an electromechanical model of the infarcted ovine LV and
been associated with rapid regional distention. The mechanisms found that APD dispersion in the border zone, which may provide
by which ischemia-induced mechanical dysfunction can result in the substrate for arrhythmias in the setting of myocardial infarc-
VPBs and induce reentry were examined by Jie et al.32 A 3D tion, was created by the combined effects of stretch and reduced
electromechanical model of the beating rabbit ventricles with electrical connectivity.
regional ischemia (4 minutes post occlusion) was developed and
contained a central ischemic zone (CIZ), a border zone (BZ), and
a normal zone (NZ). In both BZ and CIZ, cells underwent sig- Cardiac Resynchronization Therapy
nificant stretch during contraction, which led to depolarizations
due to the opening of SACs there, while such depolarizations Mechanisms Regulating Pump Dyssynchrony and the
were absent in NZ. The depolarizations resulted in mechanically Response to Cardiac Resynchronization Therapy
induced VPBs originating from the ischemic border (particularly Heart failure patients often exhibit contractile dyssynchrony due
in the LV endocardium, where fiber strain and strain rate were to electrical intraventricular delay, which diminishes the heart
the largest), but not from CIZ, although the magnitude of the systolic function. Cardiac resynchronization therapy (CRT) is a
depolarizations was larger there, because in the latter, ischemic clinical treatment that re-coordinates contraction by applying
injury suppressed excitability. VPBs then traveled intramurally appropriately timed pacing stimuli to the ventricles. Although
until emerging from the ischemic border on the epicardium, CRT reduces morbidity and mortality, 30% of patients fail to
initiating reentry (Figure 36-5). The study by Jie et al32 provided respond to the therapy. Current dyssynchrony indices used to
the first evidence that mechanically induced membrane depolar- identify potential responders to CRT have poor predictive capa-
izations and their spatial distribution within the ischemic region bility, reflecting an incomplete understanding of the electrome-
are possible mechanisms by which mechanical activity contrib- chanical behavior involved in dyssychronous heart failure.
utes to the origin of spontaneous arrhythmias. Ventricular electromechanical simulations offer an opportunity
The role of mechanoelectrical feedback via SACs in altering to elucidate the mechanisms that underlie heart failure dyssyn-
the electrical activity in myocardial infarction has also been chrony and to provide novel therapeutic strategies.
Kerckhoffs et al34 employed a computational model of failing

canine ventricles to assess the sensitivity of current clinical indices
quantifying mechanical dyssynchrony to various abnormalities
responsible for contractile dysfunction. These indices include the
echocardiography-based metric, the time to peak shortening
(WTpeak), and the MRI-based metrics of circumferential unifor-
mity ratio estimate (CURE) and internal stretch fraction (ISF),
which measure regional strain magnitudes. The abnormalities
examined were dilatation, dyssychronous activation, decreased
inotropy, and prolonged relaxation. A sensitivity analysis of these
abnormalities revealed that all indices were sensitive to dyssy-
chronous activation; however, the synergistic effect of dyssychro-
nous activation and dilatation resulted in greatest changes in ISF
and CURE. These simulation results also demonstrated that ISF 20 50
and CURE were better indices of mechanical dyssynchrony A Electromechanical delay (ms)
because of their sensitivity to regional strain inhomogeneity. The
study concluded that if geometry and electrical activation
sequence, which can be established by a clinical evaluation, are
the major determinants of regional cardiac function, and if the
hard-to-measure material properties play a minor role, patient-
specific simulations are feasible and could be performed to
suggest optimal CRT strategies tailored to the individual.
Using an electromechanical model of the human ventricles,
Niederer et al35 performed a parameter sensitivity analysis to
determine the mechanisms that regulate CRT response. The
simulation research revealed that the degree of length depen-
dence on tension and the minimum length of tension generation
(the length at which no active tension is generated) were the
significant parameters that determined CRT efficacy. At baseline
(pre-CRT), attenuating length-dependent tension regulation
augmented dyssynchrony in tension development, fiber shorten-
ing, and dispersion time to the peak in tension development rate
(tpeak). Because synchronization of tpeak following CRT was great-
est when length dependence was reduced, the change in the
maximum rate of change in LV pressure (CRT response) was the 0 20
largest. Because the minimum length of tension generation is B Increase in dP/dtmax (%)
unaltered in heart failure, these results suggest that reduced
length dependence of tension is a mechanism underpinning CRT
30
efficacy.
Increase in dP/dtmax (%)
Optimizing the Response to CRT R = –0.86

Suboptimal placement of the LV lead constitutes a major reason 20
underlying the high nonresponse rate to CRT. To date, no clear
consensus has been reached as to where the LV pacing lead
should be placed to achieve optimal CRT response. Studies have
indicated that the site of latest electrical or latest mechanical 10
activation is associated with greater hemodynamic benefit in
CRT; however, recent human data36 suggest a lack of concor-
dance between the site of latest electrical or mechanical activation
and the CRT response. 0
Constantino et al37 proposed a new strategy to determine the 0 10 20 30 40 50
LV pacing location that optimizes the response to CRT: targeting
the regions with the longest electromechanical delay. An MRI- Longitudinal distance between LV pacing
based electromechanical model of canine dyssychronous heart C site and region with longest EMD (mm)
failure (DHF), which incorporated DHF-associated remodeling Figure 36-6. Optimization of cardiac resynchronization therapy (CRT) response
aspects such as altered ventricular structure, slowed conduction, based on electromechanical delay distribution. A, Transmural short-axis electrome-
deranged calcium handling, and reduced stiffness, was employed. chanical delay maps during left bundle branch block generated with a
CRT was delivered by pacing at the RV apex, with the LV pacing magnetic resonance imaging (MRI)-based electromechanical model of
dyssynchronous heart failure. In each map, the anterior and posterior walls are at
electrode placed at 18 different epicardial sites along the LV free
the top and bottom, respectively. B, Map of the percentage increase in
wall. With the use of transmural electromechanical delay maps maximal rate of change in pressure as a function of the left ventricular (LV) pacing
(Figure 36-6, A), the region with the longest electromechanical site. Red dots denote LV pacing sites. C, Correlation of longitudinal distance
delay during left bundle branch block was determined to be the between LV pacing site and region with longest electromechanical delay, and
endocardial surface of the lateral wall between the base and the percentage increase in dP/dtmax.
midventricles. Figure 36-6, B presents CRT response as a func-
tion of LV pacing location. Maximal hemodynamic benefit (Images modified with permission from Constantino J, Hu Y, Trayanova N: A compu-
occurred when the LV pacing site was located near the base, tational approach to understanding the electromechanical activation sequence in the
which was within the region of longest electromechanical delay. normal and failing heart, with translation to the clinical practice of CRT. Prog Biophys
The relationship between LV pacing location relative to the Mol Biol. 110:372–379, 2012)
longest electromechanical delay region and the corresponding

45°
36
CRT response is quantified in Figure 36-6, C. Improvement in
maximal rate of change in LV pressure strongly correlated with
the longitudinal distance between the LV pacing site and the
Helix angle
center of the region with longest electromechanical delay. This LV
study proposed a new approach to determining the optimal CRT RV
pacing location that accounts for the relation between electrical
and mechanical activation sequences in the DHF heart.
Alternatively, computational models can be used to test new
pacing protocols in efforts to optimize CRT. Multi-site CRT, A –45°
which applies pacing stimuli via one RV lead and two LV leads,
has been shown to hold great promise as an alternative strategy
in improving CRT response. In addition, quadripolar leads, in
which two pacing stimuli can be delivered from one lead, have
recently been developed. Niederer et al38 aimed to determine
whether multi-site CRT with a quadripolar lead resulted in
improved CRT response in a model of human electromechanics.
The study found that multi-site CRT conferred a greater
improvement in dP/dtmax, as compared with conventional CRT, B
only in hearts with an infarct. Because the presence of scar often
results in poor CRT response, multi-site CRT with quadripolar
leads may thus improve the response to CRT in ischemic cardio-
myopathy patients. These studies provide convincing evidence
that cardiac electromechanical models can be used for model-
guided optimization of CRT lead position and pacing protocols
in the near future.
Abnormal Mechanical Deformation and Its

Role in Vulnerability to Electrical Shocks C 10 ms 180 ms 210 ms
and Defibrillation Figure 36-7. A, Short-axis view of ventricular geometry and fiber helix angle for
undeformed (left) and dilated (right) rabbit ventricles. Transmembrane potential
Clinical studies have demonstrated that patients with dilated, distributions in the undeformed (B) and dilated (C) ventricles following a mono-
volume- or pressure-overloaded hearts have elevated defibrilla- phasic truncated exponential shock of 10-ms duration applied at 120 ms after the
tion thresholds (DFTs). The mechanisms underlying these find- last pacing stimuli. This electrode configuration results in a uniform electrical field;
ings are not well understood. Using a rabbit ventricular thus virtual electrode polarization (VEP) formation depends only on ventricular
electromechanics model, Trayanova et al39 conducted simulations geometry and fiber orientation. Shock-end transmembrane potential distribution
of vulnerability to strong shocks and defibrillation under condi- is at 10 ms. Differences in VEP are marked by black ovals. The postshock reentry is
tions of LV dilatation, and determined the mechanisms by which illustrated in the 180- and 210-ms panels.
mechanical deformation may lead to increased vulnerability and (Images modified with permission from Trayanova N, Gurev V, Constantino J, et al:
elevated DFT. Studies of defibrillation mechanisms have demon- Mathematical models of ventricular mechano-electric coupling and arrhythmia. In
strated that following shock delivery, ventricular geometry and Kohl P, Sachs F, Franz MR (eds): Cardiac Mechano-electric Feedback and Arrhythmias.
fiber orientation determine the large-scale distribution and mag- Oxford, 2011, Oxford University Press, pp 256–258.)
nitude of the virtual electrode polarization (VEP) induced by the
strong shock.40 Thus, ventricular dilatation could affect VEP
through changes in ventricular geometry and fiber architecture,
leading to changes in the upper limit of vulnerability and DFT. Initial Efforts Toward Patient-Specific
The model revealed that LV dilatation indeed results in geomet- Electromechanical Modeling of the Heart
ric alterations, with the most prominent change noted in septal
fiber orientation (Figure 36-7, A). Transmembrane voltage maps With advancements in computational techniques and medical
following shock application, shown in Figure 36-7, B,C for the imaging and image processing, the groundwork for patient-
undeformed and dilated ventricles, respectively, illustrate the dif- specific modeling of electromechanics has been laid, and the era
ferent postshock responses in the two cases. At the end of the of personalized computational cardiac medicine is fast approach-
shock (see Figure 36-7, B, C; 10 ms), VEP in the bulk of the ing. The first step in conducting patient-specific cardiac electro-
septal myocardium was different: A larger excitable gap was seen. mechanical simulations is to construct the computational model
Consequently, although no reentrant activity was observed in the of the heart from the patient’s medical images. In addition to the
undeformed ventricles (see Figure 36-7, B), a postshock sustained challenges involved in reconstructing heart geometry from in
reentry was formed on the RV side of the septum in the deformed vivo clinical scans of low resolution for any other applications, a
ventricles (see Figure 36-7, C), indicating increased vulnerability specific hurdle in constructing the geometric mesh of a patient’s
to electrical shocks. Similarly, DFT in the dilated ventricles was heart for electromechanical applications is defining the unloaded
found to be 38%±20% higher than that in the normal ventricles state of the heart, because the heart is constantly loaded during
(for an implantable cardioverter-defibrillator [ICD] electrode image acquisition. Aguado-Sierra et al41 used an iterative
configuration), which showed good agreement with experimental estimation scheme to approximate the unloaded geometry from
findings. Results of the study suggest that not only ventricular end-diastolic geometry and ventricular pressures. To help accel-
geometry but rather rearrangement of fiber architecture in the erate the clinical adoption of personalized simulations, Lamata
deformed ventricles is responsible for the increased vulnerability et al16 developed a robust method that uses a variation of the
to electrical shocks and for reduced defibrillation efficacy in the warping technique to accurately and quickly construct patient-
dilated ventricles. specific meshes from a template heart in a matter of minutes.
Last, the patient’s fiber and sheet orientation needs to be incor- myofilament model have been systematically adjusted to achieve
porated into the model. To date, no imaging method can obtain match between measured and simulated results. These initial
the fiber/sheet architecture from in vivo images. Hence, the first results in the development of patient-specific cardiac electrome-
patient-specific electromechanical models have incorporated chanical models attest to the potential utility of cardiac simula-
fiber orientations adapted from animal hearts or have used rule- tion in the diagnosis and treatment of disorders in the pumping
based methods17 to assign the fiber orientation. Image transfor- function of the heart.
mation algorithms for fiber estimation from in vivo hearty scans
have shown significant promise.18 An example of patient fiber
orientation estimated with such image transformation algorithms
is shown in Figure 36-8, A. Concluding Remarks
Equally important is the process of parameterizing and enrich-
ing the electromechanical model of the heart with the patient’s It can be argued that cardiac electromechanical models are the
own clinical data. Validating the model with patient-specific most sophisticated organ models to date. Electromechanical
metrics personalizes the model and establishes its predictive capa- models now integrate knowledge of the intricate processes across
bilities. The types of electrophysiological data typically used for the scales of biologic complexity and reveal behavior resulting
clinical validation are the standard ECG, total electrical activa- from cooperative interactions across temporal and spatial scales.
tion time, and activation maps determined from endocardial The studies discussed in this chapter demonstrate how electro-
recordings. Figure 36-8, B shows measured and simulated iso- mechanical modeling can advance our understanding of cardiac
chronal maps of ventricular electrical activation from a patient- function and dysfunction. Specifically, we showcase modeling
specific electromechanical model by Sermesant et al.42 To studies that have provided insight into the mechanisms by which
personalize the mechanical problem of the patient’s model heart, mechanoelectrical coupling contributes to arrhythmia induction
the simulation data are compared with pressure and volume and maintenance under a variety of pathologic conditions, and
curves and gross wall motion from cineMRI. Temporal traces of into those pertaining to dyssynchronous electromechanical acti-
LV and RV pressures from a patient-specific electromechanical vation. Modeling studies focusing on optimization of CRT dem-
heart model by Aguado-Sierra et al41 are presented in Figure onstrate that patient-specific cardiac electromechanical modeling
36-8, C, and a comparison of cineMRI data with simulated wall holds high promise in improving the efficacy and providing
motion as calculated from a patient-specific heart model by Nie- enhanced benefits of the clinical procedure.
derer et al35 is shown in Figure 36-8, D. In these models, the Although the use of biophysically based electromechanical
material properties of the myocardium and parameters of the heart models in tailor-made diagnosis, treatment planning, and
Measured depolarization time Simulated depolarization time

25 50 75 100 25 50 75 100
A B 0 120 0 120
Normalized time
120 Simulated LV pressure
Measured LV pressure 0/1 0.33 0.67
100 Simulated RV pressure
Pressure (mm Hg)
Measured RV pressure Base

80 1
MRI slice number
60
1
40 2
2 3
20 MRI slice
number
0
0 200 400 600 800 3 Simulated
Apex wall motion
Time (ms)
C D
Figure 36-8. Patient-specific electromechanical modeling of the heart. A, Mapping. B, Measured (left) and simulated (right) endocardial ischronal maps of electrical activa-
tion with a patient-specific model. C, Left and right ventricular pressure traces from clinical measurements and a patient-specific model. D, Ventricular wall motion (red
lines) simulated by a patient-specific model superimposed on the patient’s cineMRI.
(A, Modified with permission from Vadakkumpadan F, Arevalo H, Ceritoglu C, et al: Image-based estimation of ventricular fiber orientations for personalized modeling of cardiac
electrophysiology. IEEE Trans Med Imaging 31:1051–1060, 2012. B, Modified with permission from Sermesant M, Chabiniok R, Chinchapatnam P, et al: Patient-specific electro-
mechanical models of the heart for the prediction of pacing acute effects in CRT: A preliminary clinical validation. Med Image Anal 16:201–215, 2012. C, Modified with permission
from Aguado-Sierra J, Krishnamurthy A, Villongco C, et al: Patient-specific modeling of dyssynchronous heart failure: A case study. Prog Biophys Mol Biol 107:147–155, 2011.
D, Modified with permission from Niederer SA, Plank G, Chinchapatnam P, et al: Length-dependent tension in the failing heart and the efficacy of cardiac resynchronization
therapy. Cardiovasc Res 89:336–343, 2011.)
prevention of sudden cardiac death and pump dysfunction is model levels, and to advance algorithms and approaches to high-
36
slowly becoming a reality, a number of challenges need to be speed simulation, are of critical importance for implementing
addressed to clear the way for the translation of patient-specific heart modeling at the bedside. Despite these challenges, the
heart modeling into the clinic. Clinical heart imaging data are expectation is that computational modeling of cardiac electrome-
currently of low spatial resolution, posing a challenge for the chanical function will continue to grow because it is providing a
construction of heart models in cases of myocardial structural new, quantitative approach to understanding, detecting, and
remodeling, such as in ischemic cardiomyopathy or interstitial treating cardiac disease.
fibrosis, and hindering the development of efficient segmentation
algorithms with minimum manual intervention. Furthermore,
although the initial application of patient-specific heart models
in the clinic is expected to address optimization of therapies Acknowledgments
whereby rhythm or pump dysfunction is strongly dependent on
patient heart structural remodeling, broader usage of the models This work was supported by National Institutes of Health (NIH)
in cardiac treatment planning will require the development of grant R01-HL103428, and National Science Foundation Grant
capabilities for informing and adjusting the models, on the fly, IOS-1124804 to NAT, and by NIH fellowship F31-HL103090
with patient-specific cardiac electrophysiological and mechanical to JC.
information. Finally, efforts to modularize and interface multiple
15. Prassl AJ, Kickinger F, Ahammer H, et al: Auto- 31. Kuijpers NH, Potse M, van Dam PM, et al: Mech-
References matically generated, anatomically accurate meshes anoelectrical coupling enhances initiation and
for cardiac electrophysiology problems. IEEE affects perpetuation of atrial fibrillation during
1. Plank G, Zhou L, Greenstein JL, Cortassa S, et al: Trans Biomed Eng 56:1318–1330, 2009. acute atrial dilation. Heart Rhythm 8:429–436,
From mitochondrial ion channels to arrhythmias 16. Lamata P, Niederer S, Nordsletten D, et al: An 2011.
in the heart: Computational techniques to bridge accurate, fast and robust method to generate 32. Jie X, Gurev V, Trayanova N: Mechanisms of
the spatio-temporal scales. Philos Transact A Math patient-specific cubic hermite meshes. Med Image mechanically induced spontaneous arrhythmias in
Phys Eng Sci 366:3381–3409, 2008. Anal 15:801–813, 2011. acute regional ischemia. Circ Res 106:185–192,
2. Noble D, Rudy Y: Models of cardiac ventricular 17. Bayer JD, Blake RC, Plank G, Trayanova NA: A 2010.
action potentials: Iterative interaction between novel rule-based algorithm for assigning myocar- 33. Wall ST, Guccione JM, Ratcliffe MB, et al: Elec-
experiment and simulation. Philos Transact A dial fiber orientation to computational heart tromechanical feedback with reduced cellular con-
Math Phys Eng Sci 359:1127–1142, 2001. models. Ann Biomed Eng. In press. nectivity alters electrical activity in an infarct
3. Trayanova NA, Rice JJ: Cardiac electromechanical 18. Vadakkumpadan F, Arevalo H, Ceritoglu C, et al: injured left ventricle: A finite element model study.
models: From cell to organ. Front Physiol 2:43, Image-based estimation of ventricular fiber orien- Am J Physiol Heart Circ Physiol 302:H206–H214,
2011. tations for personalized modeling of cardiac 2012.
4. Nordsletten DA, Niederer SA, Nash MP, electrophysiology. IEEE Trans Med Imaging 34. Kerckhoffs RC, Omens JH, McCulloch AD, et al:
et al: Coupling multi-physics models to cardiac 31:1051–1060, 2012. Ventricular dilation and electrical dyssynchrony
mechanics. Prog Biophys Mol Biol 104:77–88, 19. Trayanova NA: Whole-heart modeling: Applica- synergistically increase regional mechanical non-
2011. tions to cardiac electrophysiology and electrome- uniformity but not mechanical dyssynchrony: A
5. Holzapfel GA, Ogden RW: Constitutive modelling chanics. Circ Res 108:113–128, 2011. computational model. Circ Heart Fail 3:528–536,
of passive myocardium: A structurally based 20. Land S, Niederer SA, Smith NP: Efficient com 2010.
framework for material characterization. Philos putational methods for strongly coupled cardiac 35. Niederer SA, Plank G, Chinchapatnam P, et al:
Transact A Math Phys Eng Sci 367:3445–3475, electromechanics. IEEE Trans Biomed Eng Length-dependent tension in the failing heart and
2009. 59:1219–1228, 2012. the efficacy of cardiac resynchronization therapy.
6. Vetter FJ, McCulloch AD: Three-dimensional 21. Provost J, Gurev V, Trayanova N, et al: Mapping Cardiovasc Res 89:336–343, 2011.
analysis of regional cardiac function: A model of of cardiac electrical activation with electromechan- 36. Spragg DD, Dong J, Fetics BJ, et al: Optimal left
rabbit ventricular anatomy. Prog Biophys Mol Biol ical wave imaging: An in silico-in vivo reciprocity ventricular endocardial pacing sites for cardiac
69:157–183, 1998. study. Heart Rhythm 8:752–759, 2011. resynchronization therapy in patients with isch-
7. Stevens C, Remme E, LeGrice I, et al: Ventricular 22. Provost J, Lee WN, Fujikura K, et al: Imaging the emic cardiomyopathy. J Am Coll Cardiol 56:774–
mechanics in diastole: Material parameter sensitiv- electromechanical activity of the heart in vivo. Proc 781, 2010.
ity. J Biomech 36:737–748, 2003. Natl Acad Sci U S A 108:8565–8570, 2011. 37. Constantino J, Hu Y, Trayanova N: A computa-
8. Legrice IJ, Hunter PJ, Smaill BH: Laminar struc- 23. Kerckhoffs RC, Bovendeerd PH, Kotte JC, et al: tional approach to understanding the electrome-
ture of the heart: A mathematical model. Am J Homogeneity of cardiac contraction despite physi- chanical activation sequence in the normal and
Physiol 272:H2466–H2476, 1997. ological asynchrony of depolarization: A model failing heart, with translation to the clinical prac-
9. Vadakkumpadan F, Arevalo H, Prassl AJ, et al: study. Ann Biomed Eng 31:536–547, 2003. tice of CRT. Prog Biophys Mol Biol 110:372–379,
Image-based models of cardiac structure in health 24. Usyk TP, McCulloch AD: Relationship between 2012.
and disease. Wiley Interdiscip Rev Syst Biol Med regional shortening and asynchronous electrical 38. Niederer SA, Shetty AK, Plank G, et al: Biophysi-
2:489–506, 2010. activation in a three-dimensional model of ven- cal modeling to simulate the response to multisite
10. Gurev V, Lee T, Constantino J, et al: Models of tricular electromechanics. J Cardiovasc Electro- left ventricular stimulation using a quadripolar
cardiac electromechanics based on individual physiol 14:S196–S202, 2003. pacing lead. Pacing Clin Electrophysiol 35:204–
hearts imaging data: Image-based electromechani- 25. Gurev V, Constantino J, Rice JJ, et al: Distribution 214, 2012.
cal models of the heart. Biomech Model Mechano- of electromechanical delay in the heart: Insights 39. Trayanova N, Gurev V, Constantino J, et al: Math-
biol 10:295–306, 2011. from a three-dimensional electromechanical ematical models of ventricular mechano-electric
11. Bishop MJ, Plank G, Burton RA, et al: Develop- model. Biophys J 99:745–754, 2010. coupling and arrhythmia. In Kohl P, Sachs F, Franz
ment of an anatomically detailed mri-derived 26. Niederer SA, Smith NP: The role of the Frank- MR, editors: Cardiac Mechano-electric Feedback
rabbit ventricular model and assessment of its Starling law in the transduction of cellular work to and Arrhythmias. Oxford, 2011, Oxford University
impact on simulations of electrophysiological func- whole organ pump function: A computational mod- Press, pp 256–258.
tion. Am J Physiol Heart Circ Physiol 298:H699– eling analysis. PLoS Comput Biol 5:e1000371, 2009. 40. Trayanova N, Constantino J, Ashihara T, et al:
H718, 2010. 27. Land S, Niederer S, Aronsen JM, et al: An analysis Modeling defibrillation of the heart: Approaches
12. Helm PA, Tseng HJ, Younes L, et al: Ex vivo 3D of deformation-dependent electromechanical cou- and insights. IEEE Rev Biomed Eng 4:89–102,
diffusion tensor imaging and quantification of pling in the mouse heart. J Physiol 590(Pt 2011.
cardiac laminar structure. Magn Reson Med 18):4553–4569, 2012. 41. Aguado-Sierra J, Krishnamurthy A, Villongco C, et
54:850–859, 2005. 28. Keldermann RH, Nash MP, Gelderblom H, et al: al: Patient-specific modeling of dyssynchronous
13. McDowell KS, Vadakkumpadan F, Blake RC, et al: Electromechanical wavebreak in a model of the heart failure: A case study. Prog Biophys Mol Biol
Methodology for patient-specific modeling of human left ventricle. Am J Physiol Heart Circ 107:147–155, 2011.
atrial fibrosis as a substrate for atrial fibrillation. J Physiol 299:H134–H143, 2010. 42. Sermesant M, Chabiniok R, Chinchapatnam P,
Electrocardiol 45:640–645, 2012. 29. Hu H, Sachs F: Stretch-activated ion channels in et al: Patient-specific electromechanical models of
14. Nickerson D, Smith N, Hunter P: New develop- the heart. J Mol Cell Cardiol 29:1511–1523, 1997. the heart for the prediction of pacing acute effects
ments in a strongly coupled cardiac electrome- 30. Hu Y, Gurev V, Constantino J, et al: Novel mecha- in CRT: A preliminary clinical validation. Med
chanical model. Europace 7(suppl 2):118–127, nisms for the suppression of scroll wave breakup Image Anal 16:201–215, 2012.
2005. via mechano-electric feedback. In preparation.
Neural Control of Cardiac PART VI
Electrical Activity
CaV1.2 and β-Adrenergic

Regulation of Cardiac Function 37
Jérôme Leroy and Rodolphe Fischmeister
CHAPTER OUTLINE Molecular Mechanisms of Protein Kinase

Overview of the β-Adrenergic Regulation Regulation of L-Type Calcium Channels
of L-Type Calcium Channels 371
LTCCs display three distinct gating modes upon depolarization:
Molecular Mechanisms of Protein Kinase Regulation mode 1 corresponds to brief openings, mode 2 corresponds to
of L-Type Calcium Channels 371 long lasting openings, and mode 0 corresponds to a silent mode
Compartmentation of cAMP/PKA Regulation because of unavailability.5 PKA phosphorylation of CaV1.2 chan-
nels results in a shift of the channel from the gating mode 1 to
of L-Type Calcium Channels 374
mode 2.6 As a result, ICa,L density is increased twofold to three-
β-Adrenergic Regulation of L-Type Calcium fold, and its voltage dependence shifted slightly toward hyperpo-
Channels in Pathologic Situations 376 larized potentials (Figure 37-2, A). Voltage steady-state activation
and inactivation in adult mouse ventricular myocytes are pre-
sented in Figure 37-2. Activation for ICa,L starts at –40 mV and
is maximal around 5 mV while inactivation begins at –45 mV and
is maximal at approximately 0 mV. The overlap of the two curves
Overview of the β-Adrenergic Regulation defines a “window current” between –40 and 0 mV (i.e., near the
of L-Type Calcium Channels action potential plateau phase). The β-adrenergic stimulation
leads to an increased window current because of the effect of PKA
In cardiac cells, the L-type calcium channel (LTCC) current also phosphorylation on channel activation. As presented in Figure
called ICa,L underlies the plateau phase of the action potential (AP). 37-2, B, isoproterenol application at a maximal concentration of
Upon depolarization, ICa,L reflects calcium influx via the CaV1.2 100 nM shifts the activation by 5 mV toward negative potentials,
channels. This current initiates cardiac contraction by gating the whereas a minor shift of availability of approximately 2.5 mV is
ryanodine receptor, thereby triggering the calcium release from observed. During maintained depolarizations, ICa,L decreases with
the sarcoplasmic reticulum.1 Among several regulatory pathways time, a phenomenon named inactivation, which depends on time,
of this current, the best described is the β-adrenergic stimulation, voltage, and intracellular calcium.7 Because β-AR activation
which contributes to the positive inotropic effects of catechol- enhances calcium entry, it also accelerates ICa,L inactivation,8 and
amines. To date, three β-adrenergic receptors (β-ARs), respec- calcium-dependent inactivation becomes the main mechanism by
tively β1-AR, β2-AR, and β3-AR, have been cloned,2 and this major which the channel inactivates.9 Overall, β-AR stimulation pro-
achievement has led to the 2012 Nobel prize award to Robert motes ICa,L and thus the calcium entry during action potential that
Lefkowitz and Brian Kobilka for paving the road to the current is partially responsible for its positive inotropic effects.
understanding of their structures and functions. The classical The LTCC current ICa,L in the working myocardium is medi-
pathway for β-AR receptor signaling is activation of adenylyl ated by the predominantly expressed CaV1.2 channels. These
cyclases via Gαs, resulting in increased intracellular cyclic adenos- channels are multimeric proteins composed of a central subunit,
ine monophosphate (cAMP) levels. The primary target of cAMP α1C, the pore-forming subunit, which determines the main bio-
is the cAMP-dependent protein kinase (PKA) that in turn phos- physical and pharmacologic properties of the channel. This
phorylates the CaV1.2 channels among other key proteins of the subunit is a protein of approximately 240 kDa with 24 transmem-
excitation-contraction coupling (Figure 37-1). Although direct brane segments according to its hydropathic profile, organized
modulation by G proteins of LTCCs was first suggested to par- into four repeated domains (I to IV) of six transmembrane seg-
tially mediate the upregulation of ICa,L upon β-AR activation, it ments each (1 to 6), with intracellular N- and a large C-termini.4
has been clearly established that a cAMP phosphorylation medi- Like other high-voltage–gated calcium channels, CaV1.2 associ-
ated by PKA is responsible for this increase.3,4 ates with a largely extracellular disulfide-linked α2-δ subunit of
This chapter reviews the literature on the β-AR regulation of 170 kDa.10 It also binds CaVβ subunits to its α interaction domain
LTCCs, with emphasis on recent information on the molecular present in its intracellular I-II loop via their guanylate kinase–like
mechanisms of PKA regulation of CaV1.2 channels and the local domain.11 Four genes encode four CaVβ subunits (β1-4),
compartmentation of β-AR/cAMP/PKA signaling around these but CaVβ2 is thought to be the main isoform expressed in the
channels. We conclude with an overview of such modulation in heart.12 Both α2-δ13, and CaVβ14 auxiliary subunits influence the
a pathologic context. Additional details concerning LTCCs can biophysical properties and increase the trafficking of the channel
be found in Chapter 2. at the plasma membrane. γ(4,6-8) subunits are also expressed in
371
372 NEURAL CONTROL OF CARDIAC ELECTRICAL ACTIVITY
β-AR AC Cav1.2
Ext
Int P
Gβγ Gαs
cAMP
ATP
PKA
Figure 37-1. β-Adrenergic modulation of cardiac L-type calcium channels. Modulation of cardiac CaV1.2 channels by β-adrenergic receptor (β-AR) stimulation occurs via a
G protein pathway. β-ARs couple to heterodimeric G proteins. Agonist catecholamine binding to β-AR activates the stimulatory heterotrimeric G protein by inducing the
exchange of guanosine diphosphate (GDP) for guanosine-5′-triphosphate (GTP) on GαS. GαS dissociates from its Gβγ partner to stimulate the adenylyl cyclase (AC) present
in the plasma membrane. The AC catalyses the conversion of adenosine triphosphate (ATP) in cyclic 3’,5’ monophosphate (cAMP) that in turn activates the cAMP dependent
protein kinase (PKA) to phosphorylate CaV1.2.
1.0
Relative conductance
100 ms
Availability Activation
6 pA/pF 0.5
0.0
–60 –50 –40 –30 –20 –10 0 10 20
A B Em (mV)
Control
Isoprenaline 100 nM
Figure 37-2. The effect of β-adrenergic stimulation on ICa,L in adult mice ventricular cardiomyocytes. A, Example of current traces of ICa,L recorded using the patch-clamp
technique under whole-cell conditions at 0 mV (holding potential at –50 mV) obtained before and after activation of the β-adrenergic receptors with 100 nM isoproterenol.
An approximately twofold increase of ICa,L amplitude is induced by the application of the agonist. B, Influence of β-adrenergic stimulation on cardiac calcium channel
activation and availability: ICa,L availability and steady-state inactivation were measured by applying a 200-ms pulse from –60 to 60 mV, followed by a 3-ms repolarization
to –50 mV, before a 200-ms test pulse to 0 mV. The present curves are mean data obtained from 16 different adult mice ventricular myocytes before and after application
of 100 nM isoproterenol. Steady-state inactivation curves were obtained by normalizing the peak current at each test potential to the maximal current. Activation curves
were derived from current–voltage relations and described by: G / Gmax = 1 / (1+exp[{V½,act − V} / k]). A and B were obtained in similar recording conditions as described
by Leroy et al.91
ventricular cells and, like γ1 for the skeletal muscle calcium the difficulty of reconstituting such modulation in heterologous
channel CaV1.1, can interact with the CaV1.2 channel to modu- overexpression systems. The α1C-subunit of CaV1.2 channels
late its function when coexpressed in HEK-293 cells; however, exhibits multiple potential PKA phosphorylation sites in the N-
the exact role for these accessory proteins on CaV1.2 in native and the C-terminal regions (Figure 37-3).2 Despite the fact that
cardiac tissues remains to be determined.15 α1C is a substrate for phosphorylation by PKA in vitro,16,17 several
attempts to mimic the adrenergic stimulation of CaV1.2 channels
in expression systems have failed.18-20 This failure led to the
α-1 Subunit hypothesis that at least one missing link in heterologous systems
would preclude the reconstitution of such modulation of overex-
If the main effects of β-AR stimulation on voltage dependence pressed CaV1.2 channels. One of these links could be an A-kinase-
and amplitude of the ICa,L and their consequences for the fight- anchoring protein (AKAP) that allows the cAMP modulation and
or-flight response are well documented in the literature, the the PKA phosphorylation of serine 1928 in the C-terminus of
molecular events that mediate the increase of CaV1.2 activity CaV1.2 channels in HEK293 cells.21 A conserved leucine zipper
during a sympathetic stimulation remain elusive; this is due to motif in the C-terminus of CaV1.2 identified in native cardiac
CaV1.2 AND β-ADRENERGIC REGULATION OF CARDIAC FUNCTION 373
α2-δ
Cav1.2
α1C
ss
ss
37
I II III IV
12345 6
+ + + +
+ + + +
N
P 70
0
AID S1
04
P T17 C
Ahnak GK 8
S47 9 P P
S47
928 15
S1
β2a AP
AK
A
P PK
Figure 37-3. Subunit structure of cardiac L-type CaV1.2 channel and phosphorylation sites for PKA mediated upregulation of ICa,L. The α1C-subunit constitutes the pore-
forming subunit of the cardiac CaV1.2 channel. It presents four domains (I to IV) composed of six transmembrane segments (S1 to S6), with S4 exhibiting positively charged
residues that confer their sensitivity to potential and a pore loop between S5 and S6 of each domains. The intracellular I-II loop of the α1C-subunit associates with the β
interaction domain (BID) in the guanylate kinase–like domain of the intracellular CaVβ2a auxiliary subunit via its α interaction domain (AID). CaV1.2 associates also with the
largely extracellular glycosylphosphatidylinositol (GPI)-anchored α2-δ subunit. The C-terminal part of the α1C undergoes proteolytic cleavage by calpain at residue 1821.
The cleaved distal fragment associates monocovalently with the proximal C-terminal part to inhibit the channel. This necessary association for β-AR regulation of CaV1.2 is
relieved by the phosphorylation of serine 1700 by the PKA tethered in the distal C-terminus by AKAP15 to produce the upregulation of channel activity. PKA also phos-
phorylates serine 1704 to control its basal activity while phosphorylation of serine 1928 serves an undetermined regulatory function. PKA also phosphorylates CaVβ2a on
the two serine residues 478 and 479. This subunit associates to the giant ahnak cytosqueletal protein that could also be an important factor for the cAMP/PKA regulation
of CaV1.2 channels.
cells directly anchors a low molecular weight AKAP15-PKA upregulation of the current by cAMP/PKA pathway, whereas
complex to ensure a fast and efficient β-adrenergic modulation expression of mutated CaVβ2a at serine 478 and 479, but not at
of ICa,L (see Figure 37-3),22 allowing its phosphorylation at serine position 459, was unable to do so.28 The authors concluded that
1928 when β-ARs are stimulated.23 Surprisingly, the C-terminal the CaVβ2a ancillary subunit was the main target for PKA to
part of rabbit CaV1.2 undergoes proteolytic processing by calpain mediate the β–AR stimulation of ICa,L; however, this conclusion
at residue 1821, leaving two size forms of the α1C-subunit of these has been questioned by experiments realized in a more physio-
channels expressed in cardiac cells,17 whereas this distal 37 to logic context. Overexpression of a CaVβ2a construct mutated for
50-kDa peptide contains the serine 1928 phosphorylated by PKA its PKA phosphorylation sites in ventricular cardiac cells did not
and the binding site for AKAP15. In fact, this fragment consti- prevent the cAMP/PKA modulation of ICa,L.14 Nonetheless, the
tutes a potent autoinhibitory domain that covalently associates CaVβ2 subunit can associate with the giant cytoskeletal protein of
with the proximal C-terminal part of α1C, reducing its open prob- 700 kDa named ahnak, which emerged as an important player in
ability and shifting the voltage dependence of activation to depo- the β-AR regulation of cardiac ICa,L (see Figure 37-3). Through
larized potentials.24 The role of serine 1928 in ICa,L upregulation its interaction with CaVβ2, ahnak would serve as a brake on ICa,L
by β-AR has since been challenged. A first study showed that a that would be relieved by PKA phosphorylation of the ancillary
DHP-resistant CaV1.2 channels mutated at position 1928 subunit and the cytoskeletal protein.29 Interestingly, ahnak poly-
(S1928A) displays an unaltered response to the β-AR agonist morphism occurs, and the genetic variant generated interferes
isoproterenol when overexpressed in isolated ventricular myo- with the β-AR stimulation of ICa,L by reducing the CaVβ2 interac-
cytes.25 Moreover, the generation of a S1928A knock-in mouse tion with ahnak.29 However, if the involvement of the CaVβ or
model confirmed that the phosphorylation of this residue by PKA other binding partners such as ahnak cannot be ruled out com-
is not required for the functional effects of β-AR stimulation on pletely, recent studies have reaffirmed the importance of the
CaV1.2 in cardiac cells, because these mice exhibit an essentially proteolytically cleaved distal C-terminus of CaV1.2 for its upreg-
conserved response of ICa,L to isoproterenol and typical chrono- ulation upon β-AR stimulation. In addition, a new model for the
tropic and inotropic responses to β-AR stimulation.26 Therefore, molecular basis of β-AR stimulation of CaV1.2 has been pro-
although phosphorylation of serine 1928 within the C-terminus posed.30 Overexpression in TsA-201 cells of a truncated CaV1.2
of CaV1.2 definitely occurs when β-ARs are stimulated,17,21,23 it channel at position A1800 (the site of in vivo proteolytic cleavage
does not correlate with the functional effects observed on ICa,L. previously determined for skeletal muscle CaV1.1 channel31) with
As a result, PKA would phosphorylate other PKA sites within the α2δ1 and CaVβ1b and the distal C-terminal part of the channel
channel or another binding partner to mediate the increase of its (peptide from 1821-2171) produced a functional channel that is
activity. autoinhibited. Two presumed sites for PKA phosphorylation
were identified upstream from the proteolytic site—a serine at
position 1700 and a threonine 1704 at the interface of the distal
β-Subunit and the proximal C-terminal parts. Although a mutation of S1928
confirmed that its phosphorylation is not required for the
Three serines (S459, S478, and S479) of the cardiac the CaVβ2a increase of the current, substitution of T1704 and S1700 for
subunit were identified as putative PKA phosphorylation sites.27 alanines revealed that phosphorylation of T1704 is required for
Coexpression of this subunit with a CaV1.2 lacking the C-terminal basal CaV1.2 channel activity, whereas S1700 is crucial for its
part including the serine 1928 in TsA-201 cells allowed an cAMP/PKA-induced upregulation. In this scheme, the AKAP15
is still required to coordinate the PKA phosphorylation of S1700 serotonin 5-HT4 receptors,57 which are also coupled to increases
that disrupts the interaction of the noncovalently distal in force of contraction58 or ICa,L59 in atria but not in healthy
C-terminus, thus relieving the inhibition of the channel.30 This ventricles.
assumption is reinforced by the fact that mice expressing a CaV1.2
channel deleted for the distal C-terminus display altered cAMP/
PKA regulation accompanied with reduced expression of Role of A-Kinase-Anchoring Proteins
AKAP15.32
Cyclic AMP signaling components are organized into multipro-
tein complexes, an arrangement that increases both efficiency and
specificity of the transduction cascade. AKAPs have an essential
Compartmentation of cAMP/PKA role in these arrangements. AKAPs form a large family of proteins
Regulation of L-Type Calcium Channels comprising more than 50 members whose primary function is to
anchor PKA in the vicinity of its substrates, thus ensuring the
In light of the knowledge accumulated over the years, it is evident preferential phosphorylation of a limited number of targets.60,61 As
that intracellular cAMP is not uniformly distributed, and nor is discussed earlier, AKAP15 (also called AKAP 7 or AKAP15/18) is
PKA uniformly activated within cardiomyocytes upon β-AR stim- the main AKAP controlling the PKA phosphorylation of LTCCs.
ulation.33 On the contrary, a tight cAMP/PKA compartmentation However, in a recent study, CaV1.2 phosphorylation and β-AR
is required for adequate processing and targeting of the informa- stimulation of ICa,L was found to be unchanged in mice in which
tion generated at the cell surface, conferring the specificity of the the gene encoding this protein was inactivated, suggesting that
response to various hormones linked to Gαs-coupled receptors.34,35 PKA is anchored by a different protein to the channel.62 Further-
In the case of β-ARs, several processes contribute to a localized more, AKAP5 was found to target ACs, PKA and phosphatases
cAMP/PKA response: catecholamines activate different β-AR within caveolae to allow specific PKA phosphorylation of the
subtypes located at different places of the cell surface (e.g., caveo- subpopulation of channels present in this compartment upon
lae, T-tubules); Gαs activation of different adenylyl cyclase (AC) β-AR stimulation.63 Although AKAPs share in common their
isoforms can lead to cAMP synthesis at different locations; cAMP ability to bind PKA, they are remarkably diverse scaffold proteins.
diffusion may be restricted because of localized phosphodiesterase Within each signalosome, AKAPs couple PKA to different sub-
(PDE) activity; anchoring of PKA to AKAPs position PKA at dif- strates, enhancing the rate and fidelity of their phosphorylation by
ferent subcellular compartments to selectively phosphorylate a the kinase. Importantly, AKAPs not only bind PKA but act as
local pool of proteins for specific cellular processes.36,37 In addi- scaffold proteins for other signaling components, including phos-
tion, AKAPs also ensure that PKA is coupled to its upstream phatases 1 and 2,64,65 Epac,66 adenylyl cyclases,67,68 and PDEs.69
activators, including membrane β-ARs and ACs, and to signal Recently, it has been demonstrated that the phosphoinositide
termination enzymes such as PDEs and phosphatases.36 3-kinase p110γ, that was shown recently tether PKA,70 orches-
trates multiprotein complexes including different PDEs to control
cardiac CaV1.2 phosphorylation during β-AR stimulation.71 The
Receptor Specificity combination of PDEs and phosphatases present in individual
AKAP complexes will affect the duration, amplitude, and spatial
Although only three types of β-adrenergic receptors (β-ARs) have extent of cAMP/PKA signaling. Thus, by bringing together dif-
been cloned (β1-, β2-, and β3-ARs), the effect of catecholamines ferent combinations of upstream and downstream signaling mol-
in the human heart is generally attributed to β1- and β2-ARs. ecules, AKAPs provide the architectural infrastructure for
β1- and β2-ARs are highly homologous receptors and are specialization of the cAMP signaling network.61,66,72
both positively coupled to AC/cAMP/PKA cascade, ICa,L and
cardiac performance33; however, they exert opposite effects on
hypertrophy38,39 and apoptosis,40,41 and their respective contribu- Role of Phosphodiesterases
tion varies significantly depending on the cardiac tissue, patho-
physiologic state, age, or developmental stage.42 Part of the Localized cAMP signals can be generated by the interplay
difference is because β2-ARs couple not only to Gαs but also to between discrete cAMP production sites and restricted diffusion
Gαi proteins, and this confines the cAMP-dependent signal to the within the cytoplasm. Restricted diffusion of cAMP can be
membrane compartment and to activation of the LTCCs.43 achieved by several means. A first possibility is that physical bar-
Another difference between the two β-AR subtypes is their loca- riers are created by specialized membrane structures within the
tion at the cell surface, with β2-ARs present in the caveolae/lipid cytoplasm. This method was initially proposed to explain
rafts44,45 of the transverse-tubular structure46 and β1-ARs distrib- the differences in cAMP concentration elicited by PGE1 at the
uted throughout both caveolae–lipid rafts and nonlipid raft mem- plasma membrane and in the bulk cytosol of HEK293 cells,
brane domains44 and in both plasma and T-tubular membranes.46 although an experimental proof that this actually occurs is still
Accordingly, the β2-AR downstream activation of ICa,L is sensitive lacking.73 Another important mechanism that limits cAMP diffu-
to disruption of caveolae by cholesterol depletion, whereas the sion is cAMP degradation by PDEs, which appears to be critical
β1-AR stimulatory effect is not.47 Therefore, β2-AR stimulation to the formation of dynamic microdomains that confer specificity
exerts a local activation of LTCCs, whereas β1-AR stimulation of the response.35,60
leads to activation of LTCCs in the distance.43,48,49 Cardiac cAMP PDEs degrading belong to five families (PDE1
The β3-AR differs from β1-and β2-AR subtypes in its molecu- to PDE4 and PDE8) that can be distinguished by distinct enzy-
lar structure and pharmacologic functions.50 Expression of β3-AR matic properties and pharmacology.74 Among these families
was demonstrated in human myocardium at the mRNA51,52 and various enzymes were shown to degrade cAMP to allow a fine
protein levels.52-55 Interestingly, β3-AR activation produces a tuning of ICa,L regulation by PKA in the heart. Although PDE2 is
negative inotropic effect in human endomyocardial biopsies from not highly expressed in cardiomyocytes, it controls LTCC activity
transplanted hearts50,51 and in left ventricular samples from failing in various species, including human atrial myocytes.75-78 This
and nonfailing explanted hearts50,54 that is due to activation of the enzyme is activated by cGMP, and stimulation of guanylyl cyclase
NO/cGMP pathway, but increases ICa,L and contractility in strongly decreases local cAMP levels controlling ICa,L with only
human atrial tissue via the cAMP/PKA pathway (Figure 37-4).56 modest effects on its global concentration, suggesting the exis-
This effect is reminiscent of the contractile effects of the tence of a cAMP microdomain including β-AR and LTCCs under
37
AR
AC
3-
LT
β
C
C
Gs
cAMP PKA
ATP Ca2+
Stimulatory effectin atrium
AR
3-
LT
β
C
C
Gi/o S
eNO
NO sGC
Ca2+
GTP cGMP PKG
Inhibitory effects in ventricle
Figure 37-4. Opposite effect of β3-adrenergic receptor stimulation on L-type calcium channel (LTCC) activity in a human atrium and ventricle. In the atrium (top), the β3-
adrenergic receptor (β3-AR) is positively coupled to the Gs-protein and to adenylyl cyclase (AC). Upon activation of the β3-AR, cAMP synthesis is increased, which leads to
PKA phosphorylation of LTCCs and a stimulation of Ca2+ influx. In the human ventricle (bottom), the β3-AR is coupled to the endothelial nitric oxide synthase (eNOS), pre-
sumably via the Gi/o-protein. Activation of eNOS leads to NO production, which activates soluble guanylyl cyclase (sGC) and increases intracellular cGMP concentration.
This leads to PKG-phosphorylation of LTCCs and an inhibition of Ca2+ influx.
tight control of PDE2.79 Contrary to PDE2, PDE3 is inhibited by PDE3 inhibitor milrinone in guinea pig perfused hearts. Whereas
cGMP; this explains in part why cGMP at low concentration can each of these treatments increased intracellular cAMP and pro-
also increase basal ICa,L as shown in human atrial myocytes.78,80 duced positive inotropic and lusitropic effects, differences in the
In rodents, PDE3 and PDE4 are the major contributors to phosphorylation pattern of PLB, TnI, and MyBP-C by PKA
the total cAMP-hydrolytic activity34,81 and PDE4 is dominant to were observed.98 These results were attributed to a functional
modulate β-AR regulation of cAMP levels.49,82-84 Multiple PDE4 cellular compartmentation of cAMP and PKA substrates owing
variants associate with β-ARs,85-87 RyR2,88 SERCA2,89,90 ICa,L,91 to a different expression of PDEs at the membrane and in the
and IKs72 to exert local control of ECC. In larger mammals, PDE3 cytosol.92 In canine ventricular myocytes, an increase in particu-
activity is dominant in microsomal fractions,92-94 and PDE3 inhib- late but not total cAMP correlated to an increase in Ca2+ transient
itors exert a potent positive inotropic effect.95 Selective inhibition amplitude and decay kinetics.99 In response to β-AR stimulation,
of PDE3 with milrinone has been shown to improve cardiac approximately 45% of the total cAMP was found in the particu-
contractility in patients with congestive heart failure.96 The late fraction, but this fraction declined to less than 20% when
role of PDE4 is less well defined, but evidence is emerging that IBMX was added to Iso, although cAMP production was up to
PDE4 could also have an important role in these species. In the threefold to fourfold greater. These results show that cAMP-
canine heart, a large PDE4 activity is found in the cytoplasm,92 but PDEs reside predominantly in the cytoplasm, where they prevent
PDE4 is also present in microsomal fractions, where it accounts excessive cAMP accumulation upon β-AR activation. Thus,
for approximately 20% of the activity.93 Recent studies have indi- PDEs appear to be important to maintain the specificity of the
cated that PDE4 is expressed in the human ventricle where, β-AR response by limiting the amount of cAMP diffusing from
similar to rodents, it associates with β-ARs, RyR2, and phosphol- membrane to cytoplasm. Similar results were obtained when
amban.81,88 Moreover, PDE4 is the main PDE modulating LTCC studying the effect of PDE inhibition on ICa,L regulation by local
activity in rodent cardiomyocytes,77,84 and PDE4 was recently application of Iso in frog ventricular myocytes.100 In the absence
shown to control ICa,L, ECC and arrhythmias in the human of IBMX, the application of Iso to half of the cell increased ICa,L
atrium.97 half maximally, corresponding to activation of the channels
Early evidence of the contribution of PDEs to intracellular located in the same part of the cell as the β-AR agonist. When
cyclic nucleotide compartmentation was obtained by comparing IBMX was added with Iso on half of the cell, the effect of Iso was
the effects of the nonselective β-AR agonist Iso, or the nonselec- greatly potentiated because in this condition, LTCCs in the
tive PDE inhibitor, 3-isobutyl-1-methylxanthine (IBMX), or the remote part of the cells could be recruited. These results suggest
that PDE activity is important for the definition of local cAMP common is the autonomic nervous system.110 In pathologic con-
pools involved in the β-AR stimulation of LTCCs. Subsequent ditions such as atrial fibrillation (AF), cardiac hypertrophy or
studies using ratiometric FRET biosensors to monitor cAMP heart failure (HF), these factors are united to trigger fatal cardiac
directly have shown that the second messenger increases prefer- arrhythmias. These pathologies are accompanied with structural
entially in discrete microdomains corresponding to the dyad changes, electrophysiological remodeling, and abnormal β-AR
region under β-AR stimulation, and that cAMP diffusion is activation at tissue and cellular levels promoting aberrant electri-
limited by PDE activity.101,102 Other studies using recombinant cal activities and modulation. Modified excitation-contraction
cyclic nucleotide gated channels to measure cAMP generated at coupling owing to altered calcium homeostasis in pathologic con-
the plasma membrane identified specific functional coupling of ditions is a major cause of arrhythmias.111 As described in the
individual PDE families, mainly PDE3 and PDE4, to β1-AR, previous paragraph, a fine tuning of the LTCC (i.e., of calcium
β2-AR, PGE1-R, and Glu-R as a major mechanism enabling cycling) and its regulation in discrete subcellular compartments
cardiac cells to generate heterogeneous cAMP signals in response is required to achieve a normal cardiomyocyte function. In phys-
to different hormones.84 iopathologic conditions, deregulation of this coupling leads to
Although PDE4 regulates ICa,L in cardiomyocytes,77,84 the calcium waves and calcium alternans to promote reentrant
molecular identity of the PDE4 regulating the LTCC was arrhythmias and triggered activities.111,112 At the level of the car-
unveiled only recently.91 In mouse cardiomyocytes, PDE4B and diomyocyte in AF113 or in HF,114 calcium handling is perturbed,
PDE4D, but not PDE4A, were found to be part of a CaV1.2 leading to afterdepolarizations. When action potential duration
signaling complex. However, in mice deficient for the Pde4d gene (APD) is excessively prolonged, early afterdepolarization (EAD)
(Pde4d–/–), basal or β-AR stimulated ICa,L were not different from can occur during the repolarization phase, whereas arrhythmo-
wild-type mice, whereas in Pde4b–/– mice the β-AR response of genic delayed afterdepolarization (DAD) can occur when the cell
ICa,L was increased, together with an increase in cell contraction is fully returned to its resting potential. Non-reentrant mecha-
and Ca2+ transients.91 Upon β-AR stimulation in vivo, catheter- nisms involve triggered activities from either EADs or DADs.
mediated burst pacing triggers ventricular tachycardia in Pde4b–/– When APD is prolonged, owing to alteration of Na+ or K+ con-
mice but not in wild type.91 Thus, PDE4B is the main PDE4 ductances (as in long QT syndromes or upon antiarrhythmic
isoform regulating cardiac LTCC activity and has a key role treatments), ICa,L recovery from inactivation occurs during the
during β-AR stimulation of cardiac function. PDE4B, by limiting plateau phase of the AP which causes EADs.115 This is particu-
the amount of Ca2+ that enters the cell via the LTCCs, prevents larly true when ICa,L window current is increased upon β-AR
Ca2+ overload and arrhythmias. activation, thus increasing calcium entry during the plateau phase
and allowing its reactivation.115,116 EADs are also promoted by
increased calmodulin-dependent protein kinase II (CaMKII)
Role of Phosphatases activity in hypertrophy and HF. CaMKII is activated by increased
intracellular calcium levels upon β-AR activation via PKA-
Formation of inside-out patches from rabbit ventricular myo- dependent and independent mechanisms.117,118 CaMKII in turn
cytes causes run-down of LTCC activity that is blocked by phosphorylates CaV1.2 on its CaVβ2a subunit, triggering afterde-
okadaic acid, a serine/threonine phosphatase inhibitor.103 This polarizations.119,120 However, intracellular calcium overload, by
observation provided early functional evidence that a phospha- activating the reverse mode of the Na+/Ca2+ exchange, generates
tase is anchored in close proximity to the channel that counteracts a depolarizing current during the late AP phase 2 and phase 3,
upregulation of CaV1.2 by phosphorylation. Later, two major which is believed to act synergistically with ICa,L to induce
cardiac serine–threonine phosphatases,104 phosphatase 2A (PP2A) EADs.121,122 Similarly, DADs occur at high intracellular calcium
and 2B (PP2B or calcineurin), were found to associate with load, and their incidence is increased by both rapid pacing and
cardiac CaV1.2.105,106 For PP2A, two attachment sites were identi- β-AR stimulation.114,123 The main triggering mechanism for
fied within the C-terminus of the α1-subunit:107,108 one region DADs is spontaneous calcium release activating a transient
spans residues 1795 to 1818 and the other spans residues 1965 to inward current (ITI) because of excess diastolic calcium handled
1971. PP2B binds immediately downstream of residues 1965 to by sarcolemmal Na+/Ca2+ exchange. Large enough DADs can
1971 without competition between these two phosphatases for eventually sufficiently depolarize the membrane potential to
binding to this rather narrow region.108 Injection of a peptide that reach threshold and trigger an AP. Again, increased LTCC activ-
contains residues 1965 to 1971 and displaces PP2A but not PP2B ity contributes to DAD occurrence and Ca2+-evoked arrhythmias.
from endogenous CaV1.2 increases basal and β-AR stimulated For example, mutations in the LTCC have been associated with
cardiac ICa,L.108 Similarly, inhibition of PP2B with cyclosporin A a number of inherited arrhythmia syndromes.124 Increased LTCC
or the calcineurin autoinhibitory peptide increases cardiac ICa,L.109 owing to CaV1.2 mutations such as those depicted in the Timothy
This indicates that anchoring of PP2A and PP2B on cardiac syndrome when reinforced upon β-AR stimulation is a source for
Cav1.2 negatively regulates LTCC activity, most likely by coun- intracellular calcium disorders triggering DADs and severe
terbalancing basal and stimulated phosphorylation that is medi- arrhythmias.125 These observations demonstrate that a fine tuning
ated by PKA and possibly other kinases. of the β-AR modulation of calcium entry via CaV1.2 channels in
the myocytes is required to maintain proper calcium homeostasis
and electrical activity of the heart.
β-Adrenergic Regulation of L-Type Calcium

Channels in Pathologic Situations Atrial Fibrillation
Atrial fibrillation (AF) is an extremely common cardiac arrhyth-
CaV1.2 Involvement in Early and Delayed mia, most prevalent in elderly people, that is profoundly influ-
Afterdepolarizations enced by the autonomic nervous system, especially by the
adrenergic component.126 It is accompanied by APD shortening
According to the Coumel’s triangle, the production of a clinical and intracellular calcium homeostasis disorders.113,127 Decreased
arrhythmia requires three ingredients: an arrhythmogenic sub- depolarizing ICa,L contributes to such APD shortening by favor-
strate, a trigger factor, and modulation factors of which the most ing repolarization thus reentry substrates. If reductions in ICa,L
have been observed consistently in atrial myocytes from patients channels is well described in animals models for HF, it seems
37
with permanent AF128-130 or from patients in sinus rhythm with a largely accepted that ICa,L amplitude is maintained especially in
high risk of AF,131,132 the exact mechanism for such downregula- failing human cardiomyocytes.159,160 This apparent discrepancy
tion is not clearly established. This reduction might be primarily might be explained by an increased open probability of single
the result of transcriptional downregulation of the α1C-subunit of CaV1.2 channels because of either their increased phosphoryla-
LTCCs via a Ca2+-dependent calmodulin-calcineurin-NFAT tion161,162 to altered PKA and phosphatase activities155,163 or
system at the cost of the APD reduction observed in AF.133 increased expression of CaVβ ancillary subunits.164,165 Further-
Besides, it has been also associated with diminished expression of more, cardiac remodeling induced by chronic activation of β-AR
the CaVβ and α2-δ accessory subunits,134 which are essential for signaling activates CaMKII,166,167 which is known to increase the
α1C trafficking to the plasma membrane. Normal trafficking of CaV1.2 channel activity120 and participate in calcium influx
CaV1.2 might as well be impaired by a zinc binding protein remodeling in heart failure.168 Blunted β-AR stimulation of ICa,L
(ZnT-1) upregulated in patients with AF.135 Upregulation of a in human failing cardiomyocytes has been consistently
microRNA, miR-328, has also been correlated to AF producing reported,162,169 probably because of more abundant heteromeric
downexpression of α1C and CaVβ1.136 Oxidative stress that occurs Gi/oα proteins,170-172 leading to an increased β2-R receptor cou-
in AF could also participate to decrease LTCC activity by induc- pling with Gi antagonizing the β1-AR stimulation of ICa,L.173 In
ing S-nitrosylation of α1C137 and by promoting its β-AR stimula- addition, increased Gi/oα modifies PP1 and PP2A activities,
tion to trigger arrhythmogenic EADs.138 ICa,L amplitude is clearly which are two phosphatases that control LTCC phosphoryla-
decreased in AF, but its β-AR regulation is not systematically tion.163 GRK2 could also contribute to the remodeling of the
modified.139 Although β-AR receptor density is not impaired in β-AR stimulation of ICa,L in HF, because knockout mice for this
AF140, polymorphism of β1-AR occurs in patients with AF, leading kinase appear to be more resistant to adverse remodeling follow-
to decreased β-AR cascade,141 that could partially explain the ing myocardial infarction, but they demonstrated an increased
decreased LTCC activity in AF. Another plausible mechanism has basal ICa,L amplitude and blunted β-AR stimulation.174 Further-
been proposed as a decreased phosphorylation of the CaV1.2 more, overexpression of βARKct, a peptide derived from the
channel because of suppress to the increased phosphatase 1 or GRK2 C-terminus, is able to increase β-AR stimulation of ICa,L
phosphatase 2A activities observed in AF.130,142 This might be also independently of PKA in normal and failing cardiomyocytes by
influenced by the hyperphosphorylation by PKA of the peptide sequestering Gβγ proteins.175 Moreover, expression of a PDE4
inhibitor 1 that is prominent in AF to increase its inhibitory effect isoform—PDE4B, which modulates β-AR stimulation of ICa,L91—
on PP-1.142,143 Furthermore, depressed expression of the major has been shown to be decreased in a rat model of compensated
PDE4 isoform detected in human atria, PDE4D, correlates with hypertrophy,176 suggesting that not only the production but the
aging a well-known favoring factor for AF development.97 enzymatic degradation of the cAMP controlling ICa,L is affected
Although PDE3 is the main cAMP-degrading enzyme in human in pathological conditions. Thus, in hypertrophy and HF, the
atria, PDE4 substantially contributes to the enzymatic control of β-AR regulation of CaV1.2 channels is altered, contributing to a
β-AR stimulation of ICa,L in human atria, and its inhibition leads dysregulation of calcium handling promoting calcium-induced
to arrhythmias because of dysregulated intracellular calcium arrhythmias. Accurate calcium influx is not only a determinant
homeostasis.97 All these observations suggest that altered β-AR for normal cardiomyocyte function during the excitation-
modulation of CaV1.2 channels happens in AF to contribute to contraction coupling; recent evidence in the literature high-
such arrhythmia. lighted a possible role of CaV1.2 channels in gene transcription
and prohypertrophic signaling.177 Few studies have stated that
chronic increase calcium influx via CaV1.2 channels is deleterious
Hypertrophy and Heart Failure by promoting prohypertrophic signaling pathways. For example,
mice overexpressing the α1C-subunit of CaV1.2 channels exhibit
In human HF, a chronic activation of β-AR induced by the eleva- cardiac growth and cardiomyopathy hypertrophy178 and similarly,
tion of circulating catecholamine levels occurs and contributes to overexpression of CaVβ2a leads to cardiac hypertrophy by increas-
the progression of the disease,144,145 stressing the necessity for a ing ICa,L, and activation of calcineurin/nuclear factor of activated
strict control of β-AR signaling. This hypothesis is further dem- T cells (NFAT) and CaMKII/HDAC signaling pathways.179
onstrated in animal models with exacerbated β-AR/cAMP/PKA These assumptions are confirmed by the fact that LTCC blockers
signaling and that develop cardiac hypertrophy and heart can prevent cardiac remodeling in animal subjected to pressure
failure.146-148 A hallmark of heart failure is β-AR desensitization overload180 and by the attenuation of cardiac hypertrophy
with a loss of β-AR signaling compartmentation35,144 and intracel- observed when the expression of CaVβ2 subunit is inhibited.181
lular calcium cycling perturbations,114 accompanied by profound Astonishingly, decreasing the expression of α1C in mice also pro-
alterations of the structure of the cardiomyocytes notably a loss duces cardiac hypertrophy and HF, but this deleterious effect
of the T-tubules.149,150 Desensitization of β-AR is promoted by its appears not to be connected directly to the diminished ICa,L
phosphorylation by the G-protein–coupled receptor kinase 2 amplitude, but more possibly to the compensatory neuroendo-
(GRK2), a serine/threonine kinase upregulated in hypertrophy crine stress induced to balance the decreased contractility.182
and HF.151 The number of functional β-AR is reduced in HF, and Recently, it has been proposed that PKA phosphorylation of
their localization is changed, with a redistribution to cell crests the serine 1700 of α1C relieves its autoinhibition by inducing the
of the β2-AR subtype normally localized in the T-tubules, whereas dissociation of its proteolytically cleaved and covalently reassoci-
β1-R remain uniformly distributed at the membrane.46 Interest- ated distal C-terminus from the proximal C-terminus.30 Interest-
ingly, 80% of the CaV1.2 channels are located in the T-tubules152 ingly, the C-terminus part of α1C can translocate to the nucleus
where they are targeted by the protein BIN-1.153 The number of to act as a transcription factor for its own expression and to
T-tubular LTCCs is decreased in failing myocytes,154,155 a reduc- activate hypertrophic signaling.183 A possible role of its C-terminus
tion associated with a down-regulation of BIN-1 in human failing in hypertrophy and HF is reaffirmed by recent studies showing
cardiomyocytes.156 This must contribute to the decreased EC that deletion of the terminal part of the α1C-subunit results in HF
coupling157 that might be also affected by the architectural reor- in vivo.32,184 It is tempting to speculate that chronic β-AR stimula-
ganization of the dyadic cleft.157,158 Nonetheless, there is no real tion promotes hypertrophy and HF by favoring a permanent
consensus in the literature concerning a reduction of ICa,L ampli- dissociation of the C-terminus of the pore-forming subunit of
tude in failing cardiomyocytes. If the reduction of CaV1.2 CaV1.2 channels to induce cardiac remodeling.
channels requires membrane targeting of PKA 40. Communal C, Singh K, Sawyer DB, et al: Oppos-
References and phosphorylation of channel subunits. Neuron ing effects of β1- and β2-adrenergic receptors on
19(1):185–196, 1997. cardiac myocyte apoptosis - Role of a pertussis
1. Bers DM: Cardiac excitation-contraction cou- 22. Hulme JT, Lin TW, Westenbroek RE, et al: toxin-sensitive G proteins. Circulation
pling. Nature 415(6868):198–205, 2002. β-adrenergic regulation requires direct anchoring 100(22):2210–2212, 1999.
2. van der Heyden MA, Wijnhoven TJ, Opthof T: of PKA to cardiac Cav1.2 channels via a leucine 41. Zaugg M, Xu WM, Lucchinetti E, et al:
Molecular aspects of adrenergic modulation of zipper interaction with A kinase-anchoring β-Adrenergic receptor subtypes differentially
cardiac L-type Ca2+ channels. Cardiovasc Res protein 15. Proc Natl Acad Sci U S A 100(22): affect apoptosis in adult rat ventricular myocytes.
65(1):28–39, 2005. 13093–13098, 2003. Circulation 102(3):344–350, 2000.
3. Hartzell HC, Méry P-F, Fischmeister R, et al: 23. Hulme JT, Westenbroek RE, Scheuer T, et al: 42. Brodde OE, Bruck H, Leineweber K: Cardiac
Sympathetic regulation of cardiac calcium current Phosphorylation of serine 1928 in the distal adrenoceptors: Physiological and pathophysio-
is due exclusively to cAMP-dependent phosphor- C-terminal domain of cardiac Cav1.2 channels logical relevance. J Pharmacol Sci 100:323–337,
ylation. Nature 351(6327):573–576, 1991. during β1-adrenergic regulation. Proc Natl Acad 2006.
4. Catterall WA: Voltage-gated calcium channels. Sci U S A 103(44):16574–16579, 2006. 43. Chen-Izu Y, Xiao RP, Izu LT, et al: Gi-dependent
Cold Spring Harb Perspect Biol 3(8):a003947, 24. Hulme JT, Yarov-Yarovoy V, Lin TW, et al: Auto- localization of β2-adrenergic receptor signaling to
2011. inhibitory control of the Cav1.2 channel by its L-type Ca2+ channels. Biophys J 79(5):2547–2556,
5. Hess P, Lansman JB, Tsien RW: Different modes proteolytically processed distal C-terminal 2000.
of Ca channel gating behaviour favoured by dihy- domain. J Physiol 576(Pt 1):87–102, 2006. 44. Xiang Y, Rybin VO, Steinberg SF, et al: Caveolar
dropyridine Ca agonists and antagonists. Nature 25. Ganesan AN, Maack C, Johns DC, et al: localization dictates physiologic signaling of β2-
311(5986):538–544, 1984. β-Adrenergic stimulation of L-type Ca2+ channels adrenoceptors in neonatal cardiac myocytes.
6. Yue DT, Herzig S, Marban E: β-adrenergic stim- in cardiac myocytes requires the distal carboxyl J Biol Chem 277(37):34280–34286, 2002.
ulation of calcium channels occurs by potentiation terminus of alpha1C but not serine 1928. Circ Res 45. Calaghan S, Kozera L, White E: Compartmen-
of high-activity gating modes. Proc Natl Acad Sci 98(2):e11–8, 2006. talisation of cAMP-dependent signalling by cave-
U S A 87(2):753–757, 1990. 26. Lemke T, Welling A, Christel CJ, et al: Unchanged olae in the adult cardiac myocyte. J Mol Cell
7. Cens T, Rousset M, Leyris JP, et al: Voltage- and β-adrenergic stimulation of cardiac L-type Cardiol 45(1):88–92, 2008.
calcium-dependent inactivation in high voltage- calcium channels in Cav1.2 phosphorylation site 46. Nikolaev VO, Moshkov A, Lyon AR, et al: β2-
gated Ca2+ channels. Prog Biophys Mol Biol 90(1- S1928A mutant mice. J Biol Chem 283(50):34738– Adrenergic receptor redistribution in heart failure
3):104–117, 2006. 34744, 2008. changes cAMP compartmentation. Science
8. Argibay JA, Fischmeister R, Hartzell HC: Inacti- 27. Gerhardstein BL, Puri TS, Chien AJ, et al: Iden- 327(5973):1653–1657, 2010.
vation, reactivation and pacing dependence of tification of the sites phosphorylated by cyclic 47. Balijepalli RC, Foell JD, Hall DD, et al: Localiza-
calcium current in frog cardiocytes: correlation AMP-dependent protein kinase on the β2 subunit tion of cardiac L-type Ca2+ channels to a caveolar
with current density. J Physiol 401:201–226, of L-type voltage-dependent calcium channels. macromolecular signaling complex is required
1988. Biochemistry (Mosc) 38(32):10361–10370, for β2-adrenergic regulation. Proc Natl Acad Sci
9. Findlay I: Physiological modulation of inactiva- 1999. U S A 103(19):7500–7505, 2006.
tion in L-type Ca2+ channels: one switch. J Physiol 28. Bunemann M, Gerhardstein BL, Gao TY, et al: 48. Xiao RP, Lakatta EG: β1-Adrenoceptor stimula-
554(Pt 2):275–283, 2004. Functional regulation of L-type calcium channels tion and β2-adrenoceptor stimulation differ in
10. Davies A, Hendrich J, Van Minh AT, et al: Func- via protein kinase A-mediated phosphorylation of their effects on contraction, cytosolic Ca2+, and
tional biology of the alpha(2)delta subunits of the β2 subunit. J Biol Chem 274(48):33851–33854, Ca2+ current in single rat ventricular cells. Circ
voltage-gated calcium channels. Trends Pharma- 1999. Res 73:286–300, 1993.
col Sci 28(5):220–228, 2007. 29. Haase H, Alvarez J, Petzhold D, et al: Ahnak is 49. Nikolaev VO, Bunemann M, Schmitteckert E,
11. Dolphin AC: Beta subunits of voltage-gated critical for cardiac CaV1.2 calcium channel func- et al: Cyclic AMP imaging in adult cardiac
calcium channels. J Bioenerg Biomembr tion and its β-adrenergic regulation. FASEB J myocytes reveals far-reaching β1-adrenergic but
35(6):599–620, 2003. 19(14):1969–1977, 2005. locally confined β2-adrenergic receptor-mediated
12. Colecraft HM, Alseikhan B, Takahashi SX, et al: 30. Fuller MD, Emrick MA, Sadilek M, et al: Molec- signaling. Circ Res 99(10):1084–1091, 2006.
Novel functional properties of Ca2+ channel β ular mechanism of calcium channel regulation in 50. Rozec B, Gauthier C: β3-Adrenoceptors in the
subunits revealed by their expression in adult rat the fight-or-flight response. Sci Signal 3(141):ra70, cardiovascular system: Putative roles in human
heart cells. J Physiol 541(2):435–452, 2002. 2010. pathologies. Pharmacol Ther 111:652–673, 2006.
13. Fuller-Bicer GA, Varadi G, Koch SE, et al: Tar- 31. Hulme JT, Konoki K, Lin TW, et al: Sites of 51. Gauthier C, Tavernier G, Charpentier F, et al:
geted disruption of the voltage-dependent proteolytic processing and noncovalent associa- Functional β3-adrenoceptor in the human heart.
calcium channel alpha2/delta-1-subunit. Am J tion of the distal C-terminal domain of Cav1.1 J Clin Invest 98:556–562, 1996.
Physiol Heart Circ Physiol 297(1):H117–H124, channels in skeletal muscle. Proc Natl Acad Sci 52. Moniotte S, Vaerman J, Kockx MM, et al: Real-
2009. U S A 102( 14):5274–5279, 2005. time RT-PCR for the detection of beta-
14. Miriyala J, Nguyen T, Yue DT, et al: Role of Cavβ 32. Fu Y, Westenbroek RE, Yu FH, et al: Deletion of adrenoceptor messenger RNAs in small human
subunits, and lack of functional reserve, in protein the distal C terminus of Cav1.2 channels leads to endomyocardial biopsies. J Mol Cell Cardiol
kinase A modulation of cardiac Cav1.2 channels. loss of β-adrenergic regulation and heart failure 33(12):2121–2133, 2001.
Circ Res 102(7):e54–64, 2008. in vivo. J Biol Chem 286(14):12617–12626, 53. Chamberlain PD, Jennings KH, Paul F, et al: The
15. Yang L, Katchman A, Morrow JP, et al: Cardiac 2011. tissue distribution of the human β3-adrenoceptor
L-type calcium channel (Cav1.2) associates with 33. Xiang YK: Compartmentalization of beta- studied using a monoclonal antibody: direct evi-
gamma subunits. FASEB J 25(3):928–936, 2011. adrenergic signals in cardiomyocytes. Circ Res dence of the β3-adrenoceptor in human adipose
16. Puri TS, Gerhardstein BL, Zhao XL, et al: Dif- 109(2):231–244, 2011. tissue, atrium and skeletal muscle. Int J Obes
ferential effects of subunit interactions on protein 34. Rochais F, Abi-Gerges A, Horner K, et al: A spe- Relat Metab Disord 23(10):1057–1065, 1999.
kinase A- and C-mediated phosphorylation of cific pattern of phosphodiesterases controls the 54. Moniotte S, Kobzik L, Feron O, et al: Upregula-
L-type calcium channels. Biochemistry (Mosc) cAMP signals generated by different Gs-coupled tion of β3-adrenoceptors and altered contractile
36(31):9605–9615, 1997. receptors in adult rat ventricular myocytes. Circ response to inotropic amines in human failing
17. De Jongh KS, Murphy BJ, Colvin AA, et al: Spe- Res 98(8):1081–1088, 2006. myocardium. Circulation 103(12):1649–1655,
cific phosphorylation of a site in the full-length 35. Fischmeister R, Castro LRV, Abi-Gerges A, et al: 2001.
form of the alpha 1 subunit of the cardiac L-type Compartmentation of cyclic nucleotide signaling 55. De Matteis R, Arch JR, Petroni ML, et al: Immu-
calcium channel by adenosine 3’,5’-cyclic in the heart: The role of cyclic nucleotide phos- nohistochemical identification of the β3-
monophosphate-dependent protein kinase. Bio- phodiesterases. Circ Res 99(8):816–828, 2006. adrenoceptor in intact human adipocytes and
chemistry (Mosc) 35(32):10392–10402, 1996. 36. Perino A, Ghigo A, Scott JD, et al: Anchoring ventricular myocardium: effect of obesity and
18. Perez-Reyes E, Yuan WL, Wei XY, et al: Regula- proteins as regulators of signaling pathways. Circ treatment with ephedrine and caffeine. Int J Obes
tion of the cloned L-type cardiac calcium channel Res 111(4):482–492, 2012. Relat Metab Disord 26(11):1442–1450, 2002.
by cyclic-AMP-dependent protein kinase. FEBS 37. Diviani D, Maric D, Lopez IP, et al: A-kinase 56. Skeberdis VA, Gendvilienë V, Zablockaitë D, et
Lett 342(2):119–123, 1994. anchoring proteins: Molecular regulators of the al: β3-adrenergic receptor activation increases
19. Zong XG, Schreieck J, Mehrke G, et al: On the cardiac stress response. Biochim Biophys Acta human atrial tissue contractility and stimulates
regulation of the expressed L-type calcium 2012. the L-type Ca2+ current. J Clin Invest 118:3219–
channel by cAMP-dependent phosphorylation. 38. Morisco C, Zebrowski DC, Vatner DE, et al: 3227, 2008.
Pflügers Arch 430(3):340–347, 1995. β-Adrenergic cardiac hypertrophy is mediated 57. Schoemaker RG, Du XY, Bax WA, et al:
20. Mikala G, Klockner U, Varadi M, et al: cAMP- primarily by the β1-subtype in the rat heart. J Mol 5-Hydroxytryptamine stimulates human isolated
dependent phosphorylation sites and macroscopic Cell Cardiol 33(3):561–573, 2001. atrium but not ventricle. Eur J Pharmacol
activity of recombinant cardiac L-type calcium 39. Ahmet I, Krawczyk M, Heller P, et al: Beneficial 230:103–105, 1993.
channels. Mol Cell Biochem 185(1-2):95–109, effects of chronic pharmacological manipulation 58. Jahnel U, Rupp J, Ertl R, et al: Positive inotropic
1998. of β-adrenoreceptor subtype signaling in rodent response to 5-HT in human atrial but not in ven-
21. Gao TY, Yatani A, Dell’Acqua ML, et al: cAMP- dilated ischemic cardiomyopathy. Circulation tricular heart muscle. Naunyn Schmiedebergs
dependent regulation of cardiac L-type Ca2+ 110(9):1083–1090, 2004. Arch Pharmacol 346:482–485, 1992.
59. Ouadid H, Seguin J, Dumuis A, et al: Serotonin 79. Dittrich M, Jurevicius J, Georget M, et al: Local effects of milrinone withdrawal. Circulation 73(3
increases calcium current in human atrial myo- response of L-type Ca2+ current to nitric oxide in Pt 2):III205–12, 1986.
cytes via the newly described 5-hydroxytryptamine4
receptors. Mol Pharmacol 41:346–351, 1992.
60. Scott JD, Santana LF: A-kinase anchoring pro-
frog ventricular myocytes. J Physiol 534(1):109–
121, 2001.
80. Kirstein M, Rivet-Bastide M, Hatem S, et al:
97. Molina CE, Leroy J, Xie M, et al: Cyclic AMP
phosphodiesterase type 4 protects against atrial
arrhythmias. J Am Coll Cardiol 59(24):2182–
37
teins: getting to the heart of the matter. Circula- Nitric oxide regulates the calcium current in iso- 2190, 2012.
tion 121(10):1264–1271, 2010. lated human atrial myocytes. J Clin Invest 98. Rapundalo ST, Solaro RJ, Kranias EG: Inotropic
61. Kritzer MD, Li J, Dodge-Kafka K, et al: AKAPs: 95(2):794–802, 1995. responses to isoproterenol and phosphodiesterase
The architectural underpinnings of local cAMP 81. Richter W, Xie M, Scheitrum C, et al: Conserved inhibitors in intact guinea pig hearts: comparison
signaling. J Mol Cell Cardiol 2011. expression and functions of PDE4 in rodent and of cyclic AMP levels and phosphorylation of sar-
62. Jones BW, Brunet S, Gilbert ML, et al: Cardio- human heart. Basic Res Cardiol 106(2):249–262, coplasmic reticulum and myofibrillar proteins.
myocytes from AKAP7 knockout mice respond 2011. Circ Res 64:104–111, 1989.
normally to adrenergic stimulation. Proc Natl 82. Mongillo M, McSorley T, Evellin S, et al: Fluo- 99. Hohl CM, Li Q: Compartmentation of cAMP in
Acad Sci U S A 109(42):17099–17104, 2012. rescence resonance energy transfer-based analysis adult canine ventricular myocytes—Relation to
63. Nichols CB, Rossow CF, Navedo MF, et al: Sym- of cAMP dynamics in live neonatal rat cardiac single-cell free Ca2+ transients. Circ Res 69:1369–
pathetic stimulation of adult cardiomyocytes myocytes reveals distinct functions of compart- 1379, 1991.
requires association of AKAP5 with a subpopula- mentalized phosphodiesterases. Circ Res 100. Jurevicius J, Fischmeister R: cAMP compartmen-
tion of L-type calcium channels. Circ Res 95(1):65–75, 2004. tation is responsible for a local activation of
107(6):747–756, 2010. 83. Rochais F, Vandecasteele G, Lefebvre F, et al: cardiac Ca2+ channels by β-adrenergic agonists.
64. Marx SO, Reiken S, Hisamatsu Y, et al: PKA Negative feedback exerted by PKA and cAMP Proc Natl Acad Sci U S A 93(1):295–299, 1996.
phosphorylation dissociates FKBP12.6 from the phosphodiesterase on subsarcolemmal cAMP 101. Zaccolo M, Pozzan T: Discrete microdomains
calcium release channel (Ryanodine receptor): signals in intact cardiac myocytes. An in vivo study with high concentration of cAMP in stimulated
Defective regulation in failing hearts. Cell using adenovirus-mediated expression of CNG rat neonatal cardiac myocytes. Science
101(4):365–376, 2000. channels. J Biol Chem 279(50):52095–52105, 295(5560):1711–1715, 2002.
65. Marx SO, Kurokawa J, Reiken S, et al: Require- 2004. 102. Nikolaev VO, Lohse MJ: Monitoring of cAMP
ment of a macromolecular signaling complex for 84. Leroy J, Abi-Gerges A, Nikolaev VO, et al: Spa- synthesis and degradation in living cells. Physiol-
β adrenergic receptor modulation of the KCNQ1- tiotemporal dynamics of β-adrenergic cAMP ogy (Bethesda) 21:86–92, 2006.
KCNE1 potassium channel. Science 295(5554): signals and L-type Ca2+ channel regulation in 103. Ono K, Fozzard HA: Phosphorylation restores
496–499, 2002. adult rat ventricular myocytes: Role of phospho- activity of L-type calcium channels after rundown
66. Dodge-Kafka KL, Soughayer J, Pare GC, et al: diesterases. Circ Res 102(9):1091–1100, 2008. in inside-out patches from rabbit cardiac cells.
The protein kinase A anchoring protein mAKAP 85. Baillie GS, Sood A, McPhee I, et al: β-Arrestin- J Physiol 454:673–688, 1992.
co-ordinates two integrated cAMP effector path- mediated PDE4 cAMP phosphodiesterase 104. Herzig S, Neumann J: Effects of serine/threonine
ways. Nature 437:574–578, 2005. recruitment regulates β-adrenoceptor switching protein phosphatases on ion channels in excitable
67. Efendiev R, Samelson BK, Nguyen BT, et al: from Gs to G i. Proc Natl Acad Sci U S A membranes. Physiol Rev 80(1):173–210, 2000.
AKAP79 interacts with multiple adenylyl cyclase 100(3):941–945, 2003. 105. Davare MA, Horne MC, Hell JW: Protein phos-
(AC) isoforms and scaffolds AC 5 and 6 to AMPA 86. Richter W, Day P, Agraval R, et al: Signaling from phatase 2A is associated with class C L-type
receptors. J Biol Chem 285(19):14450–14458, β1- and β2-adrenergic receptors is defined by calcium channels (CaV1.2) and antagonizes
2010. differential interactions with PDE4. Embo J channel phosphorylation by cAMP-dependent
68. Bauman AL, Soughayer J, Nguyen BT, et al: 27(2):384–393, 2008. protein kinase. J Biol Chem 275(50):39710–
Dynamic regulation of cAMP synthesis through 87. De Arcangelis V, Liu R, Soto D, et al: Differential 39717, 2000.
anchored PKA-adenylyl cyclase V/VI complexes. association of phosphodiesterase 4D isoforms 106. Tandan S, Wang Y, Wang TT, et al: Physical and
Mol Cell 23(6):925–931, 2006. with β2-adrenoceptor in cardiac myocytes. J Biol functional interaction between calcineurin and
69. Dodge KL, Khouangsathiene S, Kapiloff MS, Chem 284(49):33824–33832, 2009. the cardiac L-type Ca2+ channel. Circ Res
et al: mAKAP assembles a protein kinase A/PDE4 88. Lehnart SE, Wehrens XHT, Reiken S, et al: Phos- 105(1):51–60, 2009.
phosphodiesterase cAMP signaling module. phodiesterase 4D deficiency in the ryanodine 107. Hall DD, Feekes JA, Arachchige Don AS, et al:
EMBO J 20(8):1921–1930, 2001. receptor complex promotes heart failure and Binding of protein phosphatase 2A to the L-type
70. Perino A, Ghigo A, Ferrero E, et al: Integrating arrhythmias. Cell 123(1):23–35, 2005. calcium channel Cav1.2 next to Ser1928, its main
cardiac PIP3 and cAMP signaling through a PKA 89. Kerfant BG, Zhao D, Lorenzen-Schmidt I, et al: PKA site, is critical for Ser1928 dephosphoryla-
anchoring function of p110gamma. Mol Cell PI3KY is required for PDE4, not PDE3, activity tion. Biochemistry (Mosc) 45(10):3448–3459,
42(1):84–95, 2011. in subcellular microdomains containing the sarco- 2006.
71. Ghigo A, Perino A, Mehel H, et al: PI3Kγ Pro- plasmic reticular calcium ATPase in cardiomyo- 108. Xu H, Ginsburg KS, Hall DD, et al: Targeting of
tects against catecholamine-induced ventricular cytes. Circ Res 101(4):400–408, 2007. protein phosphatases PP2A and PP2B to the
arrhythmia through PKA-mediated regulation of 90. Beca S, Helli PB, Simpson JA, et al: Phosphodi- C-terminus of the L-type calcium channel Ca
distinct phosphodiesterases. Circulation PMID: esterase 4D regulates baseline sarcoplasmic retic- v1.2. Biochemistry (Mosc) 49(48):10298–10307,
23008439 (in press), 2012. ulum Ca2+ release and cardiac contractility, 2010.
72. Terrenoire C, Houslay MD, Baillie GS, et al: independently of L-type Ca2+ current. Circ Res 109. Santana LF, Chase EG, Votaw VS, et al: Func-
The cardiac IKs potassium channel macromo 109(9):1024–1030, 2011. tional coupling of calcineurin and protein kinase
lecular complex includes the phosphodiesterase 91. Leroy J, Richter W, Mika D, et al: Phosphodies- A in mouse ventricular myocytes. J Physiol 544(Pt
PDE4D3. J Biol Chem 284(14):9140–9146, 2009. terase 4B in the cardiac L-type Ca2+ channel 1):57–69, 2002.
73. Rich TC, Fagan KA, Nakata H, et al: Cyclic complex regulates Ca2+ current and protects 110. Coumel P, Leenhardt A: Mental activity, adrener-
nucleotide-gated channels colocalize with adeny- against ventricular arrhythmias. J Clin Invest gic modulation, and cardiac arrhythmias in
lyl cyclase in regions of restricted cAMP diffusion. 121(7):2651–2661, 2011. patients with heart disease. Circulation 83(4
J Gen Physiol 116(2):147–161, 2000. 92. Weishaar RE, Kobylarz-Singer DC, Steffen RP, Suppl):II58–70, 1991.
74. Mika D, Leroy J, Vandecasteele G, et al: PDEs et al: Subclasses of cyclic AMP-specific phospho- 111. Weiss JN, Nivala M, Garfinkel A, et al: Alternans
create local domains of cAMP signaling. J Mol diesterase in left ventricular muscle and their and arrhythmias: from cell to heart. Circ Res
Cell Cardiol 52:323–329, 2012. involvement in regulating myocardial contractil- 108(1):98–112, 2011.
75. Méry P-F, Pavoine C, Pecker F, et al: Erythro-9- ity. Circ Res 61:539–547, 1987. 112. Eisner DA, Li Y, O’Neill SC: Alternans of
(2-hydroxy-3-nonyl)adenine inhibits cyclic 93. Lugnier C, Muller B, Lebec A, et al: Character- intracellular calcium: mechanism and signifi-
GMP-stimulated phosphodiesterase in isolated ization of indolidan-sensitive and rolipram- cance. Heart Rhythm 3(6):743–745, 2006.
cardiac myocytes. Mol Pharmacol 48(1):121–130, sensitive cyclic nucleotide phosphodiesterases 113. Wakili R, Voigt N, Kaab S, et al: Recent advances
1995. in canine and human cardiac microsomal frac- in the molecular pathophysiology of atrial
76. Rivet-Bastide M, Vandecasteele G, Hatem S, et tions. J Pharmacol Exp Ther 265:1142–1151, fibrillation. J Clin Invest 121(8):2955–2968,
al: cGMP-stimulated cyclic nucleotide phospho- 1993. 2011.
diesterase regulates the basal calcium current in 94. Smith CJ, Huang R, Sun D, et al: Development 114. Pogwizd SM, Bers DM: Cellular basis of trig-
human atrial myocytes. J Clin Invest 99(11):2710– of decompensated dilated cardiomyopathy is asso- gered arrhythmias in heart failure. Trends Car-
2718, 1997. ciated with decreased gene expression and activity diovasc Med 14(2):61–66, 2004.
77. Verde I, Vandecasteele G, Lezoualc’h F, et al: of the milrinone-sensitive cAMP phosphodiester- 115. January CT, Riddle JM: Early afterdepolariza-
Characterization of the cyclic nucleotide phos- ase PDE3A. Circulation 96(9):3116–3123, tions—Mechanism of induction and block—A
phodiesterase subtypes involved in the regulation 1997. role for L-type Ca2+ current. Circ Res 64(5):977–
of the L-type Ca2+ current in rat ventricular myo- 95. Osadchii OE: Myocardial phosphodiesterases and 990, 1989.
cytes. Br J Pharmacol 127(1):65–74, 1999. regulation of cardiac contractility in health and 116. Volders PGA, Kulcsar A, Vos MA, et al: Similari-
78. Vandecasteele G, Verde I, Rucker-Martin C, et al: cardiac disease. Cardiovasc Drugs Ther 21:171– ties between early and delayed afterdepolariza-
Cyclic GMP regulation of the L-type Ca2+ 194, 2007. tions induced by isoproterenol in canine
channel current in human atrial myocytes. 96. Monrad ES, Baim DS, Smith HS, et al: Assess- ventricular myocytes. Cardiovasc Res 34(2):348–
J Physiol 533(2):329–340, 2001. ment of long-term therapy with milrinone and the 359, 1997.
117. Pereira L, Métrich M, Fernández-Velasco M, L-type calcium channels, in. Ann N Y Acad Sci arrhythmia in patients with arrhythmogenic right
et al: The cAMP binding protein Epac modulates Feb(1188):87–95, 2010. ventricular cardiomyopathy. Heart Rhythm
Ca2+ sparks by Ca2+/calmodulin kinase signalling 136. Lu Y, Zhang Y, Wang N, et al: MicroRNA-328 9(6):961–967, 2012.
pathway in rat cardiac myocytes. J Physiol 583(Pt contributes to adverse electrical remodeling in 157. Gomez AM, Valdivia HH, Cheng H, et al: Defec-
2):685–694, 2007. atrial fibrillation. Circulation 122(23):2378–2387, tive excitation-contraction coupling in experi-
118. Mangmool S, Shukla AK, Rockman HA: 2010. mental cardiac hypertrophy and heart failure.
β-Arrestin-dependent activation of Ca2+/calmod- 137. Carnes CA, Janssen PM, Ruehr ML, et al: Atrial Science 276(5313):800–806, 1997.
ulin kinase II after β1-adrenergic receptor stimu- glutathione content, calcium current, and con- 158. Crossman DJ, Ruygrok PN, Soeller C, et al:
lation. J Cell Biol 189(3):573–587, 2010. tractility. J Biol Chem 282(38):28063–28073, Changes in the organization of excitation-
119. Wu Y, Temple J, Zhang R, et al: Calmodulin 2007. contraction coupling structures in failing human
kinase II and arrhythmias in a mouse model of 138. Gaur N, Rudy Y, Hool L: Contributions of ion heart. PLoS One 6(3):e17901, 2011.
cardiac hypertrophy. Circulation 106(10):1288– channel currents to ventricular action potential 159. Pitt GS, Dun W, Boyden PA: Remodeled cardiac
1293, 2002. changes and induction of early afterdepolariza- calcium channels. J Mol Cell Cardiol 41(3):373–
120. Koval OM, Guan X, Wu Y, et al: Cav1.2 beta- tions during acute hypoxia. Circ Res 105(12):1196– 388, 2006.
subunit coordinates CaMKII-triggered cardio- 1203, 2009. 160. Benitah JP, Alvarez JL, Gomez AM: L-type Ca2+
myocyte death and afterdepolarizations. Proc 139. Van Wagoner DR, Pond AL, Lamorgese M, et al: current in ventricular cardiomyocytes. J Mol Cell
Natl Acad Sci U S A 107(11):4996–5000, 2010. Atrial L-type Ca2+ currents and human atrial Cardiol 48(1):26–36, 2010.
121. Sipido KR, Volders PGA, de Groot SHM, et al: fibrillation. Circ Res 85(5):428–436, 1999. 161. Handrock R, Schroder F, Hirt S, et al: Single-
Enhanced Ca2+ release and Na/Ca exchange activ- 140. Gaspo R, Sun H, Fareh S, et al: Dihydropyridine channel properties of L-type calcium channels
ity in hypertrophied canine ventricular and beta adrenergic receptor binding in dogs with from failing human ventricle. Cardiovasc Res
myocytes—Potential link between contractile tachycardia-induced atrial fibrillation. Cardiovasc 37(2):445–455, 1998.
adaptation and arrhythmogenesis. Circulation Res 42(2):434–442, 1999. 162. Chen X, Zhang X, Harris DM, et al: Reduced
102( 17):2137–2144, 2000. 141. Parvez B, Chopra N, Rowan S, et al: A common effects of BAY K 8644 on L-type Ca2+ current in
122. Spencer CI, Sham JS: Effects of Na+/Ca2+ β1-adrenergic receptor polymorphism predicts failing human cardiac myocytes are related to
exchange induced by SR Ca2+ release on action favorable response to rate-control therapy in abnormal adrenergic regulation. Am J Physiol
potentials and afterdepolarizations in guinea pig atrial fibrillation. J Am Coll Cardiol 59(1):49–56, Heart Circ Physiol 294(5):H2257–H2267, 2008.
ventricular myocytes. Am J Physiol Heart Circ 2012. 163. Kashihara T, Nakada T, Shimojo H, et al: Chronic
Physiol 285(6):H2552–H2562, 2003. 142. El-Armouche A, Boknik P, Eschenhagen T, et al: receptor-mediated activation of Gi/o proteins
123. Priori SG, Chen SR: Inherited dysfunction of Molecular determinants of altered Ca2+ handling alters basal t-tubular and sarcolemmal L-type
sarcoplasmic reticulum Ca2+ handling and in human chronic atrial fibrillation. Circulation Ca2+ channel activity through phosphatases in
arrhythmogenesis. Circ Res 108(7):871–883, 114(7):670–680, 2006. heart failure. Am J Physiol Heart Circ Physiol
2011. 143. Wittköpper K, Fabritz L, Neef S, et al: Constitu- 302(8):H1645–H1654, 2012.
124. Napolitano C, Antzelevitch C: Phenotypical tively active phosphatase inhibitor-1 improves 164. Hullin R, Khan IFY, Wirtz S, et al: Cardiac
manifestations of mutations in the genes encoding cardiac contractility in young mice but is deleteri- L-type calcium channel beta-subunits expressed
subunits of the cardiac voltage-dependent L-type ous after catecholaminergic stress and with aging. in human heart have differential effects on single
calcium channel. Circ Res 108(5):607–618, J Clin Invest 120(2):617–626, 2010. channel characteristics. J Biol Chem
2011. 144. Lohse MJ, Engelhardt S, Eschenhagen T: What 278(24):21623–21630, 2003.
125. Sung RJ, Wu YH, Lai NH, et al: β-adrenergic is the role of β-adrenergic signaling in heart 165. Hullin R, Matthes J, von Vietinghoff S, et al:
modulation of arrhythmogenesis and identifica- failure? Circ Res 93(10):896–906, 2003. Increased expression of the auxiliary β2-subunit of
tion of targeted sites of antiarrhythmic therapy in 145. Movsesian MA, Bristow MR: Alterations in ventricular L-type Ca2+ channels leads to single-
Timothy (LQT8) syndrome: a theoretical study. cAMP-mediated signaling and their role in the channel activity characteristic of heart failure.
Am J Physiol Heart Circ Physiol 298(1):H33– pathophysiology of dilated cardiomyopathy. Curr PLoS One 2(3):e292, 2007.
H44, 2010. Top Dev Biol 68:25–48, 2005. 166. Hoch B, Meyer R, Hetzer P, et al: Identification
126. Workman AJ: Cardiac adrenergic control and 146. Engelhardt S, Hein L, Wiesmann F, et al: Pro- and expression of delta-isoforms of the multifunc-
atrial fibrillation. Naunyn Schmiedebergs Arch gressive hypertrophy and heart failure in beta1- tional Ca2+/calmodulin-dependent protein kinase
Pharmacol 381(3):235–249, 2010. adrenergic receptor transgenic mice. Proc Natl in failing and nonfailing human myocardium.
127. Greiser M, Lederer WJ, Schotten U: Alterations Acad Sci U S A 96(12):7059–7064, 1999. Circ Res 84(6):713–721, 1999.
of atrial Ca2+ handling as cause and consequence 147. Antos CL, Frey N, Marx SO, et al: Dilated car- 167. Zhang R, Khoo MS, Wu Y, et al: Calmodulin
of atrial fibrillation. Cardiovasc Res 89 (4):722– diomyopathy and sudden death resulting from kinase II inhibition protects against structural
733, 2011. constitutive activation of protein kinase A. Circ heart disease. Nat Med 11(4):409–417, 2005.
128. Legrand B, Hatem S, Deroubaix E, et al: Calcium Res 89(11):997–1004, 2001. 168. Wang Y, Tandan S, Cheng J, et al: Ca2+/
current depression in isolated human atrial myo- 148. Iwase M, Bishop SP, Uechi M, et al: Adverse calmodulin-dependent protein kinase
cytes after cessation of chronic treatment with effects of chronic endogenous sympathetic drive II-dependent remodeling of Ca2+ current in pres-
calcium antagonists. Circ Res 69(2):292–300, induced by cardiac Gs alpha overexpression. Circ sure overload heart failure. J Biol Chem
1991. Res 78(4):517–524, 1996. 283(37):25524–25532, 2008.
129. Van Wagoner DR, Pond AL, McCarthy PM, et 149. Lyon AR, Macleod KT, Zhang Y, et al: Loss of 169. Chen X, Piacentino VI, Furukawa S, et al: L-type
al: Outward K+ current densities and Kv1.5 T-tubules and other changes to surface topogra- Ca2+ channel density and regulation are altered in
expression are reduced in chronic human atrial phy in ventricular myocytes from failing human failing human ventricular myocytes and recover
fibrillation. Circ Res 80(6):772–781, 1997. and rat heart. Proc Natl Acad Sci U S A 106 after support with mechanical assist devices. Circ
130. Christ T, Boknik P, Wohrl S, et al: L-type Ca2+ (16):6854–6859, 2009. Res 91(6):517–524, 2002.
current downregulation in chronic human atrial 150. Orchard C, Brette F: T-tubules and sarcoplasmic 170. Neumann J, Schmitz W, Scholz H, et al: Increase
fibrillation is associated with increased activity of reticulum function in cardiac ventricular myo- in myocardial Gi-proteins in heart failure. Lancet
protein phosphatases. Circulation 110(17):2651– cytes. Cardiovasc Res 77(2):237–244, 2008. 2(8617):936–937, 1988.
2657, 2004. 151. Rengo G, Lymperopoulos A, Leosco D, et al: 171. Xiao RP, Zhang SJ, Chakir K, et al: Enhanced Gi
131. Nattel S, Maguy A, Le Bouter S, Yeh YH: GRK2 as a novel gene therapy target in signaling selectively negates β2-adrenergic recep-
Arrhythmogenic ion-channel remodeling in the heart failure. J Mol Cell Cardiol 50(5):785–792, tor (AR)- but not β1-AR-mediated positive inotro-
heart: heart failure, myocardial infarction, and 2011. pic effect in myocytes from failing rat hearts.
atrial fibrillation. Physiol Rev 87(2):425–456, 152. Brette F, Salle L, Orchard CH: Differential mod- Circulation 108(13):1633–1639, 2003.
2007. ulation of L-type Ca2+ current by SR Ca2+ release 172. El-Armouche A, Zolk O, Rau T, et al: Inhibitory
132. Dinanian S, Boixel C, Juin C, et al: Downregula- at the T-tubules and surface membrane of rat ven- G-proteins and their role in desensitization of the
tion of the calcium current in human right atrial tricular myocytes. Circ Res 95(1):e1–7, 2004. adenylyl cyclase pathway in heart failure. Cardio-
myocytes from patients in sinus rhythm but with 153. Hong TT, Smyth JW, Gao D, et al: BIN1 local- vasc Res 60(3):478–487, 2003.
a high risk of atrial fibrillation. Eur Heart J izes the L-type calcium channel to cardiac 173. He JQ, Balijepalli RC, Haworth RA, et al: Cross-
29(9):1190–1197, 2008. T-tubules. PLoS Biol 8(2):e1000312, 2010. talk of beta-adrenergic receptor subtypes through
133. Qi XY, Yeh YH, Xiao L, et al: Cellular signaling 154. He JQ, Conklin MW, Foell JD, et al: Reduction Gi blunts beta-adrenergic stimulation of L-type
underlying atrial tachycardia remodeling of in density of transverse tubules and L-type Ca2+ Ca2+ channels in canine heart failure. Circ Res
L-type calcium current. Circ Res 103(8):845–854, channels in canine tachycardia-induced heart 97(6):566–573, 2005.
2008. failure. Cardiovasc Res 49(2):298–307, 2001. 174. Raake PW, Vinge LE, Gao E, et al: G protein-
134. Bosch RF, Scherer CR, Rub N, et al: Molecular 155. Horiuchi-Hirose M, Kashihara T, Nakada T, coupled receptor kinase 2 ablation in cardiac
mechanisms of early electrical remodeling: tran- et al: Decrease in the density of t-tubular L-type myocytes before or after myocardial infarction
scriptional downregulation of ion channel sub- Ca2+ channel currents in failing ventricular myo- prevents heart failure. Circ Res 103(4):413–422,
units reduces ICa,L and Ito in rapid atrial pacing in cytes. Am J Physiol Heart Circ Physiol 2008.
rabbits. J Am Coll Cardiol 41(5):858–869, 2003. 300(3):H978–H988, 2011. 175. Volkers M, Weidenhammer C, Herzog N,
135. Beharier O, Etzion Y, Levi S, et al: The involve- 156. Hong TT, Cogswell R, James CA, et al: Plasma et al: The inotropic peptide βARKct improves
ment of ZnT-1, a new modulator of cardiac BIN1 correlates with heart failure and predicts βAR responsiveness in normal and failing
cardiomyocytes through GβY-mediated L-type preceded by a hypercontractile state and diastolic cal effects in models of cardiac hypertrophy. Circ
calcium current disinhibition. Circ Res 108(1):27– failure in the alpha1CTG mouse. PLoS One Res 101(2):166–175, 2007.
39, 2011.
176. Abi-Gerges A, Richter W, Lefebvre F, et al:
Decreased expression and activity of cAMP phos-
4(1):e4133, 2009.
179. Chen X, Nakayama H, Zhang X, et al: Calcium
influx through Cav1.2 is a proximal signal for
182. Goonasekera SA, Hammer K, Auger-Messier M,
et al: Decreased cardiac L-type Ca2+ channel
activity induces hypertrophy and heart failure in
37
phodiesterases in cardiac hypertrophy and its pathological cardiomyocyte hypertrophy. J Mol mice. J Clin Invest 122(1):280–290, 2012.
impact on β-adrenergic cAMP signals. Circ Res Cell Cardiol 50(3):460–470, 2011. 183. Schroder E, Byse M, Satin J: L-type calcium
105(8):784–792, 2009. 180. Semsarian C, Ahmad I, Giewat M, et al: The channel C terminus autoregulates transcription.
177. Best JM, Kamp TJ: Different subcellular popula- L-type calcium channel inhibitor diltiazem pre- Circ Res 104(12):1373–1381, 2009.
tions of L-type Ca2+ channels exhibit unique regu- vents cardiomyopathy in a mouse model. J Clin 184. Domes K, Ding J, Lemke T, et al: Truncation of
lation and functional roles in cardiomyocytes. Invest 109(8):1013–1020, 2002. murine Cav1.2 at Asp-1904 results in heart failure
J Mol Cell Cardiol 52(2):376–387, 2012. 181. Cingolani E, Ramirez Correa GA, et al: Gene after birth. J Biol Chem 286(39):33863–33871,
178. Wang S, Ziman B, Bodi I, et al: Dilated cardio- therapy to inhibit the calcium channel β subunit: 2011.
myopathy with increased SR Ca2+ loading physiological consequences and pathophysiologi-
Cholinergic and
Constitutive Regulation
of Atrial Potassium Channel 38
Dobromir Dobrev, Niels Voigt, and Stanley Nattel
coupled to Gi-proteins) can activate cardiac IK,ACh channels. Based

CHAPTER OUTLINE
on their regulation and biophysics, IK,ACh channels are also des-
G-Protein Cycle May Contribute to the Generation of ignated G-protein–activated inwardly rectifying K+ (GIRK)
Constitutive IK,ACh 387 channels. In contrast to ventricular tissue, IK,ACh channels are
highly expressed in the atria, including SA and atrioventricular
Regulation of IK,ACh Gating by PIP2 388 nodes, where they contribute to vagal regulation of heart rate
Constitutively Active IK,ACh 388 (negative chronotropic effect) and conduction from the atria to
the ventricles (negative dromotropic effects).10-12
Phosphorylation-Dependent IK,ACh Regulation 388
IK,ACh Channels as Potential Therapeutic Targets
in Atrial Fibrillation 389 M2-Receptor–Dependent IK,ACh Facilitates Initiation
and Maintenance of Atrial Fibrillation
Stimulation of the vagal nerve, the principal cardiac arm of the In atrial myocytes, IK,ACh channel activation leads to hyperpolar-
parasympathetic nervous system, reduces heart rate and slows ization of the resting membrane potential and shortening of the
conduction in the atrioventricular node, thereby tuning the heart action-potential duration (APD).13 ACh-induced APD shorten-
rate to “rest-and-digest” activities. In 1921, Otto Loewi found ing and resting membrane potential hyperpolarization create an
that the vagal effects on the heart are mediated by release of arrhythmogenic substrate facilitating the induction of atrial
acetylcholine (ACh) from the parasympathetic synapses, and ACh fibrillation (AF; see also chapter 45 of this book).14 It is well
became the first neurotransmitter ever discovered.1 However, it known that vagal nerve stimulation promotes AF in animal
took more than 50 years until it was suggested that ACh activates models and patients by facilitating the initiation and maintenance
a specific population of K+ channels (ACh-gated IK,ACh; Figure of reentry circuits.15 Reentry, the most established basic mecha-
38-1, A) leading to hyperpolarization of the cell membrane, nism of AF, is described as continuous impulse propagation
thereby decreasing pacemaker activity in sinoatrial (SA) node around a functional barrier or an anatomical obstacle. An impor-
cells.2 It was also found that IK,ACh conductance is voltage depen- tant requirement for reentry is that the initially activated tissue
dent, with high K+ conductance at hyperpolarized membrane zone regains excitability while the electrical impulse propagates
potentials (at which the current is typically inward) and small around the reentry circuit, explaining why reduced effective
conductance at depolarized membrane potentials associated with refractory period or decreased conduction velocity can provide a
outward current. This typical current-voltage (IV) relationship substrate for AF maintenance. APD shortening induced by IK,ACh
designates IK,ACh as an inward-rectifier K+ current (see Figure activation reduces effective refractory period, thereby favoring
38-1, B and C) similar to IK1, which is active in the absence of any AF initiation by vagal stimulation. In knockout mice lacking the
receptor agonists, and IK,ATP, which is activated by reduced intra- Kir3.4 channel subunit, M-receptor stimulation does not induce
cellular ATP levels. Cardiac IK,ACh channels are heterotetramers, AF, clearly suggesting that the AF facilitating effects of vagal
usually consisting of two Kir3.1 and two Kir3.4 channel nerve activation are exclusively mediated by IK,ACh.16 In addition,
subunits.3 mathematical modeling studies suggest that increases in an
IK,ACh channels are activated by ACh binding to type-2 mus- inward-rectifier K+ current such as IK,ACh have a significant role
carinic receptors (M2-receptors), which causes dissociation of in AF promotion.17 Because of their ability to hyperpolarize atrial
inhibitory Gi-proteins, thereby increasing IK,ACh open probability cardiomyocytes and remove voltage-dependent Na+-current (INa)
via direct interaction of G-protein βγ-subunits with the channel inactivation, enhanced inward-rectifier K+ currents are more
(Figure 38-2, A, B).4,5 After current activation by M2-receptor effective in stabilizing and accelerating AF-sustaining reentry
stimulation, atrial IK,ACh shows a characteristic biphasic desensiti- circuits (rotors) than are changes in other ionic currents (e.g.,
zation that starts within a few seconds (Figure 38-3, middle). reduced L-type Ca2+ currents) that produce a similar degree of
Although both Kir3.1 and Kir3.4 channel homomers produce APD shortening. In vivo evidence has been obtained to support
similar peak-current densities upon M2-receptor stimulation the validity of these modeling data.18
when expressed in HEK-cells, IK,ACh desensitization was observed Interestingly, AF-related atrial remodeling is associated with
in Kir3.1 homomers only. Moreover, the desensitization process reduced maximum activation of IK,ACh upon M-receptor stimula-
depends on the presence of the Kir3.1 subunit within the channel tion,13,19-22 which might partly result from reduced expression
complex.6 Besides channel subunit composition, kinetics of levels of Kir3.1 and Kir3.4 subunits in AF patients.13,21 The
G-protein cycle, membrane content of the anionic phospholipid reduced agonist-dependent IK,ACh in AF patients could be a pro-
phosphatidylinositol-4,5-bisphosphate (PIP2), and phosphoryla- tective mechanism against the profibrillatory effects of vagal
tion processes also contribute IK,ACh regulation and desensitiza- nerve stimulation.
tion (see Figure 38-3). Left-to-right atrial gradients of inward-rectifier K+ currents
Apart from canonical M2-receptor–mediated activation of contribute to AF pathophysiology. There is clinical and experi-
IK,ACh, purinergic A1,7,8 and sphingolipid Edg-39 receptors (also mental evidence that certain cases of paroxysmal AF (pAF) and
383
Basal conditions Agonist dependent

IK,ACh activation
ACh
K+ IK1 K+ IK,ACh
M2 M2
β γ
αi β γ αi
GIRK-
GDP channel
GTP
A
Sinus rhythm
# # # # # # #
*
3 pA
0.1 s
Inward Outward Inward Outward
Basal –100 –80 –60 –40 –20 0 20 40 Basal –100 –80 –60 –40 –20 0 20 40
current VM (mV) current VM (mV)
10 pA/pF IK,ACh 10 pA/pF
B [K+]o = 20 mM CCh (2 µM) [K+]o = 20 mM
Atrial fibrillation
# # # # # # #
* * * * ** * * * * **
3 pA
Vc = –120 mV 0.1 s
–100 –80 –60 –40 –20 0 20 40 –100 –80 –60 –40 –20 0 20 40
Basal Basal
VM (mV) VM (mV)
current current
10 pA/pF IK,ACh 10 pA/pF

C [K+]o = 20 mM CCh (2 µM) [K+]o = 20 mM
Figure 38-1. Basal and muscarinic (M2)-receptor activated inward-rectifier potassium (K+) currents in sinus rhythm (SR) and chronic atrial fibrillation (cAF). A, Left panel,
Physiologically G-protein–activated inward-rectifier K+ (GIRK) channels are supposed to be closed in the absence of M-receptor agonists and only the basal inward-rectifier
potassium current IK1 is active. B and C, Representative single-channel (upper tracings) and whole-cell recordings (lower tracings) in patients with SR (B) and cAF (C). The
single-channel recordings were performed at –100 mV in a cell-attached configuration. Filled and empty arrow heads indicate closed and open levels of IK,ACh, respectively.
In the absence of M-receptor agonists (basal conditions), myocytes from the cAF patients (C) exhibit both IK1 (#) and constitutive IK,ACh openings (*), whereas the latter is a
rare event in myocytes from patients with SR (B). Besides increased expression of IK1 channels, constitutive IK,ACh,c activity can contribute to the increased whole-cell basal
inward-rectifier current in cAF (compare left lower tracings in B and C)
(Replotted with permission from Dobrev D, Friedrich A, Voigt N, et al: The G protein-gated potassium current IK,ACh is constitutively active in patients with chronic atrial fibrillation.
Circulation 112:3697–3706, 2005. Figure was produced using Servier Medical Art.)
chronic AF (cAF) are maintained by high-frequency reentrant these dominant frequency gradients.22 In a sheep model of ACh-
sources (rotors) with a consistent left-to-right dominant fre- mediated, pacing-induced AF, the left-to-right (LA-RA) domi-
quency gradient, particularly in pAF.23-25 Because increased nant frequency gradient parallels a left-to-right IK,ACh gradient,
inward-rectifier currents enhance rotor frequency, the left-to- supporting the hypothesis that an unequal LA-RA distribution of
right atrial gradient in the basal inward-rectifier current present inward-rectifier K+ currents can contribute to AF maintenance.26
in pAF, but not in cAF or sinus rhythm (SR), can contribute to However, in atria from patients with SR, there is a right-to-left
Cholinergic and Constitutive Regulation of Atrial Potassium Channel 385
38
R R
αi
αi β γ
β γ GDP
GDP
GTP
P
GDP
R NDPK
αi
A ADP ATP
GTP
β γ
K+ IK,ACh
β γ
β γ
NDPK B R
NDPK B
αi
GTP
R R R
NDPK B NDPK B NDPK B
αi αi ADP ATP αi
β γ β γ P β γ
GDP GDP
P H266 H118
C
Figure 38-2. Muscarinic (M2)-receptor and nucleoside diphosphate kinase (NDPK)-dependent activation of Gi-proteins. A, Agonist binding to M2-receptors triggers GDP/
GTP (guanosine-5′-triphosphate and guanosine-5′-diphosphate) exchange at the Gαi-subunit of the heterotrimeric Gi-proteins. Consecutive dissociation of the Gαi- and
Gβγ-subunits leads to activation of IK,ACh by binding of the Gβγ-subunit to the IK,ACh channel (B). Intrinsic GTPase activity of the Gαi-subunit hydrolyses GTP into GDP resulting
in reassociation of the heterotrimeric G-protein and restorage of the initial state. NDPKs provide the GTP, which is necessary for G-protein dissociation, by a phosphotransfer
from ATP to GDP. C, In the subpopulation of heterotrimeric Gi-proteins complexed with NDPK B, a phosphotransfer from ATP to His118 in NDPK-B, and subsequently onto
His266 of the Gβ, results in a high energetic phosphate, which promotes the formation of GTP and leads to receptor independent G-protein activation. This can contribute
to the development of receptor-independent constitutive IK,ACh activity.
(Figure was produced using Servier Medical Art.)

RGS αi
β γ
GDP
RGS
GTP
ACh ACh GDP P
K+ IK,ACh
M2 M1/3 K+ IK,ACh
β γ β γ R
αi αq Fast and
GTP Na+ GTP intermediate desensitization.
RGS αi β γ
CCh, 2 µM
Basal B GTP
GIRK 1/4
GIRK 1/4
current
GIRK 1/4
IP3
PLC
5 pA/pF
β γ
PIP2
N β γ 60 s IK,ACh
Na+
ACh
β γ C K+ IK,ACh
Na+ enhances PIP2 DAG
A binding at GIRK4 only
M2
PKG
GRKII β γ β γ PKC
αi P
P GTP PKA
CaMKII
PP2A
PP1
IK,ACh PP2A
C
Figure 38-3. Mechanisms contributing to IK,ACh desensitization. Upon muscarinic (M2)-receptor activation, IK,ACh is activated by direct binding of liberated Gβγ-subunits to
the IK,ACh channel, followed by a biphasic desensitization (fast and intermediate desensitization). A, The binding of Gβγ-subunits to the IK,ACh channel subunits Kir3.1 and
Kir3.4 strengthens their interaction with cell membrane–located phosphatidyl inositol 4,5-bisphosphate (PIP2), thereby increasing IK,ACh open probability. Similarly, increased
intracellular Na+ also strengthens the interaction of PIP2 with the channel leading to receptor-independent current activation. In contrast, stimulation of Gq-coupled M1/3-
receptors activates phospholipase-C (PLC), thereby lowering the PIP2 membrane content resulting in IK,ACh inhibition and contributing to fast IK,ACh desensitization. B, Regulator
of G-protein signaling proteins (RGS) accelerate the GTP hydrolysis rate of the Gαi-subunit, thereby contributing to faster desensitization and reducing agonist-independent
constitutive IK,ACh,c activity (see also Figure 38-2, A). C, The phosphorylation of muscarinic (M) receptors and IK,ACh channels is controlled by various kinases and phosphatases,
whereby channel phosphorylation increases IK,ACh, while concomitant phosphorylation of M-receptors reduces IK,ACh. Whereas channel dephosphorylation is supposed to
contribute to the fast phase (green), the intermediate phase (blue) involves progressive receptor phosphorylation by a G-protein–coupled receptor kinase (GRK), which
uncouples the receptor from the G-protein.
(Figure was produced using Servier Medical Art.)
atrial gradient of IK,ACh current that is absent in pAF and cAF.22 also Chapter 45). Therefore, agonist-independent IK,ACh,c is
The lack of RA-dominant agonist-activated IK,ACh could have a expected to increase atrial vulnerability to tachyarrhythmias and
permissive role for LA-dominant drivers in pAF and cAF, par- to promote persistence of AF. Accordingly, inhibition of IK,ACh
ticularly in vagal contexts. with the highly-selective IK,ACh blocker tertiapin reverses the APD
abbreviation and prevents the AF promotion in dogs with atrial
tachycardia remodeling (ATR) that also develop increased
Agonist-Independent, Constitutively Active IK,ACh IK,ACh,c.30 IK,ACh is almost absent in ventricles; therefore, it is a
Can Contribute to Atrial Fibrillation Maintenance promising atrial-selective anti-AF target that lacks proarrhythmic
side effects in the ventricles.31
It is well recognized that IK,ACh can possess agonist-independent In whole-cell patch-clamp experiments, electrophysiological
“resting” activity, with a much lower opening frequency than properties of IK,ACh (i.e., current-voltage relationship) are
agonist-induced IK,ACh.27 Constitutive IK,ACh current (IK,ACh,c) may comparable with other inward-rectifier currents such as IK1 or
underlie a major part of basal K+ conductance in SA node cells, IK,ATP; therefore, single-channel recordings are often used as a
which lack IK1, creating an important role in regulating heart direct index of increased constitutive IK,ACh,c activity in atrial
rate.28 In the normal heart, constitutive IK,ACh activity is low in myocytes from patients with cAF and dogs with ATR.19,32-34
atrial myocytes, but can increase substantially with cardiac Due to their short opening-times and a characteristic single-
pathology. Agonist-independent IK,ACh,c activity increases in atrial channel conductance of approximately 40 pS IK,ACh single-
myocytes from patients and animal models of AF, whereas channel openings are clearly different from openings of IK1 or
maximum M-receptor activation of IK,ACh is reduced (see Figure IK,ATP (Table 38-1). Figure 38-1 shows representative recordings
38-1, C).19,20,22,29 IK,ACh,c might contribute to APD shortening, of IK,ACh in the presence and absence of a muscarinic-receptor
which is a hallmark of the AF-related electrical remodeling14 (see agonist in atrial myocytes from patients with SR and cAF.
Table 38-1. Single-Channel Characteristics of Inward-Rectifier K+ Currents in Human Atrial Myocytes: Acetylcholine/Carbachol (10 µM) Activated
IK,ACh in Patients with Sinus Rhythm
GS (pS) Mean open time (ms) Mean closed time (ms) Popen (%) References
38
42.0 ± 1.2 1.7 ± 0.5 1.4 ± 0.2 2.7 ± 0.4 35
18.0 ± 4.4 30.0 ± 7.6
46.0 ± 2.0 τopen (ms) = 2.5 ± 0.2 τclose = 4.2 ± 0.2 – 36
49.4 ± 3.6 τopen (ms) = 3.9 ± 0.1 31.6 ± 1.5 19
Constitutively active IK,ACh in cAF
36.3 ± 2.5 2.4 ± 0.2 5.4 ± 0.7 19, 32
IK1 in SR patients
27.0 ± 1.4 8.7 ± 2.7 0.7 ± 0.1 75.0 ± 23.0 35

66.0 ± 32.0
27.0 ± 2.0 τopen (ms) = 30.2 ± 4.0 – 56.0 ± 6.0 37
29.4 ± 1.9 τopen (ms) = 7.8 ± 0.3 – 11.4 ± 0.7 19
IK,ATP in SR patients
73.0 ± 2.5 1.4 ± 0.2 0.3 ± 0.1 – 35
GS, single-channel conductance; Popen, channel open probability; τopen/τclose, time constant for open and closed time distributions.
In pico siemens (pS)/S = 1 V/A.
Inclusion of the nonselective M-receptor agonist carbachol the initial inactive state.39 Gα-subunits are subdivided into Gαs-
(10 µM) in the pipette solution strongly activated IK,ACh in both subunits, which stimulate adenylate cyclases, Gαi-subunits, which
groups, causing frequent channel openings. In the absence of inhibit adenylate cyclases, and Gαq-subunits, which activate
M-receptor agonists, constitutive IK,ACh,c openings are apparent phospholipase-C (see later). IK,ACh channels are activated by direct
in cAF, whereas they occur only sporadically in myocytes from binding of βγ-subunits originating from Gi-protein.4,38 It is unclear
SR patients.19 whether Gβγ-subunits originating from Gs- or Gq-proteins also
In dogs subjected to AF mimicking, atrial tachycardia remod- activate IK,ACh channels under physiological conditions.
eling IK,ACh also develops agonist-independent constitutive IK,ACh,c Although Gαi-subunits do not directly activate IK,ACh channels,
activity, suggesting that the development of IK,ACh,c in patients they certainly have an important role in IK,ACh regulation. Under
with cAF can result from the high atrial rate rather than from the resting conditions, the GDP-bound Gαi-subunit is associated
underlying heart disease.29,30,34 In addition, the alterations in with the Kir3.1 channel subunit, thereby creating a “preformed
single-channel properties of IK,ACh,c are comparable between dogs complex” between the heterotrimeric G-protein and the IK,ACh
with ATR and cAF patients, suggesting a common molecular channel.40 In this condition, Gβγ-subunits are always in close
basis.34 The complex regulation of IK,ACh points to several possible proximity to the channel and can thus effectively and rapidly
mechanisms that could contribute to the development of IK,ACh,c. activate IK,ACh upon Gαi-dissociation. The interaction between
The following discussion summarizes the current knowledge GαI and Kir3.1 is therefore an important contributor to the
about the molecular regulation of agonist-dependent IK,ACh in specificity between receptor stimulation and IK,ACh activation. In
atrial myocytes, particularly focusing on the putative mechanisms addition, GDP-bound Gαi can act as a Gβγ-scavenger, continu-
underlying constitutive IK,ACh activity in AF. ously chelating Gβγ-subunits from IK,ACh channels, thereby pre-
venting agonist-independent IK,ACh activity.41 It is unknown
whether reduced Gαi-expression or a higher Gβγ-subunit avail-
ability contribute to the enhanced constitutive IK,ACh activity in
G-Protein Cycle Can Contribute to the patients with AF.
Generation of Constitutive IK,ACh The cycling of Gi-proteins is fine-tuned by multiple mecha-
nisms. Regulator of G-protein signaling (RGS) proteins, which
M2-Receptor–Dependent IK,ACh accelerate the GTPase activity of the Gα-protein subunit, accel-
erate the inactivation kinetics of receptor-activated IK,ACh (see
Inward-rectifier IK,ACh channels are activated through stimulation Figure 38-3, B).42 [Na+]i also accelerates the G-protein cycle,43,44
of appropriate Gi-protein–coupled receptors (M2-receptors), although direct binding of Na+ to Kir3.4 could activate IK,ACh
resulting in the dissociation of heterotrimeric Gi-proteins and channels (see Figure 38-3, A).45-47,47a In addition, nucleoside
consecutive binding of Gβγ-subunits to the channel (see Figure diphosphate kinases (NDPKs), which classically transfer phos-
38-2, A, B).38 Heterotrimeric G-proteins are composed of two phates from ATP to GDP by a ping-pong mechanism involving
functional units, the guanosine-5′-diphosphate and guanosine-5′- the formation of a high-energy phosphate intermediate on
triphosphate (GDP/GTP) binding Gα-subunit and the Gβγ- histidine-118, might also contribute to the replenishment of the
dimer. Upon stimulation of G-protein–coupled receptors, GDP GTP necessary for the G-protein cycle (see Figure 38-2, B, C).39
(bound to the Gα-subunit under resting conditions) is released
and replaced by GTP. The resulting conformational changes lead
to a dissociation of Gα and Gβγ subunits, which then initiate and Constitutively Active IK,ACh
regulate multiple intracellular pathways. Intrinsic GTPase activity
of the Gα-subunit results in GTP hydrolysis into GDP followed Previous studies showed that the M-receptor antagonist atropine
by reassembly of the G-protein subunits and reestablishment of does not abolish constitutive IK,ACh,c activity in atrial myocytes
from cAF patients or ATR-dogs, suggesting that a receptor- neutralize a negatively charged aspartate in the C-terminus of
independent mechanism might be implicated in the formation of Kir3.4, but not in Kir3.1, allowing stronger binding of PIP2 (see
constitutive IK,ACh,c.19 Whether increased receptor-independent Figure 38-3, A). Mintert et al46 found that increases in [Na+]i
dissociation of Gαi- and Gβγ-subunits contributes to generation resulted in higher constitutive IK,ACh,c activity in rat atrial cardio-
of constitutive IK,ACh,c in cAF patients is not known. Neither myocytes overexpressing the Kir3.4 subunit.46 However, it is
pertussis toxin, which uncouples Gi-proteins from the receptor, unknown whether changes in PIP2 membrane content or
nor the absence of GTP, which is necessary for the action of the subunit composition of the IK,ACh-channel contribute to
G-proteins, affected the IK,ACh-like current component of the clinically relevant constitutive IK,ACh,c activity and AF-associated
basal inward-rectifier K+ current in atrial myocytes from ATR IK,ACh remodeling. Nevertheless, rapid atrial rate-related
dogs and atrial myocytes from cAF patients,19,29 suggesting that increases in subsarcolemmal [Na+]i are likely to occur during
receptor-independent dissociation of Gαi- and Gβγ-subunits AF in vivo, potentially contributing to constitutive IK,ACh,c
might not be a major contributor to constitutive IK,ACh,c in AF. activity.60
However, a fraction of G-protein βγ-subunits forms complexes
with the B-isoform of NDPKs and these complexes may cause
receptor-independent G-protein activation (see Figure 38-2, B).39
This process involves NDPK-B–mediated phosphorylation of Phosphorylation-Dependent IK,ACh Regulation
the Gβ-subunit at histidine-266. Subsequently the phosphate is
transferred onto GDP bound to the Gα-subunit, and the formed M2-Receptor–Dependent IK,ACh
GTP leads to receptor-independent G-protein activation.
Because NDPK-B accumulates in the atrial membrane fraction The GIRK channel forms a macromolecular complex that is
of cAF patients48 and is involved in IK,ACh regulation,49 this composed of the catalytic subunits of protein kinase A (PKA) and
receptor-independent mechanism of G-protein activation could C (PKC), the Ca2+/calmodulin-regulated protein kinase II
be a major contributor to the development of constitutive IK,ACh,c (CaMKII) and the type-1 and type-2A protein phosphatases (PP1
in patients with cAF, although the NDPK hypothesis has also and PP2A).61 In addition, the Kir3.1 and Kir3.4 channel subunits
been challenged as a possible explanation for agonist-independent possess phosphorylation sites for PKA, PKC, CaMKII, and pos-
channel activation.50 sibly for PKG,61,62 suggesting that channel phosphorylation status
might play a crucial role in the regulation of IK,ACh activity (see
Figure 38-3, C).
Inhibition of CaMKII, PKC, and PKG reduced agonist-
Regulation of IK,ACh Gating by PIP2 activated IK,ACh current in human atrial myocytes, suggesting that
CaMKII-, PKC- and PKG-mediated channel phosphorylation
M2-Receptor Dependent IK,ACh may stabilize the binding of Gβγ-subunits to the channel, thereby
increasing IK,ACh.20 The regulation of IK,ACh by PKC is isoform
IK,ACh channels require PIP2 to maintain physiological properties specific. The PKC family contains many isoforms, including con-
and removal of PIP2 completely runs down the IK,ACh current.51 ventional isoforms (cPKCs), which are activated by Ca2+ and
Under resting conditions, the interaction between IK,ACh-channels DAG, and novel isoforms (nPKCs), which require DAG but not
and PIP2 is weak, and IK,ACh activity is low. Binding of Gβγ- Ca2+ for activation.63 There are at least five major PKC isoforms
subunits to IK,ACh channels strengthen the PIP2-channel subunit in the heart: cPKCα, cPKCβI, cPKCβII, nPKCδ, and nPKCε.
interaction, strongly increasing IK,ACh activity (see Figure 38-3, Purified cPKC isoforms strongly reduce GTPγS-activated IK,ACh
A). In atrial myocytes, the membrane PIP2 levels close to the channel, whereas nPKC isoforms have the opposite effect.33 The
IK,ACh channel are dynamically regulated by Gαq-protein–coupled relative expression-levels of stimulatory and inhibitory PKC iso-
receptors, the stimulation of which results in PIP2 breakdown by forms can vary depending on species, diseases, or other condi-
activating the PIP2 hydrolyzing enzyme phospholipase-C (see tions, potentially accounting for discrepant results between
Figure 38-3, A). IK,ACh is also regulated by Gq-protein–coupled neonatal rat and canine atrial myocytes, in which PKC activation
α1-adrenergic-, angiotensin-II type-1 (AT1)-, and endothelin-A inhibits IK,ACh and human atrial myocytes, for which PKC
(ETA) receptors, suggesting that GIRK-channels integrate appears to stimulate IK,ACh.61,64 PKA activity is expected to be
hormone and neurotransmitter signals from different path- reduced via Gαi-coupled M2-receptor activation consequent to
ways.52,53 In addition, there is also evidence for the existence of ACh-application.
Gαq-coupled M1- and M3-receptors in atria.54,55 Activation of Upon application of ACh, direct binding of liberated Gβγ-
Gαq-coupled M1- and M3-receptors and the associated PIP2 subunits to the IK,ACh channel causes rapid activation of IK,ACh,
depletion are suggested to contribute to the fast desensitization followed by desensitization with a typical biphasic pattern (see
of M2-receptor–activated IK,ACh current (see Figure 38-3, B).56,57 Figure 38-3, middle).62 Although other mechanisms such as RGS
However, the existence of functional non–M2-receptors in atria proteins and PIP2 depletion owing to activation of Gq-protein–
is still controversial.58,59 coupled M1/M3-receptors could also contribute, it has been sug-
gested that the fast phase of desensitization is due to channel
dephosphorylation,65,66 whereas the slower (intermediate) phase
involves progressive M-receptor phosphorylation by a G-protein–
Constitutively Active IK,ACh coupled receptor type-2 kinase, which uncouples the receptor
from the G-protein (see Figure 38-3, C).67
Activation of IK,ACh channels via Gβγ-subunits is mediated via Because there is evidence for increased PKA, CaMKII, and
stronger binding of PIP2 (see Figure 38-3, A), and enhanced PIP2 PKC activity in cAF patients,20,68 IK,ACh phosphorylation might
levels in the IK,ACh channel microdomain activate the IK,ACh also be involved in disease-associated remodeling of IK,ACh.
channel, even in the absence of Gβγ-subunits.51 As a result, However, inhibition of CaMKII, PKC, PKG, and PKA does not
agonist-independent IK,ACh,c activity can result from a Gβγ- affect M-receptor–activated IK,ACh in patients with cAF.20 In addi-
independent increase in the PIP2-IK,ACh channel interaction. In tion, activities of PP1 and PP2A are higher in cAF than in control
addition, IK,ACh channels containing Kir3.4 subunits can be acti- SR patients and do not translate into altered protein
vated by an increase in [Na+]i in a Gβγ-subunit–independent phosphorylation.69-71 Thus, the lack of contribution of the these
manner (see Figure 38-3, A).45-47 Na+ ions are thought to kinases to M-receptor–activated IK,ACh in cAF might result from
opposing increases in channel dephosphorylation. Alternatively, Constitutive
38
the channel could be hyperphosphorylated because of abnormal IK,ACh activity
control of kinase–phosphatase signaling in the macromolecular (IK,ACh,c)
complex,61 and inhibition of a single kinase might not reduce the
channel’s phosphorylation state below the threshold required to
K+ IK,ACh
impair agonist-activated IK,ACh. Further work is needed to verify
these hypotheses.
P
Constitutively Active IK,ACh,c PKCε
(membrane
PKCα P fraction)
Because activation of IK,ACh requires ATP, modified
phosphorylation-dependent channel regulation could contribute
to constitutive IK,ACh,c activity.62,66 In single-channel patch-clamp
experiments, excision of the patch and removal from the intact
cell produces a cell-free piece of sarcolemma attached to the Calpains
pipette and containing the IK,ACh channel (inside-out configura-
tion), with the cytosolic side facing the bath solution. Under Ca2+
these conditions, IK,ACh,c open probability is strongly reduced,
suggesting that an intracellular component necessary for consti-
tutive IK,ACh,c activity is washed out.33 In the presence of phospha-
tase inhibitors, excision of the patch only slightly reduces the
open probability of constitutive IK,ACh,c. These experiments point
to a crucial role of phosphorylation in IK,ACh,c. Inhibition of PKC
reduces basal current, which contains constitutive IK,ACh,c activity,
Atrial rate
in patients with cAF but not SR, further suggesting that PKC-
mediated phosphorylation of IK,ACh channels is involved in main-
taining constitutive IK,ACh,c.20 The expression level of PKCε, Figure 38-4. Protein kinase C (PKC) isoform shift contributes to development of
constitutively active IK,ACh (IK,ACh,c). Atrial tachycardia induces Ca2+-independent trans-
which stimulates IK,ACh,33 is increased in patients with cAF, whereas
location of PKCε from the cytosol to the membrane and causes Ca2+/calpain–
PKCα, PKCβI, and PKCδ remains unchanged.20 These data dependent breakdown of PKCα. Both reduced PKCα expression and PKCε
indicate the involvement of PKC isoform imbalance in the devel- membrane translocation contribute to increased IK,ACh,c.
opment of agonist-independent IK,ACh,c. Consistent with this
hypothesis, Makary et al.33 showed that changes in the balance (Figure was produced using Servier Medical Art.)
between stimulatory nPKC and inhibitory cPKC isoforms con-
tribute to increased constitutive IK,ACh,c activity in AF.33 The IK,ACh Channels as Potential Therapeutic
cPKC (inhibitory) phenotype predominates in control dogs, Targets in Atrial Fibrillation
whereas the nPKC (stimulatory) phenotype prevails in ATR dogs
(Figure 38-4).33 This idea is supported by a reduced total Constitutive IK,ACh activity has emerged as a potentially interest-
expression of PKCα (cPKC isoform) and increased membrane ing therapeutic target for AF treatment. Because of much higher
translocation of PKCε (nPKC isoform), in atrial myocytes from IK,ACh expression in the atria in comparison to the ventricles,
ATR dogs. In control myocytes, blocking cPKC isoforms selective IK,ACh inhibition could have antiarrhythmic effects in the
increases constitutive IK,ACh,c, suggesting tonic cPKC-mediated atria, without ventricular side effects.72 Because some drugs with
inhibition of IK,ACh because of predominance of inhibitory PKCα anti-AF efficacy (e.g., amiodarone, flecainide, quinidine, vera-
isoforms. cPKC inhibition has no effect in ATR dogs. In contrast, pamil) are also inhibitors of IK,ACh,4,21,73,74 their clinical effective-
blocking PKCε does not affect constitutive IK,ACh,c in control ness may result in part from blockade of IK,ACh,c.
myocytes, but it reduces IK,ACh,c in ATR myocytes to levels seen Recent research on structural motifs in K+ channel drug inter-
in control dogs. These data indicate that a reduced tonic inhibi- actions can allow the development of more specific IK,ACh-channel
tory effect of PKCα and an increased stimulatory effect of PKCε blocking agents for AF therapy. In addition to a central conduction
are the major determinants of enhanced constitutive IK,ACh,c activ- pathway formed by four channel subunits, both the amino and
ity in AF. carboxyl termini of the subunits form a second cytoplasmic pore
Rapid in vitro pacing of isolated atrial myocytes from control that extends the ion permeation pathway toward the intracellular
dogs reproduces the phenotype seen in ATR dogs, with a PKC side of the membrane.75,76 Tertiapin, a peptide toxin from the
isoform switch and increased constitutive IK,ACh,c activity, demon- honey bee venom, inhibits IK,ACh-channel selectively77-79 by binding
strating that the PKC isoform switch in ATR could be attribut- to the external vestibule of the central conduction pathway.80,81 In
able to the rapid atrial activity per se.33 In addition, the cellular contrast, chloroquine blocks inward-rectifier K+ channels by
in vitro pacing experiments revealed that PKCα downregulation interacting with the center of the cytoplasmic conduction pore via
likely results from atrial tachycardia-induced Ca2+ loading and interference with negatively charged aromatic residues within a
subsequent activation of the Ca2+-dependent protease calpain, central cavity.82 These results suggest that compounds targeting
which might increase the degradation of PKCα isoforms. In the specific cytoplasmic domain residues of ACh-dependent
contrast, the increased membrane translocation of PKCε in ATR inward-rectifier K+ channels might constitute chemical leads for
myocytes seems to be Ca2+ independent (see Figure 38-4), sug- the design of novel selective IK,ACh channel blockers.
gesting that two independent mechanisms combine to cause con- During the past few years, several IK,ACh-selective drugs have
stitutive IK,ACh,c activity in AF. Although there is only indirect been developed (e.g., NIP-142, NTC-801). However off-target
evidence that phosphorylation-dependent mechanisms contrib- effects on IK,ACh channels in the SA node and peripheral tissues
ute to the increase of constitutive IK,ACh,c in human AF,20 these (i.e., gastrointestinal tract, genitourinary system) could limit the
posttranslational mechanisms could provide interesting new value of directly targeting the channel pore.72 An improved
therapeutic targets for atrial-selective and pathology-specific understanding of the molecular basis for increased constitutive
treatment of AF. IK,ACh,c activity in cAF might allow for atrial-selective and
disease-specific targeting of IK,ACh in AF. However, such strategies 07CVD03), the German Federal Ministry of Education
are still in their infancy, and much more work is needed before and Research through the Atrial Fibrillation Competence
this approach can be tested for efficacy and value in clinical AF. Network (grant 01Gi0204) and German Centre for Cardiovas-
cular Research, the Deutsche Forschungsgemeinschaft
(grant Do 769/1-3), the European Union (European Network for
Acknowledgments Translational Research in Atrial Fibrillation, EUTRAF, grant
261057), the Canadian Institutes of Health Research (MGP6957
This work was supported by the Foundation Leducq (European- and MOP44365), and the Quebec Heart and Stroke
North American Atrial Fibrillation Research Alliance, grant Foundation.
a mouse knockout model. J Am Coll Cardiol 32. Voigt N, Makary S, Nattel S, et al: Voltage-clamp-
References 37(8):2136–2143, 2001. based methods for the detection of constitutively
17. Pandit SV, Berenfeld O, Anumonwo JM, et al: active acetylcholine-gated IK,ACh channels in the
1. Loewi O: Über humorale Übertragbarkeit der Ionic determinants of functional reentry in a 2-D diseased heart. Methods Enzymol 484:653–675,
Herznervenwirkung. Pflügers Arch 189(1):239– model of human atrial cells during simulated 2010.
242, 1921. chronic atrial fibrillation. Biophys J 88(6):3806– 33. Makary S, Voigt N, Maguy A, et al: Differential
2. Noma A, Trautwein W: Relaxation of the ACh- 3821, 2005. protein kinase C isoform regulation and increased
induced potassium current in the rabbit sinoatrial 18. Katsouras G, Sakabe M, Comtois P, et al: Differ- constitutive activity of acetylcholine-regulated
node cell. Pflügers Arch 377(3):193–200, 1978. ences in atrial fibrillation properties under vagal potassium channels in atrial remodeling. Circ Res
3. Krapivinsky G, Gordon EA, Wickman K, et al: nerve stimulation versus atrial tachycardia remod- 109(9):1031–1043, 2011.
The G-protein-gated atrial K+ channel IK,ACh is a eling. Heart Rhythm 6(10):1465–1472, 2009. 34. Voigt N, Maguy A, Yeh YH, et al: Changes in IK,ACh
heteromultimer of two inwardly rectifying K+- 19. Dobrev D, Friedrich A, Voigt N, et al: The G single-channel activity with atrial tachycardia
channel proteins. Nature 374(6518):135–141, protein-gated potassium current IK,ACh is constitu- remodelling in canine atrial cardiomyocytes. Car-
1995. tively active in patients with chronic atrial fibrilla- diovasc Res 77(1):35–43, 2008.
4. Hibino H, Inanobe A, Furutani K, et al: Inwardly tion. Circulation 112(24):3697–3706, 2005. 35. Heidbüchel H, Vereecke J, Carmeliet E: Three
rectifying potassium channels: their structure, 20. Voigt N, Friedrich A, Bock M, et al: Differential different potassium channels in human atrium.
function, and physiological roles. Physiol Rev phosphorylation-dependent regulation of constitu- Contribution to the basal potassium conductance.
90(1):291–366, 2010. tively active and muscarinic receptor- Circ Res 66(5):1277–1286, 1990.
5. Yamada M, Inanobe A, Kurachi Y: G protein regu- activated IK,ACh channels in patients with chronic 36. Sato R, Hisatome I, Wasserstrom JA, et al:
lation of potassium ion channels. Pharmacol Rev atrial fibrillation. Cardiovasc Res 74(3):426–437, Acetylcholine-sensitive potassium channels in
50(4):723–760, 1998. 2007. human atrial myocytes. Am J Physiol 259(6 Pt
6. Bender K, Wellner-Kienitz MC, Inanobe A, et al: 21. Voigt N, Rozmaritsa N, Trausch A, et al: Inhibition 2):H1730–H1735, 1990.
Overexpression of monomeric and multimeric of IK,ACh current may contribute to clinical efficacy 37. Koumi S-i, Backer CL, Arentzen CE: Character-
GIRK4 subunits in rat atrial myocytes removes fast of class I and class III antiarrhythmic drugs in ization of inwardly rectifying K+ channel in human
desensitization and reduces inward rectification of patients with atrial fibrillation. Naunyn Schmiede- cardiac myocytes : alterations in channel behavior
muscarinic K+ current (IK,ACh). Evidence for func- bergs Arch Pharmacol 381(3):251–259, 2010. in myocytes isolated from patients with idiopathic
tional homomeric GIRK4 channels. J Biol Chem 22. Voigt N, Trausch A, Knaut M, et al: Left-to-right dilated cardiomyopathy. Circulation 92(2):164–
276(31):28873–28880, 2001. atrial inward rectifier potassium current gradients 174, 1995.
7. Dobrev D, Wettwer E, Himmel HM, et al: in patients with paroxysmal versus chronic atrial 38. Logothetis DE, Kurachi Y, Galper J, et al: The beta
G-Protein beta(3)-subunit 825T allele is associated fibrillation. Circ Arrhythm Electrophysiol gamma subunits of GTP-binding proteins activate
with enhanced human atrial inward rectifier potas- 3(5):472–480, 2010. the muscarinic K+ channel in heart. Nature
sium currents. Circulation 102(6):692–697, 23. Lazar S, Dixit S, Marchlinski FE, et al: Presence 325(6102):321–326, 1987.
2000. of left-to-right atrial frequency gradient in parox- 39. Wieland T: Interaction of nucleoside diphosphate
8. Kurachi Y, Nakajima T, Sugimoto T: On the mech- ysmal but not persistent atrial fibrillation in kinase B with heterotrimeric G protein betagamma
anism of activation of muscarinic K+ channels by humans. Circulation 110(20):3181–3186, 2004. dimers: consequences on G protein activation and
adenosine in isolated atrial cells: involvement of 24. Sanders P, Berenfeld O, Hocini M, et al: Spectral stability. Naunyn Schmiedebergs Arch Pharmacol
GTP-binding proteins. Pflugers Arch 407(3):264– analysis identifies sites of high-frequency activity 374(5-6):373–383, 2007.
274, 1986. maintaining atrial fibrillation in humans. Circula- 40. Riven I, Iwanir S, Reuveny E: GIRK channel acti-
9. Himmel HM, Meyer Zu Heringdorf D, Graf E, et tion 112(6):789–797, 2005. vation involves a local rearrangement of a pre-
al: Evidence for Edg-3 receptor-mediated activa- 25. Swartz MF, Fink GW, Lutz CJ, et al: Left versus formed G protein channel complex. Neuron
tion of IK,ACh by sphingosine-1-phosphate in human right atrial difference in dominant frequency, K+ 51(5):561–573, 2006.
atrial cardiomyocytes. Mol Pharmacol 58(2):449– channel transcripts, and fibrosis in patients devel- 41. Yamada M, Ho YK, Lee RH, et al: Muscarinic K+
454, 2000. oping atrial fibrillation after cardiac surgery. Heart channels are activated by beta gamma subunits and
10. Verkerk AO, Geuzebroek GS, Veldkamp MW, Rhythm 6(10):1415–1422, 2009. inhibited by the GDP-bound form of alpha subunit
et al: Effects of acetylcholine and noradrenalin on 26. Sarmast F, Kolli A, Zaitsev A, et al: Cholinergic of transducin. Biochem Biophys Res Commun
action potentials of isolated rabbit sinoatrial and atrial fibrillation: IK,ACh gradients determine 200(3):1484–1490, 1994.
atrial myocytes. Front Physiol 3:174, 2012. unequal left/right atrial frequencies and rotor 42. Doupnik CA, Davidson N, Lester HA, et al: RGS
11. Gaborit N, Le Bouter S, Szuts V, et al: Regional dynamics. Cardiovasc Res 59(4):863–873, 2003. proteins reconstitute the rapid gating kinetics of
and tissue specific transcript signatures of ion 27. Sakmann B, Noma A, Trautwein W: Acetylcholine gbetagamma-activated inwardly rectifying K+
channel genes in the non-diseased human heart. J activation of single muscarinic K+ channels in iso- channels. Proc Natl Acad Sci U S A 94(19):10461–
Physiol 582(Pt 2):675–693, 2007. lated pacemaker cells of the mammalian heart. 10466, 1997.
12. Koumi S, Wasserstrom JA: Acetylcholine-sensitive Nature 303(5914):250–253, 1983. 43. Rishal I, Keren-Raifman T, Yakubovich D, et al:
muscarinic K+ channels in mammalian ventricular 28. Ito H, Ono K, Noma A: Background conductance Na+ promotes the dissociation between Galpha
myocytes. Am J Physiol 266(5 Pt 2):H1812–H1821, attributable to spontaneous opening of muscarinic GDP and Gbeta gamma, activating G protein-
1994. K+ channels in rabbit sino-atrial node cells. J gated K+ channels. J Biol Chem 278(6):3840–3845,
13. Dobrev D, Graf E, Wettwer E, et al: Molecular Physiol 476(1):55–68, 1994. 2003.
basis of downregulation of G-protein-coupled 29. Ehrlich JR, Cha TJ, Zhang L, et al: Characteriza- 44. Yakubovich D, Rishal I, Dascal N: Kinetic model-
inward rectifying K+ current (IK,ACh) in chronic tion of a hyperpolarization-activated time- ing of Na+-induced, Gbetagamma-dependent acti-
human atrial fibrillation: decrease in GIRK4 dependent potassium current in canine vation of G protein-gated K+ channels. J Mol
mRNA correlates with reduced IK,ACh and musca- cardiomyocytes from pulmonary vein myocardial Neurosci 25(1):7–19, 2005.
rinic receptor-mediated shortening of action sleeves and left atrium. J Physiol 557(Pt 2):583– 45. Ho IH, Murrell-Lagnado RD: Molecular mecha-
potentials. Circulation 104(21):2551–2557, 2001. 597, 2004. nism for sodium-dependent activation of G
14. Wakili R, Voigt N, Kaab S, et al: Recent advances 30. Cha TJ, Ehrlich JR, Chartier D, et al: Kir3-based protein-gated K+ channels. J Physiol 520 Pt 3:645–
in the molecular pathophysiology of atrial fibrilla- inward rectifier potassium current: potential role in 651, 1999.
tion. J Clin Invest 121(8):2955–2968, 2011. atrial tachycardia remodeling effects on atrial 46. Mintert E, Bosche LI, Rinne A, et al: Generation
15. Chen PS, Tan AY: Autonomic nerve activity and repolarization and arrhythmias. Circulation of a constitutive Na+-dependent inward-rectifier
atrial fibrillation. Heart Rhythm 4(3 Suppl):S61– 113(14):1730–1737, 2006. current in rat adult atrial myocytes by overexpres-
S64, 2007. 31. Dobrev D, Nattel S: New antiarrhythmic drugs sion of Kir3.4. J Physiol 585(Pt 1):3–13, 2007.
16. Kovoor P, Wickman K, Maguire CT, et al: Evalu- for treatment of atrial fibrillation. Lancet 47. Rosenhouse-Dantsker A, Sui JL, Zhao Q, et al: A
ation of the role of IK,ACh in atrial fibrillation using 375(9721):1212–1223, 2010. sodium-mediated structural switch that controls
the sensitivity of Kir channels to PtdIns(4,5)P(2). messenger PIP2 leads to K+-current desensitiza- human chronic atrial fibrillation. Circulation
Nat Chem Biol 4(10):624–631, 2008. tion. Nat Cell Biol 2(8):507–514, 2000. 114(7):670–680, 2006.
47a. Voigt N, Heijman J, Trausch A, et al: Impaired
Na+-dependent regulation of acetylcholine-
activated inward-rectifier K+ current modulates
58. Hellgren I, Mustafa A, Riazi M, et al: Muscarinic
M3 receptor subtype gene expression in the
human heart. Cell Mol Life Sci 57(1):175–180,
71. Greiser M, Halaszovich CR, Frechen D, et al:
Pharmacological evidence for altered src kinase
regulation of ICa,L in patients with chronic atrial
38
action potential rate dependence in patients with 2000. fibrillation. Naunyn Schmiedebergs Arch Pharma-
chronic atrial fibrillation. J Mol Cell Cardiol 2013 59. Oberhauser V, Schwertfeger E, Rutz T, et al: Ace- col 375(6):383–392, 2007.
[Epub ahead of print]. tylcholine release in human heart atrium: influence 72. Dobrev D, Carlsson L, Nattel S: Novel molecular
48. Abu-Taha I, Voigt N, Nattel S, et al: Nucleoside of muscarinic autoreceptors, diabetes, and age. Cir- targets for atrial fibrillation therapy. Nat Rev Drug
diphosphate kinase B is a novel receptor- culation 103(12):1638–1643, 2001. Discov 11(4):275–291, 2012.
independent activator of G-protein signaling in 60. Grandi E, Pandit SV, Voigt N, et al: Human atrial 73. Kurachi Y, Nakajima T, Ito H, et al: AN-132, a new
clinical and experimental atrial fibrillation. Heart action potential and Ca2+ model: sinus rhythm and class I anti-arrhythmic agent, depresses the
Rhythm 9(5 (Supplement):S397, 2012. chronic atrial fibrillation. Circ Res 109(9):1055– acetylcholine-induced K+ current in atrial myo-
49. Kaibara M, Nakajima T, Irisawa H, et al: Regula- 1066, 2011. cytes. Eur J Pharmacol 165(2-3):319–322, 1989.
tion of spontaneous opening of muscarinic K+ 61. Nikolov EN, Ivanova-Nikolova TT: Coordination 74. Kurachi Y, Nakajima T, Sugimoto T: Quinidine
channels in rabbit atrium. J Physiol 433:589–613, of membrane excitability through a GIRK1 signal- inhibition of the muscarine receptor-activated K+
1991. ing complex in the atria. J Biol Chem channel current in atrial cells of guinea pig.
50. Sorota S, Chlenov M, Du XY, et al. ATP-dependent 279(22):23630–23636, 2004. Naunyn Schmiedebergs Arch Pharmacol
activation of the atrial acetylcholine-induced K+ 62. Medina I, Krapivinsky G, Arnold S, et al: A switch 335(2):216–218, 1987.
channel does not require nucleoside diphosphate mechanism for G beta gamma activation of IK,ACh. 75. Furutani K, Ohno Y, Inanobe A, et al: Mutational
kinase activity. Circ Res 82(9):971–979, 1998. J Biol Chem 275(38):29709–29716, 2000. and in silico analyses for antidepressant block of
51. Huang CL, Feng S, Hilgemann DW: Direct acti- 63. Steinberg SF: Structural basis of protein kinase C astroglial inward-rectifier Kir4.1 channel. Mol
vation of inward rectifier potassium channels by isoform function. Physiol Rev 88(4):1341–1378, Pharmacol 75(6):1287–1295, 2009.
PIP2 and its stabilization by Gbetagamma. Nature 2008. 76. Whorton MR, MacKinnon R: Crystal structure of
391(6669):803–806, 1998. 64. Yeh YH, Ehrlich JR, Qi X, et al: Adrenergic control the mammalian GIRK2 K+ channel and gating
52. Cui S, Ho WK, Kim ST, et al: Agonist-induced of a constitutively active acetylcholine-regulated regulation by G proteins, PIP2, and sodium. Cell
localization of Gq-coupled receptors and G potassium current in canine atrial cardiomyocytes. 147(1):199–208, 2011.
protein-gated inwardly rectifying K+ (GIRK) chan- Cardiovasc Res 74(3):406–415, 2007. 77. Jin W, Klem AM, Lewis JH, et al: Mechanisms
nels to caveolae determines receptor specificity of 65. Kim D: Mechanism of rapid desensitization of of inward-rectifier K+ channel inhibition
phosphatidylinositol 4,5-bisphosphate signaling. J muscarinic K+ current in adult rat and guinea pig by tertiapin-Q. Biochemistry 38(43):14294–14301,
Biol Chem 285(53):41732–41739, 2010. atrial cells. Circ Res 73(1):89–97, 1993. 1999.
53. Choisy SC, James AF, Hancox JC: Acute desensi- 66. Shui Z, Boyett MR, Zang WJ. ATP-dependent 78. Jin W, Lu Z: A novel high-affinity inhibitor for
tization of acetylcholine and endothelin-1 activated desensitization of the muscarinic K+ channel in rat inward-rectifier K+ channels. Biochemistry
inward rectifier K+ current in myocytes from the atrial cells. J Physiol 505 ( Pt 1):77–93, 1997. 37(38):13291–13299, 1998.
cardiac atrioventricular node. Biochem Biophys 67. Shui Z, Boyett MR, Zang WJ, et al: Receptor 79. Jin W, Lu Z: Synthesis of a stable form of tertiapin:
Res Commun 2012. kinase-dependent desensitization of the muscarinic a high-affinity inhibitor for inward-rectifier K+
54. Wang H, Han H, Zhang L, et al: Expression of K+ current in rat atrial cells. J Physiol 487(Pt channels. Biochemistry 38(43):14286–14293, 1999.
multiple subtypes of muscarinic receptors and cel- 2):359–366, 1995. 80. Ramu Y, Klem AM, Lu Z: Short variable sequence
lular distribution in the human heart. Mol Phar- 68. Voigt N, Li N, Wang Q, et al: Enhanced sarcoplas- acquired in evolution enables selective inhibition
macol 59(5):1029–1036, 2001. mic reticulum Ca2+ leak and increased Na+-Ca2+ of various inward-rectifier K+ channels. Biochem-
55. Perez CC, Tobar ID, Jimenez E, et al: Kinetic and exchanger function underlie delayed afterdepolar- istry 43(33):10701–10709, 2004.
molecular evidences that human cardiac muscle izations in patients with chronic atrial fibrillation. 81. Ramu Y, Xu Y, Lu Z: Engineered specific and high-
express non-M2 muscarinic receptor subtypes that Circulation 125(17):2059–2070, 2012. affinity inhibitor for a subtype of inward-rectifier
are able to interact themselves. Pharmacol Res 69. Christ T, Boknik P, Wohrl S, et al: L-type Ca2+ K+ channels. Proc Natl Acad Sci U S A
54(5):345–355, 2006. current downregulation in chronic human atrial 105(31):10774–10778, 2008.
56. Jan LY, Jan YN: Heartfelt crosstalk: desensitization fibrillation is associated with increased activity of 82. Noujaim SF, Stuckey JA, Ponce-Balbuena D, et al:
of the GIRK current. Nat Cell Biol 2(9):E165– protein phosphatases. Circulation 110(17):2651– Specific residues of the cytoplasmic domains of
E167, 2000. 2657, 2004. cardiac inward rectifier potassium channels are
57. Kobrinsky E, Mirshahi T, Zhang H, et al: Receptor- 70. El-Armouche A, Boknik P, Eschenhagen T, et al: effective antifibrillatory targets. FASEB J
mediated hydrolysis of plasma membrane Molecular determinants of altered Ca2+ handling in 24(11):4302–4312, 2010.
Pulmonary Vein Ganglia and the
Neural Regulation of the Heart Rate 39
Manuel Zarzoso and Sami F. Noujaim
several cardiac regions, coronary vasculature, and the major

CHAPTER OUTLINE
intrathoracic and cervical vessels.5
Autonomic Innervation of the Heart 393
Autonomic Regulation of Pacemaker Activity 393
Intrinsic Nervous System
Pulmonary Vein Ganglia and Heart Rate Control 394
Initially, the ICNS was thought to be composed of parasympa-
Pulmonary Vein Ganglia and Atrial Fibrillation 396
thetic postganglionic neurons and their axonal projections,
alongside with intramyocardial chromaffin cells, which act as a
simple relay region under the control of the central nervous
Autonomic Innervation of the Heart system. However, more recent studies have shown that the ICNS
represents the final relay center for the coordination of regional
The traditional view of the sympathetic and parasympathetic cardiac function and is composed of sensory (afferent), intercon-
branches of the autonomic nervous system is that they exert fine- necting (local circuit), and motor (adrenergic an cholinergic
tuned reciprocal influences on the heart. Sympathetic stimulation efferent) neurons. These neurons communicate with intratho-
increases heart rate, electrical conductivity, and contractility, racic extracardiac ganglia, forming a distributive network that
whereas in general, parasympathetic stimulation produces oppos- processes centripetal and centrifugal neuronal impulses for
ing inhibitory effects. However, the cardiac regulation driven by cardiac control, under the influence of the central nervous system,
the two branches of the autonomic nervous system is complex. and circulating catecholamines.2
In fact, the electromechanical function of the heart is subject to The number of cardiac ganglia is variable and species depen-
the influence of not one, but two separate autonomic nervous dent.6 The location, shape, and size are also variable, but in many
systems—one extrinsic and the other intrinsic. In the extrinsic mammals, including humans, intrinsic cardiac ganglia are usually
system, the primary site for regulating sympathetic and parasym- distributed at specific atrial regions: around the sinoatrial node
pathetic (vagal) outflow to the heart and blood vessels is the (SAN), the roots of caval and pulmonary veins (PVs), and near
medulla, which is located in the brainstem above the spinal cord. the atrioventricular node.7,8 Intrinsic cardiac ganglia are also
The intrinsic cardiac nervous system (ICNS) is composed of present within the ventricles, although in a smaller number com-
ganglionated plexi distributed at various locations within the pared with the atria.9
heart, including the epicardium, myocardium and endocardium.1 The ICNS regulates several aspects of cardiac function such
Efferent fibers of extrinsic autonomic system enter the heart as heart rate, atrial and ventricular refractoriness, conduction,
through the hilum and then connect with and modulate the activ- contractility and blood flow.10 Furthermore, the ICNS modulates
ity of the intrinsic cardiac ganglionated plexi. Both systems are intrathoracic and central cardiovascular-cardiac reflexes and
susceptible to neuromodulatory influences from various inputs, coordinates parasympathetic and sympathetic efferent postgan-
including the central nervous system, the baroreceptor and che- glionic neuronal input to the heart.11
moreceptor reflexes, and the local interneuronal interactions It has been suggested that intrinsic cardiac ganglionated plex-
within the heart itself.2 uses exert influence over adjacent myocardial regions,12-14 where
vagal deceleration of heart rate can be mediated selectively by
neurons located at the junction of the right atrium and superior
Extrinsic Nervous System vena cava, whereas the effects on atrioventricular nodal transmis-
sion can be controlled by the neurons of a fat pad at the junction
Efferent preganglionic parasympathetic innervation originates of the inferior vena cava and the inferior left atrium.13
mainly in the nucleus ambiguous of the medulla, whereas some On the other hand, Armour2 proposed that intrinsic cardiac
neurons are located in the dorsal motor nucleus and the regions ganglia in atrial or ventricular ganglionated plexuses can selec-
between these two nuclei. Neurons originating in these areas tively influence the electrical and mechanical properties in adja-
project their axons to form synapses with postganglionic neurons cent tissues and cardiac chambers. For example, it has been
located throughout the various atrial or ventricular ganglionated reported that the cholinergic neurons of the right atrial ganglion-
plexi.2 ated plexuses can decrease the rate of discharge of SAN, depress
On the other hand, sympathetic innervation originates in the the atrioventricular node conduction, and affect ventricular con-
interomediolateral nucleus of spinal cord and segments C1-C3, tractility.15 In addition, the repolarization of ventricular muscle
C7-C8, and T1-T4,3 where preganglionic axons advance to form can be influenced by atrial and ventricular ganglia.16
synapses with the sympathetic postganglionic neurons of the
intrathoracic ganglia (left and right stellate ganglia, cranial tho-
racic sympathetic chain ganglia, middle and superior cervical
ganglia and mediastinal ganglia) and the intrinsic cardiac ganglia.4 Autonomic Regulation of Pacemaker Activity
In addition to this efferent component, the extrinsic nervous
system contains afferent neurons that transmit mechanosensitive The rhythmic and spontaneous contraction of the heart is initi-
and chemosensitive information from the local environment of ated by the periodic electrical discharges of the SAN. The control
393
of the autonomic nervous system over the SAN is essential for

the regulation of the heart rate.
Pulmonary Vein Ganglia and Heart
Parasympathetic stimulation produces negative chronotropic Rate Control
effects mediated by the release of acetylcholine (ACh) from para-
sympathetic postganglionic neurons, launching a second mes- Neuroanatomy
senger signaling cascade resulting in the modification of ion
channels activities, negative regulation of cAMP production, and The pulmonary vein ganglia (PVG) are located at the roots of
positive regulation of cAMP hydrolysis.17 The neurotransmitter the pulmonary veins and form a circuit via interconnecting nerve
ACh binds to the muscarinic receptor M2 and activates the inhibi- fibers (Figure 39-1). They have received special attention because
tory G-protein, whose αi subunit inhibits adenylyl cyclase activ- they might have a role in promoting pathophysiological condi-
ity, reduces intracellular cAMP levels and protein kinase A (PKA) tions such as atrial fibrillation (AF).26
activity, and produces an overall decrease of the funny current Recent studies have focused on the macroscopic and micro-
(If) and the L-type calcium current (ICa-L).18,19 However, the βγ- scopic neuroanatomy of the PVs and have provided detailed
subunit directly activates the ACh-sensitive inward-rectifier descriptions of nerve distribution and characteristics. Chevalier
potassium current (IKACh). In pacemaker cells, the integration of et al27 found that, in human hearts, nerve fibers and ganglia have
these effects leads to a relatively more hyperpolarized maximal distinct distribution patterns in the PVs, with a higher nerve
diastolic potential, slower phase 4 depolarization, and smaller density at the ostia compared with the distal PVs, and are more
action potential amplitude, producing a decrease in the rate of abundant epicardially, rather than endocardially. In addition, Tan
discharge of the SAN.17 et al28 performed immunostaining of 192 PV-atrial segments
Conversely, sympathetic stimulation increases the SAN rate from eight human hearts using anti–tyrosine hydroxylase and
of discharge via activation of β-adrenergic receptors, enhancing anti–choline acetyltransferase antibodies to label adrenergic and
the activity of several ion channels as well as intracellular calcium cholinergic elements, respectively. They found a similar
release and cycling.17 Activation of β1 receptors stimulates adeny- d noted adrenergic and cholinergic immunofluorescence colocal-
lyl cyclase activity and results in an increase in intracellular cAMP ization in approximately 90% of ganglia. A significant proportion
concentration, activating PKA, which phosphorylates membrane of ganglia (30%) expressed both tyrosine hydroxylase–positive
proteins and enhances ICa-L, If, the slow and rapid delayed rectifier and choline acetyltransferase–positive cells simultaneously. More
currents (IKs, and IKr), and the sodium calcium exchange current.20- recently, Vaitkevicius et al29 investigated in detail the character-
23
Collectively, these effects increase the pacemaker cell’s action istics and distribution of the neural routes by which autonomic
potential upstroke velocity, decrease the action potential dura- nerves supply the human PVs in 35 intact (nonsectioned) left
tion, increase the slope of diastolic depolarization, and conse- atrial-PV complexes stained with the Karnovsky–Roots acetyl-
quently increase the SAN activation rate.24 cholinesterase precipitation reaction. They found that three
MPV Posterior
OTrV surface
LPV
LSPV
Anterior OTrV
surface LIPV
OTrV
Inferior
surface
LCV
LD
Inferior
CS surface
Inferior surface
LV of left atrium
1 mm
A B 1mm
Figure 39-1. Neuroanatomical characterization of pulmonary veins (PVs) whole-mount preparations. A, A macrophotograph illustrating the neuroanatomy of the left dorsal
neural subplexus at the base of the PVs in a mouse heart stained histochemically for acetylcholinesterase. White arrowheads point to the intrinsic cardiac ganglia, and black
arrowheads indicate topographically comparable nerves at the coronary sinus. Dashed lines demarcate limits of the heart hilum. B, A macrophotograph of the inferior
surface of human embryonic left PVs showing the course of epicardial ganglionated nerves (extending from the middle and left dorsal neural subplexi to the PV roots).
Dotted line indicates limits of the cardiac hilum; black arrowheads indicate epicardial ganglia. CS, coronary sinus; LCV, left cranial (left azygos) vein; LPV, left pulmonary vein;
LV, left ventricle; MPV, middle pulmonary vein; LD, left dorsal subplexus; LIPV, left inferior pulmonary vein; LSPV, left superior pulmonary vein; OTrV, orifice of tributaries of
the pulmonary vein.
(Modified from Rysevaite K, Saburkina I, Pauziene N, et al: Morphologic pattern of the intrinsic ganglionated nerve plexus in mouse heart. Heart Rhythm 8:448–454, 2011.)
Pulmonary Vein Ganglia and the Neural Regulation of the Heart Rate 395
epicardial subplexi located at the inferior portion of PVs are the posterior atrial ganglia (found in a fat pad on the rostral surface
39
sole source of nerve supply (nerves extend to the PVs from the of the right atrium) have a modest role in producing vagal bra-
cardiac ganglionated plexus only), whereas free sensory nerve dycardia,33 but can mediate more pronounced parasympathetic
endings aggregate subendothelially. Based on the correlation effects.36,37 A third fat pad located between the medial superior
between the areas of epicardial ganglia and the number of somas vena cava and aortic root appears to be the “head station” of vagal
they contain, it was estimated that approximately 8000 intrinsic fibers that project to both atria.38
nerve cells (2000 associated with each PV) contribute to the In humans, RF ablation of PVG produces a bradycardic
neural control of PVs in humans.29 response,39,40 and PVG can be stimulated by high-frequency
Nerve fibers originating from PVG can extend to different trains of pulses to elicit a decrease in heart rate.26 The bradycardic
regions of the heart. Puodziukynas et al.30 examined in sheep, the responses have been attributed to an evoked vagal reflex, but the
long-term effects of radiofrequency ablation of PVG on the link between these ganglia and SAN has not been fully elucidated.
structure of epicardial nerves located distally from the ablation Hou et al41 investigated the interactions between the extrinsic
sites. Ablation of PVG resulted in the degeneration of remote and intrinsic cardiac nervous system in the context of SAN modu-
epicardial nerves after 2 to 3 months. There was disorganization lation. They proposed a complex interaction between the RAGP
in the neural structures of the dorsal left atrium, coronary sinus, and the superior left ganglionated plexus (SLGP), located adja-
ventricle, and atrioventricular node. These experiments sug- cent to the base of the left superior pulmonary vein between the
gested that there could be anatomical links between the PVG and left atrial appendage and the left pulmonary artery. They also
areas distal to the PVs surrounding. Based on that information, suggested that the main neural pathway between the left vago-
mouse whole-mount atrial preparations31 were used to show that sympathetic trunk and the SAN traverses the SLGP and RAGP
PVG form a circuit via interconnecting nerve fibers. Most impor- sequentially before proceeding to the SAN, and the stimulation
tant, nerves emerged from the PV ganglionic circuit and advanced of the right and left vagosympathetic chains produced a negative
toward the SAN area and innervate it. The data demonstrated chronotropic effect (measured as the mean of 20 beats after stim-
that there was a direct neuroanatomical communication between ulation) reflecting only a parasympathetic response.
the PVG and the SAN. We studied the effects of PVG high-frequency stimulation of
different train durations (200 to 2000 ms) on SAN activity modu-
lation in the isolated murine heart.31 The immediate effect of
Control of Heart Rate PVG stimulation was a significant increase in the mean P-P
interval of the first three beats after stimulation (Figure 39-2, A),
It is well known that the SAN can be regulated by intrinsic whereas PVG stimulation under autonomic blockade with 1 µM
ganglia located in its vicinity, and this regulation has been atropine and 0.5 µM propranolol abolished all PVG stimulation
assumed to be parasympathetic.10,32-34 Indeed, experimental effects (see Figure 39-2, D). High-frequency stimulation of the
studies have shown that ganglia of the fat pad located near the right atrial appendage did not affect the cardiac cycle length (see
right superior pulmonary vein, at the junction of the right atrium Figure 39-2, B, E). When these data were plotted as a percentage
and superior vena cava (right atrial ganglionated plexus [RAGP]), of increase in P-P interval with respect to pre-stimulation versus
mediate a selective negative chronotropic effect.35 Similarly, the the train duration, the effects of PVG stimulation were more
500 Pre PVG stimulation 500 Pre RAA stimulation PVG stimulation
90
Post PVG stimulation Post RAA stimulation PVG stimulation A + P
∆P-P interval (%)
400 400
P-P interval (ms)
P-P interval (ms)
* 70
* * * ** ** **
300 * * 300 50 **
* *
30 * *
200 200
10
100 100
–10
0 0 –30
200 500 700 1000 1250 1500 2000 200 500 700 1000 1250 1500 2000 200 500 700 1000 1250 1500 2000
A Train duration (ms) B Train duration (ms) C Train duration (ms)
500 Pre PVG stimulation A + P 500 Pre RAA stimulation A + P RAA stimulation (PVG group)
90
Post PVG stimulation A + P Post RAA stimulation A + P RAA stimulation (PVG A + P)
∆P-P interval (%)
400 400
P-P interval (ms)
P-P interval (ms)
70
300 300 50
30
200 200
10
100 100 –10
0 0 –30
200 500 700 1000 1250 1500 2000 200 500 700 1000 1250 1500 2000 200 500 700 1000 1250 1500 2000
D Train duration (ms) E Train duration (ms) F Train duration (ms)
Figure 39-2. Effects of pulmonary vein ganglia (PVG) stimulation heart rate. A, The average of three P-P intervals after stimulation at the different frequencies used, before
stimulation, and immediately after stimulation (*P < 0.05 vs. prestimulation; n = 12). B, High-frequency stimulation in the right atrial appendage (RAA) produced no effect
on heart rate (not significant vs. prestimulation; n = 23), ensuring that the parameters used were subthreshold for atrial cells but sufficient to stimulate the ganglia, as shown
in A. D, Autonomic blockade with 1 µM atropine (A) and 0.5 µM propranolol (P) abolished all PVG stimulation effects (n = 16). High-frequency stimulation applied in the
RAA produced no effect on heart rate (D, NS vs. prestimulation; n = 16). C and F, The P-P interval percentage of increase with respect to prestimulation in each train dura-
tion. *P < 0.05 vs. PVG stimulation A + P; #P < 0.05 vs. 200 ms train. Error bars indicate SEM.
pronounced with the longest trains of stimulation (see Figure stimulation caused downward (68.2% of cases) or upward (31.8%)
39-2, C). displacement of the earliest activation site, where the origin of
Analysis of the time course of beat-to-beat changes during the the activation shifted by 0.80 ± 0.23 mm upward or 1.14 ±
first 20 intervals following high frequency stimulation showed 0.34 mm downward with respect to control (Figure 39-4). These
that upon PVG with 200-ms trains, the initial slowing of heart results are consistent with the dominant parasympathetic effects
rate was followed by a progressive return toward baseline (Figure of PVG.
39-3). Autonomic blockade with atropine and propranolol abol- The mechanism by which pacemaker shifts occur after para-
ished PVG stimulation effects. Therefore, these experiments sympathetic stimulation has been studied using numerical simu-
indicate that the SAN can be directly modulated by remotely lations.51 It was proposed that the SAN is composed of electrically
located ganglia of the PVs. coupled oscillators (pacemaker cells) with different intrinsic
firing rates. Through reciprocal phase-dependent interactions,
the coupled oscillators mutually entrain, resulting in the emer-
Modulation of Sinoatrial Node Activation Pattern gence of an origin of activation. Those mutually entrained oscil-
lators, or pacemaker cells, respond to exogenous perturbations
The location of the SAN earliest site of activation is not fixed. It such as ACh through changes in maximum diastolic potential,
can shift to different zones in response to stimuli such as sympa- action potential duration, and cycle length. In the simulations by
thetic and parasympathetic stimulation, changes in temperature, Michaels et al,51 ACh caused a downward shift in the dominant
pharmacologic agents, and modifications in extracellular ions. pacemaker region, in addition to an increase in the cycle length
For example, in vivo studies in the dog reported that stimula- of the array. By the 10th beat, the array returned to its control
tion of the stellate ganglia and the vagus nerve cause cranial and pattern and rate of activation, which is similar to our experimen-
caudal shifts, respectively, of the earliest activation site.42 tal findings31 (see Figures 39-3, 39-4). It has also been suggested
Schuessler et al43 reported that vagal stimulation, from either the that nerve stimulation can produce a pacemaker shift from the
right or left vagosympathetic trunk, caused a downward shift in center to the periphery of the SAN because of electrophysiologi-
the earliest activation site. In isolated rabbit and canine SAN cal and neuroanatomical heterogeneities inherent to the
preparations, it has been shown that the site of earliest activation SAN.44,46,52-56
shifted downward or upward or remained the same after nerve
stimulation.43-46 This finding is in accord with what has been
shown recently in isolated rabbit and murine SAN preparations
subjected to direct high-frequency stimulation or isoproterenol– Pulmonary Veins Ganglia
ACh administration.47-50 and Atrial Fibrillation
Optical mapping of whole-heart preparations was used to
examine the role of PVG stimulation on the modulation of SAN It is accepted that the PVs are an important source of ectopic
cycle length and pattern of activation. In these experiments, PVG beats, initiating frequent paroxysms of AF.57 It has also been
postulated that a possible mechanism by which ectopic foci arise
from the PVs involves the enhanced activity of their intrinsic
70 PVG stimulation nerves58.
PVG stimulation A + P The role of the PVs and the associated ganglionated plexus as
50
a trigger of AF has received considerable attention during the last
∆P-P interval (%)
few years. Lemola et al59 showed that intact PVs are not needed
for the maintenance of experimental cholinergic AF, and they
30
proposed that the PVG, not the PVs themselves, are important
in vagally mediated AF promotion. Neural factors are believed
10
to play a key role in the initiation of paroxysmal AF. Although
both limbs of the autonomic nervous system have a role in AF,
–10 cholinergic stimulation is thought to be the main factor for spon-
taneous AF initiation.60 To determine the relative role of the
–30 intrinsic cardiac nerve activity in triggering atrial arrhythmias
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
and AF, Choi et al61 performed continuous in vivo nerve record-
Interbeat interval ings from extrinsic cardiac nerve activity (vagus nerve and left
Figure 39-3. Time-course effects (20 beats after stimulation) of pulmonary vein stellate ganglion) and the ICNS (superior left ganglionated plexi
ganglia stimulation with a 200-ms high-frequency stimulation trains. Error bars at the base of the left superior pulmonary vein and Ligament of
indicate SEM. Marshall) in dogs submitted to rapid atrial pacing. It was found
SVC SVC
PVs PVs
RAA RAA
A B
1 mm Before stimulation 1 mm After stimulation
Figure 39-4. Modulation of the sinoatrial node activation pattern induced by pulmonary vein ganglia (PVG) stimulation in isolated mouse heart preparations. The leading
pacemaker site shifted upward in 32% (7 of 22) and downward in 68% (15 of 22) of observations, after applying 1000 ms train of stimulation to the PVG. Panels A and B
show isochrone maps of representative example of a downward shift after PVG stimulation. The site of earliest activation is depicted in red. Isochrone lines: 1 ms. SVC,
Superior vena cava; PVs, pulmonary veins; RAA, right atrial appendage.
Pulmonary Vein Ganglia and the Neural Regulation of the Heart Rate 397
that activity of the ICNS always preceded the onset of paroxysmal parasympathetic effects on the sinus node and atrial myocardium,
39
atrial tachyarrhythmias, suggesting that these intrinsic ganglia suggesting that reinnervation occurs after ganglionated plexi
are an invariable trigger of paroxysmal AF and atrial tachycardias. ablation. More recently, Puodziukynas et al30 determined the
Once initiated, paroxysmal AF leads to atrial remodeling through long-term effects of RFA at the roots of the PVs on the structure
alteration of electrical and structural properties of the atria, of epicardial nerves located distally from the PVs. After 2 to 3
which facilitates maintenance and recurrence of AF.62 months, alterations in the structure and degeneration of remote
Radiofrequency ablation of the PVG in human and experi- atrial and ventricular nerves occurred, which suggests that long-
mental studies can result in successful denervation and preven- term autonomic dysfunction is a potential risk of RFA of PVG.
tion of AF inducibility.39,63 Circumferential PVs ablation, which In humans, an upregulation of nerve growth factor expression has
produces a concomitant autonomic denervation by PVG destruc- been reported after RFA. Although RFA did not increase trans-
tion, seems to correlate well with a lower recurrence of AF epicardiac nerve growth factor immediately after the procedure,
sodes.39 Lu et al61 showed that high-frequency stimulation applied significant systemic increases were observed on the first day after
to PVG can initiate PV rapid firings and promote paroxysmal AF, the procedure, and they persisted for at least 2 days.71 Further-
and ablation of the ganglionated plexuses eliminated those rapid more, ablation of free nerves can result in reinnervation after a
firings, suggesting an important role of the ICNS in the genesis few months, and ablation of ganglia can cause denervation hyper-
of AF in structurally normal hearts. sensitivity and autonomic dysfunction.29
Selective ablation of intrinsic cardiac ganglia in the PVs using Alternative approaches to ablation of the PVG are being
radiofrequency ablation (RFA) has been proposed as a single developed. Li et al72 and Shen et al73 found that AF can be sup-
treatment of patients with AF40,64-66 or in combination with PV pressed by low-level, high-frequency stimulation of vagosympa-
isolation.67,68 These studies showed that RFA of selected sites thetic trunks. On the other hand, Yu et al74 demonstrated that
where high-frequency stimulation induces vagal reflexes can the function of ganglionated plexuses can be suppressed with the
prevent AF recurrence64-66; however, anatomic PVG ablation intravascular administration of magnetic nanoparticles contain-
yields significantly lower success rates compared with the coming the neurotoxic agent N-isopropyl acrylamide monomer. Con-
bined PVG and PV ablation.65,67 sequently, targeted drug delivery or inhibitory low-level,
Although studies have demonstrated the efficacy of PVG abla- high-frequency stimulation can achieve autonomic denervation
tion, other factors must be considered. Studies in dogs showed and could potentially be used as means to treat AF without the
that nerve sprouting occurred within 2 hours after RFA in the risks associated with ablation.
right atrium, and sprouting persisted for at least 1 month after
the intervention, even at remote sites in relation to the place of
RFA69. Sakamoto et al70 demonstrated that removal of the epicar-
dial fat pads and ganglionated plexi ablation greatly reduced the Acknowledgment
effects of vagal stimulation in the atrial conduction system and
the atrial myocardium. Four weeks later, there was a return of Supported in part by NIH grant ROOHL 105574 to SFN.
Am J Physiol Heart Circ Physiol 279:H1201– 21. Brown HF, DiFrancesco D, Noble SJ: How does
References H1207, 2000. adrenaline accelerate the heart? Nature 280:235–
11. Armour JA, Kember GC: Cardiac Sensory neurons. 236, 1979.
1. Marron K, Wharton J, Sheppard MN, et al: Dis- In Armour JA, Ardell JL, editors: Basic and Clinical 22. Kameyama M, Hoffman F, Trautwein W: On the
tribution, morphology, and neurochemistry of Neurocardiology, New York, 2004, Oxford Uni- mechanisms of β-adrenergic regulation of the Ca
endocardial and epicardial nerve terminal arboriza- versity Press, pp 79–117. channel in guinea pig heart. Pflugers Arch
tions in the human heart. Circulation 92:2343– 12. Chen J, Wasmund SL, Hamdan MH: Back to the 405:285–293, 1985
2351, 1995. future: The role of the autonomic nervous system 23. Zhang YH, Hancox JC: Regulation of cardiac Na+-
2. Armour JA: Potential clinical relevance of the ‘little in atrial fibrillation. Pacing Clin Electrophysiol Ca2+ exchanger activity by protein kinase
brain’ on the mammalian heart. Exp Physiol 29:413–421, 2006. phosphorylation–still a paradox? Cell Calcium
93:165–176, 2008. 13. Gatti PJ, Johnson TA, Phan P, et al: The physio- 45:1–10, 2009.
3. Kawashima T: The autonomic nervous system of logical and anatomical demonstration of function- 24. DiFrancesco D: The role of the funny current in
the human heart with special reference to its origin, ally selective parasympathetic ganglia located in pacemaker activity. Circ Res 106:434–446, 2010.
course, and peripheral distribution. Anat Embryol discrete fat pads on the feline myocardium. J Auton 25. Rysevaite K, Saburkina I, Pauziene N, et al: Mor-
(Berl) 209:425–438, 2005. Nerv Syst 51:255–259, 1995. phologic pattern of the intrinsic ganglionated
4. Armour JA: Myocardial ischaemia and the cardiac 14. Gatti PJ, Johnson TA, Massari VJ: Can neurons in nerve plexus in mouse heart. Heart Rhythm 8:448–
nervous system. Cardiovasc Res 41:41–54, 1999. the nucleus ambiguus selectively regulate cardiac 454, 2011.
5. Armour JA: Cardiac neuronal hierarchy in health rate and atrio-ventricular conduction? J Auton 26. Nishida K, Maguy A, Sakabe M, et al: The role of
and disease. Am J Physiol Regul Integr Comp Nerv Syst 57:123–127, 1996. pulmonary veins vs. Autonomic ganglia in different
Physiol 287:R262–R271, 2004. 15. Yuan BX, Ardell JL, Hopkins DA, et al: Differen- experimental substrates of canine atrial fibrillation.
6. Pauza DH, Pauziene N, Pakeltyte G, et al: Com- tial cardiac responses induced by nicotine sensitive Cardiovasc Res 89:825–833, 2011.
parative quantitative study of the intrinsic cardiac canine atrial and ventricular neurones. Cardiovasc 27. Chevalier P, Tabib A, Meyronnet D, et al: Quanti-
ganglia and neurons in the rat, guinea pig, dog and Res 27:760–769, 1993. tative study of nerves of the human left atrium.
human as revealed by histochemical staining for 16. Cardinal R, Page P, Vermeulen M, et al: Spatially Heart Rhythm 2:518–522, 2005.
acetylcholinesterase. Ann Anat 184:125–136, 2002. divergent cardiac responses to nicotinic stimulation 28. Tan AY, Li H, Wachsmann-Hogiu S, et al: Auto-
7. Pauza DH, Skripka V, Pauziene N: Morphology of of ganglionated plexus neurons in the canine heart. nomic innervation and segmental muscular discon-
the intrinsic cardiac nervous system in the dog: A Auton Neurosci 145:55–62, 2009. nections at the human pulmonary vein-atrial
whole-mount study employing histochemical 17. Mangoni ME, Nargeot J: Genesis and regulation junction: Implications for catheter ablation of
staining with acetylcholinesterase. Cells Tissues of the heart automaticity. Physiol Rev 88:919–982, atrial-pulmonary vein junction. J Am Coll Cardiol
Organs 172:297–320, 2002. 2008. 48:132–143, 2006.
8. Arora RC, Waldmann M, Hopkins DA, et al: 18. DiFrancesco D, Ducouret P, Robinson RB: Mus- 29. Vaitkevicius R, Saburkina I, Rysevaite K, et al:
Porcine intrinsic cardiac ganglia. Anat Rec A carinic modulation of cardiac rate at low acetylcho- Nerve supply of the human pulmonary veins: An
Discov Mol Cell Evol Biol 271:249–258, 2003. line concentrations. Science 3;243:669–671, 1989 anatomical study. Heart Rhythm 6:221–228, 2009.
9. Batulevicius D, Skripka V, Pauziene N, et al: 19. Fischmeister R, Hartzell HC: Mechanism of action 30. Puodziukynas A, Kazakevicius T, Vaitkevicius R, et
Topography of the porcine epicardiac nerve plexus of acetylcholine on calcium current in single cells al: Radiofrequency catheter ablation of pulmonary
as revealed by histochemistry for acetylcholinester- from frog ventricle. J Physiol 376:183–202, vein roots results in axonal degeneration of distal
ase. Auton Neurosci 138:64–75, 2008. 1986. epicardial nerves. Auton Neurosci 167:61–65,
10. Tsuboi M, Furukawa Y, Nakajima K, et al: Inotro- 20. Trautwein W, Kameyama M: Intracelullar control 2012.
pic, chronotropic, and dromotropic effects medi- of calcium and potassium currents in cardiac cells. 31. Zarzoso M, Rysevaite K, Milstein M, et al: Nerves
ated via parasympathetic ganglia in the dog heart. Jpn Heart J S1:31–50, 1986. projecting from the intrinsic cardiac ganglia of the
pulmonary veins modulate sinoatrial node pace- the rabbit sinus node. J Pharmacol Exp Ther spontaneous atrial fibrillation in dogs. J Am Coll
maker function. Cardiovascular Research [Epub 214:417–422, 1980. Cardiol 43:483–490, 2004.
ahead of print] 2013. 46. Vinogradova TM, Fedorov VV, Yuzyuk TN, et al: 61. Choi EK, Shen MJ, Han S, et al: Intrinsic cardiac
32. Lazzara R, Scherlag BJ, Robinson MJ, et al: Selec- Local cholinergic suppression of pacemaker activ- nerve activity and paroxysmal atrial tachyarrhyth-
tive in situ parasympathetic control of the canine ity in the rabbit sinoatrial node. J Cardiovasc Phar- mia in ambulatory dogs. Circulation 121:2615–
sinoatrial and atrioventricular nodes. Circ Res macol 32:413–424, 1998. 2623, 2010.
32:393–401, 1973. 47. Fedorov VV, Hucker WJ, Dobrzynski H, et al: 62. Shen MJ, Choi EK, Tan AY, et al: Neural mecha-
33. Mick JD, Wurster RD, Duff M, et al: Epicardial Postganglionic nerve stimulation induces temporal nisms of atrial arrhythmias. Nat Rev Cardiol 9:30–
sites for vagal mediation of sinoatrial function. Am inhibition of excitability in rabbit sinoatrial node. 39, 2012.
J Physiol 262:H1401–H1406, 1992. Am J Physiol Heart Circ Physiol 291:H612–H623, 63. Lu Z, Scherlag BJ, Lin J, et al: Autonomic mecha-
34. Gray AL, Johnson TA, Ardell JL, et al: Parasym- 2006. nism for initiation of rapid firing from atria and
pathetic control of the heart. Ii. A novel intergan- 48. Abramochkin DV, Kuzmin VS, Sukhova GS, et al: pulmonary veins: Evidence by ablation of gangli-
glionic intrinsic cardiac circuit mediates neural Modulation of rabbit sinoatrial node activation onated plexi. Cardiovasc Res 84:245–252, 2009.
control of heart rate. J Appl Physiol 96:2273–2278, sequence by acetylcholine and isoproterenol inves- 64. Scanavacca M, Pisani CF, Hachul D, et al: Selective
2004. tigated with optical mapping technique. Acta atrial vagal denervation guided by evoked vagal
35. Randall WC, Ardell JL, Wurster RD, et al: Vagal Physiol (Oxf) 196:385–394, 2009. reflex to treat patients with paroxysmal atrial fibril-
postganglionic innervation of the canine sinoatrial 49. Fedorov VV, Chang R, Glukhov AV, et al: Complex lation. Circulation 114:876–885, 2006.
node. J Auton Nerv Syst 20:13–23, 1987. interactions between the sinoatrial node and 65. Pokushalov E, Romanov A, Artyomenko S, et al:
36. Randall DC, Brown DR, McGuirt AS, et al: Inter- atrium during reentrant arrhythmias in the canine Ganglionated plexi ablation for longstanding per-
actions within the intrinsic cardiac nervous system heart. Circulation 122:782–789, 2010. sistent atrial fibrillation. Europace 12:342–346,
contribute to chronotropic regulation. Am J 50. Glukhov AV, Fedorov VV, Anderson ME, et al: 2010.
Physiol Regul Integr Comp Physiol 285:R1066– Functional anatomy of the murine sinus node: 66. Mikhaylov E, Kanidieva A, Sviridova N, et al:
R1075, 2003. High-resolution optical mapping of ankyrin-b het- Outcome of anatomic ganglionated plexi ablation
37. Johnson TA, Gray AL, Lauenstein JM, et al: Para- erozygous mice. Am J Physiol Heart Circ Physiol to treat paroxysmal atrial fibrillation: A 3-year
sympathetic control of the heart. I. An 299:H482–H491, 2010. follow-up study. Europace 13:362–370, 2011.
interventriculo-septal ganglion is the major source 51. Michaels DC, Matyas EP, Jalife J: Mechanisms of 67. Katritsis DG, Giazitzoglou E, Zografos T, et al:
of the vagal intracardiac innervation of the ven- sinoatrial pacemaker synchronization: A new Rapid pulmonary vein isolation combined with
tricles. J Appl Physiol 96:2265–2272, 2004. hypothesis. Circ Res 61:704–714, 1987. autonomic ganglia modification: A randomized
38. Chiou CW, Eble JN, Zipes DP: Efferent vagal 52. Mabe AM, Hoover DB: Structural and functional study. Heart Rhythm 8:672–678, 2011.
innervation of the canine atria and sinus and atrio- cardiac cholinergic deficits in adult neurturin 68. Ohkubo K, Watanabe I, Okumura Y, et al: Com-
ventricular nodes. The third fat pad. Circulation knockout mice. Cardiovasc Res 82:93–99, 2009. bined effect of pulmonary vein isolation and abla-
95:2573–2584, 1997. 53. Beau SL, Hand DE, Schuessler RB, et al: Relative tion of cardiac autonomic nerves for atrial
39. Pappone C, Santinelli V, Manguso F, et al: densities of muscarinic cholinergic and beta- fibrillation. Int Heart J 49:661–670, 2008.
Pulmonary vein denervation enhances long-term adrenergic receptors in the canine sinoatrial node 69. Okuyama Y, Pak HN, Miyauchi Y, et al:
benefit after circumferential ablation for paroxys- and their relation to sites of pacemaker activity. Nerve sprouting induced by radiofrequency cath-
mal atrial fibrillation. Circulation 109:327–334, Circ Res 77:957–963, 1995. eter ablation in dogs. Heart Rhythm 1:712–717,
2004. 54. Roberts LA, Slocum GR, Riley DA: Morphological 2004.
40. Pokushalov E, Romanov A, Shugayev P, et al: study of the innervation pattern of the rabbit sino- 70. Sakamoto S, Schuessler RB, Lee AM, et al: Vagal
Selective ganglionated plexi ablation for paroxys- atrial node. Am J Anat 185:74–88, 1989. denervation and reinnervation after ablation of
mal atrial fibrillation. Heart Rhythm 6:1257–1264, 55. Opthof T: The mammalian sinoatrial node. Car- ganglionated plexi. J Thorac Cardiovasc Surg
2009. diovasc Drugs Ther 1:573–597, 1988. 139:444–452, 2010.
41. Hou Y, Scherlag BJ, Lin J, et al: Interactive atrial 56. Boyett MR, Honjo H, Kodama I: The sinoatrial 71. Kangavari S, Oh YS, Zhou S, et al: Radiofrequency
neural network: Determining the connections node, a heterogeneous pacemaker structure. Car- catheter ablation and nerve growth factor concen-
between ganglionated plexi. Heart Rhythm 4:56– diovasc Res 47:658–687, 2000. tration in humans. Heart Rhythm 3:1150–1155,
63, 2007. 57. Haïssaguerre M, Jaïs P, Shah DC, et al: Spontane- 2006.
42. Goldberg JM: Intra-sa-nodal pacemaker shifts ous initiation of atrial fibrillation by ectopic beats 72. Li S, Scherlag BJ, Yu L, et al: Low-level vagosym-
induced by autonomic nerve stimulation in the originating in the pulmonary veins. N Engl J Med pathetic stimulation: A paradox and potential new
dog. Am J Physiol 229:1116–1123, 1975. 339:659–666, 1998 modality for the treatment of focal atrial fibrilla-
43. Schuessler RB, Bromberg BI, Boineau JP: Effect of 58. Patterson E, Po SS, Scherlag BJ, et al: Triggered tion. Circ Arrhythm Electrophysiol 2:645–651,
neurotransmitters on the activation sequence of the firing in pulmonary veins initiated by in vitro auto- 2009.
isolated atrium. Am J Physiol 258:H1632–H1641, nomic nerve stimulation. Heart Rhythm 2:624– 73. Shen MJ, Shinohara T, Park HW, et al: Continu-
1990. 631, 2005. ous low-level vagus nerve stimulation reduces stel-
44. Shibata N, Inada S, Mitsui K, et al: Pacemaker shift 59. Lemola K, Chartier D, Yeh YH, et al: Pulmonary late ganglion nerve activity and paroxysmal atrial
in the rabbit sinoatrial node in response to vagal vein region ablation in experimental vagal atrial tachyarrhythmias in ambulatory canines. Circula-
nerve stimulation. Exp Physiol 86:177–184, 2001. fibrillation: Role of pulmonary veins versus auto- tion 123:2204–2212, 2011.
45. Mackaay AJ, Op’t Hof T, Bleeker WK, et al: Inter- nomic ganglia. Circulation 117:470–477, 2008. 74. Yu L, Scherlag BJ, Dormer K, et al: Autonomic
action of adrenaline and acetylcholine on cardiac 60. Sharifov OF, Fedorov VV, Beloshapko GG, et al: denervation with magnetic nanoparticles. Circula-
pacemaker function. Functional inhomogeneity of Roles of adrenergic and cholinergic stimulation in tion 122:2653–2659, 2010.
Neural Activity and Atrial
Tachyarrhythmias 40
Peng-Sheng Chen, Lan S. Chen, and Shien-Fong Lin
pathways. Sympathetic nerves are located primarily around blood

CHAPTER OUTLINE
vessels and between myocytes. The nerve fibers are oriented
Cardiac Nerves 399 along the long axis of myocytes. The cardiac branches of the
vagus nerve, which are preganglionic fibers, make synaptic con-
Neural Activity and Atrial Tachyarrhythmias 400
nections with ganglion cells in the ganglionated plexi (the intrin-
Persistent Atrial Fibrillation 403 sic cardiac nervous system). Cardiac nerves can be demonstrated
by labeling nerve-specific markers such as S100 protein (marker
Neural Activity and Ventricular Rate Control
of Schwann cells), neurofilament, synaptophysin, protein gene
During Persistent Atrial Fibrillation 405 product 9.5, and various regulatory neuropeptides (e.g.,
Neuromodulation for Atrial Tachyarrhythmia Control 406 neuropeptide-Y) using immunohistochemistry techniques.
Sympathetic nerves can be identified by immune-labeling tyro-
Conclusions 407 sine hydroxylase (TH) and parasympathetic nerves by acetylcho-
linesterase or cholineacetyltransferase (ChAT).
Atrial tachyarrhythmia, including atrial fibrillation (AF), is a Intrinsic Cardiac Nerves

major public health problem. Many studies in animal models and
in human patients have suggested that the activities of the auto- In addition to the extrinsic cardiac nerves, the heart is also richly
nomic nervous system has an important role in the generation innervated by an extensive intrinsic cardiac nervous system.1,2
and maintenance of atrial tachyarrhythmias. However, the mech- The intrinsic cardiac nervous system includes sensory, intercon-
anisms by which autonomic activation induce atrial tachyarrhyth- necting, and autonomic neurons that communicate with each
mias remain poorly understood. This gap in knowledge is in part other and with the extrinsic cardiac nervous system. The nerve
due to the limited availability of information on the anatomical structures of the intrinsic cardiac nerves are found in various parts
structures of the autonomic nerves that innervate the heart, the of the heart, but mostly in the ganglionated plexi within epicar-
general absence of information on spontaneous autonomic nerve dial fat pads. Among the ganglionated plexi, the right-atrial gan-
discharges in ambulatory animals or humans and a limited avail- glionated plexi innervates the sinus node, whereas the inferior
ability of animal model of spontaneous atrial tachyarrhythmias. vena cava–inferior atrial ganglionated plexi (at the junction of
The latter limitations have prevented the investigators from inferior vena cava and the left atrium) innervates the atrioven-
studying the temporal relationship between autonomic nerve dis- tricular node. Another region that is richly innervated is located
charges and spontaneous atrial tachyarrhythmias. This chapter at the pulmonary vein (PV)-left atrium (LA) junction. Radiofre-
summarizes the data obtained over the past few years in the quency catheter ablation at these sites can potentially result in
understanding of the anatomy and physiology of the autonomic successful denervation and prevent the inducibility of AF.3
nerves, and attempts to relate the neural activities to the genera- However, preserving (rather than ablating) the anterior epicardial
tion and maintenance of atrial tachyarrhythmias. There will also fat pad during coronary arterial bypass surgery decreases inci-
be a brief discussion of the use of neuromodulation in the preven- dence of postoperative atrial fibrillation.
tion and treatment of atrial arrhythmias.
Coexistence of Sympathetic and Parasympathetic

Nerves in the Same Structure
Cardiac Nerves
A common misunderstanding of the autonomic nervous system
Extrinsic cardiac nerves is that the nerve structures are either sympathetic or parasympa-
thetic. For example, the term vagal tone is generally used to
The preganglionic sympathetic nerves that innervate the heart describe the level of activity in the parasympathetic nervous
arise in the upper four or five segments of thoracic spinal cord. system. Vagal denervation was used to describe the successful
These preganglionic sympathetic nerves pass through white rami elimination of bradycardiac responses during catheter ablation of
communicantes, enter the sympathetic trunk, and terminate in AF. The fact, however, is that the vagus nerves and almost all
the superior cervical ganglion, the middle cervical ganglion (if other cardiac nerve structures contain both sympathetic (adren-
present), and the cervicothoracic (stellate) ganglion. These ergic) and parasympathetic (cholinergic) components. It is not
ganglia give off cardiac nerves that join with the cardiac branches possible to stimulate or ablate one branch of the autonomic
of the vagus nerve and form cardiac plexus. The cardiac plexus, nervous system without affecting the other. Tan et al4 performed
which is divided into a superficial (ventral) and deep (dorsal) immunostaining of tissues from the human PV-LA junction. The
cardiac plexus, then gives branches of the coronary and atrial authors found that adrenergic and cholinergic nerves coexist in
plexuses to innervate the heart. The sympathetic nerves are dis- all ganglionated plexi. It is also possible for the same neuron to
tributed in the superficial epicardial layer throughout most sur- express both TH and ChAT. These findings indicate that it is
faces and penetrate into myocardium along coronary arterial impossible to target either sympathetic or parasympathetic nerves
399
A B
C D
E F
Figure 40-1. Immunocytochemical staining of the cervical vagus nerve. A and C, Examples of the nerves sectioned transversely. Other panels show nerves sectioned
longitudinally (craniocaudal). Each cervical vagus nerve contains multiple parallel nerve bundles, most staining positively (brown) for cholineacetyltransferase (ChAT), as
shown in A and B. However, a small percentage of the nerve bundles, primarily at the periphery of the nerves, stained positively for tyrosine hydroxylase (TH; arrows in C
and D). E, In addition, TH-positive ganglion cells are also present in the nerves. F, These cells were ChAT negative. These findings indicate that bundles of sympathetic
nerves are present within the cervical vagus nerves. The presence of sympathetic ganglion cells in the vagus nerve suggests that the cervical vagus nerve is also an impor-
tant source of sympathetic innervation. A to C, Original magnification ×100. D to F, Original magnification ×40.
(From Park HW, Shen MJ, Han S, et al: Neural control of ventricular rate in ambulatory dogs with pacing induced sustained atrial fibrillation. Circ Arrhythm Electrophysiol
5:571–580, 2012.)
selectively during radiofrequency catheter-ablation procedures. parasympathetic nerve structure. Therefore, vagal tone includes
Sympathetic nerve fibers are also present in the thoracic vagus both sympathetic and parasympathetic components.
nerve.5 More recently, Park et al.6 performed TH and ChAT
staining of the left cervical vagus nerve (Figure 40-1). ChAT
positive nerve structures formed a majority of the cervical vagus
nerve (see Figure 40-1, A, B). However, a small amount of Neural Activity and Atrial Tachyarrhythmias
TH-positive nerves were also present at the edge of the nerve
bundles (see Figure 40-1, C, D). Unexpectedly, the authors iden- Recording Neural Activities in Ambulatory Animals
tified sympathetic neurons in the vagus nerve (see Figure 40-1,
E), indicating that the cervical vagus nerve was a source of sym- Jung et al7 continuously recorded the activity of stellate ganglia
pathetic innervation. The same neurons stained negative for in healthy dogs for an average of 41.5 days and documented that
ChAT (see Figure 40-1, F). The presence of both TH-positive both the heart rate and the stellate ganglion nerve activity
neurons and TH-positive nerve fibers is consistent with the (SGNA) showed a circadian variation. Ogawa et al5 and Tan et al8
notion that the vagus nerve is a mixed sympathetic and then applied the same methods to record vagus nerve activity
(VNA). These earlier studies showed several findings about nerve contractions and nonsustained ventricular tachycardia.9 A second
40
discharges that were previously unknown. First of all, there are important observation is that, in ambulatory animals, the nerve
fundamentally two different types of nerve activities (Figure structures often activate either simultaneously or alternatively,
40-2). The vast majority of the nerve activities were the low- suggesting a close coordination among the nerve activities from
amplitude burst discharge activities (LABDA), with an amplitude different parts of the autonomic nervous system. For example,
less than 0.2 mV and variable duration. A second type of nerve the left and right stellate ganglion usually activate together.7
activity is the high-amplitude spike discharge activity (HASDA) Similarly, the VNA can activate with SGNA.5,8 The VNA can
with amplitude of greater than 0.2 mV (average, 1.4 mV). There also activate simultaneously with the ganglionated plexi.11-13
is usually a nearly isoelectric interval between the spikes, with Figure 40-3, A, shows simultaneous discharges of the right and
obvious depolarization shifts in some of the episodes. The left stellate ganglion in a normal dog. Figure 40-3, B, shows
HASDA has a frequency of approximately 6.6 Hz, and there is simultaneous recording of both extrinsic and intrinsic nerve
an average of 6.7 spikes per run.5 HASDA episodes were rare, activities in a dog with intermittent rapid atrial pacing. Note that
with an average of approximately 15 episodes per 24 hours. both extrinsic nerve activities (SGNA and VNA) activated
However, when they were observed, there was a high likelihood together with one of the intrinsic nerve structure (superior left
of both atrial and ventricular arrhythmias.5,9,10 Figure 40-2, A, ganglionated plexus, SLGPNA), but not the ligament of Marshall
shows examples of LABDA and HASDA in a normal dog. Note ganglionated plexus. The VNA and SLGPNA activation patterns
that LABDA in the SGNA can accelerate the heart rate. HASDA were almost mirror images of each other, suggesting that these
usually occurs during LABDA and can further accelerate the two structures closely coordinate with each other. Another
heart rate. Figure 40-2, B, shows that a HASDA episode imme- important finding is that the SGNA in Figure 40-3, A, resulted
diately precedes the premature atrial contraction in a dog with in less apparent heart rate acceleration than that shown in Figure
pacing-induced heart failure.5 In another study, we have observed 40-2, whereas in Figure 40-3, B, there was sinus rate acceleration
multiple episodes of HASDA-induced premature ventricular associated with SGNA. Subsequent studies showed that right
LABDA
0.2
SGNA
120 bpm 162 bpm

1
ECG
1 2 3 4 5 6 7 8 9 10 s
0.6 HASDA
SGNA
0
168 bpm 192 bpm
2
ECG
A 1 2 3 4 5 6 7 8 9 10 s
2.5
ECG
–2.5
0.02
VNA
–0.02
0.3
SGNA
–0.3
B 0 2 4 6 8 10 12 14 16 18 20 (sec)
Figure 40-2. Two types of nerve activities in ambulatory dogs. A, Patterns of nerve activities form a normal dog. The upper panel shows low-amplitude burst discharge
activity (LABDA), which accounts for the vast majority of nerve activities in all nerve structures. In this example, LABDA accelerated the heart rate. The lower panel shows
high-amplitude spike discharge activity (HASDA), which further accelerated the heart rate. Units for stellate ganglion nerve activity and electrocardiogram (ECG) are given
in millivolts (mV) in this and all other figures. B, Patterns of nerve activities from a dog with pacing-induced heart failure. The premature atrial contraction (arrow on ECG
channel) was preceded immediately by simultaneous sympathovagal discharges. The sympathetic nerve activity show spiky discharges (HASDA).
(A, From Zhou S, Jung BC, Tan AY, et al: Spontaneous stellate ganglion nerve activity and ventricular arrhythmia in a canine model of sudden death. Heart Rhythm 5:131–139,
2008. B, From Ogawa M, Tan AY, Song J, et al: Cryoablation of extrinsic cardiac sympathetic nerves markedly reduces atrial arrhythmias in ambulatory dogs with pacing-induced
heart failure. Heart Rhythm 5:S54, 2008.)
ECG
0
–8
Left SGNA
0
–5
0
Right SGNA
3
0
0 30
A Seconds
(mv)
0
ECG
–0.2
0.025
LOMNA
–0.025
0.01
SLGPNA
–0.01
0.1
SGNA
–0.1
0.01
VNA
–0.01
B 4 sec
Figure 40-3. Coordinated activation among different nerve structures. A, An ambulatory dog with simultaneous recording of left and right stellate ganglion nerve activity
(SGNA). The arrow points to the onset of right SGNA, which slightly preceded onset of left SGNA. B, Simultaneous recording of both extrinsic and intrinsic nerve activities.
Note the similarities between the nerve activities recorded by the superior left ganglionated plexus nerve activity (SLGPNA) and the vagus nerve activity (VNA). LOMNA,
Ligament of Marshall nerve activity.
(A, From Jung BC, Dave AS, Tan AY, et al: Circadian variations of stellate ganglion nerve activity in ambulatory dogs. Heart Rhythm 3:78–85, 2006. B, From Choi E-K, Shen MJ,
Han S, et al: Intrinsic cardiac nerve activity and paroxysmal atrial tachyarrhythmia in ambulatory dogs. Circulation 121:2615–2623, 2010.)
anterior ganglionated plexus has an important role in heart rate innervations and extensive atrial nerve sprouting. Such neural
control.13 Therefore, recording extrinsic nerve activity alone remodeling could, in turn, promote the electrical remodeling
might not be sufficient in determining the mechanisms of heart caused by rapid atrial pacing.15 However, because human AF is
rate control in ambulatory animals. The complex interactions not induced by electrical stimulations, it is not known whether
among different autonomic nerve structures is one of the mecha- this model adequately simulates human AF remains.
nisms by which heart rate variability measurements in general fail
to accurately predict the instantaneous sympathetic and parasym-
pathetic nerve activities. In dogs with heart failure, the relation- Simultaneous Sympathovagal Discharges and
ship between nerve discharge and heart rate control is further Paroxysmal Atrial Tachyarrhythmias
uncoupled because of the sinus node dysfunction.13 Therefore,
there is little relationship between the heart rate variability To determine whether autonomic nerve discharges preceded par-
parameters and the actual autonomic nerve discharge patterns in oxysmal AF and other atrial tachyarrhythmias, Tan et al8
heart failure.14 implanted a pacemaker and a radiotransmitter in dogs to simul-
taneously record nerve activities of the left stellate ganglion and
left vagal nerve as well as a surface electrocardiogram (ECG) over
Canine Models of Atrial Tachyarrhythmias a period of several weeks. The authors then performed intermit-
tent rapid atrial pacing and monitored the nerve activity when
We used two different canine models to study the spontaneous the pacemaker was turned off. They found that there is a circa-
atrial arrhythmias. One is rapid ventricular pacing to induce heart dian variation of the frequencies of atrial tachyarrhythmias
failure. A significant amount of spontaneous atrial tachyarrhyth- (Figure 40-4, A), similar to that found in human patients with
mias are observed in this heart failure model.5 The second model symptomatic paroxysmal AF. They found that simultaneous sym-
uses intermittent rapid atrial pacing to cause electrical remodel- pathovagal discharges are a common trigger for premature atrial
ing and paroxysmal atrial tachyarrhythmias. This model is par- contractions (see Figure 40-4, B) and the most frequent trigger
ticularly suitable for the study of neural activation because it is of paroxysmal atrial tachyarrhythmias (see Figure 40-4, C).8
associated with a heterogeneous increase of sympathetic Cryoablation of bilateral stellate ganglia and of the superior
80
40
No of episodes
60
40
20
0
12 am–4 am 4 am–8 am 8 am–12 pm 12 pm–4 pm 4 pm–8 pm 8 pm–12 am
A PAC PAT PAF
LA 8
(mV) 0
.02
VNA
(mv) –.02
.08
SGNA
(mV) –.08 2 sec
B
LA 2.0
(mV) –2.0
.05
VNA
(mv) –.05
0.6
SGNA
(mV) –0.6
C 2 sec
Figure 40-4. Paroxysmal atrial arrhythmias in dogs with intermittent rapid atrial pacing. A, Circadian incidence of paroxysmal arrhythmias (PAC, PAT, and PAF combined)
over a 24-hour period. B, The arrow points to the PAC. C, PAT induced by simultaneous sympathovagal discharge. PAC, premature atrial contraction; PAF, paroxysmal atrial
fibrillation; PAT, paroxysmal atrial tachycardia.
(From Tan AY, Zhou S, Ogawa M, et al: Neural mechanisms of paroxysmal atrial fibrillation and paroxysmal atrial tachycardia in ambulatory canines. Circulation 118:916–925,
2008.)
cardiac branches of the left vagus nerve eliminated all episodes

Persistent Atrial Fibrillation
of paroxysmal AF and atrial tachycardias, indicating a causal
relationship between ANS activity and the generation of parox- The mechanism by which some patients develop persistent (sus-
ysmal atrial tachyarrhythmias. Similarly, in a canine model of tained) AF remains unclear. For patients with paroxysmal AF,
pacing-induced heart failure, simultaneous sympathovagal dis- approximately half of them progress to sustained AF after 25 years
charge was the most frequent trigger of atrial tachyarrhythmias,5 of follow-up.19 However, the time-to-progression varies consid-
which could be prevented by cryoablation of the stellate ganglion erably among individuals who progress from paroxysmal to sus-
and the T2-T4 thoracic sympathetic ganglia.16 Choi et al11 tained AF. Furthermore, many patients with chronic AF do not
recorded both left extrinsic nervous system activity (SGNA and have documented paroxysmal AF before diagnosis. It is apparent
VNA) and intrinsic nerve activity (the superior left ganglionated that there are large individual variations in the susceptibility to
plexi and the ligament of Marshall). They found that the vast progression of AF. Intermittent rapid atrial pacing in large
majority of atrial tachyarrhythmia episodes were preceded by animals can initially induce paroxysmal AF. However, if pacing
simultaneous discharges of both extrinsic and intrinsic nervous continues, sustained AF is induced.8, 11 Rapid pacing causes short-
systems, whereas a small percentage (10% to 20%) of episodes ening of the effective refractory periods. However, the time
was preceded by intrinsic nerve activity alone without the par- course of changes in atrial refractoriness did not exactly parallel
ticipation of the extrinsic nervous system. In all dog studies, the development of sustained AF, indicating that other factors
intrinsic cardiac nerve activities invariably preceded atrial tachyar- might also be important in the progression to sustained AF. We
rhythmia episodes. An example is shown in Figure 40-5. The analyzed long-term recordings of nerve activities in ambulatory
importance of intrinsic cardiac nervous system in generating par- dogs to determine the duration of intermittent rapid atrial pacing
oxysmal atrial tachyarrhythmias is further supported by Nishida needed to induce sustained AF (>48 hours).12 We found that
et al,17 who reported that ganglionated plexi ablation reduced the there are two differential patterns of interactions among cardiac
inducibility of AF in dogs with rapid atrial pacing. Another inter- autonomic structures (Figure 40-6, A). Among them, dogs with
esting observation in this study11 is that the activities of intrinsic a linear sympathovagal correlation (group 1) nerves are associated
cardiac nerves might contaminate local atrial electrograms, with a faster development of sustained AF than those with
resulting in recordings similar to that of complex fractionated L-shaped sympathovagal correlation (see Figure 40-6, B). Figure
atrial electrograms. These findings might explain the clinical 40-6, C and D, shows the correlation between VNA and SLGPNA
efficacy of ablative therapy that targets those intrinsic cardiac for group 1 and group 2 dogs, respectively. Figure 40-6, E and
ganglia18 or sites with complex fractionated atrial electrograms. F, shows examples of nerve activities that correspond to Figure
(mV)
Sinus arrhythmia Atrial tachycardia a
0.2 0.16 P = 0.07 mV
ECG
–0.2
1.0 a b
LA
–1.0
0.02
LOMNA –0.08
–0.02 0.16 P = 0.12 mV
0.08
SLGPNA –0.08
0.2
SGNA
–0.2
–0.08
0.02
VNA –0.02
b 400 ms
A 4 sec
(mV)
Sinus arrhythmia Atrial tachycardia
0.03
ECG
–0.03
1.0
LA
–1.0
LOMNA 0.1
–0.1
0.03
SLGPNA
–0.03
0.2
SGNA
–0.2
0.03
VNA –0.03
B 4 sec
Figure 40-5. Induction of PAT by extrinsic and intrinsic cardiac nerve activities. A, An example in which ICNA occurred before ECNA and a PAT episode. The magnified
pseudo-electrocardiogram shows the different P wave morphologies during sinus rhythm (Aa) and PAT (Ab). B, Simultaneous ICNA and SGNA leading to the onset of PAT.
ECNA, Extrinsic cardiac nerve activity; ICNA, intrinsic cardiac nerve activity; PAT, paroxysmal atrial tachycardia; SGNA, stellate ganglion nerve activity.
HR acceleration
0.5
2500 500 ECG
–0.5
SGNA (mV-s)
VNA (mV-s)
0.2
SGNA
1250 250 –0.2
0.04
VNA
–0.04
0 0 0.01
SLGPNA
0 250 500 0 125 250 –0.01
A VNA (mV-s) C SLGPNA (mV-s) E (mV) 4 sec
HR deceleration
0.2
2500 200 ECG
–0.2
SGNA (mV-s)
VNA (mV-s)
0.4 a
SGNA
1250 100 –0.4
b
0.04
VNA
–0.04
0.03
0 0 SLGPNA
0 250 500 0 100 200 –0.03
(mV) 10 sec
B VNA (mV-s) D SLGPNA (mV-s) F
Figure 40-6. Patterns of autonomic interactions. A, Representative SGNA-VNA scatter plot of a group 1 dog. Each dot represents an SGNA-VNA pair of nerve activity inte-
grated over 1 minute. The entire plot has 1440 data points to cover in a 24-hour period. B, Representative SGNA-VNA scatter plot from a group 2 dog. C, Representative
VNA-SLGPNA scatter plot from a group 1 dog. D, Representative VNA-SLGPNA scatter plot from a group 2 dog. E, An example of simultaneous sympathovagal coactivation
(black arrows) observed in a group 1 dog that led to heart rate acceleration. The arrowhead shows independent SLGPNA. F, An example of a recording from a group 2 dog
showing that simultaneously increased VNA and SLGPNA (black arrows) resulted in heart rate deceleration. ECG, Electrocardiogram; SGNA, stellate ganglion nerve activity;
VNA, vagal nerve activity; SLGPNA, superior left ganglionated plexus nerve activity.
(From Shen MJ, Choi EK, Tan AY, et al: Patterns of baseline autonomic nerve activity and the development of pacing-induced sustained atrial fibrillation. Heart Rhythm 8:583–589,
2011.)
40-6, A and B, respectively. Group 1 dogs had more paroxysmal also known that the inferior vena cava–inferior atrial ganglionated
40
atrial tachycardias at baseline and faster induction of sustained plexus (IVC-IAGP) is important in modulating AV node conduc-
AF by rapid atrial pacing compared with group 2 dogs. These tion, and that direct electrical stimulation of this GP can slow VR
findings show that baseline nerve activity patterns can predict the during AF in human patients.20 However, none of these studies
durations needed to induce sustained AF. Different forms of was performed in the ambulatory state with direct nerve record-
sympathovagal discharge patterns might also be present in human ing. Therefore, the relative importance of right vagal nerve activ-
patients, and differential ANS discharge patterns may predeter- ity (RVNA), left vagal nerve activity (LVNA), and IVC-IAGP
mine which patients will be at greater risks of progression from nerve activity (IVC-IAGPNA) in VR control during AF in
paroxysmal to sustained AF. ambulatory animals remains poorly understood. To fill this gap
in knowledge, Park et al6 recorded bilateral cervical VNA
and IVC-IAGPNA during baseline sinus rhythm and during
pacing-induced sustained AF in six ambulatory dogs. Integrated
Neural Activity and Ventricular Rate Control nerve activities and average VR were measured every 10 seconds
during Persistent Atrial Fibrillation over 24-hour periods. It was found that that the LVNA was associ-
ated with VR reduction during AF in five of six dogs and RVNA
In most patients with AF, rate control is not inferior to rhythm in one of six dogs. Figure 40-7 shows typical examples. Figure
control as a management strategy. However, the mechanisms of 40-7, A, shows that five dogs showed VR reduction with com-
ventricular rate (VR) control during AF remain unclear. It is also bined LVNA and IVC-IAGPNA. Figure 40-7, B, is from one dog
generally accepted that autonomic nervous system inputs, espe- showing VR reduction with combined RVNA discharge and IVC-
cially the vagal tone, are important in modulating the AV node IAGPNA discharge. Figure 40-7, C, shows IVC-IAGP discharge
conduction. Left vagal nerve stimulation has been proposed as a alone, without other autonomic nerve activity, is sufficient to
method to control VR during AF. In addition to vagal nerves, it is cause transient AV conduct delay. Figure 40-7, D, shows that
0.05
RVNA –0.05
0.02
LVNA
–0.02
IVC- 0.04
IAGPNA –0.04
4
LEGM
–4
A Slow VR Slow VR Slow VR Slow VR
0.04
RVNA
–0.04
0.02
LVNA
–0.02
IVC- 0.03
IAGPNA –0.03
2
LEGM
–2 Slow VR
B
0.04
RVNA
–0.04
0.02
LVNA
–0.02
IVC- 0.06
IAGPNA –0.06
4
LEGM
–4 Slow VR Slow VR Slow VR Slow VR
C
0.2
RVNA
–0.2
0.02
LVNA –0.02
0.3
IVC-
IAGPNA –0.3
4
LEGM
–4 Slow VR Rapid VR
D (mV) 5 sec
Figure 40-7. Right vagal nerve activity (RVNA), left vagal nerve activity (LVNA), inferior vena cava–inferior atrial ganglionated plexus nerve activity (IVC-IAGPNA), and local
electrograms (LEGM) during sustained atrial fibrillation. IVC-IAGPNA with LVNA (A) or RVNA (B) coactivation was associated with reduced VR. C, Independent IVC-IAGP was
associated with slow VR. D, RVNA activation after IVC-IAGPNA withdrawal was associated with rapid VR. The LEGM shows ventricular electrograms and T waves.
(From Park HW, Shen MJ, Han S, et al: Neural control of ventricular rate in ambulatory dogs with pacing-induced sustained atrial fibrillation. Circ Arrhythm Electrophysiol
5:571–580, 2012.)
RVNA activation induces rapid heart rates, suggesting selective 1 V below the threshold needed to reduce heart rate is known to
activation of the sympathetic fibers within the right cervical vagal be effective in suppressing AF induction in open-chest–anesthe-
nerve. When RVNA is not firing, the IVC-IAGP activation tized dogs.24, 25 We hypothesize that vagal stimulation could
slowed the VR. These studies show that IVC-IAGPNA is invari- achieve its antiarrhythmic effects by suppressing sympathetic
ably associated with VR reduction during AF. In comparison, outflow to the heart. To test this hypothesis, we implanted a
right or left VNA was associated with VR reduction only when it neurostimulator in 12 dogs to stimulate left cervical vagus nerve
coactivated with the IVC-IAGPNA. These studies also suggest and a radiotransmitter for continuous recording of left SGNA,
that vagus nerves do not directly innervate the AV node; rather, it left thoracic VNA, and ECGs. Group 1 dogs (n = 6) underwent
activates IVC-IAGP to control the VR during AF. 1 week of continuous LL-VNS. Group 2 dogs (n = 6) underwent
intermittent rapid atrial pacing followed by active or sham
LL-VNS on alternate weeks. We found that integrated SGNA
was significantly reduced during LL-VNS in group 1. The reduc-
Neuromodulation for Atrial tion was most apparent at 8:00 am, along with a significantly
Tachyarrhythmia Control reduced heart rate (Figure 40-8). LL-VNS did not change
VNA. We also found that LL-VNS causes structural remodeling
Animal studies suggest that increasing the vagal tone may of the left stellate ganglion. Normal stellate ganglion naturally
be beneficial for controlling heart failure and ventricular contains both TH-positive and TH-negative ganglion cells. The
arrhythmias.21,22 Spinal cord stimulation, which enhances para- density of TH-negative nerves in the left stellate ganglion 1 week
sympathetic activity, improves ventricular function and reduces after cessation of LL-VNS were significantly more than that in
ventricular arrhythmias in a canine postinfarction heart failure normal control dogs. The frequencies of paroxysmal atrial fibril-
model.23 While most of these previous studies used vagal stimula- lation and tachycardia during active LL-VNS were significantly
tion with stimulus strength sufficient to reduce heart rate, low- lower than when the LL-VNS was turned off. These findings
level vagus nerve stimulation (LL-VNS) with stimulus strength show that LL-VNS suppresses SGNA and reduces the incidences
SGNA SGNA
12
†
10 * 25
8 20
mV-s
6 15
mV-s
4 10
2 5
0 0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
A B Hour
HR HR
100
80
100
60 90
Bpm
Bpm
40 80
20 70
0 60
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
C D Hour
Baseline Baseline
During VNS During VNS
After VNS After VNS
Figure 40-8. Effects of low-level vagus nerve stimulation (LL-VNS) on stellate ganglion nerve activity (SGNA) and heart rate (HR). A, Chronic LL-VNS significantly reduced
SGNA over 24 hours. The SGNA normalized to baseline level after cessation of LL-VNS. B, Hourly averages of SGNA show that the reduction in integrated SGNA was particu-
larly striking at 8:00 AM. All values are averaged over 5 days and six dogs. C, The administration and cessation of chronic LL-VNS did not change the overall heart rate. D,
Hourly averages of HR reveal that the morning surge of HR (arrowhead) was markedly attenuated during LL-VNS. *P < 0.05.
(From Shen MJ, Shinohara T, Park HW, et al: Continuous low-level vagus nerve stimulation reduces stellate ganglion nerve activity and paroxysmal atrial tachyarrhythmias in
ambulatory canines. Circulation 123:2204–2212, 2011.)
of paroxysmal atrial tachyarrhythmias in ambulatory dogs. Sig- Neuromodulation can be effective in controlling the atrial
40
nificant neural remodeling of the left stellate ganglion is evident tachyarrhythmias. One method is to ablate the stellate ganglia
1 week after cessation of chronic LL-VNS. partially to reduce the sympathetic outflow and reduce atrial
arrhythmia. A second method is to perform LL-VNS to reduce
SGNA and thereby control atrial arrhythmia. A third method is
to perform radiofrequency catheter ablation around the PV-LA
Conclusions junction to control atrial arrhythmias by modulating the cardiac
intrinsic nervous system.
Autonomic nervous system activation invariably precedes the
onset of paroxysmal atrial tachyarrhythmias, and the preexisting
sympathovagal discharge patterns determines the duration of
rapid pacing needed to induce persistent AF in ambulatory dogs. Acknowledgments
The intrinsic nervous system activation invariably precedes the
onset of atrial tachyarrhythmia. Simultaneous discharges from The authors thank Medtronic (Minneapolis, MN), St. Jude Inc.
the sympathetic and vagal nerves of the extrinsic nervous system (St. Paul, MN), and Cryocath (Houston, TX) for donating the
are also commonly observed before the onset of atrial tachyar- research equipment to the laboratory. Peng-Sheng Chen was a
rhythmias. The extrinsic and intrinsic nervous systems also work consultant for Cyberonics, which manufactures and sells cervical
together to control the VR during sustained AF. However, vagal vagal nerve stimulators. This work was supported by National
nerves do not directly affect the AV conduction. They work Institutes of Health grants P01 HL78931, R01s HL78932, and
through the IVC-IAGP to reduce the ventricular rate during AF. R01 HL71140 and a Medtronic-Zipes Endowment.
arrhythmia in a canine model of sudden death. 18. Pokushalov E, Romanov A, Shugayev P, et al:
References Heart Rhythm 5:131–139, 2008. Selective ganglionated plexi ablation for paroxys-
10. Ogawa M, Tan AY, Song J, et al: Cryoablation of mal atrial fibrillation. Heart Rhythm 6:1257–1264,
1. Armour JA: Potential clinical relevance of the ‘little extrinsic cardiac sympathetic nerves markedly 2009.
brain’ on the mammalian heart. Exp Physiol reduces atrial arrhythmias in ambulatory dogs with 19. Jahangir A, Lee V, Friedman PA, et al: Long-term
93:165–176, 2008. pacing-induced heart failure. Heart Rhythm 5:S54, progression and outcomes with aging in patients
2. Ardell JL: The cardiac neuronal hierarchy and 2008. with lone atrial fibrillation: A 30-year follow-up
susceptibility to arrhythmias. Heart Rhythm 11. Choi E-K, Shen MJ, Han S, et al: Intrinsic cardiac study. Circulation 115:3050–3056, 2007.
2010. nerve activity and paroxysmal atrial tachyarrhyth- 20. Rossi P, Bianchi S, Barretta A, et al: Post-operative
3. Po SS, Nakagawa H, Jackman WM: Localization mia in ambulatory dogs. Circulation 121:2615– atrial fibrillation management by selective
of left atrial ganglionated plexi in patients with 2623, 2010. epicardial vagal fat pad stimulation. J Interv Card
atrial fibrillation. J Cardiovasc Electrophysiol 12. Shen MJ, Choi EK, Tan AY, et al: Patterns of base- Electrophysiol 24:37–45, 2009.
20:1186–1189, 2009. line autonomic nerve activity and the development 21. Zhang Y, Popovic ZB, Bibevski S, et al: Chronic
4. Tan AY, Li H, Wachsmann-Hogiu S, et al: Auto- of pacing-induced sustained atrial fibrillation. vagus nerve stimulation improves autonomic
nomic innervation and segmental muscular discon- Heart Rhythm 8:583–589, 2011. control and attenuates systemic inflammation and
nections at the human pulmonary vein-atrial 13. Shinohara T, Shen MJ, Han S, et al: Heart failure heart failure progression in a canine high-rate
junction: Implications for catheter ablation of decreases nerve activity in the right atrial ganglion- pacing model. Circ Heart Fail 2:692–699, 2009.
atrial-pulmonary vein junction. J Am Coll Cardiol ated plexus. J Cardiovasc Electrophysiol 24(4):404– 22. De Ferrari GM, Schwartz PJ: Vagus nerve stimula-
48:132–143, 2006. 412, 2012. tion: From pre-clinical to clinical application:
5. Ogawa M, Zhou S, Tan AY, et al: Left stellate 14. Piccirillo G, Ogawa M, Song J, et al: Power spec- Challenges and future directions. Heart Fail Rev
ganglion and vagal nerve activity and cardiac tral analysis of heart rate variability and autonomic 16:195–203, 2011.
arrhythmias in ambulatory dogs with pacing- nervous system activity measured directly in 23. Lopshire JC, Zhou X, Dusa C, et al: Spinal cord
induced congestive heart failure. J Am Coll Cardiol healthy dogs and dogs with tachycardia-induced stimulation improves ventricular function and
50:335–343, 2007. heart failure. Heart Rhythm 6:546–552, 2009. reduces ventricular arrhythmias in a canine postin-
6. Park HW, Shen MJ, Han S, et al: Neural control 15. Lu Z, Scherlag BJ, Lin J, et al: Atrial fibrillation farction heart failure model. Circulation 120:286–
of ventricular rate in ambulatory dogs with pacing begets atrial fibrillation: Autonomic mechanism for 294, 2009.
induced sustained atrial fibrillation. Circ Arrhythm atrial electrical remodeling induced by short-term 24. Li S, Scherlag BJ, Yu L, et al: Low-level vagosym-
Electrophysiol 5:571–580, 2012. rapid atrial pacing. Circ Arrhythm Electrophysiol pathetic stimulation: A paradox and potential new
7. Jung BC, Dave AS, Tan AY, et al: Circadian varia- 1:184–192, 2008. modality for the treatment of focal atrial fibrilla-
tions of stellate ganglion nerve activity in ambula- 16. Ogawa M, Tan AY, Song J, et al: Cryoablation of tion. Circ Arrhythm Electrophysiol 2:645–651,
tory dogs. Heart Rhythm 3:78–85, 2006. stellate ganglia and atrial arrhythmia in ambulatory 2009.
8. Tan AY, Zhou S, Ogawa M, et al: Neural mecha- dogs with pacing-induced heart failure. Heart 25. Yu L, Scherlag BJ, Li S, et al: Low-level vagosym-
nisms of paroxysmal atrial fibrillation and paroxys- Rhythm 6:1772–1779, 2009. pathetic nerve stimulation inhibits atrial fibrillation
mal atrial tachycardia in ambulatory canines. 17. Nishida K, Maguy A, Sakabe M, et al: The role of inducibility: Direct evidence by neural recordings
Circulation 118:916–925, 2008. pulmonary veins vs. Autonomic ganglia in different from intrinsic cardiac ganglia. J Cardiovasc Elec-
9. Zhou S, Jung BC, Tan AY, et al: Spontaneous stel- experimental substrates of canine atrial fibrillation. trophysiol 22:455–463, 2010.
late ganglion nerve activity and ventricular Cardiovasc Res 89:825–833, 2011.
Sympathetic Innervation, Denervation,
and Cardiac Arrhythmias 41
Marmar Vaseghi, Olujimi A. Ajijola, Aman Mahajan,
and Kalyanam Shivkumar
the cardiac plexus. The cardiac plexus, which is divided into the
CHAPTER OUTLINE
superficial (ventral) and deep (dorsal) cardiac plexus, gives rise to
Anatomy of Cardiac Sympathetic Innervation 409 sympathetic nerve branches that run along the coronary arteries
on the epicardium and penetrate the myocardium with the vascu-
Sympathetic Modulation of Myocardial Excitability 409
lature. Thus, the nerves are located primarily around blood vessels
Neural Remodeling in the Presence of Myocardial and between myocytes oriented along their long axis. Further, a
Pathology 410 gradient in sympathetic innervation is thought to exist from base
to apex, as well as from epicardium to endocardium.7,8
Proarrhythmic Effects of Sympathetic Stimulation 412 Data collected over the past four decades indicate that exten-
Effect of Neuraxial Modulation in Reducing Risk sive neural processing occurs in the cardiac plexus, otherwise
of Ventricular Arrhythmias 413 known as the intrinsic cardiac nervous system. This plexus contains
afferent neurons, interconnecting interneurons, and sympathetic
Surgical Sympathetic Denervation in Humans 414 and parasympathetic efferent postganglionic neurons.9 Further-
Conclusion 415 more, it is composed of seven ganglionic subplexi, containing
more than 800 epicardial ganglia. Each subplexus or group of
subplexi innervates different chambers of the heart. One sub-
plexus innervates the right ventricle, three innervate the left ven-
The autonomic nervous system plays a significant role in the tricle, and the rest innervate the atria. The highest density of
genesis and maintenance of ventricular arrhythmias, and power- epicardial ganglia, approximately 50%, exists near the hilum of
fully modulates the underlying substrate in a dynamic manner.1-3 the heart, especially on the dorsal and dorsolateral surfaces of the
Blockade of the sympathetic nervous system, whether through left atrium. The number of neurons in these ganglia decreases
medications or neuraxial modulation, has been associated with a with age, from approximately 94,000 neurons in neonates and
reduction in the risk of sudden cardiac death and the burden of children to 43,000 intrinsic neurons in the plexus of adults.10 A
ventricular arrhythmias. In this chapter, the anatomy of the complex feedback regulatory system allows the cardiac nervous
cardiac sympathetic nervous system and pathological changes system to modulate sympathetic and parasympathetic input to the
associated with neural remodeling in the setting of scar and car- heart.
diomyopathy will be reviewed. Subsequently, the role of sympa- Cardiac nerves can be demonstrated by labeling nerve-specific
thetic activation in promoting arrhythmias, and the effects of markers such as S100 protein, neurofilament, and synaptophysin
neuraxial modulation in reducing the burden of these arrhyth- using immunohistochemistry techniques,5,6 and myelinated fibers
mias will be discussed. Finally, cardiac denervation in humans, can be identified by myelin markers, such as myelin basic protein.
including feasibility, surgical techniques, and complications, will Sympathetic nerves can be identified by immunolabeling tyrosine
be considered. hydroxylase. Tyrosine hydroxylase and myelin basic protein stain-
ing have confirmed the presence of myelinated sympathetic
nerves on all surfaces of the epicardium and endocardium.11
Anatomy of Cardiac Sympathetic Innervation

The preganglionic sympathetic neurons that innervate the heart Sympathetic Modulation of
reside in the intermediate zone of the thoracic spinal cord.4 These Myocardial Excitability
preganglionic sympathetic nerves pass through the white rami
communicantes, enter the sympathetic trunk, and terminate in The major neurotransmitter mediating sympathetic response in
the cervicothoracic ganglion as well as in the T2-T4 ganglia. Of the heart is norepinephrine. Epinephrine release from intracar-
note, preganglionic neurons may synapse on neurons within the diac neural endings is negligible.12 Along the length of terminal
ganglia at the same thoracic level or may travel within the sympa- axons are a series of localized swellings known as varicosities. Most
thetic chain and synapse on neurons of ganglia at other spinal of the norepinephrine storage vesicles in a terminal axon are
levels. The preganglionic neurotransmitter within the ganglia is concentrated in these varicosities, which act as specialized sites
acetylcholine. The right and left cervicothoracic ganglia, which of norepinephrine storage and release.13 The overall effect of
are often a product of the fusion of the C8 and T1 ganglia, are norepinephrine release through multiple signaling pathways is
called the left stellate (LSG) and right stellate (RSG) ganglia. LSG shortening of the ventricular action potential duration (APD) and
and RSG, along with the ganglia of T2-T4, give rise to postgan- the refractory period.14 Most norepinephrine undergoes reuptake
glionic axons that target organs including the heart, the esopha- into nerve terminals by the presynaptic norepinephrine trans-
gus, the trachea, and the head and neck (Figure 41-1). porter. A small fraction diffuses into the vascular space, where it
Postganglionic nerve fibers from these ganglia then join the can be measured in coronary sinus blood. Norepinephrine spill-
branches of the vagus nerve to form the left and right cardiac over (both interstitial and in the coronary sinus) can be used to
(cardiopulmonary) nerves, destined for the heart.5,6 These nerves infer that sympathetic outflow to the heart can also be assessed
subsequently form a network of neurons on the epicardium, called in humans.15
409
Brain and
LSG higher centers
T2 Ganglion
LSG
T3 Ganglion Thoracic
Aorta T1-T4 RSG
T4 Ganglion
Spinal
cord
SAN
NE NE
Left
ventricle
Adrenal
medulla
Figure 41-1. The sympathetic chain and in particular the left stellate ganglion (LSG) 2/3 E
and the ganglia of thoracic spinal 1-4 (T1-T4) are located in a paravertebral position
1/3 NE
behind the parietal pleura.
(Reproduced with permission from Abrahams PH, Spratt JD, Boon J: McMinn’s Clinical
Atlas of Human Anatomy, 6th ed, St Louis, 2007, Mosby.)
Kidney
Sympathetic stimulation, predominantly mediated by post- Figure 41-2. Multiple organs mediate sympathetic outflow to the normal heart
synaptic myocardial β-adrenergic receptors, has important effects including the brain, the spinal cord, the sympathetic chain (left stellate ganglion
on chronotropy, dromotropy, lusitropy, and inotropy. Discharge [LSG], right stellate ganglion [RSG], and ganglia of T1-T4), the adrenal medulla by
of the sinoatrial (SA) node and atrioventricular (AV) nodal con- secreting catecholamines, and the renal nerves. E, Epinephrine; LSG, left stellate
duction are augmented, increasing chronotropy. In the atria and ganglion; NE, norepinephrine; RSG, right stellate ganglion; SAN, sinoatrial node;
T1-T4, thoracic spinal level 1 through 4.
ventricles, contractility and relaxation are enhanced. Both β1 and
β2 subtypes are present at a ratio of approximately 5 : 1 in the
healthy human heart.16-18 Alpha adrenoreceptors are mainly
present in the vascular wall, but are also found in ventricular innervation; certain viable sites shorten their repolarization,
myocardium, where they account for approximately 15% of while others show no response. All denervated areas show dener-
cardiac adrenergic receptors.18 vation super-sensitivity, defined as an exaggerated response to
Cardiac sympathetic activation is complex and is mediated norepinephrine infusion.21 The cellular mechanisms for this
through multiple organs at multiple levels. The brain (higher response do not involve differences in the β-adrenergic receptor
centers), brain stem, spinal cord, sympathetic ganglia, adrenal or the α-subunit of stimulatory G-protein density.22,23
medulla, and renal nerves can increase the net sympathetic In patients with recurrent ventricular arrhythmias undergoing
output to the heart, dynamically modulating cardiac function ventricular tachycardia (VT) ablation procedures, activation
(Figure 41-2). recovery interval (ARI) measurements show a reduced response
to indirect sympathetic stimulation via nitroprusside in dense scar
and in the viable peri-infarct myocardium, suggesting denerva-
tion.24 These sites demonstrate an exaggerated response to iso-
Neural Remodeling in the Presence proterenol infusion, suggesting the presence of denervation
of Myocardial Pathology super-sensitivity in humans (Figure 41-3). Further, the response
to sympathetic stimulation is extremely heterogeneous, with a
Myocardial infarction (MI) and heart failure, in addition to greater than 2-fold increase in dispersion in ARI with nitroprus-
leading to scar formation, cause remodeling of cardiac nerves. side infusion.24
Initial denervation, followed by nerve-sprouting and trans- These studies were important because they demonstrated that
differentiation of sympathetic nerves, leads to heterogeneity in transmural MI and heart failure not only can alter the myocardial
repolarization and can modulate reentry. substrate for arrhythmias, they can also disrupt innervation to
histologically viable myocardium, leading to denervation super-
sensitivity and a non-uniform electrophysiological response to
Cardiac Neural Remodeling and Denervation sympathetic stimulation. This contributes to the genesis of ven-
tricular arrhythmias in both acute and chronic MI.
Myocardial infarction causes death of sympathetic fibers within
the scar and loss of efferent sympathetic innervation at nonin-
farcted apical sites.19 Norepinephrine depletion in the scar is Cardiac Neural Remodeling and Nerve Sprouting
accompanied by increased production in noninfarcted basal
areas.20 In response to LSG stimulation, viable myocardial sites After studies showed acute denervation, evidence of nerve sprout-
apical to an infarct show evidence of heterogeneity in ing and heterogeneous hyperinnervation was observed in chronic
Sympathetic Innervation, Denervation, and Cardiac Arrhythmias 411
10 mV
41
Recording
catheter
AICD
CS
RAO
5 mV 2
1 3
4
5
6 7 8
9
10
12 14
15
11 13
Recording
RV catheter
AICD
CS
LAO
0 mV
P = .016
20 P = .08
10
10
Delta ARI (ms)
20
-30 Isoproterenol
-40
Nitroprusside
-50
-60
-70
P = .01
-80
P = .0002
-90
Normal hearts CM-normal site CM-border zones CM-scars

Figure 41-3. A recording multi-electrode catheter on fluoroscopy (left upper panels) and electroanatomical map (right upper panel) in this patient with ischemic cardiomy-
opathy (ICM) and a large anteroapical scar are used to record unipolar electrograms from scar, border zone, and viable myocardium. On the electroanatomical map, the
purple areas represent viable tissue (normal voltage) and gray represents dense scar. All other colors represent border zones (0.5 mV < voltage < 1.5 mV). The delta activa-
tion recovery interval (ARI; change in ARI from baseline) within the cardiomyopathic and normal hearts is shown in the lower panel. Note that in response to isoproterenol,
the delta ARI is greatest in the CM-normal site (viable myocardium) and scar regions of the cardiomyopathic heart. The border zones within each patient are the least
responsive to isoproterenol. On the other hand, in response to nitroprusside, the scar and the CM-NL tissue appear to be the least responsive, even paradoxically increasing
their ARI in comparison with border-zone regions. Therefore, it appears that the most denervated regions have the greatest response to catecholamines, consistent with
denervation super-sensitivity.
MI and heart failure models. Peripheral nerve injury resulting in “swarm-like” pattern in the periphery of necrotic tissues and in
Wallerian degeneration, which leads to nerve growth factor perivascular regions have been observed, and these changes were
(NGF), triggered regeneration via nerve sprouting.25,26 Along more prominent in patients with a history of ventricular
with denervation, MIBG studies have shown localized arrhythmias.29-31 These border zones of infarcts have also been
re-innervation in injured myocardium in both ischemic and shown to be frequent sites of origin of inducible ventricular
nonischemic cardiomyopathy patients.27,28 In human hearts with tachycardia (VT)/ventricular fibrillation (VF).32,33 Furthermore,
cardiomyopathy, local increases in sympathetic nerves in a nerve sprouting can occur in a noninfarct setting such as stem
cell transplantation,34 in radiofrequency ablation,35 and in rapid and nerve sprouting has been reported.40 Increased stellate gan-
pacing–induced heart failure in dogs36 and in hypercholesterol- glion nerve activity (SGNA) immediately after MI was associated
emic rabbits, where it has been shown to cause QTc dispersion, with intramyocardial nerve sprouts, as well as increased neuronal
increased heterogeneity of repolarization, and significantly size and synaptic density in the LSG and RSG.47 Although SCD
increased episodes of VF, both spontaneous and induced.37 Fur- was not observed in this study, Ogawa et al and Zhou et al had
thermore, NGF infusion into the LSG promotes nerve sprouting previously shown that sympathetic nerve discharges tend to
in dogs with MI and complete AV block, and is associated with precede ventricular arrhythmias in the same dog model of MI.48,49
increased incidence of VF.38 Infusion of NGF into the RSG has Similarly, in human cadavers with evidence of cardiac scar, neu-
not been shown to increase the risk of sudden cardiac death ronal number in the LSG is increased as compared with cadavers
(SCD).39 Of note, NGF and GAP-43 levels are increased in the without evidence of scar.50 Compared with cadavers with normal
LSG of these dogs 3 days after MI, without a concomitant hearts, those with cardiomyopathy demonstrate increased neuro-
increase in mRNA levels, indicating possible retrograde trans- nal size and synaptophysin density (Figure 41-4).51 These results
portation of these proteins to the LSG, which then triggers nerve provide a possible mechanistic link between neural remodeling
sprouting at noninfarcted LV sites.40 (nerve sprouting within the myocardium) and remodeling of the
Heart failure can cause trans-differentiation of cardiac sym- stellate ganglion and ventricular arrhythmias, although further
pathetic nerves. Cholinergic trans-differentiation of nerve sprouts studies are required. Interaction between areas of denervation,
by production of interleukin-6 cytokines through a gp-130 sig- regional nerve sprouting and trans-differentiation (neural remod-
naling pathway has been demonstrated in rodents,41,42 potentially eling), electrical remodeling due to heart failure, and electroana-
further promoting heterogeneity in repolarization. However, the tomical remodeling of the stellate ganglia all combine to create
exact ramifications of this sympathetic rejuvenation and plasticity a substrate that can be conducive for ventricular arrhythmias and
are yet unknown. Heart failure is also known to cause remodeling sudden cardiac death (Figure 41-5).
of cardiac ion channels, including increased L-type Ca (ICaL)
density, decreased potassium current, decreased Ito density, and
changes in Cl and Ca transporters and enzymes in the border
zones surrounding the infarct.43-46 Thus, sympathetic stimulation Proarrhythmic Effects of
could result in complex effects on the APD and restitution, Sympathetic Stimulation
which, along with increased ICaL density, can lead to intracellular
Ca2+ overload–induced triggered activity, potentiating the risk of Many studies have suggested that sympathetic nerve stimulation,
spontaneous ventricular arrhythmias. particularly LSG stimulation, is proarrhythmic. Greater increases
in the amplitude of early afterdepolarization have been observed
with LSG stimulation.52 Priori et al showed that LSG stimulation
Extracardiac Neural Remodeling caused delayed afterdepolarizations in vivo in cat hearts, suggest-
ing triggered activity as the mechanism of ventricular arrhyth-
In addition to cardiac neural remodeling, electroanatomical mogenesis.53 Further, an increase in dispersion with sympathetic
remodeling of the LSG in the setting of MI and heart failure has stimulation, both during ischemia and in normal canine and
been described. Increased nerve density along with increased porcine hearts, has been observed.54-57 In the porcine heart, LSG
mRNA levels of NGF and GAP-43 in the LSG of dogs with MI stimulation increased ARI dispersion by 4-fold and caused VF in
NL ICM NICM
400
Mean neuronal area (µm2)
350
300
250
NL ICM NICM
Figure 41-4. Human cardiomyopathy is associated with stellate ganglion neuronal hypertrophy. Shown in the figure are representative thionine-stained right stellate
ganglion neurons from normal controls (NL) and from patients with ischemic (ICM) and nonischemic cardiomyopathy (NICM). Neurons from ICM and NICM are larger and
show a more granular appearance compared with normal controls. (Magnification 20×; scale bar 50 µm.)
Stellate ganglion
41
remodeling
es
t o kin
Cy GF EADs
N +
Myocardial Scar & Sympathetic
infarction denervation activation
NG
Cy F DADs
to +
kin VT/VF
es
Heterogenous Dispersion of
substrate repolarization
Normal Scar
Nerve
sprouts
Figure 41-5. The schematic shows the various changes that occur as a result of myocardial infarction that cause neural remodeling and increased sympathetic activation.
Myocardial infarction causes scar formation and denervation leading to release of nerve growth factor (NGF) and cytokines by the surrounding myocardium. This causes
remodeling of the stellate ganglion and neural remodeling of the myocardium, including nerve sprout formation and cholinergic trans-differentiating, further causing
heterogeneity in repolarization with sympathetic stimulation. Sympathetic activation then causes EADs, DADs, and increased dispersion in this substrate, leading to ven-
tricular arrhythmias. DAD, delayed afterdepolarization; EAD, early afterdepolarization; NGF, neural growth factor.
25% of the animals.57 In open-chest dogs, electrical stimulation cord, the sympathetic chain, and the β-adrenergic receptors of
of LSG, the left middle cervical or left caudal pole of the cardio- the myocardium, and it can be attempted by renal denervation
pulmonary nerve, or the ventrolateral nerve caused VT in 13 of (Figure 41-6). General anesthesia, sedation, and intubation can
22 normal dog hearts, with isochronal mapping showing the suppress or significantly reduce the burden of ventricular arrhyth-
earliest electrical excitation occurring on the posterior aspect of mias and implantable cardioverter-defibrillator (ICD) shocks by
the ventricles.58 Electrical stimulation of the left ansa subclavia reducing the sympathetic drive in electrical storm.62-65 At the
in open-chest dogs during left circumflex occlusion increased the spinal cord level, Issa et al demonstrated that in a canine model
incidence of VF from 35% to 73%.59 The incidence of induced of ischemic cardiomyopathy, spinal cord stimulation at T1-T2
ventricular arrhythmias in dogs with myocardial ischemia segments reduced the incidence of ventricular arrhythmia from
increased from 54% to 68% and 63% with LSG and bilateral 59% to 23% during ischemia.66 A simultaneous decrease in heart
stellate stimulation, respectively. RSG stimulation had no signifi- rate and systolic blood pressure, consistent with the antisympa-
cant effect on the incidence of arrhythmias in this study.60 Further, thetic effects of spinal cord stimulation, was also observed.67-70
by measuring local VF intervals, Opthof et al showed that LSG Intrathecal clonidine, when delivered via a catheter at T2-T4
stimulation can increase dispersion in refractoriness by shorten- spinal segments, also significantly reduces the occurrence of ven-
ing refractoriness across nonischemic sites while either not tricular tachycardia and fibrillation during transient myocardial
changing or increasing refractoriness at ischemic sites during ischemia.71 The benefit of thoracic epidural anesthesia (TEA) was
coronary occlusion, thereby increasing dispersion across the reported in series of patients with cardiomyopathy and refractory
ischemic border by 14% to 59%.55 Finally, subthreshold LSG VT. In 66% of the patients, TEA reduced the burden of arrhyth-
stimulation has been shown to increase nerve sprouting in dogs mias by 80%.72
with MI. These dogs have a greater number of episodes of VT In humans, postganglionic sympathetic blockade, with medi-
compared with controls, suggesting that additional mechanisms cations (β-adrenergic receptor blockers) or via percutaneous stel-
underlie the increase in dispersion seen in infarcted hearts.61 late blockade (usually injection of bupivacaine 0.25%), has also
been reported to reduce ventricular arrhythmias. Among patients
early after MI with recurrent VF (electrical storm), those treated
with sympathetic blockade had improved outcomes as compared
Effect of Neuraxial Modulation in Reducing with those treated with the standard Advanced Cardiac Life
Risk of Ventricular Arrhythmias Support (ACLS) protocol.73 Sympathetic blockade was estab-
lished with the use of LSG blockade in 6 patients and with infu-
As augmented sympathetic tone increases the risk of arrhythmias, sions of either propranolol or esmolol in 21 patients without
interventions that aim to reduce sympathetic tone diminish the antiarrhythmic therapy, as recommended by ACLS. One-week
risk of sudden cardiac death and ventricular arrhythmias. and 1-year mortality were significantly higher in the group
undergoing standard ACLS protocol compared with the
sympathetic blockade group (82% vs. 22% at 1 week, 95% vs.
Strategies and Techniques of Cardiac 33% at 1 year, respectively).73 It is important to note that in
Sympathetic Denervation and Effects humans with percutaneous RSG or LSG blockade, no significant
on Ventricular Arrhythmias changes in hemodynamics or ejection fraction during rest or
with exercise have been reported.74 Finally, renal denervation has
Modulation and blockade of the sympathetic nervous system can been reported to reduce the burden of ventricular arrhythmias in
be attempted at multiple levels, including the brain, the spinal two patients, as well as electrical storm (one with nonobstructive
Brain and higher

General centers
anesthesia
LSG
TEA,
SCS, and RSG
intrathecal
Thoracic
clonidine Focal VF
T1-T4
Spinal
cord
SAN
NE
Cervicothoracic NE
sympathectomy
Adrenal
medulla
Beta blockers
2/3 E
1/3 NE
Macro
Renal denervation reentry
Kidney Functional VT
and VF
Figure 41-6. Neuraxial modulation of the cardiac sympathetic nervous system can occur at multiple levels. Sympathetic outflow from the brain and higher centers can be
blocked via general anesthesia. Thoracic epidural anesthesia, spinal cord stimulation, and intrathecal clonidine modulate the preganglionic neuronal output from the spinal
cord. Cervicothoracic sympathectomy targets the left stellate ganglia (LSG), right stellate ganglia (RSG), and T1-T4 ganglia of the sympathetic chain. Beta-adrenergic recep-
tors at the level of the myocardium can be blocked via β-blockers, and renal denervation at the level of the renal arteries can decrease sympathetic output from the kidneys.
All these therapies have been shown to reduce the burden of ventricular arrhythmias.
in the United States and Europe, it is important to keep in mind

hypertrophic cardiomyopathy and another with nonischemic that left cervicothoracic sympathectomy has been a successful
cardiomyopathy).75 treatment option for patients with recurrent angina refractory to
medical therapy, in sympathetically mediated pain syndromes,
and in patients with long QT syndrome and catecholamin-
Effects of Left and Bilateral Surgical ergic polymorphic VT with repeated episodes of ventricular
Cervicothoracic Sympathectomy arrhythmia.80-90 Left cardiac sympathetic denervation can reduce
on Ventricular Arrhythmias cardiac events in long QT syndrome. However, both in the car-
diomyopathy population and in patients with long QT syndrome
In 1983 Schwartz et al showed that the incidence of ventricular and catecholaminergic polymorphic ventricular tachycardia
fibrillation was decreased from 66% to zero by performing left (CPVT), left and bilateral cervicothoracic sympathectomies are
stellectomy in post-MI dogs.76 Further, Stramba-Badiale et al reserved for patients whose arrhythmias are refractory to medical
showed that ventricular fibrillation threshold (VFT) decreases therapy and/or who are receiving frequent ICD shocks. Most of
with vagotomy or right ganglion stellectomy, while VFT increases these cardiomyopathy patients have already undergone one or
with left stellectomy.77 Schwartz et al showed that left cervico- multiple ablation procedures for treatment of VT.
thoracic sympathectomy reduced the risk of sudden cardiac death
from 21.3% to 2.7% in high-risk MI patients with at least one
episode of ventricular tachycardia or fibrillation.78 This effect was Surgical Technique
equivalent in magnitude to β-receptor blocker medications. In
patients with refractory arrhythmias and cardiomyopathy, left Left and bilateral cervicothoracic sympathectomy involves
and bilateral cervicothoracic sympathectomy reduced the burden removal of the lower one-third to one-half of the left or bilateral
of arrhythmia in more than 60% of patients.72,79 On the basis of stellate ganglia and the thoracic ganglia of T2-T4. This proce-
these study findings, a suggested approach for sympathetic mod- dure provides adequate cardiac denervation usually with absence
ulation, particularly in the setting of VT storm, is shown in of or minimal Horner’s syndrome in long QT syndrome, CPVT,
Figure 41-7. and cardiomyopathy.72,79,82,83,85,90 This procedure is now most
commonly performed using video-assisted thoracoscopic surgical
(VATS) techniques, thereby reducing perioperative morbidity
and duration of hospitalization.85,87,91 For left cervicothoracic
Surgical Sympathetic Denervation in Humans sympathectomy, the patient is placed in a left lateral position,
under single-lung ventilation. Three 1-cm incisions are made in
Although left and bilateral cervicothoracic sympathectomies have the subaxillary region for introduction of endoscopic instru-
been performed in small series of patients with cardiomyopathy ments. The stellate and thoracic ganglia are located behind the
parietal pleura, in the paravertebral position. Then, the lower

Patient with severe
41
one-third to one-half of the stellate ganglion, along with the
ventricular
arrhythmias
chain from T2-T4, is completely removed. Histopathologic con-
firmation of neuronal cell bodies is obtained during the proce-
dure. The procedure can be typically performed on one side in
less than 60 minutes.82
Reversible causes YES
suppressed
arrhythmias? Risks and Complications of
Devise optimal Cervicothoracic Sympathectomy
NO medical therapy and
consider substrate
Left cervicothoracic sympathectomy can be accompanied imme-
mapping if indicated.
Pharmacologic YES
diately postoperatively by ipsilateral Horner’s syndrome.
therapy However, when the VATS approach is used, it occurs rarely and
successful? has been temporary because the sympathetic fibers directed
toward the ocular regions usually cross the upper portion of the
NO LSG, which is typically spared in this procedure.82,85 Mild lower-
ing of the left eyelid appears transiently after surgery, but seldom
Intubation YES persists.82
and sedation The effect of left cervicothoracic sympathectomy on LV func-
successful? tion has been studied using echocardiography. In patients with
Stabilize patient
NO and plan elective sympathetically mediated pain syndrome but normal ventricular
catheter mapping function, hemodynamic parameters including heart rate and
and ablation. blood pressure remain unchanged, although afterload reduction—
Thoracic epidural YES a desired effect in cardiomyopathy patients—does occur.89 Fur-
anesthesia (TEA)
successful? thermore, global and regional systolic function also remains
unchanged, although isovolumic relaxation time is slightly pro-
Adjust NO longed.88,89 This small effect has not compromised ventricular
TEA function, and stroke volume has actually increased as a result of
level afterload reduction.89
Emergent catheter Devise optimal medical
mapping and NO therapy and ICD
ablation. VT implantation if not
remains inducible? already present. Conclusion
YES The sympathetic nervous system plays a major role in the patho-
genesis of ventricular arrhythmias. Therapies that inhibit the
Cardiac YES Proceed with cardiac effects of this system have been shown to reduce the burden of
transplant transplant work up arrhythmias. Left and/or bilateral sympathetic denervation tech-
candidate? and listing. niques are feasible and safe options for patients with refractory
arrhythmias who have not responded to medical therapy and/or
NO
ventricular tachycardia ablation. Further, they may be the only
If arrhythmia is option for patients with refractory arrhythmias who are not can-
Plan cardiac LV origin, or didates for cardiac transplantation. Randomized studies with
Perform left cardiac
sympathetic LV-only scar. longer follow-up are required to assess the long-term risks and
sympathetic
denervation benefits of these procedures.
denervation (LCSD)
(CSD).
If arrhythmia is
RVOT, A RVC,
PMVT, RV only scar, Acknowledgments
or no cardiac scar Consider bilateral
cardiac sympathetic Some of the research reported in this chapter is supported by the
denervation (BCSD)
National Heart, Lung, and Blood Institute grant R01HL084261
(Dr. Shivkumar) and American Heart Association AHA
Figure 41-7. A suggested approach for neuraxial modulation of a patient with
#11FTF6559994 (Dr. Vaseghi).
ventricular tachycardia (VT) storm is shown.
cardiac arrhythmias. Pacing Clin Electrophysiol logical implications. Anat Sci Int 86:30–49,
References 6:1210–1220, 1983. 2011.
4. Jänig W: Functional anatomy of the peripheral 7. Angelakos E, King M, Millard R: Regional distri-
1. Schwartz PJ, Vanoli E, Stramba-Badiale M, et al: sympathetic and parasympathetic system. In Jänig bution of catecholamines in the heart of various
Autonomic mechanisms and sudden death: New W, editor: The Integrative Action of the Auto- species. Ann N Y Acad Sci 156:219–240, 1969.
insights from analysis of baroreceptor reflexes in nomic Nervous System: Neurobiology of Homeo- 8. Hutchins GD, Miller MA, Zips DP: Neurocardiac
conscious dogs with and without a myocardial stasis, Cambridge, 2006, Cambridge University imaging. In Zipes DP, Jalife J, editors: Cardiac
infarction. Circulation 78:969–979, 1988. Press, pp 13–34. Electrophysiology: From Cell to Bedside, Philadel-
2. Vaseghi M, Shivkumar K: The role of the auto- 5. Janes RD, Brandys JC, Hopkins DA, et al: Anatomy phia, 2009, Saunders Elsevier, pp 391–403.
nomic nervous system in sudden cardiac death. of human extrinsic cardiac nerves and ganglia. Am 9. Armour JA: Myocardial ischaemia and the cardiac
Prog Cardiovasc Dis 50:404–419, 2008. J Cardiol 57:299–309, 1986. nervous system. Cardiovasc Res 41:41–54, 1999.
3. Zipes DP, Barber MJ, Takahashi N, et al: Influence 6. Kawashima T: Anatomy of the cardiac nervous 10. Pauza DH, Skripka V, Pauziene N, et al: Morphol-
of the autonomic nervous system on the genesis of system with clinical and comparative morpho- ogy, distribution, and variability of the epicardiac
neural ganglionated subplexuses in the human ventricular arrhythmia. Circulation 101:1960– 52. Ben-David J, Zipes DP: Differential response to
heart. Anat Rec 259:353–382, 2000. 1969, 2000. right and left ansae subclaviae stimulation of early
11. Marron K, Wharton J, Sheppard MN, et al: Dis- 32. Dillon SM, Allessie MA, Ursell PC, et al: Influ- afterdepolarizations and ventricular tachycardia
tribution, morphology, and neurochemistry of ences of anisotropic tissue structure on reentrant induced by cesium in dogs. Circulation 78:1241–
endocardial and epicardial nerve terminal arboriza- circuits in the epicardial border zone of subacute 1250, 1988.
tions in the human heart. Circulation 92:2343– canine infarcts. Circ Res 63:182–206, 1988. 53. Priori SG, Mantica M, Schwartz PJ: Delayed after-
2351, 1995. 33. Gough WB, Mehra R, Restivo M, et al: Reentrant depolarizations elicited in vivo by left stellate gan-
12. Esler M, Jennings G, Lambert G, et al: Overflow ventricular arrhythmias in the late myocardial glion stimulation. Circulation 78:178–185, 1988.
of catecholamine neurotransmitters to the circula- infarction period in the dog: Correlation of activa- 54. Opthof T, Coronel R, Vermeulen JT, et al: Disper-
tion: Source, fate, and functions. Physiol Rev tion and refractory maps. Circ Res 57:432–442, sion of refractoriness in normal and ischaemic
70:963–985, 1990. 1985. canine ventricle: effects of sympathetic stimulation.
13. Gabella G: Fine structure of post-ganglionic nerve 34. Pak HN, Qayyum M, Kim DT, et al: Mesenchymal Cardiovasc Res 27:1954–1960, 1993.
fibers and autonomic neuroeffector junctions. In stem cell injection induces cardiac nerve sprouting 55. Opthof T, Dekker LR, Coronel R, et al: Interac-
Burnstock G, Hoyle CHV, editors: Autonomic and increased tenascin expression in a Swine model tion of sympathetic and parasympathetic nervous
Neuroeffector Mechanisms, Chur, 1995, Harwood of myocardial infarction. J Cardiovasc Electro- system on ventricular refractoriness assessed by
Academic Publishers, pp 1–31. physiol 14:841–848, 2003. local fibrillation intervals in the canine heart. Car-
14. Rubart M, Zipes DP: Mechanisms of sudden 35. Okuyama Y, Pak HN, Miyauchi Y, et al: Nerve diovasc Res 27:753–759, 1993.
cardiac death. J Clin Invest 115:2305–2315, sprouting induced by radiofrequency catheter abla- 56. Opthof T, Misier AR, Coronel R, et al: Dispersion
2005. tion in dogs. Heart Rhythm 1:712–717, 2004. of refractoriness in canine ventricular myocardium:
15. Meredith IT, Eisenhofer G, Lambert GW, et al: 36. Cha YM, Redfield MM, Shah S, et al: Effects of Effects of sympathetic stimulation. Circ Res
Cardiac sympathetic nervous activity in congestive omapatrilat on cardiac nerve sprouting and struc- 68:1204–1215, 1991.
heart failure: Evidence for increased neuronal nor- tural remodeling in experimental congestive heart 57. Vaseghi M, Zhou W, Shi J, et al: Sympathetic
epinephrine release and preserved neuronal uptake. failure. Heart Rhythm 2:984–990, 2005. innervation of the anterior left ventricular wall by
Circulation 88:136–145, 1993. 37. Liu YB, Wu CC, Lu LS, et al: Sympathetic nerve the right and left stellate ganglia. Heart Rhythm
16. Bristow M: Changes in myocardial and vascular sprouting, electrical remodeling, and increased vul- 9:1303–1309, 2012.
receptors in heart failure. J Am Coll Cardiol nerability to ventricular fibrillation in hypercholes- 58. Cardinal R, Savard P, Armour JA, et al: Mapping
22:61A–71A, 1993. terolemic rabbits. Circ Res 92:1145–1152, 2003. of ventricular tachycardia induced by thoracic
17. Bengel FM, Schwaiger M: Assessment of cardiac 38. Cao JM, Chen LS, KenKnight BH, et al: Nerve neural stimulation in dogs. Can J Physiol Pharma-
sympathetic neuronal function using PET imaging. sprouting and sudden cardiac death. Circ Res col 64:411–418, 1986.
J Nucl Cardiol 11:603–616, 2004. 86:816–821, 2000. 59. Euler DE, Nattel S, Spear JF, et al: Effect of sym-
18. Riemann B, Shafers M, Law M, et al: Radioligands 39. Zhou S, Cao JM, Tebb ZD, et al: Modulation of pathetic tone on ventricular arrhythmias during
for imaging myocardial alpha- and beta- QT interval by cardiac sympathetic nerve sprout- circumflex coronary occlusion. Am J Physiol
adrenoreceptors. Nuklearmedizin 42:4–9, 2003. ing and the mechanisms of ventricular arrhythmia 249:H1045–H1050, 1985.
19. Inoue H, Zipes DP: Time course of denervation of in a canine model of sudden cardiac death. J Car- 60. Gantenberg NS, Hageman GR: Enhanced induc-
efferent sympathetic and vagal nerves after occlu- diovasc Electrophysiol 12:1068–1073, 2001. tion of ventricular arrhythmias during sympathetic
sion of the coronary artery in the canine heart. Circ 40. Zhou S, Chen LS, Miyauchi Y, et al: Mechanisms stimulation before and during coronary artery
Res 62:1111–1120, 1988. of cardiac nerve sprouting after myocardial infarc- occlusion. Int J Cardiol 34:75–83, 1992.
20. Li W, Knowlton D, Van Winkle DM, et al: Infarc- tion in dogs. Circ Res 95:76–83, 2004. 61. Swissa M, Zhou S, Gonzalez-Gomez I, et al: Long-
tion alters both the distribution and noradrenergic 41. Kanazawa H, Ieda M, Kimura K, et al: Heart term subthreshold electrical stimulation of the left
properties of cardiac sympathetic neurons. Am J failure causes cholinergic transdifferentiation of stellate ganglion and a canine model of sudden
Physiol Heart Circ Physiol 286:H2229–H2236, cardiac sympathetic nerves via gp130-signaling cardiac death. J Am Coll Cardiol 43:858–864,
2004. cytokines in rodents. J Clin Invest 120:408–421, 2004.
21. Barber MJ, Mueller TM, Henry DP, et al: Trans- 2010. 62. Burjorjee JE, Milne B: Propofol for electrical
mural myocardial infarction in the dog produces 42. Kimura K, Ieda M, Fukuda K: Development, mat- storm; a case report of cardioversion and suppres-
sympathectomy in noninfarcted myocardium. Cir- uration, and transdifferentiation of cardiac sympa- sion of ventricular tachycardia by propofol. Can J
culation 67:787–796, 1983. thetic nerves. Circ Res 110:325–336, 2012. Anaesth 49:973–977, 2002.
22. Kammerling JJ, Green FJ, Watanabe AM, et al: 43. Huang B, Qin D, El-Sherif N: Early down- 63. Dijkman B, den Dulk K, Wellens HJ: Management
Denervation supersensitivity of refractoriness in regulation of K+ channel genes and currents in the of electrical instability after ICD implantation.
noninfarcted areas apical to transmural myocardial postinfarction heart. J Cardiovasc Electrophysiol Pacing Clin Electrophysiol 18:148–151, 1995.
infarction. Circulation 76:383–393, 1987. 11:1252–1261, 2000. 64. Mulpuru SK, Patel DV, Wilbur SL, et al: Electrical
23. Warner MR, Wisler PL, Hodges TD, et al: Mech- 44. Gidh-Jain M, Huang B, Jain P, et al: Differential storm and termination with propofol therapy: A
anisms of denervation supersensitivity in regionally expression of voltage-gated K+ channel genes in left case report. Int J Cardiol 128:e6–e8, 2008.
denervated canine hearts. Am J Physiol 264:H815– ventricular remodeled myocardium after experi- 65. Tavernier R, Derycke E, Jordaens L: An arrhyth-
H820, 1993. mental myocardial infarction. Circ Res 79:669– mia storm dependent on adrenergic drive late
24. Vaseghi M, Lux RL, Mahajan A, et al: Sympathetic 675, 1996. after implantation of an internal cardioverter defi-
stimulation increases dispersion of repolarization 45. Qin D, Zhang ZH, Caref EB, et al: Cellular and brillator. Eur J Pacing Electrophysiol 7:33–36,
in humans with myocardial infarction. Am J ionic basis of arrhythmias in postinfarction remod- 1997.
Physiol Heart Circ Physiol 302:H1838–H1846, eled ventricular myocardium. Circ Res 79:461– 66. Issa ZF, Zhou X, Ujhelyi MR, et al: Thoracic spinal
2012. 473, 1996. cord stimulation reduces the risk of ischemic ven-
25. Sjoberg J, Kanje M: The initial period of peripheral 46. Heath BM, Xia J, Dong E, et al: Overexpression of tricular arrhythmias in a postinfarction heart
nerve regeneration and the importance of the local nerve growth factor in the heart alters ion channel failure canine model. Circulation 111:3217–3220,
environment for the conditioning lesion effect. activity and beta-adrenergic signalling in an adult 2005.
Brain Res 529:79–84, 1990. transgenic mouse. J Physiol 512(Pt 3):779–791, 67. Foreman RD, Linderoth B, Ardell JL, et al: Modu-
26. Fu SY, Gordon T: The cellular and molecular basis 1998. lation of intrinsic cardiac neurons by spinal cord
of peripheral nerve regeneration. Mol Neurobiol 47. Han S, Kobayashi K, Joung B, et al: Electroana- stimulation: Implications for its therapeutic use in
14:67–116, 1997. tomic remodeling of the left stellate ganglion after angina pectoris. Cardiovasc Res 47:367–375, 2000.
27. Rabinovitch MA, Rose CP, Rouleau JL, et al: myocardial infarction. J Am Coll Cardiol 59:954– 68. Mannheimer C, Eliasson T, Andersson B, et al:
Metaiodobenzylguanidine [131I] scintigraphy 961, 2012. Effects of spinal cord stimulation in angina pectoris
detects impaired myocardial sympathetic neuronal 48. Zhou S, Jung BC, Tan AY, et al: Spontaneous stel- induced by pacing and possible mechanisms of
transport function of canine mechanical-overload late ganglion nerve activity and ventricular action. Br Med J 307, 1993.
heart failure. Circ Res 61:797–804, 1987. arrhythmia in a canine model of sudden death. 69. Sanderson JE, Ibrahim B, Waterhouse D, et al:
28. Parthenakis FI, Prassopoulos VK, Koukouraki SI, Heart Rhythm 5:131–139, 2008. Spinal electrical stimulation for intractable angina:
et al: Segmental pattern of myocardial sympathetic 49. Ogawa M, Zhou S, Tan AY, et al: Left stellate Long-term clinical outcome and safety. Eur Heart
denervation in idiopathic dilated cardiomyopathy: ganglion and vagal nerve activity and cardiac J 15:810–814, 1994.
Relationship to regional wall motion and myocar- arrhythmias in ambulatory dogs with pacing- 70. Eliasson T, Augustinsson LE, Mannheimer C:
dial perfusion abnormalities. J Nucl Cardiol 9:15– induced congestive heart failure. J Am Coll Cardiol Spinal cord stimulation in severe angina pectoris-
22, 2002. 50:335–343, 2007. presentation of current studies, indications, and
29. Vracko R, Thorning D, Frederickson RG: Nerve 50. Wood A, Docimo S, Elkowitz DE: Cardiovascular clinical experience. Pain 65:169–179, 1996.
fibers in human myocardial scars. Hum Pathol disease and its association with histological changes 71. Issa ZF, Ujhelyi MR, Hildebrand KR, et al: Intra-
22:138–146, 1991. of the left stellate ganglion. Clin Med Insights thecal clonidine reduces the incidence of ischemia-
30. Vracko R, Thorning D, Frederickson RG: Fate of Pathol 3:19–24, 2010. provoked ventricular arrhythmias in a canine
nerve fibers in necrotic, healing, and healed rat 51. Ajijola OA, Wico J, Lambert HW, et al: Chronic postinfarction heart failure model. Heart Rhythm
myocardium. Lab Invest 63:490–501, 1990. myocardial infarction is associated with neural 2:1122–1127, 2005.
31. Cao JM, Fishbein MC, Han JB, et al: Relationship remodeling in human stellate ganglia. Heart 72. Bourke T, Vaseghi M, Michowitz Y, et al: Neur-
between regional cardiac hyperinnervation and Rhythm 9:261, 2012. axial modulation for refractory ventricular
arrhythmias: Value of thoracic epidural anesthesia antiadrenergic interventions. J Cardiovasc Electro- syndromes. Pacing Clin Electrophysiol 26:870–
and surgical left cardiac sympathetic denervation. physiol 3:2–16, 1992. 873, 2003.
73.
Circulation 121:2255–2262, 2010.
Nademanee K, Taylor R, Bailey WE, et al: Treat-
ing electrical storm: Sympathetic blockade versus
79. Ajijola OA, Lellouche N, Bourke T, et al: Bilateral
cardiac sympathetic denervation for the manage-
ment of electrical storm. J Am Coll Cardiol 59:91–
86. Epstein AE, Rosner MJ, Hageman GR, et al: Pos-
terior left thoracic cardiac sympathectomy by surgi-
cal division of the sympathetic chain: An alternative
41
advanced cardiac life support-guided therapy. Cir- 92, 2012. approach to treatment of the long QT syndrome.
culation 102:742–747, 2000. 80. Tygesen H, Wettervik C, Claes G, et al: Long-term Pacing Clin Electrophysiol 19:1095–1104, 1996.
74. Gardner MJ, Kimber S, Johnstone DE, et al: The effect of endoscopic transthoracic sympathicotomy 87. Ouriel K, Moss AJ: Long QT syndrome: an indica-
effects of unilateral stellate ganglion blockade on on heart rate variability and QT dispersion in tion for cervicothoracic sympathectomy. Cardio-
human cardiac function during rest and exercise. J severe angina pectoris. Int J Cardiol 70:283–292, vasc Surg 3:475–478, 1995.
Cardiovasc Electrophysiol 4:2–8, 1993. 1999. 88. Lobato EB, Kern KB, Paige GB, et al: Differential
75. Ukena C, Bauer A, Mahfoud F, et al: Renal sym- 81. Stephenson EA, Berul CI: Electrophysiological effects of right versus left stellate ganglion block
pathetic denervation for treatment of electrical interventions for inherited arrhythmia syndromes. on left ventricular function in humans: an echocar-
storm: First-in-man experience. Clin Res Cardiol Circulation 116:1062–1080, 2007. diographic analysis. J Clin Anesth 12:315–318,
101:63–67, 2012. 82. Schwartz PJ, Priori SG, Cerrone M, et al: Left 2000.
76. Schwartz PJ, Billman GE, Stone HL: Autonomic cardiac sympathetic denervation in the manage- 89. Schlack W, Dinter W: Haemodynamic effects of a
mechanisms in ventricular fibrillation induced by ment of high-risk patients affected by the long-QT left stellate ganglion block in ASA I patients: An
myocardial ischemia during exercise in dogs with syndrome. Circulation 109:1826–1833, 2004. echocardiographic study. Eur J Anaesthesiol 17:79–
healed myocardial infarction: An experimental 83. Li C, Hu D, Shang L, et al: Surgical left cardiac 84, 2000.
preparation for sudden cardiac death. Circulation sympathetic denervation for long QT syndrome: 90. Collura CA, Johnson JN, Moir C, et al: Left
69:790–800, 1984. Effects on QT interval and heart rate. Heart cardiac sympathetic denervation for the treatment
77. Stramba-Badiale M, Lazzarotti M, Schwartz PJ: Vessels 20:137–141, 2005. of long QT syndrome and catecholaminergic poly-
Development of cardiac innervation, ventricular 84. Bhandari AK, Scheinman MM, Morady F, et al: morphic ventricular tachycardia using video-
fibrillation, and sudden infant death syndrome. Am Efficacy of left cardiac sympathectomy in the treat- assisted thoracic surgery. Heart Rhythm 6:752–759,
J Physiol 263:H1514–H1522, 1992. ment of patients with the long QT syndrome. Cir- 2009.
78. Schwartz PJ, Motolese M, Pollavini G, et al: Pre- culation 70:1018–1023, 1984. 91. Josephs LG, Menzoian JO: Technical consider-
vention of sudden cardiac death after a first myo- 85. Li J, Wang L, Wang J: Video-assisted thoraco- ations in endoscopic cervicothoracic sympathec-
cardial infarction by pharmacologic or surgical scopic sympathectomy for congenital long QT tomy. Arch Surg 131:355–359, 1996.
Arrhythmia Mechanisms PART VII
Dominant Frequency and the

Mechanisms of Initiation and
Maintenance of Atrial Fibrillation
42
Felipe Atienza and Omer Berenfeld
a different approach that consisted of analysis of recordings of

CHAPTER OUTLINE
the arrhythmia in the frequency domain.9-18
Background 419 The central objective of this chapter is to discuss experimental
and clinical data from our laboratory supporting the hypothesis
Activation of Inward-Rectifier Potassium Channels
that acute AF in sheep and in some groups of human patients is
to Establish Underlying Mechanisms of AF 420 not a totally random phenomenon. As will be demonstrated, the
The Significance of DF Gradients for Possible spatiotemporal organization of waves and dominant frequency
Mechanisms and Ablation Outcomes 424 (DF) in the isolated sheep heart suggests that AF maintenance
depends on localized reentrant sources in the left atrium (LA)
CFAEs and Dominant Frequency—Dependency and fibrillatory conduction in its periphery.9-14 Motivated by these
and Significance 428 studies, we translated the analysis on the organization of DF to
Summary and Future Directions 429 human AF. Using electroanatomic mapping in humans, we gen-
erated three-dimensional, whole-atrial DF maps and found that
AF reentrant sources are localized primarily to the pulmonary
veins in the case of paroxysmal AF but elsewhere in the case of
Background chronic AF.15-18 An electropharmacologic approach was used to
demonstrate the reentrant nature of this process and to success-
Current Strategies and Shortcomings fully terminate AF targeting maximal DF (DFmax) sites in
in AF Treatment paroxysmal AF patients.16,17 Finally, simultaneous time and
frequency-domain analyses in paroxysmal AF demonstrated that
Atrial fibrillation (AF) is the most common cardiac arrhythmia in electrogram fractionation at the posterior LA wall is a reflection
humans, is associated with increased morbidity and mortality, and of fibrillatory conduction—a consequence of the dynamic inter-
has become a public health problem of the first order.1 However, action between high-frequency reentrant sources and the atrial
currently available treatments for AF are less than satisfactory. anatomy—and should not be considered a target for ablation.18
On one hand, available antiarrhythmic drugs for the management
of AF are not sufficiently effective and are hampered by signifi-
cant cardiac and extracardiac side effects that frequently offset Hypothesized Mechanisms Initiating and
their therapeutic benefits.2 On the other hand, the demonstration Maintaining AF
of AF triggers in the atrial sleeves of the pulmonary veins (PVs)3
led to the adoption of interventional approaches to AF treatment Localized Ectopic Triggers Versus Reentry
that, by creating a set of circumferential lesions around the PV The mechanisms of human AF are complex and are still poorly
ostia, achieved substantial clinical success.4 However, the effec- understood.7,8 The notion that a localized source of reentrant
tiveness of this therapy among the more prevalent and highly activity could maintain AF was first postulated by Lewis in the
heterogeneous persistent and long-term persistent AF popula- early part of the 20th century. A few decades later, the single
tions has been disappointing.5,6 This is mainly due to the incom- reentrant mechanism evolved into a more complex proposition,
plete understanding of the mechanisms underlying this complex which hypothesized that AF is maintained by randomly propagat-
arrhythmia.7,8 Traditionally, the study of AF has been based in the ing multiple wavelets that meander, usually for a short time, then
time-domain analysis of the signals recorded by a number of annihilate and perpetuate the regeneration of new wavelets.7,19
electrodes positioned at different locations. However, the ever- Nevertheless, toward the end of the 20th century, clinical studies
changing nature of AF showing extremely complex spatiotempo- provided evidence in support of a local driver mechanism of some
ral activation precludes a straightforward interpretation. Thus, cases of AF by observing sustained focal activity at the PVs, coro-
our understanding of the mechanisms of AF maintenance has nary sinus (CS), or superior vena cava that initiated and main-
been hindered in part by our inability to reproducibly quantify tained AF and could be eliminated by discrete ablation.3,20 To this
atrial rate and activation patterns during ongoing AF. To enhance date, whether such sites are automatic, triggered, or reentrant
our understanding of AF maintenance mechanisms, we followed remains unresolved. Overall, it has been established that the
419
420 ARRHYTHMIA MECHANISMS
remodeling induced by persistent AF is probably increasing both substrate for fibrillatory propagation on the RA free wall.13,35
the prevalence of spontaneous ectopic firing of atrial cells and the Further support for this hypothesis was sought by monitoring the
propensity for impulse reentry in the atrial substrate.21 Several direction of conduction along the preferential interatrial path-
lines of experiments have contributed to the assessment of the ways.13 Figure 42-1, B illustrates an example of left-to-right
role of reentrant and ectopic activity in AF: pharmacologic22,23 propagation along Bachmann’s bundle during AF, where wave
and gene transfer24 reduction of outward K+ currents that prolong fronts propagated from left to right in 81% and 80% of the
action potential duration (APD), thus suggesting a reduced pro- activations along Bachmann’s bundle and the inferoposterior
pensity for reentry by suppressing AF. On the other hand, reduc- interatrial pathway, respectively, in the direction of a decrease in
ing inward Ca2+ current by stabilizing the sarcoplasmic reticulum DF value.13 These data suggest that in our model, AF results from
release of Ca2+ can prevent AF, suggesting an important role for impulses generated at high frequency by sources in the LA that
ectopic activity in the arrhythmia.25,26 Overall, from single reentry propagate along interatrial pathways to activate the RA in a spa-
and multiple wavelets to single source focus propositions, we still tially complex manner. To further clarify the nature of LA-to-RA
lack a unifying hypothesis capable of convincingly explaining the conduction, we used drugs that modified AF frequency. As shown
mechanisms responsible for initiation and maintenance of the in Figure 42-1, C, D, the reduction in the fastest LA frequency
different types of AF. after D600 administration was followed by an increase in RA DF,
whereas increasing acetylcholine (ACh) concentrations increased
the LA-to-RA frequency gradient. An increase in the source
The Dominant Frequency Concept frequency results in intermittent blockade caused by sink-to-
and Translation from Bench to Bedside source mismatch between the interatrial pathways and the RA
components. In contrast, when the source frequency was reduced,
Characterizing the rate, regularity, and spatial patterns of electri- RA DF increased (remaining lower than the DF of the LA),
cal activation is the cornerstone of arrhythmia diagnosis and which indicates lesser mismatch at lower frequencies. Such a
treatment. Early studies on regional differences in activation result is incompatible with the multiple wavelet hypothesis and
cycle length (CL) demonstrated that measurements of AF CL can be explained only by a frequency-dependent change in fibril-
could contribute to discerning AF mechanisms and treatment.27-30 latory propagation away from a source in the LA, allowing a
However, time-domain analyses of the signals recorded during greater number of waves to reach the RA at lower source frequen-
ongoing AF are complex and are considerably subjective, often cies. Thus it becomes evident that the RA frequency and, conse-
leading to inaccurate activation pattern characterization.18,31,32 In quently, the LA-to-RA frequency gradient are determined by the
contrast, analysis in the frequency domain offers an alternative LA frequency.
way to visualize AF electrograms for estimating the rate and While both atria are excited in a 1 : 1 manner at frequencies
regularity of activation without the need to manually measure as high as 300 beats/min (5 Hz) during atrial flutter, AF ensues
intervals in the time domain.31,32 Fourier analysis provides the in the RA when stationary reentrant sources in the LA rotate at
spectral decomposition of any time series into its constituent frequencies greater than 7 Hz.12,13 Does such a transformation
sinusoidal waveforms at all frequencies. The fast Fourier trans- depend on the input frequency? This suggests that there must be
form (FFT) enables estimation of the power spectrum of activity a critical frequency at which the 1 : 1 input-output relation
and provides a description of the stationary properties of the between the LA and RA breaks down. In Figure 42-2, we show
entire signal, where the highest peak corresponds to the domi- results of optical mapping experiments and DF analysis that dem-
nant frequency (DF). When the activation amplitude and the CL onstrate the principle of frequency-dependent breakdown in con-
are stable during the recording period, the DF corresponds to 1/ duction.13 The preparation was subjected to periodic pacing at
CL and provides a robust estimation of CL.31 Which should be Bachmann’s bundle to simulate activity arriving from a periodic
the preferable method for quantifying the local activation rate? LA source into the RA. Although stimulation at 5.0 Hz resulted
To answer this question, we recently compared time- and in 1 : 1 activation of the entire RA, a heterogeneous distribution
frequency-domain methods, taking optical mapping recordings of DF domains was established during pacing at 7.7 Hz, with
as a reference, and found a more robust correlation of the local frequencies ranging between 3.5 and 7.7 Hz.
activation rate in electrical versus optical signals by DF quantifi- In Figure 42-2, B, DFs are plotted as a function of the pacing
cation than by the average CL quantification.32 frequency, showing that below 6.7 Hz, the activation response
The notion that the spatiotemporal complexity of wave propa- was 1 : 1 everywhere in the RA. Above the breakdown frequency
gation during AF was reflected in the local activation rate disper- of 6.7 Hz, a large DF dispersion manifested as multiple domains
sion led us to characterize DF in the power spectrum throughout whose individual frequencies were equal to or lower than the
both atria; we found the DFs to be organized in discrete domains.9 pacing frequency. In addition, we found that intermittent block
Most notably, during the arrhythmia, the activation frequencies resulted in significant loss of consistency in the beat-to-beat
in certain areas of the LA were consistently faster than those in direction of wave-front propagation, which provided a direct
any other areas.9-12 In a search for the mechanisms underlying explanation for the difficulty involved in finding an origin of the
such frequency distribution in space, subsequent studies demon- activation during fibrillatory conduction.13
strated the way interatrial pathways (i.e., Bachmann’s bundle and
the inferoposterior interatrial pathway) mediate fibrillatory con-
duction and the establishment of frequency gradients between Activation of Inward-Rectifier Potassium
the LA and the right atrium (RA).12,13,33,34 In Figure 42-1, A, a DF
map with a high degree of organization is shown, with discrete
Channels to Establish Underlying
frequency domains and an LA-to-RA frequency gradient.12 For Mechanisms of AF
the most part, the DF was highest at the LA (18.8 Hz), interme-
diate at the right end of Bachmann’s bundle (14.5 Hz) and lowest ACh and AF Frequency in the Sheep Model
at the RA (9.8 Hz). This pattern of DF domain distribution
persisted for several minutes, with less than about 20% temporal The general working hypothesis that acute AF results from the
variability.12 Moreover, results indicated that the largest decay in activity of a small number of high-frequency reentrant sources
frequency occurred at the junction between Bachmann’s bundle localized in one atrium, with fibrillatory conduction to the other
and the RA, suggesting that the intricate three-dimensional atrium, is based primarily on results obtained in the isolated,
architecture of the pectinate muscle network might be the Langendorff-perfused, sheep heart.9-12 We localized the sources
Dominant Frequency and the Mechanisms of Initiation and Maintenance of Atrial Fibrillation 421
LA RA
BB left BB right
42
25
20 9.8
14.1
18.8
18.7 14.5
15 7.5
10
5
14.8 14.1 9.8
0 Hz 18.8
IPP left IPP right
A Biatrial EG
1 sec
BB right
BB left
B 100 ms
Control ACh0.2 M
25
* D600 * ACh0.5 M
25 20
20
Frequency (Hz)
15
15
10
10
5 5
0 0
C LA STP RA DF D LA STP RA DF
Figure 42-1. Relation between activity in the left atrium and the right atrium A, LA-to-RA gradient in dominant frequency (DF). DF maps of the epicardial
surfaces of LA and RA, with values of DF along Bachmann’s bundle (BB) and the inferoposterior pathway (IPP). Numbers are expressed in Hertz. The colored areas of the
DF maps indicate the optical mapping fields. B, Left-to-right directionality of impulse propagation during AF. Recordings from a bi-atrial ECG (top trace) and three bipolar
electrodes along BB, the tracing on the bottom being most leftward. Effect of D600 (C) and acetylcholine (D) on LA and RA frequencies. LA STP, Frequencies of LA spatio-
temporal periodicity; RA DF, right atrial dominant frequency. ACh, Acetylcholine; EG, electrogram; LA, left atrium; RA, right atrium.
(Reproduced from Mansour M, Mandapati R, Berenfeld O, et al: Left-to-right gradient of atrial frequencies during acute atrial fibrillation in the isolated sheep heart. Circulation
103:2631-2636, 2001.)
that maintain AF by a combined use of optical mapping and recordings) provides direct evidence that it was the mechanism
frequency analysis. Figure 42-3, A, shows an AF episode in which underlying AF (Figure 42-3, B).11 To assess the nature of the
the site of the high-frequency periodic activity was localized to mechanism determining the DF of an episode, we studied the
the LA and persisted for the entire length of the episode (25 ACh dose dependence of its frequency.36 Figure 42-3, C, presents
minutes).11 The fact that the frequency of this reentrant source results in which a direct ACh dose dependence of rotor frequency
(rotor) was equal to the highest and narrowest DF peak (most was demonstrated in both atria, with the frequency being higher
regular signal) recorded from all sites (optical and electrical in the LA than in the RA at all concentrations.
5.0 Hz 7.7 Hz Figure 42-2. The “breakdown frequency” concept in a sheep heart
Left, Endocardial and epicardial DF maps of the same isolated RA preparation
SVC 3.5 paced at 5.0 and 7.7 Hz. Note the appearance of the heterogeneous DF domains
5.5
CT CT at 7.7 Hz. Right, Response of DFs versus pacing rate (n = 5). Each symbol repre-
7.8 sents one experiment. Pacing BB at rates below ≈6.7 Hz results in 1 : 1 activation.
At higher rates, the number of domains increases but the value of the DFs
J J
Endo
decreases.
(Reproduced from Berenfeld O, Zaitsev AV, Mironov SF, et al: Frequency-dependent

6.3 breakdown of wave propagation into fibrillatory conduction across the pectinate
muscle network in the isolated sheep right atrium. Circ Res 90:1173-1180, 2002.)
1 cm
TV
7.8
5.0
5.0
Epi
5.5
6.3
4 5 6 7 8
A Frequency (Hz)
10
6.7 Hz
8
Response DFs (Hz)
0
0 2 4 6 8 10
B Pacing frequency (Hz)
Several studies have suggested that AF is characterized by

incomplete reentry and multiple unstable reentrant cir- ACh and Adenosine Slow Rate
cuits,9,29,35,37,38 which seems to be consistent with the multiple of Automatic Pacemaking
wavelet hypothesis.7,19 However, in the presence of a single high-
frequency source of stable reentry, electrical activity elsewhere Adenosine and ACh are known to activate the same Kir3.x sub-
could still present as incomplete or as short-lived and unstable family of inward-rectifier potassium channels through different
reentrant circuits because of fibrillatory conduction.13 For signaling pathways.40 The current that arises from such activation
instance, Schuessler et al.39 found that in an isolated canine right is the same and may be termed IK,ACh or IK,Ado, depending on
atrial preparation, increasing concentrations of ACh converted whether ACh or adenosine is the agonist.41,42 By increasing K+
multiple reentrant circuits into a single, relatively stable high- conductance in the atrium, both ACh and adenosine hyperpolar-
frequency reentry that generated fibrillatory conduction. In ize the cell membrane and abbreviate the action potential dura-
agreement with this, our results (see Figure 42-3) support the tion and refractory period, causing sinus slowing and acceleration
hypothesis that a single source or a small number of sources of of activation frequency.43 In our sheep model of AF, ACh increased
ongoing reentrant activity is the mechanism underlying AF in atrial activation frequency (rotor DF) in a dose-dependent
this setting.36 manner, with the effect being larger in LA than in RA, leading
67 42
ms
0
A 1 cm
LA Pseudo ECG 14.7 Hz
Power
250 ms
B 0 Frequency (Hz) 60
35
LA
Rotor Frequency (Hz)
25
RA
15
5
0 1 2 3 4
C ACh Concentration (µM)
Figure 42-3. Reentrant sources of AF A, Isochrone map of optical activity from the free wall of the LA during sustained AF showing a vortex that rotated clockwise
at a period of 68.6 ± 8.9 ms (≈14.7 Hz). B, Optical pseudo-ECG of LA during the same episode of AF with its corresponding power spectrum. C, ACh dose-response curves
created using rotation frequencies from the five longest living rotors for each of six experiments during AF at each ACh concentration. An increase in rotor frequency with
ACh, as well as LA (red) predominance over RA (green), is visible.
(A, B, Reproduced from Mandapati R, Skanes A, Chen J, et al: Stable microreentrant sources as a mechanism of atrial fibrillation in the isolated sheep heart. Circulation 101:194-
199, 2000. C, Reproduced from Ortiz J, Niwano S, Abe H, et al: Mapping the conversion of atrial flutter to atrial fibrillation and atrial fibrillation to atrial flutter: Insights into
mechanisms. Circ Res 74:882-894, 1994.)
to an increase in the LA-to-RA frequency gradient (see Figure and the underlying mechanisms responsible for AF maintenance
42-3, C).36 Besides, adenosine inhibits isoproterenol-stimulated remained unknown. We analyzed the effects of adenosine infu-
ICa,L, delayed rectifier K+ current, chloride current (Icl), the tran- sion on local activation frequency at different sites of the fibril-
sient inward current, and the pacemaker current If in rabbit lating atria to determine whether such sites are automatic,
sinoatrial (SA) node43 and atrioventricular (AV) nodal cells,44 with triggered, or reentrant and whether changes in the driver activity
a small inhibitory effect on basal L-type calcium current (ICa,L).42 would alter spatial frequency gradients.15 The experimental
As a consequence, afterdepolarizations and triggered arrhythmic observation of ACh dose-dependent acceleration of rotor fre-
activity induced by catecholamines are attenuated or abolished quency enabled translation of animal experiments to the patient
by adenosine.42 with the use of adenosine.36 Thus, we determined the effects of
adenosine infusion on DF at varying locations of both atria
during ongoing AF.15 We also generated baseline DF maps of the
Adenosine and AF Frequency in Patients LA using real-time spectral analysis that allowed determination
of the specific DFmax sites likely to harbor the AF drivers. Figure
Several clinical studies have confirmed the existence of a hierar- 42-4 shows a representative example wherein the primary DFmax
chical organization in the rate of activation of different regions site located near the right inferior pulmonary vein (RIPV) (Panel
in the atria of patients with paroxysmal and chronic AF.15-17,45,46 A) significantly accelerated from 4.64 Hz at baseline to 6.35 Hz
However, the reasons for such a specific frequency distribution at the peak of the adenosine effect (Panel B). This observation
Baseline Adenosine
A B
LSPV
4.64 Hz
RSPV 6.35 Hz
LIPV
RIPV
3.66 Hz
3.66 Hz
50.0 mm/sec 50.0 mm/sec
Bip Bip
V5 V5
4.64 Hz 6.35 Hz
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
Figure 42-4. Accelerating effect of adenosine on DFmax sites A, Real-time DF map of LA (posterior view) before infusion of adenosine, from a paroxysmal AF
patient. Red arrow indicates primary DFmax site near the RIPV. Baseline recording at the primary DFmax site with its power spectrum and simultaneous V5 reference. B, Record-
ing at the primary DFmax site with power spectrum and simultaneous V5 reference during peak adenosine effect, showing increase in DF. LSPV, LIPV, RSPV, RIPV, Left/right
superior/inferior pulmonary veins (PVs); Bip, bipolar catheter.
(Reproduced from Atienza F, Almendral J, Moreno J, et al: Activation of inward rectifier potassium channels accelerates atrial fibrillation in humans: Evidence for a reentrant
mechanism. Circulation 114:2434-2442, 2006.)
clearly rules out late phase 3 early afterdepolarization (EAD)- different locations in paroxysmal compared with persistent AF
induced triggered activity due to pause-dependent calcium over- patients.15,49
load after termination of rapid excitation.47-49 In this patient, the
arrhythmia terminated during ablation at the primary DFmax site,
supporting again the critical role of such sites as AF drivers. In a The Significance of DF Gradients for Possible
larger cohort, we analyzed the effect of adenosine on the activa-
tion rate in specific atrial regions. In paroxysmal AF patients,
Mechanisms and Ablation Outcomes
adenosine infusion increased local DF, particularly at the PV/LA
region, amplifying a left-to-right frequency gradient (Figure Factors Determining the Presence of DF Gradients
42-5).15 In persistent AF, baseline DF was significantly higher in Experimental and Human AF
than in paroxysmal AF patients at all atrial sites, and the highest-
frequency sources accelerated by adenosine were located in either Structural Factors
atrium but not at PV sites.15 In sum, adenosine infusion increased Experimental results have shown that the three-dimensional
frequency primarily at sites that activated at the highest rate at atrial structure, particularly with regard to heterogeneous wall
baseline, demonstrating that those sites are involved in the main- thickness and the networks of pectinate muscles in both LA and
tenance of AF. Adenosine-induced acceleration of activation fre- RA, may be an important factor that helps to establish the glob-
quency points toward reentry as the mechanism of AF maintenance ally aperiodic activity that characterizes AF.13,35 Simplified com-
in these patients and all but rules out an automatic or triggered puter simulation allows the prediction that although discrete
mechanism.15,49 These results support the idea that AF in humans sites of periodic activity may be responsible for the dominant
is maintained by high-frequency reentrant sources that have peak in the power spectrum, frequency-dependent patterns of
9
8 42
7
-40
Inward rectifier current (pA/pF)

6
Frequency (Hz)
-30 5
Paroxysmal AF
4
-20
3
-10 2
1
0 0
0.5 PV-LAJ HRA
1.5
Left atrium Right atrium
9
8
7
-40
Inward rectifier current (pA/pF)
Frequency (Hz)
6
-30 5
Persistent AF
4
-20
3
-10 2
1
0 0
0.5
PV-LAJ HRA
1.5
Left atrium Right atrium
Basal current Basal DF

Total current (basal current +CCh-mediated Peak Adenosine DF
current increase) in presence of Carbachol
Figure 42-5. IK,ACh current distribution and density parallel DF spatial distribution and adenosine response in paroxysmal and persistent AF
patients Right panels, Mean ± SD of DF for and high RA (HRA) at baseline (white) and peak adenosine effect (black) in paroxysmal and persistent AF patients. Left panels,
Inward-rectifier currents in RA and LA myocytes from paroxysmal and persistent AF patients (mean ± SD). White, Basal current in absence of PV-LA junction (PV-LAJ)
carbachol (CCh). Black, Total current (basal current + CCh-mediated current increase) in the presence of CCh.
(Right panels, Reproduced from Atienza F, Almendral J, Moreno J, et al: Activation of inward rectifier potassium channels accelerates atrial fibrillation in humans: Evidence for a
reentrant mechanism. Circulation 114:2434-2442, 2006. Left panels, Reproduced from Atienza F, Almendral J, Moreno J, et al: Activation of inward rectifier potassium channels
accelerates atrial fibrillation in humans: Evidence for a reentrant mechanism. Circulation 114:2434-2442, 2006; and Voigt N, Trausch A, Knaut M, et al: Left-to-right atrial inward
rectifier potassium current gradients in patients with paroxysmal versus chronic atrial fibrillation. Circ Arrhythm Electrophysiol 3:472-480, 2010.)
intermittent conduction block across sites of tissue thickening monolayers with an increasing myofibroblast/myocyte ratio57
may explain the regional differences in activation frequencies suggests that fibrosis by itself is an important factor in increasing
seen in many cases of AF.50 When acting in combination, the the complexity of the propagation and the slowing of the excita-
latter two ingredients (i.e., a discrete periodic source and complex tion frequency, independent of other ionic and structural remod-
patterns of conduction block) may in fact underlie the globally eling factors. Overall, structural remodeling (i.e., fibrosis),
aperiodic activity of AF and explain not only the reduction in the observed in both clinical and experimental AF, significantly alters
DF value but also an increase in fractionation of electrograms.14 atrial tissue composition and function and modifies the dynamics
However, several other factors may be also responsible for inter- during AF. However, the precise mechanisms underlying AF in
mittent blockades of propagation leading to spatial dispersion of the setting of atrial fibrosis are not fully elucidated.58
DFs. Increased fibrosis across the atria has also been associated
with AF incidence51 and affects the dynamics of activation during Ionic Factors
AF.52 The accumulation of fibrosis modifies the electrophysiolog- Both the stability of the spatial distribution of DFs and the sig-
ical properties of the atrium by reducing the velocity of the nificant differences in activation frequency and organization
impulse, providing a substrate for reentry and AF initiation53,54 between LA and RA suggest that the underlying electrophysio-
with reduced DF.55 In particular, increasing amounts of collagen logical properties of the former enable it to support greater acti-
I, are associated with a lower LA DF during AF after cardiac vation frequencies than are supported by the latter. As shown in
surgery.56 The finding of reduced frequency of reentry in Figure 42-3, the different responses of rotor frequency in LA and
RA are somehow related to the ACh-modulated potassium a sterile pericarditis model, Kumagai et al.29 identified in the
current, IK,ACh.36 A numeric simulation study using canine atrial septum dominant unstable reentrant circuits of very short CL
cells with realistic ionic and coupling properties showed that that maintained AF and could be successfully terminated by focal
indeed IK,ACh is a determinant of the frequency and stability of ablation.
rotors during AF.59 Computer simulations suggest that activation These observations were further confirmed in patients with
at extremely fast rates by stationary rotors may be the result of chronic AF and mitral valve disease in whom intraoperative epi-
the strong repolarizing influence exerted by their core, which cardial mapping and focal cryoablation performed at sites with
abbreviates the action potential duration (APD) in its proximity.60 rapid repetitive activation were able to successfully terminate AF
Consequently, the tissue close to the core achieves very short in a significant (80%) proportion of patients.28,65 Indirect evi-
CLs, whereas far from the core, the myocardium cannot conduct dence for the role of high-frequency atrial sites in maintaining
at the rate of the rotor, and nonuniform (i.e., other than 1 : 1) AF was also provided by the observation that sequential ablation
conduction develops, contributing to the gradient in DFs of sites showing the shortest atrial CL caused progressive slowing
observed during fibrillation. To provide more definitive evidence of AF frequency, leading to termination in 75% of patients with
for such a contention, we studied the effects of ACh on IK,ACh paroxysmal AF.30,66 In addition, ablation of visually determined
density of sheep myocytes and found that IK,Ach density is signifi- atrial electrograms with very short CL (≤120 ms), with or without
cantly higher in LA than RA myocytes.36 Thus, differences in the fractionation, was associated with AF termination in 100% par-
functional expression of IK,ACh channels in the atria, by allowing oxysmal and 91% persistent AF patients, along with significant
stabilization of the dominant rotor in the region with greatest tachycardia CL prolongation.67 These studies were hampered by
APD abbreviation (i.e., LA), are responsible for the LA-to-RA a limited sample size and relatively low mapping density and by
gradients of excitation frequency and fibrillatory conduction inaccurate assessment of atrial activation rate based on CL analy-
observed in this model. Taken together, patch-clamp and com- ses in the time domain.31,32
puter simulation results support the hypothesis that in the sheep In contrast, DF analysis of the bipolar signal of atrial activity
heart, LA myocytes can adapt to higher excitation frequencies during AF in the frequency domain provides objective character-
than RA myocytes. ization of the spatial distribution of excitation frequency.31,32 The
Our clinical observation of the presence of LA-to-RA DF use of spectral analyses of multipolar recordings in the atria
gradients15-17,45,46 can be explained by findings of recent studies that enabled demonstration of the existence of a hierarchical organi-
analyzed the distribution of voltage-dependent K+ currents of zation in the rate of activation of different regions in the atria of
human cells of RA and LA in sinus rhythm and in AF patients. patients with paroxysmal and chronic AF, with high-frequency
Voigt et al.61 demonstrated that patients with paroxysmal AF had sites usually localized to the LA driver AF, while the remaining
inward-rectifier potassium current densities that were twofold sites in the atria were passively activated.15-18,45,46,68 In a retrospec-
larger in LA than in RA cardiomyocytes, whereas chronic AF tive blind analysis of the correlation between atrial DF distribu-
patients had greater basal currents but no significant LA-RA dif- tion and ablation outcome, Sanders et al.15 found that ablation at
ferences (see Figure 42-5).16,61 Application of the muscarinic PVs harboring DFmax sites resulted in an increase in AF CL in
receptor agonist carbachol (CCH) activated IK,ACh, leading to an 89% of cases, in both paroxysmal and permanent AF. However,
increase in total current density responsible for the increase in K+ eventual arrhythmia termination occurred during ablation in
conductance. In patients with paroxysmal AF, CCH amplified the 88% of paroxysmal patients but in none with permanent AF.
LA-to-RA inward-rectifier current gradient (left panel), while ade- Most (87%) cases of termination of arrhythmia episodes were
nosine infusion, by accelerating and stabilizing rotors at the associated with ablation at a DFmax site, preferentially localized
chamber with higher current densities (LA), amplified the to the PV antra.
LA-to-RA DF gradient (right panel).16,61 In contrast, persistent AF Thus, because spectral analysis identifies localized sites of
patients had greater basal currents but no significant LA-RA dif- high-frequency activity during AF, and because highest-frequency
ferences at baseline and after total current increase following sites are associated with AF termination, we used a combination
CCH (left panel); similarly, no significant LA-RA DF differences of real-time DF mapping and RF ablation to determine the safety
were noted at baseline, and after adenosine infusion, DFs increased and long-term outcome of prospectively targeting DFmax sites in
only in the high RA (HRA) (right panel).16,61 In addition, other patients.17 For this purpose, we performed high-density mapping
chamber-specific differences in K+ current distribution are respon- of the atria by sequentially acquiring simultaneous bipolar elec-
sible for the larger basal inward-rectifier current in LA and RA of trograms during ongoing AF (Figure 42-6).16,17 Once DFmax sites
persistent AF patients without a significant LA-RA DF gradient, were identified, we performed adenosine infusions to confirm
such as the increase in voltage-independent inward-rectifier cur- their role in driving AF and targeted them for ablation, followed
rents (IK1 and the agonist-independent component [constitutive] by circumferential PV isolation (CPVI), enabling AF acute ter-
of the IK,ACh), together with the bi-atrial increase in the voltage- mination in 72% paroxysmal and 11% persistent AF patients.16,17
dependent IKs inward-rectifier current.61-64 In sum, unequal left- During follow-up, this ablation strategy resulted in long-term SR
to-right distribution and density of inward-rectifier K+ currents in maintenance in 75% of paroxysmal and 50% of persistent AF
paroxysmal versus persistent AF provide a basis for differences in patients. Radiofrequency ablation of DFmax sites leading to elimi-
excitation frequency at baseline, the different atrial locations of nation of LA-to-RA DF gradients was associated with a higher
DFmax sites, and variable responses to adenosine infusion. probability of remaining free of AF during long-term follow-up.
The proportion of AF recurrences in persistent AF was higher in
patients with untargeted DFmax sites, mostly because of safety
The Consequence and Meaning of Ablation of High concerns (e.g., left atrial appendage), pointing to the important
Dominant Frequency Sites Maintaining AF role of extrapulmonary sites in persistent AF maintenance. In
addition, we showed for the first time that real-time DF mapping
The first demonstration of the role of ablation of high-frequency is feasible and can be safely performed as a complement to con-
sites in terminating AF came from the experimental work of ventional AF ablation procedures.17 These results are in agree-
Morillo et al.27 Using a canine chronic AF model, these investiga- ment with those of other retrospective69,70 and prospective71
tors demonstrated that cryoablation at sites where AF cycle studies that analyzed the effects of ablation on atrial DFs, showing
lengths were consistently shorter, usually at the posterior left a more favorable outcome in patients with a baseline LA-to-RA
atrial wall near the PV-LA junctions, significantly prolonged DF gradient and significant frequency reduction in both atria
atrial CL and successfully restored sinus rhythm. Similarly, using following ablation.
Dominant Frequency-Bipolar
6.84 42
*
I
SVC
II
AI > 128 Points LSPV
V1
V5
ABL d
ABL p
Lasso 1,2
LIPV Lasso 3,4

RIPV
4.64 Lasso 5,6
R Lasso 7,8
Lasso 9,10
Lasso 11,12
Lasso 13,14
Lasso 15,16
Lasso 17,18
Lasso 19,20
A HRA
CS 1,2
6.84
B
Figure 42-6. Real-time DFmax site ablation A, Atrial DF map (posterior view; CARTO system) in a paroxysmal AF patient. Purple, Primary DFmax site (red arrow) on right
intermediate PV (RIPV). Red dots, Circumferential ablation line. B, Bipolar recording (top) of primary DFmax site and its power spectrum (bottom) before ablation. C, Surface
ECG leads and intracardiac lasso catheter electrograms within RIPV; ablation catheter in the encircled area, CS, and HRA catheter during isolation of right-sided PVs. Catheters
recording outside the encircled area (CS, HRA) show conversion to sinus rhythm (star), whereas the lasso catheter inside the RIPV demonstrates ongoing AF, showing that
this high-frequency site was responsible for AF maintenance. CS, Coronary sinus; DF, dominant frequency; HRA, high right atrium; PV, pulmonary vein; also LSPV, LIPV, RIPV,
left/right superior/inferior pulmonary veins (PVs).
(Reproduced from Atienza F, Almendral J, Jalife J, et al: Real-time dominant frequency mapping and ablation of dominant frequency sites in atrial fibrillation with left-to-right
frequency gradients predicts long-term maintenance of sinus rhythm. Heart Rhythm 6:33-40, 2009.)
that is needed for mapping and ablation, demonstrated that even

Potential Therapeutic Implication for DFmax with no trend of acceleration or slowing of the activation rate,
Site Ablation in Patients with Paroxysmal the instantaneous rate fluctuated by about ±1 Hz. The presence
and Persistent AF of such fluctuations possibly reflected spatial shifts in the loca-
tions of the discrete drivers responsible for high-DF sites, with
The aforementioned data clearly indicate that DFmax sites play an concomitant transient alterations in local rates of activation.
important role in the maintenance of AF in a significant number Indeed, Schuessler et al.72 used bipolar electrograms from epicar-
of patients. To what extent should one expect to be able to deter- dial sites intraoperatively during AF and reported that in half of
mine the role of DFmax sites in arrhythmia maintenance in the AF patients, the location of the highest DF changed during the
patient population at large? To address this question, consider- recording period. Nevertheless, among patients who underwent
ation should be given to the relatively low resolution of currently real-time DF mapping and ablation, in 66% of those re-studied
available mapping systems and the substantial temporal limita- because of AF recurrences, a second DF map showed nonsignifi-
tion imposed by the need to sequentially acquire the electrograms cant differences in the locations and values of DFmax sites, indicat-
needed to generate the DF maps. In this regard, studies on tem- ing acceptable long-term stability of DFs and rendering them a
poral stability of DFs at or near the PVs46 and the entire atria15 potential target for ablation to terminate AF.17
have found no significant change over periods between 10 and The real-time DF mapping and ablation strategy could be
20 seconds. However, an assessment of the stability in the CS especially useful in two situations: (1) In paroxysmal AF patients,
over a period of 50 minutes, which is more relevant to the time it may facilitate identification of the atrial site driving AF, enabling
a more limited ablation; and (2) the strategy of combining CPVI activity detection (far-field signals) specifically located at ana-
and DFmax site ablation may be particularly useful in persistent tomic junctures such as the coronary sinus or the septum. Simi-
AF patients because of the higher prevalence of extrapulmonary larly, Jadidi et al.89 found that electrogram fractionation and
sources found in this context.16-18 This pathophysiologically based voltage alterations were predominantly functional in nature, with
strategy, when directed to selectively eliminate the sites respon- most CFAE sites occurring at regions of wave collision, calling
sible for AF maintenance, could reduce the risks imposed by into question the role played by CFAEs in the perpetuation of
extensive empirical ablation procedures and increase the long- persistent and paroxysmal AF.
term success rate.8,17
Intermittency of Fractionated Activity in PAF

CFAEs and Dominant Frequency— Patients and Acceleration of Activation Rate
Dependency and Significance Previous studies demonstrated that the atria respond to activation
rate increments with progressive deterioration of stable direc-
The Controversy of Complex Fractionated tionality and electrogram fractionation and suggest a possible
Electrograms During AF relationship between fractionated electrograms and short AF
cycle lengths.14,90 Rostock et al.91 found that the occurrence
In 2004, Nademanee et al.67 proposed that areas with of fractionated electrograms anywhere in the atria was signifi-
complex fractionated electrograms (CFAEs) were critical sites cantly associated with AF CL shortening. On the other hand,
for perpetuation of AF, and that their elimination with radiofre- interventions leading to AF CL prolongation, such as pharmaco-
quency ablation was linked with a high probability of sinus logic autonomic blockade92 or PV isolation,30,82 significantly
rhythm maintenance. However, recent studies targeting decreased the proportion of fractionated electrograms in the
CFAEs during ongoing AF at multiple atrial sites, as a stand- atria. However, in those studies, AF initiation and transitions to
alone strategy or following CPVI, have yielded highly heteroge- fractionation were not analyzed, and no mechanistic explanations
neous results.73-78 As a consequence, only 50% of the panel that for this phenomenon were proposed. We systematically studied
prepared the current Consensus Statement on AF ablation rec- the mechanisms of fractionated electrogram formation on the
ommended the use of this ablation strategy as a complement to posterior left atrial wall (PLAW) in pacing-induced human par-
pulmonary vein isolation in patients with long-standing persis- oxysmal AF and analyzed transitions between organized patterns
tent AF.79 and changes in electrogram morphology.18 For this purpose, we
Several reasons may account for these conflicting results, analyzed how the first term of the CFAE definition (fractionated
including the somewhat fuzzy definition of the CFAE concept, atrial electrograms composed of two or more deflections) relates
the subjective assessment of this phenomenon, the multiple eti- to the second term of the definition (atrial electrograms with a
ologies that underlie CFAE formation, and finally their uncertain very short CL [≤120 ms]). We found that organized and frag-
role in AF maintenance.18,80 In the original publication,67 CFAEs mented electrograms are intermittently recorded at the PLAW,
were empirically defined as (1) atrial electrograms that are frac- and highly organized patterns are recorded 31% of the time.
tionated and composed of two deflections or more and/or have Moreover, we observed that transitions from organized to frac-
a perturbation of the baseline with continuous deflections from tionated electrograms were preceded by progressive CL
a prolonged activation complex; or (2) atrial electrograms with a shortening.
very short CL (<120 ms) with or without multiple potentials As shown in Figure 42-7 (left panel, A to E), during transitions
when compared with the atrial CL recorded from other parts of to fractionation, significant inter-beat interval shortening is
the atria. However, despite the intrinsic difficulties in assessing noted, along with an increase in electrogram duration and in the
these features of electrogram characteristics during ongoing AF, number of spikes. Computer simulations in Figure 42-7 (right
they were only visually assessed by most of the studies,67,73-77 panel, F and G) reproduced this observation, whereby as the rotor
which imposed a significant degree of subjectivity on their evalu- drifts toward the recording catheter, the inter-beat interval short-
ation. Further attempts to quantify fractionation using a variety ens, resulting in electrogram widening and wavebreaks. Thus, the
of software methods were faced with doubts regarding their accu- transition to fractionation is not random and reflects functional
racy.32,78,81,82 But, by far, the most important contributor to the deterioration in the atrial conduction properties in response to
confusing CFAE scenario is the fact that multiple events, not periodic input acceleration, at least in paroxysmal AF patients.
always related to the arrhythmia maintenance mechanisms, can Similarly, Narayan et al. found that CFAEs could be preceded by
eventually cause fractionation of the electrograms.80 Whereas AF acceleration and APD alternans.88 Overall, these results dem-
some fractionated signals might represent critical zones related onstrate that, contrary to what may be implied in the original
to AF maintenance (i.e., high-frequency sources “driving” CFAE definition, the local frequency of activation is the main
AF),14,83 others might be passive and unrelated to the primary determinant of electrogram fragmentation at the PLAW in
arrhythmia mechanism84,85; others still might appear to be located patients with paroxysmal AF.
in areas surrounding the autonomic ganglionated plexi.86 Experi-
mental high-resolution studies have shown that most fractionated
signals are found at the periphery of high-frequency AF drivers Patterns of Propagation and Fractionated
and that rotor meandering might also underlie, at least in part, Electrograms in Patients
the electrogram fractionation at close proximity to the source.14,83
This might explain the success of some CFAE ablation proce- Despite the spatiotemporal complexity of wave propagation
dures that may have produced an anatomic obstacle around the during AF, experimental studies have demonstrated that AF is
highest DF site. In contrast, other studies have found that certain deterministic, in that waves propagating from the PVs into the
CFAE types are unrelated to the primary arrhythmia mechanism PLAW trigger reentry by breaking at boundaries of the septo-
and simply represent transient pivoting, wave-front collision, or pulmonary bundle.85,93 We recently confirmed these results in a
wave fractionation.84,87 Narayan et al.88 used monophasic action translational study from bench93 to bedside18 showing that orga-
potential recordings to reduce the burden of far-field contribu- nized AF phases are characterized by incoming wave patterns of
tion and found that most (67%) CFAEs were due to nonlocal activation at the PLAW, where the earliest site of activation was
110 6
42
R2 0.9 R2 0.96 D 2
Electrogram duration (ms)
100 Slope 4.65(IG95% 3.42-5.94) Slope 0.26(IG95% 0.22-0.3) 11 12

5 13 3
20
Number of spikes
90 10 19 14
4
15
80 4 9 1817 16
5
70 8 6
7
3
60
50 2 MR MR
A 1 2 3 4 5 6 7
Electrogram number
8 9 10
B 1 2 3 4 5 6
Electrogram number
7 8 9 10
250
R2 0.94
Slope 3.86(IG95% 4.4-2.9)
* * F
200
200
19-20
Sistolic interval
Sistolic interval
1 SP 3 SPs
150 17-18
180 15-16
100
13-14
113 104 97 92 87 83 79 89 99
160 50 11-12
9-10
0
C D
1 2 3 4 5 6 7 8 9 10 Pre- During- Post-
Intervals between electrograms fragmentation fragmentation fragmentation 7-8
5-6
SI 3-4 110 103 97 93 88 84 83 82 103
0-2
0 0.2 0.4 0.6 0.8 1 1.2
First 7 not fragmented (Pre-fragmentiation phase) Last 3 fragmented CFAEs

E (during-fragmentiation phase) G Time (sec)
Figure 42-7. Electrogram fractionation dependence on acceleration induced by a distant approaching source A-E: Electrogram characteristics and
systolic interval (SI) during transitions from organized to fragmented electrograms, as shown in E. A, Electrogram duration. B, Number of spikes. C, SI. D, Mean values of
SI before and during fragmentation and after resumption of organization. E, Electrogram recording during transition from organized to fragmented phase. Both the mean
electrogram duration (A) and the number of spikes (B) increased gradually as the systolic interval (C) decreased and the pattern changed from organized to fragmented.
F, Two snapshots of computer simulation showing a mother rotor (MR) drifting toward a 20-electrode catheter (D-20). G, Corresponding 10 pseudo-bipoles recorded at
the catheter. The rotor drift toward the catheter is accompanied by systolic interval shortening (from 113 to 79 ms) and electrogram widening, followed by wavebreak with
formation of two additional short-lived rotors. Red squares, Singularity point (SP).
(Reproduced from Atienza F, Calvo D, Almendral J, et al: Mechanisms of fractionated electrograms formation in the posterior left atrium during paroxysmal atrial fibrillation in
humans. J Am Coll Cardiol 57:1081-1092, 2011.)
located at the highest DF site (Figure 42-8, A, B). In contrast, discrete high-frequency sources and to identify their location
during organized-to-fragmented electrogram transitions, the might help improve the efficacy of ablation procedures. The
PLAW activation pattern changed, and as the rate of the pene- aforementioned studies15-18 served as the basis for the design of
trating wave increased, the electrograms widened and double the randomized RADAR-AF trial94 (www.clinicaltrial.gov [identi-
electrograms began to appear, yielding one of two different fier: NCT00674401]), which will test the following hypotheses:
patterns: (1) reentry pattern around a line of functional block, (1) In patients with paroxysmal AF, a limited ablative approach
in which the reentrant excitation wave front circulated targeting only high-frequency left atrial sites will have efficacy
clockwise around a pivoting point located anatomically at the similar to that of the standard strategy of CPVI and will have a
septopulmonary bundle (Figure 42-8, C); or (2) activation break- lower risk profile; and (2) In patients with chronic AF, combined
through across a line of slow conduction. These results demon- high-frequency site ablation/CPVI may offer higher success rates
strate that periodic impulses originating at the PV-LA junction compared with the CPVI strategy alone. Additionally, recent
propagate in highly recurrent directions repetitively toward the mapping studies based on forward processing of intracardiac95
center of the PLAW, while electrogram fractionation is due to and body surface96 recordings, as well as studies based on the
wave-front acceleration ahead of drifting rotors on the PLAW inverse solutions of surface recordings,97 suggest that a panoramic
and/or rotor meandering. Thus the dynamic interaction between assessment of the patterns of AF activation could enhance our
high-frequency reentrant sources and the atrial anatomy is understanding of the relationship between the local dynamics of
responsible for the phenomenon termed fibrillatory conduction, or the arrhythmia and the substrate on an individual basis. The
CFAEs.18 second area of advance is the study of the broad range of chronic
conditions that are responsible for electrically or structurally
remodeling the atria. Experimental data from our laboratory55,56
and others52 indicate that the spatiotemporal organization of
Summary and Future Directions fibrillatory waves depends on the type of remodeling. It is envi-
sioned that the use of advanced analysis methods will allow better
In the future, the combined use of time- and frequency-domain correlation of the dynamics of arrhythmia with the remodeled
measures, including DF mapping and electrogram fractionation, atria and will facilitate the translation of knowledge on ionic
should help elucidate AF mechanisms, leading to the develop- mechanisms of AF into the clinical setting,36,61 where chronic
ment of more effective treatment strategies, mainly in two spe- conditions underlying AF currently pose the greatest therapeutic
cific areas. The ability to determine whether AF is maintained by challenge.
1 2
A C
0 ms 0 ms
-200 ms -200 ms
4
3
ECGI (54) 100 mg -300 -200 -100 0 100 200 300 400 500
ECG aVF (57)
REF 1 sec update (40)
Hi +0.26mV
ROV CIR 13–14 (84)
Lo–0.22mV
B
400 mm/sec
1 ECG V6 (66) 50 200 250
4 ABL D-2 (23) 0 ms 0 ms
5 ABL 3-4 (23)
6 CIR D-2 (45)
7 CIR 2-3 (45)
8 CIR 3-4 (45) -200 ms -200 ms

9 CIR 4-5 (45)
10 CIR 5-6 (45)
11 CIR 6-7 (45)

5 6
200 ms
4 3 12 CIR 8-9 (45)
5
6 15 1413 2 13 CIR 9-10 (45)
16 12 14 CIR 10-11 (45)
7 20 D
17
8 18 19 11 15 CIR 11-12 (45)
9 10 16 CIR 12-13 (45)
0 ms 17 CIR 13-14 (45)
18 CIR 14-15 (45)
19 CIR 15-16 (45)
20 CIR 16-17 (45)
21 CIR 17-18 (45)
22 CIR 18-19 (45)

0 ms 0 ms
-200 ms
23 CIR 19-20 (45)
24 HIS D-2 (35)
25 SC D-2 (37)
26 SC 3-4 (37)
-200 ms -200 ms
Figure 42-8. Patterns of activation of the posterior left atrial wall during transitions to fragmentation A, Left atrial dominant frequency (DF) map (NaVX
System, posterior view). White arrow points to highest DF site (10.8 Hz) at the left inferior pulmonary vein (LIPV) antrum. B, Posterior left atrial wall activation map during
organized phase before fragmentation (right) shows an incoming wave pattern of activation progressing from closest to highest DF site at the LIPV (left, white) to the right
(purple-blue). C, Snapshots of wave propagation at the posterior left atrial wall (PLAW) during transitions to fragmentation (sequence 1-6): purple, unactivated regions; white,
advancing activation. Reentrant circuit with a clockwise propagation around a pivoting point located to the right edge of the septopulmonary bundle.
(Reproduced from Atienza F, Calvo D, Almendral J, et al: Mechanisms of fractionated electrograms formation in the posterior left atrium during paroxysmal atrial fibrillation in
humans. J Am Coll Cardiol 57:1081-1092, 2011.)
fibrillation. Circ Arrhythm Electrophysiol 3:32–38, Langendorff-perfused sheep heart. J Cardiovasc

References 2010. Electrophysiol 11:869–879, 2000.
5. Weerasooriya R, Khairy P, Litalien J, et al: Cath- 11. Mandapati R, Skanes A, Chen J, et al: Stable micro-
1. Fuster V, Ryden LE, Cannom DS, et al: 2011 eter ablation for atrial fibrillation: Are results main- reentrant sources as a mechanism of atrial fibrilla-
ACCF/AHA/HRS focused updates incorporated tained at 5 years of follow-up? J Am Coll Cardiol tion in the isolated sheep heart. Circulation
into the ACC/AHA/ESC 2006 Guidelines for the 57:160–166, 2011. 101:194–199, 2000.
management of patients with atrial fibrillation: A 6. Chao TF, Tsao HM, Lin YJ, et al: Clinical 12. Mansour M, Mandapati R, Berenfeld O, et al: Left-
report of the American College of Cardiology outcome of catheter ablation in patients with to-right gradient of atrial frequencies during acute
Foundation/American Heart Association Task nonparoxysmal atrial fibrillation: Results of 3-year atrial fibrillation in the isolated sheep heart. Circu-
Force on Practice Guidelines developed in partner- follow-up. Circ Arrhythm Electrophysiol 5:514– lation 103:2631–2636, 2001.
ship with the European Society of Cardiology and 520, 2012. 13. Berenfeld O, Zaitsev AV, Mironov SF, et al:
in collaboration with the European Heart Rhythm 7. Jalife J: Deja vu in the theories of atrial fibrillation Frequency-dependent breakdown of wave propa-
Association and the Heart Rhythm Society. J Am dynamics. Cardiovasc Res 89:766–775, 2011. gation into fibrillatory conduction across the pec-
Coll Cardiol 57:e101–e198, 2011. 8. Atienza F, Martins RP, Jalife J: Translational tinate muscle network in the isolated sheep right
2. Camm J: Antiarrhythmic drugs for the mainte- research in atrial fibrillation: A quest for mechanis- atrium. Circ Res 90:1173–1180, 2002.
nance of sinus rhythm: risks and benefits. Int J tically based diagnosis and therapy. Circ Arrhythm 14. Kalifa J, Tanaka K, Zaitsev AV, et al: Mechanisms
Cardiol 155:362–371, 2012. Electrophysiol 5:1207–1215, 2012. of wave fractionation at boundaries of high-
3. Haissaguerre M, Jais P, Shah DC, et al: Spontane- 9. Skanes AC, Mandapati R, Berenfeld O, et al: Spa- frequency excitation in the posterior left atrium of
ous initiation of atrial fibrillation by ectopic beats tiotemporal periodicity during atrial fibrillation in the isolated sheep heart during atrial fibrillation.
originating in the pulmonary veins. N Engl J Med the isolated sheep heart. Circulation 98:1236– Circulation 113:626–633, 2006.
339:659–666, 1998. 1248, 1998. 15. Sanders P, Berenfeld O, Hocini M, et al: Spectral
4. Cappato R, Calkins H, Chen SA, et al: Updated 10. Berenfeld O, Mandapati R, Dixit S, et al: Spatially analysis identifies sites of high-frequency activity
worldwide survey on the methods, efficacy, and distributed dominant excitation frequencies reveal maintaining atrial fibrillation in humans. Circula-
safety of catheter ablation for human atrial hidden organization in atrial fibrillation in the tion 112:789–797, 2005.
16. Atienza F, Almendral J, Moreno J, et al: Activation 35. Gray RA, Pertsov AM, Jalife J: Incomplete reentry 57. Zlochiver S, Munoz V, Vikstrom KL, et al: Elec-
of inward rectifier potassium channels accelerates and epicardial breakthrough patterns during atrial trotonic myofibroblast-to-myocyte coupling
atrial fibrillation in humans: Evidence for a reen-
trant mechanism. Circulation 114:2434–2442,
2006.
fibrillation in the sheep heart. Circulation 94:2649–
2661, 1996.
36. Sarmast F, Kolli A, Zaitsev A, et al: Cholinergic
increases propensity to reentrant arrhythmias in
two-dimensional cardiac monolayers. Biophys J
95:4469–4480, 2008.
42
17. Atienza F, Almendral J, Jalife J, et al: Real-time atrial fibrillation: I-K,I-ACh gradients determine 58. Burstein B, Nattel S: Atrial fibrosis: Mechanisms
dominant frequency mapping and ablation of dom- unequal left/right atrial frequencies and rotor and clinical relevance in atrial fibrillation. J Am
inant frequency sites in atrial fibrillation with left- dynamics. Cardiovasc Res 59:863–873, 2003. Coll Cardiol 51:802–809, 2008.
to-right frequency gradients predicts long-term 37. Ortiz J, Niwano S, Abe H, et al: Mapping the 59. Kneller J, Zou R, Vigmond EJ, et al: Cholinergic
maintenance of sinus rhythm. Heart Rhythm 6:33– conversion of atrial flutter to atrial fibrillation and atrial fibrillation in a computer model of a two-
40, 2009. atrial fibrillation to atrial flutter: Insights into dimensional sheet of canine atrial cells with realis-
18. Atienza F, Calvo D, Almendral J, et al: Mechanisms mechanisms. Circ Res 74:882–894, 1994. tic ionic properties. Circ Res 90:E73–E87, 2002.
of fractionated electrograms formation in the pos- 38. Ikeda T, Wu TJ, Uchida T, et al: Meandering and 60. Pandit SV, Berenfeld O, Anumonwo JMB, et al:
terior left atrium during paroxysmal atrial fibrilla- unstable reentrant wave fronts induced by acetyl- Ionic determinants of rotor dynamics during
tion in humans. J Am Coll Cardiol 57:1081–1092, choline in isolated canine right atrium. Am J chronic atrial fibrillation in humans: a simulation
2011. Physiol 273:H356–H370, 1997. study. Biophys J 88:3806–3821, 2005.
19. de Groot NM, Houben RP, Smeets JL, et al: Elec- 39. Schuessler RB, Grayson TM, Bromberg BI, et al: 61. Voigt N, Trausch A, Knaut M, et al: Left-to-right
tropathological substrate of longstanding persis- Cholinergically mediated tachyarrhythmias atrial inward rectifier potassium current gradients
tent atrial fibrillation in patients with structural induced by a single extrastimulus in the isolated in patients with paroxysmal versus chronic atrial
heart disease: Epicardial breakthrough. Circulation canine right atrium. Circ Res 71:1254–1267, fibrillation. Circ Arrhythm Electrophysiol 3:472–
122:1674–1682, 2010. 1992. 480, 2010.
20. Jais P, Haissaguerre M, Shah DC, et al: A focal 40. Khositseth A, Clapham DE, Ackerman MJ: Intra- 62. Caballero R, de la Fuente MG, Gomez R, et al: In
source of atrial fibrillation treated by discrete cellular signaling and regulation of cardiac ion humans, chronic atrial fibrillation decreases the
radiofrequency ablation. Circulation 95:572–576, channels. In Zipes DP, Jalife J, editors: Cardiac transient outward current and ultrarapid compo-
1997. Electrophysiology: From Cell to Bedside, Philadel- nent of the delayed rectifier current differentially
21. Wakili R, Voigt N, Kaab S, et al: Recent advances phia, 2004, WB Saunders. on each atria and increases the slow component of
in the molecular pathophysiology of atrial fibrilla- 41. Kabell G, Buchanan LV, Gibson JK, et al: Effects the delayed rectifier current in both. J Am Coll
tion. J Clin Invest 121:2955–2968, 2011. of adenosine on atrial refractoriness and arrhyth- Cardiol 55:2346–2354, 2010.
22. Machida T, Hashimoto N, Kuwahara I, et al: mias. Cardiovasc Res 28:1385–1389, 1994. 63. Dobrev D, Friedrich A, Voigt N, et al: The G
Effects of a highly selective acetylcholine-activated 42. Belardinelli L, Shryock JC, Song Y, et al: Ionic protein-gated potassium current IK,ACh is consti-
K+ channel blocker on experimental atrial fibrilla- basis of the electrophysiological actions of adenos- tutively active in patients with chronic atrial fibril-
tion. Circ Arrhythm Electrophysiol 4:94–102, ine on cardiomyocytes. FASEB J 9:359–365, 1995. lation. Circulation 112:3697–3706, 2005.
2011. 43. Zaza A, Rocchetti M, DiFrancesco D: Modulation 64. Dhamoon AS, Jalife J: The inward rectifier current
23. Diness JG, Sørensen US, Nissen JD, et al: Inhibi- of the hyperpolarization-activated current (I(f)) by IK1 controls cardiac excitability and is involved in
tion of small-conductance Ca2+-activated K+ adenosine in rabbit sinoatrial myocytes. Circula- arrhythmogenesis. Heart Rhythm 2:316–324,
channels terminates and protects against atrial tion 94:734–741, 1996. 2005.
fibrillation. Circ Arrhythm Electrophysiol 3:380– 44. Wang D, Shryock JC, Belardinelli L: Cellular basis 65. Yamauchi S, Ogasawara H, Saji Y, et al: Efficacy of
390, 2010. for the negative dromotropic effect of adenosine on intraoperative mapping to optimize the surgical
24. Amit G, Kikuchi K, Greener ID, et al: Selective rabbit single atrioventricular nodal cells. Circ Res ablation of atrial fibrillation in cardiac surgery. Ann
molecular potassium channel blockade prevents 78:697–706, 1996. Thorac Surg 74:450–457, 2002.
atrial fibrillation. Circulation 121:2263–2270, 45. Sahadevan J, Ryu K, Peltz L, et al: Epicardial 66. Drewitz I, Willems S, Salukhe TV, et al: Atrial
2010. mapping of chronic atrial fibrillation in patients: fibrillation cycle length is a sole independent pre-
25. Dobrev D, Nattel S: New insights into the molecu- Preliminary observations. Circulation 110:3293– dictor of a substrate for consecutive arrhythmias in
lar basis of atrial fibrillation: Mechanistic and 3299, 2004. patients with persistent atrial fibrillation. Circ
therapeutic implications. Cardiovasc Res 89:689– 46. Lazar S, Dixit S, Marchlinski FE, et al: Presence Arrhythm Electrophysiol 3:351–360, 2010.
691, 2011. of left-to-right atrial frequency gradient in parox- 67. Nademanee K, McKenzie J, Kosar E, et al: A new
26. Cheng Y, Zhan Q, Zhao J, Xiao J: Stabilizing ysmal but not persistent atrial fibrillation in approach for catheter ablation of atrial fibrillation:
ryanodine receptor type 2: A novel strategy for the humans. Circulation 110:3181–3186, 2004. Mapping of the electrophysiologic substrate. J Am
treatment of atrial fibrillation. Med Sci Monit 47. Burashnikov A, Antzelevitch C: Reinduction of Coll Cardiol 43:2044–2053, 2004.
16:HY23–JY26, 2010. atrial fibrillation immediately after termination of 68. Lin YJ, Tai CT, Kao T, et al: Frequency analysis in
27. Morillo CA, Klein GJ, Jones DL, et al: Chronic the arrhythmia is mediated by late phase 3 early different types of paroxysmal atrial fibrillation.
rapid atrial pacing: Structural, functional, and elec- afterdepolarization-induced triggered activity. Cir- J Am Coll Cardiol 47:1401–1407, 2006.
trophysiological characteristics of a new model of culation 107:2355–2360, 2003. 69. Lazar S, Dixit S, Callans DJ, et al: Effect of pul-
sustained atrial fibrillation. Circulation 91:1588– 48. Patterson E, Lazzara R, Szabo B, et al: Sodium- monary vein isolation on the left-to-right atrial
1595, 1995. calcium exchange initiated by the Ca2_ transient: dominant frequency gradient in human atrial fibril-
28. Harada A, Sasaki K, Fukushima T, et al: Atrial An arrhythmia trigger within pulmonary veins. lation. Heart Rhythm 3:889–895, 2006.
activation during chronic atrial fibrillation in J Am Coll Cardiol 47:1196–1206, 2006. 70. Lemola K, Ting M, Gupta P, et al: Effects of two
patients with isolated mitral valve disease. Ann 49. Atienza F, Jalife J: Reentry and atrial fibrillation. different catheter ablation techniques on spectral
Thorac Surg 61:104–112, 1996. Heart Rhythm 4(3 Suppl):S13–S16, 2007. characteristics of atrial fibrillation. J Am Coll
29. Kumagai K, Uno K, Khrestian C, et al: Single site 50. Jalife J, Berenfeld O, Skanes A, et al: Mechanisms Cardiol 48:340–348, 2006.
radiofrequency catheter ablation of atrial fibrilla- of atrial fibrillation: Mother rotors or multiple 71. Matsuo S, Yamane T, Date T, et al: Real-time
tion: Studies guided by simultaneous multisite daughter wavelets, or both? J Cardiovasc Electro- dominant frequency analysis of the pulmonary vein
mapping in the canine sterile pericarditis model. physiol 9(8 Suppl):S2–S12, 1998. in patients with paroxysmal atrial fibrillation.
J Am Coll Cardiol 36:917–923, 2000. 51. Rohr S: Myofibroblasts in diseased hearts: new Pacing Clin Electrophysiol 35:28–37, 2012.
30. Haissaguerre M, Sanders P, Hocini M, et al: players in cardiac arrhythmias? Heart Rhythm 72. Schuessler RB, Kay MW, Melby SJ, et al: Spatial
Changes in atrial fibrillation cycle length and 6:848–856, 2009. and temporal stability of the dominant frequency
inducibility during catheter ablation and their rela- 52. Everett TH, Olgin JE: Atrial fibrosis and the of activation in human atrial fibrillation. J Electro-
tion to outcome. Circulation 109:3007–3013, mechanisms of atrial fibrillation. Heart Rhythm cardiol 39:S7–S12, 2006.
2004. 4(3 Suppl):S24–S27, 2007. 73. Oral H, Chugh A, Good E, et al: Radiofrequency
31. Berenfeld O: Quantifying activation frequency in 53. Weber KT, Pick R, Jalil JE, et al: Patterns of myo- catheter ablation of chronic atrial fibrillation
atrial fibrillation to establish underlying mecha- cardial fibrosis. J Mol Cell Cardiol 21(Suppl guided by complex electrograms. Circulation
nisms and ablation guidance. Heart Rhythm 5):121–131, 1989. 115:2606–2612, 2007.
4:1225–1234, 2007. 54. Anyukhovsky EP, Sosunov EA, Plotnikov A, et al: 74. Oral H, Chugh A, Good E, et al: Randomized
32. Berenfeld O, Ennis S, Hwang E, et al: Time- and Cellular electrophysiologic properties of old canine evaluation of right atrial ablation after left atrial
frequency-domain analyses of atrial fibrillation atria provide a substrate for arrhythmogenesis. ablation of complex fractionated atrial electro-
activation rate: the optical mapping reference. Cardiovasc Res 54:462–469, 2002. grams for long-lasting persistent atrial fibrillation.
Heart Rhythm 8:1758–1765, 2011. 55. Tanaka K, Zlochiver S, Vikstrom KL, et al: Spatial Circ Arrhythm Electrophysiol 1:6–13, 2008.
33. Antz M, Otomo K, Arruda M, et al: Electrical con- distribution of fibrosis governs fibrillation wave 75. Oral H, Chugh A, Yoshida K, et al: A randomized
nections between the right atrium and the left dynamics in the posterior left atrium during heart assessment of the incremental role of ablation of
atrium via the musculature of the coronary sinus. failure. Circ Res 101:839–847, 2007. complex fractionated atrial electrograms after
Circulation 98:1790–1795, 1998. 56. Swartz MF, Fink GW, Lutz CJ, et al: Left versus antral pulmonary vein isolation for long-lasting
34. Dolber PC, Spach MS: Structure of the canine right atrial difference in dominant frequency, K(+) persistent atrial fibrillation. J Am Coll Cardiol
Bachmann’s bundle related to propagation of channel transcripts, and fibrosis in patients devel- 53:782–789, 2009.
excitation. Am J Physiol 257:H1446–H1457, oping atrial fibrillation after cardiac surgery. Heart 76. Di Biase L, Elayi CS, Fahmy TS, et al: Atrial fibril-
1989. Rhythm 6:1415–1422, 2009. lation ablation strategies for paroxysmal patients:
Randomized comparison between different tech- 83. Zlochiver S, Yamazaki M, Kalifa J, et al: Rotor 91. Rostock T, Rotter M, Sanders P, et al: High-
niques. Circ Arrhythm Electrophysiol 2:113–119, meandering contributes to irregularity in electro- density activation mapping of fractionated electro-
2009. grams during atrial fibrillation. Heart Rhythm grams in the atria of patients with paroxysmal atrial
77. Hayward RM, Upadhyay GA, Mela T, et al: Pul- 5:846–854, 2008. fibrillation. Heart Rhythm 3:27–34, 2006.
monary vein isolation with complex fractionated 84. Konings KT, Kirchhof CJ, Smeets JR, et al: High- 92. Knecht S, Wright M, Matsuo S, et al: Impact of
atrial electrogram ablation for paroxysmal and density mapping of electrically induced atrial fibril- pharmacological autonomic blockade on complex
nonparoxysmal atrial fibrillation: A meta-analysis. lation in humans. Circulation 89:1665–1680, 1994. fractionated atrial electrograms. J Cardiovasc Elec-
Heart Rhythm 8:994–1000, 2011. 85. Roberts-Thomson KC, Stevenson IH, Kistler PM, trophysiol 21:766–772, 2010.
78. Dixit S, Marchlinski FE, Lin D, et al: Randomized et al: Anatomically determined functional conduc- 93. Klos M, Calvo D, Yamazaki M, et al: Atrial
ablation strategies for the treatment of persistent tion delay in the posterior left atrium. J Am Coll septopulmonary bundle of the posterior left atrium
atrial fibrillation: RASTA study. Circ Arrhythm Cardiol 51:856–862, 2008. provides a substrate for atrial fibrillation initiation
Electrophysiol 5:287–294, 2012. 86. Lin J, Scherlag BJ, Zhou J, et al: Autonomic mech- in a model of vagally mediated pulmonary
79. Calkins H, Kuck KH, Cappato R, et al: 2012 HRS/ anism to explain complex fractionated atrial elec- vein tachycardia of the structurally normal
EHRA/ECAS Expert Consensus Statement on trograms (CFAE). J Cardiovasc Electrophysiol heart. Circ Arrhythm Electrophysiol 1:175–183,
Catheter and Surgical Ablation of Atrial Fibrilla- 18:1197–1205, 2007. 2008.
tion: Recommendations for patient selection, pro- 87. Ortiz J, Niwano S, Abe H, et al: Mapping the 94. Jalife J, Atienza F, Lopez-Salazar B, et al: Molecu-
cedural techniques, patient management and conversion of atrial flutter to atrial fibrillation and lar, cellular and pathophysiological mechanisms of
follow-up, definitions, endpoints, and research trial atrial fibrillation to atrial flutter: Insights into human atrial fibrillation. Nat Clin Pract Cardio-
design. Europace 14:528–606, 2012. mechanisms. Circ Res 74:882–894, 1994. vasc Med 6:15–21, 2009.
80. de Bakker JM, Wittkampf FH: The pathophysio- 88. Narayan SM, Wright M, Derval N, et al: Classify- 95. Narayan SM, Krummen DE, Rappel WJ: Clinical
logic basis of fractionated and complex electro- ing fractionated electrograms in human atrial mapping approach to diagnose electrical rotors
grams and the impact of recording techniques on fibrillation using monophasic action potentials and and focal impulse sources for human atrial fibrilla-
their detection and interpretation. Circ Arrhythm activation mapping: Evidence for localized drivers, tion. J Cardiovasc Electrophysiol 23:447–454,
Electrophysiol 3:204–213, 2010. rate acceleration, and nonlocal signal etiologies. 2012.
81. Aizer A, Holmes DS, Garlitski AC, et al: Standard- Heart Rhythm 8:244–253, 2011. 96. Berenfeld O, Guillem M, Climent A, et al: Deter-
ization and validation of an automated algorithm 89. Jadidi AS, Duncan E, Miyazaki S, et al: Functional mination of maximal frequency sites and frequency
to identify fractionation as a guide for atrial fibril- nature of electrogram fractionation demonstrated gradients in patients with atrial fibrillation by
lation ablation. Heart Rhythm 5:1134–1141, by left atrial high-density mapping. Circ Arrhythm body surface potential mapping. Heart Rhythm
2008. Electrophysiol 5:32–42, 2012. 9(5S):S487, 2012.
82. Roux JF, Gojraty S, Bala R, et al: Effect of pulmo- 90. Tai CT, Chen SA, Tzeng JW, et al: Prolonged 97. Cuculich PS, Wang Y, Lindsay BD, et al: Nonin-
nary vein isolation on the distribution of complex fractionation of paced right atrial electrograms in vasive characterization of epicardial activation in
fractionated electrograms in humans. Heart patients with atrial flutter and fibrillation. J Am humans with diverse atrial fibrillation patterns.
Rhythm 6:156–160, 2009. Coll Cardiol 37:1651–1657, 2001. Circulation 122:1364–1372, 2010.
Rotors in Human Atrial Fibrillation
Sanjiv M. Narayan and Wouter-Jan Rappel
43
for human AF, and conclude with current clinical applications of
CHAPTER OUTLINE
this approach.
Triggers Interact With Sustaining Mechanisms
to Cause Human Atrial Fibrillation 433
Mapping Rotors During Human Atrial Fibrillation 433 Triggers Interact With Sustaining Mechanisms
Sustaining Rotors and Focal Impulses for Human to Cause Human Atrial Fibrillation
Atrial Fibrillation 435
The fact that AF may be triggered by ectopic beats1 draws an
Clinical Implications 438 intuitively attractive parallel with the mechanisms for “simple”
Conclusions 440 supraventricular (SVT) and ventricular (VT) tachycardias, which
are also triggered by ectopic beats. In those arrhythmias, triggers
engage specific mechanisms of dual atrioventricular (AV) nodal
physiology to maintain AV node reentry,17 an accessory pathway
Mechanistic concepts for human atrial fibrillation (AF) have for maintaining AV reciprocating tachycardia,18 or a slowly con-
evolved substantially. In the past 15 years, meticulous mapping ducting isthmus to maintain ventricular tachycardia.19
has established that human AF may be triggered by ectopic focal We hypothesized that AF may be caused when triggers engage
beats, while the combination of bioengineering analyses with specific AF-maintaining mechanisms that may be created dynam-
detailed mapping has recently provided compelling evidence that ically via conduction block, leading to localized reentry, or by
AF, once triggered, is maintained by relatively few stable electri- triggering repetitive focal beats. To address this hypothesis, we
cal rotors and focal sources that lie in patient-specific locations. studied the rate dependence of monophasic action potential dura-
These advances in our mechanistic understanding have already tion (APD) in the left and right atria, and bi-atrial patterns of
led to advances in therapy and improved outcomes for patients rate-dependent bi-atrial conduction slowing in the left and right
with AF. atria.20,21 and propagation en route to AF using multipolar basket
In seminal work, Haïssaguerre1 reported that localized ectopy catheters.
from the pulmonary veins (PVs) may trigger AF. This discovery In a series of studies, we recently demonstrated that human
launched the field of potentially curative AF ablation, with PV AF onset is consistently preceded by alternans and complex oscil-
isolation as its cornerstone.2 Nevertheless, the mechanisms that lations in APD, which create a milieu of heightened repolariza-
perpetuate AF, once triggered, remained undefined.3,4 The mul- tion dispersion immediately before AF onset. In patients with
tiwavelet hypothesis proposed that multiple spatially meandering persistent and paroxysmal AF, we have found that sustained rapid
electrical waves cause AF.5 However, this did not explain consis- rates produce marked alternans and complex oscillations in
tent patterns of propagation observed in patients with AF,6,7 the APD22 preceding AF,23 independently of APD restitution. In par-
termination of AF after localized ablation,2,8 or the opposite allel, bi-atrial conduction velocity slows dynamically (restitution)
finding—that extensive ablation to constrain wavelets may have just before AF onset at the location where AF is initiated.24
little acute impact.2,9 The alternative localized source hypothesis is In summary, these observations show that the onset of human
based on animal and in silico experiments in which rapid localized AF is preceded by dynamic localized conduction slowing and
spiral waves (rotors)4,10 or focal sources7 directly cause AF. exaggerated repolarization dispersion, both of which may be
Although rotors in human AF have been disputed,5 and direct subtle or concealed at baseline. These conditions favor reentry
evidence has been lacking,4 rotors have been indirectly suggested and have provided the functional underpinnings for our studies
at sites of high dominant frequency where ablation terminates of human AF mechanisms.
paroxysmal AF11,12 and, if interatrial rate gradients are abolished,
improves outcome,11 and by AF mapping.13
Recent data from our14 and other laboratories15 show that
human AF is indeed perpetuated by a small number of stable
Mapping Rotors During Human
rotors or focal sources14,15 in individuals with paroxysmal, persis- Atrial Fibrillation
tent, and long-standing persistent AF. The mechanistic role of
stable AF source rotors has been supported by the acute termina- Prior Mapping Studies of Human AF
tion and subsequent non-reinducibility of AF by brief targeted
ablation (focal impulse and rotor modulation [FIRM]) before any Several mechanisms for the maintenance of human AF have been
other intervention. Clinically, this approach has been shown to proposed, including disorganized multiwavelet reentry,5,25 spa-
substantially improve the long-term elimination of AF compared tially localized reentrant26 or focal7 sources, and mixed patterns.13
with conventional ablation alone.16 Divergence in these hypotheses in large part may reflect mapping
In this chapter, we summarize the evidence for localized elec- that has not always met “classical” requirements: to broadly map
trical rotors and focal sources for human AF, in the context of chambers of interest, at sufficient spatiotemporal resolution to
the historical background in this area. We then describe identify varying patterns, over long enough periods to capture
approaches to, and results from, mapping rotors and focal sources variability, then to use interventions to demonstrate that
433
proposed mechanisms are causal and are not bystanders. Failure

to apply these criteria even to simple supraventricular17,18,27 and
ventricular19 arrhythmias is a well-recognized cause of incorrect
diagnosis and potentially undesirable therapy.
λ
Many clinical studies over the past decade show that human
AF is spatially non-uniform. For instance, human AF exhibits
Rrotor
consistent activation patterns,28 consistent rate, or dominant fre-
quency gradients within and between atria,6,29 as well as electro-
cardiographic (ECG) spectra suggesting conserved global 2Rlocus
spatiotemporal organization for at least days30 within and between
Figure 43-1. Snapshot of a Computer Simulation in Homogeneous
patients. These data have long supported the notion that human
Tissue A rotor, the locus of migration of its tip, and its coherent domain of rota-
AF is maintained by spatially localized mechanisms, further sup- tional activity (i.e., tissue controlled in a 1 : 1 fashion by the rotor). The model
ported by the fact that human AF may terminate with ablation parameters were chosen such that beyond this region of tissue, the rotor breaks
at defined triggers, drivers,1 and other localized regions.12,31 Such down into other spatiotemporal structures as the result of discordant alternans. The
regions may arise in either atrium32 and are difficult to identify a relevant length scales in the problem are the size of the domain of locus migration
priori, but can be ablated in both atria by a systematic stepwise of the rotor tip, Rlocus, the coherent domain (of detectable rotational activation), Rrotor,
approach.8 and the wavelength of the rotor, λ.
Higher-resolution mapping, on the other hand, has produced
surprisingly inconsistent results. In seminal intraoperative
mapping studies in AF patients, Cox, Schuessler, et al33,34 found Figure 43-1 also illustrates the importance of mapping a suf-
stable reentry within disordered AF that were interrupted by ficiently large field-of-view. Attempting to map a rotor with a
lesions that formed the basis for the Maze procedure35 and that field-of-view smaller than its trajectory of migration (using a
still underpin many current ablation lesion sets.2 Conversely, in small mapping plaque) may lead to results that are difficult to
separate human AF studies,36 Allessie et al found no consistency interpret. Hence, if the locations of putative AF sources are
within disordered AF and concluded that AF was attributed to unknown, then as much of the atrial surface as possible should
multiple reentrant waves (as in early computer models37) with be mapped. Finally, the temporal resolution must be able to
“focal” events reflective of transmural breakthrough. However, distinguish activation between neighboring recording sites and
these studies mapped <20% of the dilated atria in these patients, thus can be found by dividing the spatial resolution by the
and did not apply interventions to prove causality of disordered dynamic conduction velocity. For example, for a spatial resolu-
activation. More recent epicardial mapping has revealed localized tion of 5 mm and a range of conduction velocity of 50 to
high-frequency regions in AF patients consistent with sources.7,38 150 cm/s,42 the required temporal resolution is 3.3 to 10 ms.
Schilling et al39 and more recently Cuculich et al13 used mathe- We estimated the length scales λ, Rlocus, and Rrotor, and thus
matical inverse solutions to map the atria via noncontact the required spatial resolution, to map potential localized sources
approaches (Ensite 3000™, St Jude Medical, Minneapolis, Min- of human AF, based on animal studies and observations of human
nesota, and EcVue™, Cardioinsight, Cleveland, Ohio, respec- AF. Animal models of AF show varying mechanisms, including
tively) in AF.13 However, such studies have yet to use focused localized spiral waves (rotors),43,44 focal sources,45 or nonlocalized
ablation to establish causality of each proposed mechanism and waves.10 In experiments showing rotors, the length scale for the
exclude bystanders. reentrant path (Rlocus) ranges from 1 cm to >3 cm,10,46,47 requiring
a minimum resolution of 0.5 cm for mapping. Spatial organiza-
tion in some models controlled tissue areas >5 cm2,44 correspond-
Requirements to Map Human AF ing to a rotor length scale and wavelength >2 cm. Thus, once the
location of a rotor is identified, reentry requires mapping a field-
Without a priori knowledge of the spatiotemporal organization of-view of at least 2.5 × 2.5 cm, with a resolution of ≈1 cm.
of human AF, it has been difficult to propose “design” require- In humans, the length scale of the reentrant path may be
ments for mapping. In 2001, we set out to study human AF by estimated from the concept of tissue wavelength48 as the product
designing an approach suitable for mapping spatially meandering of minimum conduction velocity and the shortest refractory
multiple wavelets or for identifying potential localized sources period. In AF patients, minimum (dynamic) conduction velocity
(rotors or focal sources). Our initial hypothesis was that localized in left and right atria is ≈40 cm/s42 and the minimum atrial refrac-
sources do not exist in human AF. tory period ≈100 to 110 ms,20,23 resulting in a minimum wave-
length of ≈4 to 5 cm, diameter of ≈1.5 cm, and minimum required
Detection Design Requirements spatial resolution of ≈1 cm.
We reasoned that the factor most relevant to mapping spatial
resolution is the length scale of the mapped event. Figure 43-1 Numeric Validation of Design Requirements for AF Mapping
illustrates a rotor with limited movement of the rotor tip (in a To test these design criteria for mapping human AF, we per-
locus of migration) and wavebreak to fibrillation, generated in a formed in silico data validation using an electrophysiological
computer simulation. Electrophysiological model parameters model. The model simulates wave propagation using the monodo-
were chosen such that the spiral breaks down far from the migra- main equation:
tion locus.3,4,23 The rotor controls activation in a “coherent dV
domain” in 1 : 1 fashion with length scale Rrotor, beyond which = ∇ ⋅ D∇V − I ion C m
activation breaks down into complex spatiotemporal patterns. A dt
coherent rotor domain surrounded by incoherent activity can where V is the membrane voltage, Cm represents the membrane
also be generated in simulations of heterogeneous tissue.40,41 To capacitance, D is the diffusion tensor, and Iion represents the
map the rotor core, the required spatial resolution is comparable membrane currents. For the purposes of illustrating propagation
with the length scale of the reentrant path Rlocus. It is important in silico, we computed membrane currents using the 3- and
to note, however, that to simply detect rotational activity around 4-variable Fenton-Karma (FK) model.49,50
the core (the rotor), the required spatial resolution is coarser Figure 43-2 illustrates spiral wave reentry in a 200 × 200-node
and is comparable with its wavelength, λ, which is much larger simulation area with a physical size of 5 × 5 cm, corresponding
than Rlocus. to a spatial resolution Δx of 0.25 mm (Figure 43-2, A). The spiral
Rotors in Human Atrial Fibrillation 435
Sustaining Rotors and Focal Impulses

0 20
40
0
20
40
for Human Atrial Fibrillation
43
16 15
Recently, our group and others showed that human paroxys-
mal, persistent, and long-standing persistent AF are predomi-
nantly sustained by localized rotors or focal sources. In a
multicenter experience that currently includes more than 200
patients, stable rotor or focal sources were identified in >98% of
patients, in diverse bi-atrial locations that were stable for pro-
60
longed periods in each individual. Direct targeted ablation was
A B 60 applied to each mapped source (FIRM) before any other inter-
vention, which led acutely to the termination or substantial orga-
nization of AF in the vast majority of cases. Patients who received
20
0 0 20 FIRM together with conventional ablation had substantially
40 greater freedom from AF on rigorous long-term follow-up com-
40
pared with those receiving conventional ablation alone.
Identification of Rotors and Focal Impulses

in Human AF: Focal Impulse and Rotor
60 Mapping
60
C D We designed FIRM mapping on the basis of the above consider-
ations including the requirement to apply ablation to prove pro-
Figure 43-2. Effect of Mapping Resolution on a Simulated Clockwise posed AF mechanisms.14 Accordingly, FIRM mapping is
Rotor Simulation on a 200 × 200 grid. A, Spatial resolution Δx = 0.25 mm, corre- performed during clinical electrophysiology study by advancing
sponding to a 5 × 5-cm domain. The activation is plotted using a gray scale, with 64-pole basket catheters to the right atrium and, after transseptal
white corresponding to depolarized and black corresponding to repolarized tissue. puncture, to the left atrium. Contact electrodes provide spatial
The tip of the rotor meanders as shown in red. The green symbols are isochrones resolution of 4 to 6 mm along each spline, and ≈4 to 10 mm
superimposed onto the snapshots. These isochrones are 20 ms apart and are
between splines (higher resolution at the poles than at the
computed as all grid points activated within 1 ms. Scalebar = 1 cm. B, Isochrones
computed on a 20 × 20 grid obtained by coarsening the original grid. Isochrones
equator). Heparin is infused to maintain activated clotting time
are again 20 ms apart with the order green, blue, purple, and red; scalebar repre- >350 ms. In early work, we mapped both atria simultaneously
sents spatial resolution (Δx = 2.5 mm). C, Isochrones computed on an 8 × 8 grid (Figure 43-3), but we now use one basket: first in the right atrium
providing spatial resolution Δx = 6.25 mm. D, Isochrones computed on a 4 × 4 grid while performing transseptal puncture, then in the left atrium.
providing spatial resolution Δx = 12.5 mm. Note that rotational activity (organized Multi-site electrograms are recorded with a temporal resolution
domain) of the rotor remains detectable at all resolutions. of 1 ms as unipoles or as overlapping bipoles to reduce far-field
artifact, and are filtered at 0.05 to 500 Hz. AF data are exported
digitally as multiple 1-minute epochs over a period of tens of
is single-armed with a wavelength larger than Lrotor/2 and a period minutes.
of 90 ms. The computed locus of migration of the rotor tip is A novel system (RhythmView™, Topera Inc., Palo Alto, Cali-
illustrated in red in Figure 43-2, A, and consists of a complex fornia) is then used to analyze multipolar AF signals.14,16,51 Elec-
meandering trajectory with a length scale of ≈1 cm. Thus, this trograms are filtered to exclude noise and far-field signals guided
simulated rotor has Llocus ≈1 cm, Lrotor >5 cm, and λ >2.5 cm. by the reported rate dynamics of human right and left atrial
Activation times for each node were determined using a voltage APD20,22,23,52 to estimate minimum activation time, and conduc-
threshold (10% maximal) and were stored at temporal resolution tion velocity20,24 to map propagation paths. The system produces
Δt = 1 ms. Activation times were used to compute isochrones sepa- isopotential movies of AF over successive cycles, as the primary
rated by 20 ms (Figure 43-2, B, green). To simulate coarser record- display modality used to guide ablation, as discussed later.
ing resolution, we coarsened stored activation time intervals but However, for illustration, single-cycle isochronal snapshots are
did not rerun simulations. Figure 43-2, B-D illustrates isochrones provided throughout the chapter.
for the same rotor at resulting spatial resolutions of Δx = 2.5 mm
(20 × 20 grid), Δx = 6.25 mm (8 × 8 grid), and Δx = 12.5 mm (4 × 4
grid), corresponding to isochronal intervals of 3 ms, 7 ms, and 18 Demonstration of AF Rotors and Focal Sources
ms, respectively. Notably, all values of Δx preserved rotational
activity of the organized domain of the rotor. Propagation can be mapped in various rhythms. For orientation,
Figure 43-3 illustrates bi-atrial propagation in sinus rhythm. The
Summary of Design Requirements for Human AF Mapping atria are projected onto grids, with the right atrium cut vertically
These estimates and numeric simulations suggest that a spatial through the tricuspid valve, opening its lateral and medial halves.
resolution of ≈0.5 cm may be able to resolve rotor migration, and The left atrium is cut horizontally through the mitral valve,
a spatial resolution of ≈1 cm may capture rotational activity of a opening its superior and inferior halves. On these projections,
rotor in human AF. The corresponding temporal resolution sinus activation is represented by color-coded isochrones emanat-
should be at least 6 ms. These considerations laid the foundation ing from the sinoatrial node (in red), crossing Bachmann’s bundle
for our direct contact focal impulse and rotor mapping (FIRM) to the left atrium (in blue).
approach,14 which we recently described to identify patient- In AF, isopotential movies were analyzed to identify rotors
specific sustaining rotors and focal sources for human AF. Local- as rotational activity around a center, or focal impulses as cen-
ized AF sources were identified at electrophysiological study and trifugal activation from a point of origin, or laminar activation or
their causal role verified by AF termination by patient-specific disorganized patterns that fell outside the three categories previ-
targeted ablation alone. ously described. Rotors and focal sources were diagnosed only if
Right atrium Left atrium

Anterior
Sinus node Late
Mitral
110
Lateral left atrium

Lateral tricuspid
Septum
0 ms
IVC Inferior Early
1cm Mitral annulus
A
B C
Figure 43-3. Anatomical Reference and Nomenclature for FIRM Mapping A, Fluoroscopic view of 64-pole basket catheters in the right and left atria. An
ablation catheter and a coronary sinus catheter are shown, along with an esophageal temperature probe. The image shows a subcutaneous continuous ECG monitor
(Reveal XT™, Medtronic, Minneapolis, Minnesota) used to categorically establish elimination of AF after ablation in clinical series. B, Sinus rhythm map on bi-atrial schematic.
Activation at basket electrodes, shown as dots, is displayed as a color-coded map from the sinus node to the lateral inferior left atrium. The right atrium is opened between
its poles, with the tricuspid annulus opened laterally and medially; the left atrium is opened along its equator, with the mitral annulus opened superiorly and inferiorly. The
pulmonary vein ostia are indicated by dashed lines.
RIGHT ATRIAL AF ROTOR WITH ELECTROGRAMS IN ORGANIZED DOMAIN LEFT ATRIAL AF ROTOR WITH ELECTROGRAMS IN ORGANIZED DOMAIN
ROTOR Superior vena cava 177 ms Superior mitral I

I Late aVF
Late B
1 V1 180 V1
180 C 180 ms
aVF 160
Activation times/ms
2 160
Activation times/ms
D 140 E3
3 140 D2
Lateral
Medial tricuspid
Septal
Lateral tricuspid
120 D3
120 E
4 C3 100 D4
100 F
C4 80 D6
5 80
D6 G 60 E6
6 60
40
40 E5 H
7 20 G5
20 A
F4 G3
8 8 7 6 5 4 3 2 1 Early
A B C D E F G H Early F3
F2 1:1 activation Inferior mitral
Inferior vena cava 1:1 activation breaks down ROTOR
0 400 800 1200 1600 2000 1000 1400 1800 2200 2600 3000
breaks down
A Time, ms B Time/ms
Figure 43-4. Right Atrial and Left Atrial Spiral Waves (Rotors) During Human AF on FIRM Mapping A, Counterclockwise rotor in posterolateral right
atrium, with collision (white double lines) beyond the coherent domain of the spiral arm. Electrograms around the right atrium rotor site indicate sequential activation of
counterclockwise rotation with variability at cycle length ≈177 ms. B, Counterclockwise rotor in midposterior left atrium, with collision beyond the spiral arm (double lines).
Pulmonary vein ostia indicated by dashed lines. ABL, Ablation electrogram; AF, atrial fibrillation; CS, coronary sinus.
stable and sustained throughout several recording epochs (>100 or collision. In each case, electrograms in the coherent domain
cycles), to exclude transient rotational and/or focal activations exhibit sequential 1 : 1 activation.
that represent fibrillatory conduction.5 Figure 43-5, A illustrates a right atrial AF rotor with disorga-
We observed sustained rotors and focal sources in nearly all nization peripherally and in the left atrium. Figure 43-5, B illus-
patients (98/101, or 98%) with paroxysmal, persistent, and long- trates, for comparison, right atrial clockwise reentry in reverse
standing persistent AF.16 Subjects exhibited 2.1 ± 1.0 sources typical atrial flutter, showing 1 : 1 propagation in both atria.
concurrently that were more prevalent in individuals with persis- Figure 43-5, C shows a repetitive focal source driving AF with
tent than paroxysmal AF (2.2 ± 1.0 vs. 1.7 ± 0.9; P = .03). AF peripheral disorganization. For comparison, Figure 43-5, D illus-
sources were sustained for hundreds of cycles analyzed in “time trates the distinct bi-atrial propagation maps of a “simple” focal
lapse” fashion, representing tens of minutes (i.e., thousands of atrial tachycardia in a patient without AF, in which propagation
cycles). Sources lay in diverse atrial locations, and it was surpris- is 1 : 1 throughout ipsilateral and contralateral atria.
ing to note that 24% were found in the right atrium. The preva-
lence of rotors was greater than the prevalence of focal beats.14
In patients in whom both spontaneous AF and induced AF were
observed, AF propagation movies converged once AF had been Electrophysiological Evidence That Rotors
sustained for >10 minutes in either case. No complications were and Focal Sources Are Primary Sustaining
noted during mapping.16 Mechanisms for AF
Figure 43-4 illustrates rotors in the (A) right and (B) left atria
in patients during AF, showing head-to-tail (red-to-blue) activa- Two lines of evidence were used to prove that rotors and focal
tion in the coherent domain, and peripheral disorganization and/ sources are causal mechanisms for human AF and not mere
RA AF ROTOR (CW) RA (REVERSE TYPICAL FLUTTER)
Rotor
Right atrium Left atrium Left PVs
Late
200
RA LA Late
200 43
Lateral tricuspid value
Activation times, ms
Lateral
Septal
Macro
reentry
0 0
Inferior vena cava Posterior MV Early Early
ECG I ECG I
CS
A CS B
LA FOCAL AF SOURCE LA FOCAL ATRIAL TACHYCARDIA
Right atrium Left atrium Right atrium Left atrium
Superior vena cava Anterior MV Left PVs Superior vena cava Anterior MV
100 300

Lateral
Septal
Septal
0 0
Inferior vena cava Posterior MV Inferior vena cava Posterior MV
C D
Figure 43-5. AF Sources Are Distinct from Simple Rhythms on FIRM Mapping A, Clockwise right atrial rotor in AF, showing noncoherent (non-1 : 1) propa-
gation in left atrium. Conversely, (B) clockwise reverse typical flutter with 1 : 1 propagation to the left atrium. C, Repetitive focal beat during paroxysmal AF (in low septal
left atrium), with activation to remaining left atrium and fibrillatory conduction to right atrium (CL ≈100 ms). In contrast, (D) focal atrial tachycardia (nonfibrillatory) from
the high posterior left atrium differs from AF by showing 1 : 1 activation centrifugally to the ipsilateral then the contralateral atria (CL, 300 ms).
bystanders. First, we carefully mapped activation to show that a wide variety of energy sources for FIRM ablation cases, includ-
activation emanates from sustained sources to the remaining atria ing cryoenergy. In each case, the catheter is maneuvered near
for each cycle of AF. Second, we targeted rotors and/or focal basket electrodes overlying each source using fluoroscopy, then
sources directly for ablation to acutely modulate AF. ablation is applied for ≈30 s at each site. The catheter is moved
within the area representing the center of rotation or focal origin,
Mapping Activation Sequence for the acute end point of AF termination or ≤10 minutes, which-
Rotors exhibited 1 : 1 activation within their coherent domains, ever comes first. AF termination is followed by attempted non-
and focal sources showed 1 : 1 activation from their origins, with reinducibility of AF. Only if AF is subsequently non-reinducible
distal disorganization and/or collision from isochronal analyses do we consider the end point of AF termination achieved, unlike
(see Figures 43-4 through 43-7) and isopotential movies. Direc- prior reports in which AF reinduction is rarely attempted. If AF
tionality was analyzed to further establish that propagation ema- did not terminate, we assessed for abruptly increased AF cycle
nated from rotors or focal sources during AF and not was directed length >10%. In our validation studies, an AF cycle length
to them. Directionality53 was computed by analyzing isochronal increase >10% indicates elimination of a secondary AF source, as
AF snapshots in 10-ms bins, then calculating vectors between suggested in computer simulations by others (using a 3% to 4%
successive isochronal contours at the spatial boundaries of each cutpoint).54,55 Our acute end point was thus the composite of AF
bin. Figure 43-6 shows (A) a left atrial AF rotor with (B) activa- termination or ≥10% slowing.
tion emanating to remaining atria, and (C) a left atrial focal AF The acute effects of targeted ablation limited to rotor or focal
source with (D) activation emanating to remaining atria. Analysis sources were reported in the CONFIRM trial,16 in patients aged
of consecutive cycles revealed that directionality was conserved 63 ± 9 years, of whom 81% had persistent AF, and in an initial
close to the center of each source but not farther away. These report from first cases at other centers,15 in patients aged 58 ± 12
data support the notion that sustained sources control and drive years, of whom 92% had persistent AF.
fibrillatory activity over multiple cycles. FIRM ablation achieved the acute end point in 86% (31/36)
of patients receiving FIRM-guided ablation in the CONFIRM
Targeted Ablation Limited to Rotors and Focal Sources trial16 and in 100% (12/12) of patients in the first external series.15
Acutely Terminates AF Figure 43-7 shows AF termination by FIRM ablation at (A) a
Recent computational advances enabled us to perform FIRM posterior left atrial rotor, (B) a right atrial rotor, and (C) a left
mapping for human AF in near-real time at electrophysiological atrial focal beat. Preliminary findings from a larger multicenter
study. This provided the opportunity to target these regions for FIRM-guided ablation experience show similar results (manu-
ablation prospectively,16 as illustrated in a recent video case script in preparation).
report.51 The AF termination end point alone was observed in 56%16
Radiofrequency energy was targeted directly at identified and 67%15 of patients before PV isolation after a median of
rotors and focal sources before PV isolation. Radiofrequency <5 min FIRM ablation at the primary source. In the CONFIRM
energy was delivered using a 3.5-mm-tip irrigated catheter trial, this was achieved with a median of 2.5 min ablation at the
(Thermocool, Biosense-Webster, Diamond Bar, California) at 25 primary source (interquartile range [IQR], 1 to 3 min).16 More-
to 35 W, or, in heart failure subjects, an 8-mm-tip nonirrigated over, AF termination by FIRM was predominantly to sinus
catheter (Blazer, Boston Scientific, Natick, Massachusetts) at 40 rhythm (23/28 terminations in pooled studies; 82%15,16), unlike
to 50 W, target 52° C. Other centers have subsequently employed conventional (nonguided) ablation, in which AF terminates
ROTOR IN LEFT ATRIUM IN AF ROTOR DRIVES AF

190
0 ms
A B
FOCAL IMPULSE IN LEFT ATRIUM IN AF FOCAL SOURCE DRIVES AF

100
0 ms
ECG I
CS
C D
Figure 43-6. Propagation Emanates from Localized Sources to Remaining Atria During Human AF A, Left atrial AF rotor. B, Directionality shows propa-
gation emanating from rotor to remaining atria. C, Left atrial AF focal source. D, Directionality shows propagation emanating from focal source to remaining atria. The
arrows indicate the activation direction53 between isochrones (color bar).
typically to an atrial tachycardia (<13% to sinus rhythm55). were causal rather than bystanders from unmapped and unidenti-
Notably, AF sources were stable over time in each patient. AF fied mechanisms.
sources were persistent over the time scale required for mapping Our results extend elegant prior studies suggesting rotors in
and ablation (mean, 115 ± 57 minutes). AF sources were con- human AF. Atienza et al reported regions of high spectral domi-
served for months in a subset of subjects who failed conventional nant frequency near the pulmonary veins in patients with
ablation in the control limb of CONFIRM, then re-presented paroxysmal AF whose frequency increased after adenosine admin-
for FIRM-guided ablation.16 istration,26 consistent with a reentrant mechanism. Using nonin-
vasive body surface mapping (ECG imaging [ECGI]), Cuculich
et al13 detected rotational activity in AF patients, but in a minority
Discussion of patients, typically of only 1 rotation, and with very short cycle
length (<75 ms).13 These differences, which may reflect the
These data demonstrate that human AF is sustained by localized respective mapping approaches, require further definition.
rotors and focal sources in patients with paroxysmal, persistent,
and long-standing persistent AF. Notably, localized sources were
observed in patients undergoing first AF ablation and in those
with recurrent AF after prior ablation. The observation that brief Clinical Implications
targeted ablation (FIRM) limited to rotors or focal sites was able
to terminate AF and render it noninducible, before anatomically The presence of a small number of stable and temporally con-
based ablation of PV triggers, provides strong evidence for the served rotors or focal sources in any given patient with AF is
mechanistic role of patient-specific AF sources in maintaining the consistent with many mapping studies, and suggests a novel
arrhythmia. approach to classifying AF and tracking the natural history of
Our finding that human AF is sustained by a small number (1 AF—by noting an increase or a decrease in the number of sources.
to 3) of rotors and focal sources that are spatially stable over The precise pathophysiology that underlies the formation of AF
time10 was initially surprising and differed from previous studies. sources requires further study, which may lead to future diagnos-
A detailed recent epicardial mapping study was interpreted to tic or therapeutic applications. However, the most immediate
show that AF is maintained by numerous complex nonrepeating therapeutic application is the direct elimination of AF sources.
activation fronts,36 yet that study used limited field-of-view Accordingly, we hypothesized that targeted FIRM ablation at
(<20% of dilated atrial surfaces) that may miss AF sources outside patient-specific AF rotors and focal sources would improve long-
the mapped field, and did not prove that the disordered waves term freedom from AF compared with conventional ablation.
LEFT ATRIAL ROTOR IN AF FIRM: TERMINATES AF TO
43
SINUS RHYTHM (<1 MIN)
Right atrium Left atrium AF SINUS
190
Rotor
(FIRM
site)
0 ms CS
A B 1 sec
RIGHT ATRIAL ROTOR, LA FOCAL BEAT IN AF FIRM: TERMINATES AF TO

SINUS RHYTHM (5.5 MIN)
AF SINUS
160
Rotor
(FIRM
site)
0 ms CS
C D 1 sec
LEFT ATRIAL FOCAL SOURCE IN AF FIRM: TERMINATES AF TO SINUS RHYTHM (6 MIN)

Right atrium Left atrium I
aVF AF SINUS
100
V6
Focal
(FIRM)
CS
0 ms
E F 1 sec
Figure 43-7. Acute Termination of AF to Sinus Rhythm by Focal Impulse and Rotor Modulation (FIRM) Ablation A, Left atrial rotor with counter-
clockwise activation (red to blue) and disorganized right atrium in AF. B, FIRM ablation at left atrium rotor terminated AF to sinus rhythm in <1 min. C, Right atrium rotor
(clockwise) and simultaneous left atrium focal source during persistent AF. D, FIRM ablation at right atrium rotor terminated AF to sinus rhythm in 5.5 min. E, Left atrium
focal source in AF. F, FIRM ablation at left atrium focal source terminated AF to sinus rhythm. All patients are AF-free on implanted cardiac monitors at >1 y. Scalebar = 1 s.
CS, coronary sinus electrogram. Atrial orientations as in Figure 43-3.
(Adapted from Narayan et al,16 with permission.)
generated in the CONFIRM trial. In more recent series, repeat

Ablation of Localized Sources—the CONFIRM Trial mapping is possible.
Conventional ablation was then performed: in FIRM-blinded
The CONFIRM trial (CONventional ablation with or without subjects, immediately after basket data were acquired, and in
Focal Impulse and Rotor Modulation) was a prospective case FIRM-guided subjects, after FIRM-guided ablation was com-
cohort study16 that enrolled pateints at 107 AF ablation procedures pleted. Conventional ablation2 comprised wide area circumferen-
who had failed at least one antiarrhythmic medication with stantial ablation to isolate the left and right pulmonary veins in pairs,
dard indications. Of this population, 36 underwent FIRM-guided with verification of pulmonary vein isolation using a circular
ablation then conventional ablation, and 71 underwent conven- mapping catheter.
tional ablation alone (1 : 2 allocation). The only exclusion criterion Patients were evaluated after the single index procedure in
was an unwillingness to consent. Patients had paroxysmal AF (epi- clinic quarterly for up to 2 years. Repeat ablation was not permit-
sodes that self-terminate in <7 days), persistent AF (episodes that ted at any time, even in the first 3 months (“blanking” period),
last >7 days and require electrical cardioversion), and long- and antiarrhythmic medications were discontinued at 3 months.
standing persistent AF (persistent AF with >1 year of continuous Recurrent arrhythmias were detected using implanted subcutane-
AF).2 Overall, three-quarters of patients had persistent or long- ous ECG monitors whenever possible—Reveal XT™ (Medtronic,
standing persistent AF (conventional, 66%; FIRM-guided, 81%), Minneapolis, Minnesota), or clinically indicated pacemakers/
and subjects had a wide range of ages (20 to 81 years), left ventricu- defibrillators with AF detection algorithms. Remaining subjects
lar ejection fractions (20% to 75%), and comorbidities. received intermittent ambulatory ECG monitoring at each visit.
FIRM was performed in all patients with spontaneous or
induced AF (n = 101, including all FIRM-guided patients). As has
been noted, maps of spontaneous and induced AF converge by 5 CONFIRM Trial Results
to 10 min. FIRM-guided ablation was performed as described for
the acute end point of AF termination or AF cycle length increase The single-procedure freedom from AF was 82.4% in FIRM-
>10% by FIRM ablation of ≤10 min per source. When AF ter- guided patients compared with 44.9% in FIRM-blinded patients
minated, we attempted to reinitiate AF; if successful, FIRM was (P < .001) after 273 days (median IQR, 132 to 681). Notably, no
repeated for ≤3 sources (≤30 min), but only 1 FIRM map was FIRM-guided case recurred after ≈7 months. FIRM-guided
therapy maintained its benefit over FIRM-blinded therapy for FREEDOM FROM ATRIAL FIBRILLATION
first-time ablation cases (P < .001; Figure 43-8 for patients off 1.0
antiarrhythmic medications) and for all prespecified subgroups.
FIRM could not be completed in a small number of patients P = .015 1st Ablation
because of esophageal heating, phrenic nerve capture, or other
0.8
factors; these patients were included as failures in the intention-
to-treat results. Moreover, FIRM-guided patients had a greater P = .003 all cases
Event-free survival
number of comorbidities and more rigorous follow-up than
0.6
FIRM-blinded patients (subcutaneous ECG monitors in 88.2%
vs. 26.1%; P < .001), suggesting that the results may actually
understate the relative benefits of FIRM-guided ablation. Patient
safety was equivalent, with no difference in adverse events 0.4
between groups. Finally, the acute and chronic end point was FIRM-Blind
achieved in patients with all forms of AF (paroxysmal, persistent, FIRM-Guided
and long-standing persistent AF). However, complete FIRM 0.2 FIRM-Blind, 1st Ablation
mapping was not possible in patients with atria larger than the FIRM-Guided, 1st Ablation
largest basket diameter (55 to 60 mm), who were less likely to
achieve the acute end point. 0.0
0 200 400 600
Days
Discussion of CONFIRM Trial Figure 43-8. Cumulative Freedom from Atrial Fibrillation in Patients
Off Antiarrhythmic Medications For all cases (bold lines), and for those at
The CONFIRM trial showed for the first time that brief ablation first ablation (dashed lines). Intention-to-treat analysis and P values reflect the com-
(FIRM) limited to patient-specific AF-maintaining sources was plete follow-up period.
able to acutely terminate persistent or paroxysmal AF before
conventional ablation in a majority of patients, and, when fol- (Adapted from Narayan et al,16 with permission.)
lowed by conventional ablation, it substantially increased long-
term AF elimination after a single procedure compared with
conventional AF ablation alone. FIRM-guided ablation. Second, the maximal current basket size
The CONFIRM trial is notable for many reasons. First, it (55 to 60 mm) places an upper limit on patients in whom com-
reports among the highest single procedure success rate of any plete atrial mapping is achievable. Although the basket has
AF ablation trial to date, verified using implanted ECG monitors suboptimal resolution, the theoretical design considerations
that are far more sensitive than symptoms or the intermittent above and the size of single-ablation lesions (≈5 to 7-mm diam-
ECG monitoring used in most prior ablation trials. Second, the eter) may limit the need for substantially higher resolution.
CONFIRM trial is one of the few AF trials to date9,56,57 to Third, a randomized controlled trial with a greater representa-
compare a novel ablation strategy with a control strategy of con- tion of female subjects is required and is already under way.
ventional ablation58 (not to antiarrhythmic medications). Third, Fourth, future studies are required to define the benefit of mech-
CONFIRM is the first trial to identify and directly target dem- anistically targeted FIRM ablation alone, without ablation at AF
onstrated AF mechanisms, proven by achievement of the acute trigger sites; such studies are also under way.
end point in a patient-tailored fashion. Fourth, FIRM-guided
ablation was more effective than conventional ablation in all
patient subgroups, including those undergoing their first proce-
dure and those with AF despite prior conventional ablation (see Conclusions
Figure 43-8).
Results of the CONFIRM trial may explain many observa- Recent data provide compelling evidence that human AF is main-
tions from conventional AF ablation. First, most sources lay in tained by a small number of stable, patient-specific localized
the left atrium. Nevertheless, the presence of right atrial sources rotors or focal sources. AF sources were detected using FIRM
in one-quarter of patients may explain the 70% to 80% success mapping, which combines wide-field-of-view bi-atrial contact
ceiling2 of conventional (mostly left atrial) ablation for AF, and recordings of AF with computational analyses. Rotors and focal
the need for right atrial ablation in some patients.32 Second, sources were present in diverse bi-atrial locations outside the
diverse source locations are consistent with reports that wide- pulmonary veins, were sustained for thousands of cycles, and were
spread ablation may be required.8 Third, the higher number of spatially stable over prolonged periods. Their mechanistic role
sources in persistent than in paroxysmal AF is consistent with was demonstrated by targeted ablation (FIRM) that rapidly termi-
lower success rates and more difficult procedures in persistent AF nated AF and rendered it noninducible in most patients. FIRM-
patients. Fourth, how widespread conventional ablation can miss guided ablation (FIRM plus conventional ablation) nearly doubled
localized sources may be explained by AF source locations often long-term AF elimination compared with conventional ablation
remote from typical lesion sets, and the fact that conventional alone with rigorous monitoring.
ablation may dwell briefly at each location (≈30-60 s) before the These results explain many observations on clinical AF, includ-
catheter is moved, but FIRM targets sources for >≈2.5 min of ing the 70% to 80% success “ceiling” of multiple conventional
ablation (similar to ablation for AV node reentry, atrial tachycar- ablation procedures (that may miss right atrial sources), the spa-
dias, or other “simple” arrhythmias17-19). tiotemporal stability of AF in individual patients, prior indirect
CONFIRM has several limitations. The first is its nonran- evidence of localized AF sources, and the fact that termination of
domized design, although subjects were enrolled consecutively AF during conventional ablation may occur at any stage while,
and were treated prospectively for prespecified end points. In conversely, extensive ablation often has little acute impact on AF.
fact, FIRM-guided subjects had a higher prevalence of persistent Future studies are required to address many remaining ques-
AF, a greater number of comorbidities, and more intense moni- tions. Mechanistically, the pathophysiology that leads to the forma-
toring than FIRM-blinded subjects, thus actually biasing the tion of AF sources is unclear and may result from structural fibrosis
study against and likely underestimating the benefit of or scar or functional properties such as gradients of repolarization
or conduction restitution. Clinically, future studies should deter- including FIRM ablation alone, in several AF patient populations.
43
mine whether technical improvements can further improve Finally, future studies may clarify whether FIRM mapping opens
mapping accuracy and outcomes of FIRM ablation in patients with the possibility of additional patient-tailored therapies for AF, such
the most dilated atria. Randomized controlled trials are under way as pharmacologic and regenerative therapies for this highly preva-
to define the benefit of different FIRM ablation strategies, lent disease with major public health and societal impact.
14. Narayan SM, Krummen DE, Rappel W-J: Clinical and surface electrocardiographic correlations—a
References mapping approach to diagnose electrical rotors and pilot study. Heart Rhythm 5:686–693, 2008.
focal impulse sources for human atrial fibrillation 30. Xi Q, Sahakian AV, Ng J, et al: Atrial fibrillatory
1. Haïssaguerre M, Jais P, Shah DC, et al: Spontane- (cover article). J Cardiovasc Electrophysiol 23:447– wave characteristics on surface electrogram: ECG
ous initiation of atrial fibrillation by ectopic beats 454, 2012a. to ECG repeatability over twenty-four hours in
originating in the pulmonary veins. N Engl J Med 15. Shivkumar K, Ellenbogen, Kenneth A, et al: Acute clinically stable patients. J Cardiovasc Electro-
339:659–666, 1998. termination of human atrial fibrillation by identi- physiol 15:911–917, 2004a.
2. Calkins H, Kuck KH, Cappato R, et al: 2012 HRS/ fication and catheter ablation of localized rotors 31. Nademanee K, McKenzie J, Kosar E, et al: A new
EHRA/ECAS expert consensus statement on cath- and sources: First multicenter experience of focal approach for catheter ablation of atrial fibrillation:
eter and surgical ablation of atrial fibrillation: Rec- impulse and rotor modulation (FIRM) ablation. J Mapping of the electrophysiologic substrate. J Am
ommendations for patient selection, procedural Cardiovasc Electrophysiol 23(12):1277–1285, Coll Cardiol 43:2044–2053, 2004a.
techniques, patient management and follow-up, 2012. 32. Hocini M, Nault I, Wright M, et al: Disparate
definitions, end points, and research trial design: A 16. Narayan SM, Krummen DE, Shivkumar K, et al: evolution of right and left atrial rate during abla-
report of the Heart Rhythm Society (HRS) task Treatment of atrial fibrillation by the ablation of tion of long-lasting persistent atrial fibrillation. J
force on catheter and surgical ablation of atrial localized sources: The conventional ablation for Am Coll Cardiol 55:1007–1016, 2010.
fibrillation. Developed in partnership with the atrial fibrillation with or without focal impulse and 33. Schuessler RB, Grayson TM, et al: Cholinergically
European Heart Rhythm Association (EHRA), a rotor modulation (CONFIRM) trial. J Am Coll mediated tachyarrhythmias induced by a single
registered branch of the European Society of Car- Cardiol 60:628–636, 2012d. extrastimulus in the isolated canine right atrium.
diology (ESC) and the European Cardiac Arrhyth- 17. Jackman WM, Beckman KJ, McClelland JH, et al: Circ Res 71(5):1254–1267, 1992.
mia Society (ECAS); and in collaboration with the Treatment of supraventricular tachycardia due to 34. Cox JL, Schuessler RB, Boineau JP: The surgical
American College of Cardiology (ACC), the Amer- atrioventricular nodal reentry by radiofrequency treatment of atrial fibrillation. I. Summary of the
ican Heart Association (AHA), the Asia Pacific catheter ablation of slow-pathway conduction. N current concepts of the mechanisms of atrial flutter
Heart Rhythm Society (APHRS), and the Society Engl J Med 327:313–318, 1992. and atrial fibrillation. J Thorac Cardiovasc Surg
of Thoracic Surgeons (STS). Endorsed by the gov- 18. Jackman WM, Wang XZ, Friday KJ, et al: Catheter 101:402–405, 1991.
erning bodies of the American College of Cardiol- ablation of accessory atrioventricular pathways 35. Weimar T, Bailey MS, Watanabe Y, et al: The Cox-
ogy Foundation, the American Heart Association, (Wolff-Parkinson-White syndrome) by radiofre- Maze IV procedure for lone atrial fibrillation: A
the European Cardiac Arrhythmia Society, the quency current. N Engl J Med 324:1605–1611, single center experience in 100 consecutive patients.
European Heart Rhythm Association, the Society 1991. J Interv Card Electrophysiol 31:47–54, 2011.
of Thoracic Surgeons, the Asia Pacific Heart 19. Stevenson WG, Khan H, Sager P, et al: Identifica- 36. Allessie MA, de Groot NM, Houben RP, et al:
Rhythm Society, and the Heart Rhythm Society. tion of reentry circuit sites during catheter mapping Electropathological substrate of long-standing per-
Heart Rhythm 9:632–696.e621, 2012. and radiofrequency ablation of ventricular tachy- sistent atrial fibrillation in patients with structural
3. Nattel S: New ideas about atrial fibrillation 50 cardia late after myocardial infarction. Circulation heart disease: Longitudinal dissociation. Circ
years on. Nature 415:219–226, 2002. 88:1647–1670, 1993. Arrhythm Electrophysiol 3:606–615, 2010.
4. Vaquero M, Calvo D, Jalife J: Cardiac fibrillation: 20. Narayan SM, Kazi D, Krummen DE, et al: Repo- 37. Moe GK, Rheinboldt W, Abildskov J: A computer
From ion channels to rotors in the human heart. larization and activation restitution near human model of atrial fibrillation. Am Heart J 67:200–220,
Heart Rhythm 5:872–879, 2008. pulmonary veins and atrial fibrillation initiation: A 1994.
5. Allessie MA, de Groot NM, Houben RP, et al: The mechanism for the initiation of atrial fibrillation by 38. Wu T-J, Doshi RN, Huang H-LA, et al: Simulta-
electropathological substrate of longstanding per- premature beats. J Am Coll Cardiol 52:1222–1230, neous biatrial computerized mapping during per-
sistent atrial fibrillation in patients with structural 2008c. manent atrial fibrillation in patients with organic
heart disease: Longitudinal dissociation. Circ 21. Gong Y, Xie F, Stein K, et al: Mechanism underly- heart disease. J Cardiovasc Electrophysiol 13:571–
Arrhythm Electrophysiol 3:606–615, 2010. ing initiation of paroxysmal atrial flutter/atrial 577, 2002.
6. Lazar S, Dixit S, Marchlinski FE, et al: Presence fibrillation by ectopic foci: A simulation study. Cir- 39. Schilling RJ, Kadish AH, Peters NS, et al: Endo-
of left-to-right atrial frequency gradient in parox- culation 115:2094–2102, 2007. cardial mapping of atrial fibrillation in the human
ysmal but not persistent atrial fibrillation in 22. Narayan SM, Bode F, Karasik PL, et al: Alternans right atrium using a non-contact catheter. Eur
humans. Circulation 110:3181–3186, 2004. of atrial action potentials as a precursor of atrial Heart J 21:550–564, 2000.
7. Sahadevan J, Ryu K, Peltz L, et al: Epicardial fibrillation. Circulation 106:1968–1973, 2002b. 40. Samie FH, Berenfeld O, Anumonwo J, et al: Rec-
mapping of chronic atrial fibrillation in patients: 23. Narayan SM, Franz MR, Clopton P, et al: Repo- tification of the background potassium current: A
Preliminary observations. Circulation 110:3293– larization alternans reveals vulnerability to human determinant of rotor dynamics in ventricular fibril-
3299, 2004. atrial fibrillation. Circulation 123:2922–2930, lation. Circ Res 89:1216–1223, 2001.
8. Haissaguerre M, Sanders P, Hocini M, et al: 2011b. 41. Baher A, Qu Z, Hayatdavoudi A, et al: Short-term
Catheter ablation of long-lasting persistent atrial 24. Lalani G, Schricker A, Gibson M, et al: Dynamic cardiac memory and mother rotor fibrillation. Am
fibrillation: Critical structures for termination. conduction slowing precedes human atrial fibrilla- J Physiol Heart Circ Physiol 292:H180–H189,
J Cardiovasc Electrophysiol 16:1125–1137, tion initiation: Insights from bi-atrial basket 2007.
2005a. mapping on transitions to atrial fibrillation. J Am 42. Lalani GG, Schricker A, Gibson M, et al: Atrial
9. Oral H, Pappone C, Chugh A, et al: Circumferen- Coll Cardiol 59:595–606, 2012. conduction slows immediately before the onset of
tial pulmonary-vein ablation for chronic atrial 25. Konings K, Kirchhof C, Smeets J, et al: High- human atrial fibrillation: A bi-atrial contact
fibrillation. N Engl J Med 354:934–941, 2006. density mapping of electrically induced atrial fibril- mapping study of transitions to atrial fibrillation. J
10. Skanes AC, Mandapati R, Berenfeld O, et al: Spa- lation in humans. Circulation 89:1665–1680, 1994. Am Coll Cardiol 59:595–606, 2012.
tiotemporal periodicity during atrial fibrillation in 26. Atienza F, Almendral J, Moreno J, et al: Activation 43. Chou CC, Chang PC, Wen MS, et al: Epicardial
the isolated sheep heart. Circulation 98:1236– of inward rectifier potassium channels accelerates ablation of rotors suppresses inducibility of
1248, 1998. atrial fibrillation in humans: Evidence for a reen- acetylcholine-induced atrial fibrillation in left pul-
11. Atienza F, Almendral J, Jalife J, et al: Real-time trant mechanism. Circulation 114:2434–2442, monary vein-left atrium preparations in a beagle
dominant frequency mapping and ablation of dom- 2006. heart failure model. J Am Coll Cardiol 58:158–166,
inant frequency sites in atrial fibrillation with left- 27. Feld GK, Fleck RP, Chen PS, et al: Radiofrequency 2011.
to-right frequency gradients predicts long-term catheter ablation for the treatment of human type 44. Cherry EM, Fenton FH: Visualization of spiral
maintenance of sinus rhythm. Heart Rhythm 6:33– 1 atrial flutter: Identification of a critical zone in and scroll waves in simulated and experimental
40, 2009. the re-entrant circuit by endocardial mapping tech- cardiac tissue. New J Phys 10:125016, 2008.
12. Sanders P, Berenfeld O, Hocini M, et al: Spectral niques. Circulation 86:1233–1240, 1992a. 45. Ryu K, Shroff SC, Sahadevan J, et al: Mapping of
analysis identifies sites of high-frequency activity 28. Gerstenfeld E, Sahakian A, Swiryn S: Evidence for atrial activation during sustained atrial fibrillation
maintaining atrial fibrillation in humans. Circula- transient linking of atrial excitation during atrial in dogs with rapid ventricular pacing induced heart
tion 112:789–797, 2005. fibrillation in humans. Circulation 86:375–382, failure: Evidence for a role of driver regions.
13. Cuculich PS, Wang Y, Lindsay BD, et al: Nonin- 1992. J Cardiovasc Electrophysiol 16:1348–1358, 2005.
vasive characterization of epicardial activation in 29. Dibs SR, Ng J, Arora R, et al: Spatiotemporal char- 46. Hill BC, Courtney KR: Design of a multi-point
humans with diverse atrial fibrillation patterns. acterization of atrial activation in persistent human laser scanned optical monitor of cardiac action
Circulation 122:1364–1372, 2010. atrial fibrillation: Multisite electrogram analysis potential propagation: Application to microreentry
in guinea pig atrium. Ann Biomed Eng 15:567– sustaining rotors abruptly terminates persistent catheter ablation. J Am Coll Cardiol 51:1003–
577, 1987. atrial fibrillation to sinus rhythm with elimination 1010, 2008.
47. Kirchhof C, Chorro F, Scheffer GJ, et al: Regional on followup. Heart Rhythm 9:1436–1439, 56. Wazni OM, Marrouche NF, Martin DO, et al:
entrainment of atrial fibrillation studied by high- 2012b. Radiofrequency ablation vs antiarrhythmic drugs
resolution mapping in open-chest dogs. Circula- 52. Narayan SM, Franz MR: Quantifying fractionation as first-line treatment of symptomatic atrial fibril-
tion 88:736–749, 1993. and rate in human atrial fibrillation using mono- lation: A randomized trial. JAMA 293:2634–2640,
48. Rensma P, Allessie M, Lammers W, et al: Length phasic action potentials: Implications for substrate 2005.
of excitation wave and susceptibility to reentrant mapping. Europace 9:vi89–vi95, 2007e. 57. Wilber DJ, Pappone C, Neuzil P, et al: Compari-
atrial arrhythmias in normal conscious dogs. Circ 53. Kalifa J, Tanaka K, Zaitsev AV, et al: Mechanisms son of antiarrhythmic drug therapy and radiofre-
Res 62:395–410, 1988. of wave fractionation at boundaries of high- quency catheter ablation in patients with
49. Fenton F, Karma A: Vortex dynamics in three- frequency excitation in the posterior left atrium of paroxysmal atrial fibrillation: A randomized con-
dimensional continuous myocardium with fiber the isolated sheep heart during atrial fibrillation. trolled trial. JAMA 303:333–340, 2010.
rotation: Filament instability and fibrillation. Circulation 113:626–633, 2006. 58. Oral H, Chugh A, Yoshida K, et al: A randomized
Chaos 8:20–47, 1998. 54. Haissaguerre M, Lim KT, Jacquemet V, et al: Atrial assessment of the incremental role of ablation of
50. Fenton FH, Cherry EM, Hastings HM, et al: Mul- fibrillatory cycle length: Computer simulation and complex fractionated atrial electrograms after
tiple mechanisms of spiral wave breakup in a model potential clinical importance. Europace 9(Suppl antral pulmonary vein isolation for long-lasting
of cardiac electrical activity. Chaos 12:852–892, 6):vi64–vi70, 2007. persistent atrial fibrillation. J Am Coll Cardiol
2002. 55. Takahashi Y, O’Neill MD, Hocini M, et al: 53:782–789, 2009.
51. Narayan SM, Patel J, Mulpuru S, et al: Focal Characterization of electrograms associated
impulse and rotor modulation (FIRM) ablation of with termination of chronic atrial fibrillation by
Atrial Ischemia and Fibrillation 44
Jérôme Kalifa, Uma Mahesh R. Avula, and Masatoshi Yamazaki
anterior, intermediate/marginal, and posterior atrial branches

CHAPTER OUTLINE
(Spalteholz classification).23 Angiographically, in a right anterior
Atrial Perfusion: Anatomy and Physiology 443 oblique (RAO) projection, the right anterior atrial branch is
slender, arises at a variable distance from the conus artery, and
Atrial Fibrillation and Myocardial Ischemia-Infarction 445
courses to the left and superior aspect of the atrium.12 It gives off
branches to both atria and the interatrial septum and terminates
by encircling the lower portion of the superior vena cava (SVC).
Atrial fibrillation (AF) is the most common arrhythmia in adults When it is the main atrial branch, it supplies the sinoatrial node
and affects more than 4 million Americans.1,2 This arrhythmia is (SAN). The right intermediate atrial branch, also called the mar-
associated with such adverse consequences as decreased quality ginal branch, arises from the acute right margin of the heart,
of life, decompensated cardiovascular status, and stroke.3,4 Despite ascends over the anterolateral surface of the RA, and supplies
the availability of antiarrhythmic drugs and interventional thera- surrounding atrial tissues (see Figure 44-1). In 13.3% of cases, it
pies, AF still represents a therapeutic challenge, and costs related is the main atrial branch that supplies the sinus node.13,14 The
to AF management steadily increase.5-9 AF may be caused by any origin of this branch is marked by the presence of a small fatty
cardiac condition; however, a predominance of heart failure and excrescence or a cardiac vein adjacent to the artery that drains
coronary artery disease (CAD) has been noted.1,10-14 Regarding into the RA wall.15 Angiographically, the right intermediate atrial
CAD-associated AF, clinical and experimental studies show that branch is not easy to see because of its small size and variability.
both acute and chronic CAD lead to onset and perpetuation of For the same reason, the right posterior atrial branch is difficult
AF.11,15-18 For instance, AF is a well-known complication of acute to locate. When present, the right posterior branch arises on the
myocardial infarction,18 and CAD is a significant risk factor for posterior aspect of the RA and supplies the right posterior atrial
AF.8 Although AF after ventricular myocardial infarction might wall and the left atrial (LA) posterior surface. Notably Busquet
also be triggered by an increase in intra-atrial pressure in the et al proposed another classification scheme in which right atrial
context of acute ventricular dysfunction,19,20 various works have branches are classified on the basis of cross-sectional diameter
shown that isolated atrial infarction is common, representing into two rather than three groups—the major and accessory
about 20% of autopsy-proven atrial infarctions.21,22 Recent years branches.15 Major branches are usually found in groups occupy-
have seen a growth of interest in further understanding the mech- ing anterior, lateral, and posterior positions relative to the tricus-
anisms by which atrial ischemia, acute or prolonged, may favor pid orifice, consistent with the Spalteholz classification. Accessory
onset and perpetuation of AF. In this chapter, we review the branches arise from the RCA in the AV groove and have an
anatomical and physiological background of atrial coronary per- average diameter of less than 0.5 mm and a length not exceeding
fusion and present the most recent insights demonstrating a caus- 10 mm. They are covered by fatty tissue in the AV groove and
ative link between atrial ischemia and AF. supply solely the lower atrial wall and the adjacent right
ventricle.
Left Atrial Branches

Atrial Perfusion: Anatomy and Physiology Arteries supplying the LA are among the earliest branches of the
left coronary artery (LCA), usually from the left circumflex
Available literature on the detailed anatomy of atrial coronaries (LCX), and originate along the AV groove. Similar to right atrial
is scarce, and unstandardized nomenclature makes any attempt circulation, left atrial branches are usually classified into the left
at gaining understanding challenging. In general the sinus nodal anterior, intermediate/marginal, and posterior atrial branches.
artery (SNA), the atrioventricular nodal artery (AVNA), and The left anterior atrial branch arises on the anterior aspect of the
several other arterial branches in the atria have been described. LA (see Figure 44-1). When it is the main atrial branch, it courses
Although these play a significant role in the pathophysiology of upward along the LA via the anterior interatrial groove to reach
atrial fibrillation and other tachyarrhythmias, they have no unani- the SAN. In RAO projection, it is an early branch that ascends
mous characterization in the cardiology literature, most likely leftward and upward.12 The remaining left atrial branches are
owing to the variability of their origin, course, and termination. variable and supply adjacent atrial tissues along their course over
Figure 44-1 shows schematically the atrial coronary branches and the LA. Another variant is the left atrial circumflex artery (LACX),
anastomoses that have been reported in the literature. which may branch early from either the LCX or its main atrial
branch (Figure 44-1).12,13 Early in its course, the LACX ascends
slightly through the LA and travels along its lower margin paral-
Atrial Coronary Arteries and Branches lel to the left AV groove. It extends around the left heart margin
and terminates in the posterior wall of the LA. Sometimes, the
Right Atrial Branches LACX extends farther, crossing the crux of the heart along
Arteries supplying the right atrium (RA) are among the earliest the right AV groove to give off right posterior atrial branches.
branches of the right coronary artery (RCA) after the conus Rarely, the LACX may even supply the SAN as the main atrial
artery, and they originate along the right atrioventricular (AV) branch itself, which, in a study of 118 patients, occurred 11% of
groove (see Figure 44-1). They are usually classified into the right the time.13
443
Right and Left Atrial Anastomoses and the Controversial rejected the notion that the anastomosis was formed by a single
Kugel’s Artery major vessel.21,22 The debate was further confounded by their
Several anastomoses between atrial coronary arterial systems usage of the name “Kugel’s artery” for their descriptions of dif-
have been described. Generally, these anastomoses exist as small ferent anastomotic networks. Reconciling these diverging view-
intra-atrial or atrioventricular branches, or as a single vessel points, T.N. James described only two variations of Kugel’s
known as Kugel’s artery—a major transatrial pathway that bridges vessel.20 He reported an arterial connection between either the
the right and left coronary systems (see Figure 44-1). As the proximal LCX or RCA with vessels in the crux. Later, these were
largest atrial artery, the SNA also serves as a major anastomotic referred to as the left and right Kugel’s artery, respectively, and
channel, connecting to right and left coronary systems through- are commonly accepted today as such.23 Angiographically, the left
out its course. It forms a vascular loop around the base of the anterior oblique projection best identifies this anastomotic
SVC, where it sends tributaries to small branches of the left and network, which appears as a flag-like rectangular termination.24
right intermediate arteries.15,20 When present, the posterior ter-
mination of the LACX can also anastomose with the distal RCA
or AVNA.16 Many authors who accepted the importance of an Atrial Perfusion Territories
anastomotic network between the left and right coronary systems
Although the reports described here have focused on a descrip-
tion of the origin, course, and termination of atrial coronary
arteries, the precise extent of atrial perfusion territories has
11
received little attention. Specifically, a delineation of atrial perfu-
sion territories at the posterior wall of the LA (PLA) and the
4
5
pulmonary veins (PVs), which are crucial regions for AF mainte-
9 nance, has received little attention. Our group conducted an
SAN 8
investigation in isolated ovine hearts to address this question.24
3 In good accordance with human descriptions,33 we and others
RA 2 LA first established that in sheep hearts, three atrial coronary
7
branches—left SNA (LSNA), right SNA (RSNA), and atrial
10 branches of the LCX—contribute to the perfusion of the PLA
AORTIC including PV regions.24,25 We then delineated atrial coronary
SINUS perfusion territories by selective perfusion of Congo red into the
RSNA, and Evan’s blue into the LCX (Figure 44-2, left and middle
1 6
panels) or the LSNA (right panel). Both atria including the PLA
Figure 44-1. Atrial arterial branches. 1, Right coronary artery; 2, right anterior atrial and the PVs were then dissected for acquisition of photographic
branch (main atrial branch in this figure); 3, right intermediate branch; 4, right snapshots. As shown in Figure 44-2, each atrial branch
posterior branch; 5, Kugel arterial anastomosis; 6, left circumflex artery; 7, left ante- (RSNA, LSNA, and LCX branches) irrigated a well-defined per-
rior atrial branch; 8, left intermediate atrial branch; 9, left posterior branch; 10, left fusion territory at the PLA and within the pulmonary
atrial circumflex; 11, atrioventricular nodal artery. LA, Left atrium; RA, right atrium;
SAN, sinoatrial node.
veins. However, the contribution of individual arteries to the
perfusion of the PLA-PV myocardium varied between speci-
(Adapted from Boppana S et al: Atrial coronary arteries: Anatomy and atrial perfusion mens, and three equally important anatomical variants were most
territories J Atrial Fibrillation September 2011.) often observed: triple-vessel PLA perfusion (Figure 44-2 left;
Triple vessel PLA perfusion LSNA dominant PLA perfusion Balanced double vessel PLA perfusion
(RSNA+LSNA+LCX) (LSNA+LCX) (RSNA+LSNA)
LSNA LSNA
region region LSNA
RSPV Roof region
RIPV LSPV
Septum Lateral
RSNA
RSNA region
LIPV
region
LCX
S region S S
R L R L LCX LCX
R L
I Base Coronary sinus region region
I I
Red: RSNA region, White: LSNA region, Blue: LCX region Red: RSNA region, Blue: LSNA region,
White: LCX region
Figure 44-2. Atrial coronary perfusion territories in sheep: Main anatomical variants. Left, Triple-vessel PLA perfusion; middle, LSNA-dominant PLA perfusion; right, balanced
double-vessel PLA perfusion. Black solid line delineates regional perfusion territories. Left panel, Selective perfusion of Congo red into the RSNA, and Evans blue into the
LCX; middle panel, selective perfusion of Congo red into the RSNA (no left atrial staining) or the LSNA (right panel). Right panel, Selective perfusion of Congo red into the
RSNA and Evans blue into the LSNA. LIPV, left inferior pulmonary vein; LSNA, left sinus node artery; LSPV, left superior pulmonary vein; PLA, posterior wall of the left atrium;
RIPV, right inferior pulmonary vein; RSNA, right sinus node artery; RSPV, right superior pulmonary vein.
(Adapted from Yamazaki M et al: Left atrial coronary perfusion territories in isolated sheep hearts: implications for atrial fibrillation maintenance. Heart Rhythm 7:1501–1508,
2010.)
29.5%), double-vessel PLA perfusion (right; 23.5%), and one- directed toward the area of atrial infarction.32 In addition, various
44
vessel PLA perfusion (middle; 29.5%). These data suggest that clinical reports have demonstrated a relationship between atrial
the complexity of atrial vessel origin and course may translate infarction, atrial arrhythmias, and AMI.31,41 Supraventricular
into intricate perfusion areas in regions well known to harbor tachyarrhythmias including AF have been reported in up to 74%
electrical sources of AF initiation and maintenance. This also of AMI cases,21,42 while AF per se occurred in 33% to 55% of
indicates that advanced coronary artery disease may associate these patients.21,42 In fact, arrhythmias are considered a major
with atrial infarctions involving variable regions of the atria. diagnostic criterion when a P-Ta-segment depression is present
in leads I, II, and III or in precordial leads (see Table 44-1).30
Besides, it was recently shown that in patients with both par-
Role of Atrial Myocardium Superfusion oxysmal and persistent AF but no proven atrial infarction, iso-
lated atrial myocardial perfusion abnormalities, coronary flow
The thickness of the atrial wall has been reported to range from reserve impairment, and microvascular dysfunction occur.43-45
0.5 to 4 mm26,27; thus it is conceivable that the blood present Skalidis et al assessed atrial coronary flow velocities using an
within the atrial cavities succeeds in superfusing the thinnest intracoronary Doppler technique in patients with lone recurrent
regions. However, available literature indicates that oxygen may AF (LRAF).44,45 Coronary flow reserve was then calculated by
diffuse to only a few layers of myocytes. A histologic study of comparing baseline and maximal coronary flow (hyperemia) with
capillary density in the left atrial free wall and the left atrial respect to flow in LCX and LACX.45 Investigators found that the
appendage of patients in chronic AF showed that the maximal coronary flow reserve of the left atrial coronary branch (LACB)
oxygen diffusion distance was estimated at 15 to 20 µM.28 Others in patients with LRAF was significantly reduced compared with
have suggested that up to 500 µM of myocardial wall may be LCX flow in the same patients. In addition, LACB coronary flow
adequately superfused.29 Nonetheless, a large proportion of the reserve was significantly reduced compared with coronary flow
atrial wall remains dependent on an efficient coronary perfusion, reserve in control subjects.45
and a perfusion impediment will translate into an infarcted region
spanning most of the atrial wall.
Ventricular Ischemia, Atrial Ischemia
The question is whether ventricular ischemia with a subsequent
increase in intra-atrial pressure provides an altered electrophysi-
Atrial Fibrillation and Myocardial ological milieu that is conducive to AF, or whether this is a direct
Ischemia-Infarction ischemic insult to atrial myocytes that is arrhythmogenic. In the
latter case, the onset of an atrial ischemia-infarction would
Atrial Infarction, Decreased Atrial Coronary Flow explain why AF is seen in only about 20% of acute coronary
Atrial infarction has been extensively described in the literature,

although it remains an elusive and difficult diagnosis to make
clinically.30,31 Although atrial infarction is fairly common in the Table 44-1. ECG Diagnostic Criteria for Atrial Infarction
setting of acute myocardial infarction (AMI),32-34 the rate of 17%
reported by Cushing et al in 1942 in the largest study to date may Major Criterion:
be the best estimate of its occurrence.21 However, these numbers, • PTa-segment elevation >0.5 mm in leads V5 and V6 with reciprocal
derived from postmortem studies, suggest that the actual clinical depression of Pta segment in V1 and V2 leads
incidence rate might be higher.35 In fact, it was suggested that the • Pta-segment elevation>0.5 mm in lead I with reciprocal depressions
pathophysiological role of atrial ischemia-infarction in AF onset in leads II and III
is greatly underestimated.36,37 Electrocardiographic recognition • Pta-segment depression >1.5 mm in precordial leads and 1.2 mm in
of atrial infarction was first established by Liu et al in 1961.32 leads I, II, III, associated with any atrial arrhythmia
Overall, isolated atrial infarction may be diagnosed on the elec-
trocardiogram (ECG) as changes in P wave shape and dura- Minor Criterion:
tion.21,38-40 Over the years, the criteria for clinical diagnosis of
atrial infarction have been refined.35 Figure 44-3 depicts that PR • Abnormal P waves (W-shaped, M-shaped, irregular, notched)
segment depression or elevation may be expected in the presence Adapted from Liu CK, Greenspan G, Piccirillo RT: Atrial infarction of the heart.
of an atrial infarction (Table 44-1).32 Analysis of the ECGs shows Circulation 23:331–338, 1961.
that the PR- or P-Ta-segment, when plotted vectorially, is
R R R
T
P T T P
P
Q Q
S Q S
S
P wave abnormality P wave abnormality

Normal PR-segment elevation PR-segment depression
Figure 44-3. Normal aspect and P wave morphology and PR-segment abnormalities related to an isolated atrial infarction.
(Adapted from Mendes RG, Evora PR: Arq Bras Cardiol 72:333–342, 1999.)
syndromes.21 A study conducted by Alasady et al accordingly (Figure 44-4, A). An increased Na+/Ca2+ exchanger current was
indicates that coronary artery disease of atrial branches is an found in cells dissociated from the right atrial infarction border
independent predictor of new-onset AF in AMI patients.46 The zone (BZ; Figure 44-4, B) in which APD was unchanged. Accord-
study enrolled patients who developed AF within 7 days of acute ingly, spontaneous discharges were significantly more numerous
myocardial infarction and matched patients of similar age, sex, in BZ cells (Figure 44-4, C). Finally, reentrant circuits anchored
and left ventricular ejection fractions from the same cohort. at the infarction border zone were visualized with optical mapping
Follow-up angiograms and echocardiograms were obtained in techniques.51 Altogether, substantial experimental evidence shows
both groups. Univariate analyses showed highly statistically sig- that atrial ischemia-infarction predisposes to onset and perpetua-
nificant association of right atrial branch, left atrial branch, or tion of AF. Mechanistically, both spontaneous focal discharges and
sinoatrial branch disease in patients with AF as compared with reentrant circuits are likely to initiate and maintain AF. In addi-
patients without AF. After multivariate analyses, right atrial tion, one may speculate that atrial ischemia-infarction, which may
branch disease and left atrial branch disease remained powerful remain silent for extended periods, contributes to preclinical
predictors of new onset of AF in AMI patients, even after adjust- remodeling conducive to AF.
ment for indices of left atrial stretch and pressure rise.46
Atrial Ischemia–Maintained Fibrillation:

Atrial Ischemia–Maintained Fibrillation: Proposed Mechanistic Perspectives
Electrophysiological Mechanisms
Despite advances in our understanding of atrial ischemia–related
Experimentally, Lammers et al showed that hypoxia of isolated AF, several mechanistic aspects remain to be investigated. First,
superfused rabbit atrial preparations results in a transient increase it is still unclear whether right and left atrial ischemia are equally
in atrial effective refractory periods (ERPs), and a decrease in conducive to AF and respond to similar underlying mechanisms.
conduction velocity.47 Jayachandran et al in opposite findings Also, the ionic mechanisms and intracellular pathways involved
indicated that proximal right coronary artery occlusion, which in acute versus chronic atrial ischemia remain to be clarified.
causes posterior left ventricular and left atrial ischemia, reduces Besides, currents that are known to be critical for ventricular
atrial ERPs—an effect not altered by the adenosine triphosphate infarction–related arrhythmias may play a distinct role in
(ATP)-dependent K+-channel blocker glibenclamide.48 More the atrial response to ischemia. For instance, it was demonstrated
recently, Sinno et al demonstrated that regional acute right atrial in mice that the molecular structure, pharmacology, and
ischemia causes myocardial necrosis, increases AF inducibility, sensitivity of atrial ATP-sensitive potassium (KATP) channels are
and adversely modulates regional impulse propagation.49 The largely different from those of ventricular myocytes.52-54 Thus it
same group suggested that geranylgeranylacetone, an orally active could be expected that ischemia and ATP depletion operate dif-
inducer of the cardioprotective heat shock protein (HSP), pre- ferently within the atria.52,53 Finally, it was suggested that in the
vents ischemia-induced atrial conduction abnormalities and sup- ventricle the late sodium current is increased under conditions of
presses ischemia-related AF.50 Nishida et al further indicated that ischemia-infarction, after which tumor necrosis factor (TNF)-α
an 8-day chronic right atrial infarction in a canine model is also a triggers an increase in late sodium current.55,56 It remains to be
substrate for AF initiation and maintenance. In this study, the elucidated whether a similar phenomenon occurs after atrial
authors observed spontaneous episodes of atrial tachyarrhythmia ischemia.
Lead I 5s
(2 Hz, 1min)
Pacing Post-pacing follow-up (1 min)
+80
Li+ (mV) 10 sec CTL
mV
0.5
Spontaneous activity (events/min)
+40
–150 –100 –50 0 50 15
CTL 0 **
10
–40
Li+ ** –1.0
***
pA/pF
*** –80
100 pA
*** –2.0
*** +80
10 sec
5
(mV) BZ
BZ 500 ms *** +40 0
CTL BZ
+40 mV *** –3.0 0
–80 mV CTL: n=26 cells/5 dogs
CTL (16 cells/5 dogs) –40 BZ: n=39 cells/8 dogs
–80 mV BZ (23 cells/5 dogs)
B –120 mV C –80
Figure 44-4. A, Recurrent atrial tachycardia recorded by 24-hr Holter electrocardiography (ECG) in a dog with myocardial infarction (MI). Dashed lines indicate atrial tachy-
cardias and atrial ectopic complexes. B, Upper and lower left panels, Recordings of sodium calcium exchange (NCX) currents in control (CTL; upper) and border-zone (lower)
cells. Right panel, Mean ± SEM (standard error of the mean) NCX current density in CTL and border-zone (BZ) cells. C, Upper and lower left panels, action potential (AP)
recordings in CTL and MI BZ cells isolated from sham and right atrial MI dogs, respectively. Right panel, Frequency of spontaneous activity. **P < .01.
(Adapted from Nishida, Kunihiro, et al: Mechanisms of atrial tachyarrhythmias associated with coronary artery occlusion in a chronic canine model. Circulation 123(2):137–146,
2011.)
ACUTE CHRONIC Figure 44-5. A, Mechanisms of coronary artery injury

after radiofrequency ablation (RFA). B, Convective
RFA
Progressive
scarring
Nidus
formation
cooling and the “shadow effect” during RFA cause
electrode-tissue temperature discrepancy. Convective
cooling confers a protective effect for coronary arteries
44
(CAs) against thermal energy from RFA, and the “shadow
effect” can prevent transmural lesion formation.
Late (Adapted from Castaño, Adam, et al: Coronary artery

Coronary Coagulation Collagen Direct coronary pathophysiology after radiofrequency catheter ablation:
spasm dysregulation shrinkage trauma thrombus Review and perspectives. Heart Rhythm 8(12):1975–
1980, 2011.)
Electrode
Blood flow
Convective cooling
Heart chamber
Myocardium
Lesion
“Shadow effect”
“Heat sink”
B Coronary perfusion convective cooling
is positioned close to a vessel, coronary blood flow within and

Atrial Perfusion During Atrial Fibrillation surrounding the vessel provides a protective feature by prevent-
Radiofrequency Ablation (RFA) ing substantive heating of the vascular endothelium,72 and the
susceptibility of CAs to thermal damage is inversely proportional
Because of the proximity of coronary arteries (CAs) to commonly to the electrode-to-artery distance.73 Convective cooling may
ablated sites, RFA may alter vascular integrity and function.57 account for the paucity of reported CA complications or measur-
Figure 44-5, A describes how RFA may cause CA injury acutely able changes in coronary arteriograms performed before and
and subacutely. In the acute setting, RF energy can result in after RFA.74 Also this effect may explain why forming transmural
coronary spasm, direct vessel trauma, and thromboembolism. RFA lesions in regions neighboring a coronary vessel can prove
Spasm is the most common mechanism of coronary injury due difficult, if not impossible. As an example, it was recently shown
to RFA,58 which is supported by reports of RFA-triggered acute that interposition of the circumflex artery between the mitral
arterial occlusions responsive to nitrates.59-62 Spasm may be due isthmus and the coronary sinus is associated with a lower prob-
to an RF-induced increase in autonomic activity at nerve termi- ability of achieving complete mitral isthmus block.75 Further,
nals in the left atrium,63-66 or to disruption of the vascular wall Fuller et al demonstrated in a rabbit model that flow through
followed by a severe modulation of vascular tone.67 Another even small intramyocardial vessels can prevent transmural lesion
mechanism of RFA-induced CA damage proposes heat-induced formation and preserve conduction through an RF lesion, thereby
collagen shrinkage and subsequent vessel narrowing.68-70 Convec- preventing complete conduction block.76 However, preexisting
tive cooling (Figure 44-5, B) is known to influence lesion forma- narrowed atherosclerotic CAs in patients with CAD are at
tion. Initially described in the setting of anti-malignancy increased risk for thermal injury. In a patient who underwent RFA
hyperthermia treatments,71 convective cooling results from the for atrial flutter, undocumented upstream right CA stenosis,
flow of intracardiac and microvascular blood, which creates a which limited flow and decreased convective cooling, was believed
“heat sink” (see Figure 44-5, B). When an RF electrode to have been the mechanism of CA injury.77
electrophysiology, American Heart Association. 9. Stewart S, Murphy N, Walker A, et al: Cost of an

References Circulation 93:1262, 1996. emerging epidemic: An economic analysis of atrial
5. Oral H, Knight BP, Tada H, et al: Pulmonary vein fibrillation in the UK. Heart 90:286, 2004.
1. Kannel WB, Benjamin EJ: Status of the epidemiol- isolation for paroxysmal and persistent atrial fibril- 10. Roy D, Talajic M, Dubuc M, et al: Atrial fibrillation
ogy of atrial fibrillation. Med Clin North Am lation. Circulation 105:1077–1081, 2002. and congestive heart failure. Curr Opin Cardiol
92:17–40, 2008. 6. Sanders P, Berenfeld O, Hocini M, et al: Spectral 24:29, 2009.
2. Genovesi S, Pogliani D: Prevalence of atrial fibril- analysis identifies sites of high-frequency activity 11. Kannel WB, Abbott RD, Savage DD, et al:
lation and associated factors in a population of maintaining atrial fibrillation in humans. Circula- Coronary heart disease and atrial fibrillation: The
long-term hemodialysis patients. Am J Kidney Dis tion 112:789, 2005. Framingham study. Am Heart J 106:389–396,
46:897–902, 2005. 7. Wolf PA, Mitchell JB, Baker CS, et al: Impact of 1983.
3. Sanoski CA: Clinical, economic, and quality of life atrial fibrillation on mortality, stroke, and medical 12. Benjamin EJ, Levy D, Vaziri SM, et al: Indepen-
impact of atrial fibrillation. J Manag Care Pharm costs. Arch Intern Med 158:229, 1998. dent risk factors for atrial fibrillation in a
15:S4, 2009. 8. Miyasaka Y, Barnes ME, Gersh BJ, et al: Secular population-based cohort. JAMA 271:840, 1994.
4. Prystowsky EN, Benson DW, Jr, Fuster V, et al: trends in incidence of atrial fibrillation in Olmsted 13. Benjamin EJ, Wolf PA, D’Agostino RB, et al:
Management of patients with atrial fibrillation: County, Minnesota, 1980 to 2000, and implications Impact of atrial fibrillation on the risk of death:
A statement for healthcare professionals from on the projections for future prevalence. Circula- The Framingham heart study. Circulation 98:946–
the subcommittee on electrocardiography and tion 114:119, 2006. 952, 1998.
14. Krahn AD, Manfreda J, Tate RB, et al: The natural fibrillation after coronary artery bypass grafting. 59. Solomon AJ, Tracy CM, Swartz JF, et al: Effect on
history of atrial fibrillation: Incidence, risk factors, Thorac Cardiovasc Surgeon 46:115–120, 1998. coronary artery anatomy of radiofrequency cathe-
and prognosis in the manitoba follow-up study. Am 38. Hellerstein HK: Atrial infarction with diagnostic ter ablation of atrial insertion sites of accessory
J Med 98:476–484, 1995. electrocardiographic findings. Am Heart J 36:422– pathways. J Am Coll Cardiol 21:1440–1444, 1993.
15. James TN: Myocardial infarction and atrial 430, 1948. 60. Simon RD, Gill JS: Coronary ischemia induced by
arrhythmias. Circulation 24:761–776, 1961. 39. Lazar EJ, Goldberger J, Peled H, et al: Atrial radiofrequency ablation in the left atrium. J Car-
16. Crenshaw M, Brian S, Ward M, et al: Atrial fibril- infarction: Diagnosis and management. Am Heart diovasc Electrophysiol 14:186–190, 2003.
lation in the setting of acute myocardial infarction: J 116:1058–1063, 1988. 61. Bardy GH, Ivey TD, Coltorti F, et al: Develop-
The GUSTO-I experience. J Am Coll Cardiol 40. Mendes RGG, Evora PRB: Atrial infarction is a ments, complications and limitations of catheter-
30:406–413, 1997. unique and often unrecognized clinical entity. Arq mediated electrical ablation of posterior accessory
17. Wong CK, White HD, Wilcox RG, et al: New Bras Cardiol 72:333–342, 1999. atrioventricular pathways. Am J Cardiol 61:309–
atrial fibrillation after acute myocardial infarction 41. Shakir DK, Arafa SOE: Right atrial infarction, 316, 1988.
independently predicts death: The GUSTO-III atrial arrhythmia and inferior myocardial infarc- 62. Lesh MD, Van Hare GF, Schamp DJ, et al: Cura-
experience. Am Heart J 140:878–885, 2000. tion form a missed triad: A case report and review tive percutaneous catheter ablation using radiofre-
18. Goldberg RJ, Yarzebski J, Lessard D, et al: Recent of the literature. Can J Cardiol 23:995, 2007. quency energy for accessory pathways in all
trends in the incidence rates of and death rates 42. Cristal N, Peterburg I, Inbar-Yanai I: Atrial infarc- locations: Results in 100 consecutive patients. J Am
from atrial fibrillation complicating initial acute tion leading to rupture. Br Heart J 41:350–353, Coll Cardiol 19:1303–1309, 1992.
myocardial infarction: A community-wide perspec- 1979. 63. Yamashita E, Tada H, Tadokoro K, et al: Left atrial
tive. Am Heart J 143:519–527, 2002. 43. Kochiadakis G, Skalidis E, Kalebubas M, et al: catheter ablation promotes vasoconstriction of the
19. Tsang T, Barnes ME, Bailey KR, et al: Left atrial Effect of acute atrial fibrillation on phasic coronary right coronary artery. Pacing Clin Electrophysiol
volume: Important risk marker of incident atrial blood flow pattern and flow reserve in humans. Eur 30(Suppl 1):S98–S102, 2007.
fibrillation in 1655 older men and women. Mayo Heart J 23:734, 2002. 64. Pauza DH, Pauziene N, Pakeltyte G, et al: Com-
Clin Proc 76:467, 2001. 44. Skalidis EI, Kochiadakis GE, Igoumenidis NE, et al: parative quantitative study of the intrinsic cardiac
20. Moller JE, Hillis GS, Oh JK, et al: Left atrial Phasic coronary blood flow velocity pattern and flow ganglia and neurons in the rat, guinea pig, dog and
volume: A powerful predictor of survival after acute reserve in the atrium: Regulation of left atrial myo- human as revealed by histochemical staining for
myocardial infarction. Circulation 107:2207, 2003. cardial perfusion. J Am Coll Cardiol 41:674, 2003. acetylcholinesterase. Ann Anat 184:125–136,
21. Cushing E, Feil H, Stanton E, et al: Infarction of 45. Skalidis EI, Hamilos MI, Karalis IK, et al: Isolated 2002.
the cardiac auricles (atria): Clinical, pathological, atrial microvascular dysfunction in patients with 65. Pauziene N, Pauza DH: Electron microscopic
and experimental studies. Br Heart J 4:17, 1942. lone recurrent atrial fibrillation. J Am Coll Cardiol study of intrinsic cardiac ganglia in the adult
22. Wartman W, Souders J: Localization of myocardial 51:2053, 2008. human. Ann Anat 185:135–148, 2003.
infarcts with respect to the muscle bundles of the 46. Alasady M, Abhayaratna WP, Leong DP, et al: 66. Saburkina I, Rysevaite K, Pauziene N, et al: Epi-
heart. Arch Pathol 50:329, 1950. Coronary artery disease affecting the atrial cardial neural ganglionated plexus of ovine heart:
23. Spalteholz W: Die Arterien der Herzwand: anato- branches is an independent determinant of atrial Anatomic basis for experimental cardiac electro-
mische Untersuchungen an Menschen-und Tier- fibrillation after myocardial infarction. Heart physiology and nerve protective cardiac surgery.
herzen. S. Hirzel, 1924. Rhythm 8:955–960, 2011. Heart Rhythm 7:942–950, 2010.
24. Yamazaki M, Morgenstern S, Klos M, et al: Left 47. Lammers W, Kirchhof C, Bonke F, et al: Vulner- 67. Demaria RG, Page P, Leung TK, et al: Surgical
atrial coronary perfusion territories in isolated ability of rabbit atrium to reentry by hypoxia: Role radiofrequency ablation induces coronary endothe-
sheep hearts: Implications for atrial fibrillation of inhomogeneity in conduction and wavelength. lial dysfunction in porcine coronary arteries. Eur J
maintenance. Heart Rhythm 7:1501–1508, 2010. Am J Physiol Heart Circul Physiol 262:H47, 1992. Cardiothorac Surg 23:277–282, 2003.
25. Yalcin B, Kirici Y, Ozan H: The sinus node artery: 48. Jayachandran JV, Zipes DP, Weksler J, et al: Role 68. Aoyama H, Nakagawa H, Pitha JV, et al: Compari-
Anatomic investigations based on injection- of the Na+/H+ exchanger in short-term atrial elec- son of cryothermia and radiofrequency current in
corrosion of 60 sheep hearts. Interact Cardiovasc trophysiological remodeling. Circulation 101: safety and efficacy of catheter ablation within the
Thorac Surg 3:249, 2004. 1861–1866, 2000. canine coronary sinus close to the left circumflex
26. Beinart R, Abbara S, Blum A, et al: Left atrial wall 49. Sinno H, Derakhchan K, Libersan D, et al: Atrial coronary artery. J Cardiovasc Electrophysiol
thickness variability measured by CT scans in ischemia promotes atrial fibrillation in dogs. Cir- 16:1218–1226, 2005.
patients undergoing pulmonary vein isolation. culation 107:1930–1936, 2003. 69. Gorisch W, Boergen KP: Heat-induced contraction
J Cardiovasc Electrophysiol 22:1232–1236, 2011. 50. Sakabe M, Shiroshita-Takeshita A, Maguy A, et al: of blood vessels. Lasers Surg Med 2:1–13, 1982.
27. Yamazaki M, Mironov S, Taravant C, et al: Hetero- Effects of a heat shock protein inducer on the atrial 70. Kang JX, Xiao YF, Leaf A: Free, long-chain, poly-
geneous atrial wall thickness and stretch promote fibrillation substrate caused by acute atrial ischae unsaturated fatty acids reduce membrane electrical
scroll waves anchoring during atrial fibrillation. mia. Cardiovasc Res 78:63–70, 2008. excitability in neonatal rat cardiac myocytes. Proc
Cardiovasc Res 94:48–57, 2012. 51. Nishida K, Qi XY, Wakili R, et al: Mechanisms of Natl Acad Sci U S A 92:3997, 1995.
28. Corradi D, Callegari S, Benussi S, et al: Regional atrial tachyarrhythmias associated with coronary 71. Haines D: The biophysics and pathophysiology of
left atrial interstitial remodeling in patients with artery occlusion in a chronic canine model. Circu- lesion formation during radiofrequency catheter
chronic atrial fibrillation undergoing mitral-valve lation 110:972778–972771, 2011. ablation. In Zipes DP, editor: Cardiac Electro-
surgery. Virchows Arch 445:498–505, 2004. 52. Glukhov AV, Flagg TP, Fedorov VV, et al: Differ- physiology: From Cell to Bedside, Philadelphia,
29. Ripplinger MC: Thesis dissertation: The Role of ential katp channel pharmacology in intact mouse 2004, WB Saunders, pp 1021–1023.
Myocardial Heterogeneity in Maintenance and heart. J Mol Cell Cardiol 48:152–160, 2010. 72. Chatelain P, Zimmermann M, Weber R, et al:
Termination of Cardiac Arrhythmias. Washington 53. Fedorov VV, Glukhov AV, Ambrosi CM, et al: Acute coronary occlusion secondary to radiofre-
University, Department of Biomedical Engineer- Effects of KATP channel openers diazoxide and quency catheter ablation of a left lateral accessory
ing, May 2008. ProQuest, 2008. pinacidil in coronary-perfused atria and ventricles pathway. Eur Heart J 16:859–861, 1995.
30. Wong A, Marais H, Jutzy K, et al: Isolated atrial from failing and non-failing human hearts. J Mol 73. D’Avila A, Gutierrez P, Scanavacca M, et al: Effects
infarction in a patient with single vessel disease of Cell Cardiol 51:215–225, 2011. of radiofrequency pulses delivered in the vicinity of
the sinus node artery. Chest 100:255, 1991. 54. Flagg TP, Kurata HT, Masia R, et al: Differential the coronary arteries: Implications for nonsurgical
31. Nielsen FE, Andersen HH, Gram-Hansen P, et al: structure of atrial and ventricular KATP: Atrial transthoracic epicardial catheter ablation to treat
The relationship between ECG signs of atrial KATP channels require SUR1. Circ Res 103:1458, ventricular tachycardia. Pacing Clin Electrophysiol
infarction and the development of supraventricular 2008. 25:1488–1495, 2002.
arrhythmias in patients with acute myocardial 55. Kawada H, Niwano S, Niwano H, et al: Tumor 74. Hindricks G: The Multicentre European Radiofre-
infarction. Am Heart J 123:69–72, 1992. necrosis factor-alpha downregulates the voltage quency Survey (MERFS): Complications of radio-
32. Liu C, Greenspan G, Piccirillo RT: Atrial infarc- gated outward K+ current in cultured neonatal rat frequency catheter ablation of arrhythmias. The
tion of the heart. Circulation 23:331–338, 1961. cardiomyocytes: A possible cause of electrical Multicentre European Radiofrequency Survey
33. Sivertssen E, Hoel B, Bay G, et al: Electrocardio- remodeling in diseased hearts. Circ J 70:605–609, (MERFS) Investigators of the Working Group on
graphic atrial complex and acute atrial myocardial 2006. Arrhythmias of the European Society of Cardiol-
infarction. Am J Cardiol 31:450–456, 1973. 56. Janczewski AM, Kadokami T, Lemster B, et al: ogy. Eur Heart J 14:1644–1653, 1993.
34. Gardin JM, Singer DH: Atrial infarction: Impor- Morphological and functional changes in cardiac 75. Yokokawa M, Sundaram B, Garg A, et al: Impact
tance, diagnosis, and localization. Arch Intern Med myocytes isolated from mice overexpressing TNF- of mitral isthmus anatomy on the likelihood of
141:1345, 1981. alpha. Am J Physiol Heart Circ Physiol 284:H960– achieving linear block in patients undergoing cath-
35. Mayuga R, Singer D: Atrial infarction: Clinical H969, 2003. eter ablation of persistent atrial fibrillation. Heart
significance and diagnostic criteria. Practical 57. Roberts-Thomson KC, Steven D, Seiler J, et al: Rhythm 8:1404–1410, 2011.
Cardiol 11:142–160, 1985. Coronary artery injury due to catheter ablation in 76. Fuller IA, Wood MA: Intramural coronary vascu-
36. Bunc M, Starc R, Podbregar M, et al: Conversion adults: Presentations and outcomes. Circulation lature prevents transmural radiofrequency lesion
of atrial fibrillation into a sinus rhythm by coronary 120:1465–1473, 2009. formation: Implications for linear ablation. Circu-
angioplasty in a patient with acute myocardial 58. Pons M, Beck L, Leclercq F, et al: Chronic left lation 107:1797–1803, 2003.
infarction. Eur J Emerg Med 8:141, 2001. main coronary artery occlusion: A complication of 77. Ouali S, Anselme F, Savoure A, et al: Acute coro-
37. Pehkonen E, Honkonen E, Makynen P, et al: Ste- radiofrequency ablation of idiopathic left ventricu- nary occlusion during radiofrequency catheter
nosis of the right coronary artery and retrograde lar tachycardia. Pacing Clin Electrophysiol ablation of typical atrial flutter. J Cardiovasc Elec-
cardioplegia predispose patients to atrial 20:1874–1876, 1997. trophysiol 13:1047–1049, 2002.
The Molecular Pathophysiology
of Atrial Fibrillation 45
Stanley Nattel, Niels Voigt, and Dobromir Dobrev
is a well-recognized factor.6 Obesity is increasingly recognized as

CHAPTER OUTLINE
an AF risk factor,7 with obstructive sleep apnea, often associated
Etiologic Determinants 449 with obesity, also noted as an important contributor. Autonomic
tone may set the conditions for AF initiation and maintenance.
General Mechanisms and AF Forms 449
The AF-promoting properties of vagal activation are well known,
Molecular Control Mechanisms 450 and increasing evidence suggests an important role for combined
sympathovagal discharge.8
Future Directions 455
Atrial fibrillation (AF) is a highly prevalent and clinically relevant

arrhythmia, for which all current therapeutic approaches have General Mechanisms and AF Forms
important limitations. Most significant advances in arrhythmia
management have been based on an understanding of the under- AF can be maintained by rapid focal firing or by reentrant activity
lying mechanisms. An improved understanding of the mechanis- (Figure 45-2). To maintain AF, focal ectopic activity must be
tic basis of AF has evolved over the past 15 years, particularly sustained, to produce rapid “driver” activity that is conducted
with respect to molecular aspects. The purpose of this chapter is heterogeneously to generate fibrillatory conduction and the
to review recent findings in the molecular pathophysiology of AF, irregular activity typical of AF.9,10 Focal activity can also be tran-
and to discuss their potential value for improving management. sient, producing isolated atrial extrasystoles or self-limited tachy-
We will begin with an overview of the etiologic determinants cardias. Transient focal activity can contribute to AF generation
of AF (Figure 45-1), then discuss briefly the principal mecha- by acting as a trigger to initiate reentry in a vulnerable substrate.
nisms contributing to AF, and finally review the molecular basis Clinical AF can be paroxysmal (self-terminating), persistent (ter-
of specific arrhythmia determinants. minating only with medical intervention), or permanent (con-
tinuing despite medical therapy). Repetitively firing focal ectopic
drivers are believed to produce paroxysmal forms. Reentrant
activity generates more persistent AF, tending to become resis-
Etiologic Determinants tant as the substrate evolves, as well as more fixed and irrevers-
ible.11,12 One factor contributing to evolution of the substrate is
Heart Disease that AF induces remodeling, related to the rapid atrial rate and
to other factors like cardiac dysfunction, neurohumoral changes,
Most patients with AF have associated cardiac disease.1 Cardiac and consequences of atrial metabolic disturbances. The remodel-
senescence is a major predisposing factor, largely mediated by ing induced by chronic long-standing AF can involve both func-
structural remodeling, which causes fibrotic alterations and tional and structural changes that promote a transition toward
microconduction slowing, as well as atrial enlargement.2 Heart complex reentrant mechanisms.
failure (HF), hypertensive heart disease, valvular disease, and
ischemic heart disease are major contributors to AF occurrence.
Less common conditions leading to AF include pericarditis, myo- Focal Ectopic Activity
carditis, atrial tumors (primarily myxomas), and various
cardiomyopathies. Several mechanisms produce abnormal impulse formation and
can cause focal ectopic activity. Spontaneous automatic activity
depends on the balance between inward and outward currents
Genetic Determinants during phase 4 of the action potential. Increased phase 4 inward
currents carried by Na+ or Ca2+, particularly time-dependent acti-
Knowledge of genetic determinants of AF has increased rapidly vating currents like the “funny current” If, and/or decreased
over the past 10 years.3,4 Disease-causing mutations have been phase 4 outward currents, produce spontaneous phase 4 depolar-
established and the underlying pathophysiology studied.3 Mono- ization. When it reaches threshold potential, the cell fires, gen-
genic forms have high penetrance and provide important mecha- erating automatic ectopic activity.
nistic insights into AF mechanisms. Genome-wide association Focal ectopic activity may also result from afterdepolariza-
studies (GWASs) provide new insights into genetic variation– tions, which are subdivided into early afterdepolarizations (arising
based population determinants of AF, while raising challenging before the end of phase 3) and delayed afterdepolarizations
pathophysiological issues.4 (DADs), which occur after full repolarization. Cell Ca2+ handling
is crucial for normal contractility (Figure 45-3, A). DADs (Figure
45-3, B) are thought to be the most important cause of focal atrial
Extracardiac Contributors ectopic firing. DADs are caused by a diastolic Ca2+ leak from the
sarcoplasmic reticulum (SR) via SR Ca2+-release channels or
A variety of extracardiac conditions can affect AF occurrence. ryanodine receptors (RyRs; RyR2 is the cardiac form). Systolic Ca2+
Heavy alcohol consumption promotes AF,5 and hyperthyroidism release through RyR2s causes cardiac contraction. Relaxation is
449
Causes of Atrial Fibrillation Figure 45-1. Etiologic contributors to atrial fibrillation.
Heart disease
Coronary artery disease Heart failure
Valvular heart disease Hypertensive heart disease
Pericarditis /Myocarditis Myxoma Aging heart Cardiomyopathic Heart Disease
ATRIAL FIBRILLATION Diabetes
Abnormalities in cardiac Alcohol, drugs Thyroid dysfunction

structure or function
Autonomic tone
Altered ion channel
function Sleep apnea Obesity
Genetic factors Extrinsic factors
mediated by diastolic Ca2+ removal from the cytosol into the SR 45-4, B), reentry is maintained by rotors established by tissue
by a Ca2+-uptake pump, the SR Ca2+-adenosine triphosphatase excitability properties (depending on both conduction and refrac-
(ATPase; SERCA). RyR2s are sensitive to both cytosolic and toriness conditions), which determine rotor rapidity, stability, and
intraluminal SR Ca2+ concentration; diastolic releases result when size (greater excitability generates smaller, more stable, and faster
SR concentrations are excessive, or when RyR2s have an abnor- rotors). Anatomical obstacles or complexities can favor reentry
mally low threshold for Ca2+ release. Excess cytosolic Ca2+ is by anchoring reentry circuits. Figure 45-4, C illustrates the effect
handled by the sarcolemmal Na+/Ca2+-exchanger (NCX), which of structural remodeling. Progressive atrial dilatation creates
moves three Na+ ions (charge +3) into the cell for each Ca2+ ion longer conduction pathways for reentry. Tissue fibrosis slows
(charge +2) extruded into the extracellular space, generating net conduction and creates conduction barriers that favor the devel-
inward current (called transient inward current [Iti]) that depolar- opment of stable rotors and/or multiple simultaneous irregular
izes the cell, producing a DAD. When DADs reach threshold, reentry circuits that can sustain AF. In addition, fibroblast prolif-
they induce premature action potential firing (dashed line in eration can promote arrhythmogenesis via cardiomyocyte-
Figure 45-3, B). Repeated DADs can generate focal atrial tachy- fibroblast interactions that alter AP properties and slow
cardias. RyR2 function is regulated by channel phosphorylation: conduction.
Hyperphosphorylation enhances RyR Ca2+ sensitivity and pro-
motes DAD formation. Calsequestrin (CSQ) is the principal
Ca2+-storage buffer of the SR. Inadequate CSQ function/
expression increases free SR Ca2+ concentration and promotes
Molecular Control Mechanisms
diastolic RyR2 Ca2+ release.13
Early afterdepolarizations (EADs) result from spontaneous Molecular Control of Cell Ca2+ Handling
depolarization during phase 2 or 3 of the action potential (AP), and DAD Generation
generally when AP duration (APD) is excessively prolonged.
With very long APs, L-type Ca2+ currents may have enough time Normal cell Ca2+ handling is crucial for cellular contraction and
to recover from inactivation, carrying inward Ca2+ current to relaxation (see Figure 45-3, A; see also Chapter 16 of this book).
generate an EAD. Abnormal SR Ca2+ handling is typically seen in AF patients.14-19
Defective Ca2+ handling promotes spontaneous RyR2-mediated
diastolic SR-Ca2+ releases in atrial cells from patients with chronic
Reentry AF.14,15,18,19 Figure 45-5 summarizes the detailed molecular patho-
biology of DAD-inducing diastolic RyR2 Ca2+ release. Protein
Two conceptual frameworks for understanding the basis of func- kinase A (PKA) phosphorylation of RyR2 at Ser280817 and Ca2+
tional reentry are shown in Figure 45-4 (top). The leading circle calmodulin–dependent kinase II (CaMKII) phosphorylation at
model (Figure 45-4, A) posits reentry around a central zone that Ser2814 are increased in dogs and goats with pacing-induced AF
is continuously activated by centripetal waves emanating from a and in AF patients.14,19-21 CaMKII activity is normally autoinhib-
reentering activation wave front. In this model, reentry estab- ited. Ca2+-calmodulin binding removes autoinhibition, activating
lishes itself in a circuit for which the dimension equals the CaMKII and causing autophosphorylation that activates CaMKII
distance traveled during one refractory period (wavelength; and makes it Ca2+ independent. Similar activation may result
refractory period times conduction velocity). When the wave- from CaMKII oxidation. Changes in RyR2 phosphorylation state
length is small because of slow conduction or brief refractoriness, at PKA and CaMKII sites may result not only from changed
multiple circuits can be accommodated in the atria and spontane- kinase activity, but also from alterations in dephosphorylating
ous self-termination is unlikely. In the spiral wave model (Figure enzyme phosphatases.22 These posttranslational alterations
The Molecular Pathophysiology of Atrial Fibrillation 451
Atrial Fibrillation – Triggers and Substrates remodeling (ATR) and is greater in atria than in ventricles.28
45
IP3R2-coupled amplification of atrial SR Ca2+-release events and
Focal Transient related arrhythmogenesis may thus contribute to AF-related
Trigger + Substrate
firing ectopic activity.
Congestive heart failure (CHF) is a very important cause of
Su
Single Reentry AF. Focal drivers and triggered activity play a role in CHF-
sta
circuit related AF.29 In experimental dilated cardiomyopathy, CHF

ine
reentry
increases SR Ca2+ load and reduces calsequestrin expression,
d
Fibrillatory
conduction Driver thereby promoting spontaneous SR Ca2+ release.30 Coronary
Remodeling Multiple artery disease (CAD) is also an important risk factor for AF. Atrial
circuit ischemia promotes AF maintenance.31 In a dog model of chronic
reentry occlusive coronary artery disease affecting the atrium, frequent
spontaneous atrial ectopy is associated with an increased inci-
ATRIAL dence of atrial cardiomyocyte-triggered activity.31 Triggered
FIBRILLATION activity is likely due to spontaneous SR Ca2+-release events and
A increased NCX function in cardiomyocytes from the ischemic
border zone.31
Atrial Fibrillation – Forms Two AF-promoting genetic variants have been linked to DAD
Ectopic mechanisms: (1) a mutation of the gene encoding the adapter
activity protein ankyrin-B (long QT syndrome-4 [LQTS4]), which
Focal
Drivers causes multiple proteins to be poorly addressed to their mem-
AF
brane targets, altering Ca2+ handling and leading to DADs/

Triggers Remodeling triggered activity32,33; and (2) a predicted loss-of-function single-
al
AF
nucleotide polymorphism (SNP) of the SLN gene encoding the

sm
Ca2+-binding protein sarcolipin, which could increase SR Ca2+

oxy
F
ent
t A
load, thereby affecting DAD susceptibility.34

Par
sist
Beta-adrenoceptor activation phosphorylates RyR2, promot-

nen
Per
ing diastolic SR Ca2+-release events.34 Conditions that directly

ma
cause DAD-promoting abnormalities in Ca2+ handling may

Per
require adrenergic stimulation to induce Ca2+ sparks and trig-

Functional Reentry gered activity.31 Spontaneous AF paroxysms occur in dog models
Substrate
Fixed of autonomic hyperinnervation,36 with sympathovagal discharge
Substrate preceding AF paroxysms.37 Vagal activation promotes arrhyth-
B mogenesis by reducing APD, allowing afterdepolarizations
Figure 45-2. Mechanistic Basis of Atrial Fibrillation (AF) and Associ- induced by adrenergic stimulation to induce ectopic firing in
ated Clinical Forms A, Focal firing usually results from local ectopic activity. pulmonary veins.38-40
Organized discrete reentrant activity and focal firing can maintain AF by producing
regularly firing drivers that are conducted irregularly in the heterogeneous atrial
substrate. B, Paroxysmal (self-terminating) AF is believed to result primarily from Molecular Control of L-Type Ca2+ Current and
focal drivers, persistent AF from functional substrates, and permanent (nontermi- Pathophysiology in AF
nating) AF from fixed fibrotic/anatomically remodeled substrates. However, overlap
and partial contributions are indicated by the gray shading. AF41 and indeed all very rapid atrial tachyarrhythmias42 remodel
atrial electrical properties to promote AF initiation and mainte-
nance (ATR). A major AF-promoting component of ATR is
increase RyR2 Ca2+ sensitivity, enhancing channel open probabil- refractory period reduction due to APD abbreviation (see Figure
ity.14,17 Mice deficient in RyR2-inhibitory FK-505 binding protein 45-4, B). Reduced depolarizing L-type Ca2+ current (ICa,L) and
12.6, mice with gain-of-function mutations in RyR2, and mice increased repolarizing inward-rectifier K+ currents underlie
with constitutively phosphorylated RyR2 channels at S2814 ATR-induced APD shortening.43-50 The molecular complex basis
(S2814D mice) all exhibit increased susceptibility to pacing- of ICa,L reduction seen in AF is illustrated in Figure 45-6, A. Rapid
induced AF in association with increased atrial cell SR Ca2+ leak atrial activation induces Ca2+ loading, activating Ca2+-calmodulin/
and triggered activity.14,20,23,24 Angiotensin effects to promote AF calcineurin/nuclear factor of activated T-lymphocyte (NFAT)
may be due in part to oxidative stress acting via CaMKII oxida- signaling that causes downregulation of Cav1.2 α-subunit
tion on diastolic RyR2 Ca2+ release.25 RyR2 dysfunction can be mRNA.51-53 Other contributors to ICa,L downregulation may
induced by Ca2+ overload resulting from phospholamban hyper- include decreased expression of accessory β1-, β2a-, β2b-, β3-, and
phosphorylation, which removes phospholamban inhibition of α2δ2-subunits45,54,55; Cav1.2 dephosphorylation via type 1 (PP1)
SERCA and enhances SR Ca2+ uptake.22 Phospholamban and/or type 2A (PP2A) protein phosphatases22,45,56; and increased
hyperphosphorylation can be produced by enhanced PKA or Cav1.2 α-subunit s-nitrosylation.57 MicroRNAs (miRNAs) are
CaMKII activity, or by decreased phosphatase function. Reduced short RNA sequences that regulate cardiac gene expression by
phosphatase function can be a consequence of increased translational suppression and/or mRNA destabilization.58 They
activity of an inhibitory protein, I-1, typically caused by I-1 are centrally involved in cardiac remodeling and appear to play
hyperphosphorylation.22 a major role in AF.58 Recent work implicates increased miRNA-
Increases in NCX expression and/or function are also com- 328 in AF promotion due to ICa,L-downregulation mediated by
monly noted in AF,14,19,22,26 causing Iti resulting from any specific inhibition of translation and mRNA destabilization.59 Finally,
amount of diastolic SR Ca2+ leak to be larger in AF, likely con- impaired Cav1.2 and Cav1.3 protein trafficking induced by a
tributing to the increased risk of DADs and triggered ectopic zinc-binding protein (Znt-1) and ankyrin B reduction, respec-
activity.14,27 Cardiac IP3 receptors (IP3R2) act as Ca2+-transporting tively, may contribute to ICa,L reduction in AF.33,60
pathways and can facilitate SR Ca2+ leak to promote arrhythmo- Loss-of-function ICa,L subunit mutations would also be antici-
genesis. IP3R2 expression is increased by atrial tachycardia pated to accelerate repolarization, thereby promoting AF. In 82
A. Systolic Ca2+-handling B. Diastolic Ca2+-handling

producing cell contraction abnormalities and DAD generation
DADs
AF
200ms 200ms
Extracellular Extracellular Repolarizing
space Ca2+ space conductances
Ca2+
NCX NCX
Iti
3Na+ 3Na+ Depolarizing
conductances
Cytosol Ca2+ Cytosol
Ca2+
P P P P
CSQ RyR2 CSQ RyR2

SERCA SERCA
[Ca2+] [Ca2+]
SR SR
Systole Diastole
Figure 45-3. Cellular Ca2+ Handling and DADs A, During the plateau phase of the action potential (AP), Ca2+ enters the cell via L-type Ca2+ channels. This Ca2+ binds
to ryanodine-receptors (RyR2s), thereby triggering a much larger Ca2+ release from the sarcoplasmic reticulum (SR), which initiates cellular contraction. SR Ca2+ stores are
maintained via Ca2+ pumping into the SR by the SR Ca2+–adenosine triphosphatase (ATPase; SERCA). B, Diastolic Ca2+-handling abnormalities underlie DADs. Spontaneous
SR Ca2+ releases through RyR2 elevate cytosolic Ca2+, which is exchanged for extracellular Na+ by the Na+/Ca2+ exchanger (NCX), producing depolarizing current (transient
inward-current [Iti]). Inappropriate diastolic RyR2 Ca2+ release is produced by RyR2 hyperphosphorylation, excess SR Ca2+, or decreased SR Ca2+ binding to calsequestrin (CSQ).
Repolarizing conductances oppose Iti and suppress diastolic depolarization, so reduced diastolic K+ current can favor DADs P, Phosphate; Iti, transient inward current..
patients with Brugada syndrome/short QT electrocardiographic ATR-induced arrhythmogenic remodeling (see also Chapter 38
(ECG) phenotypes, loss-of-function mutations of the CACNA1C of this book).
and CACNB2 genes, encoding ICa,L α- and β-subunits, were The most common AF-promoting monogenic paradigm is
observed along with AF in individual patients.61 Patients with accelerated atrial repolarization due to gain-of-function K+-
full-blown short QT syndromes of various phenotypes have channel mutations like the first mutation linked to lone AF, which
reduced APDs and are predisposed to AF.3 causes a gain-of-function in KCNQ1 (encoding α-subunit of
IKs).70 Other AF-inducing gene mutations believed to act via K+-
channel gain-of-function mutations have been reported in
Molecular Control of K+ Current and Basis KCNH2 (hERG71), KCNJ2 (Kir2.172), and KCNE2 (MiRP173). An
of AF-Promoting Alterations SNP in the NOS3 (endothelial nitric oxide synthase) encoding
gene74 has been implicated in AF. NOS3 regulates vagal signaling
Inward-rectifier K+-current enhancement promotes AF mainte- and ICa,L, and could thereby contribute to reentrant AF.75,76
nance by reducing APD (favoring reentry) and stabilizing/ Vagal enhancement is known to promote clinical AF and is
accelerating arrhythmia-maintaining rotors by removing voltage- central in some cases.77 IK,ACh hyperpolarizes atrial cardiomyo-
dependent INa inactivation through membrane hyperpolariza- cytes and reduces APD in a spatially heterogeneous way. Vagal
tion.9,62 Figure 45-6, B shows how IK1 and IK,AChc are upregulated enhancement strongly favors AF initiation and persistence by
in AF. IK1 is increased as the result of upregulation of the underly- facilitating the initiation and subsequent stability of reentrant
ing Kir2.1 subunit46-50,55,63-65 caused by reduced Kir2.1-inhibitory rotors.77 Kir3.4 knock-out strongly suppresses IK,ACh and prevents
microRNAs, like miR-1, miR-26, and miR-101.65,66 Kir2.1 cholinergic AF.78
dephosphorylation (activation) via increased PP1 and PP2A func-
tion22,45,56,67 may also contribute.
Agonist-induced muscarinic receptor–mediated IK,ACh activa- Molecular Determinants of Atrial
tion is reduced in AF.46,47 However, increased agonist-independent Conduction Disturbances
(constitutive) IK,AChc is enhanced both in dog models46,49,64,68 and
in AF patients.48,50,63,69 Increased IK,AChc is due to greater IK,ACh Conduction Abnormalities Due to Ion Channel Dysfunction
channel opening probability, with no change in single-channel Conduction slowing favors reentry. Gap junctions are essential
conductance, kinetics, or density.48,64 A key role is played by for efficient cell coupling and conduction. Discrepant results
altered IK,AChc protein kinase C (PKC) phosphorylation, with about AF-related atrial gap junctional remodeling have been
increased phosphorylation by stimulatory Ca2+-dependent iso- reported in the literature.55,79,80 Some of the variability may be
forms and reduced inhibitory classical Ca2+-independent isoform due to differences in AF duration, underlying heart disease, and
function.68,69 IK,AChc inhibition suppresses atrial tachyarrhythmias species-related factors.81 Spatially heterogeneous connexin 40
in ATR preparations,49 suggesting that IK,AChc contributes to remodeling occurs in the goat model of electrically maintained
MECHANISMS OF REENTRY Factors promoting AF by inducing diastolic
45
Functional Determinants Ca2+ leak through RyR2
Extracellular
Ang-II space
CaMKII
Refractory (autoinhibited)
Core NADPHox Atrial
CaM Ca2+ rate
Ca2+
A Leading circle B Spiral wave Ox CaM P
Thr286
? CaMKII CaMKII CaMKII

(oxidized) (Ca2+/CaM-dependent) (autophosphorylated)
Structural remodeling LA enlargement
LA LA CaMKII activity
P Thr35
RA Ca2+ PP1 I-1
RA Leak Cytosol
Ser2808 Ser2814
P P
LA fibrosis PP1
RyR2
CSQ
P Ser16
PLN
P Thr17
[Ca2+] SERCA
AF maintaining SLN
SR
substrate
C
Figure 45-4. Determinants of Reentry Top, Basic concepts of reentry. The Figure 45-5. Molecular Basis of DAD-Inducing Diastolic Ca2+ Releases
leading circle model (A) posits reentry around a central zone that is continuously RyR dysfunction is caused by RyR hyperphosphorylation or excess Ca2+ loads. Phos-
activated by centripetal waves emanating from the reentering activation wave pholamban (PLN) inhibits SERCA. PLN hyperphosphorylation removes this inhibi-
front. The spiral wave model (B) describes reentry as a “rotor” established by tissue tory effect, enhances SERCA function, and can lead to Ca2+ overload. High atrial rate
excitability properties. C, Structural remodeling (atrial enlargement and fibrosis, during AF enhances cellular Ca2+ entry. Increased cell Ca2+ promotes Ca2+/calmodu-
most typically affecting the left atrium [LA]) produces relatively fixed reentry sub- lin (CaM) binding to Ca2+/calmodulin-dependent protein kinase II (CaMKII), disin-
strates that reverse poorly if at all. hibiting the catalytic subunit. After CaMKII catalytic subunit activation, oxidation at
Met281/282 or phosphorylation at Thr286 causes persistent CaMKII activity.
Inhibitor-1 (I-1) suppresses protein phosphatase 1 (PP1) function in the SR and
AF, consistent with clinical data indicating that gene variants that contributes to PLN and RyR phosphorylation. Ang-II, Angiotensin-II; NADPHox, Nico-
tinamide adenine dinucleotide phosphate oxidase; CaM, calcium/calmodulin;
affect connexin 40 promoter function may predispose to AF.82-84 CaMKII, calcium-calmodulin dependent kinase type II; CSQ, calsequestrin; PLN,
Connexin 43 dephosphorylation/lateralization occurs in CHF, phospholamban; SERCA, sarcoplasmic reticulum calcium-ATPase; SLN, sarcolipin;
but CHF-induced conduction slowing and AF promotion are RyR2, ryanodine receptor type 2.
unchanged with CHF recovery, despite disappearance of con-
nexin abnormalities.85 Recent evidence indicates that connexin 43
gene transfer can improve conduction and suppress AF in porcine
models, supporting the importance of gap junction protein
remodeling in AF.86,87 slowing. SCN5A mutations were initially associated with AF in a
INa provides the energy for conduction and governs conduc- family presenting a complex and variable phenotype, including
tion velocity (CV). INa density decreases in canine ATR, with dilated cardiomyopathy, AF, sinus node dysfunction, and conduc-
corresponding decreases in SCN5A-subunit mRNA and protein.43 tion defects.92 SCN5A mutations and SNPs were subsequently
In humans with AF, SCN5A mRNA expression appears found in idiopathic AF subjects.93,94 Loss-of-function SCN5A
unchanged.55 Atrial cardiomyocytes from AF patients showed a mutations are the most common cause of Brugada syndrome,95
slightly reduced INa.88 which typically presents as ventricular fibrillation (VF)/sudden
Atrial ischemic disease causes localized conduction slowing, death, but can also cause AF.96 Recently, mutations in Na+-
which allows for AF-sustaining local reentry stabilized around a channel β-subunits like SCN1B, -2B, and -3B have been impli-
line of conduction block.31 With acute ischemia, gap junction cated in AF.97-99
uncoupling predominates.89
A number of AF-associated gene variants affect ion channels Structural Remodeling
that control cardiac conduction. GJA5 encodes connexin 40, an Atrial fibrosis plays an important role in AF of different
atrial-selective gap junction ion channel. Connexin 40 knockout origins.11,31,100,101 The complex underlying signaling pathways are
causes conduction abnormalities and atrial arrhythmia suscepti- shown in Figure 45-7. The development of atrial fibrosis is likely
bility.90 An AF-causing GJA5 missense somatic mutation was determined by multiple signals acting simultaneously. Cardio-
identified in idiopathic AF patients.83 GJA5-promoter variants myocytes and fibroblasts interact extensively at the level of auto-
believed to decrease gene transcription increase AF crine and paracrine factors, and possibly electrically as well.11,102
susceptibility.82,84,91 Fibroblasts produce extracellular matrix (ECM) proteins and
Cardiac Na+-channel gene (SCN5A) loss-of-function muta- mediators that affect cardiomyocyte phenotype, whereas cardio-
tions cause AF, presumably via reentry-promoting conduction myocytes generate products like reactive oxygen species (ROS),
L-type Ca2+ current Inward-rectifier K+ current

remodeling remodeling
APD APD
ICa,L Hyperpolarization
P NO IK1 Basal IR IK,AChc

current
ZnT-1 PP1/PP2A GSH Calpains
Degradation
Atrial Oxidative
rate Stress P
Trafficking
PKC
PKC
Ca2+
NFAT
CaM
Cav1.2 Kir2.1
miR-101
Translation
Calcineurin NFATc3/c4 P
ion
miR-26
lat
P
ns
miR-1
r a
miR-328
T
NFATc3/c4
Transcription
A Cytosol mRNA DNA Nucleus

B mRNA
Nucleus Cytosol
Figure 45-6. Mechanisms of Ionic Current Remodeling in AF A, L-type Ca2+-current (ICa,L) downregulation. High atrial rates in AF enhance intracellular Ca2+ load
and activating calcineurin via Ca2+/calmodulin (CaM) binding. Calcineurin dephosphorylates nuclear factor of activated T-lymphocytes (NFAT), allowing it to translocate into
the nucleus and reduce mRNA levels of the ICa,L alpha subunit, Cav1.2. Breakdown of Cav1.2 protein by calpains might also contribute to reduced Cav1.2 protein expression.
The protein zinc transporter 1 (ZnT-1) impairs Cav1.2 membrane trafficking and is upregulated in AF. Increased protein phosphatase (PP) activity dephosphorylates Cav1.2
phosphorylation and may decrease ICa,L. B, Inward-rectifier K+-current upregulation. Increased Kir2.1 subunit expression is caused by decreases in inhibitory microRNAs like
miR-101, miR-26, and miR-1. Acetylcholine-regulated K+ current (IK,AChc) is increased because of increased membrane abundance of the stimulatory protein kinase C (PKC)
isoform PKCε and decreased cellular expression of the inhibitory isoform PKCα. PP, Protein phosphatase; ZnT-1, zinc transporter-1; GSH, reduced glutathione; miR, microRNA;
NFAT, nuclear factor of activated T-cells; PKC, protein kinase C; IK1, background inward-rectifier current; IK,AChc, constitutive acetylcholine-regulated potassium current; IR,
inward rectifier.
platelet-derived growth factor (PDGF), transforming growth fibroblasts into ECM-secreting myofibroblasts.110,111 Progressive
factor-β (TGF-β), and connective tissue growth factor (CTGF), fibrosis likely contributes to conduction disturbances that make
which modulate fibroblast function. long-lasting AF very difficult to treat.12
Angiotensin-II (AT-II) plays an important role in AF,103 likely PDGF stimulates fibroblast proliferation and differentia-
in large measure via fibroblast modulation. AT-II type 1 receptors tion.11,102,107 PDGF receptors contain two transmembrane
(AT1Rs) promote fibrosis via enhanced actions of TGF-β, domains that dimerize upon stimulation and then activate an
Smad2/3, Smad4, Arkadia, and activated extracellular signal– internal tyrosine kinase. Tyrosine kinase autophosphorylation of
regulated (ERK) mitogen-activated protein kinase (MAPK).104 PDGF receptors induces Ras/MEK1/2, MAPK, JAK/STAT, and
Arkadia promotes ubiquination and removal of Smad7, thereby PLC signaling. PDGF overexpression induces cardiac fibrosis
increasing TGF-β signaling by removing Smad7 antagonism.105 and dysfunction.112 Atrial-selective PDGF expression and action
In addition, AT1Rs act through Shc/Grb2/SOS to activate Ras, may contribute to the greater fibrotic responses typically seen for
which enhances MAPK phosphorylation,106 and through phos- atria versus ventricles.107
pholipase C (PLC). PLC breaks down phosphatidylinositol Connective tissue growth factor (CTGF) lies downstream to
4,5-bisphosphate (PIP2), yielding diacylglycerol (DAG) and ino- TGF-β1 and AT-II in profibrotic signaling pathways. CTGF acti-
sitol 1,4,5-trisphosphate (IP3). DAG activates PKC, and IP3 vates fibroblasts through Src kinase and MAPKs.113 CTGF has
mobilizes intracellular Ca2+—both actions contribute to remodel- emerged as a potentially central player in atrial structural
ing. The JAK/STAT pathway, also AT1R-sensitive, activates remodeling.114-116
transcription factors such as activator protein 1 (AP-1) and MicroRNAs appear to contribute importantly to atrial struc-
nuclear factor kappa light-chain enhancer of activated B cells tural remodeling. MiR-29 inhibits collagen gene expression,117
(NF-κB), which cause further cardiomyocyte remodeling. AT2R and its downregulation likely contributes to atrial fibrosis in
activation counters AT1R-mediated MAPK activation by enhanc- CHF.118 MiR-30 and miR-133 are also downregulated in CHF,
ing dephosphorylation via PP2A and phosphotyrosine phospha- suppress CTGF translation,119 and may participate in atrial fibro-
tase (PTP).106 sis.120 In contrast to miR-29, miR-20, and miR-133, atrial miR-21
TGF-β1 is a key player in cardiac fibrosis. It is secreted by expression increases in CHF.120 MiR-21 targets the Sprouty-1
both fibroblasts and cardiomyocytes.107 Overactivity of cardiac (Spry-1) gene, which promotes fibroblast MAPK phosphoryla-
TGF-β1 causes atrial-selective fibrosis, conduction abnormalities, tion and enhances fibroblast survival.121 Although the significance
and AF promotion.108 TGF-β1 mediates AT-II effects in both of this finding has been questioned,122 recent work shows that
paracrine and autocrine fashion.109 TGF-β1 acts through SMADs atrial miR-21 knock-down suppresses CHF-associated atrial
to activate fibroblasts and enhance collagen production.11,102,105 fibrosis and AF promotion.123
Rapidly firing atrial cardiomyocytes produce Ang-II and ROS, SNPs in genes determining atrial structural integrity,
acting via enhanced TGF-β production to differentiate cardiac inflammation, and neurohumoral control have been associated
miR-30
PDGF miR-133
45
TGF-β
Aldosterone
Ang-II CTGF
Extracellular Space
Oxidase
NAPDH
TGFβR
PDGFR
α β
MR
Cell membrane
Ang-II-R Ang-II-R
type 2 type 1
Grb2 SOS Src
Grb2 SOS
Shc
Shc Cytoplasm
JAK PLC SMAD 7
SMAD 2/3
PTP ROS
SMAD 2/3
PIP2
TAK1 ubiquitination
Ras
PP2a
Arkadia
DAG
MEK1/2 SMAD 4
MAPK-P SMAD 6
ERK
JNK
Spry-1 P38 STAT PKC
miR-21 ? Nucleus
TRANSCRIPTION FACTORS:
c-jun NF κβ Elk-1 AP-1 c-fos MEF2
Proliferation ECM synthesis

fibroblast Differentiation Altered MMP/TIMP function FIBROSIS
miR-29
Figure 45-7. Molecular Pathways Causing Atrial Fibrosis αβ, Integrin receptor α- and β-subunits; Ang-II, angiotensin-II; AP-1, activator protein 1; DAG, diacyl
glycerol; ERK, extracellular signal–related kinase; Grb2, growth factor receptor binding protein 2; JAK, Janus kinase; JNK, c-Jun N-terminal kinase; MAPK, mitogen-activated
protein kinase; MEF2, myocyte enhancer factor 2; MEK1/2, mitogen-activated /ERK kinase 1/2; MMP, matrix metalloproteinase; MR, mineralocorticoid receptor; NF-κβ, nuclear
factor κβ; PDGF(R), platelet-derived growth factor (receptor); PIP2, phosphatidylinositol 4,5-bisphosphate; PKC, protein kinase C; PLC, phopholipase C; PP2A, protein phos-
phatase 2A; PTP, phosphotyrosine phosphatase; Ras, rat sarcoma GTPase protein; ROS, reactive oxygen species; shc, src homologous protein; SMAD, SMA- and MAD-related
proteins; SOS, son of sevenless protein; Spry-1, sprouty 1; STAT, signal transducers and activators of transcription; TAK1, TGF-β1–activated kinase 1; TGFβR: TGF-β receptor;
TIMP, tissue inhibitor of matrix metalloproteinases.
with AF by conventional approaches. Examples include genes In addition to improved therapies per se, insights into the
encoding angiotensin-converting enzyme (ACE),124,125 matrix molecular basis of AF may allow for the identification of novel
metalloproteinase-2 (MMP2), and interleukin-10.126 However, biomarkers and could help in developing personalized medicine
these have not been replicated in hypothesis-free large-scale targeting patient-specific features. It is hoped that by the next
population studies. GWASs have implicated SNPs on chromo- edition of this book, advances in understanding the molecular
some 4q25, with PITX2c being the closest potential target pathobiology of AF will already have led to improved patient
gene, and an SNP in chromosome 16q22 near the zinc management.
finger homeobox 3 transcription factor gene (ZFHX3).38,39,127
PITX2c is involved in cardiac development, particularly sidedness
and pulmonary vein aspects,128,129 implicating possible structural
abnormalities as the way in which it may be involved in AF. Acknowledgments
ZFHX3 is a tumor suppressor130,131 that induces expression of
PDGF receptors and protects against oxidant stress,132 also sug- The authors thank Jennifer Bacchi for excellent secretarial help.
gesting structural remodeling as a potential mechanism of AF
promotion.
Funding
Supported by the Canadian Institutes of Health Research
Future Directions (MGP6957 and MOP44365), the Quebec Heart and Stroke
Foundation, the Foundation Leducq (European-North American
The largest and most important challenge related to the molecu- Atrial Fibrillation Research Alliance [ENAFRA], grant
lar basis of AF involves translating our increased knowledge into 07CVD03), the German Federal Ministry of Education and
practical clinical applications. It is widely hoped that targeting Research through Atrial Fibrillation Competence Network
the molecular mechanisms of AF will permit the development of (grant 01Gi0204), the German Centre for Cardiovascular
novel, safer, more specific treatment approaches. The ideas of Research, the Deutsche Forschungsgemeinschaft (grant Do
targeted remodeling,133 gene- and cell-based biologicals,134 769/1-3), and the European Union (European Network for
and manipulation of miRNAs135 are all related to the molecular Translational Research in Atrial Fibrillation [EUTRAF], grant
pathophysiology of AF and may lead to new targets/treatments. 261057).
between two goat models of atrial dysfunction: pacing-induced heart failure. Heart Rhythm
References AV block-induced atrial dilatation and atrial 5:1170–1177, 2008.
fibrillation. J Mol Cell Cardiol 46:385–394, 2009. 41. Wijffels MC, Kirchhof CJ, Dorland R, et al: Atrial
1. Kozlowski D, Budrejko S, Lip GY, et al: Lone 22. El-Armouche A, Boknik P, Eschenhagen T, et al: fibrillation begets atrial fibrillation: A study in
atrial fibrillation: What do we know? Heart Molecular determinants of altered Ca2+ handling awake chronically instrumented goats. Circula-
96:498–503, 2010. in human chronic atrial fibrillation. Circulation tion 92:1954–1968, 1995.
2. Nattel S: From guidelines to bench: Implications 114:670–680, 2006. 42. Shiroshita-Takeshita A, Mitamura H, Ogawa S, et
of unresolved clinical issues for basic investiga- 23. Sood S, Chelu MG, van Oort RJ, et al: Intracel- al: Rate-dependence of atrial tachycardia effects
tions of atrial fibrillation mechanisms. Can J lular calcium leak due to FKBP12.6 deficiency in on atrial refractoriness and atrial fibrillation main-
Cardiol 27:19–26, 2011. mice facilitates the inducibility of atrial fibrilla- tenance. Cardiovasc Res 81:90–97, 2009.
3. Mahida S, Lubitz SA, Rienstra M, et al: Mono- tion. Heart Rhythm 5:1047–1054, 2008. 43. Yue L, Feng J, Gaspo R, et al: Ionic remodeling
genic atrial fibrillation as pathophysiological para- 24. Li N, Wang T, Wang W, et al: Inhibition of underlying action potential changes in a canine
digms. Cardiovasc Res 89:692–700, 2010. CaMKII phosphorylation of RyR2 prevents model of atrial fibrillation. Circ Res 81:512–525,
4. Sinner M, Ellinor PT, Meitinger T, et al: induction of atrial fibrillation in FKBP12.6 1997.
Genome-wide association studies: Principles and knockout mice. Circ Res 110:465–470, 2012. 44. Van Wagoner DR, Pond AL, Lamorgese M, et al:
new insights into AF pathophysiology. Cardiovasc 25. Purohit A, Swaminathan PD, Chen B, et al: Atrial L-type Ca2+ currents and human atrial
Res 89:701–709, 2011. Angiotensin II promotes atrial fibrillation in mice fibrillation. Circ Res 85:428–436, 1999.
5. Mukamal KJ, Tolstrup JS, Friberg J, et al: Alcohol by CaMKII oxidation. Circulation 124(Suppl) 45. Christ T, Boknik P, Wohrl S, et al: L-type Ca2+
consumption and risk of atrial fibrillation in men :A14037, 2011. current downregulation in chronic human atrial
and women: The Copenhagen City Heart Study. 26. Lenaerts I, Bito V, Heinzel FR, et al: Ultrastruc- fibrillation is associated with increased activity of
Circulation 112:1736–1742, 2005. tural and functional remodeling of the coupling protein phosphatases. Circulation 110:2651–
6. Auer J, Scheibner P, Mische T, et al: Subclinical between Ca2+ influx and sarcoplasmic reticulum 2657, 2004.
hyperthyroidism as a risk factor for atrial fibrilla- Ca2+ release in right atrial myocytes from experi- 46. Ehrlich JR, Cha TJ, Zhang L, et al: Characteriza-
tion. Am Heart J 142:838–842, 2001. mental persistent atrial fibrillation. Circ Res tion of a hyperpolarization-activated time-
7. Schoonderwoerd BA, Smit MD, Pen L, et al: New 105:876–885, 2009. dependent potassium current in canine
risk factors for atrial fibrillation: Causes of ‘not- 27. Dobrev D, Voigt N, Wehrens XH: The ryanodine cardiomyocytes from pulmonary vein myocardial
so-lone atrial fibrillation.’ Europace 10:668–673, receptor channel as a molecular motif in atrial sleeves and left atrium. J Physiol 557(Pt 2):583–
2008. fibrillation: Pathophysiological and therapeutic 597, 2004.
8. Chou CC, Chen PS: New concepts in atrial fibril- implications. Cardiovasc Res 89:734–743, 2012. 47. Dobrev D, Graf E, Wettwer E, et al: Molecular
lation: Neural mechanisms and calcium dynamics. 28. Zhao ZH, Zhang HC, Xu Y, et al: Inositol-1,4,5- basis of downregulation of G-protein-coupled
Cardiol Clin 27:35–43, viii, 2009. trisphosphate and ryanodine-dependent Ca2+ sig- inward rectifying K+ current (IK,ACh) in chronic
9. Nattel S, Burstein B, Dobrev D: Atrial remodel- naling in a chronic dog model of atrial fibrillation. human atrial fibrillation: decrease in GIRK4
ing and atrial fibrillation: Mechanisms and impli- Cardiology 107:269–276, 2007. mRNA correlates with reduced IK,ACh and musca-
cations. Circ Arrhythm Electrophysiol 1:62–73, 29. Ryu K, Shroff SC, Sahadevan J, et al: Mapping of rinic receptor-mediated shortening of action
2008. atrial activation during sustained atrial fibrillation potentials. Circulation 104:2551–2557, 2001.
10. Berenfeld O, Zaitsev AV, Mironov SF, in dogs with rapid ventricular pacing induced 48. Dobrev D, Friedrich A, Voigt N, et al: The G
et al: Frequency-dependent breakdown of wave heart failure: Evidence for a role of driver regions. protein-gated potassium current IK,ACh is constitu-
propagation into fibrillatory conduction across J Cardiovasc Electrophysiol 16:1348–1358, tively active in patients with chronic atrial fibril-
the pectinate muscle network in the isolated 2005. lation. Circulation 112:3697–3706, 2005.
sheep right atrium. Circ Res 90:1173–1180, 30. Yeh YH, Wakili R, Qi XY, et al: Calcium-handling 49. Cha TJ, Ehrlich JR, Chartier D, et al: Kir3-based
2002. abnormalities underlying atrial arrhythmogenesis inward rectifier potassium current: Potential role
11. Burstein B, Nattel S: Atrial fibrosis: Mechanisms and contractile dysfunction in dogs with conges- in atrial tachycardia remodeling effects on atrial
and clinical relevance in atrial fibrillation. J Am tive heart failure. Circ Arrhythm Electrophysiol repolarization and arrhythmias. Circulation
Coll Cardiol 51:802–809, 2008. 1:93–102, 2008. 113:1730–1737, 2006.
12. de Groot NM, Houben RP, Smeets JL, et al: Elec- 31. Nishida K, Qi XY, Wakili R, et al: Mechanisms of 50. Voigt N, Friedrich A, Bock M, et al: Differential
tropathological substrate of longstanding persis- atrial tachyarrhythmias associated with coronary phosphorylation-dependent regulation of consti-
tent atrial fibrillation in patients with structural occlusion in a chronic canine model. Circulation tutively active and muscarinic receptor-activated
heart disease: epicardial breakthrough. Circula- 123:137–146, 2011. IK,ACh channels in patients with chronic atrial
tion 122:1674–1682, 2010. 32. Mohler PJ, Schott JJ, Gramolini AO, et al: fibrillation. Cardiovasc Res 74:426–437, 2007.
13. MacLennan DH, Chen SR: Store overload- Ankyrin-B mutation causes type 4 long-QT 51. Yue L, Melnyk P, Gaspo R, et al: Molecular mech-
induced Ca2+ release as a triggering mechanism cardiac arrhythmia and sudden cardiac death. anisms underlying ionic remodeling in a dog
for CPVT and MH episodes caused by mutations Nature 421:634–639, 2003. model of atrial fibrillation. Circ Res 84:776–784,
in RYR and CASQ genes. J Physiol 587(Pt 33. Cunha SR, Hund TJ, Hashemi S, et al: Defects 1999.
13):3113–3115, 2009. in ankyrin-based membrane protein targeting 52. Qi XY, Yeh YH, Xiao L, et al: Cellular signaling
14. Voigt N, Li N, Wang Q, et al: Enhanced sarco- pathways underlie atrial fibrillation. Circulation underlying atrial tachycardia remodeling of
plasmic reticulum Ca2+ leak and increased Na+- 124:1212–1222, 2011. L-type calcium current. Circ Res 103:845–854,
Ca2+ exchanger function underlie delayed 34. Nyberg MT, Stoevring B, Behr ER, et al: The 2008.
afterdepolarizations in patients with chronic atrial variation of the sarcolipin gene (SLN) in atrial 53. Sun H, Chartier D, Leblanc N, et al: Intracellular
fibrillation. Circulation 125:2059–2070, 2012. fibrillation, long QT syndrome and sudden calcium changes and tachycardia-induced con-
15. Hove-Madsen L, Llach A, Bayes-Genis A, et al: arrhythmic death syndrome. Clin Chim Acta tractile dysfunction in canine atrial myocytes.
Atrial fibrillation is associated with increased 375:87–91, 2007. Cardiovasc Res 49:751–761, 2001.
spontaneous calcium release from the sarcoplas- 35. Ogrodnik J, Niggli E: Increased Ca2+ leak and 54. Bosch RF, Scherer CR, Rub N, et al: Molecular
mic reticulum in human atrial myocytes. Circula- spatiotemporal coherence of Ca2+ release in car- mechanisms of early electrical remodeling: tran-
tion 110:1358–1363, 2004. diomyocytes during beta-adrenergic stimulation. scriptional downregulation of ion channel sub-
16. Dobrev D: Electrical remodeling in atrial fibrilla- J Physiol 588(Pt 1):225–242, 2010. units reduces I(Ca,L) and I(to) in rapid atrial
tion. Herz 31:108–112, 2006. 36. Tan AY, Zhou S, Ogawa M, et al: Neural mecha- pacing in rabbits. J Am Coll Cardiol 41:858–869,
17. Vest JA, Wehrens XH, Reiken SR, et al: Defective nisms of paroxysmal atrial fibrillation and parox- 2003.
cardiac ryanodine receptor regulation during ysmal atrial tachycardia in ambulatory canines. 55. Gaborit N, Steenman M, Lamirault G, et al:
atrial fibrillation. Circulation 111:2025–2032, Circulation 118:916–925, 2008. Human atrial ion channel and transporter subunit
2005. 37. Choi EK, Shen MJ, Han S, et al: Intrinsic cardiac gene-expression remodeling associated with val-
18. Liang X, Xie H, Zhu PH, et al: Ryanodine nerve activity and paroxysmal atrial tachyarrhyth- vular heart disease and atrial fibrillation. Circula-
receptor-mediated Ca2+ events in atrial myocytes mia in ambulatory dogs. Circulation 121:2615– tion 112:471–481, 2005.
of patients with atrial fibrillation. Cardiology 2623, 2010. 56. Greiser M, Halaszovich CR, Frechen D, et al:
111:102–110, 2008. 38. Burashnikov A, Antzelevitch C: Reinduction of Pharmacological evidence for altered src kinase
19. Neef S, Dybkova N, Sossalla S, et al: CaMKII- atrial fibrillation immediately after termination of regulation of I (Ca,L) in patients with chronic
dependent diastolic SR Ca2+ leak and elevated the arrhythmia is mediated by late phase 3 early atrial fibrillation. Naunyn Schmiedebergs Arch
diastolic Ca2+ levels in right atrial myocardium of afterdepolarization-induced triggered activity. Pharmacol 375:383–392, 2007.
patients with atrial fibrillation. Circ Res Circulation 107:2355–2360, 2003. 57. Carnes CA, Janssen PM, Ruehr ML, et al: Atrial
106:1134–1144, 2010. 39. Patterson E, Po SS, Scherlag BJ, et al: Triggered glutathione content, calcium current, and con-
20. Chelu MG, Sarma S, Sood S, et al: Calmodulin firing in pulmonary veins initiated by in vitro tractility. J Biol Chem 282:28063–28073, 2007.
kinase II-mediated sarcoplasmic reticulum Ca2+ autonomic nerve stimulation. Heart Rhythm 58. Wang Z, Lu Y, Yang B: MicroRNAs and atrial
leak promotes atrial fibrillation in mice. J Clin 2:624–631, 2005. fibrillation: new fundamentals. Cardiovasc Res
Invest 119:1940–1951, 2009. 40. Chou CC, Nguyen BL, Tan AY, et al: Intracellular 89:710–721, 2011.
21. Greiser M, Neuberger HR, Harks E, et al: Dis- calcium dynamics and acetylcholine-induced trig- 59. Lu Y, Zhang Y, Wang N, et al: MicroRNA-328
tinct contractile and molecular differences gered activity in the pulmonary veins of dogs with contributes to adverse electrical remodeling in
atrial fibrillation. Circulation 122:2378–2387, a mouse knockout model. J Am Coll Cardiol fibrillation in a Chinese GeneID population.
2010. 37:2136–2143, 2001. Biochem Biophys Res Commun 398:98–104,
60. Levy S, Beharier O, Etzion Y, et al: Molecular
basis for zinc transporter 1 action as an endoge-
nous inhibitor of L-type calcium channels. J Biol
79. Nattel S, Maguy A, Le Bouter S, et al: Arrhyth-
mogenic ion-channel remodeling in the heart:
Heart failure, myocardial infarction, and atrial
100.
2010.
Lamirault G, Gaborit N, Le Meur N, et al: Gene
expression profile associated with chronic atrial
45
Chem 284:32434–32443, 2009. fibrillation. Physiol Rev 87:425–456, 2007. fibrillation and underlying valvular heart disease
61. Antzelevitch C, Pollevick GD, Cordeiro JM, et al: 80. Dhein S, Hagen A, Jozwiak J, et al: Improving in man. J Mol Cell Cardiol 40:173–184, 2006.
Loss-of-function mutations in the cardiac calcium cardiac gap junction communication as a new 101. Li D, Fareh S, Leung TK, Nattel S: Promotion
channel underlie a new clinical entity character- antiarrhythmic mechanism: The action of antiar- of atrial fibrillation by heart failure in dogs: Atrial
ized by ST-segment elevation, short QT intervals, rhythmic peptides. Naunyn Schmiedebergs Arch remodeling of a different sort. Circulation
and sudden cardiac death. Circulation 115:442– Pharmacol 381:221–234, 2010. 100:87–95, 1999.
449, 2007. 81. Nishida K, Michael G, Dobrev D, et al: Animal 102. Lin CS, Pan CH: Regulatory mechanisms of atrial
62. Pandit SV, Berenfeld O, Anumonwo JM, et al: models for atrial fibrillation: Clinical insights and fibrotic remodeling in atrial fibrillation. Cell Mol
Ionic determinants of functional reentry in a 2-D scientific opportunities. Europace 12:160–172, Life Sci 65:1489–1508, 2008.
model of human atrial cells during simulated 2010. 103. Ehrlich JR, Hohnloser SH, Nattel S: Role of
chronic atrial fibrillation. Biophys J 88:3806– 82. Firouzi M, Ramanna H, Kok B, et al: Association angiotensin system and effects of its inhibition in
3821, 2005. of human connexin40 gene polymorphisms with atrial fibrillation: Clinical and experimental evi-
63. Voigt N, Trausch A, Knaut M, et al: Left-to-right atrial vulnerability as a risk factor for idiopathic dence. Eur Heart J 27:512–518, 2006.
atrial inward rectifier potassium current gradients atrial fibrillation. Circ Res 95:e29–e33, 2004. 104. He X, Gao X, Peng L, et al: Atrial fibrillation
in patients with paroxysmal versus chronic atrial 83. Gollob MH, Jones DL, Krahn AD, et al: Somatic induces myocardial fibrosis through angiotensin
fibrillation. Circ Arrhythm Electrophysiol 3:472– mutations in the connexin 40 gene (GJA5) in II type 1 receptor-specific Arkadia-mediated
480, 2010. atrial fibrillation. N Engl J Med 354:2677–2688, downregulation of Smad7. Circ Res 108(2):164–
64. Voigt N, Maguy A, Yeh YH, et al: Changes in 2006. 175, 2011.
IK,ACh single-channel activity with atrial tachycar- 84. Juang JM, Chern YR, Tsai CT, et al: The associa- 105. Koinuma D, Shinozaki M, Komuro A, et al:
dia remodelling in canine atrial cardiomyocytes. tion of human connexin 40 genetic polymor- Arkadia amplifies TGF-beta superfamily signal-
Cardiovasc Res 77:35–43, 2008. phisms with atrial fibrillation. Int J Cardiol ling through degradation of Smad7. EMBO J
65. Girmatsion Z, Biliczki P, Bonauer A, et al: 116:107–112, 2007. 22:6458–6470, 2003.
Changes in microRNA-1 expression and IK1 85. Burstein B, Comtois P, Michael G, et al: Changes 106. Hunyady L, Catt KJ: Pleiotropic AT1 receptor
up-regulation in human atrial fibrillation. Heart in connexin expression and the atrial fibrillation signaling pathways mediating physiological and
Rhythm 6:1802–1809, 2009. substrate in congestive heart failure. Circ Res pathogenic actions of angiotensin II. Mol Endo-
66. Luo X, Pan Z, Xiao J, et al: Critical role of 105:1213–1222, 2009. crinol 20:953–970, 2006.
microRNAs miR-26 and miR-101 in atrial electri- 86. Bikou O, Thomas D, Trappe K, et al: Connexin 107. Burstein B, Libby E, Calderone A, et al: Differ-
cal remodeling in experimental atrial fibrillation. 43 gene therapy prevents persistent atrial fibrilla- ential behaviors of atrial versus ventricular fibro-
Circulation 122(Suppl):A19435, 2010. tion in a porcine model. Cardiovasc Res 92:218– blasts: A potential role for platelet-derived growth
67. Karle CA, Zitron E, Zhang W, et al: Human 225, 2011. factor in atrial-ventricular remodeling differ-
cardiac inwardly-rectifying K+ channel Kir(2.1b) 87. Igarashi T, Finet JE, Takeuchi A, et al: Connexin ences. Circulation 117:1630–1641, 2008.
is inhibited by direct protein kinase C-dependent gene transfer preserves conduction velocity and 108. Verheule S, Sato T, Everett TT, et al: Increased
regulation in human isolated cardiomyocytes and prevents atrial fibrillation. Circulation 125:216– vulnerability to atrial fibrillation in transgenic
in an expression system. Circulation 106:1493– 225, 2012. mice with selective atrial fibrosis caused by over-
1499, 2002. 88. Sossalla S, Kallmeyer B, Wagner S, et al: Altered expression of TGF-beta1. Circ Res 94:1458–
68. Makary S, Voigt N, Maguy A, et al: Differential Na+ currents in atrial fibrillation effects of rano- 1465, 2004.
protein kinase C isoform regulation and increased lazine on arrhythmias and contractility in human 109. Attisano L, Wrana JL: Signal transduction by the
constitutive activity of acetylcholine-regulated atrial myocardium. J Am Coll Cardiol 55:2330– TGF-beta superfamily. Science 296:1646–1647,
potassium channels in atrial remodeling. Circ Res 2342, 2010. 2002.
109:1031–1043, 2011. 89. Dhein S: Cardiac ischemia and uncoupling: gap 110. Burstein B, Qi XY, Yeh YH, et al: Atrial cardio-
69. Voigt N, Makary S, Nattel S, et al: Voltage- junctions in ischemia and infarction. Adv Cardiol myocyte tachycardia alters cardiac fibroblast func-
clamp-based methods for the detection of consti- 42:198–212, 2006. tion: A novel consideration in atrial remodeling.
tutively active acetylcholine-gated IK,ACh channels 90. Hagendorff A, Schumacher B, Kirchhoff S, et al: Cardiovasc Res 76:442–452, 2007.
in the diseased heart. Methods Enzymol 484:653– Conduction disturbances and increased atrial vul- 111. Tsai CT, Tseng CD, Hwang JJ, et al: Tachycardia
675, 2010. nerability in connexin40-deficient mice analyzed of atrial myocytes induces collagen expression in
70. Chen YH, Xu SJ, Bendahhou S, et al: KCNQ1 by transesophageal stimulation. Circulation atrial fibroblasts through transforming growth
gain-of-function mutation in familial atrial fibril- 99:1508–1515, 1999. factor β1. Cardiovasc Res 89(4):805–815, 2011.
lation. Science 299:251–254, 2003. 91. Wirka RC, Gore S, Van Wagoner DR, et al: A 112. Ponten A, Folestad EB, Pietras K, et al: Platelet-
71. Hong K, Bjerregaard P, Gussak I, et al: Short QT common connexin-40 gene promoter variant derived growth factor D induces cardiac fibrosis
syndrome and atrial fibrillation caused by muta- affects connexin-40 expression in human atria and and proliferation of vascular smooth muscle cells
tion in KCNH2. J Cardiovasc Electrophysiol is associated with atrial fibrillation. Circ Arrhythm in heart-specific transgenic mice. Circ Res
16:394–396, 2005. Electrophysiol 4(1):87–93, 2011. 97:1036–1045, 2005.
72. Xia M, Jin Q, Bendahhou S, et al: A Kir2.1 gain- 92. Olson TM, Michels VV, Ballew JD, et al: Sodium 113. Gudbjartsson DF, Arnar DO, Helgadottir A, et al:
of-function mutation underlies familial atrial channel mutations and susceptibility to heart Variants conferring risk of atrial fibrillation on
fibrillation. Biochem Biophys Res Commun failure and atrial fibrillation. JAMA 293:447–454, chromosome 4q25. Nature 448:353–357, 2007.
332:1012–1019, 2005. 2005. 114. Cardin S, Libby E, Pelletier P, et al: Contrasting
73. Yang Y, Xia M, Jin Q, et al: Identification of a 93. Chen LY, Ballew JD, Herron KJ, et al: A common gene expression profiles in two canine models of
KCNE2 gain-of-function mutation in patients polymorphism in SCN5A is associated with lone atrial fibrillation. Circ Res 100:425–433, 2007.
with familial atrial fibrillation. Am J Hum Genet atrial fibrillation. Clin Pharmacol Ther 81:35–41, 115. Adam O, Lavall D, Theobald K, et al: Rac1-
75:899–905, 2004. 2007. induced connective tissue growth factor regulates
74. Fatini C, Sticchi E, Genuardi M, et al: Analysis of 94. Ellinor PT, Nam EG, Shea MA, et al: Cardiac connexin 43 and N-cadherin expression in atrial
minK and eNOS genes as candidate loci for pre- sodium channel mutation in atrial fibrillation. fibrillation. J Am Coll Cardiol 55:469–480, 2010.
disposition to non-valvular atrial fibrillation. Eur Heart Rhythm 5:99–105, 2008. 116. Ko WC, Hong CY, Hou SM, et al: Elevated
Heart J 27:1712–1718, 2006. 95. Hedley PL, Jorgensen P, Schlamowitz S, et al: expression of connective tissue growth factor in
75. Elvan A, Rubart M, Zipes DP: NO modulates The genetic basis of Brugada syndrome: A muta- human atrial fibrillation and angiotensin II-treated
autonomic effects on sinus discharge rate and AV tion update. Hum Mutat 30:1256–1266, 2009. cardiomyocytes. Circ J 75:1592–1600, 2011.
nodal conduction in open-chest dogs. Am J 96. Francis J, Antzelevitch C: Atrial fibrillation and 117. van Rooij E, Sutherland LB, Thatcher JE, et al:
Physiol 272(1 Pt 2):H263–H271, 1997. Brugada syndrome. J Am Coll Cardiol 51:1149– Dysregulation of microRNAs after myocardial
76. Mery PF, Pavoine C, Belhassen L, et al: Nitric 1153, 2008. infarction reveals a role of miR-29 in cardiac
oxide regulates cardiac Ca2+ current: Involvement 97. Watanabe H, Darbar D, Kaiser DW, et al: Muta- fibrosis. Proc Natl Acad Sci U S A 105:13027–
of cGMP-inhibited and cGMP-stimulated phos- tions in sodium channel beta1- and beta2-subunits 13032, 2008.
phodiesterases through guanylyl cyclase activa- associated with atrial fibrillation. Circ Arrhythm 118. Dawson K, Wakili R, Ordog B, et al: Potential
tion. J Biol Chem 268:26286–26295, 1993. Electrophysiol 2:268–275, 2009. role of MicroRNA-29b in atrial fibrillation-
77. Kneller J, Zou R, Vigmond EJ, et al: Cholinergic 98. Olesen MS, Jespersen T, Nielsen JB, et al: Muta- promoting fibrotic remodeling. Circulation
atrial fibrillation in a computer model of a two- tions in sodium channel β-subunit SCN3B are 122:A12545, 2010.
dimensional sheet of canine atrial cells with associated with early-onset lone atrial fibrillation. 119. Duisters RF, Tijsen AJ, Schroen B, et al: miR-133
realistic ionic properties. Circ Res 90:E73–E87, Cardiovasc Res 89(4):786–793, 2011. and miR-30 regulate connective tissue growth
2002. 99. Wang P, Yang Q, Wu X, et al: Functional factor: Implications for a role of microRNAs in
78. Kovoor P, Wickman K, Maguire CT, et al: Evalu- dominant-negative mutation of sodium channel myocardial matrix remodeling. Circ Res 104:170–
ation of the role of IK,ACh in atrial fibrillation using subunit gene SCN3B associated with atrial 178, 2009.
120. Chen Y, Wakili R, Luo X, et al: MicroRNA 125. Bedi M, McNamara D, London B, et al: Genetic anism in hepatocellular carcinoma. Histopathology
changes and atrial arrhythmogenic remodeling in susceptibility to atrial fibrillation in patients with 52:552–559, 2008.
tachycardiomyopathic heart failure. Circulation congestive heart failure. Heart Rhythm 3:808– 131. Sun X, Frierson HF, Chen C, et al: Frequent
122(Suppl):A12988, 2010. 812, 2006. somatic mutations of the transcription factor
121. Thum T, Gross C, Fiedler J, et al: MicroRNA-21 126. Kato K, Oguri M, Hibino T, et al: Genetic factors ATBF1 in human prostate cancer. Nat Genet
contributes to myocardial disease by stimulating for lone atrial fibrillation. Int J Mol Med 19:933– 37:407–412, 2005.
MAP kinase signalling in fibroblasts. Nature 939, 2007. 132. Kim TS, Kawaguchi M, Suzuki M, et al: The
456:980–984, 2008. 127. Benjamin EJ, Rice KM, Arking DE, et al: Variants ZFHX3 (ATBF1) transcription factor induces
122. Patrick DM, Montgomery RL, Qi X, et al: Stress- in ZFHX3 are associated with atrial fibrillation in PDGFRB, which activates ATM in the cytoplasm
dependent cardiac remodeling occurs in the individuals of European ancestry. Nat Genet to protect cerebellar neurons from oxidative
absence of microRNA-21 in mice. J Clin Invest 41:879–881, 2009. stress. Dis Model Mech 3:752–762, 2010.
120:3912–3916, 2010. 128. Mommersteeg MT, Brown NA, Prall OW, et al: 133. Anne W, Duytschaever M: Upstream therapy in
123. Cardin S, Guasch E, Luo X, et al: A role for Pitx2c and Nkx2-5 are required for the formation atrial fibrillation: Traveling up the river to find the
microRNA in atrial profibrillatory fibrotic and identity of the pulmonary myocardium. Circ source. Heart Rhythm 7:529–530, 2010.
remodeling associated with experimental postin- Res 101:902–909, 2007. 134. Amit G, Qin H, Donahue JK: Biological therapies
farction heart failure. Circ Arrhythm Electro- 129. Mommersteeg MT, Hoogaars WM, Prall OW, et for atrial fibrillation. J Cardiovasc Pharmacol
physiol 5(5):1027–1035, 2012. al: Molecular pathway for the localized formation 52:222–227, 2008.
124. Fatini C, Sticchi E, Gensini F, et al: Lone and of the sinoatrial node. Circ Res 100:354–362, 135. Seok HY, Wang DZ: The emerging role of
secondary nonvalvular atrial fibrillation: Role of a 2007. microRNAs as a therapeutic target for cardiovas-
genetic susceptibility. Int J Cardiol 120:59–65, 130. Kim CJ, Song JH, Cho YG, et al: Down- cular disease. BioDrugs 24:147–155, 2010.
2007. regulation of ATBF1 is a major inactivating mech-
Myofibroblasts, Cytokines, and
Persistent Atrial Fibrillation 46
Kuljeet Kaur and José Jalife
excitability, and atrial impulse propagation; and their possible

CHAPTER OUTLINE
role in the mechanisms underlying the transition from paroxys-
Cardiac Fibroblasts and Myofibroblasts 459 mal to persistent AF, as well as stabilization of the latter. In this
chapter, we present a brief and necessarily incomplete review of
Inflammation and AF 460
possible mechanisms underlying AF-induced fibrosis. We pay
Effects of Other Cytokines on Ion Channels and AF 463 particular attention to the role of activated fibroblasts (myofibro-
blasts) and the cytokines released by them in the mechanisms of
Myofibroblasts, Cytokines, and Myocyte Electrical
electrical and structural remodeling. Our objective is twofold:
Remodeling 463 first, to go over our understanding of the molecular mechanisms
Concluding Remarks 465 of fibrosis; and second, to help predict how myofibroblasts
impinge on the electrical function of atrial myocytes by means of
signaling molecules, resulting in the electrophysiological remod-
eling that contributes to AF perpetuation.
Atrial fibrillation (AF) is the most common sustained cardiac
arrhythmia seen in clinical practice. It is the most important cause
of embolic stroke and is associated with increased morbidity and
mortality.1 Yet despite more than 100 years of basic and clinical Cardiac Fibroblasts and Myofibroblasts
research, we still do not fully understand its fundamental mecha-
nisms, and we have not learned how to treat it effectively. When Cardiac fibroblasts are critical players in normal cardiac func-
AF lasts continuously for more than 7 days, it is designated as tion.8 Although they occupy a small portion of the myocardial
persistent AF.2 Spontaneous, pharmacologic, or ablative resump- tissue volume, cardiac fibroblasts account for 50% to 70% of the
tion of sinus rhythm is infrequent in persistent AF, with cells of the normal adult mammalian heart,9 and even a greater
prompt recurrences or commonly failed cardioversions. Episodes proportion in pathologic conditions where cardiac fibrosis ensues
lasting longer than 1 year are termed “long-term persistent as a result of the transformation of fibroblasts to myofibroblasts,
AF.” As to the causes of persistent AF, it is reasonable to speculate which is a key event in connective tissue remodeling (Figure
that continuous high-frequency and irregular bombardment 46-1).8,10,11 Cardiac myofibroblast proliferation and migration
of the atrial cells and tissues with fibrillatory waves during with concomitant collagenous matrix accumulation leading to
long-lasting AF leads to a modification of the molecular substrate fibrosis develops during myocardial remodeling in ischemic,
on which waves propagate, with consequent electrical and struc- hypertensive, hypertrophic, dilated cardiomyopathies,12 arrhyth-
tural remodeling, substantial enough to increase the likelihood mogenic right ventricular cardiomyopathy (ARVC),13 and heart
of perpetuation of the electrical sources that maintain the failure. Similarly, myofibroblasts are active contributors to atrial
arrhythmia. fibrosis,8 which is part of the maladaptive atrial response to AF.14
Fibrosis has been implicated in the initiation and maintenance Myofibroblasts are phenotypically transformed (active) fibro-
of arrhythmia, affecting electrical propagation through slow, dis- blasts that contain α-smooth muscle actin and express integrins,
continuous conduction with “zigzag” propagation,3,4 reduced fibronectin, and other adhesion proteins (see Figure 46-1). They
regional coupling,5 abrupt changes in fibrotic bundle size,6 and were first identified by Gabbiani and Majno, who gave them the
micro-anatomical reentry.7 In the past, most clinical, experimen- name myofibroblasts15 after demonstrating their contractility using
tal, and numeric studies regarded fibrosis as electrically insulating chemical mediators of inflammation and strips of granulation
obstacles. However, heart injury promotes fibroblast differentia- tissue. These wound-healing fibroblast-like cells are involved in
tion into myofibroblasts,8,9 which are hypercontractile and hyper- each of the fibrous tissue responses present throughout the body.
secretory of soluble proteins termed cytokines, known to affect Yet their role in the mechanisms underlying structural and elec-
myocardial function, and have been shown to electrotonically trophysiological remodeling secondary to sustained atrial fibril-
couple with myocytes in vitro.10-14 Thus, it is likely that both lation has not been fully elucidated. Much of the available
fibroblasts and fibrosis play critical roles in atrial electroanatomi- information on the specific role played by myofibroblasts in atrial
cal remodeling and further facilitate perpetuation of AF. Yet AF fibrosis and AF perpetuation is derived from what is known about
per se may promote cardiac fibrosis, and it is notable that although hypertensive heart disease, in which impaired tissue compliance
numerous studies point to a clear association between AF and and symptomatic failure are linked to interstitial fibrosis and the
fibrosis, the underlying pathologic basis for AF-induced fibrosis expression of extracellular matrix genes, including type I and III
remains poorly understood. Arguably, early identification and fibrillar collagens and fibronectin. It has been shown that fibro-
prevention of the molecular and cellular mechanisms of the atrial blast proliferation and collagen deposition are significantly and
fibrotic process may halt the progression of AF and potentially heterogeneously increased in all atrial regions after prolonged
improve response to both pharmacologic and nonpharmacologic AF,16,17 and that the development of atrial fibrosis is highly
therapy. This highlights the importance of investigating the controlled by the renin-angiotensin-aldosterone system and by
biology of the fibroblasts as “sentinel cells”15; their role in the downstream activation of soluble proteins termed cytokines, par-
fibrotic process; the impact of their interactions with myocytes; ticularly transforming growth factor-beta1 (TGF-β1) and platelet-
their effects on sarcolemmal ion channel behavior, cardiac derived growth factor (PDGF).
459
ANG-II TGF-β
AT1R
Fibroblast
TGF-βR
Ang II Ang II
TGF-β1 TGF-β1
ανβ3 ανβ3
integrin integrin
r r
ED-A αSMA αSMA
ED-A
Fibronectin Adhesion Fibronectin Adhesion
proteins proteins
Ang II
TGF-β1
ανβ3
Myofibroblasts integrin
r
αSMA
ED-A
Fibronectin Adhesion
proteins
Figure 46-1. The phenotypic transformation (activation) of fibroblasts into myofibroblasts is highly controlled by the renin-angiotensin-aldosterone system and by down-
stream activation of soluble proteins termed cytokines, particularly transforming growth factor beta-1 (TGF-β1). Myofibroblasts proliferate and hypersecrete α-smooth muscle
actin (αSMA), which makes them hypercontractile and promotes their expression of integrins and other adhesion proteins, as well as extracellular matrix genes, including
fibronectin type I and type III fibrillar collagens.
model increased the activity of the profibrotic enzymes matrix

Inflammation and AF metalloproteinase (MMP)2 and MMP9 in the atrial interstitium,
increased MMP9 mRNA expression, and increased the activity
To fully understand the role of myofibroblasts and/or cytokines of tissue inhibitor of matrix metalloproteinase (TIMP)1 and
in fibrosis and AF perpetuation, it is necessary to fully understand TIMP3.21 Figure 46-2 illustrates a likely scenario linking sus-
the role played by inflammation, which is known to play an tained AF to inflammation with subsequent electrical remodel-
important pathogenic role leading to fibrosis in several cardio- ing, fibrosis, and atrial fibrillation perpetuation. Proinflammatory
vascular diseases.18 Increasing evidence now suggests that inflam- cytokines and hormones, such as Ang-II, tumor necrosis factor-α
matory pathways are directly linked to cellular and subcellular (TNF-α), interleukin (IL)-6, and IL-8, related to cardiovascular
mechanisms known to lead also to AF.18 Activated inflammatory disease and tissue injury, promote activation of leukocytes with
cells (such as neutrophils, lymphocytes, monocytes, and resident subsequent release of cytokines and ROS. As a result, connexin
macrophages), proinflammatory cytokines, and activated platelets and ion channel dysfunction occur, along with myocyte apoptosis
are important players in this picture (Figure 46-2). Moreover, and matrix generation and turnover, which leads to electrical and
inflammatory cascades alter ion channel function in myocytes and structural remodeling that predisposes an individual to AF.18
are strongly associated with fibrosis.18 Inflammatory stimuli such Clinical studies have identified inflammation as a pathogenic
as nicotine adenine dinucleotide phosphate (NADPH)-oxidase– factor in AF, as well as a correlation between inflammatory
derived reactive oxygen species (ROS), cytokines, growth factors, markers and the incidence of AF.22 In a 5-year follow-up of an
angiotensin-II (Ang-II), and other hormones, as well as mechani- original Framingham Heart Study cohort of about 800 patients
cal stretch, are well-known triggers of fibroblast proliferation, with no history of AF, subjects’ white blood cell count was sig-
migration, and differentiation into myofibroblasts, which are nificantly associated with AF development after correction for
critical players in the development of fibrosis.18 standard risk factors for AF.23 As a standard indicator of general
The concept that AF promotes inflammation was derived inflammation, white blood cell count is also an indicator of the
from studies demonstrating that chronic human AF is associated role of systemic inflammation as a risk factor for AF. Inflamma-
with increased atrial oxidative stress and peroxynitrite forma- tory markers appear to be good indicators of failure to terminate
tion.19 Thereafter, long-term rapid atrial pacing experiments in AF after cardioversion using both pharmacologic and electrical
dogs demonstrated pacing-induced shortening of the atrial effec- methods. Inflammatory markers such as serum amyloid A (SAA)
tive refractory period, which was associated with decreased tissue and C-reactive protein (CRP) were studied before and 3 weeks
ascorbate levels and increased protein nitration. The latter is a after cardioversion in patients with persistent AF24; 33% of
biomarker of peroxynitrite formation.20 Oral ascorbate supple- patients presented with AF recurrence at 3 weeks after cardiover-
mentation attenuated all changes. In a parallel study, supplemen- sion, and levels of CRP and SAA were associated with recur-
tal ascorbate was given to 43 patients before, and for 5 days after, rence.24 Patients with persistent AF had a higher inflammatory
cardiac bypass graft surgery.20 Patients receiving ascorbate had a index than patients with proximal AF, as measured by serum CRP
16.3% incidence of postoperative AF compared with 34.9% in levels, which suggested progressive inflammation in association
control subjects. More recently, rapid atrial pacing in a porcine with AF duration/severity.25 Patients with a history of AF have
Myofibroblasts, Cytokines, and Persistent Atrial Fibrillation 461
Hypertension Heart failure Tissue injury Atrial fibrillation molecule 1 (sICAM-1), and TGF-β1 are upregulated at baseline,32
46
and increased VEGF has been shown to associate independently
with AF occurrence.33 Levels of IL-18 are also higher in patients
with persistent AF with atrial dilatation.34 Graded increases in
Ang II, TNF-α, IL-6 Il-8, etc. TNF-α and N-terminal pro-brain natriuretic peptide (NTpBNP)
have been shown in various AF groups, and have been highest in
permanent AF.35 A cross-sectional analysis of the Heart and Soul
Study patients suggested a genetic link between IL-6 polymor-
phism 174CC and AF. Patients with this polymorphism have high
PMN Neutrophil
levels of IL-6.36
Increased serum levels of cytokines and other inflammatory
markers correlate also with failure of rhythm control in AF.37 For
example, it has been shown that expression of IL-2 is associated
Fibroblasts differentiation to myofibroblasts with the outcome of cardioversion with amiodarone, and serum
with proliferation, migration: hypersecretion levels of IL-2 were lower in patients who cardioverted success-
of TGF-β1 and other cytokines fully compared with those in noncardioverted patients. In addi-
tion, the baseline level of IL-2 was a strong indicator of therapeutic
success.37 In patients with paroxysmal or persistent AF, elevated
↑Activity of MMP2, MMP9, TIMP1 plasma levels of IL-6 and CRP appear to be independent predic-
and TIMP3; ↑MMP9 mRNA, ↑ECM tors of failure of pulmonary vein isolation.38 In another study,
generation; ↑ECM turnover; ion
baseline levels of CRP, TNF-α, sICAM-1, and oxidative stress
channel and Cx dysfunction
markers were elevated in patients with AF compared with con-
trols, and were higher in patients with than in those without
persistent AF recurrence; however, IL-6 levels were equally ele-
vated in the two subgroups.39 In yet another clinical study, serum
Electrical Fibrosis concentrations of IL-6, CRP, and CD40L decreased in patients
remodeling with no AF recurrence after ablation, whereas the same indicators
did not decrease in patients who failed ablation.40 An AF-related
release of cytokines was suggested by another study, in which
levels of CRP and IL-6 were similar in AF and sinus rhythm (SR)
Perpetuation of atrial fibrillation patients when in SR. In AF patients, single episodes of AF led to
these increased protein levels.41
Figure 46-2. Inflammatory pathways involved in the perpetuation of atrial fibrilla- Along with increases in the proinflammatory proteins, AF is
tion. Hypertension, heart failure, tissue injury, or sustained atrial fibrillation (AF) associated with decreases in anti-inflammatory mediators. Per-
leads to the release of proinflammatory cytokines and hormones, such as oxisome proliferator–activated receptor-γ (PPARγ), a multifunc-
angiotensin-II, tumor necrosis factor (TNF)-α, and interleukin-6 and interleukin-8, tional nuclear receptor protein, regulates a number of transcription
which promote activation of leukocytes with subsequent activation of fibroblasts
factors, and has been shown to be anti-inflammatory.42 mRNA
into myofibroblasts, which proliferate and become hypersecretory of a number of
cytokines. As a result, connexin and ion channel dysfunction occurs, along with and protein levels of PPARγ were shown to be decreased in
myocyte apoptosis and matrix generation and turnover, which lead to electrical hypertensive patients with AF compared with hypertensive
and structural remodeling (fibrosis) that contributes to the perpetuation of AF. patients with no AF and were further decreased in persistent
versus proximal AF patients. In addition, a negative correlation
was noted between levels of inflammatory cytokines and PPARγ.43
been shown to have increased inflammatory CD45+ cell infiltra-

tion in the atrial myocardium compared with patients with no TGF-β1 and AF
AF.26 Further support for the idea that inflammatory mediators
play a significant role in AF comes from studies demonstrating Members of the TGF-β family of pleiotropic cytokines are impli-
that hydrocortisone treatment before cardiac surgery reduced the cated in a wide variety of cell functions.44 They regulate inflam-
incidence of postoperative AF.27 mation, extracellular matrix deposition, and cell proliferation,
differentiation, and growth. In mammalian species, three struc-
turally similar isoforms of TGF-β (TGF-β1, TGF-β2, and TGF-
Cytokines and Human AF β3), encoded by three distinct genes, have been identified.45 The
most prevalent is TGF-β1. It expresses widely in many cell types,
The development of interstitial fibrosis and its impact on atrial whereas the other isoforms are found in a more limited range of
contractile and electrical function have been found to be central cells and tissues.44 Binding of TGF-β family ligands induces the
for AF maintenance.17,28,29 Fibroblast proliferation and collagen association of type II and type I receptors into a heterodimeric
deposition are significantly and heterogeneously increased in all complex (Figure 46-3). The type II receptor kinase phosphory-
atrial regions after prolonged AF,16,17 a process that is highly lates the type I receptor, inducing its serine/threonine kinase
controlled by the renin-angiotensin-aldosterone system, and by activity. Receptor-regulated Smads (R-Smads) are then activated
downstream activation of soluble cytokines such as TGF-β1 and by phosphorylation by the type I receptor kinase. Activated
PDGF. Originally termed cytokines by Stanley Cohen in 1974, R-Smads form complexes with common-partner Smads (Co-
these proteins are major players in immune reactions.30 However, Smad), and translocate into the nucleus.14 Once there, these
although immune cells are a major source of cytokines, a number proteins bind other transcriptional factors, including both tran-
of other cell types, including heart cells, can synthesize and scriptional coactivators and corepressors.
release them.31 Multiple studies have demonstrated close rela- Knowledge about the effects of TGF-β1 in advancing cardiac
tionships between elevated cytokines and AF. In patients with hypertrophy and fibrosis in vivo derives in part from studies in
paroxysmal or persistent AF, serum levels of vascular endothelial transgenic mice. TGF-β1–overexpressing mice demonstrated
growth factor (VEGF), IL-8, soluble intercellular adhesion significant cardiac hypertrophy accompanied by interstitial
Angiotensin II induces fibrosis through activation of the transforming growth

factor-β (TGF-β) family signaling pathway
TGF-β1 TGF-β1
I
TGF-β1 Type r
to
recep
e II Smad2-Smad3
Typ ptor P P P
e
Ang II rec
P
P
Phospho-Smad
Smad4
1R
P
AT
P
Smad2-3-4 complex
Cytoplasm
Co-activator(s)
P
P
Target genes
Smad7
Nucleus
Transcription factor(s) Procollagen
α-SMA
SnoN/c-Ski Arkadia
Cardiac fibroblast
Figure 46-3. Angiotensin-II (Ang-II) induces fibrosis through activation of the transforming growth factor (TGF)-β–signaling pathway in the myofibroblast. Activation of the
angiotensin type 1 receptor (AT-1R) increases the expression of TGF-β1, which leads to phosphorylation of type I and type II TGF-β1 membrane receptors. This leads to the
formation of a complex that activates Smad2-Smad3 through phosphorylation, promoting consolidation of the Smad2-3-4 complex with subsequent translocation to the
nucleus, leading to increased transcription of pro-fibrillatory genes. Ang-II/AT-1R-specific activation of Arkadia-mediated poly-ubiquitination and degradation of Smad7 are
thought to decrease the inhibitory feedback regulation of TGF-β1/Smad signaling and serve as key mechanisms for AF-induced atrial fibrosis.14
fibrosis.46 On the other hand, cardiac-restricted expression of a Angiotensin-II and TGF-β1

mutant TGF-β1 that enhances local TGF-β activity was associ-
ated with atrial but not ventricular fibrosis,47 which suggested Elevations in Ang-II and TGF-β1 levels are often found under
that the susceptibility of the atrial myocardium to the fibrogenic conditions leading to progression of heart failure. Ang-II enhances
actions of TGF-β is greater than that of the ventricular myocar- TGF-β1 expression via Smad pathway activation of the transcrip-
dium. Fibrotic remodeling of the atrium in this transgenic model tion factor activator protein 1 (AP-1), and this pathway is involved
was sufficient to increase vulnerability to AF.48 A single-nucleotide in hypertrophic growth of the heart muscle and in the develop-
polymorphism at codon 25, 915 G-> C, which is associated with ment of cardiac fibrosis.53 Ang-II can also lead to fibrosis by TGF-
myocardial infarction and atrial stiffness, is also associated with β1–independent Smad-mediated activation of Ang-II type (AT-1)
increased plasma TGF-β1 levels in patients with essential hyper- receptors and mitogen activated protein kinases (MAPKs).54 It has
tension and AF.49 been reported that Ang-II–induced left ventricular (LV) remodel-
Lower atrial voltage, especially in the left atrium, has been ing and fibrosis are dependent on both extracellular signal–
associated with AF. Plasma levels of TGF-β1 have been shown to regulated kinase (ERK) and Smad activation; inhibition of either
correlate strongly with left atrial voltage and volume, suggesting pathway is equally efficacious in restoring LV function and archi-
that TGF-β1 is an indicator of AF.50 In an interesting study in tecture.55 However, it is unknown whether ERK is also involved
which right atrial appendages were taken from SR and AF patients in Smad signaling in AF-induced atrial fibrosis.
who were divided into five groups (SR; paroxysmal/chronic per- From the foregoing, it is clear that Ang-II and TGF-β1
sistent AF of <6 months; and chronic permanent AF of 7 to 24 strongly contribute to the main causes of heart failure progres-
months, 25 to 60 months, and >60 months), a gradual increase sion. In experimental animal models56-58 and in clinical studies,59-
in fibrosis was noted as AF progressed.51 Similar to collagen, 61
inhibition of Ang-II by angiotensin-converting enzyme (ACE)
progressive increases in the expression of TGF-β1 mRNA, inhibitors or angiotensin receptor antagonists prevented heart
protein, and TGF-β receptor II protein were noted. However, failure progression and reduced mortality in patients. The effects
the TGF-β1 effect appeared to decline with increasing AF dura- of Ang-II are also evident in isolated cardiac cells. In fibroblasts,
tion, which suggested a biphasic change in TGF-β1–mediated Ang-II induces proliferation and collagen synthesis.62,63 In ven-
signaling in AF.51 tricular cardiomyocytes, Ang-II as an inducer of apoptosis pro-
In an isolated cell system, pacing of an HL-1 cell line for 24 motes hypertrophy in part via TGF-β1 signaling.53 The induction
hours increased TGF-β1 release and synthesis with increasing of TGF-β1 expression under Ang-II stimulation is mediated by
pacing frequency.52 The increase in TGF-β1 expression was asso- oxygen radicals, which are produced by NADPH (reduced form)
ciated with increased oxidative stress as measured by a fluorescent oxidase. Further downstream, p38 MAPK and AP-1 are acti-
dye, and was prevented at a lower pacing frequency with TGF-β1 vated.64 Rapid atrial pacing (RAP) in rabbits for 4 weeks increased
antibody, suggesting that TGF-β1 was increasing oxidative stress. fibrosis and collagen I and III content.14 The paced atria also
The same study indicated that increased myosin loss in paced showed increased levels of TGF-β1 protein and mRNA and an
cells was caused by TGF-β1.52 increased level of activated Smad 2/3, but the TGF-β inhibitory
protein, Smad7, was decreased.14 All of these changes were pre- Another protein of interest is the intracellular inflammatory
46
vented by losartan treatment. In addition, treatment with Ang-II cytokine high-mobility group box 1 (HMGB1), which has been
did not increase the level of collagen in cells overexpressing associated with various cardiac insults.75 HMGB1 promotes
Smad7,14 suggesting that strong feedback is provided between the inflammation when released via active secretion from necrotic
Ang-II pathway and the TGF-β1 pathway (see Figure 46-3), and cells or by activation of the NF-κB pathway. HMGB1 treatment
demonstrating the significance of increased TGF-β1–mediated in Cos cells transfected with Kv4.2 or Kv4.3 cDNA decreased
pathways in fibrosis associated with rapid atrial pacing. the respective transmembrane current.76 In neonatal rat ventricu-
Ang-II participates in the development of AF-induced myo- lar myocytes, HMGB1 treatment for 24 hours decreased Ito, as
cardial fibrosis through activation of AT-1 and AT-2 receptors.65-67 well as the sustained component of the transient outward current,
AT-1 receptor antagonism significantly attenuates AF-related but it increased cell capacitance, and therefore cell size. This
fibrosis in dogs.68 In a rabbit model of atrial tachypacing, treat- protein decreased mRNA and protein levels of both Kv4.2 and
ment with losartan decreased the deposition of cardiac collagens Kv4.3 without affecting KChIP2 expression. HMGB1 modestly
in a dose-dependent manner, suggesting that activation of AT-1 inhibited L-type Ca2+ current, but not IK1.76
receptors may be an important mechanism for AF-induced atrial
fibrosis.14
Leukemia Inhibitory Factor (LIF)
In HEK-293 cells transfected with the α-subunit of the T-type
Effects of Other Cytokines on Ion calcium channel, 12-hour treatment with LIF caused a significant
Channels and AF increase in the functional expression of channels as indicated by
changes in current density.77 LIF increased phosphorylation of
Macrophage Migration Inhibitory Factor (MIF) STAT3 and ERK in these cells. Treatment with the ERK inhibi-
tor U0126 inhibited LIF-induced increased calcium current.77
MIF is another multifaceted protein that can regulate the syn-
thesis and release of other cytokines such as TNF-α, IL-1, and
interferons.69 These cytokines are found in the sera of patients
with AF. In a recent study, atrial samples from AF patients showed Myofibroblasts, Cytokines, and Myocyte
increased levels of MIF expression, and cells isolated from the Electrical Remodeling
right atrial appendage showed a significantly reduced peak Ca2+
current compared with sinus rhythm patients. mRNA and protein The significance of investigating the interactions between myofi-
levels of the pore-forming α1C subunit of L-type Ca2+ channels broblasts and myocytes from normal and diseased hearts is high-
were also decreased.70 The MIF-induced decrease in Ca2+ current lighted by the fact that myofibroblasts are the primary mediators
was confirmed in HL-1 cells. Phosphorylation of sarcoma tyro- of wound healing in the injured heart and are the dominant cell
sine kinase (Src) protein was increased by MIF treatment in the type in the infarcted scar.14,24,25,78-80 They also contribute to fibrosis
atrial cells. Yet treatment with PP1 (a specific Src antagonist) in the atria and ventricles through their ability to produce fibrillar
antagonized MIF-mediated effects on both Ca2+ current and and nonfibrillar collagens,80,81 and to ECM remodeling through
expression of L-type calcium channel (LCC) protein.70 The their production of focal adhesion-associated proteins (see Figure
results supported the involvement of MIF in the electrical remod- 46-1).79,82,83 The importance of these studies is enhanced by the
eling that accompanies AF, through impairment of L-type Ca2+ knowledge that paracrine factors released from myofibroblasts
channel expression and function via activation of c-Src kinases in lead to myocyte hypertrophy and diastolic dysfunction.8,84 The
atrial myocytes. best studied factors include TGF-β, fibroblast growth factor
(FGF)-2, and members of the IL-6 family of proteins.8,84,85
Despite the importance of myofibroblasts and fibrosis in myo-
Effects of TNF-α on Ionic Currents cardial infarction, cardiomyopathies, heart failure, and AF,
whether and how the factors previously described affect cardiac
When mice are treated with TNF-α for 6 weeks, their ventricular myocyte electrical function or excitation-contraction coupling
myocytes exhibit reduced transient outward potassium currents, has not been adequately addressed. Clearly, there is a need to
Ito and IKur.71 In addition, TNF-α increases the action potential examine electrophysiological consequences of paracrine, mechan-
duration (APD).72 The main downstream signaling molecule for ical, and electrical myocyte-myofibroblast interactions for
TNF-α, nuclear factor kappa light-chain enhancer of activated B myocyte electrophysiology, which is underscored by the recent
cells (NF-κB), has also been implicated in electrophysiological discovery that attachment of myofibroblasts to adult cardiac myo-
remodeling in cardiac cells.73 Overexpression of the NF-κB cytes in vitro results in structural remodeling of the myocytes.86,87
activator IκB kinase-β decreased KChIP2 expression and Ito. In Whether such maladaptive behavior results in electrical remodel-
addition, the classic NF-κB activator TNF-α induced NF-κB– ing has never been investigated. Finally, although dye transfer
dependent reduction of both KChIP2 and Ito. A role for NF-κB– experiments have suggested that intercellular coupling exists
mediated ionic current decrease was further supported by the fact between co-cultured adult rabbit ventricular myocytes and myo-
that a phosphorylation mutant of IκB kinase was able to inhibit fibroblasts,86,88 to our knowledge, no such evidence exists in vivo
both NF-κB and TNF-α–mediated decreases in Ito and KChIP2 in the atria or the ventricles of the adult heart. In addition, no
expression.74 dual patch clamping experiments have been conducted to rigor-
In H9c2 cells, Ang-II reduced the peak sodium current by ously quantify the conductance of junctional channels between
interfering with SCN5A transcription, which was mediated by an adult myocytes and myofibroblasts.
increase in oxidative stress. The increased oxidative stress aug- Early reports on the consequences of culturing neonatal
mented binding of NF-κB to the SCN5A promoter with conse- myocytes in cardiac fibroblast conditioned medium (FCM)
quent decreases in both SCN5A transcription and sodium inward showed increased protein expression,89 action potential duration
current.73 Altogether, these studies suggested that by activating (APD) prolongation, and KV4.2 downregulation in the myo-
NF-κB, TNF-α and other cytokines can substantially regulate cytes.90 Pedrotty et al91 showed that exposure of neonatal rat
both outward potassium and inward sodium currents in cardiac ventricular myocyte (NRVM) monolayers to FCM produced a
myocytes. dose-dependent reduction in conduction velocity, prolongation
of APD, depolarization of the resting membrane potential, and were allowed to condition fresh serum-free M199 medium for 24
a reduction in action potential upstroke velocity. Expression of hours. The resulting FCM or myocyte-conditioned medium
fibroblast proliferation, myocyte apoptosis, and Cx43 expression (MCM) was filtered and then was added to dishes containing
was not affected. However, mRNA levels of NaV1.5, Kir2.1, and freshly dissociated adult rat ventricular myocytes.
KV4.3 were reduced by exposure to the FCM.91 More recently, We determined the electrophysiological consequences of
Vazquez et al92 showed that in NRVM monolayers treated with exposing adult myocytes to control medium, FCM, or MCM for
FCM harvested from infarcted hearts, conduction velocity could 2 days. Figure 46-4, A-C shows data from current clamp experi-
be higher or lower than in NRVM treated with FCM from ments obtained during superfusion of the cultured myocytes with
normal hearts, depending on cell density. In addition, the optical Tyrode’s solution (1 mL/min) at 37° C. As shown in panel A,
APD70 was slightly shorter in the former than in the latter.92 incubation of myocytes in FCM for 48 hours resulted in signifi-
Although these effects of neonatal FCM on NRVMs are of inter- cant abbreviation of the APD at 90% repolarization (APD90) at
est, they should not be extrapolated to the adult heart, particu- basic cycle lengths (BCLs) between 200 and 2000 ms. In contrast,
larly when NRVMs are derived from the myocardial infarction incubation of the myocytes with MCM for 48 hours produced
scar. In fact, evidence in the literature suggests that the pheno- no significant changes in APD. It is interesting to note that as
typic changes produced by paracrine factors released by cardiac shown in panel B, incubation of myocytes in FCM for 48 hours
fibroblasts may be different in the developing versus the adult resulted in significant increases in maximal action potential
myocyte.93 In addition, although the use of FCM is a matter of upstroke velocity (dV/dtmax) at BCLs between 300 and 2000 ms,
importance, it leaves the uncertainty of which specific cytokines whereas the effects of MCM were not significantly different from
are producing the electrophysiological changes. Finally, despite those of controls. As shown in panel C, the increased dV/dtmax
the potential importance of both electrical and structural remod- induced by FCM likely reflected increased availability of the
eling in atrial fibrillation, a role has not been established for sodium current (INa) secondary to concomitant hyperpolarization
specific regulation by cytokines released from atrial fibroblasts in of the resting membrane potential (RMP).
atrial myocyte electrical function and/or excitation-contraction Finally, panel D summarizes data from voltage clamp experi-
coupling. Therefore, there is a need to investigate in adult atria ments comparing Ba2+ (1 mM)-sensitive current/voltage (IV)
myocytes the electrophysiological effects of specific cytokines relations obtained in control (n = 7), 48-hour FCM (n = 3), and
that are known to be released by myofibroblasts and upregulated 48-hour MCM (n = 3). The significant increases in both peak
in the fibrillating atria. inward (−100 mV) and peak outward (−50 and −60 mV) IK1
We recently began to explore the effects of FCM on cardiac current (compared with both MCM and control) provide a direct
electrical properties of adult ventricular myocytes. Myocytes and explanation for the observed effects of FCM on RMP, APD90,
fibroblasts were obtained from the hearts of adult rats and were and dV/dtmax. Whether such changes are the result of ion
cultured as described elsewhere.94 Briefly, fibroblasts were cul- channel gene expression or are induced by specific cytokines
tured in maintenance media until they became confluent. Then contained in the FCM, or whether they result from direct effects
fibroblasts were split, passaged repeatedly at a 1 : 2 ratio, and used of any given cytokine on Kir2.x channel properties, will need to
at passages 3 to 6.94-96 After the fibroblast or myocyte cultures be addressed in future experiments. Nevertheless, in an initial
were washed thoroughly with phosphate buffered saline, they attempt to address these questions, we have conducted additional
150 500
400
dV/dt (mV/ms)
100
APD90 (ms)
300
* *
* *
50 * * *
200
*
100
0 0
0 500 1000 1500 2000 2500 0 500 1000 1500 2000 2500
A BCL (ms) B BCL (ms)
0
mV **
–100 –50 50
–20
–5
RMP (mV)
–40
–60 –10
–80
–100
*** –15
C Control MCM FCM D pA/pF

Figure 46-4. Electrophysiological effects of fibroblast-conditioned medium (FCM; closed squares; n = 8), myocyte-conditioned medium (MCM; open triangles; n = 4), and
control medium (open circles; n = 5) for 48 hours. A, Basic cycle length (BCL) dependence of APD at 90% repolarization (APD90). B, BCL dependence of maximal action
potential upstroke velocity (dV/dtmax). C, Resting membrane potential (RMP). D, IV relation shows significant increases (asterisks; P < .05) in inward and outward IK1 for FCM
(black squares; n = 3) but not MCM (open triangles; n = 3).
46
Control TGF-β1 Control FGF-2
0
A B
200 150
* *
APD90 (ms)
APD90 (ms)
** * 100
100
50
0 0
500 1000 1500 2000 200 300 400 500
C BCL (ms) D BCL (ms)
Figure 46-5. Effects of (A and B) transforming growth factor (TGF)-β1 and (C and D) fibroblast growth factor (FGF)-2 on APD90 of cultured adult rat ventricular myocytes
at 48 hours. A, top, Representative action potentials in control and TGF-β1 (50 ng/mL) at BCL = 500 ms. B, APD90 vs. BCL in control (open circles; n = 8) and TGF-β1 (black
squares; n = 10). C, Representative APs in control and FGF-2 (50 ng/mL). D, APD90 vs. BCL in control (open circles; n = 5) and FGF-2 (black squares; n = 6). Calibrations in A
and C, horizontal, 200 ms; vertical, 20 mV. *P < .05; **P < .01.
current clamp experiments with two different cytokines that are of information about whether they are involved in electrical
known to be involved in myocyte-fibroblast interactions. In remodeling in the adult heart, it seems important to examine
Figure 46-5, panels A and B show APD90 data obtained from paracrine effects of myofibroblast-derived cytokines on the action
myocytes cultured in TGF-β1 (50 ng/mL) and FGF-2 (100 ng/ potential of atrial myocytes, as well as the ionic and molecular
mL), respectively, for 48 hours. Concentrations of the cytokines mechanisms. In vitro experiments should address questions
were high but were near the ranges of those used and docu- related to the effects of conditioned media derived from atrial
mented in the literature.97 Clearly, although TGF-β1 prolongs myofibroblasts (i.e., activated fibroblasts expressing α-smooth
APD90 at BCLs between 200 and 500 ms, FGF-2 produces sig- muscle actin) obtained from normal and chronically fibrillating
nificant APD abbreviation at all BCLs. In these initial experi- hearts. Thereafter, research should focus on the effects of specific
ments, we observed no significant changes in RMP or dV/dtmax cytokines and their signaling pathways that have been shown to
for TGF-β1 or FGF-2. Therefore, although the combined effects be involved in atrial remodeling (e.g., TGF-β1, PDGF). The use
of FGF-2 and TGF-β on APD90 provide a partial explanation for of clinically relevant large animal models of persistent atrial
our observations in the presence of FCM (see Figure 46-4), it is fibrillation that are capable of recapitulating the features of sus-
clear that the effects of FCM on action potential characteristics tained AF in patients, including TGF-β1 elevation, the presence
are complex, and that additional factors are involved. Further, it of interstitial fibrosis, the upregulation of a major inward-rectifier
would be important to investigate the ionic and molecular mech- current, and the frequency characteristics and spatial distribution
anisms underlying the disparate effects of these and other impor- of the fibrillatory waves will be essential in this endeavor. Clearly,
tant cytokines on myocyte electrophysiology in both atria and although both pharmacologic and ablative treatments of persis-
ventricles, as well as the mechanisms of electrical remodeling tent AF have shown some efficacy, clinical needs are far from
associated with the transition from paroxysmal to persistent AF. being met, and the field is facing many challenges. Two such
challenges are the expected increase in the incidence of AF in the
general population, and the potentially adverse long-term conse-
quences of the extensive ablative therapy that is usually needed
Concluding Remarks to terminate persistent AF. Thus, novel antiarrhythmic strategies
targeting not only fibrosis, but also the molecular mechanisms
In view of the striking effects that myofibroblast paracrine factors underlying myocyte remodeling, are highly desirable. As such,
have on atrial and ventricular function, the evidence that cyto- finding new target genes with pleiotropic effects on both myo-
kines are significantly upregulated and may play important roles cytes and nonmyocyte cells important for pathologic remodeling
in atrial structural remodeling, and the relative dearth in AF may become an important goal of therapy.
Recommendations for patient selection, proce- 5. Spach MS, Boineau JP: Microfibrosis produces
References dural techniques, patient management and electrical load variations due to loss of side-to-side
follow-up, definitions, endpoints, and research trial cell connections: A major mechanism of structural
1. Kannel WB, Wolf PA, Benjamin EJ, et al: design. Europace 14:528–606, 2012. heart disease arrhythmias. Pacing Clin Electro-
Prevalence, incidence, prognosis, and predisposing 3. de Bakker JM, van Capelle FJ, Janse MJ, et al: physiol 20:397–413, 1997.
conditions for atrial fibrillation: Population- Slow conduction in the infarcted human heart: 6. de Bakker JM, Stein M, van Rijen HV: Three-
based estimates. Am J Cardiol 82:2N–9N, ‘Zigzag’ course of activation. Circulation 88:915– dimensional anatomic structure as substrate for
1998. 926, 1993. ventricular tachycardia/ventricular fibrillation.
2. Calkins H, Kuck KH, Cappato R, et al: 2012 4. Bian W, Tung L: Structure-related initiation of Heart Rhythm 2:777–779, 2005.
HRS/EHRA/ECAS expert consensus statement on reentry by rapid pacing in monolayers of cardiac 7. Valderrabano M, Kim YH, Yashima M, et al:
catheter and surgical ablation of atrial fibrillation: cells. Circ Res 98:e29–e38, 2006. Obstacle-induced transition from ventricular
fibrillation to tachycardia in isolated swine right 29. Boldt A, Wetzel U, Lauschke J, et al: Fibrosis in 50. Kim SK, Park JH, Kim JY, et al: High plasma
ventricles: Insights into the transition dynamics left atrial tissue of patients with atrial fibrillation concentrations of transforming growth factor-beta
and implications for the critical mass. J Am Coll with and without underlying mitral valve disease. and tissue inhibitor of metalloproteinase-1: Poten-
Cardiol 36:2000–2008, 2000. Heart 90:400–405, 2004. tial non-invasive predictors for electroanatomical
8. Souders CA, Bowers SL, Baudino TA: Cardiac 30. Cohen S: Cytokine: More than a new word, a new remodeling of atrium in patients with non-valvular
fibroblast: The renaissance cell. Circ Res 105:1164– concept proposed by stanley cohen thirty years atrial fibrillation. Circ J 75:557–564, 2011.
1176, 2009. ago. Cytokine 28:242–247, 2004. 51. Gramley F, Lorenzen J, Koellensperger E, et al:
9. Baudino TA, Carver W, Giles W, et al: Cardiac 31. Kaur K, Dhingra S, Slezak J, et al: Biology of Atrial fibrosis and atrial fibrillation: The role of the
fibroblasts: Friend or foe? Am J Physiol Heart Circ TNFalpha and IL-10, and their imbalance in heart TGF-beta1 signaling pathway. Int J Cardiol
Physiol 291:H1015–H1026, 2006. failure. Heart Fail Rev 14:113–123, 2009. 143:405–413, 2010.
10. Vasquez C, Benamer N, Morley GE: The cardiac 32. Scridon A, Morel E, Nonin-Babary E, et al: 52. Yeh YH, Kuo CT, Chan TH, et al: Transforming
fibroblast: Functional and electrophysiological Increased intracardiac vascular endothelial growth growth factor-beta and oxidative stress mediate
considerations in healthy and diseased hearts. J factor levels in patients with paroxysmal, but not tachycardia-induced cellular remodelling in cul-
Cardiovasc Pharmacol 57:380–388, 2011. persistent atrial fibrillation. Europace 14:948–953, tured atrial-derived myocytes. Cardiovasc Res
11. Baum J, Duffy HS: Fibroblasts and myofibroblasts: 2012. 91:62–70, 2011.
What are we talking about? J Cardiovasc Pharma- 33. Choudhury A, Freestone B, Patel J, et al: Relation- 53. Schroder D, Heger J, Piper HM, et al: Angiotensin
col 57:376–379, 2011. ship of soluble CD40 ligand to vascular endothelial II stimulates apoptosis via TGF-beta1 signaling in
12. Manabe I, Shindo T, Nagai R: Gene expression in growth factor, angiopoietins, and tissue factor in ventricular cardiomyocytes of rat. J Mol Med (Berl)
fibroblasts and fibrosis: Involvement in cardiac atrial fibrillation: A link among platelet activation, 84:975–983, 2006.
hypertrophy. Circ Res 91:1103–1113, 2002. angiogenesis, and thrombosis? Chest 132:1913– 54. Rodriguez-Vita J, Sanchez-Lopez E, Esteban V,
13. Brown RD, Ambler SK, Mitchell MD, et al: The 1919, 2007. et al: Angiotensin II activates the smad pathway in
cardiac fibroblast: Therapeutic target in myocar- 34. Luan Y, Guo Y, Li S, et al: Interleukin-18 among vascular smooth muscle cells by a transforming
dial remodeling and failure. Annu Rev Pharmacol atrial fibrillation patients in the absence of struc- growth factor-beta-independent mechanism. Cir-
Toxicol 45:657–687, 2005. tural heart disease. Europace 12:1713–1718, 2010. culation 111:2509–2517, 2005.
14. He X, Gao X, Peng L, et al: Atrial fibrillation 35. Li J, Solus J, Chen Q, et al: Role of inflammation 55. de Boer RA, Pokharel S, Flesch M, et al: Extracel-
induces myocardial fibrosis through angiotensin II and oxidative stress in atrial fibrillation. Heart lular signal regulated kinase and smad signaling
type 1 receptor-specific arkadia-mediated down- Rhythm 7:438–444, 2010. both mediate the angiotensin II driven progression
regulation of smad7. Circ Res 108:164–175, 2011. 36. Marcus GM, Whooley MA, Glidden DV, et al: towards overt heart failure in homozygous
15. Gabbiani G, Majno G: Dupuytren’s contracture: Interleukin-6 and atrial fibrillation in patients with tgr(mren2)27. J Mol Med (Berl) 82:678–687, 2004.
Fibroblast contraction? An ultrastructural study. coronary artery disease: Data from the heart and 56. Pfeffer JM, Pfeffer MA: Angiotensin converting
Am J Pathol 66:131–146, 1972. soul study. Am Heart J 155:303–309, 2008. enzyme inhibition and ventricular remodeling in
16. Burstein B, Qi XY, Yeh YH, et al: Atrial cardio- 37. Rizos I, Tsiodras S, Rigopoulos AG, et al: heart failure. Am J Med 84:37–44, 1988.
myocyte tachycardia alters cardiac fibroblast func- Interleukin-2 serum levels variations in recent 57. Pinto YM, Pinto-Sietsma SJ, Philipp T, et al:
tion: A novel consideration in atrial remodeling. onset atrial fibrillation are related with cardiover- Reduction in left ventricular messenger RNA for
Cardiovasc Res 76:442–452, 2007. sion outcome. Cytokine 40:157–164, 2007. transforming growth factor beta(1) attenuates left
17. Burstein B, Libby E, Calderone A, et al: Differen- 38. Henningsen KM, Nilsson B, Bruunsgaard H, et al: ventricular fibrosis and improves survival without
tial behaviors of atrial versus ventricular fibroblasts: Prognostic impact of hs-CRP and IL-6 in patients lowering blood pressure in the hypertensive
A potential role for platelet-derived growth factor undergoing radiofrequency catheter ablation for tgr(mren2)27 rat. Hypertension 36:747–754, 2000.
in atrial-ventricular remodeling differences. Circu- atrial fibrillation. Scand Cardiovasc J 43:285–291, 58. Kim S, Yoshiyama M, Izumi Y, et al: Effects of
lation 117:1630–1641, 2008. 2009. combination of ace inhibitor and angiotensin
18. Friedrichs K, Klinke A, Baldus S: Inflammatory 39. Leftheriotis DI, Fountoulaki KT, Flevari PG, et al: receptor blocker on cardiac remodeling, cardiac
pathways underlying atrial fibrillation. Trends Mol The predictive value of inflammatory and oxidative function, and survival in rat heart failure. Circula-
Med 17:556–563, 2011. markers following the successful cardioversion of tion 103:148–154, 2001.
19. Mihm MJ, Yu F, Carnes CA, et al: Impaired myo- persistent lone atrial fibrillation. Int J Cardiol 59. Effect of enalapril on survival in patients with
fibrillar energetics and oxidative injury during 135:361–369, 2009. reduced left ventricular ejection fractions and con-
human atrial fibrillation. Circulation 104:174–180, 40. Osmancik P, Peroutka Z, Budera P, et al: Changes gestive heart failure: The SOLVD investigators. N
2001. in cytokine concentrations following successful Engl J Med 325:293–302, 1991.
20. Carnes CA, Chung MK, Nakayama T, et al: Ascor- ablation of atrial fibrillation. Eur Cytokine Netw 60. Pfeffer MA, Braunwald E, Moye LA, et al: Effect
bate attenuates atrial pacing-induced peroxynitrite 21:278–284, 2010. of captopril on mortality and morbidity in patients
formation and electrical remodeling and decreases 41. Marcus GM, Smith LM, Ordovas K, et al: Intra- with left ventricular dysfunction after myocardial
the incidence of postoperative atrial fibrillation. cardiac and extracardiac markers of inflammation infarction: Results of the survival and ventricular
Circ Res 89:E32–E38, 2001. during atrial fibrillation. Heart Rhythm 7:149–154, enlargement trial. The SAVE investigators. N Engl
21. Chen CL, Huang SK, Lin JL, et al: Upregulation 2010. J Med 327:669–677, 1992.
of matrix metalloproteinase-9 and tissue inhibitors 42. Choi JM, Bothwell AL: The nuclear receptor ppars 61. Pitt B, Poole-Wilson PA, Segal R, et al: Effect of
of metalloproteinases in rapid atrial pacing-induced as important regulators of T-cell functions and losartan compared with captopril on mortality in
atrial fibrillation. J Mol Cell Cardiol 45:742–753, autoimmune diseases. Mol Cells 33:217–222, 2012. patients with symptomatic heart failure: Ran-
2008. 43. Chen X, Bing Z, He J, et al: Downregulation of domised trial—the Losartan Heart Failure Survival
22. Kaski JC, Arrebola-Moreno AL: [Inflammation peroxisome proliferator-activated receptor-gamma Study Elite II. Lancet 355:1582–1587, 2000.
and thrombosis in atrial fibrillation]. Rev Esp expression in hypertensive atrial fibrillation. Clin 62. Schorb W, Booz GW, Dostal DE, et al: Angioten-
Cardiol 64:551–553, 2011. Cardiol 32:337–345, 2009. sin II is mitogenic in neonatal rat cardiac fibro-
23. Rienstra M, Sun JX, Magnani JW, et al: White 44. Dobaczewski M, Chen W, Frangogiannis NG: blasts. Circ Res 72:1245–1254, 1993.
blood cell count and risk of incident atrial fibrilla- Transforming growth factor (TGF)-beta signaling 63. Bouzegrhane F, Thibault G: Is angiotensin II a
tion (from the Framingham heart study). Am J in cardiac remodeling. J Mol Cell Cardiol 51:600– proliferative factor of cardiac fibroblasts? Cardio-
Cardiol 109:533–537, 2012. 606, 2011. vasc Res 53:304–312, 2002.
24. Barassi A, Pezzilli R, Morselli-Labate AM, et al: 45. Schiller M, Javelaud D, Mauviel A: TGF-beta- 64. Wenzel S, Taimor G, Piper HM, et al: Redox-
Serum amyloid A and C-reactive protein indepen- induced smad signaling and gene regulation: con- sensitive intermediates mediate angiotensin
dently predict the recurrences of atrial fibrillation sequences for extracellular matrix remodeling and II-induced p38 map kinase activation, AP-1
after cardioversion in patients with preserved left wound healing. J Dermatol Sci 35:83–92, 2004. binding activity, and TGF-beta expression in adult
ventricular function. Can J Cardiol 28:537–541, 46. Rosenkranz S, Flesch M, Amann K, et al: Altera- ventricular cardiomyocytes. FASEB J 15:2291–
2012. tions of beta-adrenergic signaling and cardiac 2293, 2001.
25. Yao SY, Chu JM, Chen KP, et al: Inflammation in hypertrophy in transgenic mice overexpressing 65. Boldt A, Scholl A, Garbade J, et al: ACE-inhibitor
lone atrial fibrillation. Clin Cardiol 32:94–98, TGF-beta(1). Am J Physiol Heart Circ Physiol treatment attenuates atrial structural remodeling in
2009. 283:H1253–H1262, 2002. patients with lone chronic atrial fibrillation. Basic
26. Chen MC, Chang JP, Liu WH, et al: Increased 47. Nakajima H, Nakajima HO, Salcher O, et al: Atrial Res Cardiol 101:261–267, 2006.
inflammatory cell infiltration in the atrial myocar- but not ventricular fibrosis in mice expressing a 66. Hirayama Y, Atarashi H, Kobayashi Y, et al:
dium of patients with atrial fibrillation. Am J mutant transforming growth factor-beta(1) trans- Angiotensin-converting enzyme inhibitors are not
Cardiol 102:861–865, 2008. gene in the heart. Circ Res 86:571–579, 2000. effective at inhibiting further fibrous changes in the
27. Weis F, Beiras-Fernandez A, Schelling G, et al: 48. Verheule S, Sato T, Everett TT, et al: Increased atria in patients with chronic atrial fibrillation:
Stress doses of hydrocortisone in high-risk patients vulnerability to atrial fibrillation in transgenic mice Speculation from analysis of the time course of
undergoing cardiac surgery: Effects on interleukin-6 with selective atrial fibrosis caused by overexpres- fibrillary wave amplitudes. Jpn Heart J 45:93–101,
to interleukin-10 ratio and early outcome. Crit sion of TGF-beta1. Circ Res 94:1458–1465, 2004. 2004.
Care Med 37:1685–1690, 2009. 49. Wang Y, Hou X, Li Y: Association between trans- 67. Chrysostomakis SI, Karalis IK, Simantirakis EN,
28. Kostin S, Klein G, Szalay Z, et al: Structural cor- forming growth factor beta1 polymorphisms and et al: Angiotensin II type 1 receptor inhibition is
relate of atrial fibrillation in human patients. Car- atrial fibrillation in essential hypertensive subjects. associated with reduced tachyarrhythmia-induced
diovasc Res 54:361–379, 2002. J Biomed Sci 17:23, 2010. ventricular interstitial fibrosis in a goat atrial
fibrillation model. Cardiovasc Drugs Ther 21:357– 78. Dixon IM: The soluble interleukin 6 receptor rabbit ventricular myocytes and myofibroblasts. J
365, 2007. takes its place in the pantheon of interleukin 6 Physiol 583:225–236, 2007.
68. Nakashima H, Kumagai K: Reverse-remodeling
effects of angiotensin II type 1 receptor blocker in
a canine atrial fibrillation model. Circ J 71:1977–
signaling proteins: Phenoconversion of cardiac
fibroblasts to myofibroblasts. Hypertension
56:193–195, 2010.
89. Harada M, Itoh H, Nakagawa O, et al: Significance
of ventricular myocytes and nonmyocytes interac-
tion during cardiocyte hypertrophy: Evidence for
46
1982, 2007. 79. Santiago JJ, Dangerfield AL, Rattan SG, et al: endothelin-1 as a paracrine hypertrophic factor
69. Calandra T, Bernhagen J, Metz CN, et al: MIF as Cardiac fibroblast to myofibroblast differentiation from cardiac nonmyocytes. Circulation 96:3737–
a glucocorticoid-induced modulator of cytokine in vivo and in vitro: Expression of focal adhesion 3744, 1997.
production. Nature 377:68–71, 1995. components in neonatal and adult rat ventricular 90. Guo W, Kamiya K, Yasui K, et al: Paracrine hyper-
70. Rao F, Deng CY, Wu SL, et al: Involvement of src myofibroblasts. Dev Dyn 239:1573–1584, 2010. trophic factors from cardiac non-myocyte cells
in l-type Ca2+ channel depression induced by mac- 80. Dobaczewski M, Bujak M, Zymek P, et al: Extracel- downregulate the transient outward current
rophage migration inhibitory factor in atrial myo- lular matrix remodeling in canine and mouse myo- density and kv4.2 k+ channel expression in cultured
cytes. J Mol Cell Cardiol 47:586–594, 2009. cardial infarcts. Cell Tissue Res 324:475–488, rat cardiomyocytes. Cardiovasc Res 41:157–165,
71. Grandy SA, Fiset C: Ventricular k+ currents are 2006. 1999.
reduced in mice with elevated levels of serum 81. Espira L, Czubryt MP: Emerging concepts in 91. Pedrotty DM, Klinger RY, Kirkton RD, et al:
TNFalpha. J Mol Cell Cardiol 47:238–246, 2009. cardiac matrix biology. Can J Physiol Pharmacol Cardiac fibroblast paracrine factors alter impulse
72. Fernandez-Velasco M, Ruiz-Hurtado G, Hurtado 87:996–1008, 2009. conduction and ion channel expression of neonatal
O, et al: TNF-alpha downregulates transient 82. Thum T, Gross C, Fiedler J, et al: MicroRNA-21 rat cardiomyocytes. Cardiovasc Res 83:688–697,
outward potassium current in rat ventricular myo- contributes to myocardial disease by stimulating 2009.
cytes through inos overexpression and oxidant map kinase signalling in fibroblasts. Nature 92. Vasquez C, Mohandas P, Louie KL, et al:
species generation. Am J Physiol Heart Circ 456:980–984, 2008. Enhanced fibroblast-myocyte interactions in
Physiol 293:H238–H245, 2007. 83. Cleutjens JP, Verluyten MJ, Smiths JF, et al: Col- response to cardiac injury. Circ Res 107:1011–
73. Shang LL, Sanyal S, Pfahnl AE, et al: Nf-kappaB- lagen remodeling after myocardial infarction in the 1020, 2010.
dependent transcriptional regulation of the cardiac rat heart. Am J Pathol 147:325–338, 1995. 93. Ieda M, Tsuchihashi T, Ivey KN, et al: Cardiac
scn5a sodium channel by angiotensin II. Am J 84. Kakkar R, Lee RT: Intramyocardial fibroblast fibroblasts regulate myocardial proliferation
Physiol Cell Physiol 294:C372–C379, 2008. myocyte communication. Circ Res 106:47–57, through beta1 integrin signaling. Dev Cell 16:233–
74. Panama BK, Latour-Villamil D, Farman GP, et al: 2010. 244, 2009.
Nuclear factor kappaB downregulates the transient 85. Banerjee IFJ, Souders CA, Bowers SL, et al: The 94. Gaudesius G, Miragoli M, Thomas SP, et al: Cou-
outward potassium current i(to,f) through control role of interleukin-6 in the formation of the coro- pling of cardiac electrical activity over extended
of kchip2 expression. Circ Res 108:537–543, 2011. nary vasculature. Microsc Microanal 15:415–421, distances by fibroblasts of cardiac origin. Circ Res
75. Kohno T, Anzai T, Naito K, et al: Role of high- 2009. 93:421–428, 2003.
mobility group box 1 protein in post-infarction 86. Driesen RBDG, Verheyen FK, van den Eijnde SM, 95. Miragoli M, Gaudesius G, Rohr S: Electrotonic
healing process and left ventricular remodelling. et al: Partial cell fusion: A newly recognized type modulation of cardiac impulse conduction by myo-
Cardiovasc Res 81:565–573, 2009. of communication between dedifferentiating car- fibroblasts. Circ Res 98:801–810, 2006.
76. Liu W, Deng J, Xu J, et al: High-mobility group diomyocytes and fibroblasts. Cardiovasc Res 96. Rook MB, van Ginneken AC, de Jonge B, et al:
box 1 (HMGB1) downregulates cardiac transient 68:37–46, 2005. Differences in gap junction channels between
outward potassium current (ITO) through down- 87. Driesen RBVF, Dispersyn GD, Thoné F, et al: cardiac myocytes, fibroblasts, andheterologous
regulation of kv4.2 and kv4.3 channel transcripts Structural adaptation in adult rabbit ventricular pairs. Am J Physiol Cell Physiol 263:C959–C977,
and proteins. J Mol Cell Cardiol 49:438–448, 2010. myocytes: Influence of dynamic physical interac- 1992.
77. Dey D, Shepherd A, Pachuau J, et al: Leukemia tion with fibroblasts. Cell Biochem Biophys 97. Ishibashi Y, Urabe Y, Tsutsui H, et al: Negative
inhibitory factor regulates trafficking of t-type 44:119–128, 2006. inotropic effect of basic fibroblast growth factor on
Ca2+ channels. Am J Physiol Cell Physiol 88. Chilton L, Giles WR, Smith GL: Evidence of adult rat cardiac myocyte. Circulation 96:2501–
300:C576–C587, 2011. intercellular coupling between co-cultured adult 2504, 1997.
Role of the Autonomic Nervous System
in Atrial Fibrillation 47
Jian Sun, Benjamin J. Scherlag, and Sunny S. Po
and the vagus nerves constitute most of the parasympathetic limb

CHAPTER OUTLINE
of the extrinsic CANS, whereas the neurons in the intermedio-
The Cardiac Autonomic Nervous System (CANS) 469 lateral column of the spinal cord, the stellate ganglia, and their
axons en route to the heart make up most of the sympathetic limb
Autonomic Mechanisms of AF Initiation
of the extrinsic CANS. The intrinsic CANS is composed mainly
and Maintenance 471 of sympathetic and parasympathetic nerves, as well as ganglion-
Ablating the CANS to Treat AF 472 ated plexi (GP) on the heart itself or along the great vessels in
the thorax such as the pulmonary artery, aorta, superior vena
Modulation of CANS to Treat AF Without Destroying cava, and pulmonary veins (PVs). The stellate ganglia serve as the
Autonomic Neural Elements 473 “head stage” for the sympathetic innervation of the heart. The
Perspectives 473 postganglionic sympathetic fibers, mainly from the stellate
ganglia, constitute the vast majority of the sympathetic innerva-
tion to both the atrium and the ventricle. The GP embedded in
the epicardial fat pads contain up to several hundred autonomic
neurons. The distribution of the major ventricular GP is limited
The Cardiac Autonomic Nervous to the proximal segments of the coronary arteries; they are in
System (CANS) general small and not as extensive on the ventricles.2-5
The major atrial GP are located adjacent to the pulmonary
The autonomic nervous system can be viewed as the interface vein (PV)-atrial junction, or the junction of the right atrium and
between the central nervous system and the viscera, glands, and the superior or inferior vena cava. Chiou et al discovered that the
blood vessels. It is divided into three main components: sympa- efferent parasympathetic fibers in the vagus nerves converge at a
thetic, parasympathetic, and enteric.1 Integration of neural traf- GP before innervating the heart.6 This GP, at the junction of the
ficking among the afferent and efferent autonomic nerves as well right pulmonary artery, aorta, and superior vena cava (SVC), was
as their associated autonomic neurons maintains a delicate coined as the “head stage” GP because the bradycardia response
homeostasis of the function of the viscera, vessels, and glands. In induced by vagal stimulation in canine hearts was nearly abol-
mammalian hearts, the efferent sympathetic preganglionic ished if the RPA-Ao GP was ablated.6
neurons are located in the intermediolateral columns of the gray The different nomenclature used by anatomists introduced a
matter of the spinal cord; the preganglionic fibers of these great deal of confusion for scientific communication.2-5 In this
neurons pass through or synapse with the paravertebral ganglia chapter, we use the nomenclature based on clinical anatomy, for
(e.g., the stellate ganglia). The stellate ganglia, receiving neural example, GP’s relation to PVs (Figure 47-1). The superior left
inputs mainly from spinal nerves C6-T1, are the key neural GP (SLGP) and the inferior left GP (ILGP) are located adjacent
structures for cardiac sympathetic innervation.1 The efferent to the PV-atrial junction of the left superior PV and the left
parasympathetic preganglionic neurons are located in the motor inferior PV, respectively.7 The anterior right GP (ARGP) is situ-
nuclei of the vagus nerves (e.g., nucleus ambiguus) in the brain ated at the caudal end of the sinoatrial node, near the right
stem, from which the vagus nerves carry the preganglionic para- superior PV-atrial junction. The inferior right GP (IRGP)
sympathetic fibers to the heart. The parasympathetic postgangli- extends from the inferior right PV-atrial junction to the crux
onic neurons are concentrated mainly in the ganglionated plexi of the heart near the junction of the right atrium and inferior
embedded in epicardial fat pads, and the efferent postganglionic vena cava.
parasympathetic fibers are distributed over the entire heart. The It was once thought that the ARGP specifically innervates the
afferent autonomic fibers, both sympathetic and parasympathetic, sinus node, while the IRGP at the crux of the heart innervates
course along the cardiac plexus in the thorax and eventually reach only the AV node. Recent studies indicate that the intrinsic
the sensory neurons in the nodose ganglia at the base of the skull, CANS forms a complex neural network, and GP serve as “inte-
as well as the dorsal root ganglia of the spinal cord. These afferent gration centers” to control the physiological functions of the
nerves and ganglia mediate important cardiorespiratory reflexes heart.2,3,8 For example, high-frequency stimulation (HFS; 20 Hz)
(e.g., baroreflex) and the pain sensation from the heart to the to the SLGP also markedly slowed the sinus rate (SR), proving
brain.2,3 that the ARGP is not the only GP that innervates the sinus node.
Ablation of the ARGP greatly attenuated, but did not eliminate,
the SR slowing response induced by SLGP stimulation, indica-
The Extrinsic and Intrinsic CANS tive of the role of ARGP as the gateway GP for the sinus node
and the presence of other neural pathways bypassing the ARGP.
The CANS regulates vascular tone, contractility, and electro- Ablation of the four major atrial GP and the ligament of Marshall
physiology by transducing and integrating afferent and efferent (LOM) exerts potent inhibitory effects on the activity of the
autonomic trafficking.2,3 Autonomic control of the heart is medi- CANS, supporting clinical implications targeting these GP to
ated by a highly integrated intrinsic and extrinsic CANS.2,3 The treat AF.9
extrinsic CANS mainly consists of ganglia and their axons located The interplay between the intrinsic and extrinsic CANS is not
outside the heart. The nucleus ambiguus, the dorsal vagal nucleus, well understood. The intrinsic CANS appears to function
469
LSPV
RSPV SLGP
RSPV LOM LSPV
SLGP
LAA
RIPV LAA
RIPV
ARGP
ARGP
IRGP
A B
“CALCIUM-TRANSIENT TRIGGERING” HYPOTHESIS
LSPV
RSPV
Dependent upon Action
SLGP Calcium
potential transient
- Abbreviated action
potential
- Enhanced calcium
transient
Hyperactivity of 3 Na+/1 Ca++

LOM the CANS:
Acetylcholine
LIPV Shortens action
RIPV
potential EAD
Norepinephrine
IRGP Increases Ca++
C ILGP D loading Triggered firing
“OCTOPUS” HYPOTHESIS
Trigger
Micro-reentry
GP Macro-reentry
Figure 47-1. A-C, Location of the major atrial ganglionated plexi (GP) and the ligament of Marshall (LOM). SLGP, ILGP, ARGP, IRGP: superior left, inferior left, anterior right,
and inferior right GP, respectively. LSPV, LIPV, RSPV, RIPV: left superior, left inferior, right superior, and right inferior pulmonary veins, respectively. A, Right anterior oblique
(RAO) view. B, Anteroposterior (AP) caudal view. C, Posteroanterior (PA) view. D, Ca2+ transient hypothesis. Simultaneous activation of the sympathetic and parasympathetic
systems induces early afterdepolarization and subsequent triggered firing (see text for detail). E, Octopus hypothesis. Hyperactivity of the autonomic neurons (head of the
octopus) causes excessive release of the neurotransmitters of autonomic nerves (tentacles) at multiple sites. This leads to triggered firing and reentry at multiple sites to
initiate and maintain atrial fibrillation (AF).
Role of the Autonomic Nervous System in Atrial Fibrillation 471
interdependently with as well as independently from the extrinsic long-term outcome (<50% success, 5 years, single procedure) of
47
CANS, as evidenced by its retaining nearly full control of cardiac the standard CPVI clearly indicates that a better understanding
physiology after autotransplantation.2,3 Armour elegantly of the mechanisms underlying AF initiation and maintenance is
described the intrinsic CANS as “the little brain on the heart.” crucial if more effective ablation targets are to be identified.18
Cooperative interaction between the extrinsic and intrinsic
CANS maintains a homeostasis that facilitates balanced cardiac
physiological functions. The “Ca2+ Transient Triggering” Hypothesis
Clinical studies demonstrated that activation of both the sympa-
Noncholinergic, Nonadrenergic Neurotransmitters thetic and the parasympathetic nervous system commonly pre-
in the Intrinsic CANS ceded the initiation of paroxysmal AF.19,20 This finding was later
corroborated by multiple basic studies.21-23 Patterson et al pro-
Until the past two decades, it was believed that the sympathetic posed a “Ca2+ transient triggering” hypothesis to explain the ini-
component of the intrinsic CANS is composed exclusively of tiation of rapid PV firing (see Figure 47-1). This hypothesis states
postganglionic sympathetic fibers, and that all of the cardiac that norepinephrine (by sympathetic activation) augments the
autonomic neurons are parasympathetic neurons expressing cho- Ca2+ transient, and acetylcholine (by parasympathetic activation)
linergic markers. With advances in immunohistochemistry, sub- shortens the PV activation potential duration (APD). The abbre-
populations of cardiac autonomic neurons expressing various viated APD makes the Ca2+ transient relatively prolonged, and the
neurotransmitter markers have been identified.10 The presence myocytes even more Ca2+ overloaded. This leads to activation of
of peptidergic, nitrergic, and noradrenergic neurons, along the forward mode of the Na+/Ca2+ exchanger, the formation of
with their associated neurotransmitters such as neuropeptide-Y, early afterdepolarization, and subsequent triggered firing from
vasoactive intestinal peptide (VIP), nitric oxide synthase, and the PVs (Figure 47-1D). This hypothesis also helps explain the
angiotensin II, strongly indicates that autonomic control of observation that PV firing often occurs at the distal segments of
cardiac physiology involves a milieu of neurotransmitters the PVs, where the APD is the shortest.24 Zhou et al proposed an
beyond acetylcholine and norepinephrine.10-12 Neuropeptide-Y “octopus hypothesis,” in which the autonomic neurons in the
co-released by prolonged sympathetic activation reduces acetyl- major atrial GP function like the head of an octopus, while their
choline release from the neighboring vagal nerve ending; this is axons are analogous to the tentacles.25 When the head of the
a good example of sympathovagal cross-talk.11 These noncholin- octopus becomes hyperactive, excessive release of neurotransmit-
ergic, nonadrenergic neurotransmitters often exert effects similar ters from its tentacles can initiate triggered firing and macro- and
to those of cholinergic or adrenergic agonists or antagonists. micro-reentry at multiple sites to initiate and maintain AF (Figure
Using cholinergic and adrenergic blockers to “eliminate” CANS 47-1E). The octopus hypothesis also implies that targeting the
control is an oversimplified approach.13,14 Liu et al demonstrated head of the octopus (i.e., the GP) is perhaps the most effective
that until an antagonist of VIP ([Ac-Tyr1,D-phe2]-VIP) was means of mitigating a hyperactive state of the CANS and subse-
administered, vagal stimulation continued to induce atrial fibril- quent associated arrhythmias, i.e., paroxysmal AF.
lation (AF) in canine hearts despite GP ablation+atropine+esmo
lol.12 A better understanding of the arrhythmogenic potential of
these noncholinergic, nonadrenergic neurotransmitters may Rapid Firing from Non-PV Sites: LOM and SVC
facilitate the development of new antiarrhythmic agents for treat-
ment of AF. Although PV firing accounts for nearly 90% of initiation of
paroxysmal AF, non-PV sites such as LOM and SVC are alterna-
tive sites for rapid firing and AF initiation.26,27 Notably, the initia-
tion pattern of paroxysmal AF from the LOM, SVC, and PVs is
Autonomic Mechanisms of AF Initiation remarkably similar. LOM itself is richly innervated and had been
and Maintenance named the “left atrial neural fold” by some anatomists, signifying
the abundance of its autonomic neural elements.5 However, sub-
In 1978, Coumel et al described a group of patients with AF of stantial discrepancies in the relative abundance of sympathetic
vagal origin, manifested by the absence of structural heart disease versus parasympathetic innervation have been reported.28-31 On
and nocturnal onset of AF preceded by a slow SR.15 As the vast the basis of both electrophysiological and immunohistochemical
majority of AF patients appeared not to have the typical findings findings, Ulphani et al28 found LOM to be a parasympathetic
described by Coumel, vagal AF was viewed as a rarity. Landmark conduit in normal dogs, whereas Doshi et al29 demonstrated
findings reported by Haïsseguerre et al demonstrated that parox- sympathetic predominance within the LOM in dogs with chronic
ysmal AF, in most cases, originated from rapid focal firing in the AF. In contrast, Makino et al found in human hearts that the
PVs.16 Subsequent studies verified the pathophysiological roles LOM-LSPV (left superior pulmonary vein) junction was sympa-
of the PV muscle sleeve as an ideal substrate for reentry and thetically predominant, whereas the LOM-coronary sinus junc-
discovered periodic acid–Schiff (PAS)-positive cells in the PV tion was predominantly parasympathetic.30 Such an innervation
sleeves that appear to be reminiscent of Purkinje cells.17 However, gradient was later corroborated by Lin et al, who found that HFS
these PAS-positive cells have not been shown to elicit rapid firing at the LOM-coronary sinus junction mainly elicited AF, whereas
(300 to 600 beats/min [bpm]), and the PV sleeve, despite being HFS at the LOM-LSPV junction induced ventricular tachycar-
an ideal reentrant substrate, cannot initiate reentry without a dia, atrial tachycardia, or junctional tachycardia.31 However, it is
spontaneous, well-timed premature beat. unquestionable that the LOM provides an ideal substrate for
Over the past 15 years, AF ablation has evolved from eliminat- triggered firing if sympathetic versus parasympathetic balance is
ing the focal trigger(s) within the PVs (focal PV ablation) to altered, or if both become hyperactive.
circumferentially isolating the PV-atrial antrum (circumferential Another common site for non-PV firing is the SVC-atrial
pulmonary vein isolation [CPVI]). However, the following junction, where high density autonomic innervation has been
fundamental questions have not been fully addressed: (1) Why shown.6 In patients, the site of rapid firing and successful ablation
do PVs elicit rapid firing? (2) How does PV firing initiate was usually the posterior-septal aspect of the SVC-RA junction
AF? (3) How does AF maintain itself, particularly in the first adjacent to the RPV-Ao GP, suggesting that this “head stage” GP
few hours before structural remodeling starts? The disappointing may serve as the autonomic basis for SVC firing. Lu et al
delivered HFS to this GP and elicited rapid firing and AF only Clinical studies have independently reported that
at the SVC, not at the atrium or PV.32 Ablation of this GP only cholinergic+adrenergic blockers failed to eliminate CFAE in
prolonged the refractory period of the SVC site. These findings patients undergoing AF ablation,13,14 casting doubt on the auto-
indicate that the RPA-Ao GP may provide specific autonomic nomic mechanism for CFAE. The CFAEs recorded in patients
innervation to the SVC-atrial junction, and hyperactivity of this with paroxysmal AF were significantly more sensitive to auto-
GP may serve as the basis for SVC firing. nomic blockers than the CFAEs in persistent AF patients, under-
lying the importance of other mechanisms (e.g., fibrosis)
responsible for CFAE formation. Of note, a large milieu of non-
Perpetuation of AF: the First Few Hours and Beyond cholinergic, nonadrenergic neurotransmitters has been identified
within the intrinsic CANS.10-12 Failure to eliminate CFAE by
Electrical remodeling (e.g., shortening of the refractory period) cholinergic and adrenergic blockade cannot exclude the auto-
and structural remodeling (e.g., fibrosis) are indispensable factors nomic mechanism underlying CFAE.
in the perpetuation of AF over a period of several weeks or
months.33-35 However, how AF perpetuates itself within the first
few minutes or hours after its initiation (e.g., before the occur-
rence of structural remodeling) is poorly understood. Any therapy Ablating the CANS to Treat AF
that provides early termination of AF will have a great impact on
AF therapy. Yu et al recorded the neural activity of the canine In ablating the CANS to treat AF, the most practical targets are
ARGP or SLGP during 6 hours of AF simulated by rapid atrial the major atrial GP (e.g., the head of the “octopus”). An acciden-
pacing.36 Rapid atrial pacing not only shortened the atrial refrac- tal finding was reported by Pappone et al, who described a group
tory period but also progressively enhanced neural activity within of paroxysmal AF patients undergoing CPVI.41 When a vagal
the GP, providing direct evidence of autonomic remodeling response was evoked by ablation, elimination of all evoked vagal
during AF. Yu et al proposed that electrical remodeling and auto- responses around PV ostia resulted in outstanding success (99%)
nomic remodeling form a vicious cycle. A hyperactive state of the of freedom from AF. Scanavacca et al showed the feasibility of
intrinsic CANS (e.g., GP) facilitates AF initiation. AF then short- selective GP ablation using both endocardial and epicardial
ens the refractory period further and augments GP activity to approaches, but the result of GP-only ablation was disappoint-
perpetuate AF itself. This “vicious cycle” hypothesis helps answer ing.42 When our understanding of the anatomy and physiology
the fundamental questions described above: (1) Why do PVs of the major atrial GP was advanced, the results of GP ablation
elicit rapid firing? (2) How does PV firing initiate AF? (3) How improved substantially. A major limitation of GP ablation is that
does AF maintain itself? despite consistent location adjacent to the PV-atrial junction, the
As has been discussed, AF begets electrical, structural, and extent of each “hyperactive” GP that needs to be ablated to treat
autonomic remodeling. As AF burden increases, increased GP AF is largely unknown. Pokushalov et al randomly assigned 80
activity may spread to peripheral atrial sites. Excessive neu- paroxysmal AF patients to two groups: (1) selective GP ablation
rotransmitter release locally engenders more atrial substrates guided by vagal responses induced by HFS (20 Hz), and (2) abla-
(e.g., shortened and dispersed refractory period, higher intracel- tion of GP selected by presumed anatomical locations.43 At 13.1
lular Ca2+ load and inflammation) to sustain AF. We hypothesize ± 1.9 months, 42.5% of patients with HFS-guided GP ablation
that “metastasizing” activity within the intrinsic CANS may be a and 77.5% of patients with anatomic GP ablation were free of
crucial element in facilitating the progression of AF from the symptomatic AF (P = .02). This difference may be attributed to
paroxysmal form to more advanced forms. a significantly greater number of radiofrequency applications
delivered in the latter group, covering a larger area for each GP.
Although GP-only ablation appeared to produce results similar
Autonomic Basis for Complex Fractionated Atrial to those of standard CPVI, the combination of CPVI and GP
Electrograms (CFAEs) ablation seems to lead to higher success rates than are achieved
with CPVI alone.7,44 However, the value of GP ablation is not
In 2004, Nademanee et al described the technique for targeting without debate. Adding GP ablation to CPVI introduces an
CFAE to treat patients with paroxysmal and persistent AF.37 arrhythmogenic substrate and increases the incidence of reen-
Although CFAE may simply be caused by physical properties of trant left atrial tachycardia by approximately 30%.7 This factor
myocardial fibers such as anisotropic conduction or zones of slow may lower the success rate compared with standard CPVI.45
conduction, two other hypotheses also help account for the for- Although the debate about adding GP ablation to CPVI contin-
mation of CFAE: rotor and autonomic hypotheses. The rotor ues, it cannot be overemphasized that CPVI transects the LOM,
hypothesis indicates that CFAE is formed when meandering three of four major atrial GP at the PV-atrial junction (see Figure
rotors encounter heterogeneous substrates (e.g., dispersion of 47-1), and numerous autonomic nerves. The contribution of
refractoriness).38 CFAE is therefore the by-product of the reen- autonomic denervation to the efficacy of CPVI cannot be
trant rotor, which breaks down at its boundary. Notably, infusion overlooked.
of acetylcholine into the animal heart is required to maintain a The disappointing long-term results of a single AF ablation
stable rotor, implying that the rotor hypothesis is also related to raise further questions: What is the cause of AF recurrence after
autonomic activity. The autonomic hypothesis stems from the ablation? Is the isolated or destroyed myocardium the real trigger
locations of the CFAE originally described by Nadamanee et al, or substrate for AF? Although resumption of conduction between
in which distribution of the CFAE correlates well with the PV and atrium is considered the main cause for recurrence,
locations of the major atrial GP. Lin et al showed in a canine mapping of the PV-atrial junction in patients without clinical
model that CFAE can be produced by topical application of recurrence after AF ablation is rarely done. Furlanello et al per-
acetylcholine,39 and CFAE can be eliminated by ablation of the formed catheter ablation (PV isolation ± atrial flutter ablation)
GP at a distance, indicating that activating the intrinsic CANS is on 20 competitive athletes who had very symptomatic lone AF.46
critical in the formation of CFAE. Clinical studies later corrobo- Successful PV isolation was achieved in 19 of 20 patients in the
rated that CFAE tend to occur at presumed GP sites.40 Ablation first ablation procedure. When all patients were re-studied
of GP greatly reduced the extent of CFAE distribution around regardless of arrhythmia recurrence, 62 (81%) of the previously
the GP, implying a possible causal relation between CANS activ- isolated PVs had resumed conduction. Most important, the inci-
ity and CFAE.7,40 dence of conduction recurrence did not differ between patients
Role of the Autonomic Nervous System in Atrial Fibrillation 473
with (82.5%) and without (74.6%) AF recurrence, leading to other autonomically based diseases without permanent injury to
47
questions about the real cause of AF recurrence after ablation. the myocardium or to the intrinsic CANS.
Modulation of CANS to Treat AF Without Perspectives

Destroying Autonomic Neural Elements
Up to now, the clinical benefit of GP ablation in treating AF has
The consequences of neural degeneration and regeneration after been controversial, at best. In contrast, renal artery denervation
ablation are poorly understood. If autonomic denervation and that dramatically reduced blood pressure in patients with
regeneration are critical elements in the efficacy and failure of drug-resistant hypertension has refocused attention on the
CPVI, respectively, suppression of the hyperactive CANS by role of the autonomics in cardiovascular function.52 The sug-
modulating, instead of destroying, the autonomic neural ele- gested mechanism was that mitigation of sympathetic afferent
ments may be a more effective approach. Inspired by a clinical signals to vasomotor centers in the brain reduced efferent
report by Tai et al47 showing that PV firing in paroxysmal AF sympathetic vasoconstriction of the renal arteries. Although it
patients could be inhibited by increased vagal reflex caused by remains unclear which sympathetic pathways are modified
phenylephrine-induced hypertension, the Oklahoma group by renal artery denervation, a more direct relationship
hypothesized that by taking advantage of neural plasticity, low- between autonomic modulation and AF was reported in a clinical
level vagal stimulation (LL-VS) at voltages not slowing the SR study by Pokushalov et al.53 Patients with concomitant AF and
or AV conduction may inhibit the CANS, and subsequently AF drug-resistant hypertension were randomly assigned to two
inducibility. A series of acute canine studies corroborated that groups: CPVI and CPVI plus renal artery denervation. At
LL-VS markedly lengthened the atrial and PV refractory period follow-up of 12 months, patients with CPVI alone were 29%
and inhibited AF inducibility.48-50 Notably, LL-VS was capable of AF-free, whereas among patients with pulmonary vein isolation
preventing AF initiation and terminating AF. The antiarrhythmic plus renal artery denervation, 69% were AF-free (P = .03). The
effects were mediated by suppression of neural activity of the result that renal sympathetic denervation, rather than increased
major atrial GP and the stellate ganglia.49 Long-term canine radiofrequency applications delivered to the PVs or the LA,
studies by Shen et al not only verified these findings but also markedly improved the AF ablation outcome highlights a para-
discovered that suppression of stellate ganglion activity is respon- digm change in the treatment of arrhythmia. Instead of solely
sible for the effects of LL-VS on AF.51 Notably, in acute canine targeting the myocardial substrates, modulation of CANS activ-
studies, LL-VS of 80% below threshold voltage was as effective ity may alter both neural and myocardial substrates for arrhyth-
as 10% below threshold in suppressing AF, indicating that LL-VS mias and may mitigate the arrhythmias with minimal myocardial
may be a clinically feasible approach to the treatment of AF and damage.
11. Herring N, Paterson DJ: Neuromodulators of 21. Ogawa M, Zhou S, Tan AY, et al: Left stellate
References peripheral cardiac sympatho-vagal balance. Exp ganglion and vagal nerve activity and cardiac
Physiol 94:46–53, 2009. arrhythmias in ambulatory dogs with pacing-
1. Goetz CG: Autonomic nervous system. In Text- 12. Liu Y, Scherlag BJ, Fan Y, et al: Inducibility of induced congestive heart failure. J Am Coll Cardiol
book of Clinical Neurology, ed 3, Philadelphia, atrial fibrillation after GP ablations and ‘autonomic 50:335–343, 2007.
2007, Saunders, pp 383–404. blockade’: Evidence for the pathophysiological 22. Sharifov OF, Fedorov VV, Beloshapko G, et al:
2. Armour JA: Functional anatomy of intrathoracic role of the non-adrenergic and non-cholinergic Roles of adrenergic and cholinergic stimulation in
neurons innervating the atria and ventricles. Heart neurotransmitters. J Cardiovasc Electrophysiol spontaneous atrial fibrillation in dogs. J Am Coll
Rhythm 7:994–996, 2010. 24:188–195, 2013. Cardiol 43:483–490, 2004.
3. Armour JA: The little brain on the heart. Cleveland 13. Knecht S, Wright M, Matsuo S, et al: Impact of 23. Patterson E, Lazzara R, Szabo B, et al: Sodium-
Clin J Med 74:S48–S51, 2007. pharmacological autonomic blockade on complex calcium exchange initiated by the Ca2+ transient:
4. Saburkina I, Rysevaite K, Pauziene N, et al: Epi- fractionated atrial electrograms. J Cardiovasc Elec- An arrhythmia trigger within pulmonary vein.
cardial neural ganglionated plexus of ovine heart: trophysiol 21:766–772, 2010. J Am Coll Cardiol 47:1196–1206, 2006.
Anatomic basis for experimental cardiac electro- 14. Chaldoupi SM, Linnenbank AC, Wittkampf FH, 24. Po SS, Li Y, Tang D, et al: Rapid and stable
physiology and nerve protective cardiac surgery. et al: Complex fractionated electrograms in the re-entry within the pulmonary vein as a mechanism
Heart Rhythm 7:942–950, 2010. right atrial free wall and the superior/posterior wall initiating paroxysmal atrial fibrillation. J Am Coll
5. Pauza DH, Skripka V, Pauziene N: Morphology of of the left atrium are affected by activity of the Cardiol 45:1871–1877, 2005.
the intrinsic cardiac nervous system in the dog: A autonomic nervous system. J Cardiovasc Electro- 25. Zhou J, Scherlag BS, Edwards J, et al: Gradients of
whole-mount study employing histochemical physiol 23:26–33, 2012. atrial refractoriness and inducibility of atrial fibril-
staining with acetylcholinesterase. Cells Tissues 15. Coumel P, Attuel P, Lavallée J, et al: The atrial lation due to ganglionated plexi stimulation. J Car-
Organs 172:297–320, 2002. arrhythmia syndrome of vagal origin. Arch Mal diovasc Electrophysiol 18:83–90, 2007.
6. Chiou CW, Eble JN, Zipes DP: Efferent vagal Coeur Vaiss 71:645–656, 1978. 26. Chang HY, Lo LW, Lin YJ, et al: Long-term
innervation of the canine atria and sinus and atrio- 16. Haïssaguerre M, Jaïs P, Shah DC, et al: Spontane- outcome of catheter ablation in patients with
ventricular nodes: The third fat pad. Circulation ous initiation of atrial fibrillation by ectopic beats atrial fibrillation originating from the superior
95:2573–2584, 1997. originating in the pulmonary veins. N Engl J Med vena cava. J Cardiovasc Electrophysiol 24:250–258,
7. Po SS, Nakagawa H, Jackman WM: Localization 339:659–666, 1998. 2013.
of left atrial ganglionated plexi in patients with 17. Perez-Lugones A, McMahon JT, Ratliff NB, et al: 27. Hwang C, Fishbein MC, Chen PS: How and when
atrial fibrillation. J Cardiovasc Electrophysiol Evidence of specialized conduction cells in human to ablate the ligament of Marshall. Heart Rhythm
20:1186–1189, 2009. pulmonary vein of patients with atrial fibrillation. 3:1505–1507, 2006.
8. Hou Y, Scherlag BJ, Zhou J, et al: Interactive atrial J Cardiovasc Electrophysiol 14:803–809, 2003. 28. Ulphani JS, Arora R, Cain JH, et al: The ligament
neural network: Determining the connections 18. Ouyang F, Tilz R, Chun J, et al: Long-term results of Marshall as a parasympathetic conduit. Am J
between ganglionated plexi. Heart Rhythm 4:56– of catheter ablation in paroxysmal atrial fibrillation: Physiol Heart Circ Physiol 293:H1629–H1635,
63, 2007. Lessons from a 5-year follow-up. Circulation 2007.
9. Lu Z, Scherlag BJ, Lin J, et al: Autonomic mecha- 122:2368–2377, 2010. 29. Doshi RN, Wu TJ, Yashima M, et al: Relation
nism for initiation of rapid firing from atria and 19. Bettoni M, Zimmermann M: Autonomic tone between ligament of Marshall and adrenergic
pulmonary vein: Evidence by ablation of ganglion- variations before the onset of paroxysmal atrial atrial tachyarrhythmia. Circulation 100:876–883,
ated plexi. Cardiovasc Res 84:245–252, 2009. fibrillation. Circulation 105:2753–2759, 2002. 1999.
10. Hoover DB, Isaacs ER, Jacques F, et al: Localiza- 20. Amar D, Zhang H, Miodownik S, et al: Competing 30. Makino M, Inoue S, Matsuyama TA, et al: Diverse
tion of multiple neurotransmitters in surgically autonomic mechanisms precede the onset of postmyocardial extension and autonomic innervation
derived specimens of human atrial ganglia. Neuro- operative atirial fibrillation. J Am Coll Cardiol on ligament of Marshall in humans. J Cardiovasc
science 164:1170–1179, 2009. 42:1262–1268, 2003. Electrophysiol 17:594–599, 2006.
31. Lin J, Scherlag BJ, Lu Z, et al: Inducibility of atrial 39. Lin J, Scherlag BJ, Lu Z, et al: Autonomic mecha- ing usual training schedule and sport competition.
and ventricular arrhythmias along the ligament of nism for complex fractionated atrial electrograms. J Cardiovasc Electrophysiol 19:457–462, 2008.
Marshall: Role of autonomic factors. J Cardiovasc J Cardiovasc Electrophysiol 19:835–842, 2008. 47. Tai CT, Chiou CW, Wen ZC, et al: Effect of phen-
Electrophysiol 19:955–962, 2008. 40. Katritsis D, Sougiannis D, Batsikas K, et al: Auto- ylephrine on focal atrial fibrillation originating in
32. Lu Z, Scherlag BJ, Niu G, et al: Functional proper- nomic modulation of complex fractionated atrial the pulmonary veins and superior vena cava. J Am
ties of the superior vena cava (SVC)–aorta gangli- electrograms in patients with paroxysmal atrial Coll Cardiol 36:788–793, 2000.
onated plexi: Evidence suggesting an autonomic fibrillation. J Interv Card Electrophysiol 31:217– 48. Li S, Scherlag BJ, Yu L, et al: Low level vagosympa-
basis for rapid SVC firing. J Cardiovasc Electro- 223, 2011. thetic stimulation: A paradox and potential new
physiol 21:1392–1399, 2010. 41. Pappone C, Santinelli V, Manguso F, et al: modality for the treatment of focal atrial fibrillation.
33. Wijffels MC, Kirchhof CJ, Dorland R, et al: Atrial Pulmonary vein denervation enhances long-term Circ Arrhythm Electrophysiol 2:645–651, 2009.
fibrillation begets atrial fibrillation: A study in benefit after circumferential ablation for paroxys- 49. Sha Y, Scherlag BJ, Yu L, et al: Low-level right
awake chronically instrumented goats. Circulation mal atrial fibrillation. Circulation 109:327–334, vagal stimulation: Anticholinergic and antiadrener-
92:1954–1968, 1995. 2004. gic effects. J Cardiovasc Electrophysiol 10:1147–
34. Everett TH, 4th, Olgin JE: Atrial fibrosis and the 42. Scanavacca M, Pisani CF, Hachul D, et al: Selective 1153, 2011.
mechanisms of atrial fibrillation. Heart Rhythm atrial vagal denervation guided by evoked vagal 50. Xia S, Scherlag BJ, Yu L, et al: Prevention and
4(3 Suppl):S24–S27, 2007. reflex to treat patients with paroxysmal atrial fibril- reversal of atrial fibrillation inducibility and auto-
35. Yue L, Xie J, Nattel S: Molecular determinants of lation. Circulation 114:876–885, 2006. nomic remodeling by low-level vagosympathetic
cardiac fibroblast electrical function and therapeu- 43. Pokushalov E, Romanov A, Shugayev P, et al: stimulation. J Am Coll Cardiol 57:563–571, 2011.
tic implications for atrial fibrillation. Cardiovasc Selective ganglionated plexi ablation for paroxys- 51. Shen MJ, Shinohara T, Park HW, et al: Continu-
Res 89:744–753, 2011. mal atrial fibrillation. Heart Rhythm 6:1257–1264, ous low-level vagus nerve stimulation reduces stel-
36. Yu L, Scherlag BJ, Sha Y, et al: Interactions 2009. late ganglion nerve activity and paroxysmal atrial
between atrial electrical remodeling and autonomic 44. Katritsis DG, Giazitzoglou E, Zografos T, et al: tachyarrhythmias in ambulatory canines. Circula-
remodeling: How to break the vicious cycle. Heart Rapid pulmonary vein isolation combined with tion 123:2204–2212, 2011.
Rhythm 9:804–809, 2012. autonomic ganglia modification: A randomized 52. Symplicity HTN-2 Investigators: Renal sympa-
37. Nademanee K, McKenzie J, Kosar E, et al: A new study. Heart Rhythm 8:672–678, 2011. thetic denervation in patients with treatment-
approach for catheter ablation of atrial fibrillation: 45. Kron J, Kasirajan V, Wood MA, et al: Management resistant hypertension: A randomised controlled
Mapping of the electrophysiologic substrate. J Am of recurrent atrial arrhythmias after minimally trial. Lancet 376:1903–1909, 2010.
Coll Cardiol 43:2044–2053, 2004. invasive surgical pulmonary vein isolation and gan- 53. Pokushalov E, Romanov A, Corbucci G, et al: A
38. Kalifa J, Tanaka K, Zaitsev AV, et al: Mechanisms glionic plexi ablation for atrial fibrillation. Heart randomized comparison of pulmonary vein isola-
of wave fractionation at boundaries of high- Rhythm 7:445–451, 2010. tion with versus without concomitant renal artery
frequency excitation in the posterior left atrium of 46. Furlanello F, Lupo P, Pittalis M, et al: Radiofre- denervation in patients with refractory symptom-
the isolated sheep heart during atrial fibrillation. quency catheter ablation of atrial fibrillation in atic atrial fibrillation and resistant hypertension.
Circulation 113:626–633, 2006. athletes referred for disabling symptoms prevent- J Am Coll Cardiol 60:1163–1170, 2012.
Mechanisms of Ventricular
Tachycardia and Fibrillation 48
Derek J. Dosdall and Raymond E. Ideker
or slow heart rate and have been linked to numerous ion channel
CHAPTER OUTLINE
disturbances including L-type calcium, rectifier potassium, and
Mechanisms of VT Onset and Maintenance 475 late sodium currents. The sodium/calcium exchanger current is
thought to play a critical role in the development of EADs. Chua
Transition from VT to VF 475
et al performed optical mapping of tachycardia-induced heart
VF Maintenance 476 failure and demonstrated that
heterogeneous up-regulation of apamin-sensitive K+ current
Conclusions 480
increases sensitivity to intracellular calcium.5 This leads to
heterogeneous APD shortening and possibly to late phase 3
EADs in the context of high heart rate or post-defibrillation
Sudden cardiac death (SCD) causes an estimated 300,000 deaths recovery. Stretch activation may play a role in ventricular
in the United States alone.1 Ventricular tachycardia (VT) often arrhythmias, particularly in the context of severe heart failure and
precedes the onset of ventricular fibrillation (VF). VF causes volume overload.6 Abnormal automaticity consists of abnormal
approximately one-third of sudden cardiac deaths.2 Patients at spontaneous firing action potentials that are not coupled to previ-
high risk for SCD may be implanted with an implanted ous activations.
cardioverter-defibrillator (ICD), but the largest group of victims Stable reentrant circuits around an anatomic anchor such as a
of SCD does not have risk factors that place them in the high-risk scar or a large vessel may lead to sustained VT (see Figure 48-2,
category as candidates for ICD implantation.3 To develop more A). A section of unexcitable post-infarct scar or a large vessel such
effective treatments for VT and VF, the mechanisms of VT and as the aorta or the right ventricular (RV) outflow tract may form
VF onset and maintenance must be understood. a pathway for a stable reentrant circuit. Even without an unexcit-
able core, simulations have shown that a reentrant circuit may be
formed around a functional core rather than an anatomic core
(see Figure 48-2, A). The center of the circuit may remain excit-
Mechanisms of VT Onset and Maintenance able but unexcited as the spiral wave circles around the core.
Large reentrant circuits that encircle the entire ventricles have
Although cases of idiopathic VF have been reported, most been shown to form as well, especially with cardiac dilatation or
patients with VT and VF have a substrate that increases the prob- conduction slowing.
ability of reentry. Two primary conditions lead to the onset of The reentry wavelength is the product of the conduction
VT: ectopic foci and stable reentrant circuits. Ectopic foci due to velocity and the distance of the path of the reentrant circuit.
triggered activity or abnormal automaticity may lead to VT. VT Ischemia, fibrosis, cellular uncoupling, and electrolyte imbalance
may be hemodyamically stable or unstable and often self- may cause conduction slowing. Cardiac dilatation leads to greater
terminates, whereas VF is almost always fatal if not treated by distances for wave-front conduction. When these changes occur
administration of defibrillation shocks within minutes of VF heterogeneously over the cardiac tissue, conduction slowing may
onset. Ventricular tachyarrhythmias often progress from prema- be present to such an extent that slowed conduction in one region
ture ventricular complexes (PVCs) to runs of VT, and finally to may lead to fractionation of a single wave front. The faster-
VF (Figure 48-1). Ventricular tachycardia and heart failure lead moving portions of the wave front move around the region that
to action potential duration (APD) shortening, and rapid activa- has slowed conduction. If the region of slowed conduction delays
tion rates may lead to intracellular calcium overload. High intra- the wave front for an extended time, the tissue past the area of
cellular calcium and APD shortening promote triggered activity slowed conduction may have sufficient time to become excitable.
and the initiation of VF.4 As the excitation wave front leaves the region of slowed conduc-
Focal sources that may lead to VT include triggered activity tion, it may travel rapidly through the normal area back around
and abnormal automaticity. Focal sources may be the source of to the area of slowed conduction that activated initially. If this
repetitive rapid ectopic firing or they may lead to disruption of tissue is now excitable, the wave front may reenter into the area
the normal conduction pathways and the breakup of cohesive of slowed conduction, and the process may start again and estab-
wave fronts and establishment of reentrant circuits (Figure 48-2, lish a stable reentrant circuit.
A). Delayed afterdepolarizations (DADs) cause a rise in resting
potential during diastole. If the transmembrane potential rises
above the activation threshold, a new action potential may be
launched. DADs are traditionally linked to intracellular calcium Transition From VT to VF
overload and thus are exacerbated by the rapid heart rates seen
in VT and VF. Early afterdepolarizations (EADs) may occur Heterogeneous conduction leads to the breakup of continuous
during the plateau (phase 2) or repolarization (phase 3) phase of activation wave fronts, the development of sustained reentrant
an action potential. EADs are often associated with bradycardia circuits, and the development of sustained VF. Critically timed
focal activations may break up regular rotors and cause wave-
front collisions and conduction block, leading to breakup of the
This research is supported by grants HL091138 and HL085370 from the National Heart, Lung, and Blood Institute. regularly repeating wave fronts.
475
A primary mechanism for block leading to reentry is nonuni- When the slope of the restitution curve (APD/diastolic interval
form dispersion of refractoriness. A study by Geizer et al reported [DI]) is greater than 1, an unstable positive-feedback loop results
that a series of premature stimuli that induced large spatial disin increasing oscillations in APD and DI (Figure 48-4). Once the
persion of repolarization caused VF in an in vivo dog model.7 APD and the DI become so shortened that conduction is not
This study and others have shown that block can occur even in possible, the wave front blocks and reentry may occur. A slope of
normal, homogeneous tissue when alternans is caused in APD. less than 1 of the restitution curve leads to a negative-feedback
Dispersion in APD leads to the development of conduction block loop, decreasing alternans, and to convergence of cycle length
and to fractionation of wave fronts, which, in turn, may lead to and DI to an equilibrium.
reentrant circuits and the initiation of VT and VF (Figure 48-3).
Simulations have shown that the restitution curve (the rela-
tionship between APD and the diastolic interval) is predictive of
the breakdown of a rotor to multiple unstable reentrant circuits. VF Maintenance
VF changes as global ischemia sets in and the duration of VF
SR VT VF increases. Using high-speed cinematography on canine hearts,
Wiggers first proposed four stages of VF: (1) an initial tachysys-
tolic stage (<1 s since VF onset), (2) convulsive incoordination (1
to 40 s), (3) tremulous incoordination (40 s to 3 min), and (4)
progressive atonic incoordination (>3 min).8 Other groups have
proposed 2,9 3,10 4,11 and 512 or more stages to describe the pro-
gression of VF over time. Each strategy for classifying VF activa-
tion patterns describes a change in VF activation rate and
organization at approximately 2 to 3 minutes, with a progressive
decline in activation rate and an increased incidence of conduc-
VT VT tion block thereafter.
initiation degeneration Animal studies have shown that the ability of the heart to
effectively contract is sufficiently compromised after 3 minutes
Figure 48-1. Surface ECG recording of the transition from normal sinus rhythm (SR)
to ventricular tachycardia (VT) and eventually to ventricular fibrillation (VF).
of VF that even after successful defibrillation, return of spontane-
ous circulation does not occur without drugs and/or mechanical
(Reprinted with permission from Weiss JN, Garfinkel A, Karagueuzian HS, et al: Chaos chest compressions.13,14 With the significant changes in cardiac
and the transition to ventricular fibrillation: A new approach to antiarrhythmic drug electrophysiology and function that occur after prolonged VF, it
evaluation. Circulation 99:2819-2826, 1999.) is reasonable to expect that the mechanisms of VF maintenance
Focal VT Anatomical reentry Functional reentry
A
Multiple wavelets Mother rotor
–85 mV 10 mV
B
Figure 48-2. The mechanisms of VT and VF are demonstrated in 2D simulations of transmembrane voltage A, VT may be maintained by a focal
source, around an anatomic anchor, or as a reentrant circuit around a functional anchor. B, VF may be maintained by multiple wavelet reentry or by a mother rotor (bottom
left of panel) that fractionates and breaks up into multiple daughter wavelets.
(Reproduced with permission from Qu Z, Weiss JN: Nonlinear dynamics of excitation and propagation in cardiac muscle. In Zipes DP, Jalife J [eds]: Cardiac Electrophysiology:
From Cell to Bedside. Philadelphia, Elsevier, 2009, pp 339-348.)
Mechanisms of Ventricular Tachycardia and Fibrillation 477
48
A
A Long Short
160 ms
B Short Long
70 ms
Figure 48-3. The formation of a reentrant circuit due to heterogeneity of APD An ectopic focus at B initiates in the region of short APD and propagates
outward toward the region of long APD. The ectopic beat blocks (dotted line) as it encounters refractory tissue but continues to propagate laterally such that it wraps around
the area of block and reenters through the previous region of block and back into the short APD region.
(Reproduced with permission from Qu Z, Weiss JN: Nonlinear dynamics of excitation and propagation in cardiac muscle. In Zipes DP, Jalife J [eds]: Cardiac Electrophysiology:
From Cell to Bedside. Philadelphia, Elsevier, 2009, pp 339-348.)
APD restitution slope<1 APD restitution slope>1

APD
APD
a a
CL=APD+DI b CL
A DI B DI
Figure 48-4. Progression of alternans in APD and DI to conduction block and spiral wave breakdown depends on the slope of the restitution
curve A, A spiral wave rotating at a constant cycle length (CL) undergoing a perturbation (a) that decreases DI and will tend to cause a smaller change in APD, which in
turn will cause a smaller change in DI. This negative-feedback loop will eventually lead to a return to equilibrium. B, When the restitution slope >1, a perturbation in APD
(a) leads to a larger disturbance in DI, which leads to a larger change in APD. This positive-feedback loop leads to increased variation in APD and DI until conduction is no
longer sustained, but blocks (b). This leads to spiral wave breakup and chaotic conduction.
(Reproduced with permission from Weiss JN, Garfinkel A, Karagueuzian HS, et al: Chaos and the transition to ventricular fibrillation: A new approach to antiarrhythmic drug
evaluation. Circulation 99:2819-2826, 1999.)
may change with the progression of VF. Although many studies The mother rotor causes a high-frequency, repeatable activation
have investigated the mechanisms of short-duration VF (SDVF; sequence in a single region of the heart, also called the dominant
lasting <1 min), relatively few studies have explored the mecha- domain. As the distance from the center of the motor rotor
nisms of long duration VF (LDVF; lasting >1 min). Although increases, the distance traveled by the wave front increases. The
SDVF is of clinical relevance for ICD therapy, the average time far arms of the rotor begin to fractionate and cause daughter
from collapse of patients until defibrillation in the prehospital reentrant circuits that may be short-lived and meandering. The
setting ranges from 4 to 10 minutes.15 Therefore, the mechanisms daughter wave fronts propagate onto the remainder of the
of SDVF and LDVF are of clinical significance and interest. working myocardium and do not exhibit repeating activation
Recent studies have indicated that three different mechanisms sequences (Figure 48-2, B). The breakup of the daughter wave
may be involved in sustaining VF: (1) mother rotor reentry, (2) fronts and irregular patterns of activation away from the mother
wandering wavelet reentry, and (3) focal activity in the Purkinje rotor give the electrocardiogram (ECG) the chaotic characteris-
system. VF mechanisms may vary over time and may switch tics of VF.
between mechanisms regularly during LDVF.16 Recent experimental studies and computer simulations have
demonstrated that the inward potassium rectifier current plays
an important role in rotor stability during VF.17 Increased inward-
Mother Rotor Reentry rectifier current prevents the wave front from colliding with the
wave tail and thus prevents breakup of stable rotors. Stable
Several groups have proposed that a single, stable reentrant mother rotors have been recorded in isolated sheep ventricles
circuit, called a mother rotor, is responsible for maintaining VF. and in guinea pig hearts, and recent canine studies have shown
activity that is consistent with mother rotor reentry on the endo- myocardium nearly to the epicardial surface. The development
cardium after 2 minutes of VF.16 of transmural block and retention of the ability to conduct as
LDVF progresses have been shown to represent an irregular and
heterogeneous process.21
Wandering Wavelet Reentry To investigate the role of the Purkinje system in LDVF,
Tabereaux et al. conducted direct endocardial mapping studies on
Wandering wavelet reentry was first proposed as the mechanism isolated, perfused dog hearts.22 They recorded endocardial
of VF maintenance almost 50 years ago. The initial theory pro- activation maps from the insertion of the anterolateral papillary
posed that for VF to start, a premature stimulus initiated the muscle of electrically induced VF. Purkinje activations were
arrhythmia, and heterogeneity of refractory periods in the tissue detected separately from myocardial activations, and three dis-
caused wave-front fractionation that facilitated reentry. Wander- tinct types of activations were recorded: (1) wave fronts
ing wavelet reentry consists of multiple competing wave fronts proceeding from the working myocardium to the Purkinje
that constantly undergo collisions, fractionations, and annihila- system, presumably through retrograde conduction pathways;
tions. Wave fronts meander around refractory tissue and ana- (2) focal activations that could appear in the Purkinje system or
tomic obstacles such as scar or infarct regions. When a wave front in working myocardium; and (3) activations that began outside
encounters unexcitable tissue, whether anatomic or functional in of the mapped region in the Purkinje system and proceeded to
nature, it may (1) block entirely, (2) proceed around the obstacle, initiate a new waveform in the working myocardium (Figure
or (3) fractionate and create daughter wave fronts that proceed 48-6). Coupled with the fact that the wave fronts in LDVF tend
in different directions from the original wave front (see Figure to propagate from the endocardium to the epicardium, the focal
48-2, B). Wandering wavelet reentrant pathways are short-lived activations appearing in the mapped region suggest that wave
and transient in nature. Although normal cardiac tissue sustains fronts are initiated in the Purkinje system near the endocardial
VF, the onset of VF is facilitated in diseased tissue such as border surface. The authors concluded that the Purkinje system may
zones surrounding infarcts and in tissue with anchors such as scar play a critical role in both reentrant circuits and focal activations
or anatomic boundaries that facilitate reentry. A critical mass of during LDVF.
cardiac tissue is required for the wandering wavelets to have suf- In a subsequent study, Dosdall et al performed mapping of
ficient space to collide and create daughter wavelets. The three- the same anterolateral papillary muscle section of the LV endo-
dimensional (3D) structure of the ventricles, the fiber orientation, cardium in dogs.23 They also placed plunge needles on the
and the tissue anisotropy cause the wave-front propagation to sides of the plaque. They induced VF in two groups: (1) control
become complicated and difficult to fully map through epicardial hearts and (2) hearts in which the subendocardium, including
or endocardial approaches. The hypothesis of wandering wavelet the Purkinje system, had been chemically ablated by the applica-
reentry supposes that wave fronts propagate throughout the tion of Lugol’s solution. Investigators found that although
cardiac tissue such that there is not a stable high-frequency the endocardial-epicardial rate gradient developed in the control
region that drives the activation rate. hearts, the rate gradient was eliminated in the ablated hearts.
Rogers et al. performed panoramic optical mapping of trans- The VF spontaneously terminated after 9.2 ± 3.2 minutes
membrane voltage of perfused VF in pig hearts.18 In 17 episodes in the control hearts but terminated after 4.9 ± 1.5 minutes
of mapped VF, they found that 92% of wave fronts were linked in the ablated hearts. This further supports the conclusion
through fragmentation and collision. They did not find stable that the Purkinje system plays a significant role in maintain-
rotors that lasted beyond a few seconds. Although they were able ing LDVF.
to map only the epicardial surface, investigators were able to Studies in dogs have demonstrated that periods of highly
account for most of the wave fronts without the need for intra- synchronous activity on the endocardial surface of the LV occur
mural reentry. during LDVF. Robichaux et al. inserted a 64-electrode basket
into the left ventricular (LV) endocardium and recorded electri-
cally induced VF for 10 minutes.24 They observed periods during
Purkinje Fiber Activity which focal activations in the Purkinje system spread rapidly
through the entire LV endocardium (Figure 48-7). These activa-
Involvement of the Purkinje system has largely been ignored tions resulted in rapid synchronous activation of the endocardium
when the mechanisms of VF maintenance are discussed. A and relatively long periods of inactivity between excitations.
growing body of evidence suggests that although intramural This activity deviates substantially from wandering wavelet or
reentry may be the primary driver of SDVF, the Purkinje system mother rotor reentry, during which activity would be recorded
may play a more significant role in LDVF. continuously throughout the endocardium. Another study of
Studies have shown that following 2 to 3 minutes of VF, acti- LDVF by Li et al. demonstrated that this synchronous activation
vation rates between the endocardial and epicardial surfaces of pattern originated in the Purkinje system on the endocardium,
the heart diverged.19 Although activation rates became slower for and that these wave fronts blocked at various levels as they pro-
both endocardial and epicardial tissue as VF progressed, the ceeded from the endocardium toward the epicardium.25 Research-
epicardial activation rate slowed more dramatically in epicardial ers found that this type of synchronized activity was present
tissue. This activation rate gradient develops during LDVF in approximately one-third of the time 5 to 10 minutes after the
dogs19 and humans20 but does not develop in pigs19 (Figure 48-5). onset of VF. This synchronous pattern was almost entirely elimi-
A study compared VF activation patterns between dogs and pigs nated by giving the EAD blocker, pinacidil, but the incidence of
using plunge needles distributed throughout the LV wall, and VF this Purkinje-driven activation was not altered by the DAD
was induced electrically.19 Investigators observed that although blocker flunarizine.16
an activation rate gradient emerged in dogs after a few minutes, Potential mechanisms for focal activity in the Purkinje system
this gradient did not develop in pigs. This species-dependent during LDVF and after successful defibrillation shocks are begin-
behavior together with the different Purkinje distribution in the ning to come to light. In a study on perfused rabbit ventricles,
two species led investigators to hypothesize that Purkinje fibers Maruyama et al. found spontaneous intracellular calcium levels
could be driving the later VF activation patterns in both dogs and after defibrillation on the LV endocardial surface but not on the
pigs. As is the case with humans, the Purkinje fiber system in dogs epicardium.26 Triggered activity emerged from the endocardial
is limited primarily to the endocardial surface, and in pigs these surface and was correlated with high intracellular calcium levels.
fibers arborize from the endocardial surface through the Purkinje-like potentials preceded triggered activity on the
48
6 6
5 5
2 minutes
2 minutes
4 4
3 3
2 2
1 1
6 6
5 5
6 minutes
6 minutes
4 4
3 3
2 2
1 1
6 6
5 5
10 minutes
10 minutes
4 4
3 3
2 2
1 1
Dog Pig
Figure 48-5. Plunge needle recordings from a dog and a pig during LDVF Electrode 1 was near the endocardium, and electrode 6 was near the epicardium.
In both species, the activation rate is similar on the endocardium and the epicardium at 2 minutes of VF. At 6 and 10 minutes of VF, the epicardial activation rate of the
dog slows significantly, but an activation rate gradient does not develop in the pig. Purkinje activations (arrows on the dog endocardial recordings) accompany activations
in the dog plunge needle recordings.
(Reproduced with permission from Allison JS, Qin H, Dosdall DJ, et al: The transmural activation sequence in porcine and canine left ventricle is markedly different during long-
duration ventricular fibrillation. J Cardiovasc Electrophysiol 18:1306-1312, 2007.)
0 ms 3 ms 4.5 ms
B
A
6 ms 7.5 ms 9 ms
C
C
B
10.5 ms 12 ms 13.5 ms
D
E
D
18 ms 21 ms 22.5 ms E
24 ms 25.5 ms 27 ms
150 ms
Figure 48-6. A focal activation arising from the center of the array The focal-appearing PF activation (yellow) is followed by a WVM (red) activation wave front.
The temporal derivative of individual electrode recordings shows the wave front initiating at electrode A and spreading outward toward electrodes D and E. The green
represents a nonrelated wave front.
(Reprinted with permission from Tabereaux PB, Walcott GP, Rogers JM, et al: Activation patterns of Purkinje fibers during long-duration ventricular fibrillation in an isolated
canine heart model. Circulation 116:1113-1119, 2007.)
endocardial surfaces. Spontaneous calcium release and higher focal activation, because the Purkinje system permeates the ven-
diastolic intracellular calcium levels after rapid pacing or VF may tricular wall in pigs.
lead to triggered activity in the Purkinje system.
Li et al performed high-resolution 3D mapping in pig hearts
by placing plunge needles with six electrodes at 2-mm spacing in
a 9 × 9 grid with 2-mm spacing on the anterior LV wall.27 They Conclusions
found that although reentrant circuits were common within the
mapped region during the first few minutes (Figure 48-8), by Substrate changes that facilitate focal and reentrant activity lead
3 minutes after VF the intramural reentry disappeared, and to the onset of VT. Changes in conduction velocity and disper-
the incidence of focal activations within the mapped region sion of refractoriness may lead to degeneration of VT into VF.
(activations that began de novo in the region rather than as the VF perpetuates through wandering wavelet reentry, mother rotor
result of reentry or propagation into the mapped region from reentry, and activity in the Purkinje system. The mechanisms of
outside) increased steadily throughout the first 10 minutes of VF. VF maintenance change as VF duration increases. VT and VF
Although they did not record Purkinje fibers directly, the authors remain as two of the leading causes of death. New mapping
proposed that either focal activity in the Purkinje system or techniques are leading to a greater understanding of the onset
reentrant activity involving the Purkinje system may cause this and maintenance of these arrhythmias.
VF Onset 5 Minutes VF
48
Basket Recordings
Lead II Apex
ECG RV
A B 500 ms
Figure 48-7. Recordings from an LV multi-electrode basket, an RV electrode, and a surface lead II ECG during LDVF A, Chaotic activation pat-
terns on the LV recordings immediately after VF onset are consistent with wandering wavelet reentry. B, Synchronous, nearly simultaneous activations throughout the LV
endocardium demonstrate rapidly spreading activation throughout the LV endocardium. This activity demonstrates that activations spread rapidly throughout the endo-
cardium during periods of LDVF.
(Reproduced with permission from Robichaux RP, Dosdall DJ, Osorio J, et al: Periods of highly synchronous, non-reentrant endocardial activation cycles occur during long-
duration ventricular fibrillation. J Cardiovasc Electrophysiol 21:1266-1273, 2010.)
16
Reentry incidence (% of components)
14
13
10
0
0 100 200 300 400 500 600
A VF duration (s)
45
Foci incidence (% of wave fronts)
40
35
30
25
20
15
10
Figure 48-8. Transmural 3D plunge needle mapping in pigs showed that the inci-
5 dence of intramural reentry within a mapped region went to zero in the first 2 to 3
minutes, and the incidence of focal activations increased during the first 10 minutes
0 of VF.
0 100 200 300 400 500 600
(Reproduced with permission from Li L, Jin Q, Huang J, et al: Intramural foci during long
B VF duration (s) duration fibrillation in the pig ventricle. Circ Res 102:1256-1264, 2008.)
isolated blood-perfused pig heart. Am J Physiol duration ventricular fibrillation. J Cardiovasc Elec-
References Heart Circ Physiol 293:H1617–H1628, 2007. trophysiol 18:1306–1312, 2007.
11. Huang J, Rogers JM, Killingsworth CR, et al: Evo- 20. Masse S, Farid T, Dorian P, et al: Effect of global
1. Zipes DP, Wellens HJ: Sudden cardiac death. Cir- lution of activation patterns during long-duration ischemia and reperfusion during ventricular fibril-
culation 98:2334–2351, 1998. ventricular fibrillation in dogs. Am J Physiol Heart lation in myopathic human hearts. Am J Physiol
2. Cobb LA, Fahrenbruch CE, Olsufka M, et al: Circ Physiol 286:H1193–H1200, 2004. Heart Circ Physiol 297:H1984–H1991, 2009.
Changing incidence of out-of-hospital ventricular 12. Cheng KA, Dosdall DJ, Li L, et al: Evolution of 21. Venable PW, Taylor TG, Shibayama J, et al:
fibrillation, 1980–2000. JAMA 288:3008–3013, activation patterns during long-duration ventricu- Complex structure of electrophysiological gradi-
2002. lar fibrillation in pigs. Am J Physiol Heart Circ ents emerging during long-duration ventricular
3. Myerburg RJ, Kessler KM, Castellanos A: Sudden Physiol 302:H992–H1002, 2012. fibrillation in the canine heart. Am J Physiol Heart
cardiac death: Structure, function, and time- 13. Menegazzi JJ, Ramos R, Wang HE, et al: Post- Circ Physiol 299:H1405–H1418, 2010.
dependence of risk. Circulation 85:I2–I10, 1992. resuscitation hemodynamics and relationship to 22. Tabereaux PB, Walcott GP, Rogers JM, et al: Acti-
4. Ogawa M, Morita N, Tang L, et al: Mechanisms the duration of ventricular fibrillation. Resuscita- vation patterns of Purkinje fibers during long-
of recurrent ventricular fibrillation in a rabbit tion 78:355–358, 2008. duration ventricular fibrillation in an isolated canine
model of pacing-induced heart failure. Heart 14. Geddes LA, Roeder RA, Kemeny A, et al: The heart model. Circulation 116:1113–1119, 2007.
Rhythm 6:784–792, 2009. duration of ventricular fibrillation required to 23. Dosdall DJ, Tabereaux PB, Kim JJ, et al: Chemical
5. Chua SK, Chang PC, Maruyama M, et al: Small- produce pulseless electrical activity. Am J Emerg ablation of the Purkinje system causes early termi-
conductance calcium-activated potassium channel Med 23:138–141, 2005. nation and activation rate slowing of long-duration
and recurrent ventricular fibrillation in failing 15. Valenzuela TD, Roe DJ, Nichol G, et al: Outcomes ventricular fibrillation in dogs. Am J Physiol Heart
rabbit ventricles. Circ Res 108:971–979, 2011. of rapid defibrillation by security officers after Circ Physiol 295:H883–H889, 2008.
6. Trayanova NA, Constantino J, Gurev V: Models of cardiac arrest in casinos. N Engl J Med 343:1206– 24. Robichaux RP, Dosdall DJ, Osorio J, et al: Periods
stretch-activated ventricular arrhythmias. J Elec- 1209, 2000. of highly synchronous, non-reentrant endocardial
trocardiol 43:479–485, 2010. 16. Li L, Zheng X, Dosdall DJ, et al: Long duration activation cycles occur during long-duration ven-
7. Gelzer AR, Koller ML, Otani NF, et al: Dynamic ventricular fibrillation exhibits two distinct, orga- tricular fibrillation. J Cardiovasc Electrophysiol
mechanism for initiation of ventricular fibrillation nized, endocardial activation patterns. Heart 21:1266–1273, 2010.
in vivo. Circulation 118:1123–1129, 2008. Rhythm 8(5S):S403, 2011. 25. Li L, Jin Q, Dosdall DJ, et al: Activation becomes
8. Wiggers CJ: Studies of ventricular fibrillation 17. Jalife J: Inward rectifier potassium channels control highly organized during long-duration ventricular
caused by electric shock: Cinematographic and rotor frequency in ventricular fibrillation. Heart fibrillation in canine hearts. Am J Physiol Heart
electrocardiographic observations of the natural Rhythm 6:S44–S48, 2009. Circ Physiol 298:H2046–H2053, 2010.
process in the dog’s heart: Its inhibition by potas- 18. Rogers JM, Walcott GP, Gladden JD, et al: Pan- 26. Maruyama M, Joung B, Tang L, et al: Diastolic
sium and the revival of coordinated beats by oramic optical mapping reveals continuous epicar- intracellular calcium-membrane voltage coupling
calcium. Am Heart J 5:351–365, 1930. dial reentry during ventricular fibrillation in the gain and postshock arrhythmias: Role of Purkinje
9. Thurmann M, Janney JG Jr: The diagnostic isolated swine heart. Biophys J 92:1090–1095, fibers and triggered activity. Circ Res 106:399–408,
importance of fibrillatory wave size. Circulation 2007. 2010.
25:991–994, 1962. 19. Allison JS, Qin H, Dosdall DJ, et al: The transmu- 27. Li L, Jin Q, Huang J, et al: Intramural foci during
10. Huizar JF, Warren MD, Shvedko AG, et al: Three ral activation sequence in porcine and canine left long duration fibrillation in the pig ventricle. Circ
distinct phases of vf during global ischemia in the ventricle is markedly different during long- Res 102:1256–1264, 2008.
Molecular Genetics and PART VIII
Pharmacogenomics
Genetics of Atrial Fibrillation

Steven A. Lubitz and Patrick T. Ellinor
49
CHAPTER OUTLINE
Genetic Mapping of AF
Heritability of AF 483
Genetic Mapping of AF 483 Linkage Analysis and Candidate Gene
Future Directions 488 Resequencing
Conclusions 488 The search for genetic variants underlying AF has evolved with
improvements in genetic techniques and with the recognition of
a widespread heritable basis for AF. Early investigations used
The pathophysiology of atrial fibrillation (AF) remains incom- classical linkage analysis to identify monogenic disease suscepti-
pletely characterized; however, epidemiologic studies demon- bility loci in families. Linkage analysis relies on the use of genetic
strate a heritable basis for the arrhythmia. In recent years, markers at known locations of the genome that can help deter-
appreciation of AF heritability has stimulated the search for mine which loci transmit along with disease.
the genetic underpinnings of the disease. Genetic mapping The first genetic locus for AF was described in 1997 by
techniques have identified rare mutations and common variants Brugada et al, who identified an AF susceptibility region on chro-
associated with AF. In addition to validating suspected electro- mosome 10 in a large family with autosomal dominant AF (Table
physiological mechanisms underlying AF, recent genetic discov- 49-1).8 Since this initial report, linkage analysis has repeatedly
eries have identified previously unrecognized pathways involved been used to identify genetic mutations underlying AF in a
in the development of AF. Investigators are now searching for number of large families with AF.
causal variants at many identified loci, investigating the biological Chen et al narrowed an AF susceptibility locus to a region on
mechanisms linking AF susceptibility loci to disease, and assess- chromosome 11p15 in a family with autosomal dominant AF
ing the clinical implications of recent genetic discoveries. spanning four generations.9 Investigators extended their analysis
by sequencing KCNQ1, a candidate gene in the region that
encodes a potassium channel α-subunit. In the first transmem-
brane segment of the channel, they discovered a highly conserved
Heritability of AF serine residue that was mutated to a glycine in all affected family
members. Subsequent characterization revealed that the mutant
Reports of familial clustering of AF date back to the early 1940s.1,2 protein increases the repolarizing IKs current density when
Familial AF was generally regarded as a rare condition for many expressed with the KCNE1-subunit. This gain-of-function
years thereafter. However, over the past decade a major paradigm mutation is expected to result in shortened atrial refractory
shift occurred with widespread recognition of the heritability periods and thereby promote reentry, a well-founded mechanism
underlying AF.3-7 In the community-based Framingham Heart underlying AF.51 Our group used a similar approach to map a
Study, 27% of individuals with AF had a first-degree relative with mutation in the third transmembrane domain of KCNQ1.10 The
AF confirmed by electrocardiography.7 Familial AF was associ- S209P variant demonstrated a similar gain-of-function effect
ated with a 40% increased risk of AF for another family member again implicating this as a disease susceptibility gene for AF.
over a subsequent 8-year period (Figure 49-1). The risk associ- Linkage analysis also has identified mutations in SCN5A that
ated with familial AF remained even after adjustment for estab- segregate with AF, conduction abnormalities, cardiomyopathy,
lished clinical risk factors for AF. A study from Denmark and possibly early-onset ischemic stroke.24-26 SCN5A encodes a
demonstrated that AF was more common among monozygotic sodium channel α-subunit responsible for the depolarizing INa
twins as compared with dizygotic twins, implicating a genetic current. The D1275N mutation that results in an aspartic acid
predisposition to the arrhythmia even among those raised with for asparagine substitution was independently identified by dif-
shared environmental exposures.6 ferent investigators,24-26 and associates with atrial standstill when
In numerous reports, the heritability of AF appears to be co-segregating with connexin-40 (GJA5) mutations.52
greatest among younger individuals3,5,7 and those without struc- In yet another large multigenerational family with prevalent
tural heart disease.3,4 Premature familial AF, or that occurring in AF, Hodgson-Zingman et al mapped a frameshift mutation to
family members ≤65 years of age, was associated with a 2-fold NPPA, which encodes atrial natriuretic peptide.47 The two–base-
increase in the risk of AF compared with individuals without pair deletion eliminates a stop codon, resulting in an additional
familial AF in the Framingham Heart Study.7 Nevertheless, data 12 amino acids at the carboxy terminus of the mature 28-residue
from Framingham provide evidence that the heritability of AF is long atrial natriuretic peptide (ANP) protein. Ex vivo rabbit
present in the elderly as well.7 hearts bathed in the mutant peptide demonstrated significantly
483
484 MOLECULAR GENETICS AND PHARMACOGENOMICS
7 and fibrillation. This mechanism has been described as “atrial

Cumulative incidence (%)
6
torsades,” and observations from this report underscore the het-
erogeneity of mechanisms that can lead to AF.
5 Familial AF Candidate gene association studies in patients with AF also
4
have identified variants in genes encoding sodium channel sub-
units. Both loss-of-function27 and gain-of-function28,29 mutations
3 in SCN5A have been identified in patients with lone AF. Muta-
2 tions in sodium channel β-subunits, SCN1B and SCN2B, have
No familial AF
been described that decrease sodium current amplitude and alter
1 channel gating kinetics when coexpressed with SCN5A.30 These
0 mutations are speculated to predispose to AF through shortening
of the atrial action potential duration or by conduction slowing,
0 2 4 6 8
both of which may facilitate reentry.
Time to AF (years) Mutations in GJA5, which encodes the gap junction
Figure 49-1. The risk of atrial fibrillation is increased in individuals with a first- connexin-40, have been described in 4 of 15 patients with AF
degree relative with antecedent atrial fibrillation in the Framingham Heart Study. screened in one candidate gene association study.32 The GJA5
The estimated 8-year risk of atrial fibrillation (AF) is increased by 40% among indi- mutations in three of the four patients were present in cardiac
viduals with a first-degree relative with atrial fibrillation as compared with those tissue but not in lymphocyte specimens, suggesting that these
without familial atrial fibrillation. mutations were acquired or somatic rather than inherited or
(Adapted with permission from Lubitz et al.7)
germline mutations. Other candidate gene association studies
have linked common genetic variation in GJA5 to AF.31,33
Non-ion channel variants also have been discovered in candi-
date gene association studies. In a recent large-scale cardiovascu-
shortened atrial effective refractory periods as compared with lar candidate gene association study, an intronic variant in the
those bathed with a wild-type atrial natriuretic peptide, again IL6R receptor was associated with AF.48 Mutations in LMNA,
consistent with reentry as a pathogenic model for AF. which encodes the nuclear envelope proteins lamin A and lamin
In aggregate, linkage analysis has implicated potassium C, have been identified in patients with cardiomyopathy and
(KCNQ19,10) and sodium (SCN5A24-26) channel mutations, a AF.36,37 Mutations in these proteins underlie a diverse spectrum
nuclear envelope protein (NUP15540,41), NPPA,47 and loci on of disorders that include Emery-Dreifuss syndrome, Charcot-
chromosomes 649 and 10.8,50 Newer genetic techniques such as Marie-Tooth disease, and premature aging syndromes. Never-
exome and whole genome sequencing have emerged in recent theless, LMNA mutations appear to be rare causes of AF.38
years that make mapping of these families even more efficient. Recently, loss-of-function variants in ANK2 have been identi-
However, the large, multigenerational families needed for linkage fied in kindreds with early-onset AF, in which family members
analysis remain rare. More often AF is observed in smaller fami- had frequently progressed to permanent AF.34 ANK2 encodes
lies and thus is it is often hard to establish the causality of appar- ankyrin-B, which was previously implicated in the long QT syn-
ently disease-causing mutations with such limited genetic drome in a family in which a substantial proportion of members
information. was affected by AF.35,53 Loss-of-function mutations in ANK2
appear to decrease the expression and trafficking of CaV 1.3
channels, resulting in decreased ICa,L current.34 Heterozygous
Candidate Gene Association Studies ANK2 knockout mice exhibit a lack of discrete P waves and
increased susceptibility to pacing-induced atrial arrhythmias,
In contrast to linkage mapping, in which inferences about recom- including AF.
bination are made on the basis of segregation of markers in a Similar to linkage mapping, candidate gene association studies
pedigree, investigators have also selected and screened candidate have provided insight into the pathogenic mechanisms of AF.
genes for mutations in cohorts of patients with and without AF. These studies demonstrate that mutations in ion channels and
As an extrapolation of initial studies on KCNQ1, numerous addi- genes hypothesized to be involved in AF are rare.54-56 Further-
tional candidate gene association studies have focused on cardiac more, like linkage analysis, candidate gene association studies
ion channels (see Table 49-1). have generally identified rare mutations underlying monogenic
In one report, genetic variants in KCNQ1 were identified that forms of AF that are private to individual families. The validity
predisposed to AF in a stretch-sensitive fashion,12 illustrating the and generalizability of more common variants identified in some
potential for a concealed predisposition for AF to be elicited by candidate gene association studies have been called into question
an acquired exposure (e.g., valvular disease, heart failure). In owing to the lack of replication of many associations.57
another report implicating KCNQ1, a mutation was identified in Nevertheless, the potential importance of such rare variants
a patient with long QT syndrome.13 Diverging effects on IKs were was underscored in a recent analysis, in which sequencing for
observed depending on whether the mutant protein was expressed potassium channel mutations in a sample of 80 probands with AF
with KCNE1 or KCNE2 β-subunits, each of which are differen- and 240 controls demonstrated an excess number of nonsynony-
tially expressed in the atria and ventricles, leading the authors to mous variants among those with AF.58 Despite their rarity, model-
speculate that the mutation was the cause of both long QT syn- ing of the effects of multiple potassium channel mutations
drome and AF. Gain-of-function mutations in both KCNQ111 and suggested that the combination of such variants could result in
KCNH220 have been discovered in AF in the context of short QT substantial changes in action potential duration and dispersion of
syndrome. repolarization, potentially establishing a substrate for AF.
In contrast to the enhanced atrial repolarization mechanism
invoked by most discovered potassium channel mutations, a
nucleotide substitution resulting in a premature stop codon in Genome-Wide Association Studies
KCNA5 has been described that manifested a loss-of-function of
the Kv1.5 channel protein.23 The mutation effectively abolished Over the past decade, systematic efforts to examine human
IKur and prolonged the action potential duration, increasing sus- genetic variation such as the international HapMap project59 have
ceptibility to early afterdepolarization-induced atrial tachycardia resulted in appreciation of genetic diversity in different ancestral
Genetics of Atrial Fibrillation 485
Table 49-1. Atrial Fibrillation Susceptibility Loci Identified by Linkage Analysis or Candidate Gene Association Studies
Gene or Locus Gene Product Mapping Method Presumed Mechanism Frequency 49

Potassium Channels
KCNQ1 α-subunit, IKs Linkage9,10 Enhanced repolarization (↑IKs) Rare
Screening11-13
KCNE1 β-subunit, IKs Screening14 Enhanced repolarization (↑IKs) Rare
15-17
Screening Unknown Common
KCNE2 β-subunit, IKs Screening18 Enhanced repolarization (↑IKs) Rare
KCNE5 β-subunit, IKs Screening 19
Enhanced repolarization (↑IKs) Rare
KCNH2 Kv11.1, IKr Screening20 Enhanced repolarization (↑IKr) Rare
Screening21 Unknown Common
KCNJ2 Kir 2.1, IK1 Screening22 Enhanced repolarization (↑IK1) Rare
23
KCNA5 Kv1.5 channel, IKur Screening Delayed repolarization and afterdepolarization Rare
(↓IKur)
Sodium Channel
SCN5A α-subunit, INa Linkage24-26 Unknown Rare
27-29
Screening Hyperpolarizing shift in inactivation (loss-of- Rare
function)27
Depolarizing shift in inactivation (gain-of-
function)28,29
SCN1B β-subunit, INa Screening30 ↓INa current and altered channel gating Rare
SCN2B β-subunit, INa Screening 30
↓INa current and altered channel gating Rare
Other Ion Channel/Ion Channel–Related
GJA5 Connexin40 Screening31-33 Impaired cellular transport and intercellular Common
electrical coupling, increased dispersion of
refractoriness
ANK2 Ankyrin-B Screening34 Loss-of-function reduces expression and Rare
Linkage35 membrane targeting of Cav1.3 (↓ICa,L)
Non–Ion Channel
LMNA Lamin A/C Screening36-38 Disruption of nuclear function or altered Rare
Linkage39 interaction with cytoplasmic proteins
NUP155 Nucleoporin Linkage40,41 Reduced nuclear membrane permeability, Rare
enhanced repolarization
AGT Angiotensinogen Screening42,43 Unknown Common
ACE Angiotensin-converting enzyme Screening44-46 Insertion/deletion, unknown mechanism Common
47
NPPA Natriuretic peptide precursor A Linkage Mutant peptide, enhanced repolarization Rare
IL6R Interleukin-6 receptor Candidate gene chip48 Unknown Common
Mapped Loci Without Causative Gene
6q14-q16 — Linkage49 Unknown Rare
50
10p11-q21 — Linkage Unknown Rare
10q22-q24 — Linkage8 Unknown Rare
groups. The HapMap project revealed the presence of about 10 positions were tested for association with human traits using
million common genetic variants occurring with a frequency of standard epidemiologic techniques. Unlike linkage analysis,
5% or greater in the general population, most of which were genome-wide association testing does not require multigenera-
single-nucleotide polymorphisms (SNPs). Such efforts provided tional cohorts for assessment of variant transmission. In contrast
a reference against which observed genetic variation could be to candidate gene association studies, the presence of genome-
compared in cohorts of individuals assembled for investigation. wide genotyping allowed investigators to efficiently test variants
High-density chips were developed that allowed the simultane- in genes or genetic regions that were not previously suspected of
ous genotyping of hundreds of thousands of SNPs across the being involved in AF pathogenesis.
genome, enabling efficient genetic profiling of individuals. Unique study design and interpretation considerations are
Genome-wide associations emerged in which genotypes at applicable to genome-wide association studies. First, owing to the
each of hundreds of thousands to millions of known SNP massive number of tests performed in each genome-wide
association study, stringent significance thresholds (e.g., P < 5 × first genome-wide association study for AF in 2007.61 Review of
10−8) are used to guard against false-positive test results that occur this discovery provides an example of the power of genome-wide
by chance with multiple hypothesis testing. Second, common association studies and the follow-up work involved in identifying
genetic variants associated with disease are expected to confer causal elements at a particular susceptibility locus.
small or modest disease risks, in contrast to rare and large-effect Investigators first identified genetic variants at the chromo-
genetic variants observed in monogenic forms of AF. Third, the some 4q25 locus that were highly associated with AF in approxi-
stringent significance thresholds used and the modest relative mately 550 individuals with and 4500 without AF from Iceland.61
risks expected in genome-wide association studies necessitate The association between SNPs at chromosome 4q25 and AF has
extremely large sample sizes to adequately power genetic variant been widely replicated in samples of European,61,88-91 Han
discovery. By convention genome-wide associations require rep- Chinese,61,92 African American,48,93 and Japanese65 ancestry, dem-
lication in independent samples to claim validity, making these onstrating the fundamental role of this locus in the pathophysiol-
studies some of the largest human scientific experiments ever ogy of AF.
conducted. Diverse international collaborations have formed The mechanisms linking chromosome 4q25 to AF remain
over the past several years and have contributed to the success of unclear. The top variants associated with AF lie in a region of the
genome-wide association studies. As of 2011, 1617 genome-wide genome with no known genes, and are approximately 150 thou-
associations encompassing 249 human traits had been sand base-pairs away from the nearest gene at the locus, the
published.60 paired-like homeodomain 2 transcription factor (PITX2). Among
Genome-wide association studies have identified nine genomic the genes at 4q25, PITX2 is a plausible candidate for involvement
regions associated with AF (Table 49-2 and Figure 49-2).61-65 in AF. PITX2 encodes a transcription factor that is involved in
Association signals at these loci often span tens of thousands of cardiac and pulmonary development. Mice deficient in one
base pairs and encompass both genic and nongenic regions. Asso- isoform of the protein that is expressed in the heart, Pitx2c, do
ciated SNPs are not interpreted as causal genetic variants but not form myocardial sleeves in the pulmonary veins.71 Given the
rather as markers that tag causal genetic elements located nearby. importance of pulmonary vein ectopic foci in the pathogenesis of
To date, causal variants for AF have not been discovered. Never- AF,94 the relations between PITX2 and pulmonary vein myocar-
theless, candidate genes at these loci frequently exist, and have dial cell development are particularly intriguing. Additionally,
implicated transcription factors involved in cardiac and pulmo- Pitx2c is responsible for suppressing default formation of a sinus
nary development, ion channels, and cell signaling molecules node in the developing left atrial region.72,73
(Figure 49-3). Functional work has begun to explore the mecha- Decreased expression of Pitx2c was noted in patients with AF
nisms by which these genomic regions potentially underlie AF. as compared with those with sinus rhythm,74,75 suggesting that
The most significantly associated variants associated with AF Pitx2c deficiency or downregulation may predispose to AF.
exist at the chromosome 4q25 locus, and were reported in the Increased inducibility of AF was observed in heterozygous Pitx2c
Table 49-2. Atrial Fibrillation Susceptibility Loci Identified by Genome-Wide Association Study
Chromosome Candidate Gene(s) Factors Supporting Candidate Gene

64,65
1q21 KCNN3 • Inhibition prolongs atrial effective refractory period66,67
• Inhibition reduces inducibility and duration of pacing-induced atrial fibrillation (AF) in animal models66,67
1q2465 PRRX1 • PRRX1 knockout results in abnormalities of great vessel development in a mouse model68
• PRRX1 knockout impairs pulmonary vasculature development69
• PRRX1 affects vascular smooth muscle cell differentiation70
4q2561-65 PITX2 • Necessary for pulmonary vein myocardial sleeve formation71
• Suppresses default formation of left atrial sinus node72,73
• Decreased expression in patients with AF74,75
• Deficiency in mice associated with increased susceptibility to atrial arrhythmias75
• Expression regulates potassium and sodium channel expression in HL-1 atrial myocytes74
• PITX2 affects vascular smooth muscle cell differentiation70,76
7q3165 CAV1 • Encodes a cellular membrane protein involved in signal transduction77
• Knockout associated with dilated cardiomyopathy78 and cardiac hypertrophy79
• Directly modulates endothelial cell KCNN3 trafficking80 and indirectly modulates cardiac KCNH2
trafficking81
9q2265 C9orf3 • Aminopeptidase-O
• Unclear relation to AF
10q2265 SYNPO2L/ MYOZ1 • MYOZ1 is expressed in cardiomyocytes and localizes to Z-disc82
65
14q23 SYNE2 • Located in sarcomere and involved in maintaining nuclear structural integrity; binds lamin83
• Mutations underlie some cases of Emery-Dreifuss muscular dystrophy, which is characterized by muscle
atrophy, cardiomyopathy, and cardiac conduction abnormalities
15q2465 HCN4 • Highly expressed in sinoatrial node
• HCN4 channel is responsible for the hyperpolarization-activated inward phase 4 depolarizing current
(cardiac pacemaker or funny current, If )84
• Mutations in HCN4 are associated with sick sinus syndrome and bradycardia85
16q2265 ZFHX3 • Partial deletion of 16q22 associated with congenital heart disease86
• Variants at 16q22 associated with Kawasaki disease87
PITX2
60 49
–log10 (p)
40
20 ZFHX3
KCNN3
PRRX1 CAV1 C9orf3 SYNPO2L SYNE2
HCN4
1 2 3 4 5 6 7 8 9 10 11 12 13 15 17 19 22
Chromosome
Figure 49-2. Nine Genetic Loci Associated With Atrial Fibrillation Discovered in Genome-Wide Association Studies The chromosome and position
of each of 2.2 million tested single-nucleotide polymorphisms are plotted on the x-axis, and the −log10 (P value) is plotted on the y-axis. Peaks above the dotted horizontal
line are significantly associated with atrial fibrillation at P < 5 × 10−8. The closest gene at each associated locus is indicated.
(Adapted with permission from Ellinor et al.65)
IKs KCNQ1, KCNE1, KCNE2, KCNE5

INa SCN5A, SCN1B, SCN2B IKr KCNH2
INa SCN5A IK1 KCNJ2
IKur KCNA5
? ?
CAV1
ICa,L ANK2
Cytoplasmic actin
GJA5
SYNE2 Electrical
SYNPO2L/MYOZ1 coupling
LMNA
PITX2 ? PRRX1 ? ZFHX3 NUP155

Permeability
Sarcomere
HCN4 KCNN3 NPPA

Mutant ANP
Figure 49-3. Genes Implicated in the Pathogenesis of Atrial Fibrillation A depiction of a myocyte indicates the putative location and function of each impli-
cated gene product. Genes associated with atrial fibrillation mapped via linkage or candidate gene testing are depicted in black. Potential candidate genes associated with
atrial fibrillation mapped via genome-wide association studies are depicted in blue.
(Figure courtesy Saagar Mahida, MB, ChB.)
knockout mice subjected to programmed atrial stimulation.75 activity that has been oberved,95-97 it is possible that variants exert
Pitx2c expression also appears to regulate expression of regulatory effects on distant gene targets. Indeed, independent
certain sodium and potassium channels in HL-1 atrial cardio- susceptibility signals for AF at the chromosome 4q25 locus have
myocytes, demonstrating a direct role for Pitx2c in cellular been identified in conserved noncoding regions of chromosome
electrophysiology.74 4q25.98 The number of risk alleles at three SNPs tagging these
How genetic variants at chromosome 4q25 associate with AF independent signals correlated with increased risk of AF in a
is unknown, but in light of evidence of long-range enhancer sample of about 6000 individuals with AF and 32,000 without AF.
Consideration of genotypes at these three SNPs identified a prevalence between races.102 It is important to note that causal
subset of 12% of individuals in the sample who had at least a variants at established loci remain undiscovered.
2-fold increased risk for AF, and 1% with at least a 6-fold Application of technological advancements such as next gen-
increased risk for AF relative to those with the most common eration exome and whole genome sequencing is under way in
genotypes at each of the three SNPs. large AF cohorts, and may provide a greater understanding of the
Despite data implicating PITX2 in the pathogenesis of AF, genetic architecture related to AF at associated loci. Functional
genetic variants in PITX2 have not been identified through characterization of genetic variation at AF loci will involve a
sequencing in cohorts of patients with AF.99 Mutations in PITX2 combination of bioinformatics and experimentation in cellular
are found in Axenfeld-Rieger syndrome100 and Peters anomaly,101 and animal models systems. Such follow-up work has begun, but
which are characterized by ocular abnormalities but not AF. it will take time to reveal the mechanisms by which genetic varia-
After the discovery of AF-associated variants on chromosome tion associates with AF.
4q25, genome-wide association studies demonstrated additional The potential clinical applications of genetic variants associ-
associations on chromosome 16q2262,63 and 1q21.64 In 2012, the ated with AF remain largely unexplored. Investigation into the
largest meta-analysis of genome-wide association studies of AF predictive utility of AF genetic variants for incident arrhythmia
was published, which included 6707 individuals with and 52,426 and AF-related morbidities such as stroke and heart failure is
individuals without AF.65 In addition to the three previously iden- warranted. Assessment of pharmacogenetic interactions between
tified AF susceptibility loci, six novel loci for AF were discov- AF-related therapeutics and genetic variants is necessary to deter-
ered.65 Although causal variants at these loci have not been mine whether use of such information can enhance efficacy of
identified, candidate genes exist at most of these loci and are care and minimize adverse effects.
summarized in Table 49-2.
Conclusions
Future Directions
Gains in our understanding about the genetic underpinnings of
Understanding of the genetic basis of AF has improved rapidly AF have revealed new potential biological pathways involved in
in the past several years, yet numerous knowledge gaps still exist. the pathogenesis of the arrhythmia. The field has advanced
The heritability underlying AF is incompletely explained by top rapidly over the past several years and continues to evolve at a
variants at genetic loci.7 Despite associations between variation brisk pace. Continued genetic variant discovery and biological
at some loci and AF across different ancestral groups, significant characterization of associated loci have the potential to reveal
undiscovered genetic differences may underlie differences in AF new therapeutic targets for the management of patients with AF.
12. Otway R, Vandenberg JI, Guo G, et al: Stretch- from a systematic candidate gene-based analysis
References sensitive KCNQ1 mutation: A link between of KCNH2 (HERG). Eur Heart J 29:907–914,
genetic and environmental factors in the patho- 2008.
1. Levy RL: Paroxysmal auricular fibrillation and genesis of atrial fibrillation? J Am Coll Cardiol 22. Xia M, Jin Q, Bendahhou S, et al: A Kir2.1 gain-
flutter without signs of organic cardiac disease in 49:578–586, 2007. of-function mutation underlies familial atrial
two brothers. J Mt Sinai Hosp 8:765–770, 13. Lundby A, Ravn LS, Svendsen JH, et al: KCNQ1 fibrillation. Biochem Biophys Res Commun
1942. mutation Q147R is associated with atrial fibrilla- 332:1012–1019, 2005.
2. Wolff L: Familial auricular fibrillation. N Engl J tion and prolonged QT interval. Heart Rhythm 23. Olson TM, Alekseev AE, Liu XK, et al: Kv1.5
Med 229:396–398, 1993. 4:1532–1541, 2007. channelopathy due to KCNA5 loss-of-function
3. Fox CS, Parise H, D’Agostino RB, Sr, et al: 14. Olesen MS, Bentzen BH, Nielsen JB, et al: Muta- mutation causes human atrial fibrillation. Hum
Parental atrial fibrillation as a risk factor for atrial tions in the potassium channel subunit KCNE1 Mol Genet 15:2185–2191, 2006.
fibrillation in offspring. JAMA 291:2851–2855, are associated with early-onset familial atrial 24. McNair WP, Ku L, Taylor MR, et al: SCN5A
2004. fibrillation. BMC Med Genet 13:24, 2012. mutation associated with dilated cardiomyopathy,
4. Ellinor PT, Yoerger DM, Ruskin JN, et al: Famil- 15. Lai LP, Su MJ, Yeh HM, et al: Association of the conduction disorder, and arrhythmia. Circulation
ial aggregation in lone atrial fibrillation. Hum human minK gene 38G allele with atrial fibrilla- 110:2163–2167, 2004.
Genet 118:179–184, 2005. tion: Evidence of possible genetic control on the 25. Olson TM, Michels VV, Ballew JD, et al: Sodium
5. Arnar DO, Thorvaldsson S, Manolio TA, et al: pathogenesis of atrial fibrillation. Am Heart J channel mutations and susceptibility to heart
Familial aggregation of atrial fibrillation in 144:485–490, 2002. failure and atrial fibrillation. JAMA 293:447–454,
Iceland. Eur Heart J 27:708–712, 2006. 16. Fatini C, Sticchi E, Genuardi M, et al: Analysis of 2005.
6. Christophersen IE, Ravn LS, Budtz-Joergensen minK and eNOS genes as candidate loci for pre- 26. Laitinen-Forsblom PJ, Makynen P, Makynen H,
E, et al: Familial aggregation of atrial fibrillation: disposition to non-valvular atrial fibrillation. Eur et al: SCN5A mutation associated with cardiac
A study in Danish twins. Circ Arrhythm Electro- Heart J 27:1712–1718, 2006. conduction defect and atrial arrhythmias. J Car-
physiol 2:378–383, 2009. 17. Prystupa A, Dzida G, Myslinski W, et al: MinK diovasc Electrophysiol 17:480–485, 2006.
7. Lubitz SA, Yin X, Fontes JD, et al: Association gene polymorphism in the pathogenesis of lone 27. Ellinor PT, Nam EG, Shea MA, et al: Cardiac
between familial atrial fibrillation and risk of new- atrial fibrillation. Kardiol Pol 64:1205–1211, sodium channel mutation in atrial fibrillation.
onset atrial fibrillation. JAMA 304:2263–2269, 2006; discussion 1212-1203. Heart Rhythm 5:99–105, 2008.
2010. 18. Yang Y, Xia M, Jin Q, et al: Identification of a 28. Makiyama T, Akao M, Shizuta S, et al: A novel
8. Brugada R, Tapscott T, Czernuszewicz GZ, et al: KCNE2 gain-of-function mutation in patients SCN5A gain-of-function mutation M1875T asso-
Identification of a genetic locus for familial atrial with familial atrial fibrillation. Am J Hum Genet ciated with familial atrial fibrillation. J Am Coll
fibrillation. N Engl J Med 336:905–911, 1997. 75:899–905, 2004. Cardiol 52:1326–1334, 2008.
9. Chen YH, Xu SJ, Bendahhou S, et al: KCNQ1 19. Ravn LS, Aizawa Y, Pollevick GD, et al: Gain of 29. Li Q, Huang H, Liu G, et al: Gain-of-function
gain-of-function mutation in familial atrial fibril- function in IKs secondary to a mutation in mutation of Nav1.5 in atrial fibrillation enhances
lation. Science 299:251–254, 2003. KCNE5 associated with atrial fibrillation. Heart cellular excitability and lowers the threshold for
10. Das S, Makino S, Melman YF, et al: Mutation in Rhythm 5:427–435, 2008. action potential firing. Biochem Biophys Res
the S3 segment of KCNQ1 results in familial lone 20. Hong K, Bjerregaard P, Gussak I, et al: Short QT Commun 380:132–137, 2009.
atrial fibrillation. Heart Rhythm 6:1146–1153, syndrome and atrial fibrillation caused by muta- 30. Watanabe H, Darbar D, Kaiser DW, et al: Muta-
2009. tion in KCNH2. J Cardiovasc Electrophysiol tions in sodium channel beta1- and beta2-subunits
11. Hong K, Piper DR, Diaz-Valdecantos A, et al: De 16:394–396, 2005. associated with atrial fibrillation. Circ Arrhythm
novo KCNQ1 mutation responsible for atrial 21. Sinner MF, Pfeufer A, Akyol M, et al: The non- Electrophysiol 2:268–275, 2009.
fibrillation and short QT syndrome in utero. Car- synonymous coding IKr-channel variant KCNH2- 31. Firouzi M, Ramanna H, Kok B, et al: Association
diovasc Res 68:433–440, 2005. K897T is associated with atrial fibrillation: Results of human connexin40 gene polymorphisms
with atrial vulnerability as a risk factor for idio- 51. Nattel S: New ideas about atrial fibrillation 50 73. Mommersteeg MT, Hoogaars WM, Prall OW, et
pathic atrial fibrillation. Circ Res 95:e29–e33, years on. Nature 415:219–226, 2002. al: Molecular pathway for the localized formation
32.
2004.
Gollob MH, Jones DL, Krahn AD, et al: Somatic
mutations in the connexin 40 gene (GJA5) in
52. Groenewegen WA, Firouzi M, Bezzina CR, et al:
A cardiac sodium channel mutation cosegregates
with a rare connexin40 genotype in familial atrial
of the sinoatrial node. Circ Res 100:354–362,
2007.
74. Chinchilla A, Daimi H, Lozano-Velasco E, et al:
49
atrial fibrillation. N Engl J Med 354:2677–2688, standstill. Circ Res 92:14–22, 2003. PITX2 insufficiency leads to atrial electrical and
2006. 53. Mohler PJ, Schott J-J, Gramolini AO, et al: structural remodeling linked to arrhythmogene-
33. Juang JM, Chern YR, Tsai CT, et al: The associa- Ankyrin-B mutation causes type 4 long-QT sis. Circ Cardiovasc Genet 4:269–279, 2011.
tion of human connexin 40 genetic polymor- cardiac arrhythmia and sudden cardiac death. 75. Kirchhof P, Kahr PC, Kaese S, et al: PITX2c is
phisms with atrial fibrillation. Int J Cardiol Nature 421:634–639, 2003. expressed in the adult left atrium, and reducing
116:107–112, 2007. 54. Ellinor PT, MacRae CA: Ion channel mutations Pitx2c expression promotes atrial fibrillation
34. Cunha SR, Hund TJ, Hashemi S, et al: Defects in AF: Signal or noise? Heart Rhythm 5:436–437, inducibility and complex changes in gene expres-
in ankyrin-based membrane protein targeting 2008. sion. Circ Cardiovasc Genet 4:123–133, 2011.
pathways underlie atrial fibrillation. Circulation 55. Ellinor PT, Moore RK, Patton KK, et al: Muta- 76. Yoshida T, Hoofnagle MH, Owens GK: Myocar-
124:1212–1222, 2011. tions in the long QT gene, KCNQ1, are an din and Prx1 contribute to angiotensin II-induced
35. Schott JJ, Charpentier F, Peltier S, et al: Mapping uncommon cause of atrial fibrillation. Heart expression of smooth muscle alpha-actin. Circ
of a gene for long QT syndrome to chromosome 90:1487–1488, 2004. Res 94:1075–1082, 2004.
4q25-27. Am J Hum Genet 57:1114–1122, 56. Ellinor PT, Petrov-Kondratov VI, Zakharova E, 77. Gratton JP, Bernatchez P, Sessa WC: Caveolae
1995. et al: Potassium channel gene mutations rarely and caveolins in the cardiovascular system. Circ
36. Fatkin D, MacRae C, Sasaki T, et al: Missense cause atrial fibrillation. BMC Med Genet 7:70, Res 94:1408–1417, 2004.
mutations in the rod domain of the lamin A/C 2006. 78. Zhao YY, Liu Y, Stan RV, et al: Defects in
gene as causes of dilated cardiomyopathy and 57. Sinner MF, Lubitz SA, Pfeufer A, et al: Lack of caveolin-1 cause dilated cardiomyopathy and pul-
conduction-system disease. N Engl J Med replication in polymorphisms reported to be asso- monary hypertension in knockout mice. Proc
341:1715–1724, 1999. ciated with atrial fibrillation. Heart Rhythm Natl Acad Sci U S A 99:11375–11380, 2002.
37. Sebillon P, Bouchier C, Bidot LD, et al: Expand- 8:403–409, 2011. 79. Cohen AW, Park DS, Woodman SE, et al:
ing the phenotype of LMNA mutations in dilated 58. Mann SA, Otway R, Guo G, et al: Epistatic effects Caveolin-1 null mice develop cardiac hypertrophy
cardiomyopathy and functional consequences of of potassium channel variation on cardiac repolar- with hyperactivation of p42/44 MAP kinase in
these mutations. J Med Genet 40:560–567, 2003. ization and atrial fibrillation risk. J Am Coll cardiac fibroblasts. Am J Physiol Cell Physiol
38. Brauch KM, Chen LY, Olson TM: Comprehen- Cardiol 59:1017–1025, 2012. 284:C457–C474, 2003.
sive mutation scanning of LMNA in 268 patients 59. The International HapMap Project. Nature 80. Lin MT, Adelman JP, Maylie J: Modulation of
with lone atrial fibrillation. Am J Cardiol 426:789–796, 2003. endothelial SK3 channel activity by Ca2+-depen-
103:1426–1428, 2009. 60. Hindorff LA, Junkins HA, Mehta JP, et al: A dent caveolar trafficking. Am J Physiol Cell
39. Pan H, Richards AA, Zhu X, et al: A novel muta- catalog of published genome-wide association Physiol 303:C318–C327, 2012.
tion in LAMIN A/C is associated with isolated studies. http://www.genome.gov/gwastudies. 81. Lin J, Lin S, Choy PC, et al: The regulation of
early-onset atrial fibrillation and progressive Accessed August 2, 2011. the cardiac potassium channel (HERG) by
atrioventricular block followed by cardiomyopa- 61. Gudbjartsson DF, Arnar DO, Helgadottir A, et al: caveolin-1. Biochem Cell Biol 86:405–415, 2008.
thy and sudden cardiac death. Heart Rhythm Variants conferring risk of atrial fibrillation on 82. Frey N, Olson EN: Calsarcin-3, a novel skeletal
6:707–710, 2009. chromosome 4q25. Nature 448:353–357, 2007. muscle-specific member of the calsarcin family,
40. Oberti C, Wang L, Li L, et al: Genome-wide 62. Benjamin EJ, Rice KM, Arking DE, et al: Variants interacts with multiple Z-disc proteins. J Biol
linkage scan identifies a novel genetic locus on in ZFHX3 are associated with atrial fibrillation in Chem 277:13998–14004, 2002.
chromosome 5p13 for neonatal atrial fibrillation individuals of European ancestry. Nat Genet 83. Zhang Q, Ragnauth CD, Skepper JN, et al:
associated with sudden death and variable cardio- 41:879–881, 2009. Nesprin-2 is a multi-isomeric protein that binds
myopathy. Circulation 110:3753–3759, 2004. 63. Gudbjartsson DF, Holm H, Gretarsdottir S, et al: lamin and emerin at the nuclear envelope and
41. Zhang X, Chen S, Yoo S, et al: Mutation in A sequence variant in ZFHX3 on 16q22 associates forms a subcellular network in skeletal muscle.
nuclear pore component NUP155 leads to atrial with atrial fibrillation and ischemic stroke. Nat J Cell Sci 118(Pt 4):673–687, 2005.
fibrillation and early sudden cardiac death. Cell Genet 41:876–878, 2009. 84. DiFrancesco D: The role of the funny current in
135:1017–1027, 2008. 64. Ellinor PT, Lunetta KL, Glazer NL, et al: pacemaker activity. Circ Res 106:434–446, 2010.
42. Tsai CT, Lai LP, Lin JL, et al: Renin-angiotensin Common variants in KCNN3 are associated with 85. Dobrzynski H, Boyett MR, Anderson RH: New
system gene polymorphisms and atrial fibrillation. lone atrial fibrillation. Nat Genet 42:240–244, insights into pacemaker activity: Promoting
Circulation 109:1640–1646, 2004. 2010. understanding of sick sinus syndrome. Circulation
43. Wang QS, Li YG, Chen XD, et al: Angiotensino- 65. Ellinor PT, Lunetta KL, Albert CM, et al: 115:1921–1932, 2007.
gen polymorphisms and acquired atrial fibrillation Meta-analysis identifies six new susceptibility 86. Yamamoto T, Dowa Y, Ueda H, et al: Tetralogy
in Chinese. J Electrocardiol 43:373–377, 2010. loci for atrial fibrillation. Nat Genet 44:670–675, of Fallot associated with pulmonary atresia and
44. Bedi M, McNamara D, London B, et al: Genetic 2012. major aortopulmonary collateral arteries in a
susceptibility to atrial fibrillation in patients with 66. Diness JG, Sorensen US, Nissen JD, et al: Inhibi- patient with interstitial deletion of 16q21-q22.1.
congestive heart failure. Heart Rhythm 3:808– tion of small-conductance Ca2+-activated K+ Am J Med Genet A 146A:1575–1580, 2008.
812, 2006. channels terminates and protects against atrial 87. Burgner D, Davila S, Breunis WB, et al: A
45. Fatini C, Sticchi E, Gensini F, et al: Lone and fibrillation. Circ Arrhythm Electrophysiol 3:380– genome-wide association study identifies novel
secondary nonvalvular atrial fibrillation: Role of a 390, 2010. and functionally related susceptibility loci for
genetic susceptibility. Int J Cardiol 120:59–65, 67. Skibsbye L, Diness JG, Sorensen US, et al: The Kawasaki disease. PLoS Genet 5:e1000319, 2009.
2007. duration of pacing-induced atrial fibrillation is 88. Kaab S, Darbar D, van Noord C, et al: Large scale
46. Watanabe H, Kaiser DW, Makino S, et al: ACE reduced in vivo by inhibition of small conduc- replication and meta-analysis of variants on chro-
I/D polymorphism associated with abnormal tance Ca(2+)-activated K(+) channels. J Cardiovasc mosome 4q25 associated with atrial fibrillation.
atrial and atrioventricular conduction in lone Pharmacol 57:672–681, 2011. Eur Heart J Apr 30:813–819, 2009.
atrial fibrillation and structural heart disease: 68. Bergwerff M, Gittenberger-de Groot AC, Wisse 89. Body SC, Collard CD, Shernan SK, et al:
Implications for electrical remodeling. Heart LJ, et al: Loss of function of the Prx1 and Prx2 Variation in the 4q25 chromosomal locus predicts
Rhythm 6:1327–1332, 2009. homeobox genes alters architecture of the great atrial fibrillation after coronary artery bypass
47. Hodgson-Zingman DM, Karst ML, Zingman LV, elastic arteries and ductus arteriosus. Virchows graft surgery. Circ Cardiovasc Genet 2:499–506,
et al: Atrial natriuretic peptide frameshift muta- Arch 436:12–19, 2000. 2009.
tion in familial atrial fibrillation. N Engl J Med 69. Ihida-Stansbury K, McKean DM, Gebb SA, et al: 90. Kiliszek M, Franaszczyk M, Kozluk E, et al: Asso-
359:158–165, 2008. Paired-related homeobox gene Prx1 is required ciation between variants on chromosome 4q25,
48. Schnabel RB, Kerr KF, Lubitz SA, et al: Large- for pulmonary vascular development. Circ Res 16q22 and 1q21 and atrial fibrillation in the Polish
scale candidate gene analysis in whites and African 94:1507–1514, 2004. population. PLoS ONE 6:e21790, 2011.
Americans identifies IL6R polymorphism in rela- 70. Shang Y, Yoshida T, Amendt BA, et al: Pitx2 is 91. Viviani Anselmi C, Novelli V, Roncarati R, et al:
tion to atrial fibrillation: The National Heart, functionally important in the early stages of vas- Association of rs2200733 at 4q25 with atrial
Lung, and Blood Institute’s Candidate Gene cular smooth muscle cell differentiation. J Cell flutter/fibrillation diseases in an Italian popula-
Association Resource (CARe) project. Circ Car- Biol 181:461–473, 2008. tion. Heart 94:1394–1396, 2008.
diovasc Genet 4:557–564, 2011. 71. Mommersteeg MT, Brown NA, Prall OW, et al: 92. Shi L, Li C, Wang C, et al: Assessment of associa-
49. Ellinor PT, Shin JT, Moore RK, et al: Locus for Pitx2c and Nkx2-5 are required for the formation tion of rs2200733 on chromosome 4q25 with
atrial fibrillation maps to chromosome 6q14-16. and identity of the pulmonary myocardium. Circ atrial fibrillation and ischemic stroke in a Chinese
Circulation 107:2880–2883, 2003. Res 101:902–909, 2007. Han population. Hum Genet 126:843–849, 2009.
50. Volders PG, Zhu Q, Timmermans C, et al: 72. Wang J, Klysik E, Sood S, et al: Pitx2 prevents 93. Delaney JT, Jeff JM, Brown NJ, et al: Character-
Mapping a novel locus for familial atrial fibrilla- susceptibility to atrial arrhythmias by inhibiting ization of genome-wide association-identified
tion on chromosome 10p11-q21. Heart Rhythm left-sided pacemaker specification. Proc Natl variants for atrial fibrillation in African Ameri-
4:469–475, 2007. Acad Sci U S A 107:9753–9758, 2010. cans. PLoS ONE 7:e32338, 2012.
94. Haissaguerre M, Jais P, Shah DC, et al: Spontane- 97. Kleinjan DA, van Heyningen V: Long-range 100. Semina EV, Reiter R, Leysens NJ, et al: Cloning
ous initiation of atrial fibrillation by ectopic beats control of gene expression: Emerging mecha- and characterization of a novel bicoid-related
originating in the pulmonary veins. N Engl J Med nisms and disruption in disease. Am J Hum Genet homeobox transcription factor gene, RIEG,
339:659–666, 1998. 76:8–32, 2005. involved in Rieger syndrome. Nat Genet 14:392–
95. Nobrega MA, Ovcharenko I, Afzal V, et al: Scan- 98. Lubitz SA, Sinner MF, Lunetta KL, et al: Inde- 399, 1996.
ning human gene deserts for long-range enhanc- pendent susceptibility markers for atrial fibrilla- 101. Doward W, Perveen R, Lloyd IC, et al: A muta-
ers. Science 302:413, 2003. tion on chromosome 4q25. Circulation tion in the RIEG1 gene associated with Peters’
96. Lettice LA, Heaney SJ, Purdie LA, et al: A long- 122:976–984, 2010. anomaly. J Med Genet 36:152–155, 1999.
range Shh enhancer regulates expression in the 99. Boldt LH, Posch MG, Perrot A, et al: Mutational 102. Marcus GM, Alonso A, Peralta CA, et al: Euro-
developing limb and fin and is associated with analysis of the PITX2 and NKX2-5 genes in pean ancestry as a risk factor for atrial fibrillation
preaxial polydactyly. Hum Mol Genet 12:1725– patients with idiopathic atrial fibrillation. Int J in African Americans. Circulation 122:2009–
1735, 2003. Cardiol 145:316–317, 2010. 2015, 2010.
Mechanisms in Heritable Sodium
Channel Diseases 50
Alfred L. George, Jr.
and this impacts the functional properties of certain genetic vari-

CHAPTER OUTLINE
ants.3,4 In fetal and neonatal heart, another alternatively spliced
Cardiac Sodium Channel 491 form of NaV1.5 is expressed that uses an alternative exon 6 (exon
6a; see Figure 50-1, A), resulting in several amino acid differences
Normal Sodium Channel Function 491
within a voltage-sensor domain (D1/S3-S4).1,5 This fetal-
Consequences of Sodium Channel Dysfunction 492 expressed NaV1.5 splice variant exhibits developmental regula-
tion in mouse and human heart6 and shows distinct biophysical
Delayed Repolarization 493
properties, including a more depolarized voltage dependence of
Impaired Impulse Propagation 494 activation. Other alternative splicing events have less certain bio-
logical and physiological relevance.1
Altered Intracellular Ion Homeostasis 496
Transcription of human SCN5A is under the control of a
Syndromes With Complex Mechanisms 497 promoter that precedes the first exon and other determinants of
transcriptional activity located in the first intron.7 Common poly-
morphisms in the SCN5A promoter influence the level of gene
More than a decade ago, the discovery and characterization of a expression and may have clinical relevance. In particular, a com-
mutation in the gene encoding a human cardiac voltage-gated bination of polymorphisms constituting a common haplotype
sodium channel (SCN5A) began a new era in our understanding within the promoter has been associated with longer PR interval
of the molecular and genetic basis of arrhythmias. Subsequent and QRS duration, as well as the extent of QRS widening during
work has revealed an unexpectedly large number of clinically and challenge with sodium channel blocking drugs in Asians.8 These
pathophysiologically diverse conditions associated with an ever and other observations suggest that unequal transcription from
growing number of mutations in this gene (Box 50-1). The diver- the two copies of SCN5A could influence the clinical expression
sity of clinical disorders (phenotypes) associated with SCN5A of heterozygous mutations in this gene.9
mutations is explained in part by corresponding heterogeneity in The cardiac sodium channel resides in the myocyte plasma
mutant channel dysfunction. An understanding of the functional membrane as a multi-protein complex consisting of the pore-
consequences of SCN5A mutations has driven tremendous prog- forming (α) subunit (NaV1.5), auxiliary (β) subunits, and other
ress in elucidating arrhythmia mechanisms in these genetic interacting proteins. A family of four sodium channel β-subunits
disorders. (β1, β2, β3, β4) encoded by the genes SCN1B, SCN2B, SCN3B, and
This chapter will focus primarily on molecular genetic and SCN4B are expressed in heart and likely interact with NaV1.5,
pathophysiological aspects of the heritable cardiac sodium possibly to modulate channel function or subcellular localiza-
channel diseases. As an organizational framework, molecular tion.10 Several other proteins expressed in heart also interact
mechanisms have been clustered by the predicted effect on directly or indirectly with sodium channels, including ankyrins,
cardiac electrogenesis. Other chapters provide complementary caveolin-3, calmodulin, α1-syntrophin, glycerol-3-phosphate
information about the structure and function of sodium channels, dehydrogenase 1-like protein (GPD1L), fibroblast growth factor
as well as details about the clinical manifestations of SCN5A- homologous factors, Nedd4-like ubiquitin-protein ligase, multi-
associated disorders. copy suppressor of gsp1 (MOG1), and 14-3-3η.11 Some of these
proteins have been implicated in rare cases of inherited arrhyth-
mia susceptibility, suggesting that they mediate important func-
tional interactions.12 Further information about regulation of
Cardiac Sodium Channel sodium channels in the context of multi-protein complexes is
provided in Chapter 18.
SCN5A encodes the pore-forming subunit of the major voltage-
gated sodium channel expressed in human heart. This gene spans
approximately 100 kilobases on the short arm of chromosome 3
(3p21) and comprises 28 canonical exons ranging in size from 53 Normal Sodium Channel Function
(exon 24) to 3257 (exon 28) base pairs (Figure 50-1, A). The full-
length product of SCN5A is a 2016 amino acid protein designated Cardiac voltage-gated sodium channels are critical mediators of
as NaV1.5, but other nomenclature (i.e., hH1) is also used to phase 0 depolarization and control the velocity of impulse propa-
describe recombinant forms. Alternatively spliced messenger gation through the heart. Sodium channels switch between three
RNAs (mRNAs) transcribed from SCN5A have also been detected major functional states (closed, open, and inactivated) in response
in heart.1 One splice variant expressed in adult heart is generated to changes in membrane potential. Only open channels generate
by alternative usage of splice acceptor sequences at the junction electrical current by allowing the selective passage of sodium ions
between intron 17 and exon 18, resulting in inclusion or exclu- into cells. The functional state of sodium channels can be moni-
sion of glutamine at position 1077.2 Approximately half of mature tored in electrophysiological recording experiments by using
SCN5A mRNAs in heart encode the 2015 amino acid alternative specific voltage-clamp protocols to elicit activation (closed →
form of the protein that arises from this alternative splicing event, open), inactivation (open → inactivated), and recovery from
491
Box 50-1 Disorders of the Cardiac Sodium Channel Start SCN5A (3p21) Stop
Gene, SCN5A
Ventricular Arrhythmia/Sudden Death 1 6a/6 28 10 kb
A
Congenital long QT syndrome type 3 Long-QT syndrome
Brugada syndrome type 1
Idiopathic ventricular fibrillation
Sudden unexplained nocturnal death syndrome
Sudden infant death syndrome
Impaired Cardiac Conduction
Progressive cardiac conduction system disease (Lenègre)
Familial atrioventricular block B
Congenital sick sinus syndrome Brugada syndrome – Truncating mutations
Atrial standstill
Latent Arrhythmia Susceptibility
Drug-induced LQTS
Arrhythmia susceptibility in African Americans (S1103Y)
Other
Mixed phenotypes/overlap syndromes
Dilated cardiomyopathy with arrhythmia
C
Atrial fibrillation Brugada syndrome – Non-truncating mutations
inactivation (inactivated → closed), which occur on a millisecond

time scale. In addition to these rapid transitions, sodium channels
are susceptible to slow inactivation if the membrane remains depo- D
larized for a longer time. Slow inactivation occurs on a time Conduction system disease
course of several seconds and may contribute to determining the
availability of active channels under various physiological condi-
tions. The structural basis for sodium channel function is dis-
cussed in Chapter 1, and additional details about functional
properties of sodium channels are presented in Chapter 9.
E
Consequences of Sodium Dilated cardiomyopathy
Channel Dysfunction
More than 450 SCN5A mutations have been identified in patients
with the arrhythmia predisposition syndromes presented in
Box 50-1. Types of mutations include missense (e.g., amino
acid substitutions), nonsense (e.g., premature stop codon), inser-
tions or deletions, and splice-site mutations. Some mutation F
types, particularly those predicted to truncate the encoded
protein (nonsense, insertions, or deletions affecting the reading Figure 50-1. SCN5A gene structure and mutations A, Schematic of the
frame), destroy the functionality of the channel (loss-of-function) SCN5A gene illustrating 28 canonical exons and one alternative exon (vertical lines;
by deleting critical domains or preventing protein translation red for canonical coding exons, blue for alternative exon 6a, and dark gray for non-
through nonsense-mediated decay of the mRNA. Splice-site coding exons or portions of exons). The bent arrow indicates the approximate
mutations may also severely disrupt channel function if exons are location of the transcription start site near the promoter. Locations of the transla-
skipped or intron sequences are retained in the final mRNA. tion start and stop codons are indicated. Locations of mutations associated with
LQT-3 (B), BrS (C,D), cardiac conduction disease (E), and dilated cardiomyopathy
However, a large number of SCN5A mutations result in single (F) are superimposed on a two-dimensional membrane topology map of the
amino acid substitutions for which functional predictions are sodium channel protein
difficult. Several missense mutations and a few in-frame insertion/
deletion alleles have been studied in vitro to determine the (Mutation data were obtained from the online Inherited Arrhythmias Database
molecular basis for arrhythmia predisposition in these syndromes, [http://www.fsm.it/cardmoc/] and from the literature. Images were prepared by
and common patterns of sodium channel dysfunction have Dr. Robert Abraham, Vanderbilt University, using software generously supplied by
emerged. Dr. André Linnenbank, University of Amsterdam.)
At the cellular and tissue levels, abnormal sodium channel
function can be predicted to cause delayed repolarization, impaired
impulse propagation, altered intracellular ion homeostasis, or combi-
nations of these pathophysiological effects. These pathogenic
outcomes will be used as the framework for discussion of mecha-
nisms of sodium channel dysfunction in specific disorders.
Mechanisms in Heritable Sodium Channel Diseases 493
Delayed Repolarization
50
Persistent INa
Congenital Long QT Syndrome

INa
Congenital long QT syndrome (LQTS) is an inherited condition
of abnormal myocardial repolarization. It is characterized clini-
cally by increased risk of potentially fatal ventricular arrhythmias,
especially torsades de pointes, manifesting as syncope, cardiac
arrest, and sudden unexplained death in otherwise healthy young
adults and children. The disease is typically recognized in late
childhood or early adolescence, but extreme cases may present
during early infancy or in the perinatal period.13 The syndrome Prolonged APD
most often transmitted is in families as an autosomal dominant
trait (Romano-Ward syndrome; see also Chapter 93).
Approximately 10% of autosomal dominant LQTS cases are Vm
explained by SCN5A mutations,14 and this form of the disease is
referred to as LQT-3. The proportion of sodium channel muta-
tions in LQTS cases presenting during the perinatal and neonatal
time periods may be higher.15 The first SCN5A mutation
described was an in-frame deletion of three highly conserved
amino acid residues (delKPQ 1505-1507) located within the Figure 50-2. Increased persistent sodium current in LQT-3 Upper panel
cytoplasmic loop connecting domains 3 and 4 of NaV1.516—a is a representative current recording from a cell expressing either wild type (WT)
structural domain required for fast inactivation of the channel. cardiac sodium channels (black trace) or an SCN5A mutation associated with LQT-3
Since the discovery of the first SCN5A mutation, more than (red trace) illustrating increased persistent sodium current (INa). Lower panel illus-
75 mostly missense SCN5A mutations have been associated with trates a normal ventricular action potential waveform (black line) and an illustration
LQT-3 (Figure 50-1, B). Although mutations are found through- of prolonged action potential duration (APD; red line) as occurs in LQTS.
out the channel sequence, multiple mutations are clustered in a
small number of structural regions. One of these regions is the
S4 segment of domain 4 (D4), a critical structural determinant associated with SNTA1 mutations involves activation of neuronal
of voltage sensing. Another mutation cluster can be found in the nitric oxide synthase (nNOS) complexed with NaV1.5, leading to
aforementioned cytoplasmic inactivation gate. Finally, several increased cell levels of nitric oxide and nitrosylation of the sodium
mutations are found in the proximal carboxy terminus, which channel. Activation of nNOS is the result of disrupted scaffolding
includes binding sites for several interacting proteins that can of a plasma membrane Ca2+-ATPase (PMCA4b) that normally
modulate sodium channel function.11 inhibits the enzyme in a multi-protein complex with NaV1.5 and
Persons with SCN5A mutations associated with LQTS often α1-syntrophin.28
present with distinct clinical features, including sinus bradycar- Increased persistent sodium current provides an explanation
dia, and a tendency for cardiac events to occur during sleep or for delayed repolarization in LQTS.29 Cardiac action potentials
rest.17 In addition to bradycardia, certain features of the surface last several hundred milliseconds because of a prolonged
electrocardiogram (ECG), such as narrow and late-onset peaked depolarization phase (plateau), the result of opposing inward
or biphasic T waves, may offer additional clues to the presence (mainly Na+ and Ca2+) and outward (K+) ionic currents. Repolar-
of an SCN5A mutation in the setting of LQTS.18 Cardiac events ization occurs when net outward current exceeds net inward
are less frequent in the settings of SCN5A mutations as compared current. Increased persistent sodium current will shift this balance
with the two major forms of LQTS caused by potassium channel toward inward current and will delay onset of repolarization, thus
mutations (LQT-1, LQT-2).19,20 However, in children and adults, lengthening the action potential duration and the corresponding
the risk of dying after a cardiac event is greater for LQT-3 than QT interval. Delayed repolarization predisposes to ventricular
for either LQT-1 or LQT-2. arrhythmias mainly by increasing the probability of early after-
In LQT-3, most SCN5A mutations exhibit a dominant gain- depolarization (EAD) and by increasing the dispersion of the
of-function phenotype at the molecular level characterized by action potential duration.30 These phenomena promote condi-
defective inactivation leading to increased persistent sodium tions that allow electrical signals from depolarized regions of the
current (Figure 50-2).21 Some mutant channels may also exhibit heart to prematurely reexcite adjacent myocardium that has
accelerated recovery from inactivation—a phenomenon consis- already repolarized—the basis for a reentrant arrhythmia.
tent with an unstable inactivated state. At the level of single The pathogenic effects of increased persistent sodium current
sodium channels, increased persistent sodium current has been in LQTS have also been strongly supported by computer simula-
correlated with an increased tendency for channel reopening, tions of cardiac action potentials29,31 and by electrophysiological
which may occur in bursts.21 More severe slowing of inactivation investigations of cardiac myocytes from genetically engineered
may be noted in mutations associated with clinically severe mice. Mice engineered to carry the delKPQ mutation have spon-
LQTS,13,22 while a small number of mutations may alter voltage taneous life-threatening ventricular arrhythmias.32,33 At the cel-
dependence of activation and/or inactivation with less dramatic lular level, cardiomyocytes from these LQT-3 mice exhibit
effects on the level of persistent current.23,24 prolonged action potential duration and frequent EADs—
Increased persistent sodium current has also been observed findings that were exaggerated by slow pacing rates. These cells
when wild type NaV1.5 channels were coexpressed with LQTS- have increased persistent sodium current with faster recovery
associated mutations in certain auxiliary subunits and interacting from inactivation—features that were predicted from studies per-
proteins. Specifically, mutations in SCN4B encoding the β4 aux- formed in noncardiac cells.32 Further evidence from the delKPQ
iliary subunit,25 CAV3 encoding the vesicular trafficking protein mouse suggests that delayed afterdepolarization (DAD) caused
caveolin-3,26 and SNTA1 encoding the adapter protein α1- by a Ca2+-dependent diastolic transient inward current, possibly
syntrophin,27,28 exert their pathologic effects by disturbing sodium evoked by abnormal Na+ entry through mutant channels, also
channel inactivation, leading to increased persistent current. The contributes to arrhythmia susceptibility and may account for the
mechanism responsible for increased persistent current greater lethality of this LQTS subtype.34
Knowledge of the basic mechanisms underlying LQT-3 has Other common SCN5A variants may also be proarrhythmic
prompted new ideas regarding genotype-specific treatment. because of altered sodium channel inactivation. The common
Mexiletine and other sodium channel blocking drugs can reduce variant R1193Q is common in Asians but is a rare variant (0.3%)
persistent sodium current and shorten the QT interval in LQT-3 in subjects of European ancestry.57 This variant has been reported
patients,35-38 although no data yet indicate that this therapeutic in subjects with both acquired and congenital LQTS58,59 and
strategy will reduce mortality. In LQT-3 mouse models, mexi- in sudden unexplained death syndrome (SUDS).60 This variant
letine shortens the myocyte action potential and prevents promotes an increased persistent sodium current58 and faster
arrhythmias,30,39 mainly during slow pacing rates. Suppression of inactivation.60
increased persistent current by the antianginal drug ranolazine
has been demonstrated for LQT-3 mutations in vitro,40 and rano-
lazine treatment of human with carrying the SCN5A-delKPQ Impaired Impulse Propagation
mutation shortened the QTc interval significantly.41 Propranolol,
but not the other β-adrenergic antagonists, may exhibit sodium Brugada Syndrome
channel blocking effects at high concentrations,42-44 and certain
SCN5A mutations have increased sensitivity to the drug.13 Mutations in SCN5A have been associated with 20% to 30% of
However, propranolol and other β-blockers did not suppress cases of Brugada syndrome (BrS), a heritable form of idiopathic
arrhythmias in the delKPQ mouse model.33 ventricular fibrillation.61,62 Individuals with BrS have increased
risk for potentially lethal ventricular arrhythmias (polymorphic
ventricular tachycardia or fibrillation) typically during sleep, but
Perinatal and Neonatal LQT-3 in the absence of myocardial ischemia, electrolyte abnormalities,
or structural heart disease. Increased risk for atrial fibrillation and
SCN5A mutations may cause life-threatening LQTS during ges- intraventricular conduction abnormalities may also occur in BrS.
tation (fetal LQTS) and in neonates (neonatal LQTS). Clinical Individuals with the disease may exhibit a characteristic baseline
signs suggestive of fetal LQTS include ventricular tachycardia, ECG pattern consisting of ST elevation in the right precordial
second-degree atrioventricular (AV) block, and, most commonly, leads but normal QT intervals.63 Administration of sodium
sinus bradycardia,45 but such findings may go undetected because channel blocking agents (e.g., procainamide, flecainide, ajmaline)
electrocardiographic testing of fetuses is not routine. Certain or fever may expose this ECG pattern in latent cases.64,65 Inheri-
SCN5A mutations, many of which are de novo,13,46-49 present with tance is autosomal dominant with incomplete, often low, pene-
earlier onset and more severe congenital arrhythmia syndromes trance and a substantial male predominance. A family history of
than is typical for LQT-3. The high rate of reported de novo unexplained sudden death is typical. The sudden unexplained
SCN5A mutations in the perinatal period may reflect low herita- death syndrome (SUDS) is clinically similar to BrS and causes
bility due to a survival disadvantage conferred by severe, life- sudden death, typically during sleep, in young and middle-aged
threatening phenotypes. The functional effects of certain SCN5A males in Southeast Asian countries.66 Additional information
mutations associated with fetal LQTS can also be potentiated by about BrS may be found in Chapter 92.
a developmentally regulated splice isoform involving alternative Chen et al first identified three distinct SCN5A mutations in
forms of exon 6.6 two unrelated BrS families and in a third sporadic case.61 One
Evidence indicates that occult LQTS can clinically present as mutation was missense (T1620M), and the other alleles included
sudden infant death syndrome (SIDS) (see Chapter 98). Cardiac a frameshift caused by a single nucleotide deletion and a putative
mechanisms including life-threatening arrhythmias have been splice-site defect. More than 375 mutations have been reported
suspected as risk factors for SIDS, and mutations in genes respon- in BrS (Figure 50-1, C, D). SCN5A mutations have also been
sible for LQTS have been identified in approximately 10% of identified in subjects with SUDS.60 In contrast to LQT-3, many
cases.50,51 SCN5A mutations have accounted for approximately mutations in BrS are predicted to cause protein truncation by
50% of all LQTS mutations identified in SIDS. SCN5A muta- causing frameshifts, premature stop codons, or aberrant mRNA
tions associated with SIDS promote increased persistent current splicing. These observations provided the first suggestion that
overtly or under certain conditions such as intracellular acidosis sodium channel loss-of-function is responsible for the disorder.
or in the context of a common splice variant.50,52-54 Most SCN5A mutation–positive BrS subjects are heterozy-
gous, but a few reports have described homozygous or compound
heterozygous mutation carriers.67,68 Most SCN5A mutation-
Arrhythmia Susceptibility With Common positive BrS families exhibit low penetrance (i.e., low concor-
SCN5A Variants dance between genotype and clinical disease). One attractive
hypothesis put forth to explain incomplete penetrance in BrS is
Common variants within the SCN5A coding region have been the existence of genetic modifiers. As has been mentioned,
identified in certain populations, and some of these may confer genetic variants in the SCN5A promoter may influence the clini-
increased risk of cardiac arrhythmia. One variant common in cal expression of heterozygous loss-of-function mutations.9
subjects of African descent (S1103Y, also reported as S1102Y) has Another potential mechanism involves interactions between wild
been associated with an eightfold increased risk for ventricular type and mutant sodium channel alleles. This idea was first sug-
arrhythmia.55 Functional consequences of S1103Y include gested to explain the variable clinical expression of BrS in a family
increased transient and increased persistent current sufficient to segregating a trafficking defective missense mutation (R282H) in
evoke EADs in a computational model of cardiac action poten- which an asymptomatic mutation carrier also carried a common
tials. This variant may also increase arrhythmia risk in infants. variant (H558R) on the opposite allele.69 In vitro experiments
One study suggested that African American SIDS victims are demonstrated that H558R rescued the trafficking defect through
more often homozygous for SCN5A-S1103Y when compared interactions between the two variant NaV1.5 proteins.70
with non-SIDS infants.53 This study further demonstrated that Reduced sodium current is the primary pathophysiological
SCN5A-S1103Y channels exhibit a greater level of persistent event in BrS; this is consistent with SCN5A mutations predicted
sodium current when exposed to intracellular acidosis. A subse- to encode nonfunctional channels.61,71 Some missense mutations
quent study reported a significantly higher prevalence of hetero- are nonfunctional because of impaired protein trafficking to the
zygous SCN5A-S1103Y carriers among 71 African American cell membrane or defective ion conductance.72,73 Trafficking defi-
cases of SIDS as compared with African American controls.56 cient mutants have the potential to exert a dominant-negative
effect on the wild type sodium channel.74 Other missense muta- Familial Progressive Conduction System Disease
50
tions are functional but have biophysical defects predicted to
reduce channel availability such as altered voltage dependence of Mutations in SCN5A have been associated with a heterogeneous
activation, more rapid fast inactivation, and enhanced slow group of disorders characterized by impaired cardiac conduction
inactivation.75-77 One mutation (E1053K) disrupts binding of the manifest as abnormal atrioventricular (AV) conduction (AV
channel to ankyrin-G, and this alters membrane trafficking and block), slowed intraventricular conduction, and/or atrial inexcit-
localization in cardiac myocytes (see also Chapter 9).78 Truncat- ability (atrial standstill).102-105 The degree of impaired cardiac
ing or nonfunctional missense mutations are generally associated conduction often progresses with advancing age. Inheritance may
with more severe phenotypes.79 be autosomal dominant or autosomal recessive, but in rare cases,
Reduced sodium current may also be the consequence of digenic inheritance has been observed (see later).104,106,107
mutations in other genes that are less frequent causes of BrS, Age-related defects in conduction along the His-Purkinje
including those encoding sodium channel auxiliary subunits system have been referred to as Lenègre or Lev disease and may be
(SCN1B, SCN3B),80,81 interacting proteins (MOG1),82 or caused by progressive fibrosis of specialized conducting tissues in
glycerol-3-phosphate dehydrogenase 1-like protein (GPD1L).83 the heart.108 However, clinically similar conditions may be inher-
Mutations in GPD1L have been suggested to cause suppression ited and associated with SCN5A mutations109 (see Chapter 107).
of sodium current by a protein kinase C (PKC)-dependent mech- A variety of names have been applied to this genetic subtype,
anism that is linked with the redox state of the cell.84,85 Specifi- including progressive cardiac conduction disease, familial AV block,
cally, reduced enzymatic activity of mutant GPD1L protein is and hereditary Lenègre disease. The clinical phenotypes may
associated with an NADH/NAD+ imbalance that can activate present in early childhood or in adulthood. In children, progres-
PKC, leading to phosphorylation of a specific serine residue (Ser- sive AV block is most typical, whereas adults with these disorders
1503) in NaV1.5, causing reduced channel activity. usually have intraventricular conduction abnormalities such as
Knowledge of the mechanisms underlying loss of sodium right or left bundle branch block.
channel function in BrS has inspired the testing of strategies that Twenty-two SCN5A mutations have been associated with
may have therapeutic potential. For example, trafficking defective cardiac conduction disease, including congenital sick sinus syn-
SCN5A mutants may be rescued to the plasma membrane and drome and atrial standstill (Figure 50-1, E). The first mutations
rendered functional by treatment with mexiletine.73,86 Addition- described by Schott and colleagues were predicted to be complete
ally, for certain nonsense mutations, partial restoration of sodium loss-of-function alleles.102 Several missense mutations have since
current can be achieved in vitro by pharmacologic “read-through” been characterized in vitro, and these alleles may be nonfunc-
strategies using aminoglyosides or by genetic suppression of tional or may exhibit complex biophysical abnormalities that are
eukaryotic release factors (e.g., eRF3a) that facilitate stop codon predicted to reduce peak sodium current density at physiological
recognition.87 voltages.103,110 In the case of G514C, a missense mutation segre-
Two mechanisms have been proposed to explain the cellular gating with conduction disease in a Dutch family, the mutant
basis of BrS, both implicating electrophysiological defects in the sodium channel exhibits unequal depolarizing shifts in the voltage
right ventricle.88,89 In one mechanism, a reduction in myocardial dependence of activation and inactivation such that a smaller
sodium current exaggerates differences in action potential dura- number of channels are activated at typical threshold voltages.110
tion between endocardium and epicardium.90-92 These differences Computational modeling of these changes supported reduced
occur because of unequal distribution of transient outward current conduction velocity, but the level of predicted sodium current
(Ito), a repolarizing current more prominent in the epicardial layer loss was not sufficient to evoke shortened epicardial action poten-
that contributes to the characteristic spike and dome shape of the tials, offering a plausible explanation as to why these individuals
action potential. Reduced sodium current will cause dispropor- did not manifest BrS. Other missense mutations cause enhanced
tionate shortening of epicardial action potentials because of unop- slow inactivation, a gating process that regulates channel avail-
posed Ito leading to an exaggerated transmural voltage gradient, ability over a time course of several seconds to minutes.103 Lower
dispersion of repolarization, and a substrate promoting phase 2 peak sodium current impairs conduction velocity by slowing the
reentry. This mechanism is supported by elegant work using the rate of change in membrane potential (dV/dt) during the action
arterially perfused canine ventricular wedge preparation.90,91 The potential.
second hypothesis posits that the main effect of reduced sodium As has been discussed, aging Scn5a+/− mice variably develop
current consists of slowing of impulse conduction in the right progressive conduction system fibrosis and impaired impulse
ventricle and delayed activation of the right ventricular outflow conduction through both ventricles. The molecular basis for
tract (RVOT).88,93-95 This mechanism has gained support primar- fibrosis is unknown, but it has been correlated with increased
ily from clinical observations including electroanatomic mapping expression of transforming growth factor-β1 (TGF-β1) and
studies96,97 and the observed therapeutic benefit of epicardial abla- reduced expression of the gap junction protein Cx40 in heart
tion over the RVOT.98 It is not clear whether these two hypoth- tissue.111,112 The degree of fibrosis and the severity of impaired
eses are mutually exclusive, or whether all cases of BrS originate conduction in Scn5a+/− mice correlate with the residual level of
by the same pathophysiological mechanism. sodium channel protein, but the mechanism for variable protein
Heterozygous Scn5a knockout mice (Scn5a+/−) have provided levels is unknown.113 Observations made in Scn5a+/− mice illus-
an animal model of BrS.99-101 Hearts from young Scn5a+/− mice trate a continuum of pathophysiological events associated with
exhibit conduction slowing in the right ventricle and an increased loss-of-function SCN5A mutations and help to explain the
propensity for ventricular tachycardia—features that are aggra- co-occurrence of BrS and conduction system diseases, which have
vated by flecainide. The arrhythmia has been demonstrated to been termed overlap syndromes.
originate in the RVOT in vivo and ex vivo. Although data from
investigations of Scn5a+/− mice tend to support the hypothesis of
conduction slowing as the primary electrophysiological defect in Congenital Sick Sinus Syndrome
BrS, mouse ventricular action potentials lack a plateau phase, and Atrial Standstill
making this a suboptimal model for testing the alternative mecha-
nism. In older Scn5a+/− mice, progressive fibrosis and conduction Whereas most SCN5A-linked disorders manifest clinically with
slowing occur in both ventricles, providing a model system from abnormal ventricular excitability or conduction, in less common
which to gain an understanding of conduction system disorders syndromes mutant sodium channels disproportionately affect
associated with SCN5A mutations. atrial electrical activity. In familial atrial standstill and congenital
sick sinus syndrome, SCN5A mutations cause sinus node dysfunc- dependence of activation (Figure 50-3, A). Subsequently, the
tion progressing to atrial inexcitability and symptomatic brady- same biophysical defect was observed for SCN5A-R222Q, a
cardia in children or young adults. Fetal bradycardia and irregular mutation involving a different conserved arginine residue in the
fetal heart rate may be early signs of the condition during gesta- D1/S4 segment.119,120 These observations led to the prediction
tion. Surface ECG recordings reveal an absence of P waves and that channels might be active at voltages near the cardiac myocyte
a slow AV junctional rhythm. Electrophysiological studies may resting membrane potential. Aberrant activation of sodium
reveal prolonged His-ventricular conduction indicating a more current near the myocyte resting membrane potential can be
generalized defect in cardiac conduction. In these conditions, the demonstrated in heterologous cells expressing either R814W or
atria, but not the ventricles, are refractory to electrical pacing. R222Q subjected to slow voltage ramps (Figure 50-3, B).126 This
Families with these disorders caused by SCN5A mutations phenomenon differs from that observed for LQT-3 mutations in
may exhibit recessive inheritance. In congenital sick sinus syn- that peak current activation during voltage ramps for those muta-
drome, affected individuals inherit a nonfunctional or severely tions occurs at more positive potentials.127 The tendency for
dysfunctional SCN5A mutation from one parent and a less diastolic sodium current activation may predispose myocytes to
severely impaired allele from the other parent.105 Some of these disordered intracellular sodium and calcium ion homeostasis,
mutations impair trafficking of the sodium channel protein to the leading to myocardial dysfunction.
plasma membrane.114 Heterozygous parents are asymptomatic Another SCN5A mutation associated with DCM (R219H)
but may have subclinical first-degree AV block. In addition to predisposes to altered intracellular ion homeostasis by a novel
compound heterozygous SCN5A mutations, similar phenotypes mechanism.121 This mutation replaces the most extracellular argi-
have been associated with homozygous alleles.106 Familial atrial nine residue in the D1/S4 voltage sensing segment with histidine,
standstill has also been associated with heterozygous SCN5A but this substitution has no overt effects on channel function.
mutations coinherited with a common promoter variant in the However, cells expressing this mutant sodium channel exhibited
connexin-40 gene that is predicted to reduce gene expres- rapid internal acidification when exposed to a pH gradient. This
sion.104,107 In these families, inheritance of the sodium channel finding was explained by the generation of a proton leak current
mutation alone was not sufficient to evoke a clinical phenotype— through mutant sodium channels, but the leak does not occur
a finding consistent with digenic inheritance.
The mechanisms by which SCN5A mutations impair sinus
node dysfunction have been explored using action potential mod-
eling and studies of Scn5a+/− mice.115,116 In single-cell action 1.0
potential simulations, reduced sodium current slows pacing in the Normalized conductance
peripheral cells of the sinoatrial node (SAN) but not in the central
SAN cells, which are primarily responsible for cardiac rhythm
and lack sodium channels. In multicellular simulations, reduced
sodium current impairs conduction from SAN to surrounding
0.5
atrial tissue, leading to sinoatrial exit block. These predictions
WT
are consistent with observations made in Scn5a+/− mice.116 R814W
R222Q
Altered Intracellular Ion Homeostasis 0.0

–80 –60 –40 –20 0
Dilated Cardiomyopathy With Arrhythmia
A Voltage (mV)
Mutations affecting cardiac ion channels primarily cause disor-
ders of heart rhythm, whereas familial cardiomyopathy is typi-
cally associated with mutations in genes encoding sarcomeric or
contractile proteins. An exception to this paradigm is represented WT
by SCN5A mutations associated with familial dilated cardiomy-
2 pA/pF
R814W
opathy (DCM). McNair et al reported that a missense mutation,
SCN5A-D1275N, segregated with a phenotype including DCM
of variable expression, abnormal atrioventricular conduction,
sinus node dysfunction, and atrial and ventricular arrhythmias in
a four-generation kindred.117 Subsequent studies have revealed
sec)
ten additional mutations (Figure 50-1, F).118-125 Inheritance is +40 mV (2
most consistent with autosomal dominance with incomplete pen- –120 mV to
e ramp
etrance except in sporadic cases. Most of the mutations segre- Voltag
gated in families with a heterogeneous phenotype characterized
by variable expression of DCM including peripartum cardiomy- –120 –80 –40 0
opathy, sick sinus syndrome, and atrial and ventricular arrhyth-
mias. Mutation positive subjects may have frequent, multifocal B Voltage (mV)
premature ventricular beats (PVBs) that originate in Purkinje Figure 50-3. Aberrant mutant sodium current activation in dilated
fibers.119,120 Suppression of PVBs with quinidine has been reported cardiomyopathy A, Conductance-voltage (G-V) curves illustrating the voltage
to improve myocardial contractility in some cases.119 dependence of activation for wild type (WT) NaV1.5 and two mutations (R814W,
Several patterns of sodium channel dysfunction have been R222Q) associated with dilated cardiomyopathy. Both mutations cause hyperpolar-
elucidated for mutations associated with DCM. In the first report, izing shifts in the G-V relationship. B, Activation of sodium current during a voltage
ramp from −120 mV to +40 mV for WT-NaV1.5 and R814W. Mutant channels activate
the SCN5A-R814W mutation associated with this syndrome
aberrantly at voltages near the resting membrane potential, suggesting a predis-
exhibited a novel in vitro biophysical phenotype. This mutation, position to diastolic sodium current.
which affects a conserved arginine residue in the D2/S4 voltage
sensing segment, caused a hyperpolarized shift in the voltage (Images were prepared by Tom Beckermann and Dr. Thao Nguyen.)
through the normal pathway for sodium ions. Rather, protons mutation SCN5A-E1784K, which is associated with a highly pen-
50
are conducted through the “gating pore,” which consists of an etrant form of LQT-3 and concurrent features of BrS or sinus
aqueous channel surrounding the S4 segment. An aberrant node dysfunction in some carriers.131 These distinct biophysical
proton leak current may create an unstable myocyte membrane abnormalities are predicted to predispose to ventricular arrhyth-
potential, possibly evoking PVBs, or may adversely affect con- mia through different mechanisms at opposite extremes of heart
tractile proteins through altered intracellular pH. The increased rate.31 Whereas, increased persistent current will prolong action
intracellular proton concentration may also stimulate a cascade potentials, especially at slow heart rates, enhanced slow inactiva-
of ion exchange events (Na+/H+; Na+/Ca2+) that perturb Ca2+ tion will predispose to activity-dependent loss of sodium channel
homeostasis with long-term deleterious effects on myocardial availability at fast rates. These findings were confirmed by
contractility. computational modeling and in a mouse model genetically
One particular SCN5A mutation (D1275N) associated with engineered with the SCN5A-1795insD mutation.132 In another
DCM and a variety of conduction disorders including sinus node unusual case, deletion of lysine-1500 in SCN5A was associated
dysfunction causes a complex cardiac phenotype in mice.128 Mice with the unique combination of LQTS, BrS, and impaired
engineered to be homozygous for the human SCN5A-D1275N cardiac conduction in the same family.133 This mutation impairs
mutation exhibit bradycardia and impaired atrial and ventricular inactivation causing increased persistent sodium current but also
conduction while young (3 to 12 weeks) but then progress to reduces sodium channel availability through opposite shifts in
develop spontaneous ventricular tachycardia and nonfibrotic voltage dependence of inactivation and activation.
DCM when older. Sodium currents from homozygous mutant Certain SCN5A mutations may be associated with different
mice have markedly reduced amplitude and abnormal inactiva- clinical manifestations in different families (e.g., D1275N; see
tion gating, including slow kinetics, increased persistent current, earlier)104,117,118,129 or even among members of the same family,134
and depolarized voltage dependence. Studies of heterologously perhaps because of host-specific factors such as modifier genes
expressed SCN5A-D1275N channels demonstrated none of these that influence the final clinical expression. In the mouse model
functional abnormalities, prompting speculation that myocyte- of the SCN5A-1795insD mutation, evidence of strain-dependent
specific factors are required for full expression of the phenotype. phenotype severity prompted a search for genetic modifiers of
Another potential explanation relates to genetic modifiers cardiac conduction.135 Specifically, the conduction defect was less
or environmental factors as suggested by the highly variable severe when the mutation was present in FVB/N mice than when
nature of the phenotypes associated with this mutation in the same allele was present on a 129P2 genomic background.
humans.104,117,118,129 Further, flecainide worsened bradycardia and QRS prolongation
to a greater degree in 129P2 than in FVB/N mutant mice. Dif-
ferences in functional behavior of cardiac sodium current between
the two strains were also observed. Specifically, sodium currents
Syndromes With Complex Mechanisms recorded from 129P2 myocytes exhibited a depolarized shift in
the conductance-voltage relationship, compared with FVB/N
Mutations in SCN5A have been associated with complex pheno- myocytes. This finding suggested that insufficient sodium channel
types (overlap syndromes) featuring characteristics of two or activation was the mechanism for impaired conduction velocity—
more syndromes. Mutant cardiac sodium channels associated an idea corroborated by computational modeling. Differences in
with overlap syndromes exhibit more complex functional defects. cardiac gene expression between strains may contribute to these
The in-frame insertion mutation (SCN5A-1795insD) identified differences in sodium currents. Specifically, Scn4b encoding the
in a family segregating both LQTS and BrS causes an inactiva- β4 sodium channel auxiliary subunit is expressed at a substantially
tion defect resulting in increased persistent sodium current lower level in right ventricular tissue from 129P2 mice.135 Several
typical of LQT-3 but also confers enhanced slow inactivation other divergent patterns of gene expression were also observed,
with reduced channel availability that is more consistent with but their significance was unclear. Additional studies to uncover
BrS.130 The combination of increased persistent sodium current genetic modifiers may yield new therapeutic targets and enable
with enhanced slow inactivation has also been observed with better risk stratification strategies.
6. Murphy LL, Moon-Grady AJ, Cuneo BF, et al: 12. Wilde AA, Brugada R: Phenotypical manifesta-
References Developmentally regulated SCN5A splice variant tions of mutations in the genes encoding subunits
potentiates dysfunction of a novel mutation asso- of the cardiac sodium channel. Circ Res 108:884–
1. Schroeter A, Walzik S, Blechschmidt S, et al: ciated with severe fetal arrhythmia. Heart Rhythm 897, 2011.
Structure and function of splice variants of the 9:590–597, 2011. 13. Wang DW, Crotti L, Shimizu W, et al: Malignant
cardiac voltage-gated sodium channel NaV1.5. 7. Atack TC, Stroud DM, Watanabe H, et al: Infor- perinatal variant of long-QT syndrome caused by
J Mol Cell Cardiol 49:16–24, 2010. matic and functional approaches to identifying a a profoundly dysfunctional cardiac sodium
2. Makielski JC, Ye B, Valdivia CR, et al: A ubiqui- regulatory region for the cardiac sodium channel. channel. Circ Arrhythm Electrophysiol 1:370–
tous splice variant and a common polymorphism Circ Res 109:38–46, 2011. 378, 2008.
affect heterologous expression of recombinant 8. Bezzina CR, Shimizu W, Yang P, et al: Common 14. Kapplinger JD, Tester DJ, Salisbury BA, et al:
human SCN5A heart sodium channels. Circ Res sodium channel promoter haplotype in Asian sub- Spectrum and prevalence of mutations from the
93:821–828, 2003. jects underlies variability in cardiac conduction. first 2,500 consecutive unrelated patients referred
3. Tan BH, Valdivia CR, Rok BA, et al: Common Circulation 113:338–344, 2006. for the FAMILION long QT syndrome genetic
human SCN5A polymorphisms have altered elec- 9. Park JK, Martin LJ, Zhang X, et al: Genetic vari- test. Heart Rhythm 6:1297–1303, 2009.
trophysiology when expressed in Q1077 splice ants in SCN5A promoter are associated with 15. Horigome H, Nagashima M, Sumitomo N, et al:
variants. Heart Rhythm 2:741–747, 2005. arrhythmia phenotype severity in patients with Clinical characteristics and genetic background of
4. Tan BH, Valdivia CR, Song C, et al: Partial heterozygous loss-of-function mutation. Heart congenital long-QT syndrome diagnosed in fetal,
expression defect for the SCN5A missense muta- Rhythm 9:1090–1096, 2012. neonatal, and infantile life: A nationwide ques-
tion G1406R depends upon splice variant back- 10. Kaufmann SG, Westenbroek RE, Zeckner C, et tionnaire survey in Japan. Circ Arrhythm Electro-
ground Q1077 and rescue by mexiletine. Am J al: Functional protein expression of multiple physiol 3:10–17, 2010.
Physiol Heart Circ Physiol 291:H1822–H1828, sodium channel alpha- and beta-subunit isoforms 16. Wang Q, Shen J, Splawski I, et al: SCN5A
2006. in neonatal cardiomyocytes. J Mol Cell Cardiol mutations associated with an inherited cardiac
5. Ou SW, Kameyama A, Hao LY, et al: Tetrodotoxin- 48:261–269, 2010. arrhythmia, long QT syndrome. Cell 80:805–811,
resistant Na+ channels in human neuroblastoma 11. Abriel H: Cardiac sodium channel NaV1.5 and 1995.
cells are encoded by new variants of Nav1.5/ interacting proteins: Physiology and pathophysi- 17. Schwartz PJ, Priori SG, Spazzolini C, et al:
SCN5A. Eur J Neurosci 22:793–801, 2005. ology. J Mol Cell Cardiol 48:2–11, 2010. Genotype-phenotype correlation in the long-QT
syndrome: Gene-specific triggers for life- SCN5A:DeltaKPQ mutation. Ann Noninvasive 57. Ackerman MJ, Splawski I, Makielski JC, et al:
threatening arrhythmias. Circulation 103:89–95, Electrocardiol 6:153–158, 2001. Spectrum and prevalence of cardiac sodium
2001. 38. Abriel H, Wehrens XH, Benhorin J, et al: Molec- channel variants among black, white, Asian, and
18. Zhang L, Timothy KW, Vincent GM, et al: Spec- ular pharmacology of the sodium channel muta- Hispanic individuals: Implications for arrhythmo-
trum of ST-T-wave patterns and repolarization tion D1790G linked to the long-QT syndrome. genic susceptibility and Brugada/long QT syn-
parameters in congenital long-QT syndrome: Circulation 102:921–925, 2000. drome genetic testing. Heart Rhythm 1:600–607,
ECG findings identify genotypes. Circulation 39. Tian XL, Yong SL, Wan X, et al: Mechanisms by 2004.
102:2849–2855, 2000. which SCN5A mutation N1325S causes cardiac 58. Wang Q, Chen S, Chen Q, et al: The common
19. Zareba W, Moss AJ, Locati EH, et al: Modulating arrhythmias and sudden death in vivo. Cardiovasc SCN5A mutation R1193Q causes LQTS-type
effects of age and gender on the clinical course of Res 61:256–267, 2004. electrophysiological alterations of the cardiac
long QT syndrome by genotype. J Am Coll 40. Fredj S, Sampson KJ, Liu H, et al: Molecular basis sodium channel. J Med Genet 41:e66, 2004.
Cardiol 42:103–109, 2003. of ranolazine block of LQT-3 mutant sodium 59. Hwang HW, Chen JJ, Lin YJ, et al: R1193Q of
20. Priori SG, Schwartz PJ, Napolitano C, et al: Risk channels: Evidence for site of action. Br J Phar- SCN5A, a Brugada and long QT mutation, is a
stratification in the long-QT syndrome. N Engl macol 148:16–24, 2006. common polymorphism in Han Chinese. J Med
J Med 348:1866–1874, 2003. 41. Moss AJ, Zareba W, Schwarz KQ, et al: Ranola- Genet 42:e7, 2005.
21. Bennett PB, Yazawa K, Makita N, et al: Molecular zine shortens repolarization in patients with sus- 60. Vatta M, Dumaine R, Varghese G, et al: Genetic
mechanism for an inherited cardiac arrhythmia. tained inward sodium current due to type-3 and biophysical basis of sudden unexplained noc-
Nature 376:683–685, 1995. long-QT syndrome. J Cardiovasc Electrophysiol turnal death syndrome (SUNDS), a disease allelic
22. Kambouris NG, Nuss HB, Johns DC, et al: Phe- 19:1289–1293, 2008. to Brugada syndrome. Hum Mol Genet 11:337–
notypic characterization of a novel long-QT syn- 42. Wang DW, Mistry AM, Kahlig KM, et al: Pro- 345, 2002.
drome mutation (R1623Q) in the cardiac sodium pranolol blocks cardiac and neuronal voltage- 61. Chen Q, Kirsch GE, Zhang D, et al: Genetic basis
channel. Circulation 97:640–644, 1998. gated sodium channels. Frontiers Pharmacol and molecular mechanism for idiopathic ventricu-
23. Abriel H, Cabo C, Wehrens XH, et al: Novel 1:144, 2010. lar fibrillation. Nature 392:293–296, 1998.
arrhythmogenic mechanism revealed by a 43. Besana A, Wang DW, George AL Jr, et al: 62. Akai J, Makita N, Sakurada H, et al: A novel
long-QT syndrome mutation in the cardiac Na+ Nadolol block of NaV1.5 does not explain its effi- SCN5A mutation associated with idiopathic ven-
channel. Circ Res 88:740–745, 2001. cacy in the long QT syndrome. J Cardiovasc tricular fibrillation without typical ECG findings
24. Horne AJ, Eldstrom J, Sanatani S, et al: A novel Pharmacol 59:249–253, 2011. of Brugada syndrome. FEBS Lett 479:29–34,
mechanism for LQT3 with 2:1 block: A pore- 44. Bankston JR, Kass RS: Molecular determinants of 2000.
lining mutation in Nav1.5 significantly affects local anesthetic action of beta-blocking drugs: 63. Brugada J, Brugada P: Further characterization of
voltage-dependence of activation. Heart Rhythm Implications for therapeutic management of long the syndrome of right bundle branch block, ST
8:770–777, 2011. QT syndrome variant 3. J Mol Cell Cardiol segment elevation, and sudden cardiac death.
25. Medeiros-Domingo A, Kaku T, Tester DJ, et al: 48:246–253, 2010. J Cardiovasc Electrophysiol 8:325–331, 1997.
SCN4B-encoded sodium channel b4 subunit in 45. Hofbeck M, Ulmer H, Beinder E, et al: Prenatal 64. Brugada R, Brugada J, Antzelevitch C, et al:
congenital long-QT syndrome. Circulation findings in patients with prolonged QT interval Sodium channel blockers identify risk for sudden
116:134–142, 2007. in the neonatal period. Heart 77:198–204, 1997. death in patients with ST-segment elevation and
26. Vatta M, Ackerman MJ, Ye B, et al: Mutant 46. Bankston JR, Yue M, Chung W, et al: A novel and right bundle branch block but structurally normal
caveolin-3 induces persistent late sodium current lethal de novo LQT-3 mutation in a newborn with hearts. Circulation 101:510–515, 2000.
and is associated with long-QT syndrome. Circu- distinct molecular pharmacology and therapeutic 65. Keller DI, Huang H, Zhao J, et al: A novel
lation 114:2104–2112, 2006. response. PLoS ONE 2:e1258, 2007. SCN5A mutation, F1344S, identified in a patient
27. Cheng J, Van Norstrand DW, Medeiros- 47. Kehl HG, Haverkamp W, Rellensmann G, et al: with Brugada syndrome and fever-induced ven-
Domingo A, et al: Alpha1-syntrophin mutations Images in cardiovascular medicine: Life- tricular fibrillation. Cardiovasc Res 70:521–529,
identified in sudden infant death syndrome cause threatening neonatal arrhythmia: Successful 2006.
an increase in late cardiac sodium current. Circ treatment and confirmation of clinically suspected 66. Nademanee K, Veerakul G, Nimmannit S, et al:
Arrhythm Electrophysiol 2:667–676, 2009. extreme long QT-syndrome-3. Circulation Arrhythmogenic marker for the sudden unex-
28. Ueda K, Valdivia C, Medeiros-Domingo A, et al: 109:e205–e206, 2004. plained death syndrome in Thai men. Circulation
Syntrophin mutation associated with long QT 48. Wedekind H, Smits JP, Schulze-Bahr E, et al: De 96:2595–2600, 1997.
syndrome through activation of the nNOS- novo mutation in the SCN5A gene associated with 67. Cordeiro JM, Barajas-Martinez H, Hong K, et al:
SCN5A macromolecular complex. Proc Natl early onset of sudden infant death. Circulation Compound heterozygous mutations P336L and
Acad Sci U S A 105:9355–9360, 2008. 104:1158–1164, 2001. I1660V in the human cardiac sodium channel
29. Clancy CE, Rudy Y: Linking a genetic defect to 49. Chang CC, Acharfi S, Wu MH, et al: A novel associated with the Brugada syndrome. Circula-
its cellular phenotype in a cardiac arrhythmia. SCN5A mutation manifests as a malignant form tion 114:2026–2033, 2006.
Nature 400:566–569, 1999. of long QT syndrome with perinatal onset of 68. Frigo G, Rampazzo A, Bauce B, et al: Homozy-
30. Fabritz L, Kirchhof P, Franz MR, et al: Effect of tachycardia/bradycardia. Cardiovasc Res 64:268– gous SCN5A mutation in Brugada syndrome with
pacing and mexiletine on dispersion of repolarisa- 278, 2004. monomorphic ventricular tachycardia and struc-
tion and arrhythmias in DeltaKPQ SCN5A 50. Ackerman MJ, Siu BL, Sturner WQ, et al: Post- tural heart abnormalities. Europace 9:391–397,
(long QT3) mice. Cardiovasc Res 57:1085–1093, mortem molecular analysis of SCN5A defects in 2007.
2003. sudden infant death syndrome. JAMA 286:2264– 69. Poelzing S, Forleo C, Samodell M, et al: SCN5A
31. Clancy CE, Rudy Y: Na+ channel mutation that 2269, 2001. polymorphism restores trafficking of a Brugada
causes both Brugada and long-QT syndrome phe- 51. Arnestad M, Crotti L, Rognum TO, et al: Preva- syndrome mutation on a separate gene. Circula-
notypes: A simulation study of mechanism. Cir- lence of long-QT syndrome gene variants in tion 114:368–376, 2006.
culation 105:1208–1213, 2002. sudden infant death syndrome. Circulation 70. Shinlapawittayatorn K, Dudash LA, Du XX, et al:
32. Nuyens D, Stengl M, Dugarmaa S, et al: Abrupt 115:361–367, 2007. A novel strategy using cardiac sodium channel
rate accelerations or premature beats cause life- 52. Schwartz PJ, Priori SG, Dumaine R, et al: A polymorphic fragments to rescue trafficking-
threatening arrhythmias in mice with long-QT3 molecular link between the sudden infant death deficient SCN5A mutations. Circ Cardiovasc
syndrome. Nature Med 7:1021–1027, 2001. syndrome and the long-QT syndrome. N Engl J Genet 4:500–509, 2011.
33. Fabritz L, Damke D, Emmerich M, et al: Auto- Med 343:262–267, 2000. 71. Schulze-Bahr E, Eckardt L, Breithardt G, et al:
nomic modulation and antiarrhythmic therapy in 53. Plant LD, Bowers PN, Liu Q, et al: A common Sodium channel gene (SCN5A) mutations in 44
a model of long QT syndrome type 3. Cardiovasc cardiac sodium channel variant associated with index patients with Brugada syndrome: Different
Res 87:60–72, 2010. sudden infant death in African Americans, incidences in familial and sporadic disease. Hum
34. Fredj S, Lindegger N, Sampson KJ, et al: Altered SCN5A S1103Y. J Clin Invest 116:430–435, Mutat 21:651–652, 2003.
Na+ channels promote pause-induced spontane- 2006. 72. Baroudi G, Pouliot V, Denjoy I, et al: Novel
ous diastolic activity in long QT syndrome type 3 54. Wang DW, Desai RR, Crotti L, et al: Cardiac mechanism for Brugada syndrome: Defective
myocytes. Circ Res 99:1225–1232, 2006. sodium channel dysfunction in sudden infant surface localization of an SCN5A mutant
35. Schwartz PJ, Priori SG, Locati EH, et al: Long death syndrome. Circulation 115:368–376, (R1432G). Circ Res 88:E78–E83, 2001.
QT syndrome patients with mutations of the 2007. 73. Valdivia CR, Tester DJ, Rok BA, et al: A traffick-
SCN5A and HERG genes have differential 55. Splawski I, Timothy KW, Tateyama M, et al: ing defective, Brugada syndrome-causing SCN5A
responses to Na+ channel blockade and to Variant of SCN5A sodium channel implicated in mutation rescued by drugs. Cardiovasc Res
increases in heart rate: Implications for gene- risk of cardiac arrhythmia. Science 297:1333– 62:53–62, 2004.
specific therapy. Circulation 92:3381–3386, 1995. 1336, 2002. 74. Clatot J, Ziyadeh-Isleem A, Maugenre S, et al:
36. Wang DW, Yazawa K, Makita N, et al: Pharma- 56. Van Norstrand DW, Tester DJ, Ackerman MJ: Dominant-negative effect of SCN5A N-terminal
cological targeting of long QT mutant sodium Overrepresentation of the proarrhythmic, sudden mutations through the interaction of NaV1.5
channels. J Clin Invest 99:1714–1720, 1997. death predisposing sodium channel polymor- alpha-subunits. Cardiovasc Res 96:53–63, 2012.
37. Windle JR, Geletka RC, Moss AJ, et al: Normal- phism S1103Y in a population-based cohort of 75. Dumaine R, Towbin JA, Brugada P, et al: Ionic
ization of ventricular repolarization with fle- African-American sudden infant death syndrome. mechanisms responsible for the electrocardio-
cainide in long QT syndrome patients with Heart Rhythm 5:712–715, 2008. graphic phenotype of the Brugada syndrome are
temperature dependent. Circ Res 85:803–809, conduction system disease. Am J Physiol Heart 112. Royer A, van Veen TA, Le BS, et al: Mouse model
1999. Circ Physiol 292:H399–H407, 2007. of SCN5A-linked hereditary Lenegre’s disease:
76. Wang DW, Makita N, Kitabatake A, et al:
Enhanced Na+ channel intermediate inactivation
in Brugada syndrome. Circ Res 87:E37–E43,
95. Coronel R, Casini S, Koopmann TT, et al: Right
ventricular fibrosis and conduction delay in a
patient with clinical signs of Brugada syndrome:
Age-related conduction slowing and myocardial
fibrosis. Circulation 111:1738–1746, 2005.
113. Leoni AL, Gavillet B, Rougier JS, et al: Variable
50
2000. A combined electrophysiological, genetic, histo- NaV1.5 protein expression from the wild-type
77. Rook MB, Alshinawi CB, Groenewegen WA, et pathologic, and computational study. Circulation allele correlates with the penetrance of cardiac
al: Human SCN5A gene mutations alter cardiac 112:2769–2777, 2005. conduction disease in the Scn5a(+/−) mouse
sodium channel kinetics and are associated with 96. Postema PG, van Dessel PF, de Bakker JM, et al: model. PLoS One 19:e9298, 2010.
the Brugada syndrome. Cardiovasc Res 44:507– Slow and discontinuous conduction conspire in 114. Gui J, Wang T, Jones RP, et al: Multiple loss-of-
517, 1999. Brugada syndrome: A right ventricular mapping function mechanisms contribute to SCN5A-
78. Mohler PJ, Rivolta I, Napolitano C, et al: NaV1.5 and stimulation study. Circ Arrhythm Electro- related familial sick sinus syndrome. PLoS One
E1053K mutation causing Brugada syndrome physiol 1:379–386, 2008. 5:e10985, 2010.
blocks binding to ankyrin-G and expression of 97. Postema PG, van Dessel PF, Kors JA, et al: Local 115. Butters TD, Aslanidi OV, Inada S, et al: Mecha-
Nav1.5 on the surface of cardiomyocytes. Proc depolarization abnormalities are the dominant nistic links between Na+ channel (SCN5A)
Natl Acad Sci U S A 101:17533–17538, 2004. pathophysiologic mechanism for type 1 electro- mutations and impaired cardiac pacemaking in
79. Meregalli PG, Tan HL, Probst V, et al: Type of cardiogram in Brugada syndrome: A study of elec- sick sinus syndrome. Circ Res 107:126–137,
SCN5A mutation determines clinical severity and trocardiograms, vectorcardiograms, and body 2010.
degree of conduction slowing in loss-of-function surface potential maps during ajmaline provoca- 116. Lei M, Zhang H, Grace AA, et al: SCN5A and
sodium channelopathies. Heart Rhythm 6:341– tion. J Am Coll Cardiol 55:789–797, 2010. sinoatrial node pacemaker function. Cardiovasc
348, 2009. 98. Nademanee K, Veerakul G, Chandanamattha P, Res 74:356–365, 2007.
80. Watanabe H, Koopmann TT, Le SS, et al: et al: Prevention of ventricular fibrillation epi- 117. McNair WP, Ku L, Taylor MR, et al: SCN5A
Sodium channel beta1 subunit mutations associ- sodes in Brugada syndrome by catheter ablation mutation associated with dilated cardiomyopathy,
ated with Brugada syndrome and cardiac conduc- over the anterior right ventricular outflow conduction disorder, and arrhythmia. Circulation
tion disease in humans. J Clin Invest 118: tract epicardium. Circulation 123:1270–1279, 110:2163–2167, 2004.
2260–2268, 2008. 2011. 118. Olson TM, Michels VV, Ballew JD, et al: Sodium
81. Hu D, Barajas-Martinez H, Burashnikov E, et al: 99. Papadatos GA, Wallerstein PM, Head CE, et al: channel mutations and susceptibility to heart
A mutation in the beta 3 subunit of the cardiac Slowed conduction and ventricular tachycardia failure and atrial fibrillation. JAMA 293:447–454,
sodium channel associated with Brugada ECG after targeted disruption of the cardiac sodium 2005.
phenotype. Circ Cardiovasc Genet 2:270–278, channel gene Scn5a. Proc Natl Acad Sci U S A 119. Laurent G, Saal S, Amarouch MY, et al:
2009. 99:6210–6215, 2002. Multifocal ectopic Purkinje-related premature
82. Kattygnarath D, Maugenre S, Neyroud N, et al: 100. Martin CA, Zhang Y, Grace AA, et al: In vivo contractions: a new SCN5A-related cardiac chan-
MOG1: A new susceptibility gene for Brugada studies of Scn5a+/− mice modeling Brugada syn- nelopathy. J Am Coll Cardiol 60:144–156,
syndrome. Circ Cardiovasc Genet 4:261–268, drome demonstrate both conduction and repolar- 2012.
2011. ization abnormalities. J Electrocardiol 43:433–439, 120. Nair K, Pekhletski R, Harris L, et al: Escape
83. London B, Michalec M, Mehdi H, et al: Mutation 2010. capture bigeminy: phenotypic marker of cardiac
in glycerol-3-phosphate dehydrogenase 1 like 101. Martin CA, Zhang Y, Grace AA, et al: Increased sodium channel voltage sensor mutation R222Q.
gene (GPD1-L) decreases cardiac Na+ current right ventricular repolarization gradients promote Heart Rhythm 9:1681–1688, 2012.
and causes inherited arrhythmias. Circulation arrhythmogenesis in a murine model of Brugada 121. Gosselin-Badaroudine P, Keller DI, Huang H, et
116:2260–2268, 2007. syndrome. J Cardiovasc Electrophysiol 21:1153– al: A proton leak current through the cardiac
84. Valdivia CR, Ueda K, Ackerman MJ, et al: 1159, 2010. sodium channel is linked to mixed arrhythmia and
GPD1L links redox state to cardiac excitability by 102. Schott JJ, Alshinawi C, Kyndt F, et al: Cardiac the dilated cardiomyopathy phenotype. PLoS
PKC-dependent phosphorylation of the sodium conduction defects associate with mutations in ONE 7:e38331, 2012.
channel SCN5A. Am J Physiol Heart Circ Physiol SCN5A. Nat Genet 23:20–21, 1999. 122. Cheng J, Morales A, Siegfried JD, et al: SCN5A
297:H1446–H1452, 2009. 103. Wang DW, Viswanathan PC, Balser JR, et al: rare variants in familial dilated cardiomyopathy
85. Liu M, Sanyal S, Gao G, et al: Cardiac Na+ Clinical, genetic, and biophysical characterization decrease peak sodium current depending on the
current regulation by pyridine nucleotides. Circ of SCN5A mutations associated with atrioven- common polymorphism H558R and common
Res 105:737–745, 2009. tricular conduction block. Circulation 105:341– splice variant Q1077del. Clin Transl Sci 3:287–
86. Moreau A, Keller DI, Huang H, et al: Mexiletine 346, 2002. 294, 2010.
differentially restores the trafficking defects 104. Groenewegen WA, Firouzi M, Bezzina CR, et al: 123. Rampersaud E, Siegfried JD, Norton N, et al:
caused by two Brugada syndrome mutations. A cardiac sodium channel mutation cosegregates Rare variant mutations identified in pediatric
Front Pharmacol 3:62, 2012. with a rare connexin40 genotype in familial atrial patients with dilated cardiomyopathy. Prog
87. Teng S, Gao L, Paajanen V, et al: Read through standstill. Circ Res 92:14–22, 2003. Pediatr Cardiol 31:39–47, 2011.
of nonsense mutation W822X in the SCN5A 105. Benson DW, Wang DW, Dyment M, et al: Con- 124. McNair WP, Sinagra G, Taylor MR, et al: SCN5A
gene can effectively restore expression of cardiac genital sick sinus syndrome caused by recessive mutations associate with arrhythmic dilated car-
Na+ channels. Cardiovasc Res 83:473–480, mutations in the cardiac sodium channel gene diomyopathy and commonly localize to the
2009. (SCN5A). J Clin Invest 112:1019–1028, 2003. voltage-sensing mechanism. J Am Coll Cardiol
88. Meregalli PG, Wilde AA, Tan HL: Pathophysi- 106. Lopez KN, Decker JA, Friedman RA, et al: 57:2160–2168, 2011.
ological mechanisms of Brugada syndrome: Homozygous mutation in SCN5A associated with 125. Morales A, Painter T, Li R, et al: Rare variant
Depolarization disorder, repolarization disorder, atrial quiescence, recalcitrant arrhythmias, and mutations in pregnancy-associated or peripartum
or more? Cardiovasc Res 67:367–378, 2005. poor capture thresholds. Heart Rhythm 8:471– cardiomyopathy. Circulation 121:2176–2182,
89. Wilde AA, Postema PG, Di Diego JM, et al: The 473, 2011. 2010.
pathophysiological mechanism underlying 107. Makita N, Sasaki K, Groenewegen WA, et al: 126. Nguyen TP, Wang DW, Rhodes TH, et al: Diver-
Brugada syndrome: Depolarization versus repo- Congenital atrial standstill associated with coin- gent biophysical defects caused by mutant sodium
larization. J Mol Cell Cardiol 49:543–553, heritance of a novel SCN5A mutation and con- channels in dilated cardiomyopathy with arrhyth-
2010. nexin 40 polymorphisms. Heart Rhythm mia. Circ Res 102:364–371, 2008.
90. Antzelevitch C, Yan GX, Shimizu W: Transmural 2:1128–1134, 2005. 127. Wang DW, Yazawa K, George AL, et al:
dispersion of repolarization and arrhythmogenic- 108. Lenègre J: Etiology and pathology of bilateral Characterization of human cardiac Na+ channel
ity: The Brugada syndrome versus the long QT bundle branch block in relation to complete mutations in the congenital long QT syndrome.
syndrome. J Electrocardiol 32(Suppl):158–165, heart block. Prog Cardiovasc Dis 6:409–444, Proc Natl Acad Sci U S A 93:13200–13205,
1999. 1964. 1996.
91. Yan GX, Antzelevitch C: Cellular basis for the 109. Wolf CM, Berul CI: Inherited conduction system 128. Watanabe H, Yang T, Stroud DM, et al: Striking
Brugada syndrome and other mechanisms of abnormalities—one group of diseases, many in vivo phenotype of a disease-associated human
arrhythmogenesis associated with ST-segment genes. J Cardiovasc Electrophysiol 17:446–455, SCN5A mutation producing minimal changes in
elevation. Circulation 100:1660–1666, 1999. 2006. vitro. Circulation 124:1001–1011, 2011.
92. Antzelevitch C: Ion channels and ventricular 110. Tan HL, Bink-Boelkens MT, Bezzina CR, et al: 129. Laitinen-Forsblom PJ, Makynen P, Makynen H,
arrhythmias: Cellular and ionic mechanisms A sodium-channel mutation causes isolated et al: SCN5A mutation associated with cardiac
underlying the Brugada syndrome. Curr Opin cardiac conduction disease. Nature 409:1043– conduction defect and atrial arrhythmias. J Car-
Cardiol 14:274–279, 1999. 1047, 2001. diovasc Electrophysiol 17:480–485, 2006.
93. Tukkie R, Sogaard P, Vleugels J, et al: Delay in 111. Hao X, Zhang Y, Zhang X, et al: TGF-beta1- 130. Bezzina C, Veldkamp MW, van Den Berg MP, et
right ventricular activation contributes to Brugada mediated fibrosis and ion channel remodeling are al: A single Na+ channel mutation causing both
syndrome. Circulation 109:1272–1277, 2004. key mechanisms in producing the sinus node dys- long-QT and Brugada syndromes. Circ Res
94. Zhang ZS, Tranquillo J, Neplioueva V, et al: function associated with SCN5A deficiency and 85:1206–1213, 1999.
Sodium channel kinetic changes that produce aging. Circ Arrhythm Electrophysiol 4:397–406, 131. Makita N, Behr E, Shimizu W, et al: The E1784K
Brugada syndrome or progressive cardiac 2011. mutation in SCN5A is associated with mixed
clinical phenotype of type 3 long QT syndrome. 133. Grant AO, Carboni MP, Neplioueva V, et al: duction defect or Brugada syndrome in a large
J Clin Invest 118:2219–2229, 2008. Long QT syndrome, Brugada syndrome, and French family. Circulation 104:3081–3086, 2001.
132. Remme CA, Verkerk AO, Nuyens D, et al: conduction system disease are linked to a single 135. Remme CA, Scicluna BP, Verkerk AO, et al:
Overlap syndrome of cardiac sodium channel sodium channel mutation. J Clin Invest 110:1201– Genetically determined differences in sodium
disease in mice carrying the equivalent mutation 1209, 2002. current characteristics modulate conduction
of human SCN5A-1795insD. Circulation 114: 134. Kyndt F, Probst V, Potet F, et al: Novel SCN5A disease severity in mice with cardiac sodium chan-
2584–2594, 2006. mutation leading either to isolated cardiac con- nelopathy. Circ Res 104:1283–1292, 2009.
Inheritable Potassium Channel Disease 51
Amin Al-Ahmad and Arthur A.M. Wilde
Four subtypes of LQTS (LQT1, LQT2, LQT7, and LQT13)

CHAPTER OUTLINE
are linked to mutations in genes encoding for the pore-forming
Long QT Syndrome 501 α-subunits of voltage-gated cardiac K+ channels (Kv channels),
and three subtypes (LQT5, LQT6, and LQT11) are linked to
Short QT Syndrome 506
mutations in genes encoding for one of the regulatory subunits
Sudden Infant Arrhythmia Death Syndrome 507 of Kv channels. Most mutations involve single-nucleotide substi-
tutions in the coding regions (exons) of genes that alter a codon,
Other Potassium Channel Diseases 507
leading to the replacement of one amino acid by a different one
(missense mutations; approximately two-thirds of all mutations)
Potassium (K+) channels are necessary for proper functioning of or the creation of an early (premature) stop codon resulting in
cardiac myocytes and many other human cell types. In the mid- the formation of a truncated protein (nonsense mutations).
1990s, the human ether-a-go-go related gene (HERG; now referred Single-nucleotide substitutions in the noncoding regions (introns)
to as KCNH2) and KVLQT1 (now referred to as KCNQ1), each also occur and may result in altered gene transcripts. Specific
encoding for a major cardiac K+ channel, were among the first intronic nucleotide sequences at the intron/exon (acceptor site)
genes linked to an inheritable arrhythmia syndrome.1,2 Ever and exon/intron (donor site) boundaries are crucial for the
since, many other genes encoding for cardiac K+ channels or their process whereby introns are excised from gene transcripts to
accessory (regulatory) subunits have been implicated in the etiol- create mature protein-encoding mRNAs (i.e., the splicing
ogy of various inheritable arrhythmia syndromes, including the process). Mutations within these highly conserved intronic
long QT syndrome, the short QT syndrome, sudden infant death regions may cause aberrant splicing, leading to the deletion of
syndrome, catecholaminergic polymorphic ventricular tachycar- entire exons (or exon parts) or inclusion of entire introns (or
dia, the Brugada syndrome, and familial atrial fibrillation (Table intron parts) in the mature mRNA, thereby often altering the
51-1).3 A detailed description of the molecular biology and physi- open reading frame of translation and generating a new sequence
ology of cardiac K+ channels is provided elsewhere in this book. of amino acids in the final product (i.e., frameshift). Mutations
This chapter focuses on the molecular genetics of inheritable also involve insertion or deletion of one or more nucleotides,
diseases related to cardiac K+ channels, so-called K+ channelopa- which may lead to a shift in the open reading frame, or (when a
thies. Clinical features and management of these diseases are multiple of three nucleotides are inserted or deleted) to the addi-
described in detail elsewhere in this book. tion or removal of one or more amino acids in the final product,
without affecting the reading frame.4
Experimental investigation of mutated K+ channels (or a
normal channel with a mutated regulatory subunit) in heterolo-
Long QT Syndrome gous expression systems such as Xenopus oocytes or mammalian
cell lines has unequivocally demonstrated that LQTS-related
The long QT syndrome (LQTS) is a cardiac arrhythmogenic mutations delay repolarization by reducing the outward K+
disease that causes syncope and sudden death due to torsades de current through the affected channel (loss-of-function).3 It is
pointes, a characteristic form of polymorphic ventricular tachy- important to note that loss-of-function effects of LQTS-related
cardia, and ventricular fibrillation. Inheritable (congenital) LQTS mutations in KCNQ1 and KCNH2 (causing LQT1 and LQT2,
is characterized by a prolonged heart rate corrected QT interval respectively) have also been shown in the much more native
(QTc duration) on a 12-lead electrocardiogram (ECG) due to a environment of cardiac myocytes derived from patient-specific
delay in repolarization of the ventricles, in the absence of struc- pluripotent stem cells from members of families affected with
tural heart diseases or secondary causes of a prolonged QTc these diseases.5,6 In general, mutations cause loss-of-function by
duration, such as electrolyte abnormalities, hypothermia, and use decreasing the number of functional channels in the sarcolemma
of certain drugs. Inheritable LQTS with an autosomal dominant (lower expression) or by disrupting the extent or speed of channel
pattern of inheritance has an estimated prevalence of 1 : 2500, and opening and closing (altered gating). Because α-subunits of Kv
is traditionally called the Romano-Ward syndrome (first described channels assemble as tetramers to form functional ion channels,
independently by the Italian pediatrician Cesarino Romano in it is of functional importance whether mutated channel proteins
1963 and the Irish pediatrician O. Connor Ward in 1964). The possess the ability to assemble with normal channel proteins in
very rare autosomal recessive form of LQTS, with concomitant heterozygous mutation carriers where both normal and mutated
congenital bilateral sensory neural deafness, is called the Jervell channels coexist (a normal allele is inherited from one parent and
and Lange-Nielsen syndrome (first described by Anton Jervell a mutant allele is inherited from the other parent). When mutated
and his associate Fred Lange-Nielsen in 1957).3 The autosomal subunits assemble with normal subunits, they can disturb the
dominant form is subdivided into different subtypes (so far 13 sarcolemmal expression or gating of the normal channel subunits
subtypes) according to the affected gene. The subdivision is based (i.e., a dominant-negative effect). In this case, the overall loss of
on the chronological order in which the subtypes are reported. K+ current will exceed 50%. In contrast, when mutant subunits
This subdivision, which is based on underlying molecular genetic do not participate in tetramer assembly, 50% reduction (maxi-
and pathophysiological mechanisms, is not ideal, but is followed mally) in the K+ current is anticipated (i.e., haploinsufficiency).
in this chapter. As expected, dominant-negative LQTS-linked mutations have a
501
Table 51-1. Genes Related to Inheritable Potassium Channel Diseases
Current Chromosome Gene Protein Protein Function Disease
Ito,fast 1p13.3 KCND3 Kv4.3 α-Subunit BrS

11q13.4 KCNE3 MiRP2 β-Subunit BrS
Xq22.3 KCNE5 MiRP4 β-Subunit BrS, FAF
IKs 11p15.5 KCNQ1 Kv7.1 α-Subunit LQT1, JLNS, SQT2, SIDS, FAF
21q22.12 KCNE1 MinK β-Subunit LQT5, JLNS, SIDS
Xq22.3 KCNE5 MiRP4 β-Subunit FAF
7q21-q22 AKAP9 AKAP9 Regulatory subunit LQT11
IKr 7q36.1 KCNH2 Kv11.1 α-Subunit LQT2, SQT1, SIDS, BrS, FAF
21q22.12 KCNE2 MiRP1 β-Subunits LQT6, SIDS, FAF
IK1 17q24.3 KCNJ2 Kir2.1 α-Subunit LQT7, SQT3, CPVT, FAF
IKur 12p13 KCNA5 Kv1.5 α-Subunit FAF
IK-ATP 12p11.23 KCNJ8 Kir6.1 α-Subunit SIDS, BrS
IK-ACh 11q24 KCNJ5 Kir3.4 α-Subunit LQT13
BrS, Brugada syndrome; CPVT, catecholaminergic polymorphic ventricular tachycardia; FAF, familial atrial fibrillation; JLNS, Jervell and Lange-Nielsen syndrome; LQT, long
QT syndrome type; SIDS, sudden infant death syndrome; SQT, short QT syndrome type.
worse clinical outcome than mutations that cause haploinsuffi- and protein kinase A (PKA), requiring A-kinase anchoring pro-
ciency.7,8 In addition, LQTS patients with compound mutations teins (AKAPs). As a result, IKs enables the physiological response
(≈8% of all genotyped patients) in the same gene, but particularly (i.e., abbreviation) of repolarization to fast heart rates during
in different LQTS-related genes, are shown to display a more sympathetic nerve activity.11 In 1991, two linkage studies linked
severe phenotype than single mutation carriers.9 These clinical a gene locus on chromosome 11 to LQTS in several unrelated
data indicate that the extent of decrease in cardiac K+ currents families. Five years later, positional cloning techniques estab-
may determine disease severity in LQTS. lished KCNQ1 as the chromosome 11–linked LQT1 gene.2 In
A delicate balance between inward and outward ion currents 1997, targeted mutational analysis of KCNE1 in two families with
determines the repolarization of ventricular myocytes. Substan- LQTS identified two missense mutations in KCNE1 (LQT5).12
tial differences in the expression levels of K+ channels between LQT1 is now known to account for nearly 40% of all LQTS
cardiac myocytes in different cardiac regions create a spatial dis- cases. LQT5 is rare, and mutations in KCNE1 may be responsible
persion of repolarization within the healthy ventricular myocar- for nearly 3% of all LQTS cases.13-15 In 2007, targeted mutational
dium. In LQTS, K+ current reduction leads to prolongation of analysis of the translated exons of AKAP9, encoding the A kinase
the action potential plateau phase (reflected as QT interval pro- anchoring protein type 9 (AKAP9; also called Yatiao), identified
longation on the ECG), and this allows recovery from inactiva- a missense mutation in a single patient with LQTS (LQT11). So
tion and reactivation of L-type Ca2+ channels, which produces far, only this one mutation in Yatiao has been linked to LQTS.16
early afterdepolarizations (EADs). EADs, together with an
accentuated spatial dispersion of repolarization, underlie the sub-
strate and the trigger for the development of torsades de pointes Long QT Syndrome Type 1
in LQTS.10 The consensus is that EADs initiate torsades de
pointes, and EADs are probably the initiating event in LQT2, To date, more than 300 mutations in KCNQ1 have been linked
the second most prevalent type of LQTS. However, in LQT1, to LQT1. Most mutations are missense mutations (≈70%), fol-
the most common type of LQTS, where arrhythmic events lowed by frameshift mutations (≈10%), splice-site mutations
usually and predictably start at higher heart rates, the arrhyth- (≈10%), nonsense mutations (≈5%), and in-frame deletions or
mogenic mechanism might be different. Delayed afterdepolariza- insertions (≈5%).13-15 Novel mutation detection methods have
tions (DADs), spontaneous action potentials during phase 4 of demonstrated large genomic rearrangements (i.e., copy number
the cardiac action potential, might play a role in this condition. variants) in KCNQ1, leading to the complete deletion of one or
DADs may also trigger ventricular arrhythmias in LQT7 (see more exons in LQTS patients who were mutation negative after
later). traditional point mutation analyses.17 This study suggests that the
prevalence of copy number variants in LQT1 may be higher
than was previously thought. LQT1 mutations in Kv7.1 are
IKs-Related Long QT Syndrome mostly located in the transmembrane segments (≈60%), the
intracellular C-terminus (≈30%), and the intracellular N-terminus
LQTS type 1 (LQT1), type 5 (LQT5), and type 11 (LQT11) are (≈10%).13-15 A fraction of these mutations have been investigated
linked to mutations that cause a reduction in the slowly activating in experimental settings, and these studies have revealed an array
delayed rectifier K+ current (IKs) in the heart. The α-subunit of of molecular mechanisms that underlie IKs loss-of-function
the channel responsible for IKs (Kv7.1) is encoded by KCNQ1, and (Figure 51-1) including (1) defective Kv7.1 protein synthesis, (2)
Kv7.1 proteins require the presence of their regulatory β-subunit defective trafficking of mutated Kv7.1 proteins to the sarcolemma
MinK (encoded by KCNE1) to conduct IKs. IKs is markedly and their retention in the endoplasmic reticulum, (3) impaired
enhanced by β-adrenergic stimulation through phosphorylation ability of mutated proteins to coassemble into tetrameric chan-
of Kv7.1 channels by protein kinase C (PKC), requiring MinK, nels, (4) altered biophysical properties, (5) disrupted interaction
Outward
Loss-of-function Channel K+ current
Gain-of-function
51
Altered gating
Altered selectivity Altered gating
Reduced K+ permeation
Altered channel modulation
K+
K+ K+ K+ +
K
K+ K+ K+
Defective trafficking
Golgi apparatus
Protein
Reduced translation
Increased mRNA degradation
Endoplasmic reticulum
Ribosome
Cytosol
Reduced transcription
mRNA
DNA
Nucleus
Figure 51-1. Common molecular mechanisms responsible for cardiac potassium channel loss-of-function or gain-of-function in inherited potassium channel diseases.
with regulatory proteins, and (6) defective endosomal recy- in heart rate (immediately at standing from supine position or at
cling.3,18-20 Of note, these mechanisms are not mutually exclusive. peak exercise, and during the recovery phase of treadmill exercise
Loss-of-function alterations in biophysical properties of mutated testing).22,23 In healthy subjects, QT intervals shorten at faster
IKs channels involve a slower rate of channel activation, a shift in heart rates, enabling QTc to remain within normal limits with
the voltage dependence of activation toward more depolarized decreasing R-R intervals. Moreover, QTc duration lengthening
membrane potentials (indicating later channel activation), and is also observed in LQT1 patients in response to the intravenous
accelerated deactivation (indicating faster channel closing). Dis- infusion of epinephrine.24 These clinical features indicate loss of
rupted interaction of mutated Kv7.1 proteins with the key regula- an adequate compensatory response of IKs to β-adrenergic stimu-
tory protein AKAP9 has been demonstrated to reduce Kv7.1 lation and stress hormones, most probably as the result of reduced
phosphorylation by PKA upon β-adrenergic stimulation.19 Dis- phosphorylation and disrupted endosomal recycling of mutated
rupted interaction with phosphatidylinositol-4,5-bisphosphate Kv7.1 channels. As expected, antiadrenergic therapies such as
(PIP2) results in reduced PKC-mediated phosphorylation of the β-blockers and left stellate ganglion ablation have great efficacy
Kv7.1 proteins. PIP2, a sarcolemmal lipid, increases IKs activity in LQT1 patients.25,26 β-Blockers have been shown to diminish
through phosphorylation.20 IKs is also increased by the stress QTc changes during exercise or standing and significantly reduce
hormone cortisol through the action of the serum- and the rate of cardiac events.22,25
glucocorticoid-inducible kinase 1 (SGK1). Cortisol upregulates Molecular genetics may be useful not only for diagnostic pur-
SGK1, thereby facilitating endosomal recycling of Kv7.1 chan- poses, but also for risk stratification in LQT1. In a multicenter
nels and stimulating their insertion into the sarcolemma. Muta- study in 600 LQT1 patients, significantly higher rates of cardiac
tions in Kv7.1 can disturb this process, causing further IKs events were found in patients with mutations located in trans-
reduction upon stimulation of SGK1 by cortisol.18 membrane segments or with mutations that exert dominant-
Consistent with the molecular signaling pathways, LQT1- negative effects on normal IKs channel subunits.7 Another large
related cardiac events occur often during exercise (in particular multicenter study associated mutations in highly conserved
swimming) and psychological stress, when the adrenergic tone amino acid residues in Kv7.1 with significant risk of cardiac
and plasma levels of cortisol are increased.21 In addition, the QT events.27 Moreover, a recent study in 387 LQT1 patients corre-
interval fails to shorten appropriately, and might even lengthen lated clinical phenotype with changes in biophysical properties
(paradoxical QT response), in LQT1 patients upon an increase of Kv7.1 channels caused by different KCNQ1 mutations.
Especially, a slower rate of channel activation was associated with accelerating deactivation.37 Other mechanisms include defective
increased risk for events in LQT1.28 In all these studies, the trafficking of mutated MinK proteins (and thereby Kv7.1 pro-
effects of the mutations were independent of traditional risk teins), impaired channel assembly, and reduced sensitivity to
factors (i.e., QTc, female gender, and β-blocker therapy). Muta- PIP2.20,37 Residues in the C-terminus of MinK are identified as
tion type and location may be used to predict whether a KCNQ1 key determinants for sensitivity of Kv7.1 to PIP2, and LQT5-
mutation is pathogenic or is an innocuous rare variant.29 This is linked mutations in these residues are shown to reduce PIP2-
clinically relevant in that large overlap of QTc values has been mediated phosphorylation of Kv7.1 proteins by PKC.20 Because
noted between patients with LQTS and healthy patients. Non- of its low prevalence, genotype-phenotype correlations in LQT5
missense mutations, regardless of location, have an estimated are not available but may resemble those observed in LQT1.
predictive value of more than 99% to be pathogenic, and mis- An SNP in KCNE2, D85N, located in the C-terminus of
sense mutations have a high predictive value when located in the MinK, has been associated with QT interval duration in the
transmembrane segments, the pore loop, and the C-terminus of general population,33,34 and is shown to be more prevalent in
Kv7.1 proteins.29 LQT1 patients with a mutation in the cytoplas- (genotype-negative) LQTS patients.38 Heterologous expression
mic loops are at higher risk for lethal arrhythmias than are studies revealed significant loss-of-function effects of D85N on
patients with a mutation in other regions of Kv7.1, probably KCNQ1-encoded and KCNH2-encoded currents.38 Therefore,
because of a pronounced reduction in channel activation upon D85N is suggested to be a disease-causing variant in LQTS, and
β-adrenergic stimulation.30 certainly a modifier of phenotype in LQT1 and LQT2 patients.
Recent studies have identified single-nucleotide polymor-
phisms (SNPs) in the nitric oxide 1 adaptor protein gene
(NOS1AP) as strong modulators of QT interval duration and risk Long QT Syndrome Type 11
of cardiac events in LQT1 (and other LQTS types).31,32 First,
genome-wide association studies associated SNPs in NOS1AP AKAP9 (Yatiao) has an N-terminal and a C-terminal binding
with QT interval in the general population. Next, a role for domain that interacts with the C-terminus of Kv7.1. The only
NOS1AP SNPs was found in sudden cardiac death in the general LQT11 mutation described so far (S1570L) has been found in 1
population.33,34 In 2009, a family-based association analysis linked out of 50 genotype-negative unrelated LQTS patients. S1570L
SNPs in NOS1AP with QT interval prolongation and risk for is located in the C-terminal binding domain of AKAP9. It has
cardiac arrest and sudden death in a large South African cohort been shown to disrupt, but not eliminate, the interaction between
of LQT1 patients with the A341V mutation.31 NOS1AP encodes AKAP9 and Kv7.1, leading to reduced cyclic adenosine mono-
CAPON, an accessory protein of the neural nitric oxide synthase, phosphate (cAMP)-mediated phosphorylation of Kv7.1 by PKA
which controls intracellular nitric oxide production. Functional during β-adrenergic stimulation. This is speculated to result in
studies indicate that CAPON fastens cardiac repolarization less IKs enhancement during sympathetic nerve activity.16
through inhibition of L-type Ca2+ channels.35 This provides a
rationale for the association of SNPs in NOS1AP with QT inter-
val duration. SNPs in the 3′-untranslated region (3′UTR) of Jervell and Lange-Nielsen Syndrome
KCNQ1 may also modulate phenotype severity in LQT. In a
recent study in two independent LQT1 cohorts from the Neth- Jervell and Lange-Nielsen syndrome (JLNS), the autosomal
erlands and the United States, SNPs in KCNQ1’s 3′UTR were recessive form of LQTS affecting 3 to 5 in 1 million children, is
associated with QTc duration and symptoms in an allele-specific characterized by QT interval prolongation, congenital bilateral
manner. Patients with derived SNP variants on their mutated sensory neural deafness, torsades de pointes, and sudden cardiac
KCNQ1 allele had shorter QTc and fewer symptoms, while death from very early in life. JLNS depends on homozygous or
patients with these variants on their normal KCNQ1 allele had compound heterozygous mutations on KCNQ1 or KCNE1.
longer QTc and a greater number of symptoms. Experimental Hearing loss is due to dysfunction of IKs channels in the inner
studies showed that the expression of KCNQ1’s 3′UTR with ear, where they act as K+ charge carriers for sensory transduction
derived SNP variants was less than the expression of 3′UTR with and generation of endocochlear potential in the endolymph.
ancestral SNP variants. The 3′UTR play a crucial regulatory role Most JLNS-linked mutations are found in KCNQ1 (≈90%), and
in gene expression by controlling stability and translation of mutations in KCNE1 are associated with a less severe phenotype
mRNAs. SNPs in this region were suggested to affect this func- and are more prevalent in asymptomatic patients. Although
tion of the 3′UTR, thereby altering gene expression in an allele- exceptions exist, most JLNS-linked mutations are nonsense
specific manner.36 If true, this is expected to be especially relevant mutations, frameshift mutations, and splice-site mutations (i.e.,
when one allele contains a pathogenic mutation. However, these mutations that are expected to disrupt channel assembly and lead
findings await replication in larger LQT1 cohorts before their to haploinsufficiency).39 This may explain the mild phenotype in
clinical use is proposed. heterozygous parents of JLNS patients compared with LQT1
patients. Similar to LQT1, cardiac events are triggered during
exercise and emotional stress. However, compared with LQT1,
Long QT Syndrome Type 5 JLNS is associated with longer QTc durations, greater risk of
cardiac events (at a younger age), and lower efficacy of β-blocker
The KCNE-encoded MinK exerts its regulatory effects on Kv7.1 therapy.39
via interactions between the transmembrane segments and the
C-termini of the two proteins. In general, interactions between
transmembrane segments are believed to be essential for normal IKr-Related Long QT Syndrome
channel activation, while interactions between the C-termini
regulate channel assembly and channel deactivation. LQT5 muta- LQTS type 2 (LQT2) and type 6 (LQT6) are linked to mutations
tions involve mainly missense mutations, although in-frame dele- that cause a reduction in the rapidly activating delayed rectifier
tions, nonsense mutations, and frameshift mutations are also K+ current (IKr) in the heart. The α-subunit of the channel
found.13-15 Most mutations impair the ability of MinK to modu- responsible for IKr (Kv11.1) is encoded by KCNH2, traditionally
late gating properties of Kv7.1, causing a shift in the voltage called the human ether-a-go-go related gene (HERG) because of
dependence of activation toward more depolarized potentials (in its homology to the Drosophila ether-a-go-go (eag) gene. The
particular, mutations in the transmembrane segments) and interaction of Kv11.1 with the KCNE2-encoded β-subunit
(MinK-related peptide 1; MiRP1) is thought to be required for defective trafficking and intracellular retention of mutated Kv11.1
51
normal gating and pharmacologic properties of IKr channels. In proteins include decreased affinity to heat shock protein 70
1994, linkage studies linked a gene locus on chromosome 7 to (Hsp70; known to promote maturation of Kv11.1) and increased
LQTS in nine families, and in 1995, mutations in HERG were affinity to heat shock cognate 70 (Hsc70; which promotes
found to be responsible for LQT2.1 In 1999, missense mutations proteasomal degradation of Kv11.1),46 abnormal ubiquitylation
in KCNE2 were first identified in LQTS patients.40 LQT2 is now of Kv11.1 by ubiquitin ligase Nedd4-2,47 and redistribution of
known to account for nearly 30% of all LQTS cases, and muta- mutated proteins to a microtubule-dependent quality control
tions in KCNE2 may be responsible for less than 1% of all LQTS compartment in the endoplasmic reticulum.48
cases.13-15 In contrast to IKs, which is enhanced significantly by cAMP-
mediated phosphorylation upon β-adrenergic stimulation, cAMP
has only a minor effect on IKr channel gating. The C-terminus
Long QT Syndrome Type 2 of Kv11.1 contains a cyclic nucleotide binding domain (CNBD).
Binding of cAMP to the CNBD leads to a small change in the
The IKr channel is composed of four α-subunits that are encoded voltage dependence of channel activation.49 However, disruption
by two transcripts of KCNH2: HERG1a or the alternative spliced of this minor effect of cAMP on IKr channel gating in the presence
transcript HERG1b.41 HERG1b proteins lack the first 376 amino of a mutation may provide a rationale for the characteristic occur-
acids of hERG1a. To date, more than 200 LQT2-linked muta- rence of arrhythmias in LQT2 upon sudden fright (e.g., unex-
tions in hERG1a are reported,3 and only one missense mutation pected auditory stimuli, other arousal-related stimuli).21 In
in HERG1b has recently been identified in LQT2.41 Most LQT2 agreement with this, QTc duration in LQT2 patients is length-
mutations are missense mutations (≈60%), followed by frameshift ened immediately at standing from supine position or in response
mutations (≈25%), nonsense mutations (≈10%), splice-site muta- to intravenous injection of an epinephrine bolus—conditions
tions (<5%), and in-frame deletions or insertions (≈5%).13-15 associated with abrupt β-adrenergic stimulation.23,50 In contrast,
Large genomic rearrangements (i.e., copy number variants) in QTc changes in LQT2 upon exercise testing or continuous low-
KCNH2 were also found in LQTS patients.17 LQT2-linked muta- dose infusion of epinephrine are variable and transient.22,24 In
tions are located in the intracellular C-terminus (≈40%), trans- accordance, antiadrenergic therapies (β-blockers and left stellate
membrane segments (≈30%), and intracellular N-terminus ganglion ablation) are effective in LQT2, but less so than in
(≈30%) of Kv11.1.13-15 Missense mutations have a predictive value LQT1.25,26
of 100% to be pathogenic when located in the transmembrane Although mutation-specific therapies for LQT2 are not avail-
segments, the linkers, and the pore loop of Kv11.1 proteins.29 A able in clinical settings, experimental studies have suggested
multicenter study in 858 LQT2 patients associated missense various approaches to restore the effects of a mutation on expres-
mutations in the transmembrane pore regions (segment 5, pore sion, trafficking, or gating of Kv11.1 channels. Aminoglycoside
loop, and segment 6) and the N-terminus with significantly antibiotics have been shown to partially restore the functional
greater risk of cardiac events than missense mutations located in expression of full-length proteins for some KCNH2 nonsense
the C-terminus or the transmembrane nonpore regions (seg- mutations by permitting read-through of the premature termina-
ments 1 to 4).8 Moreover, mutations in α-helical domains, where tion codons created by these mutations.51 However, how amino-
the secondary protein structure may be more highly ordered, glycoside antibiotics change stop codon at the molecular level
were associated with greater risk than mutations in β-sheet and the final effects of read-through on channel level are not yet
domains or other locations.8 In contrast, the risk of non-missense known. More important, it is well established that defective traf-
mutations was not location specific. Of note, an SNP in KCNH2, ficking of mutated Kv11.1 proteins may be restored in vitro by
K897T, located in the C-terminus of Kv11.1, has been associated culturing cells at lower incubation temperatures or in the pres-
with QT interval duration in the general population,33,34 and may ence of high-affinity IKr channel-blocking drugs (e.g., E4031,
modify the phenotype in LQT2.42 Heterologous expression cisapride).45 These interventions may stabilize the mutated pro-
studies revealed loss-of-function effects of K897T on KCNH2- teins in configurations that facilitate normal trafficking. However,
encoded currents.43 the use of such drugs is contraindicated in LQT2 patients because
The underlying mechanisms of IKr reduction have been of their potent IKr-blocking properties. Finally, interactions
studied experimentally for only a fraction of all LQT2 mutations, between the N-terminus and pore-forming domains of Kv11.1
and include (1) defective Kv11.1 protein synthesis, (2) nonsense- channels are known to regulate slow deactivation gating. Muta-
mediated decay (NMD), (3) impaired intracellular trafficking, (4) tions within the N-terminus disrupt these interactions and cause
abnormal biophysical properties (e.g., slower channel activation, accelerated deactivation. Application of small molecules contain-
faster inactivation, faster deactivation), and/or (5) reduced K+ ing the N-terminal residues of Kv11.1 has been demonstrated to
selectivity or permeation.3,44,45 NMD is an RNA surveillance restore the deactivation properties of mutated channels.52
mechanism that selectively degrades mRNA transcripts contain- However, further in vivo studies are required to unravel possible
ing premature termination codons as the result of nonsense or future clinical benefits of such interventions.
frameshift mutations.44 NMD may have a reciprocal impact on
LQT2 phenotype: although NMD may prevent mutated prod-
ucts from exerting a dominant-negative effect on normal Kv11.1 Long QT Syndrome Type 6
proteins, it may also lead to degradation of mutated mRNAs that,
if not degraded, can produce partially functional Kv11.1 channels. Only a few mutations in KCNE2 have been associated with
However, defective trafficking of mutated Kv11.1 proteins to the LQT6. Most of them are missense mutations and are located in
sarcolemma is the dominant molecular mechanism of IKr reduc- the single transmembrane segment of the protein.13-15 The pres-
tion in LQT2, in particular for mutations located in highly ence of mutated MiRP2 subunits leads to reduced current densi-
conserved domains of Kv11.1.45 Normally, Kv11.1 proteins are ties and various alterations in the gating properties of Kv11.1
modified through the addition or trimming of sugar moieties in channels, including a shift in voltage dependence of activation
the endoplasmic reticulum (core glycosylation) and in the Golgi toward more positive potentials, faster inactivation, and slower
apparatus (complex glycosylation). Most LQT2-linked mutations recovery from inactivation.40,54,55 Mutated MiRP2 subunits may
disrupt the trafficking of core-glycosylated (“immature”) Kv11.1 also change the pharmacologic response of Kv11.1 channels to
proteins to the Golgi apparatus, resulting in their retention in drugs.40 However, the role of MiRP2 in IKr modulation and ven-
the endoplasmic reticulum. Processes suggested to be involved in tricular repolarization remains controversial, especially because
the expression level of KCNE2 is very low in ventricles and only cardiac death, in the absence of cardiac structural abnormalities,
significantly high in pacemaker cells and Purkinje fibers. in several members of two unrelated families.60 In 2004, SQTS
was first linked to missense mutations in KCNH2,61 and in the
same year, a missense mutation in KCNQ1 was associated with
Long QT Syndrome Type 7 the disease.62 In 2005, a missense mutation in KCNJ2 was linked
to SQTS in a single family.63 Thus, in a short time, three subtypes
The Andersen-Tawil syndrome (ATS), also called Andersen’s syn- of SQTS have been recognized, each subtype associated with a
drome, is an uncommon disease associated with ventricular mutation in a K+ channel-encoding gene: KCNH2 (SQT1),
arrhythmias, potassium-sensitive periodic paralysis, and dysmor- KCNQ1 (SQT2), and KCNJ2 (SQT3). All SQTS-linked muta-
phic features, including short stature, scoliosis, clinodactyly, tions are missense mutations and, when studied in heterologous
hypertelorism, and subtle facial abnormalities.55 ATS is catego- expression systems, cause gain-of-function of corresponding K+
rized as LQTS type 7 (LQT7) because of the presence of promi- channels by altering gating properties. The expected increase in
nent U waves, which are difficult to distinguish from truly IKr (in SQT1), in IKs (in SQT2), or in the outward component of
prolonged QT intervals, and mild QT prolongation. The first IK1 (in SQT3) fastens repolarization and leads to QT interval
ATS cases were reported by Andersen and colleagues in 1971, shortening on ECG.
and Dr. Rabi Tawil has significantly contributed to further
description of the disease.56 In 2001, mutations in KCNJ2, encod-
ing the α-subunit (Kir2.1) of the channel responsible for the Short QT Syndrome Type 1
inward-rectifier K+ current (IK1) in the heart, were linked to
ATS.55 It is now known that KCNJ2 mutations account for ≈60% So far, three different mutations in KCNH2 are linked to SQT1.
of all ATS cases. IK1 is the major determinant of the resting The first two SQT1-linked mutations result in the same amino
membrane potential in working myocardium, and contributes to acid change, N588K, located in the pore loop region of Kv11.1.
the terminal phase of the action potential repolarization. Nearly This mutation caused a shift of inactivation toward much more
all ATS-linked mutations in KCNJ2 lead to the generation of positive potentials.61 Because inactivation is responsible for
nonfunctional channels and a dominant-negative effect on the inward rectification (i.e., a decrease in outward K+ current at
function of normal Kir2.1 proteins.56 The trafficking of mutated sustained depolarization), the shift of inactivation resulted in
Kir2.1 proteins is often normal. Moreover, ATS-linked mutations failure of rectification and an increase in IKr during the plateau
alter the sensitivity of Kir2.1 to PIP2, an essential activator of the phase of the action potential. Another SQT1-linked mutation in
channel, and, in accordance, ≈50% of all mutations are located the pore loop region, T618I, also caused a shift of inactivation
at residues that are crucial for the interaction of Kir2.1 with PIP2. toward more positive potentials, and induced faster recovery
IK1 reduction allows the Na+/Ca2+ exchanger (INa+/Ca2+), the major from inactivation.64 The third SQT1 mutation, R1135H, located
inward current during phase 4 of the cardiac action potential, to in the C-terminus of Kv11.1, caused gain-of-function by slowing
depolarize the membrane potential to the threshold of L-type channel deactivation.65 The R1135H mutation is believed to be a
Ca2+ channel activation, thereby producing spontaneous action modifier in SQTS, because not all carriers of the mutation dis-
potentials (DADs).57 This mechanism of arrhythmogenesis in played short QT intervals. Experiments in canine left ventricular
ATS (LQT7) contrasts with most other LQTS types (see earlier), wedge preparations, using an IKr agonist to mimic SQT1 muta-
in which EADs during the action potential plateau phase trigger tions, suggest that a proportionate decrease in the effective
arrhythmias. DADs may underlie frequent ventricular ectopies refractory period (secondary to action potential shortening) in
with occasionally bidirectional ventricular tachycardia and recur- combination with accentuated spatial dispersion of repolarization
rent polymorphic ventricular tachycardia in ATS. (due to spatial transmural heterogeneity of IKr expression) facili-
tates the initiation of polymorphic ventricular arrhythmias in
SQT1.66 It is important to note that QTc duration may prolong
Long QT Syndrome Type 13 in patients with SQT2 after treatment with hydroquinidine, a
nonselective cardiac K+ channel blocker.67
In 2010, a linkage study in a large Chinese family linked a het-
erozygous missense mutation in KCNJ5 to LQTS in affected
members. KCNJ5 encodes the α-subunit (Kir3.4) of a channel Short QT Syndrome Type 2
that carries the acetylcholine-sensitive inwardly rectifying K+
current (IK-Ach).58 Acetylcholine leads to opening of the channel Two mutations in KCNQ1 are associated with SQT2.62,68 The
via activation of membrane G-proteins. Heterologous expression first reported mutation, V307L, located in the pore loop of Kv7.1,
of the mutated channels revealed IK-Ach loss-of-function due to was identified in a 70-year-old patient with short QT intervals
reduced plasma membrane expression. IK-Ach is present in the and ventricular fibrillation. In heterologous expression systems,
sinoatrial node, atria, and atrioventricular node, where it plays a the mutation caused a shift of voltage-dependent activation
role in parasympathetic slowing of the heart rate and repolariza- toward more negative potentials (indicating earlier channel acti-
tion of atrial action potentials.58 However, its precise role in the vation) and accelerated the rate of activation.62 These changes are
ventricles is yet unclear. predicted to increase IKs. The second SQT2-linked mutation,
V141M in segment 1 of Kv7.1, was identified in a baby girl with
SQTS and atrial fibrillation in utero and after birth. The muta-
tion abrogated voltage-dependent gating, resulting in instanta-
Short QT Syndrome neous opening of the Kv7.1 channels. This indicates the presence
of a constitutive repolarizing current, which may explain the
The short QT syndrome (SQTS) was introduced in 2001 by association of the V141M mutation with atrial fibrillation.68 Pres-
Gussak and colleagues after the first description of constantly ervation of voltage-dependent gating in V307L mutated channels
shorter-than-normal QT intervals in one isolated case with might be a reason why this mutation does not induce atrial fibril-
syncope and sudden cardiac death at a young age, and three lation. It must be noted that KCNQ1 mutations are related to
familial cases, of which one suffered from paroxysmal atrial fibril- SQTS only in single isolated cases, and co-segregation of KCNQ1
lation.59 In 2003, Gaita and colleagues associated persistent short mutations with SQTS in large families has not yet been
QT intervals with syncope, ventricular fibrillation, and sudden demonstrated.
Short QT Syndrome Type 3 polymorphic ventricular tachycardia (CPVT), the Brugada
51
syndrome (BrS), and familial atrial fibrillation (FAF) (see
So far, two mutations in KCNJ2 are linked to SQT3. The D172N Table 51-1).3,11,75-78 The molecular genetics of these diseases
mutation, located in the transmembrane segment 1 of Kir2.1, was are described in detail elsewhere in this book. In short, CPVT
discovered in a 5-year-old girl and her father, both of whom is characterized by syncope and sudden death due to
displayed extremely short QT intervals.63 The second SQT3- exercise-induced (adrenergically mediated) bidirectional or
linked mutation (M301K, located in the C-terminus of Kir2.1) polymorphic ventricular tachycardia, and is associated with
was found in an 8-year-old girl with markedly short QT inter- mutations in RyR2-encoded cardiac ryanodine receptor/Ca2+-
vals.69 Both mutations caused larger outward IK1 at more positive release channel and CASQ2-encoded calsequestrin—proteins
potentials, indicating a larger contribution of IK1 to the terminal involved in Ca2+ homeostasis and excitation-contraction coupling
phase of action potential repolarization (because the membrane in cardiac myocytes. The ventricular arrhythmias in CPVT
potential returns from depolarized levels to hyperpolarized may resemble arrhythmias observed in patients with LQT7.
levels).63,69 Indeed, computer modeling studies showed that In 2006, a genetic study first identified mutations in KCNJ2
increased outward IK1 results in shortening of the action potential in 3 of 11 unrelated patients (with normal QT intervals)
duration as the result of abbreviation of the terminal phase of who were clinically diagnosed with CPVT, suggesting
repolarization.63 These findings are in line with the asymmetric phenotypic mimicry between CPVT and LQT7.75 However, the
shape of the T wave in SQT3 patients with a rapid descending mutations were not studied experimentally. In 2009, a KCNJ2
limb. In contrast, T waves in patients with SQT1 and SQT2 are mutation, which was found in a patient with CPVT, was shown
symmetric.61,62 to induce loss-of-function of Kir2.1 channels by abrogating
increase of IK1 upon PKA-dependent phosphorylation—not at
baseline.76 This indicates that, at least some KCNJ2 mutations
Sudden Infant Arrhythmia Death Syndrome may cause CPVT by inducing IK1 loss-of-function during
β-adrenergic stimulation by impairing Kir2.1 phosphorylation.
The term sudden infant death syndrome (SIDS), first introduced in However, the role of KCNJ2 mutations in CPVT should be
1969, refers to the sudden death of an infant, younger than 1 year further clarified.
of age, which is unexpected by clinical history, and in which a In BrS, ventricular tachycardia and ventricular fibrillation are
thorough postmortem analysis and investigation of the death accompanied by characteristic coved-type ST segment (or J
scene fail to identify an adequate cause of death.70 SIDS is point) elevation in the right precordial ECG leads V1 to V3. BrS
believed to be a multifactorial disease, with multiple risk factors is traditionally linked to loss-of-function mutations in the cardiac
contributing to its development, including gender, age, and Na+ channel, SCN5A. These mutations result in decreased INa
genetic variants (intrinsic), low birth weight and prematurity during phase 0 of the cardiac action potential, which may allow
(developmental), supine sleeping position, mild infection, the repolarizing transient outward K+ current (Ito) (active during
smoking or drug intake by parents, and socioeconomic status action potential phases 0 and 1) to cause loss of the action poten-
(extrinsic and/or environmental). Genetic variants associated tial plateau phase (by disabling the membrane potential to reach
with the occurrence of SIDS are mutations and polymorphisms voltages required for L-type Ca2+ channels to activate). Because
in genes encoding for proteins involved in the autonomic nervous Ito is more expressed in the subepicardial myocytes, this is believed
system, the immune system, fatty acid oxidation, and blood to aggravate transmural voltage gradients and provide a substrate
glucose regulation, and in genes encoding for the cardiac ion for reentrant excitation waves. Consistent with this theory, muta-
channels or their regulatory subunits.71 tions in KCND3, encoding the α-subunit of the channel respon-
In 2001, a mutation in KCNQ1, previously linked to LQTS, sible for Ito (Kv4.3), and KCNE3 and KCNE5, encoding β-subunits
was identified in an infant with SIDS.72 Ever since, numerous of the Ito channel, were recently discovered in unrelated patients
cohort studies and case reports, wherein mutation screening was with BrS.3 As expected, these mutations caused Ito gain-of-func-
performed in cardiac genes that were earlier linked to inheritable tion when studied in vitro. It is interesting to note that mutations
arrhythmia syndromes in adults, have related mutations in the in KCNH2 may also be implicated in BrS. Two BrS-linked mis-
following potassium channel encoding genes to SIDS: KCNQ1, sense mutations in KCNH2 were shown to cause gain-of-function
KCNH2, KCNJ8 (encoding adenosine triphosphate [ATP]-sensi- of the mutated Kv11.1 channels and, in computer simulation
tive K+ current channels), KCNE1, and KCNE2.71 These muta- models, loss of the action potential dome in right ventricular
tions often cause gating alterations that are expected to result in subepicardial myocytes.77 More recently, a missense mutation in
loss-of-function and/or gain-of-function.73 Indeed, some of these KCNJ8, encoding the α-subunit of the ATP-sensitive K+ channel
mutations are also found in adult patients with LQTS or SQTS.71 (Kir6.1), was identified in patients with Brugada syndrome
It is interesting to note that in a large prospective study in a (and in two other patients with ventricular fibrillation and early
population of 44 596 neonates, 59 infants displayed markedly repolarization pattern in the inferolateral ECG leads). The
prolonged QTc duration (QTc ≥460 ms), of whom 16 had muta- Kir6.1 channels are responsible for an inward-rectifying K+
tions in potassium channel encoding genes (9 KCNQ1 mutations, current (IK-ATP) during hypoxia or ischemia, when intracellular
6 KCNH2 mutations, 1 KCNE1 mutation, and 1 KCNE2 muta- ATP concentration or the ATP/adenosine diphosphate (ADP)
tion; with one infant carrying two mutations).74 Although the ratio declines. However, the BrS-linked mutation in KCNJ8
molecular pathogenesis of SIDS remains not well understood, on caused an increase in IK-ATP during normal conditions, probably
the basis of the previously mentioned findings, it is tempting to as the result of reduced sensitivity of the mutated channels to
speculate that SIDS cases suffer from inheritable arrhythmia syn- intracellular ATP.78
dromes (e.g., LQTS, SQTS), which, in conjunction with other Finally, several mutations in potassium channel genes are
intrinsic, developmental, and environmental risk factors, mani- found in families or single cases with lone atrial fibrillation (i.e.,
fest during early childhood. atrial fibrillation at a relatively young age in the absence of
cardiac structural abnormalities), including mutations in KCNQ1
and KCNE5 (involved in IKs), KCNH2 and KCNE2 (involved in
Other Potassium Channel Diseases IKr), KCNE3 (involved in Ito), KCNJ2 (responsible for IK1), and
KCNA5. 3,11 The KCNA5-encoded Kv1.5 channels conduct the
Other inheritable arrhythmia syndromes that are linked to muta- ultra-rapidly activating delayed outward-rectifying K+ current
tions in potassium channel genes comprise catecholaminergic (IKur), a repolarizing current present only in the atria. Generally,
all these mutations cause gain-of-function of the corresponding ventricular electrical activity. However, additional studies are
currents, leading to shortening of atrial action potential dura- needed to explore the pathophysiological role of potassium
tions, thereby facilitating the initiation of reentry. channel mutations in the development of arrhythmias in CPVT,
In conclusion, identification of mutations in cardiac potassium BrS, and FAF, and to develop clinically useful genotype-phenotype
channel genes in (often) isolated cases of CPVT, BrS, or FAF may correlation, risk stratification strategies, and gene/mutation-
emphasize the importance of these channels for normal atrial and specific therapies.
endosomal recycling of IKs Channels. Circ Res 35. Chang KC, Barth AS, Sasano T, et al: CAPON
References 103:1451–1457, 2008. modulates cardiac repolarization via neuronal
19. Kurokawa J, Motoike HK, Rao J, Kass RS: Regula- nitric oxide synthase signaling in the heart. Proc
1. Curran ME, Splawski I, Timothy KW, et al: A tory actions of the A-kinase anchoring protein Natl Acad Sci U S A 105:4777–4782, 2008.
molecular basis for cardiac arrhythmia: HERG Yatiao on a heart potassium channel downstream 36. Amin AS, Giudicessi JR, Tijsen AJ, et al: Variants
mutations cause long QT syndrome. Cell 80:795– of PKA phosphorylation. Proc Natl Acad Sci in the 3′untranslated region of the KCNQ1-
803, 1995. U S A 101:16374–16378, 2004. encoded Kv7.1 potassium channel modify disease
2. Wang Q, Curran ME, Splawski I, et al: Positional 20. Li Y, Zaydman MA, Wu D, et al: KCNE1 enhances severity in patients with type 1 long QT syndrome
cloning of a novel potassium channel gene: phosphatidylinositol 4,5-bisphosphate (PIP2) sen- in an allele-specific manner. Eur Heart J 33:714–
KVLQT1 mutations cause cardiac arrhythmias. Nat sitivity of IKs to modulate channel activity. Proc 723, 2012.
Genet 12:17–23, 1996. Natl Acad Sci U S A 108:9095–9100, 2011. 37. Harmer SC, Wilson AJ, Aldridge R, et al: Mecha-
3. Shimizu W, Horie M: Phenotypic manifestations 21. Schwartz PJ, Priori SG, Spazzolini C, et al: nisms of disease pathogenesis in long QT syn-
of mutations in genes encoding subunits of cardiac Genotype-phenotype correlation in the long-QT drome type 5. Am J Physiol Cell Physiol
potassium channels. Circ Res 109:97–109, 2011. syndrome: Gene-specific triggers for life- 298:C263–C273, 2010.
4. Tester DJ, Ackerman MJ: Genetic testing for threatening arrhythmias. Circulation 103:89–95, 38. Nishio Y, Makiyama T, Itoh H, et al: D85N, a
potentially lethal, highly treatable inherited 2001. KCNE1 polymorphism, is a disease-causing gene
cardiomyopathies/channelopathies in clinical prac- 22. Wong JA, Gula LJ, Klein GJ, et al: Utility of tread- variant in long QT syndrome. J Am Coll Cardiol
tice. Circulation 123:1021–1037, 2011. mill testing in identification and genotype predic- 54:812–819, 2009.
5. Moretti A, Bellin M, Welling A, et al: Patient- tion in long-QT syndrome. Circ Arrhythm 39. Schwartz PJ, Spazzolini C, Crotti L, et al: The
specific induced pluripotent stem-cell models for Electrophysiol 3:120–125, 2010. Jervell and Lange-Nielsen syndrome: natural
long-QT syndrome. N Engl J Med 363:1397– 23. Aziz PF, Wieand TS, Ganley J, et al: Genotype- history, molecular basis, and clinical outcome. Cir-
1409, 2010. and mutation site-specific QT adaptation during culation 113:783–790, 2006.
6. Itzhaki I, Maizels L, Huber I, et al: Modelling the exercise, recovery, and postural changes in children 40. Abbott GW, Sesti F, Splawski I, et al: MiRP1 forms
long QT syndrome with induced pluripotent stem with long-QT syndrome. Circ Arrhythm Electro- IKr potassium channels with HERG and is associ-
cells. Nature 471:225–229, 2011. physiol 4:867–873, 2011. ated with cardiac arrhythmia. Cell 97:175–187,
7. Moss AJ, Wataru Shimizu W, Wilde AA, et al: 24. Vyas H, Hejlik J, Ackerman MJ: Epinephrine QT 1999.
Clinical aspects of type-1 long-QT syndrome by stress testing in the evaluation of congenital 41. Sale H, Wang J, O’Hara TJ, et al: Physiological
location, coding type, and biophysical function of long-QT syndrome. Circulation 113:1385–1392, properties of hERG 1a/1b heteromeric currents
mutations involving the KCNQ1 gene. Circulation 2006. and a hERG 1b-specific mutation associated with
115:2481–2489, 2007. 25. Priori SG, Napolitano C, Schwartz PJ, et al: Asso- long-QT syndrome. Circ Res 103:e81–e95, 2008.
8. Shimizu W, Moss AJ, Wilde AA: Genotype- ciation of long QT syndrome loci and cardiac 42. Crotti L, Lundquist AL, Insolia R, et al: KCNH2-
phenotype aspects of type 2 long QT syndrome. J events among patients treated with β-blockers. K897T is a genetic modifier of latent congenital
Am Coll Cardiol 54:2052–2062, 2009. JAMA 292:1341–1344, 2004. long-QT syndrome. Circulation 112:1251–1258,
9. Itoh H, Shimizu W, Hayashi K, et al: Long QT 26. Schwartz PJ, Priori SG, Cerrone M, et al: 2005.
syndrome with compound mutations is associated Left cardiac sympathetic denervation in the man- 43. Gentile S, Martin N, Scappini E, et al: The human
with a more severe phenotype: A Japanese multi- agement of high-risk patients affected by the ERG1 channel polymorphism, K897T, creates a
center study. Heart Rhythm 7:1411–1418, 2010. long-QT syndrome. Circulation 109:1826–1833, phosphorylation site that inhibits channel activity.
10. Antzelevitch C: Role of spatial dispersion of repo- 2004. Proc Natl Acad Sci U S A 105:14704–14708, 2008.
larization in inherited and acquired sudden cardiac 27. Jons C, Moss AJ, Lopes CM, et al: Mutations in 44. Gong Q, Zhang L, Vincent GM, et al: Nonsense
death syndromes. Am J Physiol Heart Circ Physiol conserved amino acids in the KCNQ1 channel and mutations in hERG cause a decrease in mutant
293:H2024–H2038, 2007. risk of cardiac events in type-1 long-QT syndrome. mRNA transcripts by nonsense-mediated mRNA
11. Amin AS, Tan HL, Wilde AA: Cardiac ion chan- J Cardiovasc Electrophysiol 20:859–865, 2009. decay in human long-QT syndrome. Circulation
nels in health and disease. Heart Rhythm 7:117– 28. Jons C, O-Uchi J, Moss AJ, et al: Use of mutant- 116:17–24, 2007.
126, 2010. specific ion channel characteristics for risk stratifi- 45. Anderson CL, Delisle BP, Anson BD, et al: Most
12. Splawski I, Tristani-Firouzi M, Lehmann MH, et cation of long QT syndrome patients. Sci Transl LQT2 mutations reduce Kv11.1 (hERG) current
al: Mutations in the hminK gene cause long QT Med 3:76ra28, 2011. by a class 2 (trafficking-deficient) mechanism. Cir-
syndrome and suppress IKs function. Nat Genet 29. Kapa S, Tester DJ, Salisbury BA, et al: Genetic culation 113:365–373, 2006.
17:338–340, 1997. testing for long-QT syndrome: Distinguishing 46. Li P, Ninomiya H, Kurata Y, et al: Reciprocal
13. Splawski I, Shen J, Timothy KW, et al: Spectrum pathogenic mutations from benign variants. Circu- control of hERG stability by Hsp70 and Hsc70
of mutations in long-QT syndrome genes: lation 120:1752–1760, 2009. with implication for restoration of LQT2 mutant
KVLQT1, HERG, SCN5A, KCNE1, and KCNE2. 30. Barsheshet A, Goldenberg I, O-Uchi J, et al: Muta- stability. Circ Res 108:458–468, 2011.
Circulation 102:1178–1185, 2000. tions in cytoplasmic loops of the KCNQ1 channel 47. Albesa M, Grilo LS, Gavillet B, et al: Nedd4-2-
14. Tester DJ, Will ML, Haglund CM, et al: Compen- and the risk of life-threatening events: Implications dependent ubiquitylation and regulation of the
dium of cardiac channel mutations in 541 consecu- for mutation-specific response to β-blocker therapy cardiac potassium channel hERG1. J Mol Cell
tive unrelated patients referred for long QT in type 1 long-QT syndrome. Circulation 125: Cardiol 51:90–98, 2011.
syndrome genetic testing. Heart Rhythm 2:507– 1988–1996, 2012. 48. Smith JL, McBride CM, Nataraj PS, et al:
517, 2005. 31. Crotti L, Monti MC, Insolia R, et al: NOS1AP is Trafficking-deficient hERG K+ channels linked to
15. Kapplinger JD, Tester DJ, Salisbury BA, et al: a genetic modifier of the long-QT syndrome. Cir- long QT syndrome are regulated by a microtubule-
Spectrum and prevalence of mutations from the culation 120:165716–165763, 2009. dependent quality control compartment in the ER.
first 2,500 consecutive unrelated patients referred 32. Tomás M, Napolitano C, De Giuli L, et al: Poly- Am J Physiol Cell Physiol 301:C75–C85, 2011.
for the FAMILION long QT syndrome genetic morphisms in the NOS1AP gene modulate QT 49. Cui J, Melman Y, Palma E, et al: Cyclic AMP regu-
test. Heart Rhythm 6:1297–1303, 2009. interval duration and risk of arrhythmias in the lates the HERG K+ channel by dual pathways. Curr
16. Chen L, Marquardt ML, Tester DJ, et al: Mutation long QT syndrome. J Am Coll Cardiol 55:2745– Biol 10:671–674, 2000.
of an A-kinase-anchoring protein causes long-QT 2752, 2010. 50. Hekkala AM, Swan H, Viitasalo M, et al: Epineph-
syndrome. Proc Natl Acad Sci U S A 104:20990– 33. Newton-Cheh C, Eijgelsheim M, Rice KM, et al: rine bolus test in detecting long QT syndrome
20995, 2007. Common variants at ten loci influence QT interval mutation carriers with indeterminable electrocar-
17. Barc J, Briec F, Schmitt S, et al: Screening for copy duration in the QTGEN study. Nat Genet 41:399– diographic phenotype. Ann Noninvasive Electro-
number variation in genes associated with the long 406, 2009. cardiol 16:172–179, 2011.
QT syndrome. J Am Coll Cardiol 57:40–47, 2011. 34. Pfeufer A, Sanna S, Arking DE, et al: Common 51. Yao Y, Teng S, Li N, et al: Aminoglycoside antibi-
18. Seebohm G, Strutz-Seebohm N, Ureche ON, et variants at ten loci modulate the QT interval dura- otics restore functional expression of truncated
al: Long QT syndrome-associated mutations in tion in the QTSCD study. Nat Genet 41:407–414, HERG channels produced by nonsense mutations.
KCNQ1 and KCNE1 subunits disrupt normal 2009. Heart Rhythm 6:553–560, 2009.
52. Gianulis EC, Trudeau MC: Rescue of aberrant to mutations in HERG. Circulation 109:30–35, infant deaths: A definitional and diagnostic
gating by a genetically encoded PAS (Per-Arnt- 2004. approach. Pediatrics 114:234–238, 2004.
Sim) domain in several long QT syndrome mutant
human ether-á-go-go-related gene potassium chan-
nels. J Biol Chem 286:22160–22169, 2011.
62. Bellocq C, van Ginneken AC, Bezzina CR, et al:
Mutation in the KCNQ1 gene leading to the short
QT-interval syndrome. Circulation 109:2394–
71. Tfelt-Hansen J, Winkel BG, Grunnet M, et al:
Cardiac channelopathies and sudden infant death
syndrome. Cardiology 119:21–33, 2011.
51
53. Lu Y, Mahaut-Smith MP, Huang CL, et al: Mutant 2397, 2004. 72. Schwartz PJ, Priori SG, Bloise R, et al: Molecular
MiRP1 subunits modulate HERG K+ channel 63. Priori SG, Pandit SV, Rivolta I, et al: A novel form diagnosis in a child with sudden infant death syn-
gating: A mechanism for pro-arrhythmia in long of short QT syndrome (SQT3) is caused by a drome. Lancet 358:1342–1343, 2001.
QT syndrome type 6. J Physiol 551:253–262, 2003. mutation in the KCNJ2 gene. Circ Res 96:800– 73. Rhodes TE, Abraham RL, Welch RC, et al:
54. Pourrier M, Zicha S, Ehrlich J, et al: Canine ven- 807, 2005. Cardiac potassium channel dysfunction in sudden
tricular KCNE2 expression resides predominantly 64. Sun Y, Quan XQ, Fromme S, et al: A novel muta- infant death syndrome. J Mol Cell Cardiol 44:571–
in Purkinje fibers. Circ Res 93:189–191, 2003. tion in the KCNH2 gene associated with short QT 581, 2008.
55. Plaster NM, Tawil R, Tristani-Firouzi M, et al: syndrome. J Mol Cell Cardiol 50:433–441, 2011. 74. Schwartz PJ, Stramba-Badiale M, Crotti L, et al:
Mutations in Kir2.1 cause the developmental and 65. Itoh H, Sakaguchi T, Ashihara T, et al: A novel Prevalence of the congenital long-QT syndrome.
episodic electrical phenotypes of Andersen’s syn- KCNH2 mutation as a modifier for short QT inter- Circulation 120:1761–1767, 2009.
drome. Cell 105:511–519, 2001. val. Int J Cardiol 137:83–85, 2009. 75. Tester DJ, Arya P, Will M, et al: Genotypic hetero-
56. Tristani-Firouzi M, Etheridge SP: Kir 2.1 chan- 66. Patel C, Antzelevitch C: Cellular basis for arrhyth- geneity and phenotypic mimicry among unrelated
nelopathies: The Andersen-Tawil syndrome. mogenesis in an experimental model of the SQT1 patients referred for catecholaminergic polymor-
Pflugers Arch 460:289–294, 2010. form of the short QT syndrome. Heart Rhythm phic ventricular tachycardia genetic testing. Heart
57. Morita H, Zipes DP, Morita ST, et al: Mechanism 5:585–590, 2008. Rhythm 3:800–805, 2006.
of U wave and polymorphic ventricular tachycardia 67. Giustetto C, Schimpf R, Mazzanti A, et al: Long- 76. Vega AL, Tester DJ, Ackerman MJ, et al: Protein
in a canine tissue model of Andersen-Tawil syn- term follow-up of patients with short QT syn- kinase A-dependent biophysical phenotype for
drome. Cardiovasc Res 75:510–518, 2007. drome. J Am Coll Cardiol 58:587–595, 2011. V227F-KCNJ2 mutation in catecholaminergic
58. Yang Y, Yang Y, Liang B, et al: Identification of a 68. Hong K, Piper DR, Diaz-Valdecantos A, et al: De polymorphic ventricular tachycardia. Circ
Kir3.4 mutation in congenital long QT syndrome. novo KCNQ1 mutation responsible for atrial fibril- Arrhythm Electrophysiol 2:540–547, 2009.
Am J Hum Genet 86:872–880, 2010. lation and short QT syndrome in utero. Cardiovas 77. Verkerk AO, Wilders R, Schulze-Bahr E, et al: Role
59. Gussak I, Brugada P, Brugada J, et al: Idiopathic Res 68:433–440, 2005. of sequence variations in the human ether-a-go-go-
short QT interval: a new clinical syndrome? Car- 69. Hattori T, Makiyama T, Akao M, et al: A novel related gene (HERG, KCNH2) in the Brugada
diology 94:99–102, 2000. gain-of-function KCNJ2 mutation associated with syndrome. Cardiovasc Res 68:441–453, 2005.
60. Gaita F, Giustetto C, Bianchi F, et al: Short QT short-QT syndrome impairs inward rectification 78. Medeiros-Domingo A, Tan BH, Crotti L, et al:
syndrome: A familial cause of sudden death. Circu- of Kir2.1 currents. Cardiovasc Res 93:666–673, Gain-of-function mutation S422L in the KCNJ8-
lation 108:965–970, 2003. 2012. encoded cardiac K(ATP) channel Kir6.1 as a patho-
61. Brugada R, Hong K, Dumaine R, et al: Sudden 70. Krous HF, Beckwith JB, Byard RW, et al: Sudden genic substrate for J-wave syndromes. Heart
death associated with short QT syndrome linked infant death syndrome and unclassified sudden Rhythm 7:1466–1471, 2010.
Genetics and Cellular Mechanisms of
the J Wave Syndromes 52
Charles Antzelevitch
the cellular and ionic mechanisms underlying J wave syndromes.

CHAPTER OUTLINE
The clinical aspects are discussed in Chapter 96.
Definition and Clinical Characteristics of the An ER pattern in the ECG, consisting of a distinct J wave or
J Wave Syndromes 511 J point elevation, a notch or slur of the terminal part of the QRS,
and an ST-segment elevation, is commonly seen in healthy young
Cellular Basis for the J Wave and Associated males and until recent years had been viewed as benign.15,16 Our
Arrhythmogenesis 512 observation in 2000 that an ER pattern in the canine coronary-
Clinical Correlates 514 perfused wedge preparation can readily convert to one in which
phase 2 reentry gives rise to polymorphic ventricular tachycardia/
Genetics of J Wave Syndromes 515 ventricular fibrillation (VT/VF) prompted us to suggest that ER
Differentiating J Wave Deflections Caused by in some cases may predispose to malignant arrhythmias in the
Conduction vs. Repolarization Defects 517 clinic.8,17,18
A number of case reports and experimental studies have sug-
gested a critical role for the J wave in the pathogenesis of idio-
pathic ventricular fibrillation (IVF).19-28 A definitive association
between ER pattern and IVF was reported in two studies pub-
Definition and Clinical Characteristics lished in New England Journal of Medicine in 2008.29,30 These
of the J Wave Syndromes publications were followed by another study from Viskin and
coworkers31 that same year and by large population association
The J wave is a positive deflection in the electrocardiogram studies in 2009 and 2010.32-36 As is discussed in Chapter 96, a
(ECG) that occurs at the junction between the QRS complex and number of case-control and population-based studies appeared
the ST segment, also known as the J point. Because much of the between 2010 and 2012, confirming the association between ER
J wave is typically hidden inside the QRS, it may manifest as a J and IVF. It is interesting to note that recent studies have reported
point elevation, a slurring of the terminal part of the QRS, or a that the prevalence of inferior and anterior, but not lateral, ER
late delta wave following the QRS. When it becomes more accen- is significantly greater among patients who developed VT/VF
tuated, it may appear as a small secondary R wave (R′) or ST within 72 h after an acute myocardial infarction.37-40
segment elevation (Figure 52-1). A strong male predominance is observed with all of the J wave
The J wave or elevated J point has been recognized in the syndromes,8 including BrS.41 Experimental and clinical studies
ECG of humans and animals since the early 1930s. It was first have provided evidence in support of the hypothesis that testos-
described by Tomaszewski1 in 1938 in an accidentally frozen terone plays an important role in ventricular repolarization.
human. The J wave has also been termed an Osborn wave after Ezaki and coworkers42 demonstrated that ST-segment elevation
Osborn’s description of the wave in hypothermic dogs.2 is relatively small and is similar in males and females before
In both humans and animals, the appearance of a prominent puberty. After puberty, ST-segment elevation in males, but not
J wave in the ECG is pathognomonic of hypothermia3-5 and in females, increases sharply, more so in the right precordial
hypercalcemia6,7 and, more recently, as a marker for a substrate leads, subsequently decreasing gradually with advancing age. The
capable of generating life-threatening ventricular arrhythmias.8 effect of androgen-deprivation therapy on the ST segment was
A distinct J wave has been described in the ECG of subjects evaluated in 21 prostate cancer patients. Androgen-deprivation
completely recovered from hypothermia9,10 and those predis- therapy significantly decreased ST-segment elevation. These
posed to idiopathic ventricular fibrillation, but is otherwise rarely results suggest that testosterone modulates the early phase of
observed in humans under normal conditions. A distinct J wave ventricular repolarization and thus ST-segment elevation.
is commonly observed under baseline conditions in the ECG of We recently suggested a classification scheme for ER based
some animal species, such as dogs and baboons, and is greatly on the data available in the literature in 2010 (Table 52-1).8 An
accentuated under hypothermic conditions.11-13 An elevated J ER pattern manifest exclusively in the lateral precordial leads was
point, on the other hand, is commonly encountered in humans designated as Type 1; this form is prevalent among healthy male
and in some animal species under normal conditions. athletes and is thought to be associated with a relatively low level
The term early repolarization (ER) to the best of our knowl- of risk for arrhythmic events. The ER pattern in the inferior or
edge was coined by Grant et al14 to describe ST-segment devia- inferolateral leads was designated as Type 2; this form is thought
tions and associated T wave changes and was thought to result to be associated with a moderate level of risk. Finally, an ER
from premature repolarization. The ER pattern in the ECG in pattern appearing globally in the inferior, lateral, and right pre-
recent years has been shown to be associated with life-threatening cordial leads was labeled Type 3; this form is associated with the
arrhythmias, describing an entity that we termed early repolariza- highest level of risk and in some cases has been associated with
tion syndrome (ERS). Although Brugada syndrome (BrS) and ERS electrical storms.8 Type 3 ER patterns may be very similar to
differ with respect to the magnitude and lead location of abnor- those of Type 2, exhibiting inferolateral ER, except for brief
mal J wave manifestation, they are thought to represent a con- periods just before the development of VT/VF, when pronounced
tinuous spectrum of phenotypic expression termed J wave J waves are also observed in the right precordial leads.43 BrS
syndromes.8 In this chapter, we discuss the genetic basis for and represents a fourth variant in which ER is limited to the right
511
Control Control Flecainide Pinacidil NS5806 NS5806 (16 µM)

+ ACh (2 µM) (16 µM)
Endo
Epi
0.5 0.4 0.5

1.0 1.0 mV 0.5
mV mV mV mV mV
ECG
200 ms 200 ms 200 ms 200 ms 200 ms 200 ms

A B C D E F
Figure 52-1. Cellular Basis for Diversity in the Manifestation of the Early Repolarization Pattern in the ECG Each panel shows transmembrane
action potentials recorded from the epicardial and endocardial regions of an arterially perfused canine ventricular wedge preparation and a transmural electrocardiogram
(ECG) simultaneously recorded. Under the conditions indicated, the six panels illustrate the cellular basis for a J point elevation, a distinct J wave, slurring of the terminal
part of the QRS, combined J wave and ST-segment elevation, and a very prominent J wave appearing as an ST-segment elevation and giving rise to a brief episode of
polymorphic ventricular tachycardia (VT).
Table 52-1. J Wave Syndromes: Similarities and Differences
J wave syndromes
Inherited Acquired
ER In Lateral ER in Inferior or Ischemia-

Leads Inferolateral Leads Global ER Brugada Mediated Hypothermia-
ERS Type 1 ERS Type 2 ERS Type 3 Syndrome VT/VF Mediated VT/VF
Anatomic location Anterolateral left Inferior left ventricle Left and right Right Left and right Left and right
ventricle ventricles ventricle ventricles ventricles
Leads displaying I, V4-V6 II, III, aVF Global V1-V3 Any of 12 leads Any of 12 leads
J point/J wave
Response of J wave/
ST elevation to:
Bradycardia or pause N/A N/A
+
Na channel blockers N/A N/A
89,90
Sex Dominance Male Male Male Male Male Either gender
VT/VF Rare Yes23,29 Yes, electrical Yes Yes Yes
Common in healthy storms30,65
athletes15,16,50
91
Response of quinidine Limited data
to:
J wave /ST elevation
VT/VF
Response of Limited data N/A N/A
isoproterenol to:
J wave /ST elevation
VT/VF
ER, Early repolarization; ERS, early repolarization syndrome; N/A, not available; VF, ventricular fibrillation; VT, ventricular tachycardia.
Modified from Antzelevitch C, Yan GX: J wave syndromes. Heart Rhythm 7:549–558, 2010.
precordial leads. Within each category, the level of risk appears

to increase in accordance with the presence of additional electro-
Cellular Basis for the J Wave and Associated
cardiographic signatures, including a horizontal or declining ST Arrhythmogenesis
segment following the J wave or J point elevation,35,44,45 relatively
short QT intervals,46 and very prominent J waves or J point eleva- Transmural differences in the magnitude of the action potential
tion, which may appear as a coved-type ST-segment elevation notch have long been recognized as the basis for inscription of
(Box 52-1). Risk stratification strategies for BrS and ERS are the electrocardiographic J wave.47,48 The ventricular epicardial
discussed more fully in Chapter 96. (Epi) action potential (AP), particularly in the right ventricle
Genetics and Cellular Mechanisms of the J Wave Syndromes 513
Box 52-1 Risk Stratification of Patients With Early
52
50
Repolarization (ER) Pattern: Who Is at Risk?
mV
Epi
1. Association of ER pattern with sudden cardiac death (SCD),
unexplained syncope, or unexplained family history of
SCD.
2. J point or ST-segment elevation of 0.2 mV or greater in
inferior and inferolateral or global leads. 50
3. Transient J wave augmentation portends a high risk for Endo mV
ventricular fibrillation in patients with ER.
4. Appearance of distinct and prominent J waves.
5. Association of ER pattern with abbreviated QT intervals.
6. Association with horizontal or descending ST segment. 1
7. Appearance of closely coupled extrasystoles. ECG mV
8. Hyperpnea-induced ventricular tachycardia (VT)/
ventricular fibrillation (VF).
200 msec 200 msec
A B
outflow tract, displays a prominent transient outward current
(Ito)-mediated action potential notch or spike and dome morphol- 36° C 29° C
ogy. The presence of a prominent Ito-mediated action potential
notch in ventricular epicardium but not endocardium leads to the 50
development of a transmural voltage gradient that manifests as a mV
Epi
J wave or J point elevation on the ECG. Direct evidence in
support of this hypothesis was first acquired in the arterially
perfused canine ventricular wedge preparation,20 as illustrated in
Figures 52-1 and 52-2. Modulation of the notch amplitude, 50
whether due to hypothermia (Figure 52-2, C and D) or due to mV
Endo
changes in activation rate or to prematurity (Figure 52-3), gives
rise to parallel changes in the amplitude of the J wave. Factors
that influence Ito kinetics or ventricular activation sequence
(Figure 52-2, A, B) can modify the manifestation of the J wave 0.5
ECG mV
on the ECG. Whether reduced by Ito blockers such as
4-aminopyridine, quinidine, or premature activation; or aug-
mented by exposure to hypothermia, ICa and INa blockers, or Ito C D 200 msec
agonists such as NS5806 or IK-ATP agonists, changes in the mag-
nitude of the epicardial action potential notch parallel those of Figure 52-2. Effect of Ventricular Activation Sequence and Tempera-
the J wave or J point elevation (see Figure 52-1).49-52 ture on the Manifestation of the J Wave in the Pseudo-ECG
Augmentation of net repolarizing current, whether secondary Recorded for Coronary-Perfused Canine Right Ventricular Wedge
Preparations A, When the wedge preparation is stimulated from the endocardial
to a decrease in inward current or an increase in outward current, (Endo) surface, a J wave on the electrocardiogram (ECG) is temporally aligned with
accentuates the notch leading to augmentation of the J wave or the Ito-mediated epicardial action potential notch. B, When the preparation is
appearance of ST-segment elevation. A further increase in net paced from the epicardial surface, the epicardial action potential notch is coinci-
repolarizing current can result in partial or complete loss of the dent with the QRS, and a J wave is no longer observed. C and D, Hypothermia-
action potential dome, leading to a transmural voltage gradient induced J wave. Each panel shows transmembrane action potentials from the
that manifests as an accentuated J wave or an ST-segment eleva- epicardial and endocardial regions of an arterially perfused canine left ventricular
tion.18,49,50 In regions of the myocardium exhibiting a prominent wedge and a transmural ECG simultaneously recorded. C, Because of the smaller
Ito, such as the epicardium of the right ventricular outflow tract notch in the left ventricular (LV) epicardium, a distinct J wave is not seen under
(RVOT), marked accentuation of the action potential notch leads baseline conditions. The small action potential notch in epicardium but not in
endocardium is associated with an elevated J point at the R-ST junction (arrow) at
to a very pronounced J wave, generally referred to as a coved-type 36° C. D, A decrease in the temperature of the perfusate to 29° C results in an
or Type 1 ST-segment elevation, which is diagnostic of BrS when it increase in the amplitude and width of the action potential notch in epicardium
appears in the right precordial leads (Figure 52-4, B). When this but not endocardium, leading to the development of a transmural voltage gradient
pronounced manifestation appears in the inferior or lateral leads, that manifests as a prominent J wave on the ECG (arrow).
it is sometimes referred to as a variant of BrS, but in our view it
is best characterized as a more pronounced manifestation of ER. (Modified from Antzelevitch C, Yan GX: J wave syndromes. Heart Rhythm 7:549–558, 2010.)
Additional outward shift of the net current active during the
early phase of the AP can lead to loss of the AP dome (APD), thus heterogeneous, resulting in marked abbreviation of action poten-
creating dispersion of repolarization between epicardium and tial at some sites but not at others. The dome can then propagate
endocardium as well as within epicardium, between the region from regions at which it is maintained to regions where it is lost,
where the dome is lost and regions at which it is maintained giving rise to a very closely coupled extrasystole via phase 2
(Figure 52-4, C). Sodium channel blockers like procainamide, reentry (Figure 52-4, D).57 The phase 2 reentrant beat is capable
pilsicainide, propafenone, flecainide, and disopyramide cause a of initiating polymorphic VT or VF (Figure 52-4, E, F).
further outward shift of current flowing during the early phases of Although most studies point to the pathophysiology of BrS as
the action potential, and therefore are effective in inducing or caused by repolarization abnormalities, secondary to accentua-
unmasking ST-segment elevation in patients with concealed J tion of the notch in the early phases of the AP, recent data suggest
wave syndromes.53-55 Sodium channel blockers like quinidine, that the possibility of delayed depolarization in the right ven-
which also inhibits Ito, reduce the magnitude of the J wave and tricular outflow tract in some cases may provide the principal
normalize ST-segment elevation.18,56 Loss of the APD is usually substrate of the ST-segment elevation or J waves associated with
110 110
50 100 100
Action potential notch
J wave amplitude ( )
Epi mV 90 90
amplitude ( )
80 80
70 70
Endo 50 60 60
mV
50 50
40 40
ECG 0.2 30 30
mV
S1 S2 200 300 400 500 600 700 800 900 1000
A 200 ms B S1-S2 interval (ms)
Figure 52-3. Effect of Premature Stimulation on the Relationship Between Epicardial (Epi) Action Potential (AP) Notch (APN) Amplitude
and J Wave Amplitude A, Simultaneous recording of a transmural electrocardiogram (ECG) and transmembrane APs from the Epi and endocardial (Endo) regions of
an isolated arterially perfused right ventricular wedge. A significant APN in epicardium is associated with a prominent J wave (arrow) during basic stimulation (S1-S2: 4000 ms).
Premature stimulation (S1-S2: 300 ms) causes a parallel decrease in the amplitude of Epi APN and that of the J wave (arrow). B, Plot of the amplitudes of the Epi APN (open
circles) and the J wave (open squares) as a function of the S1-S2 interval. The amplitude of the Epi APN and that of the J wave are normalized to the value recorded at an
S1-S2 interval of 900 ms.
(Modified from Yan GX, Antzelevitch C: Cellular basis for the electrocardiographic J wave. Circulation 93:372–379, 1996.)
Table 52-2. Features Common to Brugada (BrS) and Early Repolarization Syndromes (ERS) and Possible Underlying Mechanisms
BrS ERS Possible Mechanism(s)
Region associated with highest arrhythmic risk RVOT Inferior myocardium Increased levels of Ito
Male predominance Yes (75%) Yes (80%) Testosterone modulation of ion currents
underlying the epicardial AP notch
Average age of first event 40 42
Dynamicity of ECG High High Autonomic modulation of ion channel currents
underlying early phases of the epicardial AP
VT/VF trigger Short-coupled PVC Short-coupled PVC Phase 2 reentry
Ameliorative response to quinidine Yes Yes Inhibition of Ito and possible vagolytic effect
Ameliorative response to isoproterenol and Yes Yes Increased ICa and faster heart rate
cilostazol
Ameliorative response to pacing Yes Yes Reduced availability of Ito due to slow recovery
from inactivation
Vagally mediated accentuation of ECG pattern Yes Yes Direct effect to inhibit ICa and indirect effect to
increase Ito (due to slowing of heart rate)
AP, Action potential; ECG, electrocardiogram; Ica, inward calcium channel current; Ito, inward calcium channel current; RVOT, right ventricular outflow tract; AP, action
potential; PVC, premature ventricular contraction; VF, ventricular fibrillation; VT, ventricular tachycardia.
BrS.58,59 The repolarization versus depolarization hypothesis con-

troversy has been documented as a point-counterpoint.60 Clinical Correlates
The net outward shift of current may extend beyond the
action potential notch causing depression of the epicardial action In patients with BrS or ERS, the manifestation of the J wave or
potential dome, thus leading to both ST-segment elevation and ER is dynamic,27,61,62 likely as the result of autonomic influences.
an accentuated J wave. Activation of the adenosine triphosphate The most prominent ECG changes appear just before the onset
(ATP)-sensitive potassium current (IK-ATP) or depression of the of VT/VF.20-27,43,61,62 Other ECG characteristics of ERS also
inward calcium channel current (ICa) can effect such a change closely match those of BrS, including the presence of accentuated
(Figure 52-5, A, B). This is more likely to manifest in the ECG J waves, ST-segment elevation, pause and bradycardia, depen-
as an ER pattern consisting of a J point elevation, slurring of the dence, and short coupled extrasystole-induced polymorphic VT/
terminal part of the QRS, and mild ST-segment elevation. The VF (Table 52-2). Vagally mediated accentuation of the ECG and
ER pattern facilitates loss of the dome secondary to agents or arrhythmic manifestations of BrS have long been appreci-
conditions that produce a further outward shift of net current, ated.61,63,64 In both BrS and ERS, bradycardia accentuates
leading to the development of ST-segment elevation, phase 2 ST-segment elevation, and tachycardia tends to normalize the ST
reentry, and VT/VF (Figure 52-5, C). Inhibition of Ito shifts net segment. VF often occurs near midnight or in the early morning
current in the inward direction, thus normalizing the ST segment hours, when heart rate is slower and parasympathetic tone is
and suppressing the J wave and arrhythmic manifestation. augmented.23,65
Control Terfenadine Terfenadine Terfenadine
Endo
52
TDR
Epi 2
50
50
mV
Epi 1 mV
EDR
1 1
ECG mV mV
A 100 ms B C D 100 ms
Terfenadine Terfenadine
Endo
Epi 2
50
Epi 1 mV
ECG 1
mV
· ·
E S1S2 1s F 1s
Figure 52-4. Cellular Basis for Electrocardiographic and Arrhythmic Manifestation of Brugada Syndrome (BrS) Each panel shows transmembrane
action potentials (APs) from one endocardial (Endo) (top) and two epicardial (Epi) sites, together with a transmural electrocardiogram (ECG) recorded from a canine coronary-
perfused right ventricular wedge preparation. A, Control (basic cycle length = 400 ms). B, Combined sodium and calcium channel block with terfenadine (5 µM) accentuates
the Epi AP notch, creating a transmural voltage gradient that manifests as an exaggerated J wave or ST-segment elevation in the ECG. C, Continued exposure to terfenadine
results in all-or-none repolarization at the end of phase 1 at some Epi sites but not others, creating a local Epi dispersion of repolarization (EDR), as well as a transmural
dispersion of repolarization (TDR). D, Phase 2 reentry occurs when the Epi AP dome propagates from a site where it is maintained to regions where it has been lost, giving
rise to a closely coupled extrasystole. E, Extrastimulus (S1-S2 = 250 ms) applied to the epicardium triggers a polymorphic ventricular tachycardia (VT). F, Phase 2 reentrant
extrasystole triggers a brief episode of polymorphic VT.
(Modified from Fish JM, Antzelevitch C: Role of sodium and calcium channel block in unmasking the Brugada syndrome. Heart Rhythm 1:210–217, 2004.)
Recent studies have found that the arrhythmic manifestations slowing of transmural conduction, so that the J point shifts to a
of ERS may be provoked by vagal maneuvers such as hyperpnea.66 lower position on the terminal part of the QRS.69
It is interesting to note that there is a higher prevalence of an ER
pattern in patients with spinal cord injury at levels of injury that
disrupt central sympathetic command of the heart (at the level of
C5 to C6), resulting in enhanced vagal tone and/or reduced Genetics of J Wave Syndromes
sympathetic tone.67
Suppression of ECG features by isoproterenol or pacing in BrS has been associated with mutations in 12 different genes (see
ER patients further supports the thesis that they share common Table 52-3). More than 300 mutations in SCN5A (Nav1.5, BrS1)
underlying electrophysiological abnormalities with BrS patients.43 have been reported in 11% to 28% of BrS probands.70-72 Muta-
However, salient diagnostic features of BrS such as provocation tions in CACNA1C (Cav1.2, BrS3), CACNB2b (Cavβ2b, BrS4),
by sodium channel blockers or positive signal averaged ECG are and CACNA2D1 (Cavα2δ, BrS9) are found in approximately 13%
rarely observed in ERS patients.30,43 An exception to this rule of probands.73,74 Mutations in the glycerol-3-phophate dehydro-
appears to apply to ERS associated with SCN5A mutations.68 genase 1–like enzyme gene (GPD1L, BrS2), SCN1B (β1-subunit
Kawata and coworkers showed that sodium channel blockers of Na channel, BrS5), KCNE3 (MiRP2, BrS6), SCN3B (β3-
attenuate ER in patients with ERS apparently as the result of subunit of Na channel, BrS7), KCNJ8 (BrS8), and KCND3
(BrS10) are more rare.75-80 Mutations in these genes lead to loss- generate an ER pattern in canine ventricular wedge preparations,
of-function in INa and ICa, as well as to gain-of-function in Ito or a rare variant of KCNJ8, responsible for the pore-forming subunit
IK-ATP. MOG1 was recently described as a new partner of NaV1.5, of the IK-ATP channel, has recently been reported in patients with
playing a role in its regulation, expression, and trafficking. A ERS as well as BrS.78,87,88 Recent studies from our group have also
missense mutation in MOG1 was also associated with BrS identified loss-of-function mutations in the α1 and β2 and α2δ
(BrS11).81 Mutations in KCNH2, KCNE5, and HCN4, although subunits of the cardiac L-type calcium channel (CACNA1C,
not causative, have been identified as capable of modulating the CACNB2, and CACNA2D1) in patients with ERS.74 The most
substrate for the development of BrS (Table 52-3).82-84 Loss-of- recent addition to the genes associated with ERS is SCN5A, the
function mutations in HCN4 causing a reduction in the pace- gene that encodes the α-subunit of the cardiac sodium channel.
maker current, If, have the potential to unmask BrS by reducing Watanabe and coworkers reported loss-of-function mutations in
heart rate.84 SCN5A in patients with VT associated with the ER pattern.68 It
The genetic basis for ERS is gradually coming into focus. The appears that SCN5A mutations are associated with a Type 3 ERS,
familial nature of the ER pattern has been demonstrated in a in which a J point or an ST-segment elevation is present in the
number of studies.34,85,86 In the Framingham Heart Study, siblings right precordial leads and in the inferior and lateral leads under
of ER subjects were twice as likely to have ER than non-ER baseline conditions or after a sodium block challenge.68
subjects (odds ratio [OR] 2.22; P < .05). In another study of more Our working hypothesis concerning the mechanism underly-
than 500 British families, ER was more than twice as likely to ing electrocardiographic and arrhythmic manifestations of the J
occur in children (OR 2.54; P = .005) if one of the parents had waves is that an outward shift in repolarizing current due to a
an ER pattern in the ECG.34,85 decrease in sodium or calcium channel currents or an increase in
ERS has been associated with mutations in six genes (Table outward currents, including Ito, IK-ATP, and IK-ACh, gives rise to both
52-4). Consistent with the findings that IK-ATP activation can the substrate and the trigger of BrS and ERS (Figure 52-6). The
Table 52-3. Genetic Basis of Brugada Syndrome (BrS)
Causative Genes
Locus Ion Channel Gene/Protein % of Probands
BrS1 3p21 INa SCN5A, Nav1.5 11-28

BrS2 3p24 INa GPD1L Rare
BrS3 12p13.3 ICa CACNA1C, Cav1.2 6.6
BrS4 10p12.33 ICa CACNB2b, Cavβ2b 4.8
BrS5 19q13.1 INa SCN1B, Navβ1 1.1
BrS6 11q13-14 Ito KCNE3, MiRP2 Rare
BrS7 11q23.3 INa SCN3B, Navβ3 Rare
BrS8 12p11.23 IK-ATP KCNJ8, Kir6.1 2
BrS9 7q21.11 ICa CACNA2D1, Cavα2d 1.8
BrS10 1p13.2 Ito KCND3, Kv4.3 Rare
BrS11 17p13.1 INa MOG1 Rare
BrS12 12p12.1 IK-ATP ABCC9, SUR2A Rare
Modulatory Genes
15q24-q25 If HCN4
7q35 IKr KCNH2, HERG
Xq22.3 Ito KCNE5 (KCNE1-like)
Table 52-4. Genetic Basis of Early Repolarization Syndrome (ERS)
Locus Ion Channel Gene/Protein % of Probands
ERS1 12p11.23 IK-ATP KCNJ8, Kir6.1

ERS2 12p13.3 ICa CACNA1C, CaV1.2 4.1
ERS3 10p12.33 ICa CACNB2b, Cavβ2b 8.3
ERS4 7q21.11 ICa CACNA2D1, Cava2d 4.1
ERS5 12p12.1 IK-ATP ABCC9, SUR2A
ERS6 3p21 INa SCN5A, NaV1.5
Early repolarization syndrome in a healthy young male Outward shift of repolarizing current
52
during early phase of the action potential
Surface J wave syndromes

ECG (V5)
IK-ATP INa, ICa

A
INa, ICa IK-ATP
Ito IK-ACh
Canine ventricular action potentials and ECG
Control Pinacidil (2 µM) Brugada Early

syndrome repolarization
?
Phase 2 reentry syndrome
Action Endo
Endo
potentials 50 Figure 52-6. J Wave Syndromes Schematic depicting our working hypothesis
Epi Epi mV that an outward shift in repolarizing current due to a decrease in sodium or calcium
channel currents or an increase in transient outward current (Ito), ATP-sensitive
potassium current (IK-ATP), or IK-ACh, or other outward currents, can give rise to accen-
tuated J waves associated with the Brugada syndrome and early repolarization
J wave J wave 0.4 syndrome. Both are thought to be triggered by closely coupled phase 2 reentrant
ECG mV extrasystoles, but in the case of early repolarization syndrome (ERS), a Purkinje
source of ectopic activity is also suspected.
B 200 msec (Modified from Antzelevitch C, Yan GX: J wave syndromes. Heart Rhythm 7:549–558,
2010.)
Pinacidil + 4-AP
50
Epi 1 mV
prominent J waves that predispose to the development of VT/
VF.8 The substrate involves a dispersion of repolarization both
within the epicardium and transmurally between the epicardium
50 and the endocardium, whereas the trigger is likely due to phase
Epi 2 mV 2 reentry, which develops as a result of the dome at some epicar-
dial sites, but not others.
In patients with BrS, the appearance of prominent J waves is
0.5 limited to the leads facing the right ventricular outflow tract,
ECG mV where Ito is thought to be most prominent. The more prominent
Ito in the right ventricular epicardium provides for a greater
C 500 msec D 300 msec outward shift in the balance of current, which promotes the
appearance of the J waves in this region of the ventricular myo-
Figure 52-5. Cellular Basis for the Early Repolarization Syndrome cardium. In the case of ERS, the appearance of prominent J waves
(ERS) A, Surface electrocardiogram (ECG) (lead V5) recorded from a 17-year-old may occur in other regions of the heart because of an outward
healthy African American male. Note the presence of a small J wave and marked shift in the balance of current in these regions secondary to an
ST-segment elevation. B, Simultaneous recording of transmembrane action poten-
increase in outward current (e.g., Ito, IK-ATP) or a decrease in
tials (APs) from epicardial (Epi) and endocardial (Endo) regions and a transmural
ECG in an isolated arterially perfused canine left ventricular wedge. A J wave in the
inward current (e.g., ICa, INa).
transmural ECG is manifest because of the presence of an AP notch in epicardium
but not endocardium. Pinacidil (2 µM), an adenosine triphosphate (ATP)-sensitive
potassium channel opener, causes depression of the AP dome in epicardium,
resulting in ST-segment elevation in the ECG, resembling the ERS. C, ATP-sensitive
Differentiating J Wave Deflections Caused by
potassium current (IK-ATP) activation in the canine right ventricular wedge prepara- Conduction vs. Repolarization Defects
tion using 2.5 µM pinacidil produces heterogeneous loss of the AP dome in epi-
cardium, resulting in ST-segment elevation, phase 2 reentry, and ventricular Perturbations of the terminal phase of the action potential gener-
tachycardia or ventricular fibrillation (VT/VF) (Brugada syndrome [BrS] phenotype).
ally referred to as J waves can arise from repolarization or depo-
D, The Ito blocker, 4-aminopyridine (4-AP), restored the Epi AP dome, reduced both
transmural and Epi dispersion of repolarization, normalized the ST segment, and
larization abnormalities. When due to the latter, the apparent J
prevented phase 2 reentry and VT/VF in the continued presence of pinacidil. wave is expected to appear as a notch interrupting the terminal
part of the QRS, with little or no ST-segment elevation. A simple
(Modified from Antzelevitch C, Yan GX: J wave syndromes. Heart Rhythm 7:549–558, way to distinguish between the two mechanisms is to examine the
2010.) effect of rate or atrial premature beats. When due to delayed
conduction, the notched appearance should become progres-
sively more accentuated with acceleration of rate or prematurity,
and when due to repolarization problems, the amplitude of the J
particular phenotype (BrS vs. ERS) depends on the part of the wave should gradually diminish. These different responses are
heart that is principally affected and the ion channels involved. due to the fact that delayed conduction almost invariably becomes
We view the J wave syndromes as a spectrum of disorders that more accentuated at faster rates or with prematurity, whereas the
involve accentuation of the epicardial action potential notch in Ito-mediated action potential notch diminishes as the result of
different regions of the heart, leading to the development of insufficient time for Ito to reactivate.
25. Shinohara T, Takahashi N, Saikawa T, et al: Char- 45. Rosso R, Adler A, Halkin A, et al: Risk of sudden
References acterization of J wave in a patient with idiopathic death among young individuals with J waves and
ventricular fibrillation. Heart Rhythm 3:1082– early repolarization: putting the evidence into per-
1. Tomaszewski W: Changement electrocar- 1084, 2006. spective. Heart Rhythm 8:923–929, 2011.
diographiques observes chez un homme mort de 26. Riera AR, Ferreira C, Schapachnik E, et al: 46. Watanabe H, Makiyama T, Koyama T, et al: High
froid. Arch Mal Coeur Vaiss 31:525–528, 1938. Brugada syndrome with atypical ECG: Downslop- prevalence of early repolarization in short QT syn-
2. Osborn JJ: Experimental hypothermia: Respiratory ing ST-segment elevation in inferior leads. J Elec- drome. Heart Rhythm 7:647–652, 2010.
and blood pH changes in relation to cardiac func- trocardiol 37:101–104, 2004. 47. Litovsky SH, Antzelevitch C: Transient outward
tion. Am J Physiol 175:389–398, 1953. 27. Shu J, Zhu T, Yang L, et al: ST-segment elevation current prominent in canine ventricular epicar-
3. Clements SD, Hurst JW: Diagnostic value of ECG in the early repolarization syndrome, idiopathic dium but not endocardium. Circ Res 62:116–126,
abnormalities observed in subjects accidentally ventricular fibrillation, and the Brugada syndrome: 1988.
exposed to cold. Am J Cardiol 29:729–734, 1972. Cellular and clinical linkage. J Electrocardiol 48. Antzelevitch C, Sicouri S, Litovsky SH, et al: Het-
4. Thompson R, Rich J, Chmelik F, et al: Evolution- 38:26–32, 2005. erogeneity within the ventricular wall: Electro-
ary changes in the electrocardiogram of severe 28. Boineau JP: The early repolarization variant— physiology and pharmacology of epicardial,
progressive hypothermia. J Electrocardiol 10:67– normal or a marker of heart disease in certain subendocardial, and M cells. Circ Res 69:1427–1449,
70, 1977. jects. J Electrocardiol 40:3.e1–e10, 2007. 1991.
5. Eagle K: Images in clinical medicine: Osborn 29. Haissaguerre M, Derval N, Sacher F, et al: Sudden 49. Antzelevitch C, Yan GX: Cellular and ionic mecha-
waves of hypothermia. N Engl J Med 10:680, 1994. cardiac arrest associated with early repolarization. nisms responsible for the Brugada syndrome.
6. Kraus F: Ueber die wirkung des kalziums auf den N Engl J Med 358:2016–2023, 2008. J Electrocardiol 33(Suppl):33–39, 2000.
kreislauf. Dtsch Med Wochenschr 46:201–203, 30. Nam GB, Kim YH, Antzelevitch C: Augmentation 50. Yan GX, Lankipalli RS, Burke JF, et al: Ventricular
1920. of J waves and electrical storms in patients with repolarization components on the electrocardio-
7. Sridharan MR, Horan LG: Electrocardiographic J early repolarization. N Engl J Med 358:2078–2079, gram: Cellular basis and clinical significance. J Am
wave of hypercalcemia. Am J Cardiol 54:672–673, 2008. Coll Cardiol 42:401–409, 2003.
1984. 31. Rosso R, Kogan E, Belhassen B, et al: J-point 51. Calloe K, Cordeiro JM, Di Diego JM, et al: A
8. Antzelevitch C, Yan GX: J wave syndromes. Heart elevation in survivors of primary ventricular fibril- transient outward potassium current activator reca-
Rhythm 7:549–558, 2010. lation and matched control subjects: Incidence and pitulates the electrocardiographic manifestations of
9. Phillipson EA, Herbert FA: Accidental exposure to clinical significance. J Am Coll Cardiol 52:1231– Brugada syndrome. Cardiovasc Res 81:686–694,
freezing: Clinical and laboratory observations 1238, 2008. 2009.
during convalescence from near-fatal hypothermia. 32. Tikkanen JT, Anttonen O, Junttila MJ, et al: Long- 52. Fish JM, Antzelevitch C: Role of sodium and
Can Med Assoc J 97:786–792, 1967. term outcome associated with early repolarization calcium channel block in unmasking the Brugada
10. Okada M, Nishimura F, Yoshino H, et al: The J on electrocardiography. N Engl J Med 361:2529– syndrome. Heart Rhythm 1:210–217, 2004.
wave in accidental hypothermia. J Electrocardiol 2537, 2009. 53. Shimizu W, Antzelevitch C, Suyama K, et al: Effect
16:23–28, 1983. 33. Sinner MF, Reinhard W, Muller M, et al: Associa- of sodium channel blockers on ST segment, QRS
11. Hugo N, Dormehl IC, Van Gelder AL: A positive tion of early repolarization pattern on ECG with duration, and corrected QT interval in patients
wave at the J-point of electrocardiograms of anaes- risk of cardiac and all-cause mortality: A population- with Brugada syndrome. J Cardiovasc Electro-
thetized baboons. J Med Primatol 17:347–352, based prospective cohort study (MONICA/ physiol 11:1320–1329, 2000.
1988. KORA). PLoS Med 7:e1000314, 2010. 54. Brugada R, Brugada J, Antzelevitch C, et al:
12. West TC, Frederickson EL, Amory DW: Single 34. Noseworthy PA, Tikkanen JT, Porthan K, et al: Sodium channel blockers identify risk for sudden
fiber recording of the ventricular response to The early repolarization pattern in the general death in patients with ST-segment elevation and
induced hypothermia in the anesthetized dog: Cor- population: Clinical correlates and heritability. right bundle branch block but structurally normal
relation with multicellular parameters. Circ Res J Am Coll Cardiol 57:2284–2289, 2011. hearts. Circulation 101:510–515, 2000.
7:880–888, 1959. 35. Tikkanen JT, Junttila MJ, Anttonen O, et al: Early 55. Morita H, Morita ST, Nagase S, et al: Ventricular
13. Santos EM, Frederick KC: Electrocardiographic repolarization: Electrocardiographic phenotypes arrhythmia induced by sodium channel blocker
changes in the dog during hypothermia. Am Heart associated with favorable long-term outcome. Cir- in patients with Brugada syndrome. J Am Coll
J 55:415–420, 1957. culation 123:2666–2673, 2011. Cardiol 42:1624–1631, 2003.
14. Grant RP, Estes EH Jr, Doyle JT: Spatial vector 36. BurashnikovAAntzelevitchCEvaluation of: [Tik- 56. Gussak I, Antzelevitch C, Bjerregaard P, et al: The
electrocardiography; the clinical characteristics of kanen JT, et al. Early repolarization: Electrocardio- Brugada syndrome: Clinical, electrophysiologic
S-T and T vectors. Circulation 3:182–197, 1951. graphic phenotypes associated with favorable and genetic aspects. J Am Coll Cardiol 33:5–15,
15. Wasserburger RH, Alt WJ: The normal RS-T long-term outcome. Circulation 123:2666-2673, 1999.
segment elevation variant. Am J Cardiol 8:184– 2011]Faculty of 1000: July 6 2011F1000.com/ 57. Krishnan SC, Antzelevitch C: Flecainide-induced
192, 1961. 11746956 arrhythmia in canine ventricular epicardium: Phase
16. Mehta MC, Jain AC: Early repolarization on scalar 37. Patel RB, Ng J, Reddy V, et al: Early repolarization 2 reentry? Circulation 87:562–572, 1993.
electrocardiogram. Am J Med Sci 309:305–311, associated with ventricular arrhythmias in patients 58. Postema PG, van Dessel PF, Kors JA, et al: Local
1995. with chronic coronary artery disease. Circ depolarization abnormalities are the dominant
17. Gussak I, Antzelevitch C: Early repolarization syn- Arrhythm Electrophysiol 3:489–495, 2010. pathophysiologic mechanism for type 1 electrocar-
drome: Clinical characteristics and possible cellular 38. Patel RB, Ilkhanoff L, Ng J, et al: Clinical charac- diogram in Brugada syndrome: A study of electro-
and ionic mechanisms. J Electrocardiol 33:299– teristics and prevalence of early repolarization cardiograms, vectorcardiograms, and body surface
309, 2000. associated with ventricular arrhythmias following potential maps during ajmaline provocation. J Am
18. Yan GX, Antzelevitch C: Cellular basis for the acute ST-elevation myocardial infarction. Am J Coll Cardiol 55:789–797, 2010.
Brugada syndrome and other mechanisms of Cardiol 110:615–620, 2012. 59. Nademanee K, Veerakul G, Chandanamattha P, et
arrhythmogenesis associated with ST segment 39. Naruse Y, Tada H, Harimura Y, et al: Early repo- al: Prevention of ventricular fibrillation episodes in
elevation. Circulation 100:1660–1666, 1999. larization is an independent predictor of occur- Brugada syndrome by catheter ablation over the
19. Bjerregaard P, Gussak I, Kotar Sl, et al: Recurrent rences of ventricular fibrillation in the very early anterior right ventricular outflow tract epicardium.
syncope in a patient with prominent J-wave. Am phase of acute myocardial infarctions. Circ Circulation 123:1270–1279, 2011.
Heart J 127:1426–1430, 1994. Arrhythm Electrophysiol 5:506–513, 2012. 60. Wilde AA, Postema PG, DiDiego JM, et al: The
20. Yan GX, Antzelevitch C: Cellular basis for the elec- 40. Tikkanen JT, Wichmann V, Junttila MJ, et al: Asso- pathophysiological mechanism underlying Brugada
trocardiographic J wave. Circulation 93:372–379, ciation of early repolarization and sudden cardiac syndrome: Depolarization versus repolarization.
1996. death during an acute coronary event. Circ J Mol Cell Cardiol 49:543–553, 2010.
21. Geller JC, Reek S, Goette A, et al: Spontaneous Arrhythm Electrophysiol 5:714–718, 2012. 61. Kasanuki H, Ohnishi S, Ohtuka M, et al: Idio-
episode of polymorphic ventricular tachycardia in 41. Shimizu W, Matsuo K, Kokubo Y, et al: Sex pathic ventricular fibrillation induced with vagal
a patient with intermittent Brugada syndrome. J hormone and gender difference—Role of testoster- activity in patients without obvious heart disease.
Cardiovasc Electrophysiol 12:1094, 2001. one on male predominance in Brugada syndrome. Circulation 95:2277–2285, 1997.
22. Daimon M, Inagaki M, Morooka S, et al: Brugada J Cardiovasc Electrophysiol 18:415–421, 2007. 62. Matsuo K, Shimizu W, Kurita T, et al: Dynamic
syndrome characterized by the appearance of J 42. Ezaki K, Nakagawa M, Taniguchi Y, et al: Gender changes of 12-lead electrocardiograms in a patient
waves. PACE 23:405–406, 2000. differences in the ST segment: Effect of androgen- with Brugada syndrome. J Cardiovasc Electro-
23. Kalla H, Yan GX, Marinchak R: Ventricular fibril- deprivation therapy and possible role of testoster- physiol 9:508–512, 1998.
lation in a patient with prominent J (Osborn) waves one. Circ J 74:2448–2454, 2010. 63. Mizumaki K, Fujiki A, Tsuneda T, et al: Vagal
and ST segment elevation in the inferior electro- 43. Nam GB, Ko KH, Kim J, et al: Mode of onset of activity modulates spontaneous augmentation of
cardiographic leads: A Brugada syndrome variant? ventricular fibrillation in patients with early repo- ST elevation in daily life of patients with Brugada
J Cardiovasc Electrophysiol 11:95–98, 2000. larization pattern vs. Brugada syndrome. Eur Heart syndrome. J Cardiovasc Electrophysiol 15:667–
24. Komiya N, Imanishi R, Kawano H, et al: Ventricu- J 31:330–339, 2010. 673, 2004.
lar fibrillation in a patient with prominent J wave 44. Junttila MJ, Sager SJ, Tikkanen JT, et al: Clinical 64. Patruno N, Pontillo D: Brugada syndrome and
in the inferior and lateral electrocardiographic significance of variants of J-points and J-waves: vasovagal syncope. PACE 29:215, 2006.
leads after gastrostomy. PACE 29:1022–1024, Early repolarization patterns and risk. Eur Heart J 65. Qi X, Sun F, An X, Yang J: A case of Brugada
2006. 33:2639–2643, 2012. syndrome with ST segment elevation through
entire precordial leads. Chin J Cardiol 32:272–273, 74. Burashnikov E, Pfeiffer R, Barajas-Martinez H, 83. Ohno S, Zankov DP, Ding WG, et al: KCNE5
2004. et al: Mutations in the cardiac L-type calcium (KCNE1L) variants are novel modulators of
66. Koutbi L, Roussel M, Haissaguerre M, et al:
Hyperpnea test triggering malignant ventricular
arrhythmia in a child with early repolarization.
channel associated J wave syndrome and sudden
cardiac death. Heart Rhythm 7:1872–1882, 2010.
75. London B, Michalec M, Mehdi H, et al: Mutation
Brugada syndrome and idiopathic ventricular
fibrillation. Circ Arrhythm Electrophysiol 4:352–
361, 2011.
52
Heart Rhythm 9:1153–1156, 2012. in glycerol-3-phosphate dehydrogenase 1 like gene 84. Ueda K, Nakamura K, Hayashi T, et al: Functional
67. Marcus RR, Kalisetti D, Raxwal V, et al: Early (GPD1-L) decreases cardiac Na+ current and characterization of a trafficking-defective HCN4
repolarization in patients with spinal cord injury: causes inherited arrhythmias. Circulation mutation, D553N, associated with cardiac arrhyth-
Prevalence and clinical significance. J Spinal Cord 116:2260–2268, 2007. mia. J Biol Chem 279:27194–27198, 2004.
Med 25:33–38, 2002. 76. Watanabe H, Koopmann TT, Le Scouarnec S, 85. Reinhard W, Kaess BM, Debiec R, et al: Heritabil-
68. Watanabe H, Nogami A, Ohkubo K, et al: Elec- et al: Sodium channel b1 subunit mutations associ- ity of early repolarization: A population-
trocardiographic characteristics and SCN5A muta- ated with Brugada syndrome and cardiac conduc- based study. Circ Cardiovasc Genet 4:134–138,
tions in idiopathic ventricular fibrillation associated tion disease in humans. J Clin Invest 118:2260–2268, 2011.
with early repolarization. Circ Arrhythm Electro- 2008. 86. Nunn LM, Bhar-Amato J, Lowe MD, et al: Preva-
physiol 4:874–881, 2011. 77. Delpón E, Cordeiro JM, Núñez L, et al: Func- lence of J-point elevation in sudden arrhythmic
69. Kawata H, Noda T, Yamada Y, et al: Effect of tional effects of KCNE3 mutation and its role in death syndrome families. J Am Coll Cardiol
sodium-channel blockade on early repolarization the development of Brugada syndrome. Circ 58:286–290, 2011.
in inferior/lateral leads in patients with idiopathic Arrhythm Electrophysiol 1:209–218, 2008. 87. Haissaguerre M, Chatel S, Sacher F, et al: Ven-
ventricular fibrillation and Brugada syndrome. 78. Medeiros-Domingo A, Tan BH, Crotti L, et al: tricular fibrillation with prominent early repolar-
Heart Rhythm 9:77–83, 2012. Gain-of-function mutation S422L in the KCNJ8- ization associated with a rare variant of KCNJ8/
70. Chen Q, Kirsch GE, Zhang D, et al: Genetic basis encoded cardiac K(ATP) channel Kir6.1 as a patho- KATP channel. J Cardiovasc Electrophysiol 20:93–
and molecular mechanisms for idiopathic ventricu- genic substrate for J-wave syndromes. Heart 98, 2009.
lar fibrillation. Nature 392:293–296, 1998. Rhythm 7:1466–1471, 2010. 88. Barajas-Martinez H, Hu D, Ferrer T, et al: Molec-
71. Schulze-Bahr E, Eckardt L, Breithardt G, et al: 79. Giudicessi JR, Ye D, Tester DJ, et al: Transient ular genetic and functional association of Brugada
Sodium channel gene (SCN5A) mutations in 44 outward current (Ito) gain-of-function mutations and early repolarization syndromes with S422L
index patients with Brugada syndrome: Different in the KCND3-encoded Kv4.3 potassium channel missense mutation in KCNJ8. Heart Rhythm
incidences in familial and sporadic disease. Hum and Brugada syndrome. Heart Rhythm 8:1024– 9:548–555, 2012.
Mutat 21:651–652, 2003. 1032, 2011. 89. Lerner DJ, Kannel WB: Patterns of coronary heart
72. Kapplinger JD, Wilde AAM, Antzelevitch C, et al: 80. Cranefield PF, Hoffman BF: Conduction of the disease morbidity and mortality in the sexes: A
A worldwide compendium of putative Brugada cardiac impulse. II. Summation and inhibition. 26-year follow-up of the Framingham population.
syndrome associated mutations in the SCN5A Circ Res 28:220–233, 1971. Am Heart J 111:383–390, 1986.
encoded cardiac sodium channel. Heart Rhythm 81. Kattygnarath D, Maugenre S, Neyroud N, et al: 90. Every N, Hallstrom A, McDonald KM, et al: Risk
6:S392, 2009. Abstract. MOG1: A new susceptibility gene for Brugada syn- of sudden versus nonsudden cardiac death in
73. Antzelevitch C, Pollevick GD, Cordeiro JM, et al: drome. Circ Cardiovasc Genet 4:261–268, 2011. patients with coronary artery disease. Am Heart J
Loss-of-function mutations in the cardiac calcium 82. Verkerk AO, Wilders R, Schulze-Bahr E, et al: 144:390–396, 2002.
channel underlie a new clinical entity characterized Role of sequence variations in the human ether-a- 91. Johnson P, Lesage A, Floyd WL, et al: Prevention
by ST-segment elevation, short QT intervals, and go-go-related gene (HERG, KCNH2) in the of ventricular fibrillation during profound hypo-
sudden cardiac death. Circulation 115:442–449, Brugada syndrome. Cardiovasc Res 68:441–453, thermia by quinidine. Ann Surg 151:490–495,
2007. 2005. 1960.
Inheritable Phenotypes Associated
With Altered Intracellular
Calcium Regulation 53
Silvia G. Priori and Carlo Napolitano
extrudes one Ca2+ ion (two positive charges) for every three Na+
CHAPTER OUTLINE
ions (three positive charges) that are transferred into the cell.
Overview of the Function of the Calcium-Handling Thus, NCX generates a net inward depolarizing current, the
System 521 transient inward current—Iti. In physiological conditions,
SERCA is responsible for approximately 63% of calcium removal,
Pathophysiology 523 and NCX mediates the remaining 37%.11 However, NCX
Experimental Therapies for Calcium-Handling Disorders 525 becomes important to remove Ca2+ in any condition of calcium
overload (e.g., in patients with genetic mutations of the RyR2
Conclusions 526 gene [CPVT] and during heart failure).11 Excessive activation of
NCX can be arrhythmogenic.
Overview of the Function of the Sarcoplasmic Reticulum Ca2+ Release Threshold

Calcium-Handling System and the Adrenergic Signaling
Growing clinical and experimental evidence highlights the rele- Activation of the adrenergic nervous system, mainly through
vance of cardiac calcium handling in the pathogenesis of inher- β-adrenergic receptors, has profound effects on calcium han-
ited arrhythmias.1 The control of Ca2+ fluxes in myocardial cells dling, and it is often the initiator of calcium-mediated arrhyth-
requires the timely coordination of several events that ultimately mogenesis.12 Adrenergic activation has two major effects on
lead to contraction. Any perturbation of this process has the calcium handling: the enhancement of the amplitude of the
potential to determine an arrhythmogenic substrate. L-type calcium current (ICa) and the increase of SR Ca2+ levels
The core events of Ca2+ ion movements in the myocardial cells induced through the activation of SERCA13. The latter is respon-
are the opening of the L-type Ca2+ channels followed by the release sible for an increase of Ca2+-transient amplitude. The effects of
of Ca2+ from the sarcoplasmic reticulum (SR) through the opening adrenergic activation can take place because of the phosphoryla-
of the ryanodine receptors (RyR2).2 This process is called calcium- tion target proteins14 induced by the activation of two enzymes
induced calcium release (CICR).3 The L-type calcium channel with kinase activity: protein kinase A14 and Ca2+ calmodulin
(CaV1.2) activation is therefore the first step of CICR. CaV1.2 kinase II (CAMKII).15 Among the several target proteins, the
belongs to the family of the voltage-gated calcium channels that adrenergic-dependent changes of Ca2+ handling are mainly due
are macromolecular complexes consisting of an ion conducting to the phosphorylation of L-type Ca2+ channel (increased current
protein (the α1-subunit) and additional accessory peptides with amplitude) phospholamban (removal of SERCA inhibition and
regulatory function called α2δ, the β1-4, and γ subunits.4 Among the increase of SR Ca reuptake) and RyR2 (increased open probabil-
biophysical properties of voltage-dependent calcium current ity and increased transients).16 Thus, protein phosphorylation is
(ICa), the inactivation process is relevant to inherited arrhythmias an important mechanism that enables adrenergic activation to
(Timothy syndrome, see Chapter 95). Two components have been enhance the SR calcium release. In physiological conditions, this
identified: voltage- (VDI) and Ca2+ (CDI)-dependent inactivation. response is useful to react to environmental stressors by improv-
The Ca2+-mediated component of inactivation is modulated by ing myocardial contractility.
intracellular (cytosolic) concentration of calcium (Ca2+). The importance of the adrenergic nervous system in the mod-
CaV1.2 channels tend to cluster in the T-tubules in close ulation of the function of the CRUs is highlighted by the evi-
proximity with the ryanodine receptors sitting across the mem- dence that mutations in the genes encoding for three of its major
brane of the SR.5 CaV1.2 activation constitutes the signal for the components, the ryanodine receptor (RyR2), cardiac calsequestrin
activation of the RyR2. This latter event can be detected as a (CASQ2), triadin (TRDN) lead to a severe inherited arrhythmo-
calcium transient, which is the sum of coordinated local releases genic syndrome characterized by life-threatening arrhythmias
that occur at specialized structures: the calcium release units induced by adrenergic stimulation, namely catecholaminergic
(CRUs; Figure 53-1). One CRU is formed by clusters of RyR2 polymorphic ventricular tachycardia (CPVT). Recently, some
receptors that are in close proximity to L-type Ca2+ channels in authors have provided data supporting a causal link between RyR2
the T-tubules.6 CRUs also include cardiac calsequestrin (CASQ2), mutations and the onset of structural cardiomyopathy. This
triadin (TRDN), and junctin (JTC) that contribute to the control chapter reviews the current evidences and pathophysiological
of the calcium release process. These peptides form a macromo- knowledge on phenotypes associated with intracellular calcium
lecular complex that acts in coordination to control Ca2+ release. handling proteins.
The number of CRUs recruited for release at each cardiac cycle
is an important modulator of the systolic Ca2+ transient ampli-
tude,7,8 and the loss of integrity of CRU is part of the pathophysi- Phenotypes Associated With Mutation in Calcium
ology of CPVT.9 Handling Proteins
During the relaxation phase, SR Ca2+ release terminates, and
Ca2+ is taken up in the SR by the SR Ca2+-ATPase (SERCA) or CPVT is the most frequent phenotype associated with mutations
extruded from the cell by the Na+/Ca2+ exchanger (NCX).10 NCX involved in intracellular calcium handling. Three genes have
521
CaV 1.2
Sarcoplasmic
reticulum
TRDN
CASQ2
JTC Plasma
membrane
RyR2
FKBP12.6
T-tubule
CPVT1 – RyR2
CPVT2 – CASQ2
CPVT3 – TRND
Figure 53-1. Schematic representation of CRU and CPVT gene localization. The figure depicts a schematic representation of a T-tubule with the LTCC sitting across the
plasmalemma juxtaposed to the RyR2/CASQ2/TRND/JTC macromolecular complex from the SR side. The SR component of CRU is involved in CPVT pathogenesis. CASQ2,
cardiac calsequestrin; CPVT, catecholaminergic polymorphic ventricular tachycardia; CRU, calcium release unit; JTC, junctin; LTCC, L-type calcium channel; RyR2, cardiac
ryanodine receptor; SR, sarcoplasmic reticulum; TRDN, triadin.
Table 53-1. CPVT Variants and Phenotypes
Arrhythmogenic Proposed Functional Consequence

Variant Gene Chromosome Clinical Presentation Mechanism of Mutations
CPVT1 RyR2 1q43 Adrenergic-induced bidirectional DADs, triggered Increased calcium sensitivity
ventricular tachycardia activity domain unzipping channel instability
Supraventricular tachycardia owing to FKBB12.6 dissociation
Idiopathic ventricular fibrillation Altered fractional release (associated with
Cardiomyopathy structural abnormalities)
CPVT2 CASQ2 1p13.3 Adrenergic-induced bidirectional DADs, triggered Impaired luminal calcium sensitivity
ventricular tachycardia activity Increased RYR2 channel open probability
Adrenergic-induced polymorphic Ultrastructural junctional SR abnormalities
ventricular tachycardia Reduced expression levels of TRDN and JCT
Supraventricular tachycardia
CPVT3 TRDN 6q22.31 Polymorphic VT, unexplained sudden n/a* Reduced triadin expression and possibly
death CASQ2 expression (based on indirect
evidence)
*Reduced TRDN expression is likely to result in a CPVT2-like phenotype.
DAD, Delayed afterdepolarization; n/a, not available.
been implicated in its pathogenesis (Table 53-1): ryanodine however, 14% of CPVT patients harbor RyR2 mutations located
receptor (RyR2 [CPVT1]),17 cardiac calsequestrin (CASQ2 outside these areas.21 Recent data suggest that there is no signifi-
[CPVT2]),18 and cardiac triadin (TRDN [CPVT3]).19 CPVT1 is cant difference of outcome of CPVT according to mutation
an autosomal dominant trait, whereas CPVT2 and CPVT3 site, including the outcome of the subgroup with mutations
are rare autosomal recessive disorders. In 2001, Priori et al17 outside the canonical clusters.22 N, RyR2 mutations have been
identified RyR2 as the gene responsible for the most frequent reported also in patients referred for idiopathic ventricular fibril-
variant of CPVT. The prevalence of RyR2 mutations in patients lation and in relatives of subjects who died suddenly.23,24 Autoso-
with a clearly diagnostic phenotype is high (≈60% to 70%).20,21 mal recessive CPVT is rarely identified in the clinical setting.
Mutations concentrate in three specific areas of the RyR2 Combined estimated prevalence of the two recessive variants is
protein (amino acids 77-466, 2246-2534, and 3778-4967); approximately 5%.
Inheritable Phenotypes Associated With Altered Intracellular Calcium Regulation 523
Clinical Manifestations and Management of CPVT important for dosage titration because the threshold for arrhyth-
53
mias in CPVT is reproducible. β-Blockers can prevent the onset
CPVT is a severe inherited arrhythmogenic disease manifesting of arrhythmia in 70% to 80% of cases.23,26 In addition, nonselec-
with adrenergically mediated arrhythmias, often leading to tive β-blockers (e.g., nadolol, propranolol) confer the highest
syncope or cardiac arrest.17,23,25-27 Although the resting electrocar- degree of protection against the onset of VT and SCD. However,
diogram is unremarkable, the reproducible inducibility of ven- the nontrivial incidence of recurrent cardiac on optimal β-blocker
tricular tachycardia (VT) during exercise stress test is the hallmark dosage (approximately 30%)22 calls for the identification of addi-
of the disease. Most patients with CPVT show a bidirectional tional therapeutic strategies. Implantable cardioverter defibrilla-
VT pattern characterized by beat-to-beat 180-degree rotation of tor (ICD) implant, left cardiac sympathetic denervation, and
the QRS axis (Figure 53-2)17,23,26; however, polymorphic tachy- other pharmacologic approaches have been proposed, and fle-
cardia or ventricular fibrillation may also be part of the picture.23,27 cainide appears to be the most promising strategy. Although the
This unstable, catecholamine-sensitive, substrate can lead to mechanisms for its effectiveness are debatable (discussed later),
sudden death as the first manifestation of the disease in up to clinical data and personal experience suggest that flecainide
30% of cases.23 Symptoms suggesting the presence of arrhyth- affords additional protection when added to β-blockers in
mias tend to manifest early in life (median age, 12 years), although CPVT.31,32
later onset is possible. The development of palpitations and syn-
copal events during adrenergic stress is a critical element to
suspect the diagnosis of CPVT, which is confirmed by the induc-
tion of bidirectional VT on an exercise stress test. In untreated Pathophysiology
patients, the occurrence of severe arrhythmias is approximately
60% to 70%, and approximately 30% experience a cardiac arrest Abnormalities of Calcium Handling in CPVT
or sudden death upon first manifestation (see Figure 53-2).23,28
CPVT patients typically present with structurally normal heart. From a mechanistic standpoint, similarities and significant differ-
However, some RyR2 mutations have been associated anecdotally ences exist among CPVT1-3 genetic variants. In all three forms
with structural cardiomyopathies. One of the first reports on of CPVT, arrhythmogenesis is a consequence of the occurrence
RyR2 gene mutations suggested an association with arrhythmo- of spontaneous calcium release (SCR; Figure 53-3). This term is
genic right ventricular cardiomyopathy (ARVC).29 More recently,
other authors have reported preliminary data linking RyR2 with
hypertrophic cardiomyopathy (HCM).30 Although careful scru-
50 µM
tiny of these clinical reports does not establish a causal link,
recent experimental findings suggest a mechanistic explanation
as to why some RyR2 variants could in principle make the heart
more prone to develop structural abnormalities (discussed later).
β-Adrenergic receptor block represents the first-choice thera-
peutic approach. The use of β-blockers is based on the evidence WT 1s
of the direct link between adrenergic activation and cardiac 8
events in CPVT. Clinical data show that β-blockers have an effect
F/F0
on the natural history of CPVT by achieving a significant reduc-

tion of cardiac events. In this context, exercise stress testing is
1
100
80 19%
RyR2R4496C+/–
* *
Event-free survival (%)
29% 8
60
F/F0
A 1
40
20
0
mV
0
0 10 20 30 40 50
n = 119 n = 105 n = 50 n = 29 n = 20 n=7
B –100
Figure 53-3. A, Representative line-scan images and Ca2+ transients in a wild type
Time from birth (years)
myocyte and a RyR2R4496C+/− myocyte at 0.2 Hz pacing in the presence of isopro-
Figure 53-2. Kaplan-Maier curve showing the cumulative occurrence of cardiac terenol (100 nM). Asterisk indicates spontaneous Ca2+ release in the RyR2R4496C+/−
arrest, sudden death, or ICD shocks in patients with catecholaminergic polymor- myocytes. Arrows indicated the field stimulations. B, Action potential recording in
phic ventricular tachycardia (CPVT) in the absence of therapy. An example of bidi- a myocyte isolated from a RyR2R4496C+/− during exposure to isoproterenol
rectional ventricular tachycardia, the typical CPVT arrhythmia, is depicted in the 30 nmol. Arrows indicate stimulated beats, which are followed by two triggered
inset. beats and delayed afterdepolarizations.
used to refer to SR Ca2+ release events that are not driven by a The use of cardiac myocytes derived from induced pluripotent
stimulated action potential. Whenever an SCR occurs and the stem cells (iPS) has been proposed recently as a means to study
levels of Ca2+ in the sarcolemma increase, the NCX activates to the consequences of mutations in cells derived directly from the
extrude the excess of ions. As mentioned earlier, NCX activation affected patients and to overcome the limitation of preceding
generates the Iti current, which is the cause of the development experimental models. Fatima et al.45 studied iPS-derived myo-
of delayed afterdepolarizations (DADs)33 and transient mem- cytes from a carrier of the F2483I mutation. Calcium imaging
brane depolarizations occurring during phase 4 of the action studies showed that, in agreement with a CPVT phenotype, iPS-
potential (see Figure 53-3). When DAD amplitude reaches the derived myocytes from the patient with CPVT presented abnor-
voltage threshold for sodium channel activation, triggered beats mal spontaneous calcium transients. Upon administration of
occur. isoproterenol, the same cells had a negative chronotropic response
The effects of CPVT mutations have been analyzed in a similar to that observed in CPVT mice.46 This finding suggests
variety of experimental settings, including lipid bilayer, heterolo- that the myocytes obtained by these authors are indeed similar
gous expression systems, murine transgenic models, and in myo- to nodal or embryonic automatic cells. The onset of DADs and
cytes derived from patient-specific induced pluripotent stem increased Ca2+ sparks frequency during adrenergic stimulation
cells.34 The following sections outline the key concepts about was observed by Jung et al.47 in myocytes from a carrier of the
arrhythmogenesis in each form of CPVT. S406L mutation. Overall, the data obtained in iPS-myocytes
from CPVT patients confirm the data observed previously in
Ryanodine Receptor, Mutations, and CPVT other experimental models, although the limitations of this
The RyR2 channel is a homotetramer; each subunit is formed by approach currently prevent significant steps in the understanding
4967 amino acids with a long (≈4300 amino acids) N-terminal of CPVT pathophysiology.34
cytoplasmic domain.35 The last 500 amino acids at the C-terminal Additional functional studies have investigated the effects of
of RyR2 form the transmembrane segments encircling the RyR2 mutation in specific anatomical structures. It is known
channel pore. The regulation of RyR2 opening and closing that Purkinje fibers are more susceptible to Ca2+ overload than
(gating) is mainly controlled by Ca2+ levels at cytoplasmic and ventricular muscle, possibly because of their greater sodium load
luminal SR sides, and it is facilitated when the Ca2+ concentration and longer action potential duration. Recently, two groups
at either side increases.36,37 Therefore, RyR2 calcium sensitivity reported that Purkinje fibers isolated from R4496C mutant mice
is an important physiological function that controls CICR. The display a greater propensity to develop intracellular Ca2+ han-
majority of RyR2 mutations identified in CPVT patients cause dling disorder than do ventricular myocytes, suggesting that
increased calcium sensitivity (defined as “gain-of-function”). focally activated arrhythmias might originate in the specialized
Marks et al. showed that mutant RyR2 exhibits an increased electrical conducting cells of the His-Purkinje system in
sensitivity to cytosolic Ca2+ after protein kinase A phosphoryla- CPVT.48,49 Direct mapping of bidirectional VT using voltage
tion. They also suggested that such change in sensitivity is due to sensitive dyes further supports this finding.50
an abnormal dissociation of FKBP12.6, a putative RyR2 stabiliz- The effect of the R4496C mutation, a typical CPVT muta-
ing protein.38 During adrenergic stimulation, phosphorylation of tion, has also been investigated at the level of sinus node.46 The
RyR2 would promote SCRs by further dissociation of FKBP12.6 rational for this study is the possible presence of lower than
with a consequent excessive increase of open probability. normal heart rate in patients with CPVT.28 In sinus cells isolated
An alternative (and strongly supported by the experimental from mice harboring the R4496C mutation, the study provided
evidence) hypothesis attributes a central role to the concept of initial evidences for a reduced pacemaker activity and impaired
SR threshold for calcium release. This mechanism is called store chronotropic response under β-adrenergic stimulation. This
overload-induced Ca2+ release to highlight the idea of the tight decreased automaticity appears to be mediated by a Ca2+-depen-
interplay between SR calcium content and release threshold. dent decrease of ICa and sarcoplasmic reticulum Ca2+ depletion
Researchers have demonstrated that several mutations of the during diastole upon adrenergic activation.46
RyR2 channel cause a reduction of the threshold for Ca2+ release;
on the contrary only few mutations are associated with a different RyR2 Mutations and Cardiomyopathy
sensitivity to cytosolic Ca2+ (i.e., they reduce the Ca2+ release The possible link between RyR2 mutations and structural abnor-
threshold).9,39,40 In the context of a lowered SR threshold (the malities has stimulated a recent experimental study.51 The authors
effect of the RyR2 mutation), β-adrenergic activation, which compared mutations possibly causing ARVC and HCM and
physiologically increases SR [Ca2+], greatly enhances the propen- focused the attention not only on the threshold for SR Ca2+
sity for SCR events because the threshold for release is reached release activation (affected in CPVT mutations), but also on the
more easily. threshold for Ca2+ release termination. The two thresholds con-
Yamamoto et al.41 have demonstrated that some mutations tribute to define the fractional release (i.e., the amount of total
disrupt the three-dimensional conformation of the channel. They SR Ca2+ released at each CICR cycle). This study showed that
initially showed that the closed state of the RyR2 channel is ARVC mutations cause reduced threshold for both activation and
stabilized by tight contacts between the central and N-terminal termination, with termination to a greater extent; this increased
regions. The presence of a reduced “stickiness” of these regions the fractional SR Ca2+ release, which is unaltered in RyR2-CPVT
is defined as “domain unzipping.” The same authors performed mutations.51 On the contrary the HCM-associated mutation
follow-up studies demonstrating that RyR2 mutations result in A1107M induces a significant reduction of fractional release
domain unzipping and enhance Ca2+ sensitivity, thus facilitating because of increased termination threshold. The authors con-
spontaneous Ca2+ release.42,43 The unzipping mechanism can clude that mutations affecting termination threshold and frac-
complement the store overload-induced Ca2+ release hypothesis tional release are associated with structural abnormalities, but
to provide a structural explanation of the reduced threshold for they could not provide an explanation for why fractional release
SR release. abnormalities should lead to rearrangements of the contractile
Although there is still debate over the subcellular mechanisms machinery.51
of RyR2 mutations, animal models consistently show that DADs Overall, the available evidence does not allow for a definitive
triggering action potentials12,44 is the cause of the onset of bidi- conclusion on the suggested link between RyR2 mutations and
rectional or polymorphic VTs. Adrenergic activation exacerbates specific cardiomyopathies, but it is possible that some mutants
the arrhythmogenic substrate, but the propensity for DAD and are associated with peculiar biophysical consequences that render
triggered beats can be observed even at baseline in isolated cells.12 the heart more prone to developing structural abnormalities.
Calsequestrin and CPVT arrhythmias leading to fatalities is possible. Therefore, several
53
CASQ2 is a 399 amino acid protein expressed at the level of the investigators have attempted novel strategies to achieve better
CRUs, where it acts as an intra-SR Ca2+ buffer and modulates protection.
RyR2 activity.52 At low Ca2+ levels (<0.6 mM) CASQ2 is present
as a monomer. However, at Ca2+ levels between 0.6 and 3.0 mM,
the protein starts to form dimers by N-terminal interactions. Remodulation of Abnormal Calcium Dynamics
Further increase of SR Ca2+ concentration (>3 mM) induces
CASQ2 polymerization by promoting additional binding through Experimental evidence has elucidated the possible role of fle-
C-terminal interactions. cainide65 on the basis of the observation that this drug reduces
CASQ2 also modulates RyR2 via triadin and junctin.53 A total the durations of channel openings and disrupts the propagation
of fourteen CASQ2 CPVT-related mutations have been described. of calcium waves. Thus, a direct effect on the RyR2 channel was
Functional characterization has shown multiple effects: reduced postulated. Subsequent work challenged this concept by provid-
CASQ expression levels,54,55 impaired polymerization and buffering evidence that the most important action of flecainide in coun-
ing capacity,56,57 and reduced CASQ2-RyR2 binding with loss of teracting a leaky RyR2 channel occurs through the sodium
RyR2 modulation.58,59 These mechanisms work in conjunction to current inhibition and its negative bathmotropic effect.44 Indeed,
generate SCRs and DADs and are further exacerbated upon flecainide does not inhibit SCRs, but it reduces the propensity
adrenergic activation that induces triggered arrhythmias. for triggered beats.44 Although the mechanism is still debated,
Although RyR2 Ca2+ sensitivity is directly proportional to the clinical effectiveness of flecainide has been confirmed by
SR-free Ca2+ content,3,60,61 the presence of CASQ2 mutations can independent investigators in small series.66,67 This approach
increase RyR2 sensitivity by impairing the polymerization capac- should be considered in addition to β-blockers when insufficient
ity or by reducing the CASQ2 expression levels. Less clear is the protection is demonstrated. Theoretical studies and anecdotal
role of calcium buffering properties, because SR calcium content clinical cases have also proposed a role for calcium channel block-
was found to be either reduced or normal in CASQ2-deficient ers as a mean to remodulate calcium dynamics in CPVT,68,69 but
mice.55,62 a clear demonstration of effectiveness is lacking.
Another interesting and puzzling effect of CASQ2 mutations
is the presence of ultrastructural abnormalities. Loss of calse-
questrin causes a spectrum of ultrastructural changes at the SR Calmodulin Kinase Pathway
level, detectable with electron microscopy. The chainlike polymer
at the level of junctional SR (representing CASQ2 protein) is Another interesting approach is that of inhibiting the effects of
absent54,62; couplons are shorter or junctional SR is fragmented β-adrenergic stimulation by acting on the downstream targets of
(Figure 53-4). RyR2 phosphorylation. Calmodulin kinase (CAMKII) phosphor-
Whether these ultrastructural abnormalities have conse- ylates the RyR2 at different sites. Moreover, CAMKII inhibition
quences for, or are the cause of, Ca2+ releasing abnormalities is reduces ICa, diastolic Ca2+ leakage, and the Iti current.70 There
still a matter of debate; however, this peculiar feature makes is also evidence that a specific CAMKII inhibitor, KN93, could
recessive CPVT a form of ultrastructural cardiomyopathy. It is prevent arrhythmias both in vitro and in vivo.71
of considerable interest to note that ultrastructural abnormalities
have also been observed in human cardiac myocytes harboring
the D307H mutation derived from iPS.63 Electron microscopy RyR2 Channel Stabilization
has shown that CPVT myocytes present an immature morphol-
ogy with less-organized myofibrils, enlarged sarcoplasmic reticu- An increased FKBP12.6 dissociation could be one of the mecha-
lum cisternae, and a reduced number of caveolae. nisms for arrhythmogenesis in the context of a mutant RyR2
channel. Accordingly, pharmacologic enhancement or restora-
Triadin and CPVT tion of FKBB12.6 binding might represent a way to revert CPVT
Only one study reporting two small human TRND families exists arrhythmias. It has been shown that the FKBP12.6 knockout
in the literature.19 As a consequence, there is only scanty direct mouse model develops stress-induced arrhythmias mainly because
evidence of the functional consequences of mutations. Data from of increased RyR2 open probability.38 In this model, the treat-
a TRDN null mouse have demonstrated the propensity toward ment with the benzothiazepine compound K201 was able to
adrenergically induced arrhythmias and, interestingly, ultrastruc- restore FKBP12.6-RyR2 binding and to prevent arrhythmias.38
tural abnormalities not dissimilar from that of CASQ2 mice.64 Similar results were also observed directly in a CPVT knock-in
Heterologous expression of the human TRDN-T59R mutant in mouse model harboring the R2474S mutation.72 More recently,
COS-7 resulted in intracellular retention and degradation of the S107 (a drug similar to K201) was found to inhibit the induc-
mutant protein.19 No calcium dynamics data are available, but ibility of atrial fibrillation in other CPVT mouse models73;
given the tight physiological activity coordination among RyR2, however, conflicting evidence has challenged these results.
CASQ2, and TRND, it is conceivable that the lack of triadin George et al.74 and Jiang et al.39 showed that mutant RyR2 could
might produce an abnormal RyR2 open kinetics by altering SR interact normally with FKBP12.6, and Liu et al.12 observed
calcium sensitivity and reducing the SR releasing threshold, as normal FKBP12.6 binding and no significant antiarrhythmic
observed for CASQ2 mutants. The evidence of the tight interplay effect of K201 both in vivo and in vitro.
between CASQ and TRDN supports this hypothesis. Indeed, it
is known that a reduction of CASQ2 expression is associated with
a reduction of TRDN levels and vice versa.54,64 Gene Therapy
Restoration of a functionally normal gene function is clearly an
attractive goal for genetic diseases. This approach has been
Experimental Therapies for attempted for autosomal recessive CPVT, and it was demon-
Calcium Handling Disorders strated that infection of an adeno-associated viral (AAV) vector
harboring wild type calsequestrin achieves long-term (at least
The use of β-blockers is accepted as the most effective approach 3 months) reexpression of properly localized CASQ2 peptides.75
for patients with CPVT, although the occurrence of severe In parallel, there was complete protection from adrenergically
WT KO INF
A B C
CASQ2 CASQ2 CASQ2
240 RyR2 240 RyR2 240 RyR2
200 200 200
Intensity
Intensity
Intensity
160 160 160
120 120 120
80 80 80
40 40 40
0 0 0
Length (µm) Length (µm) Length (µm)
WT KO INF
jSR jSR
jSR
TT
TT TT
A B C
KO INF
Control
Epinephrine 2mg/kg – 1 minute
Figure 53-4. High-magnification immunofluorescent image of showing colocalization of CASQ2 (red) and RyR2 (blue) in a wild type cell (A-WT). As expected, no CASQ2
signal is evident in the CASQ2 knockout cell (B-KO). Conversely knockout cells infected with AAV9-CASQ2 (C-inf ) show restoration of correct RyR2/CASQ2 colocalization at
the z-line level. The middle panel shows electron micrographs of exemplificative cells from wild type (WT), knockout (KO), and KO 20 weeks after AAV9-CASQ2 infection.
The electron-dense material inside the junctional sarcoplasmic reticulum (jSR) reappears in the infected cell (INF), whereas the ultrastructural abnormalities (jSR enlargement
and fragmentation) are completely reversed. When challenged with epinephrine to mimic adrenergic activation, the infected mice (lower panel, INF) show a clear antiar-
rhythmic effect in vivo compared with KO animals (lower panel, KO).
induced arrhythmias and a regression of ultrastructural abnor- release in cardiac muscle is unclear, it has been shown to suppress
malities (see Figure 53-4). Because AAV vectors have been used ventricular arrhythmias in the R2474S+/− mouse model of CPVT.77
in the clinical setting for gene therapy of cardiac disease,76 this Interestingly these data have been confirmed in cardiomyocytes
approach could possibly be transferred to clinical practice. derived from iPS of a patient with CPVT.47
Dantrolene
Conclusions
It has long been known that dantrolene suppresses spontaneous
SR Ca2+ release in skeletal muscle. Dantrolene does not seem to Inherited dysfunctions of calcium handling system are associated
inhibit the function of normal RyR2 channels; it can bind to and with a progressively growing spectrum of clinical phenotypes.
inhibit abnormal RyR2 channels with defective domain interac- The experimental studies addressing the genetic basis and
tions. Although the action of dantrolene on spontaneous SR Ca2+ pathophysiology of these disorders have clearly shown the
interdependence of the plasmalemmal component (voltage- than the arrhythmogenic disorders associated to mutations in
53
gated calcium channels and associated proteins) and the SR sodium or potassium handling systems, possibly because of
component (ryanodine receptor, calsequestrin, and other its critical physiological role. The remarkable knowledge gained
associated peptides). The perturbation of one of these compo- in the last decade allows for a better understanding of the
nents unavoidably reflects on the others’ function to cause the mechanism and the design of novel therapeutic strategies, includ-
development of highly arrhythmogenic substrate. In general, ing the proof of principle that a cure is possible through gene
the inherited disease of calcium handling appears more severe therapy.
CASQ2 is associated with autosomal recessive steroids in Purkinje fibres. J Physiol 263:73–100,
References catecholamine-induced polymorphic ventricular 1976.
tachycardia in Bedouin families from Israel. Am J 34. Priori SG, Napolitano C, Di Pasquale E, et al:
1. Napolitano C, Bloise R, Monteforte N, et al: Hum Genet 69:1378–1384, 2001. Induced pluripotent stem cell-derived cardiomyo-
Sudden cardiac death and genetic ion channelopa- 19. Roux-Buisson N, Cacheux M, Fourest-Lieuvin A, cytes in studies of inherited arrhythmias. J Clin
thies: long QT, Brugada, short QT, catecholamin- et al: Absence of triadin, a protein of the calcium Invest 2012;(in press).
ergic polymorphic ventricular tachycardia, and release complex, is responsible for cardiac arrhyth- 35. George CH, Jundi H, Thomas NL, et al: Ryano-
idiopathic ventricular fibrillation. Circulation mia with sudden death in human. Hum Mol Genet dine receptors and ventricular arrhythmias: emerg-
125:2027–2034, 2012. 21:2759–2767, 2012. ing trends in mutations, mechanisms and therapies.
2. Shannon TR, Guo T, Bers DM: Ca2+ scraps: local 20. Ackerman MJ, Priori SG, Willems S, et al: HRS/ J Mol Cell Cardiol 42:34–50, 2007.
depletions of free [Ca2+] in cardiac sarcoplasmic EHRA Expert Consensus Statement on the State 36. Sitsapesan R, Williams AJ: Gating of the native and
reticulum during contractions leave substantial of Genetic Testing for the Channelopathies and purified cardiac SR Ca(2+)-release channel with
Ca2+ reserve. Circ Res 93:40–45, 2003. Cardiomyopathies: This document was developed monovalent cations as permeant species. Biophys.J
3. Fabiato A, Fabiato F: Calcium-induced release as a partnership between the Heart Rhythm 67:1484–1494, 1994.
of calcium from the sarcoplasmic reticulum of Society (HRS) and the European Heart Rhythm 37. Sitsapesan R, Williams AJ: Regulation of the gating
skinned cells from adult human, dog, cat, rabbit, Association (EHRA). Europace 13:1077–1109, of the sheep cardiac sarcoplasmic reticulum
rat, and frog hearts and from fetal and new-born 2011. Ca(2+)-release channel by luminal Ca2+. J Membr
rat ventricles. Ann N Y Acad.Sci 307:491–522, 21. Cerrone M, Priori SG: Genetics of sudden death: Biol 137:215–226, 1994.
1978. focus on inherited channelopathies. Eur Heart J 38. Wehrens XH, Lehnart SE, Huang F, et al:
4. Napolitano C, Antzelevitch C: Phenotypical mani- 2011. FKBP12.6 deficiency and defective calcium release
festations of mutations in the genes encoding sub- 22. Cerrone M, De Giuli L, Bloise R, et al: Value of channel (ryanodine receptor) function linked to
units of the cardiac voltage-dependent L-type entire open reading frame screening of the cardiac exercise-induced sudden cardiac death. Cell
calcium channel. Circ Res 108:607–618, 2011. ryanodine receptor gene RyR2. Evidence from the 113:829–840, 2003.
5. Franzini-Armstrong C, Protasi F, Ramesh V: Italian CPVT registry. Heart Rhythm 6:S462, 39. Jiang D, Wang R, Xiao B, et al: Enhanced store
Shape, size, and distribution of Ca(2+) release units 2011. overload-induced Ca2+ release and channel sensi-
and couplons in skeletal and cardiac muscles. 23. Priori SG, Napolitano C, Memmi M, et al: Clinical tivity to luminal Ca2+ activation are common
Biophys J 77:1528–1539, 1999. and molecular characterization of patients with defects of RyR2 mutations linked to ventricular
6. Franzini-Armstrong C, Protasi F, Tijskens P: The catecholaminergic polymorphic ventricular tachy- tachycardia and sudden death. Circ Res 97:1173–
assembly of calcium release units in cardiac muscle. cardia. Circulation 106:69–74, 2002. 1181, 2005.
Ann N Y Acad Sci 1047:76–85, 2005. 24. Tester DJ, Kopplin LJ, Will ML, et al: Spectrum 40. Jiang D, Xiao B, Yang D, et al: RyR2 mutations
7. Cannell MB, Cheng H, Lederer WJ: The control and prevalence of cardiac ryanodine receptor linked to ventricular tachycardia and sudden death
of calcium release in heart muscle. Science (RyR2) mutations in a cohort of unrelated patients reduce the threshold for store-overload-induced
268:1045–1049, 1995. referred explicitly for long QT syndrome genetic Ca2+ release (SOICR). Proc Natl Acad Sci U S A
8. Satoh H, Blatter LA, Bers DM: Effects of [Ca2+]i, testing. Heart Rhythm 2:1099–1105, 2005. 101:13062–13067, 2004.
SR Ca2+ load, and rest on Ca2+ spark frequency 25. Coumel P, Fidelle J, Lucet V, et al: 41. Yamamoto T, Yano M, Xu X, et al: Identification
in ventricular myocytes. Am J Physiol 272:H657– Catecholaminergic-induced severe ventricular of target domains of the cardiac ryanodine receptor
H668, 1997. arrhythmias with Adams-Stokes syndrome in chil- to correct channel disorder in failing hearts. Cir-
9. Priori SG, Chen SR: Inherited dysfunction of sar- dren: report of four cases. Br Heart J 40:28–37, culation 117:762–772, 2008.
coplasmic reticulum Ca2+ handling and arrhyth- 1978. 42. Tateishi H, Yano M, Mochizuki M, et al: Defective
mogenesis. Circ Res 108:871–883, 2011. 26. Leenhardt A, Lucet V, Denjoy I, et al: Catechol- domain-domain interactions within the ryanodine
10. Bers DM: Cardiac excitation-contraction coupling. aminergic polymorphic ventricular tachycardia in receptor as a critical cause of diastolic Ca2+ leak in
Nature 415:198–205, 2002. children. A 7-year follow-up of 21 patients. Circu- failing hearts. Cardiovasc Res 81:536–545, 2009.
11. Pieske B, Maier LS, Bers DM, et al: Ca2+ handling lation 91:1512–1519, 1995. 43. Uchinoumi H, Yano M, Suetomi T, et al: Catechol-
and sarcoplasmic reticulum Ca2+ content in iso- 27. Hayashi M, Denjoy I, Extramiana F, et al: aminergic polymorphic ventricular tachycardia is
lated failing and nonfailing human myocardium. Incidence and risk factors of arrhythmic events in caused by mutation-linked defective conforma-
Circ Res 85:38–46, 1999. catecholaminergic polymorphic ventricular tachy- tional regulation of the ryanodine receptor. Circ
12. Liu N, Colombi B, Memmi M, et al: Arrhythmo- cardia. Circulation 119:2426–2434, 2009. Res 106:1413–1424, 2010.
genesis in catecholaminergic polymorphic ven- 28. Postma AV, Denjoy I, Kamblock J, et al: Catechol- 44. Liu N, Denegri M, Ruan Y, et al: Short communi-
tricular tachycardia. Insights from a RyR2 R4496C aminergic polymorphic ventricular tachycardia: cation: flecainide exerts an antiarrhythmic effect in
knock-in mouse model. Circ Res 99:292–298, RYR2 mutations, bradycardia, and follow up of the a mouse model of catecholaminergic polymorphic
2006. patients. J Med Genet 42:863–870, 2005. ventricular tachycardia by increasing the Thresh-
13. Ginsburg KS, Bers DM: Modulation of excitation- 29. Tiso N, Stephan DA, Nava A, et al: Identification old for triggered activity. Circ Res 109:291–295,
contraction coupling by isoproterenol in cardio- on mutations in the cardiac ryanodine receptor 2011.
myocytes with controlled SR Ca2+ load and Ca2+ gene in families affected with arrhythmogenic right 45. Fatima A, Xu G, Shao K, et al: In vitro modeling
current trigger. J Physiol 556:463–480, 2004. ventricular cardiomyopathy type 2 (ARVD2). Hum of ryanodine receptor 2 dysfunction using human
14. Hussain M, Orchard CH: Sarcoplasmic reticulum Mol Genet 10:189–194, 2001. induced pluripotent stem cells. Cell Physiol
Ca2+ content, L-type Ca2+ current and the Ca2+ 30. Fujino N, Ino H, Hayashi K, et al: A novel mis- Biochem 28:579–592, 2011.
transient in rat myocytes during beta-adrenergic sense mutation in cardiac ryanodine receptor gene 46. Neco P, Torrente AG, Mesirca P, et al: Paradoxical
stimulation. J Physiol 505(Pt 2):385–402, 1997. as a possible cause of hypertrophic cardiomyopa- effect of increased diastolic Ca2+ release and
15. Grimm M, Brown JH: Beta-adrenergic receptor thy: evidence from familial analysis. Circulation decreased sinoatrial node activity in a mouse model
signaling in the heart: role of CaMKII. J Mol Cell 114(Suppl II):164 (ABS), 2006. of catecholaminergic polymorphic ventricular
Cardiol 48:322–330, 2010. 31. van der Werf C, Hofman N, Tan HL, et al: Diag- tachycardia. Circulation 126:392–401, 2012.
16. Wehrens XH, Lehnart SE, Reiken SR, et al: Pro- nostic yield in sudden unexplained death and 47. Jung CB, Moretti A, Mederos y Schnitzler M,
tection from cardiac arrhythmia through ryanodine aborted cardiac arrest in the young: the experience et al: Dantrolene rescues arrhythmogenic RYR2
receptor-stabilizing protein calstabin2. Science of a tertiary referral center in The Netherlands. defect in a patient-specific stem cell model of cat-
304:292–296, 2004. Heart Rhythm 7:1383–1389, 2010. echolaminergic polymorphic ventricular tachycar-
17. Priori SG, Napolitano C, Tiso N, et al: Mutations 32. Hong RA, Rivera KK, Jittirat A, et al: Flecainide dia. EMBO Mol Med 4:180–191, 2011.
in the cardiac ryanodine receptor gene (hryr2) suppresses defibrillator-induced storming in cate- 48. Hilliard FA, Steele DS, Laver D, et al: Flecainide
underlie catecholaminergic polymorphic ventricu- cholaminergic polymorphic ventricular tachycar- inhibits arrhythmogenic Ca2+ waves by open state
lar tachycardia. Circulation 103:196–200, 2001. dia. Pacing Clin Electrophysiol 35:794–797, 2012. block of ryanodine receptor Ca2+ release channels
18. Lahat H, Pras E, Olender T, et al: A missense 33. Lederer WJ, Tsien RW: Transient inward current and reduction of Ca2+ spark mass. J Mol Cell
mutation in a highly conserved region of underlying arrhythmogenic effects of cardiotonic Cardiol 48:293–301, 2009.
49. Herron TJ, Milstein ML, Anumonwo J, et al: Pur- exercise-induced sudden cardiac death. Circ Res ventricular tachycardia. Heart Rhythm 4:1149–
kinje cell calcium dysregulation is the cellular 98:1151–1158, 2006. 1154, 2007.
mechanism that underlies catecholaminergic poly- 59. Houle TD, Ram ML, Cala SE: Calsequestrin 69. Sung RJ, Lo CP, Hsiao PY, et al: Targeting intra-
morphic ventricular tachycardia. Heart Rhythm mutant D307H exhibits depressed binding to its cellular calcium cycling in catecholaminergic poly-
7:1122–1128, 2010. protein targets and a depressed response to morphic ventricular tachycardia: a theoretical
50. Cerrone M, Noujaim SF, Tolkacheva EG, et al: calcium. Cardiovasc Res 64:227–233, 2004. investigation. Am J Physiol Heart Circ Physiol
Arrhythmogenic mechanisms in a mouse model of 60. Qin J, Valle G, Nani A, et al: Luminal Ca2+ regula- 301:H1625–H1638, 2011.
catecholaminergic polymorphic ventricular tachy- tion of single cardiac ryanodine receptors: insights 70. Wu Y, Temple J, Zhang R, et al: Calmodulin kinase
cardia. Circ Res 11:1039–1048, 2007. provided by calsequestrin and its mutants. J Gen II and arrhythmias in a mouse model of cardiac
51. Tang Y, Tian X, Wang R, et al: Abnormal termina- Physiol 131:325–334, 2008. hypertrophy. Circulation 106:1288–1293, 2002.
tion of Ca2+ release is a common defect of RyR2 61. Gyorke I, Gyorke S: Regulation of the cardiac 71. Liu N, Ruan Y, Denegri M, et al: Calmodulin
mutations associated with cardiomyopathies. Circ ryanodine receptor channel by luminal Ca2+ kinase II inhibition prevents arrhythmias in
Res 110:968–977, 2012. involves luminal Ca2+ sensing sites. Biophys J RyR2R4496C+/− mice with catecholaminergic
52. Gyorke S, Terentyev D: Modulation of ryanodine 75:2801–2810, 1998. polymorphic ventricular tachycardia. J Mol Cell
receptor by luminal calcium and accessory proteins 62. Knollmann BC, Chopra N, Hlaing T, et al: Casq2 Cardiol 50:214–222, 2011.
in health and cardiac disease. Cardiovasc Res deletion causes sarcoplasmic reticulum volume 72. Lehnart SE, Mongillo M, Bellinger A, et al: Leaky
77:245–255, 2008. increase, premature Ca2+ release, and catechol- Ca2+ release channel/ryanodine receptor 2 causes
53. Gyorke I, Hester N, Jones LR, et al: The role of aminergic polymorphic ventricular tachycardia. seizures and sudden cardiac death in mice. J Clin
calsequestrin, triadin, and junctin in conferring J Clin Invest 116:2510–2520, 2006. Invest 118:2230–2245, 2008.
cardiac ryanodine receptor responsiveness to 63. Novak A, Barad L, Zeevi-Levin N, et al: Cardio- 73. Shan J, Xie W, Betzenhauser M, et al: Calcium leak
luminal calcium. Biophys J 86:2121–2128, 2004. myocytes generated from CPVTD307H patients through ryanodine receptors leads to atrial fibrilla-
54. Rizzi N, Liu N, Napolitano C, et al: Unexpected are arrhythmogenic in response to beta-adrenergic tion in three mouse models of catecholaminergic
structural and functional consequences of the stimulation. J Cell Mol Med 16:468–482, 2012. polymorphic ventricular tachycardia. Circ Res
R33Q homozygous mutation in cardiac calseques- 64. Chopra N, Yang T, Asghari P, et al: Ablation of 2012.
trin: a complex arrhythmogenic cascade in a triadin causes loss of cardiac Ca2+ release units, 74. George CH, Higgs GV, Lai FA: Ryanodine recep-
knock in mouse model. Circ Res 103:298–306, impaired excitation-contraction coupling, and tor mutations associated with stress-induced ven-
2008. cardiac arrhythmias. Proc Natl Acad Sci U S A tricular tachycardia mediate increased calcium
55. Song L, Alcalai R, Arad M, et al: Calsequestrin 2 106:7636–7641, 2009. release in stimulated cardiomyocytes. Circ Res
(CASQ2) mutations increase expression of calre- 65. Watanabe H, Chopra N, Laver D, et al: Flecainide 93:531–540, 2003.
ticulin and ryanodine receptors, causing catechol- prevents catecholaminergic polymorphic ventricu- 75. Denegri M, Avelino-Cruz JE, Boncompagni S,
aminergic polymorphic ventricular tachycardia. lar tachycardia in mice and humans. Nat Med et al: Viral gene transfer rescues arrhythmogenic
J Clin Invest 117:1814–1823, 2007. 15:380–383, 2009. phenotype and ultrastructural abnormalities in
56. Valle G, Galla D, Nori A, et al: Catecholaminergic 66. Pott C, Dechering DG, Reinke F, et al: Successful adult calsequestrin-null mice with inherited
polymorphic ventricular tachycardia-related muta- treatment of catecholaminergic polymorphic ven- arrhythmias. Circ Res 2012.
tions R33Q and L167H alter calcium sensitivity of tricular tachycardia with flecainide: a case report 76. Jessup M, Greenberg B, Mancini D, et al: Calcium
human cardiac calsequestrin. Biochem J 413:291– and review of the current literature. Europace Upregulation by Percutaneous Administration of
303, 2008. 2010. Gene Therapy in Cardiac Disease (CUPID): a
57. Kalyanasundaram A, Bal NC, Franzini-Armstrong 67. van der Werf C, Kannankeril PJ, Sacher F, et al: phase 2 trial of intracoronary gene therapy of sar-
C, et al: The calsequestrin mutation CASQ2D307H Flecainide therapy reduces exercise-induced ven- coplasmic reticulum Ca2+-ATPase in patients with
does not affect protein stability and targeting to the tricular arrhythmias in patients with catecholamin- advanced heart failure. Circulation 124:304–313,
junctional sarcoplasmic reticulum but compro- ergic polymorphic ventricular tachycardia. J Am 2011.
mises its dynamic regulation of calcium buffering. Coll Cardiol 57:2244–2254, 2011. 77. Kobayashi S, Yano M, Uchinoumi H, et al: Dan-
J Biol Chem 285:3076–3083, 2009. 68. Rosso R, Kalman JM, Rogowski O, et al: Calcium trolene, a therapeutic agent for malignant hyper-
58. Terentyev D, Nori A, Santoro M, et al: Abnormal channel blockers and beta-blockers versus beta- thermia, inhibits catecholaminergic polymorphic
interactions of calsequestrin with the ryanodine blockers alone for preventing exercise-induced ventricular tachycardia in a RyR2(R2474S/+)
receptor calcium release channel complex linked to arrhythmias in catecholaminergic polymorphic knock-in mouse model. Circ J 74:2579–2584, 2010.
Pharmacologic, Genetic, and Cell PART IX
Therapy of Ion Channel Dysfunction
Pharmacologic Bases of
Antiarrhythmic Therapy 54
Juan Tamargo and Eva Delpón
exacerbation of the treated arrhythmias or generation of entirely

CHAPTER OUTLINE new ones (proarrhythmia).
Antiarrhythmic Drugs: Introduction 529 The selection of pharmacologic or nonpharmacologic thera-
+
peutic strategies differs with respect to the origin of the arrhyth-
Na Channel Blockers: Class I AADs 529 mia. AADs are used for ventricular tachycardias (VTs) in
β-Blockers: Class II AADs 533 competition with implantable cardioverter defibrillators (ICDs)
and ablation techniques as additional therapy to reduce the
K+ Channel Blockers: Class III Antiarrhythmic Effects 533 number of necessary shocks. Atrial tachyarrhythmias, particularly
Calcium Channel Blockers: Class IV AADS 535 atrial fibrillation (AF), which is the most common arrhythmia,
are the domain for drug treatment, although AF is particularly
Gap-Junction Coupling Enhancers 535 difficult to treat because of its tendency to become persistent in
Stretch-Activated Ion Channels 536 the course of the disease.
Until recently, arrhythmias were primarily considered a purely
Modulation of Ion Channel Trafficking 536 electrophysiological problem, and AADs mainly target cardiac
Targeting Intracellular Calcium Handling 536 Na+, Ca2+, and K+ ion channels (Figure 54-1). They bind to spe-
cific receptor sites within the channel drug affinity being strongly
Pharmacologic Treatment of Inherited Cardiac modulated by the channel state in a time- and voltage-dependent
Arrhythmia Syndromes 537 manner.2 In addition, some AADs modulate the autonomic tone,
Target Cardiac Remodeling: Upstream Therapies 537 primarily by antagonizing β1-adrenoceptors (β-blockers) or mus-
carinic receptors (atropine, disopyramide) or by stimulating
Other AADs 538 adenosine A1 receptors (adenosine).
Unresolved Questions and Future Strategies 538 Conventional AADs are most commonly grouped according
to the Vaughan-Williams classification in four different groups
Conclusions 539 (Table 54-1): Na+ channel-blocking drugs (class I), β-blockers
(class II), K+ channel blockers (class III) that prolong action
potential (AP) duration (APD) and refractoriness, and L-type
Ca2+ channel blockers (class IV). However, most AADs are “dirty
drugs” that in a narrow range of concentrations simultaneously
Antiarrhythmic Drugs: Introduction block different types of channels and modulate adrenergic and/
or muscarinic receptors (e.g., quinidine, disopyramide, propafe-
Cardiac arrhythmias represent a problem in current clinical prac- none, amiodarone, dronedarone). Moreover, some old (adenos-
tice; they are a major cause of morbidity and mortality and are ine, digoxin) and new AADs (ivabradine) cannot be listed in any
an increasing economic burden for health care systems.1 Treat- of the original four classes. Table 54-2 shows drug selection based
ment of cardiac arrhythmias with antiarrhythmic drugs (AADs) on the mechanisms underlying cardiac arrhythmias, and Figure
has two main objectives: relieve symptoms and complications 54-2 shows the main cellular targets for AADs.
(improve quality of life) and reduce mortality directly related to
the arrhythmia. A basic principle in pharmacology is that the best
treatment is targeted specifically to disease mechanisms. However,
in many patients the ultimate underlying mechanisms of the Na+ Channel Blockers: Class I AADs
arrhythmia remain incompletely understood. Thus, the choice of
a given AAD is empiric and based on the characteristics of the Na+ channel blockers bind and unbind in a strongly time- and
arrhythmia, the pharmacologic properties of the AAD, and, voltage-dependent manner to a receptor site within the pore-
above all, its safety profile. Moreover, triggers and arrhythmo- forming subunit (Nav1.5) of the channel when channels are in
genic substrates can vary among patients with the same arrhyth- the activated or inactivated state.2 Na+ channel blockers also slow
mia depending on the underlying structural heart disease (i.e., Na+ channel reactivation (transition from the inactivated to the
coronary artery disease [CAD], heart failure [HF], left ventricular resting state) upon repolarization, which prolongs refractoriness
[LV] hypertrophy, hypertension). This variation could explain independently from changes in APD—that is, they increase the
why AADs produce widely divergent effects on different patients effective ratio of refractory period (ERP) to APD (postrepolariza-
ranging from termination of the arrhythmia to inefficacy, to tion refractoriness). As a consequence, the effects of class I AAD
529
530 PHARMACOLOGIC, GENETIC, AND CELL THERAPY OF ION CHANNEL DYSFUNCTION
Atrial Ventricular would be expected to increase: (1) at faster rates of stimulation

(use-dependent block), because Na+ channels spend more time in
20 mV
I to1 I to1 activated and inactivated states, and the diastolic time for recov-
ery from drug-induced block is shortened; and (2) in depolarized
IKur INaL ICaL
100 ms IKr cardiac tissues (voltage-dependent block) because membrane
ICaL depolarization inactivates Na+ channels and slows recovery from
INaL IKs
block.
IKr
IK1, IKATP Clinical and experimental data suggest that block of atrial Na+
IKs INa channels can terminate AF. Indeed, propafenone and flecainide
INa IK1, I KACh, IKATP
are first-choice drugs for AF cardioversion, but only in patients
without structural heart disease. Conversely, ventricular Na+
channel block is associated with proarrhythmic effects. Class I
AADs decrease excitability and slow ventricular conduction and
Figure 54-1. Ionic currents involved in shaping of human atrial (A) and ventricular can increase all-cause mortality and, therefore, they have no role
(B) action potentials. The initial upstroke of the action potential is due to the activa- (or are contraindicated) in the primary prevention of sudden
tion of the peak inward Na+ current (INa). Repolarization is determined by the cardiac death (SCD) because of life-threatening ventricular
balance between inward-depolarizing L-type Ca2+ (ICaL) and late Na+ (INaL) currents tachyarrhythmias (VT or ventricular fibrillation [VF]) in high-
and outward-repolarizing K+ currents, including the rapidly activating and inactivat- risk patients with post–myocardial infarction.3 In secondary pre-
ing transient current (Ito1), ultrarapid (IKur), rapid (IKr), and slow (IKs) components of vention, the use of class I AADs is limited because of their low
the delayed rectifier, the inward rectifier current (IK1) and the ligand-gated currents efficacy and serious adverse effects that may offset the therapeutic
activated by a decrease in the intracellular concentration of adenosine triphos-
benefit, including proarrhythmia, multiorgan toxicity, reduced
phate (IKATP) or activated by acetylcholine (IKAch) or adenosine (IKAdo). Some currents
are exclusive to the atria (IKur) or more important (Ito1, IKACh) in the atria than in the cardiac function, or worsening coexisting diseases.4-6
ventricles and vice versa (IKr). Arrows indicate the ion movement direction. Class I drugs are subdivided into drugs with intermediate (IA),
fast (IB), and slow (IC) onset/offset kinetics of Na+ channel block
(see Table 54-1). Recently, it has been demonstrated experimen-
tally that drugs with fast onset/offset kinetics of Na+ channel
block, such as vernakalant and ranolazine, could selectively target
atrial versus ventricular Na+ channels.7,8 This targeting would
allow them to exhibit a disease-specific component of
NCX
inhibitors Digoxin
Class I Class III Na+ – Ca2+ Na+ – K+
exchanger ATPase
Na+ Kv Kir
Ca2+ 3Na+
NCX ATPase
Rycals K+ K+
Class IC
Carvedilol calstabin2 SR Connexion Gap-junction
channel enhancers
Calsequestrin
Ca2+
Ca2+
Class IV RyR2
P Ca2+
Nucleus
P
SERCA2 PLB
P
CaMKII Forward
Golgi
PKA trafficking
Ion channel
cAMP trafficking
Na+, K+ ATP K+
+ Membrane
+ AC – +
Gαs Gαi
insertion
Ivabradine β1AR M2R Adenosine

If Class II A1R KACh KACh inhibitors
KAdo
Figure 54-2. Cellular targets of antiarrhythmic drugs (AADs). Most available AADs modify the conductance of Na+ (class I), Ca2+ (class IV), and K+ ion channels (class III)
located in the sarcolemma, leading to a decrease or increase in a given ionic current, or they block β-adrenergic (class II), adenosine-A1, and muscarinic-M2 receptors. Some
AADs inhibit abnormal persistent Na+ entry and diastolic Ca2+ release via the ryanodine receptor RyR2, with minimal effects on normal channel function. Furthermore, they
can inhibit ion exchangers (Na+-Ca2+ [NCX]) or pumps (Na+-K+ ATPase), increase gap junction conductance, and correct ion channel expression within the membrane by
modulating trafficking pathways. A1R, adenosine A1 receptor; AC, adenylyl cyclase; ATPase, Na+-K+ ATPase; β1AR, β1-adrenergic receptor; CaMKII, Ca2+/calmodulin kinase II;
CSQ, calsequestrin; Gs, stimulatory G protein; M2R, muscarinic type-2 receptor; PBL, phospholamban; PKA, protein kinase A; RyR2, ryanodine receptor; SERCA2a, SR Ca2+-ATPase;
SR, sarcoplasmic reticulum.
Pharmacologic Bases of Antiarrhythmic Therapy 531
Table 54-1. Antiarrhythmic Drug Classes
Class Main Mechanism of Action

+
Subclass Other Actions and Effects
+
Drugs 54
I Block Na channels IA. Intermediate offset Block K channels Disopyramide, procainamide,
Decrease excitability and slow CV kinetics (τre 1-3 s) Prolong APD quinidine
Prolong ERP (postrepolarization ↑ QRS and QT
refractoriness) IB. Fast offset kinetics Shorten APD Lidocaine, mexiletine
(τre 300-500 ms)
IC. Slow offset kinetics Markedly slow conduction Ajmaline, flecainide,
(τre > 5 s) No effects on ventricular APD propafenone
↑ RR, PR, and QRS*
Fast atrial selective Blocks INaL and K+ channels (IKur,IK,Ach) Ranolazine
Antianginal drug
Blocks K+ channels (IKur, Ito1, IKr, IKACh) Vernakalant
II β-Adrenoceptor antagonists Reduce sinus rate and AV conduction Atenolol, carvedilol,
↑ RR and PR metoprolol, propranolol
(and others)
III Pure IKr inhibitor Prolong atrial and ventricular APD/ Dofetilide
ERP
IKr inhibitor, INaL agonist Ibutilide
↑ QT
IKr inhibitor and β-blocker D-sotalol
Block Na+, Ca2+, and several K+ channels ↑ RR, PR, QRS, and QT Amiodarone, dronedarone
and α-/β-adrenoceptors
IV Block L-type Ca2+ channels ↑ RR and PR Verapamil, diltiazem
Other Block IKAdo, inhibit cAMP-induced ICaL ↑ RR and PR Adenosine
Block Na /K ATPase, ↑ vagal activity
+ +
↑ RR and PR Digoxin
Specific If inhibitor ↑ RR Ivabradine
*RR and PR intervals of the electrocardiogram.

CV, Conduction velocity; APD, action potential duration; ERP, effective refractory period; RR, antianginal drug; ↑, increase.
atrial-selectivity because of the high atrial rate in AF, which with minor ventricular effects, also inhibit the rapid component
enhances block of atrial over ventricular Na+ channels. Indeed, it of the delayed rectifier current (IKr) and other K+ currents pro-
has been proposed that the faster the AADs dissociation kinetics longing atrial APD. This effect that enhances the use-dependent
the more selective the atrial effect during AF and the fewer the INa block by further reducing the diastolic interval at fast rates.8,9
ventricular proarrhythmic effects, because the fast dissociation
kinetics allows recovery from block at slower frequencies (e.g.,
those of the ventricles either at sinus rhythm [SR] or during AF). Late Na+ Current (INaL) Inhibition
Moreover, atrial cells exhibit a slightly more depolarized resting
membrane potential (RMP) and a more gradual phase 3 of the Na+ channels open for a few milliseconds after membrane depo-
AP, which at rapid atrial rates results in a less negative take-off larization, generating peak INa, and then most of them undergo
potential.7,8 In addition, atrial cells exhibit a more negative poten- fast inactivation. However, some Na+ channels do not inactivate
tial for half-maximum inactivation of Na+ channels than ventricu- or inactivate but reopen during depolarization generating the late
lar myocytes. Because reactivation depends on membrane Na+ current (INaL), which presents slow inactivation and recovery
potential, fewer Na+ channels recover from the inactivated state kinetics.10 INaL increases under pathologic conditions, including
during diastole in atria than in ventricles. Therefore, Na+ channel myocardial ischemia, LV hypertrophy, HF, AF, and some variants
blockers that preferentially bind to the inactivated state of the of the long QT syndrome (LQTS).10,11 In the ventricular wall, M
channel and have a fast dissociation rate will exhibit atrial selec- cells present a larger INaL than do epicardial and endocardial cells,
tivity, because steady-state drug binding and consequently and an increase in INaL increases transmural dispersion of repo-
channel block will be larger in atria than in ventricles.8 However, larization (TDR) facilitating reentry.11 During myocardial ische
block of K+ channels, particularly those preferentially or exclu- mia, enhanced INaL increases Na+ influx and, via the Na+/Ca2+
sively present in the atria, seems to be also required to meet the exchanger (NCX), produces a Na+-mediated Ca2+ overload that
“class I atrial selective” profile. Otherwise lidocaine, which is prolongs ventricular APD. APD prolongation enhances triggered
highly selective for the inactivated state and exhibits a fast onset activity that can occur during phases 2 or 3 of the AP (early
and offset kinetics, would be also atrial selective. afterdepolarizations [EADs]), whereas intracellular Ca2+ (Cai2+)
Indeed, class IA and IC, like class III AADs, also block several increase enhances triggered activity after AP repolarization
K+ channels and prolong ventricular (quinidine) or atrial (propafe- (delayed afterdepolarizations [DADs]).10 Ranolazine is an anti-
none and flecainide) APD, depending on the type of K+ channel anginal drug that, besides its effects on peak INa at fast rates,
they preferentially block (discussed later). Furthermore, amioda- blocks INaL and suppresses atrial and ventricular arrhythmias in a
rone, dronedarone, ranolazine, and vernakalant, which produce variety of conditions (e.g., ischemia, reperfusion, HF, AF, type 3
an atrial-selective INa block and in animal models suppress AF LQTS).10,11 During myocardial ischemia, ranolazine reduces
Table 54-2. Drug Selection Based on the Mechanisms Underlying Cardiac Arrhythmias
Arrhyhtmia Mechanism Target or Action Drugs
Enhanced automaticity If If blockers: adenosine, β-blockers, ivabradine

ICaL β-Blockers, diltiazem, verapamil
INCX INCX blockers, amiodarone, dronedarone
INa Class IA and IC drugs
Abnormal automaticity ICaL β-Blockers, diltiazem, verapamil
Early afterdepolarizations ICaL β-Blockers, ICaL blockers
INaL INaL blockers: amiodarone, flecainide, ranolazine
INCX INCX blockers
Shorten the APD IKr, IKs, and IKATP (nicorandil) agonists
INa Class IA and IC
Bradicardia-dependent EADs Isoproterenol
Delayed afterdepolarizations Na+-dependent Ca2+ overload INa (class I drugs) and INaL blockers
Ca2+-dependent Ca2+ overload Verapamil, RyR2 stabilizers
INCX INCX blockers
Increase threshold potential Class I drugs
Reentry Prolong APD, refractoriness Class I and III drugs
Improve conduction velocity Gap junction enhancers: rotigaptide, PUFAs
Reduce fibrosis
Atrial selective targets IKur IKur blockers: BMS-394136, XEN-D0103
IKACh Disopyramide, NTC-801
IKAChC Amiodarone, flecainide
Atrial INa Amiodarone, dronedarone, ranolazine, vernakalant
INaL Amiodarone, ranolazine
Paroxysmal supraventricular tachycardia Adenosine, β-blockers, diltiazem, verapamil (class I
and III AADs)
Abnormalities of Ca2+ handling RyR2 β-Blockers, rycals, flecainide, propafenone, dantrolene,
edavarone
CaMKII CaMKII inhibitors
Idiopathic VF Outflow tract arrhythmia due to Adenosine, β-blockers, ICa,L blockers
cAMP-mediated DADs
Left ventricular outflow tract due to His Verapamil, β-blockers, ICa,L blockers
bundle reentry
Remodeling-induced arrhythmias (upstream Fibrosis β-Blockers, RAAS inhibitors, PUFAs, statins
therapies) TGF-β1 and CTGF inhibitors
Inflammation Corticosteroids, PUFAs, statins
Oxidative stress Carvedilol, PUFAs, RAAS inhibitors, statins
Stretch Stretch-activated channel blockers, PUFAs
Ischemia-induced arrhythmias Antiischemic drugs β-Blockers, diltiazem, ivabradine, ranolazine, statins,
verapamil
IKATP IKATP blockers
APD, Action potential duration; EAD, early afterdepolarization; PUFA, polyunsaturated fatty acid; AAD, antiarrhythmic drug; DAD, delayed afterdepolarization;
VF, ventricular fibrillation; RAAS, renin-angiotensin-aldosterone system; CTGF, connective tissue growth factor; TGF, transforming growth factor.
intracellular Na+ and Ca2+ overload and improves mechanical Antiarrhythmic Effects of Class I AADs
dysfunction, but has no effect on INa, excitability or conduction
velocity.11 In non−ST-elevation acute coronary syndromes12 Considering their mechanisms of action, class I AADs are able to
ranolazine reduces non-sustained VT and supraventricular tachy- terminate arrhythmias generated by both focal ectopic activity
cardias, and preliminary evidence suggests that it safely converts and reentry. Class I AADs slow or suppress automatic activity in
paroxysmal AF and prevents AF recurrences.11 Moreover, rano- atrial and ventricular ectopic pacemaker cells that generate Na+-
lazine prevents ventricular APD (QT) prolongation, reduces dependent APs, because they slow the spontaneous diastolic
TDR, and suppresses EADs produced by IKr inhibitors and in depolarization and shift the threshold voltage to less negative
patients with type 3 LQTS. potentials. They can also suppress DADs by decreasing Na+-
dependent Cai2+ overload (INa [class I drugs] and INaL [amioda- of these arrhythmias. Moreover, β-blockers decrease the inci-
54
rone, flecainide, ranolazine] inhibitors). dence of postoperative AF, which is explained by the role of an
Describing the ultimate mechanisms responsible for the anti- increased adrenergic tone in the genesis of this arrhythmia.
arrhythmic effects of class I AADs in reentry is a big challenge. By inhibiting If and ICaL, β-blockers decrease spontaneous
Indeed, the underlying mechanism responsible for reentry itself activity of the sinoatrial node (SAN), inhibit the ectopic activity
is a matter of debate. Class I AADs markedly slow conduction of the His-Purkinje system (Purkinje fibers are particularly prone
and prolong refractoriness. The leading-circle theory of reentry to develop abnormal automaticity during increased sympathetic
predicts that because Na+ channel blockers slow conduction they activation), and suppress abnormal automaticity generated in
should, if anything, favor reentry by decreasing the wavelength cardiac cells depolarized to potentials between −60 and −40 mV
(product of refractory period and conduction velocity), resulting owing to the activation of ICaL.18 In the atrioventricular node
in proarrhythmic effects. The proarrhythmia risk increases with (AVN), they decrease conduction velocity and prolong refractori-
drugs that produce marked conduction slowing in depolarized- ness. β-Blockers are widely used to treat inappropriate sinus
ischemic cardiac tissues. Experimental and clinical evidence, tachycardia and exercise-induced and supraventricular and ven-
however, shows that class I AADs can terminate reentrant tricular arrhythmias where an increase in sympathetic tone plays
arrhythmias, such as AF, without increasing the wavelength. The a role17; they are the most effective drugs for ventricular rate
most characteristic electrophysiological change that they produce control in AF and are first-choice drugs in the treatment of
before AF termination is an increase in the temporal excitable supraventricular reentrant tachycardias where the AVN forms
gap.13,14 According to the multiple wavelets theory of the origin part of the circuit.1,17 Finally, β-blockers are effective in control-
of AF, this widening will lower the chance that reentrant waves ling the proarrhythmia induced by class I AADs, probably because
encounter areas of partially refractory tissue, so that slowing of of their bradycardic effect, which decreases the rate-dependent
conduction and fractionation of wavelets will occur less fre- conduction slowing induced by Na+ channel blockers.
quently. This widening decreases the number of fibrillation waves β-Blockers decrease all the proarrhythmic actions secondary
by promoting their fusion, which increases the chance to termi- to the increase of extracellular Ca2+ influx through L-type Ca2+
nate the arrhythmia. channels, sarcoplasmic reticulum Ca2+ release, and intracellular
However, there are clinical and experimental data demonstrat- Ca2+ overload.17,18 Thus, they suppress EADs and DADs and are
ing that reentry is maintained by the periodic activity of one or first-choice drugs in drug-induced torsades de pointes (TdP) and
a small number of functional reentrant sources with a wavefront in patients with LQTS types 1 and 2 or catecholaminergic poly-
rotating around a central core (“rotor”).15,16 The waves emerging morphic VT (CPVT).19,20 Surprisingly, and even when there are
from the rotors undergo spatially distributed fragmentation and marked pharmacologic differences among β-blockers, there are
give rise to fibrillatory conduction. The size of the spiral wave is few direct comparator data regarding their efficacy.
determined by tissue excitability and refractoriness, so that the
rotor will turn faster and in a more stable position, resulting in
higher excitability and shorter refractoriness.15,16 In this context,
Na+ channel blockers terminate reentry as they (1) enlarge center K+ Channel Blockers: Class III
of rotation, so that the rotor cannot any longer be accommodated Antiarrhythmic Effects
by the substrate; (2) decrease anchoring to functional obstacles,
increasing meander and extinction at boundaries; and (3) reduce These drugs prolong cardiac APD and refractoriness effects that,
the number of daughter waves that could provide new primary under theoretical bases, can suppress reentry. Some drugs (e.g.,
rotors.15,16 dofetilide) are considered pure class III AADs because they selec-
tively block the channels that generate the IKr (named Kv11.1 or
HERG), thus prolonging both atrial and ventricular APD and
refractoriness in the absence of effects in conduction velocity.9,21
β-BLOCKERS: CLASS II AADs Experimental data demonstrated that selective IKr inhibition pro-
duces a greater prolongation of APD and refractoriness in atria
β-Blockers include a heterogeneous group of drugs that inhibit versus ventricles, at least at normal SR frequencies. Unfortu-
sympathetic effects. Antiarrhythmic effects of β-blockers are the nately, at fast driving frequencies, the relative role of IKr in atrial
result of their electrophysiological effects (β-adrenoceptor stimu- and ventricular repolarization diminishes so much that IKr inhibi-
lation modulates several ion currents, including hyperpolarization- tion is not able to prolong APD during the arrhythmia episodes.
activated inward current [If], INa, ICaL, IK1, Ito1, IKr, and IKs), This property, known as reverse use-dependence, limits the clinical
inhibition of neurohumoral activation (and perhaps of sympa- efficacy of these drugs to stop fast arrhythmias, particularly AF.
thetic hyperinnervation and sprouting), and their antiischemic Moreover, marked IKr inhibition produces excessive and inhomo-
(reduce myocardial oxygen demands, increase subendocardial geneous ventricular APD prolongation, results in TDR, and pre-
perfusion, and improve cardiac metabolism), antihypertensive, disposes to EADs, which in turn can trigger TdP. Inhibition of
and antiproliferative effects.17,18 Indeed, β-blockers exert a benefi- IKr, and hence proarrhythmic risk, increases in the presence of
cial effect on LV remodeling in patients with myocardial infarc- bradycardia, hypokalemia, HF, or other IKr-inhibiting drugs.
tion (MI) or HF. Furthermore, in the ischemic myocardium, Indeed, many other drugs, including antihistamines, antipsychot-
β-blockers decrease dispersion of ventricular repolarization and ics, antimicrobials, and diuretics, are able to markedly block Kr
increase the VF threshold. channels, thus prolonging the QT interval (drug-induced LQTS)
Sympathetic hyperactivity is the predominant change in the and increasing the incidence of TdP and SCD. It has been pro-
autonomic tone preceding malignant ventricular arrhythmias and posed that prolongation of APD produced by pure IKr inhibitors
SCD, whereas reduced vagal activity is associated with increased is characterized by a set of four disturbances of repolarization
mortality in HF. β-Blockers remain among the few AADs that including triangulation of the AP, reverse use-dependence, insta-
are effective for both primary and secondary prevention of SCD bility, and dispersion of APD whose magnitude would help to
in different clinical conditions, including acute and chronic myo- predict drug-induced proarrhythmia risk.22
cardial ischemia, congestive HF or LV dysfunction, and hyper- HERG channel blockers gain access, from the intracellular
trophic cardiomyopathy.3,17,18 Thus, most patients who have a side of the membrane, to a binding site located within the central
propensity to develop VT/VF should receive β-blockers because cavity of the channel. The inner vestibule of HERG channels is
they are a rational, mechanism-based therapy for the treatment larger than that of other K+ channels and presents two aromatic
residues on each of the four subunits assembled to compose the be used to induce or unmask ST-segment elevation in patients
conducting pore.23 These topological characteristics explain the with concealed J wave syndromes (BrS and early repolarizing
marked pharmacologic promiscuity of HERG because they allow syndromes).29 Conversely, quinidine, which is a potent Ito1 inhibi-
drugs with rather different chemical structures to establish inter- tor, reduces the magnitude of the J wave and normalizes
actions with the aromatic moieties by π electron stacking and ST-segment elevation. The efficacy of quinidine in the preven-
block K+ efflux through the pore. Finally, it is worth mentioning tion of BrS-associated arrhythmias is currently under
that besides the “pure Kr blockers” many AADs, including amio- evaluation.
darone, quinidine, propafenone, flecainide, and ranolazine, also IKur is carried by Kv1.5 channels and is detected only in the
block HERG channels at therapeutic concentrations, an effect human atria, but not in the ventricle. Therefore, IKur inhibitors
that could contribute to their antiarrhythmic properties particu- would be expected to prolong atrial, but not ventricular APD and
larly at the atrial level (i.e., amiodarone, ranolazine). However, refractoriness.30 In myocytes from patients with SR, selective IKur
unfortunately in some cases, HERG blockade can increase their inhibition shifts the plateau phase to more positive potentials at
ventricular proarrhythmic effects (quinidine). which the IKr is rapidly activated, so that mid to late repolariza-
In healthy human atrial and ventricular myocytes driven at SR tion is accelerated and APD is slightly shortened.30 This could
frequencies, IKs seems to be small, and its pharmacologic inhibi- explain why selective IKur inhibitors do not suppress or even
tion does not alter APD.24 Importantly, the role of IKs in deter- induce AF in experimental models.8 Indeed, KCNA5 (the gene
mining APD rises in prominence at increasing beating frequencies encoding Kv1.5) loss-of-function mutations can cause AF.
or under β-adrenergic stimulation, which causes channel accu- Chronic AF reduces ICaL and markedly shortens atrial APD, so
mulation in closed states near the open state. In addition, IKs helps that IKur blockers increase the plateau potential over a range
to terminate the AP when the repolarization reserve is compro- (below 0 mV) where the activation of IKr and IKs is less pro-
mised (i.e., IKr decrease).25 Moreover, recent data demonstrated nounced, and prolong atrial APD.30 Furthermore, mathematical
that chronic AF increases by 100% IKs density in myocytes from models predict that IKur or Ito1 inhibition can induce rotor termi-
both the right and left atria.26 Mathematical models suggest that nation.31 However, chronic AF also reduces IKur, particularly in
this chronic AF-induced remodeling of IKs density contributes to the right atria; this decrease suppresses the endogenous right-to-
the atrial APD and refractoriness shortening produced by the left gradient of the current.26 As a result, the relative contribution
arrhythmia, which in turn further enhances reentry, thus favoring of IKur to repolarization in chronic AF can be small, decreasing
AF maintenance. Moreover, there are data indicating that the atrial sensitivity to IKur inhibitors.30 These findings raise
stable fast reentry sources (rotors) occur with significantly higher serious doubts on the usefulness of “pure” IKur blockers, and in
rotation frequencies, lower conduction velocities, and shorter AP fact none of those currently under development have been suc-
in cells with prominent IKs.27 In addition, the frequency-dependent cessfully launched for therapeutics.32 Conversely, vernakalant,
accumulation of IKs promotes post-repolarization refractoriness which combined blockade of IKur, Ito1, IKr, and IKACh, together with
and fibrillatory conduction of waves emanating from rotors. fast blockade of inactivated Na+ channels produces a high conver-
Therefore, it seems reasonable to propose that IKs could be an sion rate to SR in patients with recent-onset AF.
interesting target for the treatment of AF. In fact, some AADs, IK1 contributes to the final phase of repolarization and stabi-
such as propafenone, quinidine, and amiodarone, inhibit IKs—an lizes the RMP of atrial and ventricular myocytes.33 IK1 increase
effect that probably contributes to their antiarrhythmic effects.21 hyperpolarizes RMP, shortens APD, reduces spontaneous activ-
Unfortunately, selective inhibition of IKs would be difficult to ity, and accelerates and stabilizes arrhythmia-maintaining rotors,
achieve, because molecular determinants for channel block are playing a role in the genesis of AF and VF.31,33,34 IK1 exhibits dif-
similar to those required for HERG blockade and would not be ferential biophysical properties in atrial and ventricular tissues.
devoid of ventricular proarrhythmic effects. Indeed, patients car- Moreover, experimental data suggest that the Kir2.1 channel is
rying mutations in KCNQ1 or KCNE1 genes, which encoded the the major isoform underlying ventricular IK1, whereas relative
α (Kv7.1) and β (minK) subunits that form the channels, present contribution of Kir2.2 and 2.3 to atrial IK1 seems to be greater.
LQTS types 1 and 5, respectively. This finding suggests that IK1 increases in left and right atrial myocytes from patients with
excessive inhibition of IKs could also lead to a marked prolonga- chronic AF, an augmentation that is greater in the left than in the
tion of ventricular APD, particularly under conditions of increased right atria in patients with paroxysmal AF. Left-to-right ventricu-
sympathetic tone that promote EADs and TdP. lar differences in IK1 have been postulated to represent a mecha-
Ito1 (4-aminopyridine sensitive component of the transient nism for the preferential stabilization of high-frequency rotors in
outward current) determines the amplitude of the phase 1 of early the LV.34 Moreover, patients carrying gain-of-function Kir2.1
repolarization, and thus the height of the AP plateau, which in mutations exhibit type 3 short QT syndrome characterized by a
turn influences the activation of other currents involved in repo- high incidence of VF and SCD. Furthermore, flecainide and
larization (mainly ICaL, IKr, and IKs).28 Variations of Ito1 strongly propafenone at therapeutically relevant concentrations selec-
influence the shape of cardiac AP, intracellular Ca2+ transients, tively increase IK1 current generated by Kir2.1 homotetramers,
and cardiac contractility. Relative Ito1 density is much higher in an effect that could contribute to their ventricular proarrhythmic
atrial, Purkinje, epicardial, and midmyocardial cells than in endo- effects. Overall, it seems that IK1 inhibitors could be an efficacious
cardial cells. Ito1 densities are also reported to be higher in right antifibrillatory strategy. Unfortunately, IK1 inhibition also results
than in left ventricular myocytes. Prominent Ito1 leads to the in proarrhythmia. In fact, patients carrying loss-of-function
“spike and dome” morphology typical of epicardial (particularly Kir2.1 mutations exhibit congenital LQTS type 7 (Andersen-
right) and some atrial AP (Ito1 is also differentially expressed Tawil syndrome) characterized by APD prolongation, diastolic
among atrial cells). HF and chronic AF reduce Ito1 density, this depolarization, and increased propensity for EAD- and DAD-
effect being more marked in the left than in the right atria. In triggered arrhythmias.35 Moreover, the development of Kir2.1-
regions of the myocardium exhibiting a prominent Ito1, marked based antiarrhythmics is hampered because Kir2.1 channels are
accentuation of the AP notch results in a transmural voltage expressed in many excitable tissues, which could lead to organ
gradient, leading to coved ST-segment elevation diagnostic of toxicity. More promising could be selective blockade of Kir2.3
Brugada syndrome (BrS).29 Because accentuation of the AP notch over Kir2.1 channels for the treatment of AF. Kir2.3 channels are
and loss of the dome could be secondary to either a decrease in indeed preferentially inhibited by several drugs taking advantage
inward currents (INa and ICaL) or an increase in outward potassium of the low affinity of Kir2.3 channels for phosphatidylinositol
currents (mainly Ito1), sodium channel blockers such as procain- 4,5-bisphosphate (PIP2), which critically determines the channel
amide, propafenone, flecainide, ajmaline, and disopyramide can function.
Acetylcholine released from vagal nerve terminals activates in SR, or who will undergo cardioversion. Dronedarone is less
54
IKACh. Because IKACh is exclusively present in the atria, it has been effective than amiodarone in decreasing AF recurrence, but it has
proposed as a target for development of atrial selective AADs. a better safety profile. However, unlike amiodarone, dronedarone
IKACh activation slows SAN pacemaker activity, produces a het- is contraindicated in patients with HF.
erogeneous APD and ERP shortening, and hyperpolarizes RMP,
increasing Na+ channel availability. These effects create a sub-
strate for reentry, accelerate and stabilize high frequency rotors,
and contribute to initiation and maintenance of AF.34 Some AADs Calcium Channel Blockers: Class IV AADs
inhibit IKACh via the blockade of muscarinic receptors (disopyra-
mide) and may be useful in vagally mediated AF. Others inhibit Nondihydropyridine Ca2+-channel blockers (e.g., diltiazem, vera-
IKACh via GTP binding proteins or the KACh channel itself pamil) inhibit the ICaL and exhibit antiischemic and antihyperten-
(propafenone, dofetilide, and amiodarone). The new multichan- sive properties. These blockers stabilize the channel in its
nel blocker dronedarone is 100-fold more potent in blocking inactivated state and prolong its reactivation, so that their effects
IKACh than amiodarone, an effect that could contribute to its increase at fast rates (use-dependent block) and at depolarized
antiarrhythmic properties. Importantly, it has been demonstrated membrane potentials (voltage-dependent block).1
that chronic AF also reduces IKACh, reducing its relative role in In cardiac tissues with RMPs positive to −60 mV (i.e., SAN,
atrial repolarization while increasing an IKACh component (IKAChC) AVN, and ischemic tissues), ICaL is responsible for AP depolariza-
with constitutive activity (active in the absence of muscarinic tion (phase 0) and conduction velocity. Class IV AADs prolong
ligands).36 Ideally, an AAD should selectively block IKAChC (an the refractory period and slow conduction through the AVN.
atrial- and disease-specific therapeutic target) without affecting The rate-dependent effect on the AVNs is the basis for their use
the IKACh that modulates SAN and AVN. IKAChC is inhibited by to terminate or prevent reentrant supraventricular tachycardias
flecainide, AVE0118, and investigational AADs, but highly selec- whose circuit involves the AVN, being an alternative to adenos-
tive IKAChC blockers are not currently available. ine, and to control the ventricular rate in patients with AF or
In up to 80% of cases, VT/VF occurs in the setting of CAD flutter. At therapeutic concentrations, diltiazem and verapamil
and HF. Attempts to prevent arrhythmic events and SCD with decrease heart rate slightly, because their direct negative chrono-
class I and III AADs in the setting of myocardial ischemia have tropic effect is partly counteracted by their peripheral vasodilator
failed. Activation of IKATP during acute myocardial ischemia effects that produce a sympathetic-mediated reflex response.
results in shortening of the APD, accumulation of extracellular However, in patients treated with β-blockers, reflex sympathetic
K+, membrane depolarization, and slow conduction velocity. stimulation is inhibited, and diltiazem and verapamil slow heart
These effects render the ischemic heart vulnerable to reentrant rate and AVN conduction.
arrhythmias. It has been hypothesized that specific IKATP inhibi- Inhibition of ICaL suppresses automaticity generated in depo-
tors prolong ventricular refractoriness in ischemic tissues with larized ventricular cells (abnormal automatism) and EADs occur-
little or no effect on normal tissue, being truly ischemia-selective ring at the plateau phase of AP because of ICaL reactivation (as
AADs.37 Conversely, the KATP opener nicorandil shortens the marked delay in repolarization allows Ca2+ channels to recover
APD, reduces TDR, and suppresses EADs in patients with from inactivation). Thus, class IV AADs and magnesium sulfate
LQT1.38 are effective for the treatment EAD-triggered arrhythmias. On
the other hand, abnormal Cai2+ overload activates forward mode
of the NCX (Ca2+ extrusion) generating a transient inward
Amiodarone and Dronedarone current (Iti) that underlies phase 3 EADs and DADs. Class IV
AADs and AADs that block Ca2+ channels at therapeutic concen-
Amiodarone is considered the epitome of class III AADs because, trations (e.g., amiodarone, dronedarone, propafenone) suppress
in chronic treatment, it prolongs APD and refractoriness in all DADs by preventing Cai2+ overload.
cardiac tissues, an effect that is attributable to the blockade of Verapamil is a drug of choice in idiopathic outflow tract and
several K+ channels (Ito1, IKur, IKr, IKs, IK1, and IKACh). However, fascicular VT and drug-induced TdP; it can be an alternative to
as was mentioned before, amiodarone also inhibits peak INa β-blockers in CPVT and prevents VTs related to ischemia.
(inhibits inactivated Na+ channels with fast kinetics [class I]), INaL, However, class IV drugs have no effect on all-cause mortality in
and ICaL (class IV), and antagonizes α-/β-adrenergic receptors patients with prior MI and are contraindicated in congestive HF.3
(class II).1
Amiodarone presents a low proarrhythmic risk and is effective
against supraventricular and ventricular arrhythmias caused by
focal activity or reentry. It is the most effective AAD for prevent- Gap-Junction Coupling Enhancers
ing recurrences of AF, being the last-choice drug in patients with
structural heart disease, and in the prophylaxis and treatment of Connexins (Cx) are membrane proteins that assemble into hexa-
life-threatening ventricular arrhythmias in patients with MI or meric hemichannels, also known as connexons. At the gap junc-
congestive HF or after cardiac surgery, although it is being tions, connexons allow electrical flow between cardiomyocytes
replaced by ICDs. Amiodarone decreases the incidence of SCD and thus regulate intercellular coupling. Changes in Cx expres-
and is an alternative for patients who are not eligible for, or who sion, location (lateralization), and function occur in many forms
do not have access to, ICD therapy for the prevention of SCD of heart disease (i.e., ischemia, hypertrophy, HF, AF) and con-
and to inhibit unpleasant ICD shocks.3,39 However, in patients tribute to arrhythmogenesis as they slow conduction velocity,
with HF and LV ejection fraction of 35% or less, amiodarone has cause heterogeneities in repolarization, and modulate automatic-
no favorable effect on survival.3 ity. Cx40 has a critical role in mediating atrial conduction,
Because amiodarone has complex pharmacokinetics, a long whereas Cx43 seems to be the major ventricular form. Cellular
half-life, and a high incidence of organ toxicity, new amiodarone- uncoupling owing to dephosphorylation and redistribution of
like agents with a better pharmacologic profile have been devel- Cx43, together with increased fibrosis40 and decreased expression
oped. Dronedarone is an amiodarone-like drug without the of Na+ channels, are implicated in conduction slowing during
iodine moieties that exhibit class I, II, III, and IV effects, a short ischemia, increasing the risk of fatal ventricular arrhythmias.
half-life, and less lipophilicity compared with amiodarone. It is Rotigaptide increases gap junction conductance and prevents
indicated in patients with paroxysmal or persistent AF, who are acute ischemic arrhythmias. However, it partially reverses the loss
of Cx43 but does not restore normal conduction or prevent could contribute to the empirically observed antiarrhythmic
arrhythmias in the healing infarct border zone, such as after a actions produced by flecainide in some patients with Andersen-
prolonged period of gap junction remodeling.41 Thus, it seems Tawil syndrome.44
that interventions that are effective in restoring cell-to-cell cou-
pling under conditions of acute ischemia might not be effective
in a substrate that has already undergone a prolonged period of
gap junction remodeling. Gap junction–enhancing drugs can also Targeting Intracellular Calcium Handling
reduce AF vulnerability in some models of AF (chronic mitral
regurgitation and acute ischemia), but not in others (HF or Abnormal intracellular Ca2+ handling can promote both atrial and
atrial tachypacing).41 These findings suggest that the role of Cx ventricular arrhythmias; it also has a key role in cardiac electrical
in AF is disease specific and that gap junction enhancers may be and structural remodeling in patients with AF, HF, LV hypertro-
effective only when conduction slowing is due to alterations in phy, and CAD.14,46-48 Moreover, CPVT associated with mutations
gap junctions. in the sarcoplasmic reticulum Ca2+ release channel/ryanodine
receptor (RyR2) and calsequestrin (CASQ2) genes destabilize the
RyR2 channel complex and increase spontaneous Ca2+ release,
which under certain conditions (e.g., during exercise or
Stretch-Activated Ion Channels β-adrenergic stimulation) facilitates the development of DADs
and triggered arrhythmias.20,46 Thus, intracellular Ca2+-handling
Chronic stretch causes heterogeneous conduction, prolongs proteins (Ca2+-ATPase [SERCA2a], phospholamban, RyR2, and
refractoriness, and induces structural remodeling (dilatation, its accessory protein calstabin 2, NCX, and CSQ), provide targets
hypertrophy, fibrosis), which contributes to both atrial and ven- for developing new AADs.
tricular arrhythmias. Recent evidence suggests that these electro- Dysregulation of RyR2, characterized by improper SR Ca2+
physiological effects can be explained via the activation of the release and increased diastolic Ca2+ leak is related to: (1) RyR2
nonselective cationic transient receptor potential TRPC6 chan- hyperphosphorylation by protein kinase A (PKA) and Ca2+/
nels.42 In experimental models, nonselective stretch-activated calmodulin-dependent kinase II (CaMKII), which transiently dis-
channel blockers (gadolinium and the tarantula peptide GsMtx- sociates calstabin 2 from the RyR2 complex increasing open
4) suppress AF inducibility.14 Nowadays, modulation of stretch- channel probability; (2) increased sensitivity of RyR2 to activa-
activated channels as an antiarrhythmic strategy is limited by the tion by luminal Ca2+; and (3) reactive oxygen species generated
lack of effective drugs. in the diseased heart, which make the channel leaky.14,46-48
β-Adrenergic stimulation and renin-angiotensin-aldosterone
system (RAAS) can promote triggered arrhythmias via PKA- and
CaMKII-induced RyR2 hyperphosphorylation.46,47 β-Blockers
Modulation of Ion Channel Trafficking prevent PKA- and CaMKII-mediated hyperphosphorylation of
RyR2, depletion of calstabin 2, and SR Ca2+ leak being the only
Structural heart diseases and many mutations associated with effective drugs in improving survival in CPVT.20 β-Blockers and
channelopathies affect biogenesis, forward trafficking to the verapamil act synergistically to prevent stress-induced increase in
surface membrane subdomains, and degradation of cardiac sarcoplasmic reticulum Ca2+ content by reducing heart rate, Ca2+
ion channels.43 Recent evidence demonstrates that AADs influx into the cell, and Ca2+ uptake into the sarcoplasmic reticu-
inhibit ion conduction and modulate ion channel trafficking or lum.20 However, in patients with CPVT, cardiac events remain
surface density and that channel blockade and trafficking can be considerable even in this combination, so that alternative thera-
targeted independently.43,44 Some HERG channel blockers pies are needed.
(cisapride, E4031, quinidine) can rescue defective HERG Carvedilol, a nonselective β- and α1-adrenergic antagonist,
surface trafficking, whereas others (ketoconazole and fluoxetine) blocks several ion channels (IKur, IKr, ICaL, Ito1, and IKs) and exerts
can decrease HERG surface density. Furthermore, there antioxidant and antiproliferative effects. It is the only β-blocker
are drugs that do not block HERG channel conductance, that suppresses sarcoplasmic reticulum Ca2+ release by directly
but either decrease (pentamidine, probucol) or increase (fexofe- reducing RyR2 mean open time, independently of its β- or α1-
nadine) HERG surface density.45 Moreover, type 2 LQTS- antagonism or antioxidant activities, which likely contributes to
associated mutant channels can be rescued to the plasma its antiarrhythmic effects in HF patients.20,46 Carvedilol analogues
membrane with drugs (E4031, terfenadine, fexofenadine) that with minimal β-blocking activity retain the ability to suppress
likely stabilize misfolded protein in the endoplasmic reticulum, Ca2+ release and prevent CPVT without causing bradycardia.
although the drug effectiveness depends on the particular muta- Interestingly, the combination of these analogues with selective
tion present.45 β-blockers (metoprolol or bisoprolol) is more effective than each
Quinidine induces a stereospecific dose-, time- and subunit- agent alone for preventing CPVT.
dependent internalization of Kv1.5, concomitant with the classi- Stabilizers of the calstabin 2–RyR2 complex (Rycals) represent
cal pore block.43 The specific expression of Kv1.5 channels in a new antiarrhythmic approach. Rycals (such as JTV519 and
human atria raises the possibility of designing drugs to specifi- S44121) increase the affinity of calstabin-2 for PKA-
cally modulate Kv1.5 channel trafficking.43 Such drugs, in theory, hyperphosphorylated RyR2, reduce channel open probability and
can cause a rapid and reversible decrease in IKur and a prolonga- prevent arrhythmogenic diastolic Ca2+ leak without affecting
tion of atrial APD terminating AF, with minimal risk of ventricu- normal RyR2.14,20,46,48 As a consequence, JTV519 prevents APD
lar proarrhythmia. However, there is a need for a better alternans and triggered ventricular arrhythmias and reduces Ca2+
understanding of trafficking mechanisms and how chronic waves in arrhythmogenic Purkinje fibers following MI. It also
decrease of channel expression contributes to arrhythmia-induced reduces ectopic activity and DADs in pulmonary vein cardiomyo-
remodeling. cytes, decreasing AF inducibility.47 Its putative efficacy in the
Some drugs such as pentamidine and flecainide decrease and treatment of AF and ventricular arrhythmias associated with HF
increase Kir2.1 channel expression in the membrane, respec- and CPVT is currently being tested.
tively. Flecainide is also able to promote trafficking of some Flecainide and propafenone block the open state of RyR2,
Kir2.1 loss-of-function mutants increasing the amplitude of the inhibit arrhythmogenic diastolic Ca2+ waves, and suppress DADs
outward K+ current generated by the channels, if any. This effect and triggered arrhythmias in experimental models and in patients
with CPVT.20,47 In addition, their Na+ channel–blocking proper- possibility for antiarrhythmic therapies based on genotype and
54
ties prevent DADs from reaching threshold potential and from clinical presentation. However, randomized trials have not been
triggering premature beats. Furthermore, β-blocking activity of conducted because of the rarity of these conditions. Nevertheless,
propafenone can also contribute to its clinical efficacy in CPVT. therapy alleviating the mutation consequences depends on the
As a result, class IC AADs can be an alternative to β-blockers for mutation class, although most AADs lack the desired selectivity
some patients with CPVT.20,47 for a given channel.
Some RyR2 mutations can lead to an abnormally tight
domain-domain interaction that results in an erroneous activa-
tion of the channel and diastolic Ca2+ leak.46 In experimental
models, dantrolene stabilizes domain interactions within the Target Cardiac Remodeling:
RyR2 and prevents aberrant Ca2+ release and CPVT.20 The anti- Upstream Therapies
oxidant edaravone also ameliorates the defective interdomain
interaction of the RyR2 and prevents Ca2+ leak and LV remodel- Cardiac arrhythmias produce changes (remodeling) in the
ing during the development of HF.48 electrophysiological (ion channel and gap junction function
Even when strong evidence suggests that CaMKII control or expression, abnormal Ca2+ handling) and structural (hypertro-
over Ca2+ handling might be more potent than that of PKA, phy, fibrosis, dilatation) properties of the myocardium often
CaMKII inhibitors might not represent a feasible target, because associated with stretch, oxidative stress, inflammation, or isch-
the ubiquitous role of CaMKII in cellular physiology can result emia, creating a substrate for reentry and focal activity.13,14
in deleterious off-target side effects.48 In addition, some CaMKII Furthermore, aging and structural heart diseases also produce
inhibitors (KN-93) cannot discriminate between CaMKII electrophysiological and structural remodeling that increases
and CaMKIV and inhibit voltage-gated K+ and Ca2+ channels. susceptibility to arrhythmias and makes arrhythmias more drug
Therefore, to achieve an effective and safe cardiac effect, drugs resistant.14,49 However, remodeling is a disease-specific process
with tissue and isoform-specific action (CaMKIIδ) are probably presenting important differences in signaling pathways and phar-
required. An alternative strategy is small peptides like autocam- macologic responses among comorbidities. Upstream therapies to
tide-2-related inhibitory peptide (AIP) that inhibit CaMKII prevent or delay myocardial remodeling are an attractive approach
activation by blocking calmodulin binding to CaMKII. in the prophylaxis and treatment of arrhythmias associated with
HF, hypertension, or MI.13,14,48,50,51 Potential drugs in this category
include antiinflammatory agents, RAAS inhibitors (angiotensin-
converting enzyme inhibitors [ACEIs], angiotensin II receptor
Pharmacologic Treatment of Inherited antagonists [ARAs], aldosterone antagonists), omega-3 polyun-
Cardiac Arrhythmia Syndromes saturated fatty acids (PUFAs), statins, and antifibrotic agents.
During the past decade, there was an explosion of information

linking mutations in genes encoding ion channel α- and accessory- Antiinflammatory Agents
subunits and cytoskeletal molecules to inherited cardiac arrhyth-
mia syndromes, including short QT syndrome, LQTS, BrS, AF, Raised inflammatory markers (interleukin 6 and high-sensitivity
idiopathic VF, and CPVT (Table 54-3).19,29,35,38,45 Identification of C-reactive protein) are associated with a higher risk of atrial and
disease-associated genes provides important information on the ventricular arrhythmias in patients with HF or CAD.13,14,48 It has
molecular mechanisms of cardiac arrhythmias and opens the been hypothesized that inflammation contributes to the structural
Table 54-3. Pharmacologic Treatment for Inherited Arrhythmia Syndromes
Syndrome Genetic Disorder Alteration Specific Drug Therapy
Long QT syndrome LQTS1, LQTS5, JLN1/2 ↓ IKs β-Blockers, K+ channel openers (nicorandil, L-364,373)
LQTS2/LQTS6 ↓ IKr β-Blockers, K-sparing agents, K+ channel openers
LQTS3, LQTS9, LQTS10, LQTS12 ↑ INa Flecainide, mexiletine
INaL inhibitors: ranolazine
Prevention of bradycardia
LQTS7 ↓ IK1 Flecainide
LQTS8 ↑ ICaL β-Blockers, diltiazem, ranolazine, verapamil
Short QT syndrome ↑ IKr, IKs, IK1, IKATP Quinidine, disopyramide, amiodarone
↓ INa, ICaL Blockers of affected K+ channels
CPVT CPVT1, CPVT2 ↑ SR Ca2+ leak β-blockers, verapamil, RyR2 stabilizers (rycals), dantrolene,
flecainide, propafenone
J wave syndromes Brugada syndrome ↓ INa, ↓ ICaL Ito blockers, quinidine
↑ Ito, IKATP Increase ICaL: cilostazol, denopamine, isoproterenol
Early repolarization syndromes ↑ IKAch, IKATP
↓ INa, ICaL
Atrial fibrillation ↑ IKs, IKr, IKur, Ito, IK1, IKATP, INa Blockers of affected K+ channels
↓ ICaL, ↓ Cx40
CPVT, Catecholaminergic polymorphic ventricular tachycardia; SR, sarcoplasmic reticulum; ↑, inerease; ↓, decrease.
remodeling and might be a target for antiarrhythmic therapies. substrate in fibrotic hearts. Cardiac fibrosis is the final result of
However, whether inflammation is an initiating event or a conse- multiple signaling pathways that can vary in different cardiac dis-
quence in the development of arrhythmias remains controversial. eases, and many potential therapeutic antifibrotic targets have
Corticosteroids reduce the risk of postoperative AF, probably been identified, although its clinical relevance remains uncertain.
because inflammation has an important role in its pathogenesis,51
but their potential toxicity restricts their therapeutic value.
Conclusions
RAAS Inhibitors Most positive data with upstream therapies came from experi-
mental models, observational studies, and retrospective analyses
ACEIs and ARAs inhibit angiotensin II AT1 receptor–mediated of clinical data, whereas prospective randomized trials failed to
proarrhythmic effects including APD shortening, enhanced auto- demonstrate a beneficial effect on secondary prevention of AF
maticity, hypokalemia, reduced cell coupling, pressure overload, burden or cardiovascular outcomes. This can be explained because
abnormal Ca2+ handling, oxidative stress, neurohumoral activa- upstream therapies are effective at the beginning of the remodel-
tion, and structural remodeling (hypertrophy, fibrosis, ing process, but ineffective at later stages, when structural remod-
inflammation).5,14,52 eling is irreversible, or because drugs are only effective in patients
In primary prevention, ACEIs and ARAs reduce the risk of with certain comorbidities. Therefore, prospective studies are
AF in patients with HF, hypertension, and LV hypertrophy, but needed to confirm this hypothesis and to define the populations
not in patients with post–myocardial infarction.53 In secondary of patients that most likely benefit from upstream therapies.
prevention, they exert a beneficial effect after cardioversion of
persistent AF and in the prevention of paroxysmal AF, particu-
larly in patients with significant underlying heart disease (e.g., LV
dysfunction or hypertrophy) and when patients also received Other AADs
amiodarone.50,53 However, prospective randomized clinical trials
question the value of ARAs in secondary prevention of AF.50 Adenosine acting on cardiac A1 receptors activates an outward
ACEIs decrease the risk of death following a recent MI by reduc- K+ current (IKAdo) present in the atria, SAN, and AVN and inhibits
ing cardiovascular mortality. In patients with congestive HF, If and adenylyl cyclase, which indirectly results in a decrease in
ACEIs reduce all-cause mortality but not SCD.3 catecholamine-stimulated ICaL and ITI. As a result, adenosine
Aldosterone antagonists inhibit cardiac fibrosis, exert direct suppresses the pacemaker activity of SAN and depresses
antiarrhythmic actions, and reduce the risk of hypokalemia; AVN conduction, being the first-line agent for terminating par-
therefore, they represent a therapeutic option for patients with oxysmal paroxysmal supraventricular tachycardias in which the
HF, LV hypertrophy, and AF. Indeed, preliminary data suggest AV node is part of the reentry pathway. In addition, adenosine
that aldosterone inhibitors reduce the incidence of recent-onset suppresses catecholamine-stimulated EADs and DADs and
AF, VT/VF, and SCD in patients with congestive HF.50 catecholamine-mediated atrial and ventricular arrhythmias.
Digoxin inhibits Na+-K+–ATPase and decreases intracellular
Na+ concentration, which in turn, increases via the NCX both
Statins Cai2+ and contractile force. In addition, digoxin enhances vagal
tone and reduces sympathetic and RAAS tone. At therapeutic
They exert multiple antiarrhythmic effects, including direct concentrations, digoxin decreases the automaticity of the SAN
effects on cardiac ion channels, plaque stabilization, and (vagal stimulation inhibits If) and slows conduction and prolongs
antifibrotic, antiinflammatory, and antioxidant actions.54,55 In refractoriness in the AVN. This latter effect is the basis for heart
addition, they suppress triggered activity in canine pulmonary rate control in patients with AF and systolic HF, particularly in
vein preparations. Studies of statins for primary or secondary unstabilized HF patients in whom β-blockers and calcium antag-
prevention of AF do not support specific recommendations, onists are contraindicated. In permanent AF, the combination of
except perhaps for primary prevention of postoperative AF.51 digoxin with these drugs provides a satisfactory rate control both
Statins do not reduce the risk of ventricular arrhythmias or of at rest and during exercise. Digoxin has no role in ventricular
cardiac arrest and produce a modest benefit on SCD in patients arrhythmias, does not modify cardiac mortality in patients with
with CAD treated with an ICD and in patients with nonischemic congestive HF in SR, and can increase the rate of SCD.
cardiomyopathy.3,56 Raised resting heart rate is a strong independent risk factor
for cardiovascular events. The antianginal drug ivabradine, a
selective If inhibitor, added on background therapy reduces the
Omega-3 Polyunsaturated Fatty Acids risk of cardiovascular death or hospitalization for worsening HF
in patients with symptomatic HF and an LV ejection fraction of
They modulate ion channels and connexins, reduce fluctuations 35% or less with heart rate 70 beats/minute or greater.60 In addi-
in Ca2+i, and exert antiinflammatory and antioxidant actions.57 tion, ivabradine reverses LV remodeling, suggesting that it pre-
However, there is no robust evidence to support their efficacy for vents disease progression in these patients. Preliminary data
primary or secondary prevention of AF or VT/VF.3,51,57 suggest that ivabradine is a promising option in patients with
inappropriate sinus tachycardia.
Antifibrotic Agents
A variety of stimuli (i.e., fast heart rates, pressure/volume overload, Unresolved Questions and Future Strategies
oxidative stress, inflammation, ischemia, cytokines, growth factors,
neurohumoral activation) induce the proliferation of cardiac fibro- There are many unresolved questions concerning the pharmaco-
blasts and its differentiation into myofibroblasts.58,59 Myofibro- logic treatment of arrhythmias. For example, it is still unclear
blasts when electrotonically coupled to cardiomyocytes cause slow whether targeting an individual ion channel is better than target-
conduction, ectopic activity, and electric remodeling based on ing multiple ion channels. By blocking both inward and outward
paracrine interactions, which generate an arrhythmogenic currents, multichannel blockers can create steady-state
conditions that avoid large variations in AP repolarization, therapeutic target, including two-pore–domain K+ channels,
54
thereby preventing the development of electrical instability, Ca2+-activated nonselective cation channels, small-conductance
although the risk of proarrhythmia may increase. Ca2+-activated K+ channels (SK1-3 channels are expressed in the
As mentioned, VT/VF associated with SCD preferentially atria), and transient receptor potential channels of the canonical
occurs in the setting of myocardial ischemia, and it would be (TRPC1/3/6) and melastin-related (TRPM3/4/6/7) subfami-
optimal to develop ischemia-selective AADs. Because class I lies.9,14 However, because drugs used to target these currents lack
and III AADs have failed to prevent ischemia-induced arrhyth- specificity, the functional role of these channels in humans
mias, antiarrhythmic strategy has been shifted to drugs com- remains uncertain. In the same line, abnormal Ca2+ handling
monly prescribed in patients with MI or HF that prevent plaque plays a major role in arrhythmogenesis, but the developed drugs
rupture and reduce myocardial ischemia, such as antianginal do not have sufficient specificity.
drugs (β-blockers, ivabradine, ranolazine), statins, RAAS inhibi-
tors, and platelet antiaggregants, in an attempt to eliminate
potential triggers for ventricular arrhythmias and, in the long-
term, prevent the formation of myocardial scars and SCD. RAAS Conclusions
inhibitors and β-blockers might also limit the substrate for ven-
tricular arrhythmias because they prevent LV remodeling after AADs remain the mainstay of therapy for the majority of patients
MI.3 with cardiac arrhythmias, but there is an unmet need for new
Another unresolved question is the duration and timing of drugs that achieve effective suppression and prevention of life-
AAD therapy. Chronic antiarrhythmic therapy increases the risk threatening arrhythmias without the risk of proarrhythmia. The
of proarrhythmia, and a suitable timing of antiarrhythmic drug rational development of these new AADs should be the final
therapy might enhance safety without reducing efficacy. It is result of a better understanding of the molecular and cellular
unlikely that patients with infrequent, brief, and mild symptom- mechanisms involved in the genesis and maintenance of cardiac
atic arrhythmic episodes should receive long-term AAD therapy. arrhythmias in different pathologic substrates and the mecha-
As an alternative strategy, a single oral bolus dose of propafenone nisms of action of AADs.
or flecainide (“pill in the pocket”) can be administered in selected
patients to terminate recent-onset AF outside the hospital once
treatment has proved safe in the hospital. However, few prospec-
tive data are available on the relative safety of this approach, and Acknowledgments
the decision to initiate therapy outside the hospital should be
carefully individualized. The authors thank Ricardo Caballero for comment on the manu-
Because AF is the most common arrhythmia, there are strate- script and Paloma Vaquero for editorial assistance in manuscript
gies for the development of atrial-selective AADs devoid of ven- preparation. This work was supported by grants from the Mi
tricular proarrhythmic effects based on the identification of nisterio de Ciencia e Innovación (SAF2011-30088), Fondo de
atrial-specific targets. There are preliminary experimental data Investigaciones Sanitarias (PI11/01030), and Red Heracles
suggesting that several channels represent a possible atrial (RD12/0042/0011).
7. Hatem SN, Coulombe A, Balse E: Specificities of 15. Vaquero M, Calvo D, Jalife J: Cardiac fibrillation:
References atrial electrophysiology: Clues to a better under- from ion channels to rotors in the human heart.
standing of cardiac function and the mechanisms Heart Rhythm 5(6):872–879, 2008.
1. Darbar D: Standard antiarrhythmic drugs. In Zipes of arrhythmias. J Mol Cell Cardiol 48(1):90–95, 16. Comtois P, Sakabe M, Vigmond EJ, et al: Mecha-
D, Jalife J, editors: Cardiac electrophysiology: 2010. nisms of atrial fibrillation termination by rapidly
From cell to bedside, ed 5, Philadelphia, 2009, 8. Burashnikov A, Antzelevitch C: Novel pharmaco- unbinding Na+ channel blockers: insights from
Saunders Elsevier. pp 959–973. logical targets for the rhythm control management mathematical models and experimental correlates.
2. Sheets MF, Fozzard HA, Lipkind GM, et al: of atrial fibrillation. Pharmacol Ther 132(3):300– Am J Physiol Heart Circ Physiol 295(4):H1489–
Sodium channel molecular conformations and 313, 2011. H1504, 2008.
antiarrhythmic drug affinity. Trends Cardiovasc 9. Ravens, U: Antiarrhythmic therapy in atrial fibril- 17. López-Sendón J, Swedberg K, McMurray J, et al:
Med 20(1):16–21, 2010. lation. Pharmacol Ther 128(1):129–145, 2010. Expert consensus document on beta-adrenergic
3. Das MK, Zipes DP: Antiarrhythmic and nonanti- 10. Zaza A, Belardinelli L, Shryock JC: Pathophysiol- receptor blockers. Eur Heart J 25(15):1341–1362,
arrhythmic drugs for sudden cardiac death preven- ogy and pharmacology of the cardiac “late sodium 2004.
tion. J Cardiovasc Pharmacol 55(5):438–449, current.” Pharmacol Ther 119(3):326–339, 18. Workman AJ: Cardiac adrenergic control and atrial
2010. 2008. fibrillation. Naunyn Schmiedeberg Arch Pharma-
4. Zipes DP, Camm AJ, Borggrefe M, et al: ACC/ 11. Antzelevitch C, Burashnikov A, Sicouri S, et al: col 381(3):235–249, 2010.
AHA/ESC 2006 guidelines for management of Electrophysiologic basis for the antiarrhythmic 19. Cerrone M, Priori SG: Genetics of sudden death:
patients with ventricular arrhythmias and the pre- actions of ranolazine. Heart Rhythm 8(8):1281– focus on inherited channelopathies. Eur Heart J
vention of sudden cardiac death. Circulation 1290, 2011. 32(17):2109–2118, 2011.
114(10):E385–E484, 2006. 12. Scirica BM, Morrow DA, Hod H, et al: Effect of 20. Werf Cv, Zwinderman AH, Wilde AA: Therapeu-
5. Camm AJ, Kirchhof P, Lip GY, et al: Guidelines ranolazine, an antianginal agent with novel electro- tic approach for patients with catecholaminergic
for the management of atrial fibrillation: the Task physiological properties, on the incidence of polymorphic ventricular tachycardia: state of the
Force for the Management of Atrial Fibrillation of arrhythmias in patients with non ST-segment ele- art and future developments. Europace 14(2):175–
the European Society of Cardiology (ESC). Eur vation acute coronary syndrome: results from the 183, 2012.
Heart J 31(19):2369–2429, 2010. Metabolic Efficiency With Ranolazine for Less 21. Tamargo J, Caballero R, Gómez R, et al: Pharma-
6. Fuster V, Rydén LE, Cannom DS, et al: 2011 Ischemia in Non ST-Elevation Acute Coronary cology of cardiac potassium channels. Cardiovasc
ACCF/AHA/HRS Focused Updates Incorporated Syndrome Thrombolysis in Myocardial Infarction Res 62(1):9–33, 2004.
Into the ACC/AHA/ESC 2006 Guidelines for the 36 (MERLIN-TIMI 36) randomized controlled 22. Shah RR, Hondeghem LM: Refining detection of
Management of Patients With Atrial Fibrillation: trial. Circulation 116(15):1647–1652, 2007. drug-induced proarrhythmia: QT interval and
a Report of the American College of Cardiology 13. Schotten U, Verheule S, Kirchhof P, et al: Patho- TRIaD. Heart Rhythm 2(7):758–772, 2005.
Foundation/American Heart Association Task physiological mechanisms of atrial fibrillation: a 23. Perry M, Sanguinetti M, Mitcheson J: Revealing
Force on Practice Guidelines Developed in part- translational appraisal. Physiol Rev 91(1):265–325, the structural basis of action of hERG potassium
nership with the European Society of Cardiology 2011. channel activators and blockers. J Physiol
and in collaboration with the European Heart 14. Wakili R, Voigt N, Kääb S, et al: Recent 588(17):3157–3167, 2010.
Rhythm Association and the Heart Rhythm advances in the molecular pathophysiology of atrial 24. Charpentier F, Mérot J, Loussouarn G, et al:
Society. J Am Coll Cardiol 57(11):e101–e198, fibrillation. J Clin Invest 121(8):2955–2968, Delayed rectifier K+ currents and cardiac repolar-
2011. 2011. ization. J Mol Cell Cardiol 48(1):37–44, 2010.
25. Varro A, Baczkó I: Cardiac repolarization reserve: anti-arrhythmic therapy. Pharmacol Ther 50. Savelieva I, Kakouros N, Kourliouros A, et al:
a principle for understanding drug-related proar- 120(1):54–70, 2008. Upstream therapies for management of atrial fibril-
rhythmic risk. Br J Pharmacol 164(1):14–36, 38. Terzic A, Alekseev AE, Yamada S, et al: Advances lation: review of clinical evidence and implications
2011. in cardiac ATP-sensitive K+ channelopathies from for European Society of Cardiology guidelines.
26. Caballero R, de la Fuente MG, Gómez R, et al: In molecules to populations. Circ Arrhythm Electro- Part I: primary prevention. Europace 13(3):308–
humans, chronic atrial fibrillation decreases the physiol 4(4):577–585, 2011. 328, 2011.
transient outward current and ultrarapid compo- 39. Piccini JP, Berger JS, O’Connor CM: Amiodarone 51. Savelieva I, Kakouros N, Kourliouros A, et al:
nent of the delayed rectifier current differentially for the prevention of sudden cardiac death: a meta- Upstream therapies for management of atrial fibril-
on each atria and increases the slow component of analysis of randomized controlled trials. Eur Heart lation: review of clinical evidence and implications
the delayed rectifier current in both. J Am Coll J 30(10):1245–1253, 2009. for European Society of Cardiology guidelines.
Cardiol 55(21):2346–2354, 2010. 40. Severs NJ, Bruce AF, Dupont E, et al: Remodelling Part II: secondary prevention. Europace 13(5):610–
27. Muñoz V, Grzeda KR, Desplantez T, et al: Adeno- of gap junctions and connexin expression in dis- 625, 2011.
viral expression of IKs contributes to wavebreak and eased myocardium. Cardiovasc Res 80(1):9–19, 52. Goette A, Bukowska A, Lendeckel U: Non-ion
fibrillatory conduction in neonatal rat ventricular 2008. channel blockers as anti-arrhythmic drugs (reversal
cardiomyocyte monolayers. Circ Res 101(5):475– 41. Dhein S, Hagen A, Jozwiak J, et al: Improving of structural remodeling). Curr Opin Pharmacol
483, 2007. cardiac gap junction communication as a new anti- 7:219–224, 2007.
28. Niwa N, Nerbonne JM: Molecular determinants of arrhythmic mechanism: the action of antiarrhyth- 53. Schneider MP, Hua TA, Bohm M, et al: Prevention
cardiac transient outward potassium current (Ito) mic peptides. Naunyn Schmiedebergs Arch of atrial fibrillation by renin-angiotensin system
expression and regulation. J Mol Cell Cardiol Pharmacol 381(3):221–234, 2010. inhibition: a meta-analysis. J Am Coll Cardiol
48(1):12–25, 2010. 42. Dyachenko V, Husse B, Rueckschloss U, et al: 55(21):2299–2307, 2010.
29. Antzelevitch C, Yan GX: J wave syndromes. Heart Mechanical deformation of ventricular myocytes 54. Tamargo J, Caballero R, Gómez R, et al: Lipid-
Rhythm 7(4):549–558, 2010. modulates both TRPC6 and Kir2.3 channels. Cell lowering therapy with statins, a new approach to
30. Ravens U, Wettwer E: Ultra-rapid delayed rectifier Calcium 45(1):38–54, 2009. antiarrhythmic therapy. Pharmacol Ther 114(1):
channels: molecular basis and therapeutic implica- 43. Schumacher SM, Martens JR: Ion channel traffick- 107–126, 2007.
tions. Cardiovasc Res 89(4):776–785, 2011. ing: a new therapeutic horizon for atrial fibrilla- 55. Thomas G, Lerman BB: Expanding the role of
31. Pandit SV, Berenfeld O, Anumonwo JM, et al: tion. Heart Rhythm 7(9):1309–1315, 2010. statins in postoperative atrial fibrillation. Heart
Ionic determinants of functional reentry in a 2-D 44. Caballero R, Dolz-Gaitón P, Gómez R, et al: Fle- Rhythm 9(2):170–171, 2012.
model of human atrial cells during simulated cainide increases Kir2.1 currents by interacting 56. Rahimi K, Majoni W, Merhi A, et al: Effect of
chronic atrial fibrillation. Biophys J 88(6):3806– with cysteine 311, decreasing the polyamine- statins on ventricular tachyarrhythmia, cardiac
3821, 2005. induced rectification. Proc Natl Acad Sci U S A arrest, and sudden cardiac death: a meta-analysis of
32. Tamargo J, Caballero R, Gómez R, et al: IKur/Kv1.5 107(35):15631–15636, 2010. published and unpublished evidence from random-
channel blockers for the treatment of atrial fibril- 45. Sanguinetti MC: HERG1 channelopathies. ized trials. Eur Heart J 33(13):1571–1581, 2012.
lation. Expert Opin Investig Drugs 18(4):399–416, Pflugers Arch 460(2):265–276, 2010. 57. Mozaffarian D, Wu JH: Omega-3 fatty acids and
2009. 46. Venetucci L, Denegri M, Napolitano C, et al: cardiovascular disease: effects on risk factors,
33. Anumonwo JMB, Lopatin AN: Cardiac strong Inherited calcium channelopathies in the patho- molecular pathways, and clinical events. J Am Coll
inward rectifier potassium channels. J Mol Cell physiology of arrhythmias. Nat Rev Cardiol 2012. Cardiol 58(20):2047–2067, 2011.
Cardiol 48(1):45–54, 2010. [Epub ahead of print] 58. Rohr S: Arrhythmogenic implications of fibroblast-
34. Jalife J: Inward rectifier potassium channels control 47. Dobrev D, Voigt N, Wehrens XH: The ryanodine myocyte interactions. Circ Arrhythm Electro-
rotor frequency in ventricular fibrillation. Heart receptor channel as a molecular motif in atrial physiol 5(2):442–452, 2012.
Rhythm 6(11 Suppl):S44–S48, 2009. fibrillation: pathophysiological and therapeutic 59. de Jong S, van Veen TA, van Rijen HV, et al: Fibro-
35. Tristani-Firouzi M, Etheridge SP: Kir 2.1 chan- implications. Cardiovasc Res 89(4):734 –743, sis and cardiac arrhythmias. J Cardiovasc Pharma-
nelopathies: the Andersen-Tawil syndrome. 2011. col 57(6):630–638, 2011.
Pflugers Arch 460(2):289–294, 2010. 48. Tamargo J, López-Sendón J: Novel therapeutic 60. Tardif JC, O’Meara E, Komajda M, et al; SHIFT
36. Dobrev D, Friedrich A, Voigt N, et al: The G targets for the treatment of heart failure. Nat Rev Investigators: Effects of selective heart rate reduc-
protein-gated potassium current IK,ACh is constitu- Drug Discov 10(7):536–555, 2011. tion with ivabradine on left ventricular remodelling
tively active in patients with chronic atrial fibrilla- 49. Michael G, Xiao L, Qi XY, et al: Remodelling of and function: results from the SHIFT echocar-
tion. Circulation 112(24):3697–3706, 2005. cardiac repolarization: how homeostatic responses diography substudy. Eur Heart J 32(20):2507–
37. Billman GE: The cardiac sarcolemmal ATP- can lead to arrhythmogenesis. Cardiovasc Res 2515, 2011.
sensitive potassium channel as a novel target for 81(3):491–499, 2009.
Pharmacogenomics of
Cardiac Arrhythmias 55
Dan M. Roden
absorption, distribution, and renal and biliary excretion reflect

CHAPTER OUTLINE
cellular drug uptake and efflux by specific transporter3 molecules.
Principles of Pharmacogenomics 541 It is variability in function and expression of metabolizing and
transport molecules, regulated by a host of genetic and environ-
The Future: Using Pharmacogenetic Information
mental factors, that determines pharmacokinetic variability. Sim-
in Patient Management 545 ilarly, variability in the biological milieu in which drugs act can
be conceptualized as variability in the function of multiple mol-
ecules, including the target molecules with which drugs interact
Individuals vary widely in their responses to therapy with most to produce their beneficial and adverse effects, whose integrated
drugs. Indeed, response to cardiovascular drug therapy and anti- behavior determines normal and abnormal cellular and whole-
arrhythmics in particular is so highly variable that study of the organ function.
underlying mechanisms has elucidated important lessons for Some DNA variants are rare, cause specific “monogenic” dis-
understanding variable responses to drug therapy in general.1,2 eases, and have conventionally been termed mutations. More
Single nucleotide changes can, by disrupting gene product common variants are termed polymorphisms and might or might
function, produce dramatic changes in physiology: The long QT not alter function or expression of the encoded protein. As we
syndromes and inherited errors of metabolism such as alkapton- begin to understand that each human harbors thousands of DNA
uria are examples. Indeed, recognition that inborn errors of variants4,5—some common and some extremely rare—the dis-
metabolism arose from defective biotransformation of endoge- tinction between “mutation” and “rare variant” becomes increas-
nous substrates led to the suggestion more than a century ago ingly unclear, and more generic language like rare and common
that exogenous substrates (drugs) might similarly be aberrantly polymorphisms is being adopted. One critical aspect of modern
metabolized and might produce unusual actions in affected genomics is that DNA tends to be highly ancestry specific. Vari-
patients. This pharmacogenetic paradigm has been validated by the ants implicated in traits like variable drug responses in one ethnic
identification of individual patients and families with defects in group may be absent in another, or different variants in the same
the genes encoding specific drug-metabolizing enzymes. The gene may contribute.
term pharmacogenomics encompasses the idea that variability in A change in a single nucleotide, a single-nucleotide polymorphism
drug responses across individuals or populations reflects the com- (SNP), is the most common type of DNA variant. Others include
bined influences of many DNA variants across individual nucleotide insertions or deletions (indels) and duplication or dele-
genomes. tion of large stretches of DNA, termed copy number variations
(CNVs). Only about 1% of the genome is protein-coding (this
subset of DNA is termed the exome), and protein function can be
altered if a polymorphism results in a change in primary amino
Principles of Pharmacogenomics acid sequence (a nonsynonymous polymorphism). In addition, non-
coding variants can alter protein abundance through multiple
Definitions mechanisms (e.g., by changing mRNA stability, by regulating the
rate of mRNA transcription). Such regulation can arise because
Two key steps are included in the series of events that take place of polymorphisms in the promoter (the region that directly con-
between administration of a drug and manifestation of its benefi- trols gene transcription, usually directly upstream of exon 1) or
cial or adverse effects (Figure 55-1). First, drug must be delivered in more distant genomic regions. One emerging view is that
to its molecular site of action (e.g., receptor, ion channel). The polymorphisms may be physiologically silent until an environ-
magnitude of the effect at the target is determined by drug con- mental stressor is superimposed: Examples of environmental
centration, and the study of the time dependence of the concen- stressors important for arrhythmia pathophysiology include
tration of drug (and metabolites) achieved in plasma, tissue, or adrenergic stress, acute myocardial ischemia, and administration
other sites such as urine or bile is termed pharmacokinetics. The of a drug.
second major process that determines drug action has been
termed pharmacodynamics and broadly includes the processes that
must occur between the interaction of a drug with a specific Approaches to Identifying
molecular target and the manifestation of drug action at the Pharmacogenetic Variants
molecular, cellular, whole-organ, and whole-patient levels.
Because drugs act in a complex (and often abnormal) biological Drugs display variability in both efficacy and toxicity, and phar-
milieu, considerable intersubject variability in drug effects can macogenomic experiments to date have addressed both types of
arise from pharmacodynamic mechanisms. drug responses. Drug efficacy often reflects the combined effects
These principles of pharmacokinetics and pharmacodynamics of multiple pharmacokinetic and pharmacodynamic determi-
have been recognized for decades, and it is now apparent that nants; thus, identifying polymorphisms with large effect sizes
they are manifestations of the highly regulated function of indi- contributing to efficacy has been challenging. Similarly, some
vidual molecules. Thus, metabolism of a drug occurs by interac- drug toxicities reflect an extension of the biology of efficacy (e.g.,
tion of the drug with specific drug-metabolizing molecules, and excessive ventricular rate slowing with atrioventricular [AV]
541
Eliminated pharmacogenomics.8 Here, single variants that alter the function

Disease of drug-metabolizing or transport molecules may confer a very
drug
high likelihood of developing aberrantly high (or low) plasma
Metabolism, drug concentrations—and thus highly variable drug responses—
excretion Whole organ during treatment. In addition, genetically determined variations
Absorption
Dose Plasma and in drug targets (molecules with which drugs interact to achieve
body effects their therapeutic or adverse effects) may strongly modulate the
Distribution outcomes of drug therapy. Specific examples are discussed in the
following sections.
Molecular
Modulators Unbiased Approaches: The Genome-Wide Association
target
Study Paradigm
Pharmacokinetics Pharmacodynamics
In arrhythmia science, GWASs have been used to identify new
genes and pathways involved in physiological traits (like electro-
Figure 55-1. Mechanisms modulating drug actions The left side illustrates cardiographic [ECG] intervals) and susceptibility to common
pharmacokinetic variables that determine absorption, distribution, metabolism, arrhythmias like atrial fibrillation (AF) or sudden cardiac death.9,10
and elimination (often abbreviated ADME). Variability in interactions between These results, in turn, are being used to explore the role of vari-
drugs and their molecular targets, along with variability in underlying pathophysi-
ants in these genes in variable drug response. In addition, GWASs
ologies including disease processes, modulates pharmacodynamic mechanisms
(right) that contribute to net drug responses. Understanding the molecular deter- have been used to directly analyze drug responses, as described
minants of pharmacokinetics and pharmacodynamics is the first step toward iden- later.11
tifying genetic variants that contribute to drug responses. A fundamental enabling discovery for the GWAS paradigm is
the concept of linkage disequilibrium. Although each human
harbors millions of common SNPs, many are “linked” in the
nodal blocking drugs), and thus the experimental challenges are sense that knowing the specific genotype at one locus allows an
similar. However, other toxicities are not predicted by what is investigator to infer the genotype at a second locus. If an SNP at
known about the efficacy of a drug and seem to occur in a rela- one genetic locus always informs the genotype at the second SNP
tively unpredictable fashion; examples include skin rashes, statin- site, the two are said to be in complete linkage disequilibrium.
related myopathy, and drug-induced arrhythmias. In some of Thus, a platform that interrogates large numbers of SNPs need
these apparently idiosyncratic cases, single variants with relatively only include a few such “tag” SNPs to identify genotypes across
large effect sizes have been identified. an entire linkage disequilibrium block or haplotype.
Proving that a DNA variant contributes to a specific clinical The GWAS experiment starts by identifying cases and con-
phenotype (such as an unusual drug response) requires compel- trols for a specific phenotype. These can be categorical (e.g.,
ling statistical arguments and replication in multiple datasets; premature heart disease, breast cancer, drug-induced adverse
demonstration that a variant produces altered biological proper- effect, AF, restless leg syndrome) or continuous (e.g., PR dura-
ties in vitro can also serve as a supporting argument. The rapid tion, warfarin steady state dose).6 High-throughput platforms are
proliferation of polymorphism databases has led to a very large then used to determine genotypes at hundreds of thousands or
number of false-positive associations between polymorphisms millions of SNP sites in cases and controls, and tests of associa-
and variable human phenotypes—associations that are subse- tion are performed at each SNP to identify those associated with
quently not reproduced. the phenotype under study (Figure 55-2). Evidence that the
Associating genetic variants with clinical phenotypes, includ- experiment has yielded a positive result may include very low P
ing drug response, in humans has taken one of two broad values (after correction for multiple comparisons), replication,
approaches. The first is predicated on a perceived understanding and ultimately biological plausibility. SNPs are chosen because
of the fundamental physiology, pathophysiology, or pharmacol- they tag blocks of linkage disequilibrium; therefore, there is little
ogy of the phenotype under study; this is termed a candidate gene expectation that those associated with low P values are functional
approach (see Figure 55-1). The second takes advantage of themselves. Rather, they act as sign posts within the genome,
emerging high-throughput technologies by genotyping or by identifying specific loci at which functional variants may reside.
direct sequencing of large regions of DNA (up to whole exomes GWAS analyses of the distribution of normal ECG intervals
and genomes) to then determine whether there is an association (e.g., PR, QRS, QT) have been conducted in tens of thousands
between any locus interrogated and the phenotype under study. of patients and have identified genomic loci that contribute to
To date, the most widely used method in this unbiased or variability in these traits.12-18 Some of these are, in retrospect,
hypothesis-free approach is the genome-wide association study obvious from an understanding of underlying physiology. Thus,
(GWAS) paradigm.6 One clear emerging lesson of these genetic for example, strong signals are present in the KCNQ1 and KCNH2
association studies is that any result requires further validation loci (encoding potassium channels important for cardiac repolar-
both by replication and by further experiments testing the under- ization) in GWAS analyses of variability in the QT interval.
lying biology. Mutations in these genes are the most common causes of the
congenital long QT syndrome; the GWAS result demonstrates
Candidate Gene Approaches that common variants in these genes contribute to physiological
Although the candidate gene approach is intuitively very appeal- variability of QT intervals in a normal population.
ing, repeated experience over the past decade has demonstrated Other signals identified by GWAS identify genes whose role
that initially identified associations frequently failed to replicate.7 in the phenotype under study is completely unsuspected. In the
The reasons for this failure of replication are multiple: (1) The QT analyses, the strongest signal has consistently been noted
candidate variant may not, in fact, explain a large proportion of near NOS1AP, which encodes an ancillary protein for the neuro-
the variance in the phenotype under study; (2) the studies gener- nal isoform of nitric oxide synthase. Initial studies suggest that
ally involve small numbers and so are underpowered; and (3) a NOS1AP encodes a regulator of cardiac potassium and calcium
publication bias is associated with positive results, so attempts to function.19 Follow-up studies have now implicated NOS1AP vari-
replicate generally regress to the mean. ants in phenotypes beyond normal QT variability: These include
A major exception to the general “rule” that candidate gene risk for sudden cardiac death in populations,20 risk for events in
studies fail to replicate in a robust fashion is seen in patients with congenital long QT syndrome,21,22 and risk of
Pharmacogenomics of Cardiac Arrhythmias 543
Variant A Variant B
susceptibility. However, the actual size of these effects can be
55
modest, and their contribution to overall heritability of the traits
under study is often small, with odds ratios rarely exceeding 2.0.
The fact that strong GWAS signals identify common variants
with only very modest effect sizes has been a major criticism of
the approach. Nevertheless, identification of new physiological
pathways to human phenotypes represents the major triumph of
the approach and, as has been outlined earlier, can be a starting
point for new risk stratifiers in disease settings and with drug
exposure.
Other Experimental Approaches

A design intermediate between a single candidate gene interroga-
8
7 tion and an unbiased GWAS approach is a multiplexed candidate
6 gene study. In this approach, hundreds of SNPs, chosen because
-log10(P)
5 they tag major haplotype blocks within logically chosen candidate

4
3 genes, are interrogated to identify specific loci that may associate
2 with the phenotype of choice.35 This approach offers the advan-
1 tage of shifting the focus away from single variants, along with
0
Chr1 Chr2 Chr3 Chr4 Chr5 Chr6 Chr7 Chr8 Chr9 Chr10 the disadvantage that it remains a candidate gene approach albeit
Chr11 Chr12 Chr13 Chr14 Chr15 Chr16 Chr17 Chr18 Chr19 using candidates chosen often through unbiased approaches. The
Chr20 Chr21 Chr22 ChrX ChrXY ChrMT hypothesis that combinations of genetic variants contribute to a
trait can be addressed using this method.36 Another emerging
Figure 55-2. The genome-wide association study paradigm The first approach is to combine the results of genomic studies with those
step (top panel) is to assign each subject in a large cohort to case (orange) or
control (white) status. The entire cohort is then genotyped at hundreds of thou-
of other studies yielding complementary datasets. Examples
sands to millions of common polymorphic sites. The figure illustrates how a hypo- include expression profiling in cell lines or specific tissues, infor-
thetical polymorphism predicting the phenotype in question might segregate: In matics approaches,37 and evaluation of drug response in model
this case, variant B is associated with the phenotype. A statistical test of association organisms with known genetic backgrounds, such as mice or
is then performed for each polymorphism and the results displayed on a Manhat- zebrafish.9,38
tan plot (bottom): The x-axis is the chromosomal location of the polymorphism, and An emerging approach in modern genomics is direct sequenc-
the y-axis is the exponent of the P value for the individual statistical test (higher ing of candidate genes, of loci implicated as modulators of a
values denote lower P values). In the example shown, a cluster of polymorphisms phenotype by GWAS, of large candidate gene sets, and ultimately
in chromosome 1 (arrow) achieves P values less than 10−8. The red horizontal line of the whole genome. New technologies have enabled the devel-
denotes an arbitrary level of statistical significance after correction for multiple
testing.
opment of very high throughput and reasonably accurate geno-
typing for such experiments. The challenge is that the larger the
region of the genome interrogated, the larger the number of
genetic variants identified, and the ability to relate specific vari-
sudden death during treatment with some drugs.23 The strongest ants to phenotypes under study remains a major challenge in
GWAS signal for variability in PR and QRS is seen in SCN10A, this area. Nevertheless, these approaches are being explored
which encodes a sodium channel previously implicated only in in genomics and have been applied to electrophysiological
pain perception and not known to play a role in the heart. Pre- phenotypes.39,40
liminary studies have suggested multiple roles for the gene in the
heart: A contribution to late sodium current,24 regulation of the
canonical cardiac sodium channel SCN5A,25,26 and a role in neural Pharmacokinetic Mechanisms
regulation of conduction27 have been suggested.
GWASs of patients with and without AF have consistently Each drug is a substrate for one or more specific drug-metabolizing
implicated SNPs at chromosome 4q25.28-30 The nearest gene enzymes and is transported into and out of cells by specific drug
encodes the transcription factor PITX2; initial studies suggest uptake and efflux molecules. When variable drug effects arise
that PITX2c, a cardiac-specific isoform, regulates both develop- because of variability in plasma or tissue drug concentrations,
ment of the pulmonary myocardium31 and expression of other polymorphisms in the genes encoding drug-metabolizing
genes (e.g., NPPA, KCNQ1) that have been implicated in AF enzymes or transporters are high-priority candidates for explain-
susceptibility.32 These data are also being used to inform addi- ing variability in drug effects.
tional studies of variable response to AF therapy. Thus, for Most drug metabolism occurs in the hepatocyte, where drugs
example, reports have suggested that SNPs at Chr4q25 predict are biotransformed to one or more metabolites, usually by oxida-
response to drug33 or ablation34 therapy in AF. tion. Drug metabolites are generally more polar than the parent
In addition to analysis of phenotypes such as ECG intervals drug and are themselves excreted or conjugated (most commonly
or disease susceptibility, the GWAS paradigm has been used to as glucuronides) before renal or biliary excretion. Oxidation is
directly study variability in drug response. Here, the problem is usually accomplished by members of the cytochrome P-450
that precise definitions of drug response phenotypes are needed, (CYP) superfamily, the most important members of which for
and the numbers of patients that can be accrued is by nature of drug metabolism are the CYP3A family (CYP3A4 and CYP3A5),
the experiment much smaller than analyses of ECG intervals or CYP2D6, CYP2C9, and CYP2C19. Conjugation is accomplished
of arrhythmias themselves.11 Nevertheless, as is described further by uridine glucuronyl transferase, N-acetyltransferase, or a group
later, initial attempts have been made to analyze phenotypes such of methyltransferases.8
as warfarin steady state dose requirement or susceptibility to
drug-induced torsades de pointes. High-Risk Pharmacokinetics
Large GWAS analyses have identified incontrovertible statis- The term high-risk pharmacokinetics has been used to describe
tically significant associations such as those between variants in specific clinical settings in which variation in normal metabolism
NOS1AP and QT duration or between variants at 4q25 and AF or excretory pathways can confer especially important clinical
effects.41 Such variation most often arises because of genetic poly- (erythromycin, itraconazole, cyclosporine) in elevating digoxin
morphisms in the pathway or because of coadministration of concentrations likely reflect interference with this P-glycoprotein
drugs that inhibit the pathway. One high-risk setting involves use elimination pathway. Similarly, sotalol and dofetilide are elimi-
of a drug that has a narrow margin between effective dose and nated largely by renal excretion of unchanged drug, so marked
toxic dose, and that has a single predominant route of elimina- QT prolongation is a real risk in patients with renal dysfunction
tion. A second setting involves administration of a prodrug, a drug given usual drug dosages; the specific transporter is unknown.
that requires a specific metabolic pathway for its bioactivation. Patients with entirely defective P-glycoprotein have not been
Examples of prodrugs whose bioactivation depends on drug described, although mice in which the gene is disrupted have no
metabolism pathways with known polymorphisms include clopi- manifest baseline phenotype. However, because of the very
dogrel and codeine. Table 55-1 presents examples of antiarrhyth- prominent role of P-glycoprotein in maintenance of an effective
mic drugs for which variants in single genes produce large effects. blood-brain barrier, these mice display striking accumulation in
A spectacular example of high-risk pharmacokinetics has been the central nervous system and resulting toxicity when exposed
drug interactions involving substrates of CYP3A4, notably the to certain drugs, including digoxin. Polymorphisms have been
antihistamine terfenadine and the pro-motility agent cisapride.42 described that modulate the function of drug transport molecules
These compounds are high-potency QT-prolonging agents in vitro; these have been linked to variability in plasma and tissue
(resulting from block of the potassium current IKr) and ordinarily concentrations and effects of digoxin, simvastatin, and many
undergo very extensive presystemic metabolism by CYP3A (and other drugs.
no other major pathway) to non–IKr-blocking metabolites. When Patients homozygous for loss-of-function alleles in CYP2D6
drugs that inhibit CYP3A are coadministered, presystemic (5% to 10% of European and African populations) display mark-
metabolism is inhibited, concentrations of terfenadine or cis- edly enhanced β-blocking action, including bronchospasm and
apride entering the systemic circulation can increase by several bradyarrhythmia, during propafenone therapy as the result of
orders of magnitude, and torsades de pointes due to IKr block accumulation of the parent drug. Similar effects of the poor
becomes a real risk. These risks led to withdrawal of these drugs metabolizer genotype have been reported for metoprolol. Fle-
from the market. Although the activity of CYP3A4 varies among cainide is also a CYP2D6 substrate, but because it also undergoes
individual patients, the reasons for this are not completely under- renal excretion, loss of CYP2D6 activity (on a genetic basis or
stood. No common nonsynonymous SNPs are present in the through drug interactions) is not usually a problem. However,
gene, although noncoding SNPs regulating function have been CYP2D6 becomes the major determinant of flecainide elimina-
described.43,44 Functionally important polymorphisms in the tion in patients with renal failure, and this is one setting in which
coding region of the very closely related gene for CYP3A5 are flecainide toxicity can occur if CYP2D6 activity is reduced on the
expressed in enterocytes and hepatocytes. Loss-of-function basis of genomic variation or drug interactions.
alleles are more common in white or Asian patients than in
African American patients.
Another widely used drug that is eliminated largely by a single Pharmacodynamic Mechanisms
molecular mechanism is digoxin, whose hepatic and renal excre-
tion is mediated by the drug efflux transporter P-glycoprotein DNA polymorphisms can result in important functional changes
encoded by ABCB1.3 The well-recognized effects of quinidine, in drug target molecules.1,2 One example is the common R389G
amiodarone, verapamil, and numerous other drugs (substitution of glycine for arginine at position 389) variant in
ADR1 encoding the β1-adrenergic receptor. The R389 variant
demonstrated a twofold to fourfold greater increase in myocyte
contractility during exposure to β-agonists and predicted a ben-
Table 55-1. Pharmacogenetics of Antiarrhythmic Drugs eficial response of patients with heart failure receiving bucindo-
lol; in fact, clinical response in G389 carriers was no different
Genes With from response to placebo. Recent studies implicate this polymor-
Polymorphisms Possibly phism as a modulator of rate control in atrial fibrillation.45
Drug Affecting Response Inhibitor of Another example may be polymorphisms in the KCNE2 gene.
Some, but not all, studies suggest that KCNE2 can partner with
Quinidine CYP2D6, ABCB1 and modify the function of HERG, the protein whose expression
Procainamide NAT2 results in IKr. A KCNE2 polymorphism resulting in T8A increases
sensitivity to IKr blockers approximately threefold and has been
Amiodarone CYP3A4, ABCB1,
linked to susceptibility to drug-induced torsades de pointes. A
CYP2C9
variant in the SCN5A promoter has been described that reduces
Flecainide CYP2D6 promoter activity and is associated with longer QRS duration
Propafenone CYP2D6 (slower conduction) at baseline,46 and, as has been discussed,
SCN10A variants also modulate QRS. The SCN5A variant pre-
Sotalol dicted greater QRS prolongation with sodium channel-blocking
Dofetilide drug challenge in Asian patients with Brugada syndrome46 and
thus becomes a candidate for modulating risk of sudden death
Dronedarone ABCB1
seen when conduction is slowed by drugs, disease, or genetic
Metoprolol, timolol CYP2D6, ADR1 syndromes.
β-Blockers ADR1 Variable responses to warfarin reveal that this is an example
of a drug in which two genes play an important role.1,47 One is
Verapamil CYP3A4, ABCB1 CYP2C9, which is responsible for the bioinactivation of
Diltiazem CYP3A4, ABCB1 S-warfarin, the active enantiomer of the drug. Patients with
CYP2C9 variants that result in loss of function have higher plasma
Warfarin CYP2C9, VKORC1
drug levels, increased drug effect, excessive risk of bleeding at
Adenosine ordinary doses, and decreased steady state dose requirements.
Digoxin ABCB1 The specific variants and their frequencies vary by ancestry. The
second gene is VKORC1, which encodes an important component
of the warfarin drug target (the vitamin K complex). Very rare autonomic function, potassium homeostasis, or PI3 kinase signal-
55
patients have VKORC1 coding region mutations, resulting in ing are examples. A fourth set of candidate genes are those
warfarin resistance. In addition, however, common variation in responsible for metabolism and elimination of the QT-prolonging
the promoter clearly modulates VKORC1 hepatic mRNA abun- drug; in this case, genetic variants will be specific to individual
dance and can be related to warfarin sensitivity. In fact, the initial culprit drugs. For example, the antipsychotic thioridazine is
warfarin dose requirement appears more dependent on VKORC1 metabolized by CYP2D6, and evidence suggests that poor metab-
variants than on those in CYP2C9. Steady state warfarin dose olizers are at increased risk for diTdP with this drug.53
requirements vary strikingly across ethnicities, and higher dose Individual case reports and series have provided evidence that
requirements in African subjects and lower ones in Asian subjects some patients with previously unrecognized congenital long QT
have been associated with VKORC1 variation. GWASs of warfa- syndrome may present with torsades de pointes when challenged
rin steady state dose have confirmed a prominent contribution with a QT-prolonging drug. In addition, common variants in
by common variants in CYP2C9 and VKORC1 and have suggested these genes could contribute: One example is a common nonsyn-
a role for at least one other gene, CYP4F2, which is thought to onymous SNP in the cardiac sodium channel gene resulting in
be involved in vitamin K oxidation.48,49 S1103Y, detected only in African Americans and reported to
modulate the risk of a range of arrhythmias, including diTdP and
The Example of Drug-Induced Long QT Syndrome sudden infant death syndrome.54,55 Systematic evaluations of
A range of genetic approaches have been used to study the poten- patients with diTdP, surveying increasingly large sets of congeni-
tial contributions of DNA variants to risk for drug-induced tor- tal long QT syndrome and other congenital arrhythmia syn-
sades de pointes (diTdP). This serious adverse drug event occurs drome genes, have identified potentially contributory variants in
in 1% to 5% of patients treated with QT-prolonging antiar- patients with diTdP in up to 65% of cases.39,56 Another set of
rhythmics, such as sotalol, dofetilide, or quinidine, and to a much candidate variants that might modulate risk are those in the β1-
lesser extent in patients exposed to noncardiovascular drugs such adrenergic receptor gene, but systematic surveys have not sup-
as terfenadine, cisapride, erythromycin, haloperidol, methadone, ported this contention.
and many others. The mechanism whereby patients receiving A large candidate gene survey studied 1424 SNPs in 18 high-
antiarrhythmic drugs are at so much higher risk than those priority candidate genes (including congenital long syndrome
receiving the noncardiovascular drugs remains unexplained and disease genes and NOS1AP) in a set of 176 European ancestry
may reflect a contribution by concomitant disease, notably atrial patients with diTdP and two sets of controls: 837 population
fibrillation, which is frequently the indication for prescribing controls and 207 patients exposed to QT-prolonging drugs and
QT-prolonging antiarrhythmics; data suggest that the period of not developing marked QT interval changes.35 This study identi-
conversion from atrial fibrillation to sinus rhythm is one of high fied a single nonsynonymous SNP in KCNE1 that predicted
risk for QT interval dysregulation because of mechanisms that diTdP with a relatively high odds ratio, 9.0. KCNE1 encodes a
are not well understood.42 key subunit necessary for physiological IKs function, and the
An understanding of normal and abnormal cardiac repolariza- variant, D85N, has been implicated as a modulator of normal QT
tion informs a set of candidate genes that may modulate QT and interval and as a risk factor in modulating the phenotype in both
diTdP risk. Virtually all drugs that cause torsades de pointes are congenital and drug-induced TdP. A preliminary report of a
IKr blockers. The biological context in which IKr blockers act GWAS using a very similar set did not identify common genetic
includes other elements of the action potential (e.g., IKr, INa, ICa), variants with large effect sizes.57
as well as mechanisms that control normal autonomic function
and serum potassium. One recent study suggested that PI3 kinase
inhibition (a property of some anticancer drugs that have been
linked to diTdP) affects multiple ionic currents, suppressing IKr The Future: Using Pharmacogenetic
and IKs and increasing late sodium current—effects that may Information in Patient Management
prolong QT.50 This complexity is consistent with the notion that
the mechanisms that maintain a short QT interval vary among One goal of pharmacogenetic studies is to identify mechanisms
individual patients, and patients with reduced repolarization leading to large variability with existing drug therapies. These
reserve as the result of genetic or environmental factors are at results, in turn, could be used to tailor therapy with existing drugs
increased risk for developing torsades de pointes on challenge and to evaluate new drugs to ensure that high-risk situations are
with an IKr blocker. Several studies have implicated variable IKs as avoided.1,2 Another possible outcome of pharmacogenomic
playing an important role in maintaining this reserve.51 One studies is the development of readily measurable biomarkers to
report indicated that in vitro exposure to an IKr-blocking drug predict individual patient responses. In addition, identification of
paradoxically shortened action potentials. The proposed explana- new pathways to variable physiological and drug responses may
tion was decreased expression of a microRNA (miRNA), whose be the first clue to the development of new drug targets.
ordinary role is to inhibit translation of KCNQ1, a key gene in A rapidly increasing knowledge base is relating common and
the IKs complex; the decrease in miRNA thus resulted in increased rare polymorphisms to variable drug response and other pheno-
IKs expression and unexpectedly shortened action potentials.52 types such as susceptibility to disease. The ultimate hope is that
This experiment highlights the way in which complex regula- such studies will usher in a new era of personalized medicine, in
tion of multiple ion currents control cardiac repolarization, as which an individual patient’s polymorphism set will be used to
well as a potential role for miRNAs as modulators of these efficiently diagnose disease, determine disease susceptibility, and
processes. select optimal therapies. Despite an increasingly compelling body
Because two common clinical situations in which marked QT of knowledge linking variable drug effects with genetic variation,
prolongation and torsades de pointes are observed are diTdP and adoption in clinical practice has been slow. Many reasons have
the congenital long QT syndromes, one obvious set of candidate been proposed to explain this apparent paradox. One is that large
genes for mediating diTdP risk consists of the congenital long randomized clinical trials, with very few exceptions, have not
QT syndrome disease genes. A second set of candidates are those, been conducted to demonstrate the value of adding pharmaco-
such as NOS1AP, that have been implicated in variability in genetic information to routine clinical care. A second is that the
normal QT intervals by GWASs. A third set includes genes effects of drug administration are so variable that even factoring
implicated by physiological studies,38 informatics approaches,37 or out pharmacogenetic contributors still leaves variability in drug
in silico modeling as modulators of the QT: Genes involved in responses. A third possible explanation is a logistic one: It is
cumbersome to not only prescribe a drug but at the same time information related to drug response and perhaps to other impor-
obtain a genetic test whose result (often appearing days later) may tant pathophysiological phenotypes such as susceptibility to
require a second patient encounter to adjust medication dose or disease can be embedded in patients’ electronic medical records,
to change medications. In busy clinical practice environments, it to be accessed when a culprit drug is prescribed. Electronic
may be simpler to choose different drugs or to ignore pharma- systems would then advise the physician at the point of care
cogenomic influences altogether. whether the choice of drug or the drug dosage needs to be
Rapid advances in genotyping technology are now raising altered. This potential approach to incorporating genomic variant
the possibility that testing for pharmacogenomic variation data into the flow of health care is now being explored at various
can be accomplished in a “preemptive” fashion, that is, genetic academic medical centers.58
19. Chang KC, Barth AS, Sasano T, et al: CAPON 36. Lubitz SA, Sinner MF, Lunetta KL, et al: Indepen-
References modulates cardiac repolarization via neuronal dent susceptibility markers for atrial fibrillation on
nitric oxide synthase signaling in the heart. Proc chromosome 4q25. Circulation 122:976–984,
1. Roden DM, Johnson JA, Kimmel SE, et al: Car- Natl Acad Sci U S A 105:4477–4482, 2008. 2010.
diovascular pharmacogenomics. Circ Res 109:807– 20. Kao WH, Arking DE, Post W, et al: Genetic varia- 37. Berger SI, Ma’ayan A, Iyengar R: Systems pharma-
820, 2011. tions in nitric oxide synthase 1 adaptor protein are cology of arrhythmias. SciSignal 3:ra30, 2010.
2. Roden DM, Wilke RA, Kroemer HK, et al: associated with sudden cardiac death in US white 38. Milan DJ, Kim AM, Winterfield JR, et al: A drug-
Pharmacogenomics: the genetics of variable community-based populations. Circulation sensitized zebrafish screen identifies multiple
drug responses. Circulation 123:1661–1670, 119:940–951, 2009. genes, including GINS3, as regulators of myocar-
2011. 21. Crotti L, Monti MC, Insolia R, et al: NOS1AP is dial repolarization. Circulation 120:553–559, 2009.
3. DeGorter MK, Xia CQ, Yang JJ, et al: Drug trans- a genetic modifier of the long-QT syndrome. Cir- 39. Ramirez AH, Shaffer CM, Delaney JT, et al: Novel
porters in drug efficacy and toxicity. Annu Rev culation 120:1657–1663, 2009. rare variants in congenital cardiac arrhythmia
Pharmacol Toxicol 52:249–273, 2012. 22. Tomás M, Napolitano C, De Giuli L, et al: Poly- genes are frequent in drug-induced torsades de
4. Tennessen JA, Bigham AW, O’Connor TD, et al. morphisms in the NOS1AP gene modulate QT pointes. Pharmacogenomics J 2012. doi: 10.1038/
Evolution and functional impact of rare coding interval duration and risk of arrhythmias in the tpj.2012.14. [Epub ahead of print]
variation from deep sequencing of human exomes. long QT syndrome. J Am Coll Cardiol 55:2745– 40. Refsgaard L, Holst AG, Sadjadieh G, et al: High
Science 337:64–69, 2012. 2752, 2010. prevalence of genetic variants previously associated
5. Nelson MR, Wegmann D, Ehm MG, et al: An 23. Becker ML, Visser LE, Newton-Cheh C, et al: A with LQT syndrome in new exome data. Eur J
abundance of rare functional variants in 202 drug common NOS1AP genetic polymorphism is asso- Hum Genet 20:905–908, 2012.
target genes sequenced in 14,002 people. Science ciated with increased cardiovascular mortality in 41. Roden DM, Stein CM: Clopidogrel and the
337:100–104, 2012. users of dihydropyridine calcium channel blockers. concept of high risk pharmacokinetics. Circulation
6. Manolio TA: Genomewide association studies and Br J ClinPharmacol 67:61–67, 2009. 119:2127–2130, 2009.
assessment of the risk of disease. N Engl J Med 24. Yang T, Atack TC, Stroud DM, et al: Blocking 42. Roden DM: Drug-induced prolongation of the
363:166–176, 2010. SCN10A channels in heart reduces late sodium QT interval. N Engl J Med 350:1013–1022,
7. Ioannidis JP, Ntzani EE, Trikalinos TA, et al: Rep- current and is antiarrhythmic. Circ Res 111:322– 2004.
lication validity of genetic association studies. Nat 332, 2012. 43. Wang D, Guo Y, Wrighton SA, et al: Intronic
Genet 29:306–309, 2001. 25. Arnolds DE, Liu F, Fahrenbach JP, et al: TBX5 polymorphism in CYP3A4 affects hepatic expres-
8. Wang L, McLeod HL, Weinshilboum RM: drives Scn5a expression to regulate cardiac conduc- sion and response to statin drugs. Pharmacoge-
Genomics and drug response. N Engl J Med tion system function. J Clin Invest 122:2509–2518, nomics J 11:274–286, 2011.
364:1144–1153, 2011. 2012. 44. Lamba V, Panetta JC, Strom S, Schuetz EG:
9. Milan DJ, Lubitz SA, Kaab S, et al: Genome-wide 26. van den Boogaard M, Wong LYE, Tessadori F, Genetic predictors of interindividual variability in
association studies in cardiac electrophysiology: et al: Genetic variation in T-box binding element hepatic CYP3A4 expression. J Pharmacol ExpTher
recent discoveries and implications for clinical functionally affects SCN5A/SCN10A enhancer. 332:1089–1099, 2010.
practice. Heart Rhythm 7:1141–1148, 2010. J Clin Invest 122:2519–2530, 2012. 45. Parvez B, Chopra N, Rowan S, et al: A common
10. Bezzina CR, Pazoki R, Bardai A et al: Genome- 27. Verkerk AO, Remme CA, Schumacher CA, et al: beta1-adrenergic receptor polymorphism predicts
wide association study identifies a susceptibility Functional Nav1.8 channels in intracardiac favorable response to rate-control therapy in atrial
locus at 21q21 for ventricular fibrillation in acute neurons: the link between SCN10A and cardiac fibrillation. J Am Coll Cardiol 59:49–56, 2012.
myocardial infarction, Nat Genet 42:688–691, electrophysiology. Circ Res 111:333–343, 2012. 46. Bezzina CR, Shimizu W, Yang P, et al: Common
2010. 28. Gudbjartsson DF, Arnar DO, Helgadottir A, et al: sodium channel promoter haplotype in Asian sub-
11. Motsinger-Reif AA, Jorgenson E, Relling MV, Variants conferring risk of atrial fibrillation on jects underlies variability in cardiac conduction.
et al: Genome-wide association studies in pharma- chromosome 4q25. Nature 448:353–357, 2007. Circulation 113:338–344, 2006.
cogenomics: Successes and lessons. Pharmaco- 29. Kaab S, Darbar D, van Noord C, et al: Large scale 47. Klein TE, Altman RB, Eriksson N, et al:
genet Genomics 23(7):333–340, 2013. replication and meta-analysis of variants on chro- Estimation of a warfarin dose with clinical and
12. Holm H, Gudbjartsson DF, Arnar DO, et al: mosome 4q25 associated with atrial fibrillation. pharmacogenetic data: International Warfarin
Several common variants modulate heart rate, PR Eur Heart J 30:813–819, 2009. Pharmacogenetics Consortium. N Engl J Med
interval and QRS duration. Nat Genet 42:117– 30. Ellinor PT, Lunetta KL, Albert CM, et al: Meta- 360:753–764, 2009.
122, 2010. analysis identifies six new susceptibility loci for 48. Cooper GM, Johnson JA, Langaee TY, et al: A
13. Pfeufer A, Sanna S, Arking DE, et al: Common atrial fibrillation. Nat Genet 44:670–675, 2012. genome-wide scan for common genetic variants
variants at ten loci modulate the QT interval dura- 31. Mommersteeg MTM, Brown NA, Prall OWJ, with a large influence on warfarin maintenance
tion in the QTSCD study. Nat Genet 41:407–414, et al: Pitx2c and Nkx2-5 are required for the for- dose. Blood 112:1022–1027, 2008.
2009. mation and identity of the pulmonary myocardium. 49. Takeuchi F, McGinnis R, Bourgeois S, et al: A
14. Newton-Cheh C, Eijgelsheim M, Rice KM, et al: Circ Res 101:902–909, 2007. genome-wide association study confirms
Common variants at ten loci influence QT interval 32. Wang J, Klysik E, Sood S, et al: Pitx2 prevents VKORC1, CYP2C9, and CYP4F2 as principal
duration in the QTGEN study. Nat Genet 41:399– susceptibility to atrial arrhythmias by inhibiting genetic determinants of warfarin dose. PLoS
406, 2009. left-sided pacemaker specification. Proc Natl Acad Genet 5:e1000433, 2009.
15. Pfeufer A, van Noord C, Marciante KD, et al: Sci U S A 107:9753–9758, 2010. 50. Lu Z, Wu CY, Jiang YP, et al: Suppression of phos-
Genome-wide association study of PR interval. 33. Parvez B, Vaglio J, Rowan S, et al: Symptomatic phoinositide 3-kinase signaling and alteration of
Nat Genet 42:153–159, 2010. response to antiarrhythmic drug therapy is modu- multiple ion currents in drug-induced long QT
16. Sotoodehnia N, Isaacs A, de Bakker PI et al: lated by a common single nucleotide polymor- syndrome. Sci Transl Med 4:131ra50, 2012.
Common variants in 22 loci are associated with phism in atrial fibrillation, J Am Coll Cardiol 51. Roden DM: Repolarization reserve: a moving
QRS duration and cardiac ventricular conduction. 60:539–545, 2012. target. Circulation 118:981–982, 2008.
Nat Genet 42:1068–1076, 2010. 34. Husser D, Adams V, Piorkowski C, et al: Chromo- 52. Xiao L, Xiao J, Luo X, et al: Feedback remodeling
17. Chambers JC, Zhao J, Terracciano CMN, et al: some 4q25 variants and atrial fibrillation recur- of cardiac potassium current expression: A novel
Genetic variation in SCN10A influences cardiac rence after catheter ablation. J Am Coll Cardiol potential mechanism for control of repolarization
conduction. Nat Genet 42:149–152, 2010. 55:747–753, 2010. reserve. Circulation 118:983–992, 2008.
18. Denny JC, Ritchie MD, Crawford DC, et al: Iden- 35. Kaab S, Crawford DC, Sinner MF, et al: A large 53. Llerena A, Berecz R, de la Rubia A, et al: QTc
tification of genomic predictors of atrioventricular candidate gene survey identifies the KCNE1 interval lengthening is related to CYP2D6 hydrox-
conduction: Using electronic medical records as a D85N polymorphism as a possible modulator of ylation capacity and plasma concentration of thio-
tool for genome science. Circulation 122:2016– drug-induced torsades de pointes. Circ Cardiovasc ridazine in patients. J Psychopharmacol 16:
2021, 2010. Genet 5:91–99, 2012. 361–364, 2002.
54. Splawski I, Timothy KW, Tateyama M, et al: 56. Itoh H, Sakaguchi T, Ding WG, et al: Latent syndrome. Circulation 120:S580, 2009. Abstract
Variant of SCN5A sodium channel implicated in genetic backgrounds and molecular pathogenesis 1971.
risk of cardiac arrhythmia. Science 297:1333–1336,
2002.
55. Plant LD, Bowers PN, Liu Q, et al: A common
in drug-induced long QT syndrome. Circ
Arrhythm Electrophysiol 2:511–523, 2009.
57. Kaab S, Ritchie MD, Crawford DC, et al: Genome-
58. Pulley JM, Denny JC, Peterson JF, et al: Opera-
tional implementation of prospective genotyping
for personalized medicine: the design of the Van-
55
cardiac sodium channel variant associated with wide association study identifies novel genomic derbilt PREDICT project. Clin Pharmacol Ther
sudden infant death in African Americans, SCN5A regions associated with drug-induced long QT 92:87–95, 2012.
S1103Y. J Clin Invest 116:430–435, 2006.
Optogenetic Control of Heart Muscle 56
Emilia Entcheva and Ira S. Cohen
potential components for bioelectronics and a new generation of

CHAPTER OUTLINE
optical memory because of their ultra-fine spatiotemporal control
Defining Optogenetics 549 by light (i.e., their ability to address single molecules by focused
light at very fast rates). The latter is of equal interest in control
The Optogenetics Toolbox 550
of eukaryotic cells. Structurally and functionally, BR provides
Optical Control of Cardiac Function 551 good insight for optogenetics as it shares high homology with all
(class I) opsins currently in use.
Potential Future Cardiac Applications 554
Conclusions 556
The New Generation of Single-Unit
Optical Actuators
Defining Optogenetics Current-day optogenetics began with the characterization and
cloning of ChR1 and the higher-conductance light-sensitive
Optogenetics is an emerging technology for optical interrogation ion channel ChR2 from green algae by Nagel, Hegemann,
and control of biological function with high specificity and spa- Bamberg, and colleagues1,2 in 2002 and 2003. This was followed
tiotemporal resolution. Mammalian cells and tissues can be sen- in 2005 by the first robust demonstrations of the use of ChR2
sitized to respond to light by a simple and well-tolerated genetic to stimulate mammalian cells.3,4 Upon heterologous expression,
modification using microbial opsins (light-gated ion channels and these microbial ion channels provide excitatory (cation-
pumps). Fast and specific excitatory or inhibitory responses can mediated) current with relatively fast kinetics9 and can effectively
be achieved, with distinct advantages over traditional pharmaco- trigger electrical impulses (action potentials) in excitable cells
logic or electrical perturbation. The breakthrough came with the upon light stimulation at relevant physiological rates. This
discovery of fast microbial opsins that behave like gated ion chan- demonstrated usefulness in neuroscience revived interest in
nels,1,2 and the subsequent demonstration that these microbial other types of microbial opsins, discovered earlier and extensively
opsins (channelrhodopsin2, ChR2, in particular) can generate studied within the microbial photobiology field. These include
sufficient photocurrent to optically stimulate and control the chloride pump halorhodopsin (HR)10 and the BR-like
mammalian neurons with very high temporal resolution.3,4 proton pump archaerhodopsin (AR).11 Both have proved capable
Optogenetics has already been widely used to elucidate brain of providing outward/hyperpolarizing current in mammalian
circuitry and function in health and disease,5-7 yet expansion of cells.12,13
this emerging technology in cardiovascular research has surpris-
ingly remained largely unexplored.
Distinct Advantages of Optogenetics
Nature-Derived Optical Actuators What makes recent optogenetic tools (several types of microbial
opsins) more practical compared with earlier systems and com-
Genetically encoded reporters of gene expression and cell func pared with classic (electrical and chemical) ways of stimulation
tion—green fluorescent proteins (GFPs) and their derivatives— are the following distinguishing characteristics: (1) simplicity of
have found widespread use. Similarly, optogenetics necessitates expression and operation without exogenous cofactors, offering
genetic modification of the cells and tissues of interest by heter- the attraction of a single-component system, in contrast to
ologous expression of microbial opsins. Unlike GFP-based prior multi-component optical actuators, such as using G-protein–
observational/imaging tools, optogenetics also offers actuation coupled signaling and ligand-requiring opsins (chARGe)14; (2)
possibilities—active perturbation of cell function (e.g., mem- apparent minimal interference with endogenous function15
brane potential or cell signaling) with high cellular specificity and (much less than genetically encoded voltage and calcium sensors)
spatiotemporal resolution, not attainable by pharmacologic or and remarkable reliability of use on a large scale in vitro and in
electrical means. vivo; (3) high specificity compared with electrical stimulation
(i.e., selective cell type targeting due to the genetic means of
Bacteriorhodopsin manipulation); (4) very high spatiotemporal precision of
Bacteriorhodopsin (BR) is one of the simplest and best studied manipulation (i.e., fast addressing of single molecules or cells by
optoelectrical transducers from the microbial (class I) opsins, the combination of focused light and genetically defined cell
found in archae, eubacteria, fungi, and algae. It is a protein with targets; (5) robustness and range of action within the same
seven transmembrane domains that acts like a light-gated active paradigm (i.e., excitatory and inhibitory effects can be encoded);
ion pump—it captures photon energy via its covalently bound (6) lower light energy required for activation compared with
chromophore, retinal—and moves protons against their electro- prior attempts to stimulate by infrared light alone16; and (7)
chemical gradient from the cytoplasm to the extracellular space. noncontact stimulation with precise localization, allowing highly
Since their discovery,8 the microbial opsins have been viewed as parallel/high-throughput operations.
549
Brief Overview of Neuroscience Applications chromophore (light-sensing element). Unlike BR, ChR2 is an ion
channel (not a pump), and upon opening it conducts cations
Neuroscience applications have successfully used the specificity down their electrochemical gradient.
offered by optogenetics to dissect neural circuits and connectiv- The chromophore, all-trans-retinal, is covalently bound to the
ity, linking specific neuron populations to normal behaviors and ion channel, and the complex does not undergo the dissociation
to disease. These studies include applications to better under- seen for class II mammalian rhodopsins, where the retinal/opsin
stand learning,17 olfactory processing in vivo,18 depression,19 sleep complex is reassembled/disassembled upon each stimulus. Upon
disorders,20 fear,21 and addiction.22 More translational studies interaction with a photon, all-trans-retinal undergoes isomeriza-
tackled questions related to epilepsy and terminating seizures tion to 13-cis-retinal, triggering channel opening. All-trans-
using inhibitory opsins,23 controlling Parkinson’s disease with retinal, derived from intake of vitamin A–containing nutrients, is
deep brain stimulation (DBS),24,25 countering visual degeneration present only in small amounts in nonretinal and nonembryonic
in retinitis pigmentosa,26 restoring respiratory control,27 optimiz- tissues (<0.5 nmol/g).30 It was a serendipitous finding that in most
ing nerve stimulation of skeletal muscle,28 and optimizing stem vertebrate cells, enough all-trans-retinal is present naturally to
cell differentiation while taking advantage of the inherently par- form functional ChR2 complexes. This is even more surprising
allel nature of optical stimulation.29 This is a small subset of the in cell culture, where the source of vitamin A must be serum/cell
wide spectrum of studies conducted thus far; more recent and culture impurities. To date, no systematic studies have demon-
comprehensive reviews provide further information.6,7 strated if and how retinal availability varies between cell and
tissue types, and whether it can be a limiting factor in the light
responsiveness of different cell types modified with ChR2.
Similar to BR, ChR2 has seven transmembrane (TM) domains.
The Optogenetics Toolbox It has a molecular weight of 77 kDa and a total of 737 amino
acids, approximately 300 of which are located at the amino-
The set of microbial opsins, adopted for use in mammalian cells, terminus and fully define its photocurrent generation.31 The
make up the “optogenetics toolbox.” These include both proteins crystal structure of ChR2 was recently solved,31 and this revealed
that generate depolarizing/excitatory currents and proteins that that the conductive pore is defined by TM1, -2, -3, and -7, and
produce inhibitory/hyperpolarizing currents. that TM7 is critical for the interaction with retinal, while TM2
determines channel selectivity and conductance. ChR2 has a
higher energy barrier for excitation than BR (energy is inversely
Excitatory/Depolarizing proportional to the wavelength)—its spectral response peaks at
Opsins—Channelrhodopsin2 around 470 nm (570 nm for BR).
ChR2 conducts cations with differential selectivity in the fol-
ChR2 from Chlamydomonas reinhardtii, cloned by Nagel et al. in lowing order (H+ > Na+ > K+ > Ca2+, …9; Figure 56-1). More
2003,2 is the prototypical and most widely used optogenetic tool. specifically, PH/PNa = 1.062 × 106; PK/PNa = 0.427; PCa/PNa = 0.117.
Like BR, it belongs to class I microbial opsins, all of which Thus, for physiological concentrations and membrane potentials,
require the cytosolic presence of retinal, which acts as a ChR2 provides predominantly Na+-mediated inward current. It
I/pF (pA /pF)

4
ChR2
Vm (mv)
–80 –60 –40 –20 0 20 40
Light ON
–4
–20 mV
–8
Retinal – 80 mV
–12
Voltage-dependent
hν
0.3 mW/mm2
5.5 mW/mm2
H+ > Na+ > K+ > Ca2+

A B Light-dependent
Figure 56-1. Biophysical Properties of ChR2 A, ChR2 is a light- and voltage-dependent ion channel; all-trans-retinal, intracellularly available and covalently bound
to ChR2, acts like a chromophore (sensing photons) to facilitate ChR2 opening and the transport of cations with differential preference from H+ to Ca2+. B, The resultant
current is predominantly inward/excitatory with a fast peak and a sustained component; ChR2 exhibits strong inward rectification with a reversal potential around 0 mV;
shown are also selected traces for ChR2 current under different voltage clamps (inset) and irradiance levels (bottom); scale bars are 10 pA/pF and 100 ms.
(Data for the sustained component shown from Jia Z, Valiunas V, Lu Z, et al: Stimulating cardiac muscle by light: Cardiac optogenetics by cell delivery. Circ Arrhythm Electrophysiol
4:753-760, 2011.)
has a reversal potential close to 0 mV, showing inward rectifica- recently developed.41 Genetic engineering is currently used to
56
tion (i.e., minimal outward current3,9,32-35). Several competing expand and optimize available opsins in three important aspects:
theories have been put forward for the mechanism of rectifica- light sensitivity, speed, and spectral response. These efforts,
tion, including that it is a single-channel property defined by an carried out mainly in the laboratories of Deisseroth, Bamberg,
asymmetrical barrier33 or a macroscopic property resulting from Hegemann, Tsien, Boyden, and others, have resulted in ChR2
the kinetics of multiple ion species interacting with the channel.35 mutants with single-amino-acid substitutions, including the
Upon a light pulse with the proper wavelength (470 nm) and higher-conductance H134R,3 T159C,42 and ET/TC42; the Ca2+
sufficient irradiance (in mW/mm2) for excitation, ChR2 gener- permeable CatCh43 and the speed-optimized ChETA44; hybrids
ates current that peaks rapidly and then relaxes to a smaller steady of ChR1 and ChR2 (ChIEF)9; and hybrids of ChR1 and VChR1
state level (see Figure 56-1). Higher irradiance and more negative (C1V1) for red-shifted variants,45 among others. In most cases,
voltages speed the kinetics of both activation and relaxation. Even optimization in one aspect (e.g., conductance) comes with a
at room temperature, all time constants are <20 ms.32,34 The tradeoff in another aspect (e.g., speed). Extensive quantitative
single-channel conductance for the wild-type ChR2 is relatively comparisons of genetically engineered opsins can be found in
small, and the few reported values vary widely from 40 to 90 fS33,36 several excellent reviews.5,6,42 Other efforts to improve the opto-
to 0.25 to 2.42 pS,9 depending on the method of estimation. For genetics toolbox include reduction of toxicity seen with early use
comparison, the Na+ channel conductance in muscle cells is as of these opsins, better membrane targeting, better cell specificity,
large as 18 pS.37 Zimmermann et al. used freeze-fracture electron and optimization of expression.46
microscopy (EM) and particle counting as well as whole cell
conductance and capacitance measurements to estimate ChR2
density of expression (typical case) and found about 2000
channels/µm.2,36 If all the channels were opened simultaneously, Optical Control of Cardiac Function
the generated whole cell current would be between 400 pA/pF
and 25 nA/pF. As is discussed later, one of the first ChR2 single- Overview of Early Work in Cardiac Optogenetics
amino-acid mutants (H134R) was designed to increase conduc-
tance by two- to threefold compared with wild type, with minimal Since 2010, publications have begun to appear that extend opto-
slowing of kinetics.3 genetics to cardiac muscle. Arrenberg et al.47 used a zebra fish
Conceptual and quantitative understanding of the function of model to express both excitatory (ChR2) and inhibitory (HR)
ChR2 is aided by recent mathematical models, as proposed by opsins. Using structured illumination, they demonstrated the use
Hegemann and colleagues38 and modified by others.39 A four- of optogenetics to spatially map the pacemaking region in zebra
state model is currently favored, with two open states (a high- fish during development. They also presented a range of rhythm
conductance state and a low-conductance, light-adapted one) and disorders that were triggered optically. Bruegmann et al.48 pub-
two closed states. Photon absorption and isomerization of retinal lished the first cardiac mammalian application. They combined
constitutes a near-instantaneous process, so that ChR2 confor- viral expression of a ChR2 variant with a CAG promoter in
mational changes, after light sensing, determine its photocurrent mouse embryonic stem cells (mESCs), with targeted differentia-
kinetics. Current models capture light dependence well but over- tion and purification of ESC-derived cardiomyocytes for in vitro
simplify voltage dependence. demonstration of optical pacing. Furthermore, they generated
transgenic mice with cardiac ChR2 expression, in which normal
rhythm was perturbed in vivo by light pulses, and focal arrhyth-
Inhibitory/Hyperpolarizing Opsins mias were induced by long pulses.
Simultaneously and independently, our group demonstrated a
Opsins that produce inhibitory/hyperpolarizing current include nonviral optogenetic approach.34 Given the heart’s dense and
the chloride pump HR from Natronomonas pharaonis adopted for well-coupled environment, we used cell delivery to demonstrate
mammalian use, eNpHR,12 and some BR-like proton pumps (e.g., optical pacing in vitro. We called this strategy a tandem-cell-unit
Archaerhodopsin-3 [AR] from Halorubrum sodomense13) and the (TCU) approach, where dedicated (nonexcitable) donor cells
lower-wavelength-activated pump from the fungus Leptosphaeria expressing light-sensitive ion channels (ChR2) couple with car-
maculans (Mac).13 AR is the most potent inhibitory opsin to date; diomyocytes, providing light sensitivity to individual myocytes or
compared with HR and Mac, it offers larger photocurrent and a cardiac syncytium. Our study also demonstrated the first inte-
faster recovery from inactivation. Similar to HR and Mac, AR gration of high-speed/high-resolution optical imaging with
provides an outward current with an extremely negative reversal optogenetics-based actuation for a fully optical interrogation of
potential and only about a 20% drop in current when going from excitable tissue and quantitative comparison of wave propagation
0 to −120 mV. By its action, despite being an efficient proton upon optical versus electrical stimulation. Two more in vitro
pump, AR changes the H+ concentration (pH) minimally (i.e., the studies were published approximately at the same time—one
intracellular pH can increase by about 0.15 after 1 minute of using a cardiac cell line (HL-1), in which ChR2 was expressed by
continuous illumination).13 The turnover rate of AR is unknown electroporation,49 and the second using lentiviral delivery of
but by whole cell current (1 nA induced by strong illumination ChR2 into human embryonic stem cells (hESCs), followed by
in neurons13) is comparable in amplitude with the ChR2 current, cardiomyocyte differentiation.50 Both studies applied microelec-
most likely because of high density of expression, as is the case trode arrays (MEAs) to confirm electrical response upon optical
for most active pumps usually with an order of magnitude of stimulation. Abilez et al.50 combined their experiments with com-
higher density than ion channels.40 putational modeling of the function of ChR2 into cardiac tissue.
Optimization of Optical Actuators Cardiac Electrophysiology and Optogenetics

Having both excitatory and inhibitory optogenetics tools of com- Compared with neuronal electrophysiology, cardiac action
parable performance opens the possibility for optical control of potentials are longer and more complex, reflecting the close inte-
membrane potential (i.e., “shaping” the action potentials and/or gration of electrical and mechanical function with a prominent
the frequency response of the system). A method for optimized role for Ca2+ as intermediary. Physiologically relevant frequencies
tandem expression of excitatory and inhibitory opsins has been of electrical response are one to two orders of magnitude lower
than for neurons. Myocytes are big and well coupled via gap computer simulations (unpublished data) with the ten Tusscher
junctions, forming a syncytium. Excitation waves, under normal ventricular cell model,52 including a version of ChR2 modified
conditions, follow well-known paths, as reflected in the highly from earlier papers.39
regular electrocardiogram (ECG).
Given these characteristics, it is likely that optimization of
optogenetics tools and their specific application will follow dif- Genetic Modification of Cardiac Tissue
ferent routes compared with neuroscience. For example, higher-
speed opsins are not as relevant for the heart; higher-conductance Common techniques for genetic modification (including optoge-
opsins are quite relevant in light of the higher electrotonic load. netics) are shown schematically in Figure 56-3, A. Transgenic
Although in neuroscience the ideal excitatory pulses are very brief mice present an attractive experimental model and can be gener-
and information is encoded mostly by the frequency of the pulses, ated with specifically targeted cell types. Various strains of trans-
for cardiac applications, it is possible to explore longer, lower- genic ChR2-expressing mice were developed by Feng and
intensity pulses. In the cardiac literature, energy minimization colleagues for in vivo neuroscience research,53 and many of these
has been pursued via waveform optimization,51 and it is known are currently available through the Jackson Laboratory. To date,
that a typical rectangular monophasic pulse, for example, does no commercially available mice with cardiac expression of opto-
not provide the most efficient stimulus. It can be speculated that genetic tools have been produced. Witten et al.54 have developed
the conceptually different optogenetic mode of stimulation, a more elegant general strategy for cell type–specific expression
where the stimulus (the actual photocurrent induced) is shaped of opsins in rats. They used recombinase-driver rat cell lines that
via rapid real-time feedback about the membrane voltage (Figure can drive the gene expression in specific cell types with Cre
56-2, B) may be inherently optimal for stimulation and perhaps recombinase under control of relatively large regulatory regions
more energy efficient, especially at longer low-irradiance pulses. (>200 kb). This approach permits faster generation of experi-
It is more difficult to speculate about the performance of opto- mental rat models than other classical transgenic approaches and
genetic pulses for cardioversion and defibrillation; it is hoped that will likely be used in cardiac applications as well. Such Cre driver
computer modeling and in vitro experiments will provide addi- lines confer another level of selectivity, in addition to promoter-
tional insight into how the pre-stimulus state may affect the determined selectivity, when combined with viral delivery. For
outcome. Figure 56-2 illustrates the response of a ventricular cardiac applications, one can envision targeting the conduction
myocyte to comparable electrical and optical stimulation using system as a whole or regions of it (e.g., the sinoatrial node, Pur-
kinje fibers). Although specific gene expression (HCN4, Cx40)
or locally enriched transcription factors, such as Contactin-255
have been investigated for the conduction system, more work is
needed to make relevant promoters available for cell-specific
targeting, similar to the arsenal available for neuroscience.
Electrically-triggered AP
In addition to transgenic animals, gene targeting in vitro or
Optically-triggered AP in vivo can be achieved by direct DNA delivery (by electropora-
tion or other transfection methods), viral delivery, or cell delivery.
0 mV Through concerted efforts to optimize the optogenetics toolbox,
multiple laboratories use repositories like Addgene and make
20 mV
their constructs publicly available. Of the shown delivery methods

0.1 s (see Figure 56-3, A), viral delivery using lentivirus (LV) or adeno-
associated virus (AAV) is the most common in optogenetics appli-
cations6 because of the high efficiency and potential specificity
noted if a cell/tissue-specific promoter is used. Successful long-
A term expression with LV and AAV has been demonstrated not
only in mice56 and rats24 but also in primates.15 For cardiac use,
I ChR2 optimization of viral delivery will involve a search for the most
suitable AAV serotype and small promoters for cell-specific
expression (small payload for AAV of <4.7 kb limits promoter
size). The cell delivery approach (Figure 56-3, B) is based on the
TCU strategy and relies on coupling of donor cells with native
myocytes with a coupling conductance > 2nS.34 This approach is
3 pA/Pf
particularly relevant to cardiac applications involving stem cell

delivery. It does not address cell-specific labeling of native myo-
B 0.1 s cytes, but it allows for optimization of donor cells for better opsin
10 ms performance in vivo.
Figure 56-2. Optogenetic Stimulation of Cardiac Cells Human ventricu- When optogenetics were applied to cardiac tissue, it was
lar myocytes, stimulated electrically (5 ms, 10 pA/pF) and optically (10 ms, unknown a priori whether the required chromophore for ChR2
5 mW/mm2), produce very similar action potentials; the underlying ChR2 current operation can be found in sufficient amounts. For example, bulk
during a cardiac action potential is fast inward current that gets reversed by the tissue measurements obtained with chromatography and ultra-
change in voltage and becomes briefly outward for positive voltages. The inset violet (UV) techniques (high-performance liquid chromatogra-
illustrates that optogenetic stimulation is inherently waveform-optimized (com- phy [HLPC]/UV)30 indicate that cardiac muscle may contain
pared with electrical rectangular pulses) because of the built-in feedback control
smaller quantities of endogenous retinoids than are seen in liver,
by voltage (i.e., the inward current injection self-terminates once the membrane
has been depolarized).
kidney, adipose tissue, and brain. Yet the few cardiac applications
previously discussed did not use exogenous retinol, thus suggest-
(Modified from Entcheva E: Cardiac optogenetics. Am J Physiol Heart Circ 304:H1179- ing that some (possibly sufficient) quantities of retinal were avail-
H1191, 2013; the results were taken from computer simulations with the ten Tusscher able. It is interesting to note that the human embryonic kidney
myocyte model, integrated with a ChR2 model; modified from Nikolic K, Grossman N, cells (HEK293) used in our TCU experiments34 and in other
Grubb MS, et al: Photocycles of channelrhodopsin-2. Photochem Photobiol 85:400- optogenetics studies are known for their optimized retinoid
411, 2009.) machinery, but myocytes most likely are not.
Virus-mediated Cell-mediated
56
Direct plasmid
Virus
Cell
Efficacious
Plasmid
Transgenic
Safe/desirable
A Optically excitable cardiac tissue
Cardiomyocyte
Host Electrical response
cardiomyocyte
Donor cell:
ChR2
Optical stimuli
B
Figure 56-3. Inscribing Light Sensitivity in Cardiac Tissue A, The most common approaches to optogenetic transduction include the generation of a transgenic
animal or different ways of gene delivery: direct plasmid transfection, virally mediated or cell mediated; the relative efficacy and safety of these approaches is depicted.
B, The cell delivery approach works for well-coupled cells, as in the myocardium, where a tandem cell unit (TCU) can be formed between a nontransduced myocyte and
a nonexcitable donor cell. Shown are a canine adult ventricular cell and a donor ChR2 HEK cell, in which optical stimuli drive action potentials in the myocyte.
(From Entcheva E: Cardiac optogenetics. Am J Physiol Heart Circ 304:H1179-H1191, 2013, with permission; also from Jia Z, Valiunas V, Lu Z, et al: Stimulating cardiac muscle by
light: Cardiac optogenetics by cell delivery. Circ Arrhythm Electrophysiol 4:753-760, 2011.)
Challenges for Light Access in the Heart cardiac muscle). Alternatively, catheter accessibility to the endo-
cardial conduction system may provide immediate opportunities
In neuroscience applications, precise injections for optogenetic for optogenetic applications.
targeting of desired brain locations and implantation of stimulat-
ing or recording devices are performed by widely available ste-
reotactic systems. Furthermore, implantable optogenetic devices Energy for Optical Stimulation
evolved rapidly in brain research since 200724 to the current fully
integrated systems in freely moving animals, in some cases with The optical stimulus strength needed to trigger a response is
wireless powering.57 typically measured in units of irradiance (mW/mm2). Strength-
An analogous approach to the stereotactic system is not in duration curves link minimum irradiance and pulse duration and
place for cardiac gene or cell delivery. The challenge is to achieve capture the overall energy required for optogenetic stimulation
this goal in a beating heart without the firm support and point (Figure 56-4, A). The energy is influenced by a multitude of
of reference naturally offered by the skull for the brain. Optical factors, including expression levels and functionality of the
fiber conduits for imaging purposes have been developed before opsins, the host cell electrophysiological milieu (balance of depo-
for cardiac applications58; possibly this intramural optrode larizing and repolarizing currents), the cable properties of the
approach could be adopted for localized gene or cell delivery, as tissue and electrotonic load for activation, the efficiency of light
well as for optical stimulation and recording. Alternatively, for delivery/penetration, and so on. In Figure 56-4, A, strength-
optical stimulation, surface-conforming solutions, whereby the duration curves are assembled from published data for cardiac
device is moving with the contracting heart, may come from cell monolayers with the TCU approach34 and for ventricular and
recent new developments in stretchable electronics and optoelec- atrial tissue in transgenic mice48; these are compared with the
tronics59 (e.g., light-emitting diode [LED] matrices can be orga- much higher values reported for stimulation in neural applica-
nized to conform and follow accessible surfaces—epicardial or tions in vitro and in vivo.4,53,61 The results are surprising if we
endocardial—with minimally invasive procedures similar to the consider only cable properties. In such cases, it is not obvious
use of inflatable balloons). As a dense and highly scattering that neural applications would require higher energies than those
medium, the heart may require the development of red-shifted needed for cardiac use. Furthermore, atrial myocytes were found
optogenetics tools, similar to VChR1 and C1V1.45 Two-photon to express ChR2 at higher levels and to produce larger functional
excitation of ChR2 offers an alternative way of increasing wave- currents than ventricular myocytes.48 Yet when tested at the tissue
length60 (i.e., it may be possible to stimulate deeper tissue level, a puzzling result was the greater energy needed to excite
[>0.5 mm] by surface illumination even in the denseness of atrial muscle than ventricular tissue, which theoretically should
by ChR2 current with electrical current injected in a rectangular

10.00 Neural 1 pulse, the waveform in the two cases (see Figure 56-2) suggests
that optical stimulation may yield benefit for longer pulses, but
the instant-upstroke electrical pulses are more efficient at short
Irradiance (mW/ms/mm 2 )
durations. Without a doubt, stimulus delivery to the site of inter-

Atrial 2
1.00
est will profoundly affect the overall energy requirements.
Ventricular 2
Potential Future Cardiac Applications

0.10 Cardiac monolayers 3
Optical Pacing and Basic Studies of Arrhythmias
It has to be emphasized that the power of optogenetics lies in the
0.01 possibility of offering new ways to better address basic science
A 0 20 40 60 80 100 questions. Whether it can progress into more translational/
therapeutic uses remains to be seen, for neuroscience and for
cardiac applications.
As a basic science tool in cardiac research, optical pacing can
offer contact-less stimulation with higher spatiotemporal resolu-
tion and cell selectivity and a new ability for parallelization com-
pared with electrical stimulation. The possibility of combining
optical stimulation with optical readout is particularly attractive
Localized Distributed for all-optical interrogation of cardiac electrophysiology (voltage
8 or calcium; Figure 56-5, A). We demonstrated the combined use
Energy (mW * ms/mm 2 )
of high-speed ultra-high-resolution optical mapping with optical

Gene stimulation34 (see Figure 56-5, B). Because of its contact-less
nature, optical pacing naturally lends itself to parallelization and
Cell
scalability, as well as closed-loop feedback control.
4 These features can be useful for numerous research questions.
One set of problems deals with probing and confirming cell-to-
cell coupling (e.g., cardiomyocyte–fibroblast coupling, coupling
between donor [stem] cells and host cardiomyocytes) in regenera-
B tive cardiomyoplasty. Currently, no direct and specific method is
available to address these questions in vivo. Optogenetics may
Figure 56-4. Energy for Optical Stimulation of Cardiac Tissue offer solutions via selective cell type–specific expression and
A, Strength-duration curves (irradiance and pulse duration needed to pass optical stimulation, if light access problems are resolved. A
the threshold for stimulation) are assembled from published data for cardiac cell
second related set of problems deals with initiation of focal
monolayers with the TCU approach and for ventricular and atrial tissue in trans-
genic mice; these are compared with the much higher values reported for stimula- arrhythmias. Suspected common sites (e.g., automaticity at endo-
tion in neural applications in vitro and in vivo. B, Energy to stimulate depends on cardial Purkinje network locations)63 can be systematically studied
the spatial distribution of gene or cell delivery of ChR2: Computational data show by cell-specific expression and perturbation by light to induce or
that lower energy is needed for direct gene delivery in myocytes if a localized area suppress such activity. Critical contributions of different parts of
is transduced; for a more sparsely distributed expression, inert cell delivery may be the pacemaking and conduction system can be probed, as was
more efficient. demonstrated in the zebra fish study.47 Such approaches may
facilitate understanding of arrhythmia induction and may offer
(A, Data from multiple papers, including Boyden ES, Zhang F, Bamberg E, et al:
new antiarrhythmic strategies. A third set of problems suitable
Millisecond-timescale, genetically targeted optical control of neural activity. Nat Neu-
for in vitro investigation is related to mechanisms of reentrant
rosci 8:1263-1268, 2005; Bruegmann T, Malan D, Hesse M, et al: Optogenetic control
arrhythmias and their termination. Precise dynamic optical
of heart muscle in vitro and in vivo. Nat Methods 7:897-900, 2010; and Jia Z, Valiunas
probing can be used to address the exact nature of reentrant
V, Lu Z, et al: Stimulating cardiac muscle by light: Cardiac optogenetics by cell deliv-
activation and the state of the reentrant core—spiral wave versus
ery. Circ Arrhythm Electrophysiol 4:753-760, 2011.)
leading circle. The search for mechanisms of atrial fibrillation or
(B, From Boyle PM, Williams JC, Entcheva E, Trayanova NA: Spatial distribution of
ventricular fibrillation—mother rotor, wandering wavelets, or
channelrhodopsin-2 affects optical stimulation efficiency in cardiac tissue. Heart
other—can be better tackled by fine stimulation tools to establish
Rhythm 9:S254, 2012.)
vulnerability and may have real impact on the development of
better defibrillation strategies.
have presented the bigger electrotonic load. Clearly, multiple

factors are at play. Given that gene targeting will rarely result in Cardioversion
perfectly uniform expression, we also explored computationally
how the energy of stimulation will depend on the spatial distribu- Classical defibrillation works by synchronous depolarization of a
tion of gene (viral) or cell delivery (see Figure 56-4, B). The critical mass of the myocardium (>95%) using strong electrical
results show that less energy is needed for (viral) gene delivery shocks; lower-energy alternatives are pursued by the proper
in myocytes if a localized area is transduced, but for more sparsely timing of multiple electrical shocks to interact with and extin-
distributed expression, cell delivery may be more efficient.62 guish reentrant waves underlying an arrhythmic episode.64 For
A key question for cardiac pacing is whether optogenetic acti- optogenetic termination of arrhythmias (cardioversion and defi-
vation can be more energy efficient than electrical stimulation, brillation), both excitatory and inhibitory approaches will be of
potentially providing longer battery life. Our in vitro experiments interest. Global hyperpolarization (or forced repolarization) is
hinted at such a possibility.34 If one compares the charge delivered difficult to achieve by electrical fields; thus optical suppression
56
Light
Optical actuation
LED/
driver Vm
Acquisition
Visualization
Control
Ca 2+
Sample
L
PD1
Optical sensing
LS DM1
Ex F Em F
PD2
M DM2
A
Electrical stimulation Optical stimulation
STIM STIM
A A A A
B B B B
1s 2 mm 1s
B
0s 0.4 s
Figure 56-5. All-Optical Actuation, Sensing and Control of Cardiac Function A, Shown is a system for all-optical contact-less cardiac electrophysiology,
built around a microscope. Optical actuation is achieved through collimated LED-produced light; optical sensing of voltage (Vm) and calcium (Ca2+) is done by voltage-
sensitive dye (di-8-ANEPPS) and calcium-sensitive dye (Rhod-4), respectively; shown are actual records (light pulse was 50 ms, 0.2 mW/mm2). Computer controls the LED
driver and the acquisition by photodetectors (PD1, PD2), thus allowing for a closed-loop feedback control. Other components include L—objective lens; LS—light source
for imaging; Ex F, Em F—excitation and emission filters; and M, DM1, DM2—full and dichroic mirrors. B, All-optical interrogation of cardiac function over time and space
by combining high-resolution optical mapping with optogenetic actuation (from Jia et al., 2011). Waves of excitation in cardiac monolayers, triggered by electrical and
optical pacing at 0.5 Hz and captured by activation maps. Color represents time of activation; isochrones are shown in black at 0.15 s. Calcium transients (Rhod-4) in response
to electrical or optical stimulation are shown from two locations, with normalized fluorescence. Blue marks indicate time of stimulation (electrical pulses were 10 ms; optical
were 20 ms each). In panels A and B, the cell delivery approach of ChR2 was used.
(From Entcheva E: Cardiac optogenetics. Am J Physiol Heart Circ 304:H1179-H1191, 2013, with permission.)
offers a new tool. It is interesting to point out successful termina- However, efforts have been made to extend optogenetics to
tion of epileptic seizure activity by light (using inhibitory HR) in include broader control of physiological parameters, especially
hippocampal tissue slices in vitro,23 as epilepsy and cardiac protein-protein interactions and cellular signaling. An example
arrhythmias share some mechanistic similarities. Concerns of a cardiac application used flavin-binding opsins to create light-
regarding in vivo optical cardioversion and defibrillation do not induced fusion of calcium ion channels (Cav1.2) to show how
involve the magnitude of achievable photocurrents, but instead oligomerization and channel clustering may affect the current
pertain to the need for spatially distributed light delivery. produced by these channels in cardiomyocytes.65 Furthermore,
It is unclear if optogenetics can prove a “disruptive technol- in addition to the microbial class I opsins used for direct control
ogy” for in vivo pacing and defibrillation, considering the success of voltage (see Figure 56-2), the optogenetics toolkit has been
of the current devices. However, if potential safety concerns are expanded to include derivatives of the vertebrate class II
resolved and if critical benefits are demonstrated (e.g. substantial opsins, commonly referred to as OptoXR.66 These are G-
battery life extension or pain reduction for defibrillation), then protein–coupled proteins that can interact with intracellular
perhaps optical devices will be a viable alternative. messengers, including cyclic adenosine monophosphate
(cAMP), phosphoinositide-3-kinase (PI3K), and inositol-1,4,5-
triphosphate (IP3), and can provide precise optical interrogation
Beyond Optical Control of Voltage of biochemical signaling. Even though such uses may exhibit
lower temporal resolution, the selectivity and the spatial preci-
The common use of microbial opsins, as discussed here, involves sion of manipulation offered by optogenetics tools are still
control of electrical activity (i.e., transmembrane voltage). desirable.
Conclusions Acknowledgments
Optogenetics has already proved to be an indispensable research We would like to acknowledge current and past researchers
approach in neuroscience. A vast number of useful light-sensitive working on this project in our labs (Z. Jia, PhD; H. Bien, MD,
actuation tools have been produced and made publicly available. PhD; C. Ambrosi, PhD; J. Xu, BS; J.C. Williams; A. Klimas,
Cardiac applications are in their infancy, although it is already MS; J. Yu, BS; J. Kalra; X. Chen, BS; Z. Lu, MD, PhD; B.
clear that optogenetics tools will allow difficult questions to be Rosati, PhD; H. Liu, MS; J. Zuckerman, MS; C. Gordon, MS),
investigated that heretofore were unapproachable with current as well as collaborators at Stony Brook (P.R. Brink, PhD; V.
tools.67 However, new developments are needed to accelerate Valiunas, PhD; C. Gordon, MS) and at Johns Hopkins University
cardiac use of optogenetics, especially in vivo, and to help cardiac (P.M. Boyle, PhD; N.A. Trayanova, PhD). Our work on
electrophysiologists realize the full potential of this unique cardiac optogenetics is supported by NIH-NHLBI grant
approach. R01HL111649.
18. Arenkiel BR, Peca J, Davison IG, et al: In vivo delivery. Circ Arrhythm Electrophysiol 4:753–760,
References light-induced activation of neural circuitry in 2011.
transgenic mice expressing channelrhodopsin-2. 35. Gradmann D, Berndt A, Schneider F, et al: Recti-
1. Nagel G, Ollig D, Fuhrmann M, et al: Neuron 54:205–218, 2007. fication of the channelrhodopsin early conduc-
Channelrhodopsin-1: A light-gated proton channel 19. Airan RD, Meltzer LA, Roy M, et al: High-speed tance. Biophys J 101:1057–1068, 2011.
in green algae. Science 296:2395–2398, 2002. imaging reveals neurophysiological links to behav- 36. Zimmermann D, Zhou A, Kiesel M, et al: Effects
2. Nagel G, Szellas T, Huhn W, et al: ior in an animal model of depression. Science on capacitance by overexpression of membrane
Channelrhodopsin-2, a directly light-gated cation- 317:819–823, 2007. proteins. Biochem Biophys Res Commun
selective membrane channel. Proc Natl Acad Sci U 20. Carter ME, Adamantidis A, Ohtsu H, et al: Sleep 369:1022–1026, 2008.
S A 100:13940–13945, 2003. homeostasis modulates hypocretin-mediated sleep- 37. Sigworth FJ, Neher E: Single Na+ channel currents
3. Nagel G, Brauner M, Liewald JF, et al: Light acti- to-wake transitions. J Neurosci 29:10939–10949, observed in cultured rat muscle cells. Nature
vation of channelrhodopsin-2 in excitable cells of 2009. 287:447–449, 1980.
caenorhabditis elegans triggers rapid behavioral 21. Johansen JP, Hamanaka H, Monfils MH, et al: 38. Hegemann P, Ehlenbeck S, Gradmann D: Multiple
responses. Curr Biol 15:2279–2284, 2005. Optical activation of lateral amygdala pyramidal photocycles of channelrhodopsin. Biophys J
4. Boyden ES, Zhang F, Bamberg E, et al: Millisecond- cells instructs associative fear learning. Proc Natl 89:3911–3918, 2005.
timescale, genetically targeted optical control of Acad Sci U S A 107:12692–12697, 2010. 39. Nikolic K, Grossman N, Grubb MS, et al: Photo-
neural activity. Nat Neurosci 8:1263–1268, 2005. 22. Lobo MK, Covington HE 3rd, Chaudhury D, cycles of channelrhodopsin-2. Photochem Photo-
5. Mattis J, Tye KM, Ferenczi EA, et al: Principles for et al: Cell type-specific loss of BDNF signaling biol 85:400–411, 2009.
applying optogenetic tools derived from direct mimics optogenetic control of cocaine reward. 40. Opie LH: The Heart: Physiology, From Cell to
comparative analysis of microbial opsins. Nat Science 330:385–390, 2010. Circulation, ed 4, Baltimore, MD, 2003, Lippin-
Methods 9:159–172, 2012. 23. Tonnesen J, Sorensen AT, Deisseroth K, et al: cott Williams & Wilkins.
6. Yizhar O, Fenno LE, Davidson TJ, et al: Optoge- Optogenetic control of epileptiform activity. Proc 41. Kleinlogel S, Terpitz U, Legrum B, et al: A gene-
netics in neural systems. Neuron 71:9–34, 2011. Natl Acad Sci U S A 106:12162–12167, 2009. fusion strategy for stoichiometric and co-localized
7. Fenno L, Yizhar O, Deisseroth K: The develop- 24. Aravanis AM, Wang LP, Zhang F, et al: An expression of light-gated membrane proteins. Nat
ment and application of optogenetics. Annu Rev optical neural interface: In vivo control of rodent Methods 8:1083–1088, 2011.
Neurosci 34:389–412, 2011. motor cortex with integrated fiberoptic and opto- 42. Berndt A, Schoenenberger P, Mattis J, et al: High-
8. Oesterhelt D, Stoeckenius W: Rhodopsin-like genetic technology. J Neural Eng 4:S143–S156, efficiency channelrhodopsins for fast neuronal
protein from the purple membrane of Halobacte- 2007. stimulation at low light levels. Proc Natl Acad Sci
rium halobium. Nat New Biol 233:149–152, 1971. 25. Kravitz AV, Freeze BS, Parker PR, et al: Regulation U S A 108:7595–7600, 2011.
9. Lin JY, Lin MZ, Steinbach P, et al: Characteriza- of parkinsonian motor behaviours by optogenetic 43. Kleinlogel S, Feldbauer K, Dempski RE, et al:
tion of engineered channelrhodopsin variants with control of basal ganglia circuitry. Nature 466:622– Ultra light-sensitive and fast neuronal activation
improved properties and kinetics. Biophys J 626, 2010. with the Ca(2)+-permeable channelrhodopsin
96:1803–1814, 2009. 26. Busskamp V, Duebel J, Balya D, et al: Genetic catch. Nat Neurosci 14:513–518, 2011.
10. Matsuno-Yagi A, Mukohata Y: Two possible roles reactivation of cone photoreceptors restores visual 44. Gunaydin LA, Yizhar O, Berndt A, et al: Ultrafast
of bacteriorhodopsin: A comparative study of responses in retinitis pigmentosa. Science 329:413– optogenetic control. Nat Neurosci 13:387–392,
strains of Halobacterium halobium differing in pig- 417, 2010. 2010.
mentation. Biochem Biophys Res Commun 27. Gourine AV, Kasymov V, Marina N, et al: Astro- 45. Zhang F, Prigge M, Beyriere F, et al: Red-shifted
78:237–243, 1977. cytes control breathing through pH-dependent optogenetic excitation: A tool for fast neural
11. Ihara K, Umemura T, Katagiri I, et al: Evolution release of ATP. Science 329:571–575, 2010. control derived from volvox carteri. Nat Neurosci
of the archaeal rhodopsins: Evolution rate changes 28. Llewellyn ME, Thompson KR, Deisseroth K, 11:631–633, 2008.
by gene duplication and functional differentiation. et al: Orderly recruitment of motor units under 46. Gradinaru V, Zhang F, Ramakrishnan C, et al:
J Mol Biol 285:163–174, 1999. optical control in vivo. Nat Med 16:1161–1165, Molecular and cellular approaches for diversifying
12. Gradinaru V, Thompson KR, Deisseroth K: 2010. and extending optogenetics. Cell 141:154–165,
Enphr: A Natronomonas halorhodopsin enhanced for 29. Stroh A, Tsai HC, Wang LP, et al: Tracking stem 2010.
optogenetic applications. Brain Cell Biol 36:129– cell differentiation in the setting of automated 47. Arrenberg AB, Stainier DY, Baier H, et al: Opto-
139, 2008. optogenetic stimulation. Stem Cells 29:78–88, genetic control of cardiac function. Science
13. Chow BY, Han X, Dobry AS, et al: High- 2011. 330:971–974, 2010.
performance genetically targetable optical neural 30. Kane MA, Folias AE, Wang C, et al: Quantitative 48. Bruegmann T, Malan D, Hesse M, et al: Optoge-
silencing by light-driven proton pumps. Nature profiling of endogenous retinoic acid in vivo and netic control of heart muscle in vitro and in vivo.
463:98–102, 2010. in vitro by tandem mass spectrometry. Anal Chem Nat Methods 7:897–900, 2010.
14. Zemelman BV, Lee GA, Ng M, et al: Selective 80:1702–1708, 2008. 49. Hofmann B, Maybeck V, Eick S, et al: Light
photostimulation of genetically charged neurons. 31. Kato HE, Zhang F, Yizhar O, et al: Crystal struc- induced stimulation and delay of cardiac activity.
Neuron 33:15–22, 2002. ture of the channelrhodopsin light-gated cation Lab Chip 10:2588–2596, 2010.
15. Han X, Qian X, Bernstein JG, et al: Millisecond- channel. Nature 482:369–374, 2012. 50. Abilez OJ, Wong J, Prakash R, et al: Multiscale
timescale optical control of neural dynamics in the 32. Chater TE, Henley JM, Brown JT, et al: Voltage- computational models for optogenetic control of
nonhuman primate brain. Neuron 62:191–198, and temperature-dependent gating of heterolo- cardiac function. Biophys J 101:1326–1334, 2011.
2009. gously expressed channelrhodopsin-2. J Neurosci 51. Qu F, Li L, Nikolski VP, et al: Mechanisms of
16. Jenkins MW, Duke AR, Gu S, et al: Optical pacing Methods 193:7–13, 2010. superiority of ascending ramp waveforms: New
of the embryonic heart. Nat Photon 4:623–626, 33. Feldbauer K, Zimmermann D, Pintschovius V, insights into mechanisms of shock-induced vulner-
2010. et al: Channelrhodopsin-2 is a leaky proton ability and defibrillation. Am J Physiol Heart Circ
17. Huber D, Petreanu L, Ghitani N, et al: Sparse pump. Proc Natl Acad Sci U S A 106:12317–12322, Physiol 289:H569–H577, 2005.
optical microstimulation in barrel cortex drives 2009. 52. ten Tusscher KH, Noble D, Noble PJ, et al: A
learned behaviour in freely moving mice. Nature 34. Jia Z, Valiunas V, Lu Z, et al: Stimulating cardiac model for human ventricular tissue. Am J Physiol
451:61–64, 2008. muscle by light: Cardiac optogenetics by cell Heart Circ Physiol 286:H1573–H1589, 2004.
53. Wang H, Peca J, Matsuzaki M, et al: High- neural control of freely-behaving animals. J Neural stimulation efficiency in cardiac tissue. Heart
speed mapping of synaptic connectivity using pho- Eng 8:046021, 2011. Rhythm 9:S254, 2012.
tostimulation in channelrhodopsin-2 transgenic
mice. Proc Natl Acad Sci U S A 104:8143–8148,
2007.
58. Kong W, Ideker RE, Fast VG: Intramural optical
mapping of v(m) and Ca(i)2+ during long-duration
ventricular fibrillation in canine hearts. Am J
63. Robichaux RP, Dosdall DJ, Osorio J, et al: Periods
of highly synchronous, non-reentrant endocardial
activation cycles occur during long-duration ven-
56
54. Witten IB, Steinberg EE, Lee SY, et al: Physiol Heart Circ Physiol 302:H1294–H1305, tricular fibrillation. J Cardiovasc Electrophysiol
Recombinase-driver rat lines: Tools, techniques, 2012. 21:1266–1273, 2010.
and optogenetic application to dopamine-mediated 59. Kim DH, Ghaffari R, Lu N, et al: Flexible and 64. Luther S, Fenton FH, Kornreich BG, et al: Low-
reinforcement. Neuron 72:721–733, 2011. stretchable electronics for biointegrated devices. energy control of electrical turbulence in the heart.
55. Pallante BA, Giovannone S, Fang-Yu L, et al: Annu Rev Biomed Eng 14:113–128, 2012. Nature 475:235–239, 2011.
Contactin-2 expression in the cardiac Purkinje 60. Papagiakoumou E, Anselmi F, Begue A, et al: Scan- 65. Dixon RE, Yuan C, Cheng EP, et al: Ca2+ signaling
fiber network. Circ Arrhythm Electrophysiol less two-photon excitation of channelrhodopsin-2. amplification by oligomerization of l-type cav1.2
3:186–194, 2010. Nat Methods 7:848–854, 2010. channels. Proc Natl Acad Sci U S A 109:1749–
56. Adamantidis AR, Zhang F, Aravanis AM, et al: 61. Cardin JA, Carlen M, Meletis K, et al: Driving 1754, 2012.
Neural substrates of awakening probed with opto- fast-spiking cells induces gamma rhythm and con- 66. Airan RD, Thompson KR, Fenno LE, et al: Tem-
genetic control of hypocretin neurons. Nature trols sensory responses. Nature 459:663–667, porally precise in vivo control of intracellular sig-
450:420–424, 2007. 2009. nalling. Nature 458:1025–1029, 2009.
57. Wentz CT, Bernstein JG, Monahan P, et al: A wire- 62. Boyle PM, Williams JC, Entcheva E, et al: Spatial 67. Entcheva E: Cardiac optogenetics. Am J Physiol
lessly powered and controlled device for optical distribution of channelrhodopsin-2 affects optical Heart Circ 304:H1179–H1191, 2013.
Cell Therapy and Regenerative
Electrophysiology 57
Timothy J. Kamp
CHAPTER OUTLINE Cell Sources for Cardiac Repair

Introduction to Cardioregenerative Medicine 559
A wide variety of different cell sources have been investigated for
Cell Sources for Cardiac Repair 559
their ability to repair the heart in animal models and some in
Basic Mechanisms by which Cell Therapy Can Affect clinical trials. Advances in stem cell research over the past two
Cardiac Electrophysiology 560 decades have also contributed to the variety of cells types under
consideration. Cells from the recipient of the graft (autologous)
Clinical Experience with Cell Therapy for Ischemic and isolated from donors (allogeneic) have been tested (Box
Heart Disease and Arrhythmias 563 57-1). Autologous cell sources hold the advantage of not being
Conclusions 565 recognized by the immune system as foreign, but they have the
disadvantage of potentially costly individualized cell processing
and the possibility that the disease phenotype will make the
transplanted cells dysfunctional. Alternatively, allogeneic cells
may be at risk for immune rejection, but the ability to optimize
Introduction to Cardioregenerative Medicine and manufacture large batches of quality-controlled cells for use
in multiple patients could be advantageous. The cell sources vary
The human heart has a limited ability for repair after myocardial dramatically in their properties, including proliferative capacity,
infarction and other major insults. The remodeling process that potency, ability to differentiate into different cell types, ability to
occurs in response to major cell loss includes hypertrophy of survive ischemic and inflammatory insults, and secretion of sig-
remaining cardiomyocytes and fibrosis of the myocardium, which naling molecules. Furthermore, it is likely that the different cell
provide both triggers and substrate for arrhythmias. Continued sources can differentially affect the electrophysiological proper-
stress on the myocardium can lead to progressive remodeling and ties either acutely or over time.
heart failure. Although pharmacologic therapies have greatly Initial investigations in cell therapy for the heart sought the
advanced and can blunt or in some cases partially reverse the goal of remuscularizing the tissue by providing differentiated
remodeling of the failing heart, typically these therapies are only myocytes. The first cell source studied in detail was skeletal
partially effective with substantial morbidity and mortality myoblasts derived from skeletal muscle satellite cells.7,8 Satellite
remaining in part because of arrhythmias. cells can be isolated from a muscle biopsy and differentiated into
In the past decade, insights regarding the regenerative capa- myoblasts that can be expanded greatly in culture. Transplanted
bilities of the heart have offered new hope for the treatment of myoblasts formed viable grafts in animal hearts and improved
heart disease. Although the adult human heart has been described functional properties of the hearts9,10; however, the transplanted
as a postmitotic, terminally differentiated organ, new studies have cells formed skeletal muscle grafts, not cardiac muscle. This
provided evidence that the heart is a dynamic organ with turnover result had consequences because skeletal muscle lacks connexin
of cells throughout life, including cardiomyocytes.1,2 Like many expression and gap junctions; therefore, the cells did not electri-
other organ systems, tissue-specific stem and progenitor cells cally integrate into the myocardium.10,11 Alternatively, trans-
have been identified recently; they provide a source for generat- planted fetal mouse ventricular myocytes were demonstrated to
ing new cardiomyocytes and other essential cell types in the integrate into a recipient mouse heart forming intercalated discs
heart.3-5 However, endogenous cardiac stem cells are unable to between donor cells and the recipient myocardium.12 Treatment
generate adequate numbers of cardiomyocytes to repair the heart of infarcted or cryoinjured myocardium with fetal cardiomyo-
after large insults, such as myocardial infarction. Furthermore, cytes in several animal studies resulted in improved left ventricu-
the capacity of intrinsic repair by cardiac stem cells declines with lar function and limited adverse myocardial remodeling compared
age.6 Thus, the concept of delivering new, viable cells to the heart with sham control animals.13,14 However, fetal cardiomyocytes are
for repair and regeneration has been investigated aggressively poorly tolerant of acute ischemia, with the vast majority not
over the past decade for treatment after myocardial infarction and surviving the transplant. Fetal cardiomyocytes are also terminally
in heart failure. Ideally, such cell-based therapy will lead to regen- differentiated, and they do not exhibit significant cell division at
erated myocardium that exhibits normal functional properties the site of engraftment. Therefore, it is difficult to obtain ade-
and consequently reduces the risk of arrhythmias as the abnormal quate cell numbers for larger areas of damage. Finally, the major
substrate is replaced, and the conditions that trigger arrhythmias limitation of applying this strategy to clinical medicine is the
are eliminated. However, the delivery of cells to the myocardium ethical objection to the use of human fetal tissue as well as the
can also potentially introduce conditions that increase the risk for limited supply of such tissue.
arrhythmias. The purpose of this chapter is to examine the elec- In 2001, Orlic et al.15 demonstrated that bone marrow–derived
trophysiological consequences of cell therapy for heart disease lineage-negative (lin−) and c-kit+ stem cells injected after myocar-
based on existing experimental data and early clinical dial infarction in a mouse model resulted in remarkable repair of
experience. the heart accompanied by functional improvement.15 The
559
The establishment of technologies to produce human pluripo-

Box 57-1 Cell Sources for Cardiac Therapy
tent stem cells created additional considerations for cardiac cell
AUTOLOGOUS ALLOGENEIC therapy. The successful isolation of human embryonic stem cells
(ESCs) from surplus in vitro fertilization embryos by Thomson
Skeletal myoblasts Fetal cardiomyocytes et al.30 created new avenues for cardiac therapeutic applications.30
Hematopoietic stem cells Embryonic stem cells Subsequent demonstration that human ESCs cells could differ-
entiate into functional cardiomyocytes confirmed the potential
• C-kit+ lin− and derivatives
utility of this cell source for cardiac repair.31,32 Studies in animal
• Bone marrow mononuclear cells models with mouse ESCs showed that the cells could have ben-
Endothelial progenitor cells Mesenchymal stem eficial effects after myocardial infarction and generate different
• CD34+ cell*,† cell lineages following transplantation of undifferentiated
• CD133+ ESCs.33-35 However, transplanting the ESCs without differentia-
tion carries the risk of teratoma tumor formation36; therefore,
Cardiac stem/progenitor cells subsequent studies have evaluated differentiated derivatives for
• Side population repair.37,38 Even more recently, the remarkable reprogramming of
• C-kit+ somatic cells, such as dermal fibroblasts, to induced pluripotent
• Cardiosphere-derived stem cells has provided a pluripotent stem cell source that can
• Epicardial progenitors potentially be genetically identical to the patient.39,40 Initial
studies using this cell source have shown early promise,41 but
Induced pluripotent stem cells many questions remain regarding the stability of the phenotype
and derivatives* and the long-term effects of reprogramming.
As investigators more critically examined cell survival and
* Can be both autologous and allogeneic. engraftment after cell delivery, it became progressively clear that
†
Can be derived from multiple tissues including bone marrow, adipose. many of the functional benefits observed in animal models were
likely not due to simple remuscularization. The majority of trans-
planted cells, regardless of source, did not survive let alone gener-
ate new myocardium. Nevertheless, clear beneficial effects of the
therapies were observed based on the functional and structural
appealing concept that delivering stem cells to the injured heart properties of treated hearts, and a number of additional mecha-
could lead to robust regeneration of functional myocardium was nisms of benefit have been proposed (Figure 57-1). Perhaps most
put forward, but soon these results were challenged by others prominent among the potential mechanisms of benefit is the
arguing against the ability of a hematopoietic stem cell to form reported paracrine effects of certain cell populations, such as
cardiac tissue.16,17 Nevertheless, investigators considered other MSCs, to secrete molecules to promote survival of existing heart
cells sources, such as mesenchymal stem cells (MSCs) derived tissue and blunt adverse remodeling.19,42 Other potential benefi-
from bone marrow to treat the injured heart, likewise suggesting cial effects include activating endogenous stem cells, cell fusion,
the generation of new myocardium; however, this conclusion has induction of angiogenesis, antiinflammatory effects, and resyn-
also been challenged.18,19 chronizing the myocardium. The exact mechanistic effect likely
The relatively recent demonstration of rare endogenous varies with different cell sources, and these details are far from
cardiac stem and progenitor cells in the heart provides additional completely defined in animal studies. Nevertheless, these early
possible cell sources for cardiac repair. Investigators have used a animal studies have generated sufficient interest and data to
number of different techniques and cell surface markers to isolate proceed quickly to clinical trials.
endogenous cardiac progenitors. In the earliest study, progenitors
were suggested based on the ability to expel Hoechst 33342 dye
which was used with fluorescence-activated cell sorting (FACS)
to identify a rare “side population” of cells in a similar fashion as Basic Mechanisms by Which Cell Therapy
had previously been done in hematopoietic cells.20 The side pop- Can Affect Cardiac Electrophysiology
ulation cells have cardiac potential based on in vitro studies. The
cell surface marker c-kit was subsequently used to identify mul- Cellular grafts need to undergo electrical and mechanical inte-
tipotent cardiac stem cells in mouse and human hearts.21, 22 Like- gration into the myocardium for optimal benefit. Furthermore,
wise Sca-1 was identified as a cell surface marker for cardiac the functional properties of the cells ideally must match those of
progenitors, but this protein is not expressed in the humans.4 The normal myocardium. To investigate and optimize these features
transcription factor Isl-1 has also been used to define a cardiac of cell therapies, studies have been performed using in vitro and
progenitor cell population in transgenic mouse models and in the animal models. Depending on the precise details of the grafts and
human heart.5, 23 Explanted cardiac tissue can be cultured under their integration, the transplanted cells can be either proarrhyth-
conditions promoting the migration of cells that can be isolated mic or antiarrhythmic (Figure 57-2). Cell therapy can contribute
and cultured and expanded under nonadherent conditions to to the genesis of arrhythmias by affecting all three basic mecha-
form cardiospheres.24,25 These cardiospheres contain a mixed nisms of arrhythmias: reentry, abnormal automaticity, and trig-
population of progenitors with the ability to form multiple lin- gered activity. Alternatively, cell therapy can blunt arrhythmias
eages including cardiomyocytes. Finally, the epicardium has been by improving the underlying substrate and removing triggers.
reported to contain multipotent epicardial progenitors identified Careful examination of the integration and functional properties
by WT-1 or Tbx-18, which can give rise to multiple cell types in of the cellular grafts is essential for optimizing safe and effective
the myocardium.26,27 The differences between the various pro- cell therapy approaches.
genitor populations is still the subject of intense investigation,
and it is possible that different strategies identify some related
populations at different stages of maturation.28 Nevertheless, Cell Coupling and Integration
transplanting such progenitor cells holds appeal for cardiac
repair, and initial animal studies have shown functional benefit Successful regeneration of myocardium with cell therapy requires
after myocardial infarction.3, 15, 29 electromechanical integration of the new cells into the functional
Cell Therapy and Regenerative Electrophysiology 561
Figure 57-1. Possible mechanisms of benefit from cell therapy for

Transplanted cells injured myocardium.
57
Regeneration/repair Stabilization
Paracrine Cell fusion
CMs, ECs, VSM of scar
Endogenous Reactivation of
stem cells cell cycle in CMs
Neovascularization Matrix remodeling

reduce fibrosis
Ischemic
post-conditioning Reduced apoptosis
Immune
modulation
Proarrhythmic effects
• Poor coupling, condution blocks

• Electrical heterogeneities
• Abnormal automaticity
• Inexcitable tissue resulting in source-sink mismatch
• Abnormal autonomic innervation
• Abnormal ion channel expression pattern
• Tissue inflammation/edema
Antiarrhythmic effects
• Improved coupling in areas of damage

• Reduce electrical heterogeneities
• Resynchronization
• Blunt adverse remodeling
• Normalization of ion channel expression pattern
• Neovascularization with reduced ischemia
Figure 57-2. Potential effects of cardiac cell therapy on arrhythmia risk.
myocardium. However, significant barriers must be overcome in then the delivery of cells could produce wavefront breaks and
the diseased heart in order for integration to be successful. The reentry.
presence of scarring and fibrosis in the heart requires significant As an initial test of the ability of donor cells to couple with
remodeling in order for integration of transplanted cells with the native cardiomyocytes, a number of in vitro coculture experi-
functional myocardium. Delivering or homing the transplanted ments have been performed. These cocultures of donor cells with
cells to the site in the heart in need of repair is also a major chal- ventricular cardiomyocytes have highlighted different forms of
lenge. Not only must the graft integrate and couple to electromechanical coupling. In the case of cocultured human
native myocardium; ideally, it regenerates tissue with matched ESC-derived cardiomyocytes, clear coupling with rat neonatal
anisotropic conduction properties of heart. The success of ventricular myocytes has been demonstrated with the formation
graft integration will in part determine the effect on arrhythmia of connexin43 (Cx43) gap junctions.43 Synchronized contractions
risk. For example, replacing or reducing nonexcitable or slowly were observed in the cocultures, suggesting clear functional cou-
conducting tissue at the site of infarction would reduce the sub- pling. In studies coculturing MSCs with rat ventricular myocytes,
strate for reentrant arrhythmias. Alternatively, if cell therapy formation of Cx43 gap junctions was also observed, but in this
produces areas of uncoupled tissue, poorly coupled tissue, or case electrotonic conduction occurred via MSCs because these
coupled inexcitable tissue that produces source-sink mismatches, cells are not electrically excitable.44,45 In contrast, coculture of
skeletal myoblasts with neonatal rat ventricular cardiomyocytes have reached conflicting results regarding the arrhythmia risk,
failed to couple because myotubes do not express Cx43 or form but species and model differences might help to explain the dif-
gap junctions.46 Thus, islands of electrically isolated myotubes ferent results.
provided substrate for reentry involving the surrounding cardio- Overall, the studies to date emphasize the importance of
myocytes. However, genetically engineered expression of Cx43 donor cells being able to couple electrically to the native ven-
in myoblasts can lead to functional coupling with neonatal car- tricular myocardium via Cx43-containing gap junctions. In the
diomyocytes.46 These simple coculture studies highlight key dif- absence of such coupling, examples of increased incidence of
ferences in coupling that can occur after cell therapy to the arrhythmias in animal models have been observed. However, in
myocardium, but understanding the full complexity of electro- grafts that exhibit robust coupling via Cx43, some animal studies
mechanical coupling resulting from cell therapy requires study in have demonstrated a reduction in the risk of ventricular arrhyth-
intact hearts. mias in the heart after injury.
A wide range of different cell sources have been tested in
various animal models of cardiac injury; however, only a small
minority of the studies has rigorously investigated the electro- Intrinsic Properties of the Transplanted Cells
physiologic consequences of cell therapy. Transplanted skeletal
myoblasts were first examined for their ability to couple to native The effect of cell therapies to the injured heart will depend on
myocardium. Despite being transplanted into the heart, myo- the integration of the donor cells and the functional properties
blasts differentiate into skeletal myotubes lacking Cx43 and do of the transplanted cells. Given the array of different cell types
not couple to the post–myocardial infarction rat heart,47,48 which tested, there is a tremendous range of possible functional proper-
is consistent with the in vitro studies. However, genetically engi- ties for engrafted cells. Transplantation of stem or progenitor
neered expression Cx43 in skeletal myotubes can result in cou- cells can lead to the generation of diverse cellular progeny in the
pling to the native heart and reduce risks for ventricular graft (e.g., cardiomyocytes, endothelial cells, fibroblasts), which
arrhythmias in a mouse infarct model.49 In the case of bone can differentially affect the electrophysiological status of the
marrow derived MSCs, transplantation into a rat myocardial treated heart. Some of the resulting cell types are excitable,
infarction model resulted in improved electrical conduction whereas others are not. Unexcitable cells such as fibroblasts
properties around the borderzone of the infarct based on voltage or mesenchymal stem cells can still exert potent electrophysio-
optical mapping studies with the MSCs expressing multiple con- logical effects by acting to bridge nonconducting areas of
nexins including Cx43.50 This finding with MSCs parallels the in myocardium (antiarrhythmic) or to produce current sinks increas-
vitro studies suggesting that MSCs can provide electronic cou- ing heterogenous conduction patterns (proarrhythmic). Trans-
pling of cardiomyocytes. plantation of terminally differentiated cardiomyocytes will lead
Transplanted mouse fetal cardiomyocytes into the adult mouse to cardiomyocyte-dominated grafts. However, a wide range of
heart have been shown to electrically couple to the native myo- functional properties of the transplanted cardiomyocyte is pos-
cardium using a number of different approaches. Using GFP- sible given differences in the type and maturity of the engrafted
labeled fetal cardiomyocytes transplanted to a native mouse cardiomyocytes. To complicate the analysis further, the func-
heart, multiphoton microscopy was used to demonstrate syn- tional properties of the transplanted cells can change over time
chronized intracellular calcium transients comparing native adult as they adapt and respond to the native cardiac environment.
cardiomyocytes and transplanted fetal cardiomyocytes.51 Using Thus, consideration of the donor cells’ functional phenotype
an alternative imaging approach in which genetically engineered requires not only consideration of the cell preparation used for
mouse cardiomyocytes expressing a fluorescent calcium indicator transplantation, but also the resulting cellular phenotype in the
protein (GCaMP2) were transplanted to the post–myocardial cardiac graft. In addition to the presence of new cardiomyocytes
infarction mouse heart, intracellular Ca2+ transients in the trans- in the grafts, cell fusion of donor cells with native heart cells
planted cells entrained with the native cardiac electrical rhythm, could alter the properties of the native myocardium, which could
indicating coupling of transplanted cells and native heart.49 Fur- have important functional consequences.
thermore, there was an improvement in conduction in the infarct For transplanted cardiomyocytes or cardiomyocytes that dif-
region and a significant decrease in inducible ventricular tachy- ferentiate from transplanted progenitor cells, the functional
cardia in the fetal cardiomyocyte-treated hearts compared with properties can be put into perspective with well-known electro-
sham or myoblast-treated hearts.49 In a myocardial cryoinjury physiological properties seen in healthy and diseased myocar-
mouse model, GFP-labeled fetal cardiomyocytes transplanted to dium. ESC-derived cardiomyocytes in culture exhibit features
the borderzone were able to couple with native myocardium as typical of the embryonic heart with cardiomyocytes showing
determined by sharp microelectrode recordings of labeled cells, spontaneous automaticity, a more depolarized maximum diastolic
although transplanted cells remote from border zone in the area potential, and a reduced upstroke velocity.32 These features all
of cryoinjury showed spontaneous activity and were not coupled have the potential to be proarrhythmic in the transplanted heart
in most cases.52 by leading to areas of abnormal automaticity, depolarizing tissue,
The electrophysiological consequences of transplantation of and slowing conduction, respectively.53 However, there is evi-
ESC-derived cardiomyocytes have also been examined. In one dence that the functional properties of transplanted cardiomyo-
study, transplantation of mouse ESC-derived cardiomyocytes cytes can mature in situ and exhibit more hyperpolarized resting
into the post–myocardial infarction mouse heart resulted in an membrane potentials and more rapid upstroke velocities.52 In
increase in inducible ventricular tachycardia and increased mor- addition to manifesting an embryonic phenotype, it is possible
tality.38 Another study transplanting human ESC-derived cardio- that transplanted cardiomyocytes will exhibit differences in repo-
myocytes into the cryoinjured guinea pig heart suggested a strong larization from the native myocardium, creating areas of disper-
antiarrhythmic effect with a reduction in spontaneous ventricular sion of refractoriness and substrate for reentry. Because
arrhythmias and decreased inducible ventricular tachycardia.37 In repolarization is finely regulated by multiple ion channels
this later study, the human ESC-derived cardiomyocytes formed expressed in the cardiomyocytes, altering this delicate balance is
Cx43-positive gap junctions with native myocardium and using possible. Furthermore, if transplanted cardiomyocytes are injured
the genetically engineered Ca reporter construct (GCaMP3) in the transplantation process or respond adversely to the dis-
functional coupling of the grafted hESC-cardiomyocytes and eased myocardium, they can develop pathologic responses includ-
native heart was demonstrated.37 As a result, the few studies ing calcium overload and the propensity to exhibit delayed
exploring ESC-derived cardiomyocytes for cardiac cell therapy afterdepolarizations and resulting triggered arrhythmias. Despite
all these potential problems, it is encouraging that the slow turn- intramyocardial via both epicardial injection at the time of cardiac
57
over and replacement of cardiomyocytes in the normal heart surgery and via catheter-based endocardial injection. An early
results in functionally matched cardiomyocytes based on the trial delivering skeletal muscle–derived myoblasts at the time of
similar properties of isolated cardiomyocytes from the adult heart coronary artery bypass surgery provided an initial note of caution.
and lack of arrhythmias in normal hearts. Nevertheless, defining In this trial, 4 of 10 patients treated with myoblasts experienced
the functional properties of engrafted cells remains a critical and ventricular tachycardia.57 However, there was not a comparable
rarely investigated feature of cardiac cell therapy. control group, and the patient population had substantial risk for
ventricular arrhythmias. Subsequently, more than 100 phase 1
trials have been completed around the world using a wide variety
Other Cardiac Tissue Effects of Cell Therapy of cell sources; they have failed to find major adverse effects,
With Arrhythmia Relevance including arrhythmias. Different patient populations have been
studied manifesting a range of heart diseases; however, the major-
Cells delivered to the diseased myocardium can exert a number ity of effort has focused on ischemic heart disease, given the
of additional effects that can more generally affect the electro- prevalence and burden of this disease.
physiological status of the heart. These effects can be related to The current best understanding of the clinical risk of arrhyth-
the effect of cell therapy on the underlying heart disease patho- mias comes from the handful of phase 2 clinical trials shown in
physiology, which secondarily affects arrhythmia risk. Secreted Table 57-1. These trials have primarily involved two patient
molecules by the transplanted cells can have powerful paracrine populations with ischemic heart disease: post–myocardial infarc-
effects on the heart. Likewise, cell therapy can affect the status tion patients following percutaneous revascularization of the
of autonomic innervation heart, the burden of ischemia, and infarct-related artery and patients with chronic ischemic heart
cardiac synchronization. disease not in the peri-infarct period. Although both patient
In the setting of the post–myocardial infarction heart, perhaps populations reflect pathology secondary to coronary artery
the most critical effect of cell therapy is the ability to blunt disease, the state of the myocardium receiving the cells is rather
further adverse remodeling of the myocardium, which can prog- different, and the delivery approaches required are distinct.
ress to advancing LV dysfunction and heart failure with ever- Therefore, the effects of the cell therapy on arrhythmia risk could
increasing risk of life-threatening ventricular arrhythmias. For be quite different.
many of the cell sources tested to date, the inhibition of adverse
remodeling has been proposed to be due to paracrine effects of
the transplanted cells. Potential mechanisms for this beneficial Post–Myocardial Infarction Clinical Trials
effect include the ability of some stem cell sources to modulate of Cell Therapy
the inflammation present after myocardial infarction, thus
improving the remodeling process. Others have suggested that Cell therapy to treat patients after myocardial infarction is based
some stem cells can activate endogenous cardiac stem cells and on the premise that the acutely infarcted myocardium is uniquely
enhance intrinsic cardiac repair.54 Regardless, if the net result of amenable to repair and to interventions that can blunt or block
paracrine signaling is to blunt remodeling and prevent progres- chronic adverse remodeling and the progression to heart failure.
sion to heart failure, then this will have a long-term antiarrhyth- Following a myocardial infarction, the myocardium undergoes
mic effect on treated hearts relative to untreated. dynamic remodeling in multiple phases consisting of an acute
Cell therapies that successfully lead to neovascularization and inflammatory stage, followed by gradual replacement fibrosis.
repair of the heart will reduce the ischemia experienced by the The guiding concept for these cell therapy trials is that it will be
myocardium. Eliminating or reducing ischemia as a trigger for more productive and feasible to intervene in the actively remod-
arrhythmias can be of obvious benefit. Alternatively, if cell deliv- eling heart after myocardial infarction than treating myocardium
ery by the intracoronary route exacerbates ischemia, this could with chronic fibrotic scars; however, the post–myocardial infarc-
have an adverse effect on arrhythmia risk. tion heart provides fertile substrate for ventricular arrhythmias.
Delivery of MSCs to the injured heart has been associated In these trials, primary angioplasty for acute revascularization of
with areas of cardiac nerve sprouting.55 The generation of the myocardium was performed, and the patients subsequently
areas of nerve sprouting has previously been associated with an were randomized to undergo bone marrow harvest of cells with
increased risk of sudden cardiac death56; therefore, investigators either bone marrow mononuclear cells or placebo delivered to
have speculated that MSC cell therapy might generate this the myocardium. Cell delivery follows the straightforward
risk. approach of intracoronary catheter–based delivery via the infarct-
Cardiac resynchronization has become a common therapy to related artery. Thus, the cells are delivered to the infarct bed,
treat heart failure, especially in the presence of impaired conduc- although depending on the extent of reperfusion, how broadly
tion (QRS duration > 120 ms). The clinically applied form of these cells are deposited and how they distribute in the infarct is
resynchronization involves a biventricular pacemaker, but cell not currently well described, nor is the retention and survival of
therapy may be even more effective in its resynchronization, these cells in the myocardium well understood. The clinical out-
depending on the proper integration and function of the grafts. comes of this approach have varied among the trials. One promi-
Potential benefits from this effect might not be evident immedi- nent trial, REPAIR-AMI, showed an improvement in ejection
ately, but they could manifest in longer-term studies. fraction, a reduction in the combined clinical endpoint of death,
and recurrence of myocardial infarction or revascularization;
however, a significant effect on ejection fraction was not observed
in other trials (see Table 57-1). Questions have arisen regarding
Clinical Experience With Cell Therapy for differences in efficacy that have focused on cell preparation pro-
Ischemic Heart Disease and Arrhythmias cedures including the use of heparin, which has been proposed
to blunt the beneficial effects. In addition, the timing of cell
Starting in 2001, clinical trials evaluating the effects of cell delivery after myocardial infarction has been suggested to be a
therapy for various forms of heart disease have been performed. critical variable. In all these trials, the incidence of arrhythmias
The initial phase 1 trials have tested a wide range of primarily has been relatively low, and no significant difference was observed
autologous cell products. Cell delivery methods have varied, between the placebo and cell-treated patients. A metaanalysis of
including catheter-based intracoronary, intravenous, and these trials also failed to detect any difference in arrhythmias
Table 57-1. Phase 2 Randomized Placebo-controlled Cell Therapy Trials for Ischemic Heart Disease
Cell Preparation
Trial Patients Randomized and Dose Delivery Route Arrhythmic Events Primary Outcome
Post Acute MI Revascularization Trials
ASTAMI 100 patients with Autologous BMCs, Median 6 days Ventricular arrhythmias occurred No effect on LV
2006 acute anterior STEMI mean 68 × 106 post-MI by IC in 2 BMC patients and 1 control function (EF) or infarct
and PCI patient. No deaths due to size measured at 6 mo
arrhythmia measured by echo,
SPECT, and MRI
REPAIR- 204 patients with Autologous BMCs, 3-7 days post-MI Documented ventricular Improvement in EF at
AMI* 2006 acute STEMI and PCI mean dose not by IC arrhythmia or syncope in 5 4 months, reduction in
with EF < 45% stated BMC-treated and 5 in control at combined end point of
4 months. 6 deaths in placebo death, recurrence of MI
and 2 in BMC or any revascularization
procedures at 1 and
2 year
Janssens 67 patients Autologous BMCs, 7 days post-MI No differences in treatment- No improvement in EF
2006 randomized with mean 172 × 106 cells by IC related tachyarrhythmia at 4 months by MRI but
STEMI and PCI on Holter monitoring reduction in infarct size
(supraventricular and better regional
arrhythmia, n = 6 control and n = systolic function
5 BMSC; non-sustained
ventricular tachycardia,
n = 3 control, n = 0 BMCs)
BONAMI* 101 patients with Autologous BMCs, 7-10 days 2 arrhythmias for each group Improvement in
2010 STEMI and PCI with EF mean 98 × 106 cells post-MI by IC but not described. myocardial viability in
< 45%, decrease 1 sudden death in BMC and multivariate analysis at
viability on SPECT none in control 3 months but no
change in EF
Late TIME* 87 patients, post 1st Autologous BMCs, 2-3 week post MI 2 ICD placements in placebo No significant change
2011 STEMI and PCI with EF 150 × 106 cells by IC none in BMC in EF, LV volumes, or
< 45% infarct size at 6 months
Chronic Ischemic Heart Disease
MAGIC* 120 patients, Autologous skeletal At time of CABG All patients had ICD in place, At 6 mo, no
2008 EF < 35%, CABG myoblasts, low dose 27 g needle IM at trend for more arrhythmias improvement in
candidate, MI > 4wk 400 × 106 or high 30 epicardial post-op for cell group but levels echocardiographic EF,
prior dose 800 × 106 sites around out over 24 mo so no difference but was decrease in LV
akinetic area in incidence, no death from volumes in high dose
arrhythmia, no difference in use
of amiodarone
ACT34- 167 patients with C-GSF mobilization Electroanatomic No sudden cardiac death, no At 6 months lower
CMI* 2011 refractory angina not and apheresis for mapping with IM arrhythmias noted angina frequency and
amenable to CD34+ cells, low catheter-based increased exercise
revascularization dose 1 × 105/kg or delivery tolerance in low dose
high dose 5 × 105/kg but not high dose
group
FOCUS- 92 patients with Autologous BMCs Electroanatomic No sudden cardiac death, no No improvement in
CCTRN* chronic ischemic delivered IM mapping with IM arrhythmia events noted LVESV index, maximal
2012 disease not amenable 100 × 106 cells catheter-based O2 consumption or
to revascularization delivery SPECT reversibility
with symptoms,
EF < 45%
*Multicenter trial.
BMCs, Bone marrow mononuclear cells; CABG, coronary artery bypass grafting surgery; EF, ejection fraction; IC, intracoronary; IM, intramuscular; LVESV, left ventricular end
systolic volume; PCI, percutaneous coronary intervention; STEMI, ST elevation myocardial infarction.
1. Lunde K, Solheim S, Aakhus S, et al: Intracoronary injection of mononuclear bone marrow cells in acute myocardial infarction. N Engl J Med 355:1199–1209, 2006.
2. Schachinger V, Erbs S, Elsasser A, et al: Intracoronary bone marrow-derived progenitor cells in acute myocardial infarction. N Engl J Med 355:1210–1221, 2006.
3. Janssens S, Theunissen K, Boogaerts M, et al: Bone marrow cell transfer in acute myocardial infarction. Nat Clin Pract Cardiovasc Med 3 Suppl 1:S69–72, 2006.
4. Roncalli J, Mouquet F, Piot C, et al: Intracoronary autologous mononucleated bone marrow cell infusion for acute myocardial infarction: Results of the randomized
multicenter bonami trial. European Heart Journal 32:1748–1757, 2011.
5. Traverse JH, Henry TD, Ellis SG, et al: Effect of intracoronary delivery of autologous bone marrow mononuclear cells 2 to 3 weeks following acute myocardial infarction
on left ventricular function: The latetime randomized trial. Jama 306:2110–2119, 2011.
6. Menasche P, Alfieri O, Janssens S, et al: The myoblast autologous grafting in ischemic cardiomyopathy (magic) trial: First randomized placebo-controlled study of
myoblast transplantation. Circulation 117:1189–1200, 2008.
7. Losordo DW, Henry TD, Davidson C, et al: Intramyocardial, autologous cd34+ cell therapy for refractory angina. Circ Res 109:428–436, 2011.
8. Perin EC, Willerson JT, Pepine CJ, et al: Effect of transendocardial delivery of autologous bone marrow mononuclear cells on functional capacity, left ventricular
function, and perfusion in chronic heart failure: The focus-cctrn trial. Jama 307:1717–1726, 2012.
between placebo and cell-treated groups.58 These results are

Conclusions
57
encouraging but do not exclude rare, potentially lethal ventricu-
lar arrhythmias when larger patient populations are treated.
Phase 3 trials are underway in post–myocardial infarction patients Cell therapy is of growing interest in the treatment of various
and will advance current understanding. forms of heart disease, including post–myocardial infarction and
heart failure. It is part of a larger advancing field of regenerative
medicine. Basic research and studies in animal models have sug-
Chronic Ischemic Heart Disease Clinical Trials gested that the effect of cell therapies on the heart is complex
of Cell Therapy and dependent on the details of cell delivery methodology, donor
cell types, and underlying cardiac disease. Multiple mechanisms
In trials testing the effect of cell therapy on patients with chronic of benefit have been identified, including a range of paracrine
ischemic heart disease, the patient populations, cell preparations, effects, mechanical effects, and rarely generating new cardiomyo-
and approaches are more heterogenous than the post–myocardial cytes. The experimental studies have also identified multiple
infarction trials. Because patients with chronic ischemic heart potentials ways in which cell therapy can be proarrhythmic or
disease do not have robust coronary perfusion to the area of dis- antiarrhythmic. Regardless, the appealing concept of repairing or
eased myocardium, targeted intramyocardial injections typically regenerating myocardium with stem cells has led to the rapid
have been the delivery method of choice. Targeting the delivery translation from the research laboratory to clinical trials. Several
of cells has focused on areas of viable but at-risk myocardium. The different primarily autologous cell preparations have been deliv-
MAGIC trial studied patients undergoing coronary artery bypass ered via the intracoronary or intramyocardial route in patients
grafting surgery with an ejection fraction less than 35% who had with coronary artery disease in phase 1 and 2 trials. The initial
a previous but remote myocardial infarction. Autologous skeletal results from the trials have been variable and show either an
myoblasts were injected epicardially around the echocardiograph- improvement primarily in ejection fraction or no effect. The
ically determined area of akinesis, and there was no improvement trials have not identified safety concerns, including a lack of
in the primary outcome of ejection fraction. There was substantial proarrhythmia by the cell therapy. Phase 3 trials are underway;
concern about arrhythmias based on the earlier phase 1 trial; therefore, more definitive clinical data will be available in the next
therefore, all patients received an implantable cardiodefibrillator few years.
(ICD) for protection. However, after 24 months there was no Future approaches for cardiac repair will undoubtedly evolve.
significant difference in the incidence of arrhythmias between the The patient populations most amenable to this therapy are not
two groups, although there was a nonsignificant trend of more known, nor is the optimal cell type for each disease known. It is
arrhythmias after surgery in the myoblast group. The two other likely that different cell preparations will show differences in
phase 2 multicenter trials intervening on patients with chronic utility depending on the underlying cardiac disease. Will geneti-
ischemic heart disease have focused on patients who have no pos- cally engineered cells be preferable for expressing proteins that
sible revascularization options left, but have persistent symptoms help to couple, promote targeting to the area of need, or optimize
of angina or heart failure. These trials have used electroanatomic paracrine signaling? Tissue engineering applications currently
mapping of the myocardium to determine areas of viable (electri- under investigation could enable new delivery strategies using
cal signal) but noncontracting (lack mechanical signal) referred to viable patches of tissue.
as hibernating myocardium, to which the cell delivery is targeted by Understanding the risk of arrhythmias with the ongoing cell
catheter-based endocardial injection. The cell sources have dif- therapy approaches will continue to be of utmost importance to
fered with the ACT34-CMI trial using peripheral blood mobi- ensure that these therapies are safe and effective. Arrhythmia risk
lized CD34-positive cells for therapy while the FOCUS-CCTRN can be dynamic, such as an increase in risk early following cell
trial used bone marrow mononuclear cells. The ACT34-CMI delivery and potentially a decrease later. In some cases, it is useful
trial showed significant improvements in anginal symptoms and to at least transiently treat the patient with antiarrhythmic drugs
exercise capacity, but the FOCUS-CCTRN trial did not find any to lower the risk of arrhythmias. More detailed monitoring for
significant effect on end points such as maximal oxygen consump- arrhythmias during the trials is needed to clarify risk and poten-
tion. Regarding arrhythmias in these no-options patients, there tially to define one form of benefit—reduction in arrhythmias.
were no differences reported between cell-treated patients and Routine surveillance for arrhythmias is reasonable using Holter
placebo, and no sudden cardiac deaths were reported in either monitors or implantable event recorders. For patients with ICDs,
study. Thus, the efficacy of cellular therapy in this patient popula- regular interrogations and data analysis before and after therapy
tion remains unclear, but the interventions appear safe without may be useful. The optimal duration for monitoring is unknown.
any evidence for a proarrhythmic effect. Differences in cell types, Consideration of the electrophysiologic effects of cell therapy to
delivery methods, and patient populations raise many questions the diseased myocardium will continue to be essential to advance
requiring future study. this revolutionary new form of treatment.
Homing, differentiation, and fusion after grafts in heart. J Clin Invest 92(3):1548–1554,
References infarction. Proc Natl Acad Sci U S A 100(21):12313– 1993.
12318, 2003. 9. Taylor DA, Atkins BZ, Hungspreugs P, et al:
1. Kajstura J, Gurusamy N, Ogorek B, et al: Myocyte 5. Laugwitz KL, Moretti A, Lam J, et al: Postnatal Regenerating functional myocardium: Improved
turnover in the aging human heart. Circ Res isl1+ cardioblasts enter fully differentiated cardio- performance after skeletal myoblast transplanta-
107(11):1374–1386, 2010. myocyte lineages. Nature 433(7026):647–653, tion. Nat Med 4(8):929–933, 1998.
2. Bergmann O, Bhardwaj RD, Bernard S, et al: Evi- 2005. 10. Scorsin M, Hagege A, Vilquin JT, et al: Compari-
dence for cardiomyocyte renewal in humans. 6. Dimmeler S, Leri A: Aging and disease as modifiers son of the effects of fetal cardiomyocyte and skel-
Science 324(5923):98–102, 2009. of efficacy of cell therapy. Circ Res 102(11):1319– etal myoblast transplantation on postinfarction left
3. Beltrami AP, Barlucchi L, Torella D, et al: Adult 1330, 2008. ventricular function. J Thorac Cardiovasc Surg
cardiac stem cells are multipotent and support 7. Marelli D, Desrosiers C, el Alfy M, et al: Cell trans- 119(6):1169–1175, 2000.
myocardial regeneration. Cell 114(6):763–776, plantation for myocardial repair: An experimental 11. Reinecke H, MacDonald GH, Hauschka SD, et al:
2003. approach. Cell Transplant 1(6):383–390, 1992. Electromechanical coupling between skeletal and
4. Oh H, Bradfute SB, Gallardo TD, et al: 8. Koh GY, Klug MG, Soonpaa MH, et al: Differen- cardiac muscle. Implications for infarct repair.
Cardiac progenitor cells from adult myocardium: tiation and long-term survival of c2c12 myoblast J Cell Biol 149(3):731–740, 2000.
12. Soonpaa MH, Koh GY, Klug MG, et al: Formation 28. Pfister O, Oikonomopoulos A, Sereti KI, et al: from human embryonic stem cells. Nat Biotechnol
of nascent intercalated disks between grafted fetal Role of the atp-binding cassette transporter abcg2 22(10):1282–1289, 2004.
cardiomyocytes and host myocardium. Science in the phenotype and function of cardiac side popu- 44. Valiunas V, Doronin S, Valiuniene L, et al: Human
264(5155):98–101, 1994. lation cells. Circ Res 103(8):825–835, 2008. mesenchymal stem cells make cardiac connexins
13. Scorsin M, Marotte F, Sabri A, et al: Can grafted 29. Smith RR, Barile L, Cho HC, et al: Regenerative and form functional gap junctions. J Physiol 555(Pt
cardiomyocytes colonize peri-infarct myocardial potential of cardiosphere-derived cells expanded 3):617–626, 2004.
areas? Circulation 94(9 Suppl):II337–II40, 1996. from percutaneous endomyocardial biopsy speci- 45. Chang MG, Tung L, Sekar RB, et al: Proarrhyth-
14. Li RK, Jia ZQ, Weisel RD, et al: Cardiomyocyte mens. Circulation 115(7):896–908, 2007. mic potential of mesenchymal stem cell transplan-
transplantation improves heart function. Ann 30. Thomson JA, Itskovitz-Eldor J, Shapiro SS, et al: tation revealed in an in vitro coculture model.
Thorac Surg 62(3):654–660, 1996. Embryonic stem cell lines derived from human Circulation 113(15):1832–1841, 2006.
15. Orlic D, Kajstura J, Chimenti S, et al: Bone blastocysts. Science 282(5391):1145–1147, 1998. 46. Abraham MR, Henrikson CA, Tung L, et al: Anti-
marrow cells regenerate infarcted myocardium. 31. Kehat I, Kenyagin-Karsenti D, Snir M, et al: arrhythmic engineering of skeletal myoblasts for
Nature 410(6829):701–705, 2001. Human embryonic stem cells can differentiate into cardiac transplantation. Circ Res 97(2):159–167,
16. Balsam LB, Wagers AJ, Christensen JL, et al: Hae- myocytes with structural and functional properties 2005.
matopoietic stem cells adopt mature haematopoi- of cardiomyocytes. J Clin Invest 108(3):407–414, 47. Leobon B, Garcin I, Menasche P, et al: Myoblasts
etic fates in ischaemic myocardium. Nature 2001. transplanted into rat infarcted myocardium are
428(6983):668–673, 2004. 32. He JQ, Ma Y, Lee Y, et al: Human embryonic stem functionally isolated from their host. Proc Natl
17. Murry CE, Soonpaa MH, Reinecke H, et al: Hae- cells develop into multiple types of cardiac myo- Acad Sci U S A 100(13):7808–7811, 2003.
matopoietic stem cells do not transdifferentiate cytes: Action potential characterization. Circ Res 48. Rubart M, Soonpaa MH, Nakajima H, et al: Spon-
into cardiac myocytes in myocardial infarcts. 93(1):32–39, 2003. taneous and evoked intracellular calcium transients
Nature 428(6983):664–668, 2004. 33. Behfar A, Zingman LV, Hodgson DM, et al: Stem in donor-derived myocytes following intracardiac
18. Toma C, Pittenger MF, Cahill KS, et al: Human cell differentiation requires a paracrine pathway in myoblast transplantation. J Clin Invest 114(6):775–
mesenchymal stem cells differentiate to a cardio- the heart. FASEB J 16(12):1558–1566, 2002. 783, 2004.
myocyte phenotype in the adult murine heart. Cir- 34. Min JY, Yang Y, Converso KL, et al: Transplanta- 49. Roell W, Lewalter T, Sasse P, et al: Engraftment of
culation 105(1):93–98, 2002. tion of embryonic stem cells improves cardiac connexin 43-expressing cells prevents post-infarct
19. Mirotsou M, Zhang Z, Deb A, et al: Secreted function in postinfarcted rats. J Appl Physiol arrhythmia. Nature 450(7171):819–824, 2007.
frizzled related protein 2 (sfrp2) is the key 92(1):288–296, 2002. 50. Mills WR, Mal N, Kiedrowski MJ, et al: Stem cell
akt-mesenchymal stem cell-released paracrine 35. Singla DK, Hacker TA, Ma L, et al: Transplanta- therapy enhances electrical viability in myocardial
factor mediating myocardial survival and repair. tion of embryonic stem cells into the infarcted infarction. J Mol Cell Cardiol 42(2):304–314,
Proc Natl Acad Sci U S A 104(5):1643–1648, mouse heart: Formation of multiple cell types. J 2007.
2007. Mol Cell Cardiol 40(1):195–200, 2006. 51. Rubart M, Pasumarthi KB, Nakajima H, et al:
20. Hierlihy AM, Seale P, Lobe CG, et al: The post- 36. Nussbaum J, Minami E, Laflamme MA, et al: Physiological coupling of donor and host cardio-
natal heart contains a myocardial stem cell popula- Transplantation of undifferentiated murine embry- myocytes after cellular transplantation. Circ Res
tion. FEBS Lett 530(1-3):239–243, 2002. onic stem cells in the heart: Teratoma formation 92(11):1217–1224, 2003.
21. Beltrami AP, Barlucchi L, Torella D, et al: Adult and immune response. FASEB J 21(7):1345–1357, 52. Halbach M, Pfannkuche K, Pillekamp F, et al:
cardiac stem cells are multipotent and support 2007. Electrophysiological maturation and integration of
myocardial regeneration. Cell 114(6):763–776, 37. Shiba Y, Fernandes S, Zhu WZ, et al: Human es- murine fetal cardiomyocytes after transplantation.
2003. cell-derived cardiomyocytes electrically couple and Circ Res 101(5):484–492, 2007.
22. Kubo H, Jaleel N, Kumarapeli A, et al: Increased suppress arrhythmias in injured hearts. Nature 53. Zhang YM, Hartzell C, Narlow M, et al: Stem
cardiac myocyte progenitors in failing human 2012. cell-derived cardiomyocytes demonstrate arrhyth-
hearts. Circulation 118(6):649–657, 2008. 38. Liao SY, Liu Y, Siu CW, et al: Proarrhythmic risk mic potential. Circulation 106(10):1294–1299,
23. Bu L, Jiang X, Martin-Puig S, et al: Human isl1 of embryonic stem cell-derived cardiomyocyte 2002.
heart progenitors generate diverse multipotent car- transplantation in infarcted myocardium. Heart 54. Hatzistergos KE, Quevedo H, Oskouei BN, et al:
diovascular cell lineages. Nature 460(7251):113– Rhythm 7(12):1852–1859, 2010. Bone marrow mesenchymal stem cells stimulate
117, 2009. 39. Takahashi K, Tanabe K, Ohnuki M, et al: Induction cardiac stem cell proliferation and differentiation.
24. Messina E, De Angelis L, Frati G, et al: Isolation of pluripotent stem cells from adult human fibro- Circ Res 107(7):913–922, 2010.
and expansion of adult cardiac stem cells from blasts by defined factors. Cell 131(5):861–872, 55. Pak HN, Qayyum M, Kim DT, et al: Mesenchymal
human and murine heart. Circ Res 95(9):911–921, 2007. stem cell injection induces cardiac nerve sprouting
2004. 40. Yu J, Vodyanik MA, Smuga-Otto K, et al: Induced and increased tenascin expression in a swine model
25. Chimenti I, Smith RR, Li TS, et al: Relative roles pluripotent stem cell lines derived from human of myocardial infarction. J Cardiovasc Electro-
of direct regeneration versus paracrine effects of somatic cells. Science 318(5858):1917–1920, 2007. physiol 14(8):841–848, 2003.
human cardiosphere-derived cells transplanted 41. Nelson TJ, Martinez-Fernandez A, Yamada S, et 56. Cao JM, Chen LS, KenKnight BH, et al: Nerve
into infarcted mice. Circ Res 106(5):971–980, al: Repair of acute myocardial infarction by human sprouting and sudden cardiac death. Circ Res
2010. stemness factors induced pluripotent stem cells. 86(7):816–821, 2000.
26. Zhou B, Ma Q, Rajagopal S, et al: Epicardial pro- Circulation 120(5):408–416, 2009. 57. Menasche P, Hagege AA, Vilquin JT, et al: Autolo-
genitors contribute to the cardiomyocyte lineage in 42. Mangi AA, Noiseux N, Kong D, et al: Mesenchy- gous skeletal myoblast transplantation for severe
the developing heart. Nature 454(7200):109–113, mal stem cells modified with akt prevent remodel- postinfarction left ventricular dysfunction. J Am
2008. ing and restore performance of infarcted hearts. Coll Cardiol 41(7):1078–1083, 2003.
27. Cai CL, Martin JC, Sun Y, et al: A myocardial Nat Med 9(9):1195–1201, 2003. 58. Clifford DM, Fisher SA, Brunskill SJ, et al: Stem
lineage derives from tbx18 epicardial cells. Nature 43. Kehat I, Khimovich L, Caspi O, et al: Electrome- cell treatment for acute myocardial infarction.
454(7200):104–108, 2008. chanical integration of cardiomyocytes derived Cochrane Database Syst Rev 2:(CD006536), 2012.
Diagnostic Evaluation PART X
Assessment of the Patient

With a Cardiac Arrhythmia 58
Mithilesh K. Das and Douglas P. Zipes
Many patients are acutely aware of any cardiac irregularity,

CHAPTER OUTLINE whereas others are oblivious even to long runs of a rapid ven-
History Taking 567 tricular tachycardia or atrial fibrillation with rapid ventricular
rate. Often the asymptomatic patients are those referred for
Physical Examination 569 evaluation of an arrhythmia noted incidentally during assessment
Laboratory Tests 570 for another reason, such as a preathletic physical examination in
a child or adolescent, a preinsurance physical examination in an
Summary 573 adult, or a routine preoperative assessment. Patients describe
these symptoms in various ways. Most frequently, they use terms
such as a thumping or flip-flopping sensation in the chest; a fullness
The evaluation of a patient with a suspected cardiac rhythm in the throat, neck, or chest; or a pause in the heart beat, “as if
disturbance is fundamental to the role of the clinical cardiac my heart stopped or skipped a beat.” The last is most likely
electrophysiologist. The approach followed for this evaluation caused by the compensatory pause after a premature ventricular
varies from patient to patient and is influenced by the patient’s complex (PVC) or the resetting of sinus rhythm after a premature
clinical status and symptoms, but a general outline can be estab- atrial complex. Presumably, the premature beat, particularly if it
lished, as presented in this chapter. As always, the initial evalua- is a ventricular extrasystole, occurs too early to permit sufficient
tion begins with a careful history and physical examination. ventricular filling to cause a sensation when the ventricle con-
tracts. The ventricular systole that ends the compensatory pause
may be responsible for the actual palpitation and is caused by a
more forceful contraction from prolonged ventricular filling or
History Taking increased motion of the heart in the chest. Anxiety over such
symptoms is commonly the complaint that brings the patient to
Significant overlap exists among the clinical features of various the physician’s office.
rhythm disturbances, imparting a degree of imprecision to the
interpretation of the patient’s history. Despite this drawback, Skipped Beats Versus Sustained Palpitation
the history often can provide direction and diagnostic clues as the Premature atrial or ventricular complexes probably constitute the
first step in assessing the patient with, or suspected of having, a most common cause of palpitations, and patients often use the
cardiac arrhythmia. It often is the most important source of term skipped beat or dropped beat to describe them. If the prema-
information about the arrhythmia. ture complexes are frequent or particularly if a sustained tachy-
cardia is present, patients are more likely to complain of
lightheadedness, syncope or near-syncope, chest pain, fatigue, or
Symptoms and Signs shortness of breath. The presence of associated cardiovascular
problems influences the nature of the symptoms. For example, a
Major symptoms and signs of cardiac arrhythmias are palpita- supraventricular tachycardia at a rate of 180 beats/min can
tions, presyncope, syncope, and sudden cardiac death (SCD). In provoke chest pain in a patient with coronary artery disease or
this setting, nonspecific symptoms such as shortness of breath, syncope in a patient with aortic stenosis, but result in only a
weakness, and fatigue can be due to compromise in cardiac output breathless feeling in an otherwise healthy young person.
and prolonged duration of the arrhythmia or its rate, either very An important point is that patients with ventricular tachycar-
fast or very slow. Older patients with bradycardia owing to sinus dia (VT), particularly young, otherwise healthy persons, can be
node dysfunction or atrioventricular (AV) nodal block can present completely asymptomatic or experience minimal symptoms
with altered mental status and dementia. during the arrhythmic episode. The lack of significant symptoms
should not exclude the diagnosis of VT. Bradyarrhythmias have
Palpitations their own constellation of symptoms that usually includes
Awareness of an irregular heartbeat varies greatly from patient to syncope, near-syncope, and fatigue.
patient. Patients who complain of symptoms most commonly In this fashion, the clinician can obtain information about the
note palpitations, defined as sensations experienced as an unpleas- nature of the beginning and end of the tachycardia, whether
ant awareness of forceful, irregular, or rapid beating of the heart. the ventricular rhythm is regular or irregular, and the rate of the
567
568 DIAGNOSTIC EVALUATION
tachycardia. Knowledge about the typical onset and termination

Box 58-1 Differential Diagnosis of Palpitations
of the tachycardia is helpful. Abrupt, paroxysmal onset is consis-
tent with a tachycardia such as AV nodal reentrant tachycardia Cardiac Arrhythmias
(AVNRT; see Chapter 77), whereas gradual speeding and slowing Sinus tachycardia
are more in keeping with a sinus tachycardia (see Chapter 72). • Physiologic
Termination by Valsalva maneuver or carotid sinus massage sug- • Inappropriate sinus tachycardia
gests a tachycardia incorporating nodal tissue in the reentrant
• Postural orthostatic tachycardia syndrome
pathway, such as sinus node reentry, AVNRT, or AV reentrant
tachycardia (AVRT; see Chapters 77 and 76), and idiopathic right • Anxiety disorders, thyrotoxicosis, perimenopausal
ventricular outflow tract tachycardia. It often is helpful to have syndrome, pheochromocytoma
the patient tap out the cadence of the perceived palpitations, from Atrial arrhythmias
onset to termination. • Atrial premature complexes
The rate of the untreated tachycardia often narrows diagnos- • Atrial fibrillation, atrial flutter, atrial tachycardia
tic possibilities, and patients should be taught to count their Supraventricular tachycardia
radial or carotid pulse rate. Ventricular rates of 150 beats/ • Atrioventricular nodal reentry tachycardia, orthodromic
minute (bpm) should always suggest the potential diagnosis of atrioventricular reciprocating tachycardia in Wolff-
atrial flutter with 2 : 1 AV block (see Chapter 74), whereas most Parkinson-White syndrome
supraventricular tachycardias, such as those caused by AVNRT
• Permanent form of junctional reciprocating tachycardia
or AVRT, usually occur at rates exceeding 150 bpm. The rates of
VTs overlap those of the supraventricular tachycardias. Palpita- Junctional Tachycardia
tions, hot flashes, and sweating in middle-aged women suggest Ventricular tachycardia
perimenopausal syndrome. Palpitations, dizziness, and shortness • Ventricular premature complexes, ventricular couplets,
of breath on mild exertion, typically in young women with struc- nonsustained ventricular tachycardia
turally normal hearts, suggest the syndrome of inappropriate • Idiopathic
sinus tachycardia. Palpitations owing to sinus tachycardia on
• Caffeine intake
standing should point toward postural hypotension. Palpitations
and presyncope on standing can be symptoms of postural ortho- • Drug-induced: cocaine, QT-prolonging drugs
static tachycardia syndrome. Various possible causes of palpita- • Alcohol
tions are listed in Box 58-1. • Sustained ventricular tachycardia
• Monomorphic VT
Associated Cardiac or Systemic Diseases • Polymorphic VT and torsades de pointes
It also is important to establish whether the patient has structural Conduction system disease
heart disease and, if so, the diagnosis and extent of disease. • Sinus bradycardia
Certain clinical diagnoses are linked to the presence of specific • Tachycardia-bradycardia syndrome
arrhythmias. For example, the occurrence of mitral stenosis • Heart block
should suggest the possibility of atrial fibrillation, whereas a
• Pause-dependent torsades de pointes
history of a myocardial infarction or tetralogy of Fallot repair
invokes VT as a distinct prospect. Thyrotoxicosis should suggest Familial arrhythmia syndrome
atrial arrhythmias, including sinus tachycardia. At times it is • Long QT syndrome: polymorphic VT
useful to search for a family history of similar problems and to • Short QT syndrome: polymorphic VT, AF
obtain electrocardiograms (ECGs) of close family members, such • Catecholaminergic polymorphic ventricular tachycardia:
as parents, siblings, or children. Family history of palpitations, polymorphic VT, AT/AF
syncope, or SCD should be investigated carefully for inherited • Brugada syndrome: polymorphic VT, AF
cardiac arrhythmias, including atrial fibrillation, long QT syn- • Inherited cardiomyopathy with ventricular arrhythmias,
drome, short QT syndrome, catecholaminergic polymorphic VT, such as right ventricular cardiomyopathy: monomorphic
arrhythmogenic right ventricular dysplasia or cardiomyopathy and polymorphic VT
(ARVD/C), and inherited cardiomyopathy with arrhythmia.
• Pacemaker-mediated tachycardia, intermittent ventricular
Presyncope and Syncope pacing
The diagnosis of presyncope and syncope and its cause requires Metabolic syndromes
comprehensive history taking from the patient and witness. The • Hypoglycemia
differential diagnosis of syncope is lengthy and can be a warning • Electrolyte imbalance
sign of SCD (Chapter 99, Table 99-1). It is important to differenti- Structural heart disease
ate cardiac versus noncardiac causes of syncope. It is more impor- • Valvular disease
tant to differentiate a benign cause of syncope from a malignant • Cardiomyopathies
cause. Of the reflex syncopes (neurocardiogenic, carotid hypersen- • Primary pulmonary hypertension
sitivity, and situational), neurocardiogenic is the most common. It
should be differentiated from syncope owing to orthostasis, which Drug-Induced
is commonly seen in autonomic failure (e.g., due to diabetes), and • Proarrhythmia: antiarrhythmic drugs, psychotropic drugs
from syncope resulting from other cardiac causes. • Sympathomimetic agents, vasodilators, anticholinergic
When caused by a cardiac arrhythmia, onset of syncope is drugs, β-blocker withdrawal, amphetamine and other
rapid and duration is brief, with or without preceding aura, and antianxiety drugs, nicotine, caffeine
usually is not followed by a postictal confusional state. It can be
associated with bodily injury if the patient falls while unconscious. VT, Ventricular tachycardia; AF, atrial fibrillation.
Palpitations preceding syncope also support an arrhythmic cause
of syncope. Seizure activity is uncommon and occurs mostly after
a prolonged asystole. Therefore, the seizure does not begin with
the syncope. However, in epileptic seizures, convulsive
Assessment of the Patient With a Cardiac Arrhythmia 569
movements start within seconds of the onset of syncope. Tongue occur unpredictably, even in high-risk patients. Structural heart
58
biting or incontinence is also uncommon in cardiac syncope. The disease, such as coronary artery disease, cardiomyopathy, and
history of syncope should be elicited and interpreted carefully congenital heart disease, is responsible for up to 65% to 80%
because older people who have fallen might deny loss of con- cases of SCD. Approximately 5% to 10% of SCDs occur in
sciousness during the event because of brief retrograde amnesia. people with primary electrical abnormalities of the heart, such as
With vasodepressor and cardioinhibitory syncope, the episode long QT syndrome, Brugada syndrome, idiopathic ventricular
usually unfolds more slowly and can be preceded by manifesta- fibrillation, and Wolff-Parkinson-White syndrome. The remain-
tions of autonomic hyperactivity such as nausea, abdominal ing sudden deaths (15% to 20%) are due to noncardiac causes
cramping, diarrhea, sweating, or yawning. On recovery, the such as pulmonary embolism, drugs, drowning, and sudden infant
patient may be bradycardic, pale, sweaty, and fatigued, in contrast death syndrome. Therefore, careful questioning to uncover or
with the patient recovering from a Stokes-Adams attack or an elucidate a family history of SCD is indicated, and family screen-
episode of VT, who could be flushed and have a sinus tachycardia. ing for the suspected cardiac condition should be performed.
Common arrhythmic causes of syncope include bradyarrhyth-
mias caused by sinus node dysfunction or AV block and tachyar-
rhythmias, most often ventricular but also supraventricular on Precipitating Factors
occasion. Bradycardia can follow tachycardia in patients with the
bradycardia-tachycardia syndrome, and treatment of both may be Proarrhythmic Drugs
necessary. Patients usually cannot indicate a specific inciting event,
Drug-induced (orthostatic hypotension, bradyarrhythmia) but the physician should inquire about the use of potentially
and nonarrhythmic cardiac causes such as aortic stenosis, hyper- proarrhythmic drugs such as antiarrhythmic drugs, QT prolong-
trophic cardiomyopathy, pulmonary stenosis, pulmonary hyper- ing drugs, bronchodilators, histamine H1-blocking antihista-
tension, and acute myocardial infarction can be excluded by the mines, some decongestants, psychotropic agents, or other
history, physical examination, ECG, echocardiography, and other over-the-counter drugs, and the ingestion of alcohol or excessive
laboratory tests. Noncardiac causes of syncope such as hypogly- caffeine-containing foods.
cemia, transient ischemic attack, and psychogenic often can be
excluded by a careful history. Exercise, Swimming, Emotions, and Auditory Stimuli
In patients with long QT syndrome, exercise or acute emotional
reactions often precipitate a ventricular arrhythmia in those with
the LQT1 variant, whereas exercise, emotional upset, and sleep
Symptoms of or rest are culprits of fairly equal frequency in LQT2, with audi-
Neurocardiogenic Syncope Signs tory stimuli, more specific precipitators, and events occurring
more frequently during rest in LQT3. Exercise also precipitates
Provoked by prolonged • Unresponsive or semiresponsive various benign and malignant arrhythmias; these include idio-
standing, warm environment, • Facial pallor pathic right ventricular outflow tract tachycardia, exercise-
exercise, emotional stress, sleep • Weak or no pulse transiently induced polymorphic VT, and VT in patients with ARVD/C.
deprivation • Rapid weak pulse during VT Exercise can induce syncope or SCD in patients with outflow
• Dilated pupils tract obstruction and severe pulmonary hypertension.
• Convulsions due to prolonged
AV block or sinus pauses
Premonitory symptoms
• Lightheadedness, dizziness
Physical Examination
• Nausea, epigastric distress
• Palpitations The physical examination offers the opportunity to gain impor-
• Sweating tant information about the presence of associated structural heart
• Weakness disease, if any, and the nature of the arrhythmia, if present.
• Feeling warm or cold Although it is well known that patients with normal hearts can
• Dimming or blurred vision have supraventricular tachycardias, it is less commonly appreci-
ated that patients without recognizable structural heart disease
No premonitory symptoms can also have VTs that on occasion are life threatening. Thus,
reported in patients (mostly in normal results of a physical examination do not preclude the
elderly patients) who present diagnosis of VT, even in a young person. It is likely that at least
trauma associated with some of these patients have structural heart disease that is not
witnessed syncope recognized.
VT, ventricular tachycardia; AV, atrioventricular. The sex of the patient can be a clue to the nature of the
tachycardia. For example, a young woman who complains of
episodes of a regular tachycardia over many years is likely to have
AVNRT. In contrast, a young man with a similar history is more
Sudden Cardiac Death likely to have AVRT associated with the Wolff-Parkinson-White
Sudden cardiac death (SCD) causes approximately 350,000 to syndrome. Symptomatic long QT syndrome is more common in
450,000 deaths annually in the United States. It is responsible for females, whereas Brugada syndrome is more common in males.
nearly 50% of all cardiovascular-related deaths worldwide. SCD
has an incidence of 0.1% to 0.2% per year among adults older
than 35 years. Therefore, careful evaluation of patients who are Atrioventricular Dissociation
resuscitated from SCD is mandatory. SCD occurs in a majority
of patients without known heart disease as the first manifestation If the tachycardia is present during the physical examination, a
of underlying coronary artery disease. The life-threatening ven- 12-lead ECG should be obtained, if time and the patient’s clinical
tricular arrhythmias, such as sustained VT and ventricular fibril- status permit. If an ECG is not possible, a careful physical exami-
lation, are responsible for two thirds of SCDs. These arrhythmias nation can yield helpful findings. For example, the presence of
regular cannon A waves in the jugular venous pulse would be specificity, and predictive accuracy are chosen. A 12-lead ECG is
consistent with a 1 : 1 retrograde ventriculoatrial relation, as in obtained in all patients, and frequently a 24-hour ECG or 30-day
tachycardias such as AVRT, AVNRT, and some junctional tachy- event recording and stress test is helpful in exposing the arrhyth-
cardias and VTs. In contrast, the patient can have physical fea- mia. A chest roentgenogram and an echocardiogram provide
tures indicative of AV dissociation, such as intermittent cannon information about the presence of structural heart disease. The
A waves in the neck, variable intensity of the first heart sound, hierarchy of steps taken to evaluate and treat a patient suspected
and variable peak systolic blood pressure. Common arrhythmic of having an arrhythmia generally proceeds from simple, nonin-
causes of AV dissociation include VT and nonparoxysmal AV vasive, and inexpensive tests to more complex, expensive, and
junctional tachycardia, without retrograde capture of the atria. invasive studies.
Ventricular or junctional tachycardias that produce retrograde The nature of the rhythm disturbance and its effects on the
2 : 1 or Wenckebach block cause intermittent cannon A waves that patient determine the order in which the tests are performed.
recur at regular intervals. Some rhythm disturbances, such as sustained VT or ventricular
fibrillation, are hazardous in and of themselves, whereas others,
such as AVRT or AVNRT, must be evaluated according to the
Atrioventricular Block context in which they occur. AVRT or AVNRT occurring at a
rate of 180 bpm in a young patient who complains only of palpi-
In the patient with second-degree AV block, the study of neck tations or mild anxiety is approached differently from such
veins can reveal the nature of the block, but usually the findings arrhythmias precipitating angina in a patient with coronary artery
are too subtle to recognize. Patients with type I (Wenckebach) disease, syncope in a patient with aortic stenosis, or claudication
second-degree AV block can exhibit progressive increase in the in a patient with peripheral vascular disease. It is imperative to
A-C jugular venous pulse interval before the nonconducted P remember that clinical decision making must be founded on the
wave, progressive quickening of the ventricular rate, and progres- ECG interpretation of the arrhythmia in concert with the assess-
sive decrease in the intensity of the first heart sound. In type II ment of the patient. Thus, the physician evaluates and determines
second-degree heart block, the PR interval remains fixed before treatment for a patient who has a rhythm disturbance, rather than
the block, and so does the A-C interval and intensity of S1. The a rhythm disturbance in isolation.
features of AV dissociation noted earlier usually are present The principle in diagnosing and treating symptomatic patients
during complete AV block. with an undocumented cardiac rhythm disturbance is simple and
obvious. One needs merely to record the ECG during a symp-
tomatic episode and then document a causal relation between
Carotid Sinus Massage arrhythmia and symptoms. This often is easier said than done,
however, and a variety of approaches are used to achieve that
Modulating autonomic tone by carotid sinus massage during the result.
physical examination can be useful to expose the patient with the
hypersensitive carotid sinus reflex. The clinician first needs to
listen carefully over both carotids to be certain that no bruit is Rationale for the Use of Ambulatory
present, palpate lightly to determine that a normal carotid pulse Electrocardiography
is present, and then gently depress or rub the carotid sinus.
Gentle massage for approximately 10 to 15 seconds or less usually • Arrhythmia diagnosis and exclusion as a cause of infrequent
is all that is necessary to produce significant periods of sinus symptoms
arrest or AV block in susceptible patients. • ST-T changes related to myocardial ischemia
The response to carotid sinus massage or other vagal maneu-
vers can be helpful in differentiating one tachycardia from
another. In the most definitive responses, carotid sinus massage Short-Term and Long-Term Continuous Ambulatory
acutely terminates tachycardias such as AVRT, AVNRT, sinus Electrocardiographic Monitoring
node reentry, adenosine-sensitive atrial tachycardia, and idio-
pathic right ventricular outflow tract tachycardia. Carotid sinus The duration of electrocardiographic monitoring depends on the
massage can gradually slow a sinus tachycardia without termina- frequency of symptoms. Rhythm disturbances occurring with
tion and will decrease the ventricular response to atrial tachycar- great frequency are naturally easier to document than those that
dia, atrial flutter, and atrial fibrillation without termination, occur sporadically. Long-term ECG recordings in the outpatient
thereby exposing atrial activity. Carotid sinus massage transiently setting usually constitute one of the early diagnostic choices in
terminates the permanent form of AV junctional reciprocating the patient without a life-threatening cardiac arrhythmia. The
tachycardia, which then restarts when carotid massage ceases. patient with a life-threatening arrhythmia may need to be hospi-
Carotid sinus massage does not affect reentrant ventricular or talized to allow for these recordings. A long-term ECG recording
junctional tachycardias. Unfortunately, not all presentations of provides the most direct documentation of an infrequent cardiac
these tachycardias behave in such a predictable fashion, and inter- arrhythmia. Prolonged ECG recordings in patients engaged in
mediate or overlapping responses can occur. normal daily activity provide the methodology to quantitate the
frequency and complexity of the rhythm disturbance, to correlate
these alterations with symptoms, and to evaluate the effect of
appropriate pharmacologic therapy on the arrhythmia. In addi-
Laboratory Tests tion, such recordings can document alterations in the QRS-ST
and T contour. A 30-day event recorder often is helpful if
As indicated earlier, the initial assessment of the patient begins arrhythmia occurs at least once in a month. The patient can
with a careful history and physical examination. Several nonin- activate these latter devices when symptoms occur, store 30
vasive and invasive tests add to the physician’s ability to obtain seconds or more of the ECG rhythm (a memory loop provides
information about the arrhythmia. Before any test is ordered, ECG information about the arrhythmia that transpired for some
however, it is imperative to decide whether the information pro- seconds before device activation), and transmit it to a central
vided by the test is sufficiently important to justify its risk or monitoring station over the telephone. Alternatively, some
expense. Whenever possible, tests with maximal sensitivity, devices automatically record rhythms that exceed preset limits.
The automatic recorder is useful in patients who fail to perceive arrhythmia. Apart from diagnosing ventricular arrhythmia, a
58
all symptoms associated with the arrhythmia or are unable to dual-chamber implantable cardioverter defibrillator also
activate the recording system because of rapidly progressing helps in identifying cycle length, duration, and frequency of
syncope or other problems. atrial arrhythmias.
1. 24- to 48-hour ambulatory Holter monitoring. Short-term
continuous Holter monitoring may be sufficient for patients
with daily symptoms related to arrhythmia such as Diagnostic Yield of Electrocardiographic
palpitation, presyncope, or syncope. If the arrhythmia does Monitoring
not occur with sufficient frequency, then a simple 24-hour,
or even 48-hour, recording will not be useful. Newer Holter For the recording session to be specific, the patient must have
monitors also record a 12-channel ECG. Such studies are both the arrhythmia and symptoms simultaneously. If symptoms
helpful in arrhythmia characterization including atrial occur without an arrhythmia, the latter can be excluded as a
fibrillation, as well as ST-T wave changes related to cause. Recording arrhythmias without symptoms precludes a
ischemia and Brugada syndrome. definitive causal relation between symptoms and arrhythmia and
2. Long-term (15-day) Holter monitoring system. These systems reduces the specificity of the test. The sensitivity of the test is
are used for diagnosing arrhythmias occurring once or twice highly variable, depending on the prevalence of the arrhythmia.
in a week. With these devices, cardiac activity is The diagnostic value of ambulatory monitoring seems to
continuously recorded by chest electrodes that are attached depend on a number of variables, including the frequency and
to a pager-sized sensor. The sensor of the pager wirelessly duration of arrhythmia, accurate diary maintenance, and inpa-
transmits collected data to a portable monitor that analyzes tient monitoring versus outpatient monitoring. Approximately
the rhythm data. If an arrhythmia is detected by an 25% to 50% of patients experienced a symptom, but only 2% to
arrhythmia algorithm, the monitor automatically transmits 15% record a causal cardiac arrhythmia, and 35% will log a
recorded data wirelessly via the internet to a central symptom without a corresponding ECG abnormality.
monitoring station for subsequent analysis. Patient activated
data is also transmitted.
3. Event recorders (with and without loop):
a. Trans-telephonic monitoring (TTM) systems are external Correlation With Cardiac Arrhythmias on 24-Hour
event recorders without loop; they are noncontinuous and Long-Term Electrocardiographic Monitoring
ambulatory recording system.1 After activation by the
patient, an ECG is recorded and directly transmitted by Twenty-Four–Hour Holter Recordings
telephone to a receiving center. In a study of 518 patients, 24-hour Holter recordings were per-
b. Event recorders with looping memory (continuous event formed for palpitations and other symptoms related to arrhyth-
recorders [CERs]) make a continuous one-lead recording, mia. Two hundred seventy-four (53 %) had significant arrhythmias
but the rhythm strip will only be saved when a patient (41% ventricular and 20% ventricular, 8% both).2 No presenting
activates the device. Most devices can be programmed to complaint or cardiovascular diagnosis correlated closely with any
save preactivation and postactivation rhythm strips. specific cardiac arrhythmia. Major arrhythmias, including supra-
c. Autotriggered event monitors with looping memory ventricular and ventricular tachycardias, often occurred asymp-
(autotriggered CER) devices automatically recognize tomatically (in 44 of 54 and 37 of 40 patients, respectively).
prespecified high and low heart rates. One such device Among 371 patients with accurate historic logs, only 176 (47%)
performs a continuous ECG analysis combined with who had long-term electrocardiographic monitoring had typical
automatic storage of abnormal events detected in a symptoms during the monitoring period. Only 50 patients (13%)
20-minute solid-state memory with continuous loop had concurrence of their presenting complaints with an arrhyth-
analysis up to 7 days. In addition, it also records patient- mia, whereas 126 patients (34%) had their typical symptoms
trigger events. The most recent advancement in associated with a normal electrocardiogram, which may be
ambulatory arrhythmia monitoring is mobile cardiac helpful in excluding any cardiac arrhythmia as the primary cause
outpatient telemetry in which a portable sensor for their complaints.
continuously detects asymptomatic, prespecified
arrhythmias and transmits the ECG data in real-time to a Continuous Event Recording Versus 24- to 48-hour
pocket-sized monitor at the patient’s home. If the Holter Monitoring
algorithms in the monitor detect an abnormal heartbeat, Studies have shown that with CER, a diagnosis can be established
the monitor automatically transmits the patient’s ECG in 21% to 62% of the patients, compared with a maximum of
data to the monitoring center using wireless 30% with Holter monitoring. The CER is also better than the
communications. Holter monitor at excluding arrhythmias during symptoms (34%
4. Implantable autotriggered and patient-triggered loop recorders. and 2%, respectively).1
An implantable loop recorder placed beneath the skin can be
used for monitoring of the cardiac rhythm for as long as 12 Autotriggered Continuous Event Recording Versus Traditional
to 24 months. Therefore, it is useful in patients with Continuous Event Recording Versus Holter recording
infrequent symptoms. The device has both autotriggered In a larger study of 1800 patients, the autotriggered CER was
and patient-activated arrhythmia recording facilities. The compared with the traditional CER and 24-hour Holter monitor-
devices are also available for recording a specific arrhythmia, ing with 600 patients in each group. The diagnostic yields were
such as atrial fibrillation. Use of such devices has been 71%, 27%, and 6% in autotriggered-CER, patient-triggered
successful in recording tachyarrhythmias and, more CER, and 24-hour Holter monitoring groups, respectively.3
commonly, bradyarrhythmias. Arrhythmia recordings can be
sent to the analyzing center via the telephone and then to ILR Versus Noninvasive Testing
physicians via the Internet. Giada et al.4 studied 50 patients for the diagnostic yield of the
5. Pacemakers and implantable cardioverter defibrillators. Dual- use of insertable loop recorder (ILR), which was randomly com-
chamber pacemakers can record atrial and ventricular pared with conventional strategy (24-hour Holter recording, a
high-rate episodes and can be correlated with the 4-week period of CER, or electrophysiological testing if the
previous two strategies yielded negative results).4 The diagnosis identify groups of patients at greater or lower risk for SCD, they
was made in only five patients (21%) of 25 patients in the con- all suffer from an inability to predict precisely the occurrence of
ventional strategy group compared with 19 (73%) of 25 patients life-threatening arrhythmias in individual patients (see Chapters
in ILR group (with arrhythmia recording up to 1 year). 65 to 69).
Electrophysiologic Study Tests Indicated for Specific Symptoms

Naturally, in many patients, invasive electrophysiologic testing is Syncope
required to initiate the electrical abnormality. Such studies The underlying disorder can be determined using standardized
provide important information when a particular arrhythmia can clinical evaluation in up to three fourths of patients with syncope.
be initiated that is responsible for the patient’s symptoms. In In unselected populations, slightly more than one third of the
some instances, however, a tachyarrhythmia present clinically patients have neurocardiogenic syncope, one fourth have
cannot be induced in the electrophysiology laboratory. An impor- orthostatic hypotension, and the remaining patients have miscel-
tant point is that failure to demonstrate a rhythm abnormality laneous conditions. Evaluation of patients suspected of cardioin-
does not exclude the possibility that it is present on another occa- hibitory or vasodepressor syncope often includes tilt table testing.
sion and is still responsible for the patient’s symptoms. Thus, the The yield from prolonged ECG ambulatory recordings and from
sensitivity of the electrophysiological study may be low, depend- electrophysiological studies in patients experiencing syncope who
ing on the nature of the rhythm disturbance. Absence of proof is do not have structural cardiac disease is in general low and unre-
not the same as proof of absence. warding. It is important to establish the cause of the syncope, if
Ideally, the electrophysiologic study would induce only clini- possible, because patients who have syncope from a noncardiac
cally and prognostically important cardiac arrhythmias in all cause usually have an excellent prognosis, whereas those who
patients who are at risk for a spontaneous arrhythmia and in no have syncope from a cardiac cause have a greater prevalence of
patient without such a risk. Unfortunately, this is not the case, sudden cardiac death.
and it is clear that such a study, depending on the number of extra Electrophysiological testing is recommended for patients who
stimuli used, can induce nonspecific tachyarrhythmias, in particu- have syncope or near-syncope that remains unexplained after a
lar flutter and fibrillation in both the atria and ventricles. With thorough evaluation, including a complete neurologic evaluation
certain arrhythmias, the test can be highly specific. For example, and ambulatory ECG recordings. Because both bradycardia and
it would be uncommon to induce a sustained AVRT or AVNRT tachycardia can be responsible for syncope, the diagnosis or
in a patient who does not also have this arrhythmia clinically or exclusion of a specific etiologic disorder should not be based on
is at risk of having it. Furthermore, it would not be likely to anything other than an ECG recording during the syncopal
induce a sustained monomorphic VT in a patient who is not at event, or replication, during an electrophysiological study or
risk for a clinical occurrence of such an arrhythmia. other maneuvers, of a cardiac rhythm disturbance that produces
the same or similar symptoms in the patient. For example, in the
patient who has first-degree AV block and left bundle branch
Stress Testing and Other Noninvasive Studies block, episodes of VT may be responsible for the syncope, rather
than episodes of more advanced AV block.
An exercise stress test can be useful, particularly in patients who
experience symptoms when exercising (see Chapter 63). In Bradyarrhythmias
response to exercise testing, approximately one third of normal Many patients have asymptomatic bradyarrhythmias, and it is
subjects have ventricular ectopy, usually in the form of occasional important to establish that they produce symptoms in a given
uniform PVCs. These PVCs are more likely to occur at greater patient before assuming that therapy is required. Conversely, if
heart rates and are not reproducible from one test to the next. the patient becomes symptomatic when a spontaneous bradyar-
Multiform PVCs, pairs of PVCs, and VT infrequently develop rhythmia is demonstrated, further diagnostic studies might not
in response to exercise in healthy subjects. Because they can be be necessary. It is also possible that patients can be minimally
recorded in normal subjects, their presence does not establish the symptomatic but have arrhythmias that permit definitive thera-
existence of ischemia or heart disease. Ventricular ectopy gener- peutic decisions. For example, in patients who have type II
ally appears at lower heart rates (<130 bpm) in patients with second-degree AV block, the demonstration of His-Purkinje
coronary artery disease than in a normal population, and it often block, even in the minimally symptomatic or possibly asymptom-
occurs in the early recovery period. Ventricular arrhythmias are atic person, can be sufficient evidence to conclude that pacemaker
more reproducible from one test to the next in patients with therapy is indicated because of the risk for progression to com-
coronary artery disease, and more frequent PVCs (exceeding 10 plete AV block.
per minute), polymorphic PVCs, and VT are more likely to occur The patient with sinus node dysfunction can have syncope or
in patients with coronary artery disease than in persons with near-syncope but also might complain of symptoms consistent
normal hearts. PVCs at rest can be suppressed by exercise in with low cardiac output because of persistent bradycardia, such
patients with documented coronary disease; therefore, this obser- as fatigue or even manifestations of congestive heart failure.
vation does not necessarily imply a benign prognosis or absence Some patients can have associated tachycardia—producing the
of underlying structural heart disease. In normal subjects, results aforementioned bradycardia-tachycardia syndrome. Electrophys-
from consecutive exercise tests might not be reproducible, iological studies of sinus node function have low sensitivity but
whereas the test results are more reproducible in patients with relatively high specificity. Correlation of the presence of the
coronary artery disease, but not dependably so. Exercise testing bradycardia with the patient’s symptoms is of utmost importance.
may be useful to expose ECG abnormalities in patients with less Electrophysiological studies are indicated only when a causal
obvious forms of the long QT syndrome. relation between the appearance of the bradycardia and the
A variety of noninvasive tests have been developed in an patient’s symptoms cannot be established despite repeated non-
attempt to identify patients at risk for sudden arrhythmic death. invasive evaluations. It is important to keep in mind that asymp-
These tests include signal-averaged electrocardiography, heart tomatic sinus bradycardia with heart rates of 35 to 40 bpm, sinus
rate variability QT dispersion testing, T wave alternans assess- arrhythmia with pauses of 2 to 3 seconds, Wenckebach second-
ment, and baroreflex testing. Although these tests can help to degree AV block (particularly during sleep), wandering atrial
pacemaker, and junctional escape complexes can be completely interval), the tachycardia is called a short RP′ tachycardia, whereas
58
normal, especially in young people and in well-conditioned if a P wave occurs in the second half of the RR cycle, the arrhyth-
athletes. mia is called a long RP′ tachycardia. Considerations in the differ-
In patients with AV block, the scalar ECG is the most impor- ential diagnosis for a short RP′ tachycardia include AVNRT,
tant laboratory test, because the site of block usually dictates the AVRT, junctional tachycardia, and atrial tachycardia with a mark-
clinical course of the patient and whether a pacemaker is needed, edly prolonged PR interval. If no P waves or other evidence of
and the site of block usually can be determined from analysis of atrial activity are apparent, and the R-R interval is regular,
the scalar ECG. Only infrequently is an electrophysiological AVNRT is most likely. If a retrograde P wave is present in the
study indicated. Autonomic manipulation can be used to help ST segment, AVRT is most probable. Long RP tachycardias
establish the site of block. Atropine or isoproterenol shorten AV include sinus tachycardia, atypical AVNRT, permanent junctional
nodal conduction time and refractoriness, whereas vagal maneu- reciprocating tachycardia, and atrial tachycardia. The presence
vers prolong them. Little change occurs in the normal His- of conduction over an accessory pathway during sinus rhythm or
Purkinje conduction. Thus, exercise, atropine, or isoproterenol during tachycardia naturally suggests that the Wolff-Parkinson-
can shorten the PR interval and increase the ratio of conducted White syndrome with its associated accessory pathway is respon-
P waves during type I (Wenckebach) AV nodal block, whereas sible for the dysrhythmia. During VT, specific QRS contours and
these maneuvers can increase the number of blocked P waves in the presence of AV dissociation are useful in making the
type II second-degree AV block. Of note, however, important diagnosis.
overlap between the two responses is possible.
Tachyarrhythmias
As mentioned earlier, a 12-lead ECG should be obtained during Summary
tachycardia, as long as the patient’s condition is relatively stable.
If the QRS is normal and identical to that present during sinus Careful evaluation of the patient who has documented or sus-
rhythm, the tachycardia must be supraventricular, and the dif- pected cardiac arrhythmia is the prime focus of the clinical
ferential diagnosis now relates to its mechanism (see Chapters 73 cardiac electrophysiologist. Much useful information can be
77-79). The 12-lead ECG provides many diagnostic clues in this gleaned noninvasively, thereby potentially sparing many patients
regard. Supraventricular tachycardias can be classified as short from unnecessary electrophysiological testing. When indicated,
RP′ or long RP′ tachycardias, depending on the timing of the P however, such studies, particularly when coupled with radiofre-
wave in relation to the preceding R wave. When a P wave occurs quency ablation techniques, provide the most definitive informa-
closer to the preceding R wave (i.e., in the first half of the R-R tion for appropriate diagnosis and therapy.
2. Zeldis SM, et al: Cardiovascular complaints. Cor- for arrhythmia detection. Am J Cardiol 95(9):1055–
References relation with cardiac arrhythmias on 24-hour elec- 1059, 2005.
trocardiographic monitoring. Chest 78(3):456–461, 4. Giada F, et al: Recurrent unexplained palpitations
1. Hoefman E, Bindels PJ, van Weert HC: Efficacy of 1980. (RUP) study comparison of implantable loop
diagnostic tools for detecting cardiac arrhythmias: 3. Reiffel JA, Schwarzberg R, Murry M: Comparison recorder versus conventional diagnostic strategy.
systematic literature search. Neth Heart J of autotriggered memory loop recorders versus stan- J Am Coll Cardiol 49(19):1951–1956, 2007.
18(11):543–551, 2010. dard loop recorders versus 24-hour Holter monitors
Differential Diagnosis of Narrow and
Wide Complex Tachycardias 59
John Michael Miller and Mithilesh K. Das
P wave makes up a relatively large portion of the tachycardia

CHAPTER OUTLINE
cycle, as opposed to focal ATs (and all other types of SVT, during
Narrow QRS Tachycardias 575 which atrial activation begins at a discrete point as though it were
a focus).1 A major limitation of discerning P-wave morphology is
Wide QRS Tachycardias 577
the need to determine what is a P wave and what are ST segment,
Remaining Problems 579 T wave, and QRS complex. Helpful aids include finding periods
of NCT with 2 : 1 AV conduction; comparing the complex in
Summary 580
question with a sinus rhythm P-QRS-T cycle; and increasing
ECG gain (Figure 59-1).
Many patients with cardiovascular problems have sudden onset

of severe symptoms; among the variety of diagnoses, rapid tachy- History and Physical Examination
cardias perhaps are most likely to elicit symptoms in caregivers.
Part of this anxiety arises from uncertainty about the specific Patients with NCTs usually have recurrent episodes of arrhyth-
diagnosis (i.e., supraventricular tachycardia [SVT] vs. ventricular mia. The age of onset of episodes often suggests a diagnosis:
tachycardia [VT]), part from uncertainty about how to treat, and Episodes from birth onward are likely to be AV reentrant tachy-
part from uncertainty about clinical implications (e.g., Is the cardia (AVRT) using an accessory AV pathway present from
patient having a heart attack? Is he going to die before treat- birth, or AT. Onset of symptoms during or after puberty is
ment?). In this chapter, we will explore the tools available to common in AV nodal reentrant tachycardia (AVNRT). Although
address the first problem (diagnosis); after getting this correct, these scenarios are generally true, any type of NCT can occur
answers to the other problems generally flow naturally. An later in life. Symptoms include palpitations, light-headedness,
important distinguishing feature for clinical implications of a dyspnea, chest pain, and neck fullness. In many, episodes are
tachycardia episode is whether or not structural heart disease facilitated by exercise and emotional upset. Physical maneuvers
(SHD; prior infarction, cardiomyopathy, prior surgery, etc.) is such as Valsalva or breath holding can often terminate episodes.
present: In most cases of SVT, SHD is either absent or unrelated Episodes tend to become more common and longer lasting with
to the episode, but in most VT patients, SHD serves as the basis aging. Physical examination during NCT episodes shows tachy-
of, or substrate for, the arrhythmia. cardia in a conscious, often anxious patient. Blood pressure is
The first major differentiator in correctly diagnosing tachy- usually preserved. Bulging of neck veins sometimes can be per-
cardia is the width of the QRS complex: Narrow (<120 ms) QRS ceived. In patients with repaired congenital heart disease, scar-
complex tachycardias (NCTs) in adults are almost always supra- based atrial macroreentry should be suspected.
ventricular in origin (involving tissue at or above the bundle of
His), whereas wide (≥120 ms) QRS complex tachycardias (WCTs)
are often, but not always, ventricular in origin. Electrocardiographic Differential Diagnosis
Among NCTs, the differential diagnosis is based on the A : V
ratio; among those with 1 : 1 AV ratio, the timing of the P wave
Narrow QRS Tachycardias relative to a QRS complex; and P-wave morphology (Table 59-1).
Although individual variability is noted, some patterns are rela-
Diagnostic Possibilities tively constant.
A. A : V ratio:
The major categories of NCTs include those that are primarily 1. NCTs with A : V ratio >1 include AT and flutter, as well
atrial in origin (atrial tachycardia, flutter, fibrillation); those that as rare cases of AVNRT with 2 : 1 block, usually in the
are based in the atrioventricular (AV) junction; and those that His bundle (Figure 59-1, far right).
incorporate atrium and ventricle in a large circuit (accessory 2. NCTs with A : V ratio = 1 comprise a large and
pathway medicated AV reentry). In this chapter, atrial fibrillation heterogeneous group, among which are AVNRT, AVRT,
will not be considered further, but flutter and atrial tachycardia AT, and the uncommon automatic junctional tachycardia.
(AT) deserve consideration. Classic electrocardiographic atrial ATs at times may have a 1 : 1 AV ratio, but the timing
flutter is now understood to be a continuous wave front propagat- relationship between QRS complex and the subsequent P
ing either clockwise or counterclockwise around the tricuspid wave is not fixed.
annulus. Other atrial arrhythmias are termed flutter on electro- 3. NCTs with A : V ratio <1 are rare and include sinus
cardiogram (ECG) but are mechanistically distinct; these can be rhythm with simultaneous conduction over fast and slow
focal in origin or reentrant (usually large circuits bounded by AV nodal pathways, nodofascicular pathway–based
natural barriers such as valves or scar tissue). ATs can be focal reentry, and either AVNRT or junctional tachycardia
(true focus or microreentry that appears focal in its propagation with retrograde block.
pattern) or macroreentrant, incorporating significant amounts of B. R-P interval in cases with 1 : 1 A : V ratio
atrial tissue in the circuit. The latter are noteworthy in that the 1. Absence of a visible P wave (subsumed in the QRS
575
AT with 2:1 Comparison with baseline Different gain settings Narrow inverted P wave
AV conduction Sinus Left Atrial AT Standard 2x standard Exactly between QRSs
2
3
aVR
aVL
aVF
V1
V2
V3
V4
V5
V6
1 sec
Figure 59-1. Determination of P waves during narrow complex tachycardia using a variety of methods. Vertical gray bands denote the entire width of the P wave; arrows
indicate a visible P wave, AT, Atrial tachycardia; AV, atrioventricular.
Table 59-1. Narrow Complex Tachycardia Diagnostic Features
Atrioventricular (AV) Ratio
>1 (A > V) 1 (A = V) <1 (A < V)* Indeterminate (no clear P)
Diagnostic possibilities Atrial tachycardia AVNRT Sinus rhythm/1 : 1 conduction AVNRT

Atrial flutter AVRT AVNRT (Junctional tachycardia)
(AVNRT) Atrial tachycardia Junctional tachycardia (Atrial tachycardia)
(Junctional tachycardia) Nodofascicular tachycardia
R-P Interval
No Visible P Short RP Intermediate RP Long RP
Diagnostic possibilities AVNRT AVRT AVNRT Atrial tachycardia

(Atrial tachycardia) AVNRT Atrial tachycardia AVNRT
Atrial tachycardia AVRT AVRT
(Junctional tachycardia) (Junctional tachycardia) (Junctional tachycardia)
P Wave Morphology
Positive Leads 2, 3, aVF Negative Leads 2, 3, aVF Negative Lead 1 Positive All Precordial Leads
Diagnostic possibilities Atrial tachycardia AVNRT AVRT (left lateral pathway) Pulmonary vein ostia
AVRT AVRT Left atrial tachycardia
Atrial tachycardia
*All four items in this group are rare. Diagnostic possibilities are listed in order of frequency. Terms in parentheses denote rare situations.
AVNRT, Atrioventricular nodal reentrant tachycardia; AVRT, atrioventricular reentry tachycardia.
complex) is common in AVNRT (anterograde slow, 4. Long R-P NCTs are an interesting group with the same
retrograde fast pathways) but can occur in AT with a diagnostic possibilities as are seen in the other R-P
long AV conduction time. subsets, but ATs predominate; AVNRT is of the less
2. NCTs with a short R-P interval (P wave in the first common “fast-slow” variety, and accessory pathways are
one-third of the R-R interval) include AVRT, AVNRT of the very uncommon slowly conducting type.
(especially in patients >50 years old), and AT, with C. P-wave morphology2
junctional tachycardia a rare diagnosis. 1. Atrial activation in NCTs with positive P waves in the
3. Intermediate R-P interval NCTs (P wave in middle inferior leads begins near the top of the atria, including
one-third of the R-R interval) are of the same types as the upper crista terminalis, superior vena cava, and
short R-P NCTs, but AVNRT (“slow-slow”) and AT are appendage in the right atrium, and the pulmonary veins
more common than AVRT. and appendage in the left atrium, as well as cephalad
Differential Diagnosis of Narrow and Wide Complex Tachycardias 577
portions of the tricuspid and mitral annuli. Evaluation of the last portion of V1 negative are termed left bundle branch
59
precordial leads (anteroposterior) and lead 1 (left-right) block (LBBB)-type QRS, whereas the right bundle branch block
further refines the site of origin in the other two planes. (RBBB) pattern is denoted by a positive deflection in the last half
As such, ATs account for many of these, but AVRT with of V1. Some WCTs do not fit well into these descriptions, with
pathway atrial insertions on the tops of mitral or “Rs” or “qRs” patterns; in cases with this degree of ambiguity,
tricuspid annuli are also part of this group. the diagnosis is almost always VT.
2. Negative P waves in the inferior leads denote onset of
atrial activation in the lower portion of the atria (low
crista terminalis, coronary sinus os, low septum, and History and Physical Examination
tricuspid annulus in the right atrium, and low septum or
mitral annulus in the left atrium). All varieties of Patients with SVT-A, like those with NCT, have usually had
AVNRT as well as AVRT using posterior AV pathways prior episodes of arrhythmia, whereas for those presenting with
fall into this group, as do some ATs. VT, it is often their first episode. VT patients tend to have a
3. An inverted P wave in lead 1 is a reliable indicator of history of structural heart disease such as prior myocardial infarc-
left-to-right atrial activation, either from AT arising in tion, cardiomyopathy, or significant valvular disease. Symptoms
the left atrium or pulmonary veins, or from AVRT using in VT patients include palpitations, light-headedness, or syncope,
a left lateral pathway. dyspnea, and chest pain; episodes often begin without apparent
4. When all precordial leads show positive P waves, a left provocation. Physical examination during VT episodes shows
atrial or pulmonary venous source should be suspected. tachycardia in a patient who may be conscious, but often with
marginal blood pressure. Cannon A waves in neck veins and vari-
able intensity of S1 are sometimes present.
Special Cases
P-wave morphology is usually a good indicator of the origin of Electrocardiographic Differential Diagnosis
atrial activation; however, there are several situations in which
this inference should be made cautiously. These include cardiac Several features of the ECG in WCT have proven diagnostic
malposition (dextrocardia, prior pneumonectomy) and previous utility; the more important among these include the following
extensive atrial ablation or surgery, as for repair of congenital (Figure 59-3 shows several examples):
heart disease. After ablation or surgery, common arrhythmias A. QRS duration—In 1978, Wellens found that among WCTs
such as right atrial cavotricuspid isthmus-dependent flutter can with QRS duration >140 ms, 95% were VTs. Because some
have an unusual appearance, and uncommon arrhythmias (such cases of LBBB aberration have QRS >140 ms, this rule was
as left atrial macroreentry) can masquerade as more common later modified to allow SVT-A with an LBBB pattern to
ones (such as right atrial flutter).3 Finally, very rarely, VT can have QRS up to 160 ms. This criterion is deceptively
have a relatively narrow QRS (<120 ms) and be mistakenly diag- simple, in that different observers can measure very different
nosed as SVT. This occurs in several settings: children (with a QRS durations, especially when QRS onset or offset is not
narrower baseline QRS in sinus rhythm) and adults with VT
arising in the His-Purkinje system or propagating into it very
early in the QRS. In almost all of these cases, the prognosis is LBBB-type QRS RBBB-type QRS
similar to that of SVT (i.e., benign). V1 V6 V1 V6
Ventricular
tachycardia
Wide QRS Tachycardias
Diagnostic Possibilities SVT with

aberration
Although several possible causes of WCT are known (Box 59-1;
Figure 59-2), a vast majority are VT or SVT with aberrant inter-
ventricular conduction (SVT-A). The remainder of this discus- SVT with
sion will concentrate on differentiating between the two. Because preexcitation
of the wider diagnostic possibilities and the more serious nature (WPW)
of some causes, WCTs evoke greater anxiety among health care
workers than NCTs. For purposes of classification, WCTs with
SVT with
baseline
abnormal ECG
Box 59-1 Wide Complex Tachycardia Diagnostic Possibilities
• Ventricular tachycardia SVT with
• Supraventricular tachycardia with one of the following: hyperkalemia
• Aberrant interventricular conduction (His-Purkinje)
• Anterograde conduction over accessory pathway
• Abnormal baseline QRS configuration Ventricular
pacing
• Nonspecific QRS widening due to electrolyte abnormality/
drug effect
• Ventricular pacing Figure 59-2. Examples of leads V1 and V6 in both LBBB and RBBB types of QRS
• Electrocardiographic artifact complexes in different types of WCT. Similarities and differences can be appreciated
among the various causes.
1 aVR V1 V4
2 aVL V2 V5
3 aVF V3 V6
Figure 59-3. ECG of VT showing several diagnostic features, including very wide QRS; right superior axis deviation; concordant (negative) precordial R-wave pattern; R-wave
duration in V1 >30 ms; QRS onset to S wave nadir in V1 >60 ms; AV dissociation (filled black circles denoting visible P waves, hollow circles denoting less visible ones); and
fusion beats (third complex in V1).
V1 V6 to typical aberration patterns that include rSR′, rSr′,

or rR′ in lead V1; all others, including Rsr′, Rr′, qR,
Normal and monophasic R, are then VT patterns.
conduction b. Lead V6—A small amount of normal right ventricular
LBBB RBBB LBBB RBBB (RV) voltage is directed away from V6; in RBBB
aberration, this is shifted later in the QRS to enlarge
the S wave. Because it is a small vector, the R : S ratio
SVT with
aberration is >1. In VT, all of the RV voltage, and some of the
left, is directed away from V6, leading to an R : S ratio
<1 (rS, QS patterns). In some cases, a monophasic R
wave or qR is seen, inconsistent with aberration and
VT thus likely VT.
2. LBBB pattern
a. Lead V1—The LBB determines the initial part of the
Figure 59-4. Diagrammatic representation of common QRS morphologies normal QRS; when blocked, the RBB transmits to
encountered in VT and SVT-A, in leads V1 and V6, for both LBBB and RBBB QRSs. Note the ventricles, and although several left ventricular
initial portions of the QRS complex in normal and aberrated QRS complexes, con- activation patterns may ensue, the initial portion of
trasted with initial QRS forces in VT complexes. the QRS is inscribed relatively rapidly. This yields
patterns in V1 (or V2) such as QS or rS, in which the
sharply delineated. For this reason, measurements in several QRS onset to the S-wave nadir is ≤60 ms, and the
simultaneously recorded leads should always be used. duration of the R wave, if present, is ≤30 ms. Patterns
B. QRS axis—Because aberration patterns are confined to left other than these, such as QS or rS with onset QRS to
or right axis deviation (thus, −60° to +120°), WCTs with a S wave nadir >60 ms or R wave duration >30 ms,
QRS axis outside this range are likely to be VTs. This is indicate VT.5 Distinguishing scar-based VT from VT
particularly true for the axis from −90° to 180° (i.e., negative in the absence of SHD (which often resembles
complexes in leads 1, 2, and 3); this criterion has positive SVT-A) can be difficult; R-wave transition after V4
predictive accuracy (PPV) >95% for VT. and, in V1 or V2, notching of the downstroke, or onset
C. AV relationship—It has long been recognized that AV of QRS to the nadir of the S wave >90 ms, each
dissociation (ventricular rhythm faster than an independent, indicates scar-based VT.6
regular atrial rhythm) is a reliable indicator of VT, with very b. Lead V6—In true LBBB, no Q wave is present in the
rare exceptions. However, it is present (or recognizable) in lateral precordial leads; the presence of any Q wave in
only one-third of known VT ECGs,4 thus decreasing its V6 during WCT provides strong evidence for a VT
utility. Other AV relationships are helpful, such as 2 : 1 diagnosis.
retrograde conduction and retrograde Wenckebach E. Precordial leads
conduction (often more difficult to recognize than AV 1. Concordance—If all of the precordial leads have the
dissociation). The latter forms of non-1 : 1 retrograde same polarity (all positive or all negative), it has been
conduction, still diagnostic of VT, are present in about 5% said that VT is likely because aberration patterns
of cases of VT, whereas a 1 : 1 AV relationship is recognized fundamentally never have positive concordance; in some
in about 8% of cases.4 Uncommonly seen fusion and capture cases of LBBB aberration, R waves may not be seen until
beats are manifestations of a non-1 : 1 AV relationship V7 or later, leaving a concordant negative pattern.
during VT. Because concordant patterns are present in <20% of all
D. Specific patterns in ECG leads V1 and V6 (Figure 59-4): VTs, this criterion has low sensitivity. A more recent
1. RBBB pattern analysis4 found that a negative concordant pattern had
a. Lead V1—The RBB does not contribute greatly to the virtually no capacity to distinguish SVT-A from VT, but
initial part of the normal QRS; thus, when blocked, a positive concordant pattern remained a strong
the first 40 ms of the QRS is unchanged. This leads differentiator.
Differential Diagnosis of Narrow and Wide Complex Tachycardias 579
2. Brugada criteria7—In most cases of SVT-A, at least one A. VT B. VT C. SVT D. VT E. VT F. SVT

L. 2:VT L. 2:VT L. 2:SVT L. 2:SVT L. 2:SVT(VT) L. 2:SVT
59
precordial lead has an “RS” pattern, but this need not be
aVR:VT aVR:VT aVR:SVT aVR:VT aVR:VT aVR:VT
the case in VT; thus if no precordial leads in a WCT
have an RS, it is most likely VT. This is the first step of 1
the algorithm by Brugada et al; further, if an RS complex
was present, the interval from onset of R to nadir of S
>100 ms strongly suggests VT. Other steps in the 2
algorithm include evaluation for AV dissociation and
the usual V1 and V6 patterns.
F. Newer criteria—Two recent algorithms have used single 3
ECG leads (aVR or lead 2) to differentiate VT from
SVT-A. These have the virtues of simplicity and, with it,
easier memorization. aVR
1. aVR criteria—Vereckei proposed two algorithms
incorporating lead aVR. The first had four steps (a
positive result at any step makes a VT diagnosis, with Figure 59-5. Lead 2 and aVR criteria in SVT-A and VT. Lead 1, 2, 3, and aVR are shown
from several WCTs. Long dashed vertical lines indicate earliest QRS onset among
the remaining ECGs categorized as SVT-A)8: AV
all leads; shorter dashed lines denote peak in lead 2. In A and B (VT) and C (SVT),
dissociation; R wave in aVR; standard criteria in V1 and various morphologies of lead 2 and aVR are shown; both criteria correctly diagnose
V6; and the ventricular activation velocity index. The each rhythm. In D and E (both VT), lead 2 incorrectly diagnoses SVT, while aVR is
latter is the ratio of the vertical amplitude (in tenths of correct; in E, lead 2 by itself is incorrect (with an isoelectrical segment before the
millivolts) traversed in the first 40 ms of the QRS QRS downstroke), but when analyzed using the true QRS onset, lead 2 is correct.
complex, divided by the same amplitude measurement in In F, SVT is present and is correctly diagnosed by lead 2, but not by aVR (slurred
the last 40 ms, of any lead with a biphasic or triphasic downstroke of Q wave).
QRS complex. Ratios >1 indicate SVT-A, whereas ratios
≤1 indicate VT. This algorithm performed well in initial
testing but is somewhat cumbersome, and it is difficult to
remember how to make the measurements. The second AV dissociation, the rate is so rapid that dissociated P waves
proposed algorithm9 involves only aVR and thus is usually cannot be readily observed. This relatively
generally simpler. It also consists of four steps: presence uncommon arrhythmia is important because of its
of monophasic R; Q or R >40 ms in duration; slurred straightforward treatment with catheter ablation. Most
downstroke on Q wave; and the velocity index from the patients with BBR also have other forms of VT for which
previously noted algorithm (but applied only to aVR). As an ICD is indicated; thus catheter ablation is rarely
with the prior algorithm, a positive result at any step their sole therapy. If BBR is discovered during
indicates a diagnosis of VT, with a final default result electrophysiological testing, ablation can decrease ICD
being SVT-A. This algorithm, similar to the prior one, shocks.
had an overall accuracy of 91%. Neither, however, has C. Irregular VT—Like SVTs, almost all VTs have very regular
performed as well when applied to other data sets by R-R intervals to the resolution of measurement on standard
different investigators.10 ECG. A small proportion of VTs are grossly irregular, and
2. R-wave peak time—Pava et al11 proposed another simple, in these cases, the obvious alternative diagnosis is atrial
one-step criterion: the interval from QRS onset to peak fibrillation with aberration. In this case, morphology criteria
amplitude (positive or negative) in lead 2. Using a cutoff can be very helpful (as can finding AV dissociation, if atrial
of 50 ms, almost all WCTs with a shorter time to peak fibrillation is not actually present).
amplitude in lead 2 were SVT, whereas almost all WCTs Finally, in some situations, ECG criteria suggest VT in
with intervals ≥50 ms were VT. The sensitivity and patients with SVT-A. This group includes preexcited SVT
specificity were 0.98 and 0.93, respectively. Although this (during sinus rhythm, preexcitation is almost always readily
criterion appears to have many desirable features— apparent) and other typical forms of SVT in patients with sig-
simplicity, ease of application, accuracy—its performance nificant structural heart disease such as repaired congenital heart
in the hands of other investigators10 has been less disease or severe cardiomyopathy. These are relatively small but
impressive (sensitivity 0.60, specificity 0.83). Examples growing populations in which better methods of distinguishing
are shown in Figure 59-5. SVT-A from VT will become increasingly important in the
future.
Special Cases
Several special situations are known in which criteria suggest Remaining Problems
SVT-A when VT is present:
A. Relatively narrow complex VT—As noted previously, these Although numerous algorithms have been developed to aid in
VTs generally (but not always) occur in the absence of SHD diagnosis of WCT, none is 100% sensitive and specific; particu-
and in younger individuals. Correctly identifying these larly difficult cases include VT in the absence of SHD. In addi-
patients is important therapeutically because catheter tion, no criteria have performed as well in subsequent analysis
ablation can be curative in most cases. A more important and “real-world” testing as in the original publication. Thus,
error is correctly identifying the WCT as VT, while using better criteria are necessary, but research is also needed into why
implantable cardioverter-defibrillator (ICD) therapy merely the existing criteria are not as robust in practice as they initially
on the basis that the rhythm was VT. seemed to be. Potential causes include (1) misremembering of
B. Bundle branch reentrant (BBR) VT—QRS complexes are criteria (e.g., Was it 130 or 140 ms for QRS duration?); (2) incor-
identical to LBBB aberration or, more rarely, RBBB rect application of criteria (e.g., mistaking small irregularities in
aberration with left or right axis deviation. ECG distinction the ECG baseline for AV dissociation); (3) imperfections of cri-
from SVT-A is difficult, in that even though most cases have teria (especially in VT in the absence of SHD, or in SVT in
patients with severe cardiomyopathy); and (4) unwillingness to that cannot be prevented, is used in most cases of VT related to
believe the results of analysis (e.g., All the criteria suggest VT, SHD. The ECG is reasonably good for distinguishing among
but I STILL think it’s SVT). Until progress is made on each of NCTs, as long as a P wave can be clearly seen (or is clearly
these fronts, diagnostic errors will persist. absent). In cases of WCT, many algorithms have been proposed
to differentiate between the two major causes: VT and SVT-A.
Although each algorithm is introduced with great promise, each
has its limitations. The ideal algorithm would be one that is (1)
Summary easy to remember, (2) universally applicable to all WCTs, (3) easy
to apply with unequivocal results, and (4) 100% sensitive and
Arriving at the correct diagnosis of tachycardia has obvious clini- specific for VT (or SVT). Until such a tool is developed, it is
cal importance, in that current therapies can cure many disorders safest to treat the patient with WCT that cannot be readily clas-
(most SVTs and VTs in the absence of SHD), thereby preventing sified for whatever reason as though the rhythm is VT, until
further episodes. In contrast, the ICD, which reacts to episodes proven otherwise.
4. Miller JM, Das MK, Yadav AV, et al: Value of the 8. Vereckei A, Duray G, Szenasi G, et al: Application
References 12-lead ECG in wide QRS tachycardia. Cardiol of a new algorithm in the differential diagnosis of
Clin 24:439–451, 2006. wide QRS complex tachycardia. Eur Heart J
1. Shah D, Sunthorn H, Burri H, et al: Narrow, slow- 5. Kindwall KE, Brown J, Josephson ME: Electrocar- 28:589–600, 2007.
conducting isthmus dependent left atrial reentry diographic criteria for ventricular tachycardia in 9. Vereckei A, Duray G, Szenasi G, et al: New algo-
developing after ablation for atrial fibrillation: wide complex left bundle branch block morphol- rithm using only lead aVR for differential diagnosis
ECG characterization and elimination by focal RF ogy tachycardias. Am J Cardiol 61:1279–1283, of wide QRS complex tachycardia. Heart Rhythm
ablation. J Cardiovasc Electrophysiol 17:508–515, 1988. 5:89–98, 2008.
2006. 6. Wijnmaalen AP, Stevenson WG, Schalij MJ, et al: 10. Jastrzebski M, Kukla P, Czarnecka D, Kawecka-
2. Kistler PM, Roberts-Thomson KC, Haqqani HM, ECG identification of scar-related ventricular Jaszcz K: Comparison of five electrocardiographic
et al: P-wave morphology in focal atrial tachycar- tachycardia with a left bundle-branch block con- methods for differentiation of wide QRS-complex
dia: Development of an algorithm to predict the figuration. Circ Arrhythm Electrophysiol 4:486– tachycardias. Europace 14:1165–1171, 2012.
anatomic site of origin. J Am Coll Cardiol 493, 2011. 11. Pava LF, Perafan P, Badiel M, et al: R-wave peak
48:1010–1017, 2006. 7. Brugada P, Brugada J, Mont L, et al: A new time at DII: A new criterion for differentiating
3. Gerstenfeld EP, Marchlinski FE: Mapping and approach to the differential diagnosis of a regular between wide complex QRS tachycardias. Heart
ablation of left atrial tachycardias occurring after tachycardia with a wide QRS complex. Circulation Rhythm 7:922–926, 2010.
atrial fibrillation ablation. Heart Rhythm 4(3 83:1649–1659, 1991.
Suppl):S65–S72, 2007.
Mapping and Imaging 60
Vivek Y. Reddy
time. The surface ECG can be used for the reference time during
CHAPTER OUTLINE
ventricular arrhythmias, but this is less likely to be effective for
Mapping 581 atrial arrhythmias because the surface P wave is often difficult
to discern because of the ECG changes related to ventricular
Imaging 587
depolarization/repolarization. A reference catheter placed in a
Conclusions 591 stable location within the chamber is often more effective to use
for the reference timing to which all the mapping electrogram
timings are compared. For focal atrial or ventricular tachycardias,
the goal is to identify the point of earliest activation. Ideally, the
Imaging and mapping have a central role in the modern practice morphology of the unipolar electrogram is used to recognize this
of clinical cardiac electrophysiology—in particular, the treatment site; at the site of the arrhythmia focus, the unfiltered unipolar
of cardiac tachyarrhythmias with catheter ablation. On the one electrogram should exhibit a QS morphology. The presence of
hand, mapping the electrical activation of focal arrhythmias such an R wave indicates that the ablation catheter is not at the site of
as ectopic focal atrial tachycardia has long been a primary meth- origin. For reentrant arrhythmias, activation mapping can also be
odology to identify their origin. Initial approaches involved the useful to identify the pathways of activation and determine where
sequential placement and movement of multielectrode catheters to target for ablation. In general, the two areas to consider for
based solely on fluoroscopy, but this has given way to the use of ablation are (1) areas of constrained activation (often between two
electroanatomical mapping (EAM) systems to carefully catalogue areas of block such that only a short bridging lesion set is required
electrical activation information in a spatially relevant manner to to transect the circuit) and (2) areas of slow conduction where
identify the site of tachycardia origin. However, the use of EAM even a single ablation lesion at the right location is often enough
systems has provided another means to address certain arrhyth- to terminate the rhythm.
mias beyond activation mapping—namely, substrate-based abla-
tion. For example, it is possible to eliminate the multiple Pacemapping
ventricular tachycardias (VTs) that are characteristic of postmyo- During pacemapping, the mapping catheter is manipulated to
cardial infarction VT by carefully rendering a geometric model various cardiac locations from each of which pacing is performed;
of the left ventricle with a superimposed voltage amplitude map the resulting ECG morphology is then compared to the target
to localize the infarcted tissue, followed by targeted ablation arrhythmia. From a practical perspective, pacemapping is almost
within this myocardial scar. When these mapping and imaging never used for atrial arrhythmias because of the difficulty in
approaches are combined with a solid understanding of the discerning the P wave morphology. However, pacemapping is
patient’s history, and the 12-lead electrocardiograms (ECGs) used for ventricular arrhythmias in two circumstances: (1) for
during sinus rhythm versus the target arrhythmia, the likelihood focal VTs such as outflow-tract VTs in which the QRS morphol-
for clinical success of catheter ablation is optimized. This chapter ogy at the VT site of origin would provide a “12-of-12” lead
will explore the various cardiac EAM systems and imaging ECG match, and (2) during substrate-based VT ablation of scar-
approaches (preacquired and intraprocedural) available for use in related VTs. For the latter, when pacemapping is performed
clinical practice. along the scar border, the morphology would be similar, if not
identical, to the target VT morphology at the site of the scar
border from which the VT exits. Once identified, a series of
linear lesions can be placed at this location to transect the VT
Mapping circuit empirically.
Conventional Mapping Techniques Entrainment Mapping

Entrainment mapping is an extremely useful technique to provide
There are three conventional approaches to cardiac mapping: evidence that the mechanism of a particular arrhythmia is reentry
activation mapping, pacemapping, and entrainment mapping. with an excitable gap, as opposed to an automatic mechanism or
These techniques are complementary and, depending on the triggered activity.1 In addition, once the mechanism of a rhythm
mechanism of the arrhythmia, one or more of these mapping is known to be reentry, entrainment maneuvers can be performed
approaches would be appropriate. to dissect the pathway of the arrhythmia—be it of atrial or ven-
tricular origin.2 By observing the response of the return cycle at
Activation Mapping cessation of entrainment (also referred to as continuous resetting),
Activation mapping involves the sequential movement of one or it is possible to determine whether the site of pacing is an active
more single or multielectrode catheters within the chamber of part of the circuit or is an irrelevant or passive location. That is,
interest to identify the activation pattern of the arrhythmia. if within the circuit, the return cycle at the cessation of pacing
Because mapping is performed during the target arrhythmia, this would be equivalent to the tachycardia cycle length; if outside the
approach is most applicable for sustained arrhythmias, and less circuit, the return cycle would be longer than the tachycardia
so for transient or unsustained rhythms. The timing of the elec- cycle length. Furthermore, the morphology of the paced ECG
trogram at each location is determined in relation to a reference complex (again, more useful for ventricular rhythms) can be
581
compared to the arrhythmia’s ECG complex to determine system has just recently been introduced into clinical practice –
whether the area (from which entrainment is performed) is within the Rhythmia mapping system (Rhythmia Medical, Burlington,
a constrained location. This would be relevant because the con- MA). Each of these systems will be discussed. In addition, it
strained areas are typically better sites to target for ablation. should be noted that two other EAM systems have largely been
abandoned in clinical practice: the LocaLisa system (Medtronic,
Miscellaneous Mapping Approaches Minneapolis, MN), which used three orthogonal electrical fields
In addition to activation, pacemapping, and entrainment mapping, to localize catheters in space, and the Real-time Position Man-
there are a few other specialized mapping approaches that can be agement system (Boston Scientific, Natick, MA), which used an
helpful. During standard catheter mapping of an arrhythmia, the array of ultrasound transducers embedded within diagnostic and
pressure from the catheter can cause transient tissue dysfunction mapping catheters for spatial colocalization by triangulation.
that, if applied at a critical site, can terminate the tachycardia.
This can occur in certain scar-related atypical atrial flutters and The CARTO System
scar-related VTs, but the most common systematic use of this The CARTO system uses magnetic localization technology to
approach is during mapping of Mahaim tachycardias.3 The atrio triangulate the position of small sensors that are incorporated
nodal or atrioventricular pathways that potentiate these tachycar- into the tips of various diagnostic and ablation catheters. Beneath
dias are typically located at the endocardial surface and are prone the fluoroscopy table is an electromagnetic location pad that
to pressure-induced transient interruption of conduction. Accord- emits a low-intensity series of magnetic fields and allows the
ingly, one can capitalize on this phenomenon and use an approach system to precisely localize, record, and display in real time the
of “bump” mapping, during which the catheter is used to apply position of the sensors, and hence the mapping catheter tip, in
pressure to various locations along the tricuspid valve such that three dimensions (x, y, and z) as well as orientation (roll, pitch,
when conduction is transiently interrupted, ablation is performed and yaw). The location of the mapping catheter is gated to a
to eliminate the putative pathway. One problem with this reliable point in the cardiac cycle and position information
approach is the unpredictability of the time before conduction recorded relative to the location of an external reference patch
resumes. This is particularly problematic if the exact time or affixed to the patient’s back, allowing for some degree of com-
location of the bump is not known; therefore, bump-mapping pensation for both cardiac and patient movement.
must be performed carefully, ideally with use of an EAM system Although effective, the initial versions of the CARTO system
so that the location can be spatially catalogued in a precise were limited by the ability to visualize only the catheter with the
manner. magnetic sensors—that is, the mapping catheter. However, the
Cryomapping is another approach that can provide transient most recent version of the system, CARTO 3, has nine magnets
arrhythmia interruption, but in a more predictably reversible positioned below the patient table in the “locator pad,” and six
manner.4 A cryocatheter is manipulated to the site thought to be reference patches are placed on the patient’s body to allow for
critical to the arrhythmia circuit. Next, refrigerant is delivered to better compensation for patient movement. The system also has
the catheter tip, but only to achieve a tip temperature of −30 °C—a the ability to use either a gated or nongated electrical reference.
temperature that is not cold enough to ablate any appreciable Nongated timing allows for the tracking of catheter movement
amount of tissue, but cold enough to cause transient interruption during fluoroscopy. This system is also capable of tracking mul-
of electrical conduction. Cryomapping is particularly useful when tiple mapping catheters by a hybrid of magnetic location technol-
the target arrhythmia is within close spatial proximity to a critical ogy and “current-based” impedance data that enables real-time
normal structure, such as during catheter ablation of a para- tip and curve identification and tracking (termed advanced catheter
Hisian accessory pathway. location). In this system, a magnetically tracked mapping catheter
is maneuvered within the cardiac chamber while simultaneously
emitting a low-level current, to allow the system to characterize
Electroanatomical Mapping Systems impedance data within the chamber. Subsequently, any standard
multielectrode catheter connected to the system can be localized
Fluoroscopy remains fundamentally critical to performing elec- with the chamber, albeit not necessarily with the same submilli-
trophysiology procedures; however, nonfluoroscopic EAM meter level of spatial resolution possible with magnetic localiza-
systems have revolutionized catheter ablation procedures. These tion. In procedures such as AV nodal slow pathway modification
systems are able to perform several important tasks. First, they for AVNRT treatment, the distance between the His bundle
can precisely localize the mapping catheter and other diagnostic catheter and the ablation catheter can be better judged in real
multielectrode catheters to a degree of spatial accuracy that time to improve safety and efficacy.
exceeds what is possible with fluoroscopy alone. In addition, The system also enables faster, high-quality anatomic recre-
because the localization is not fluoroscopy based, the systems ation using a nongated mapping mode termed fast anatomic
provide a real-time, continuous, three-dimensional (3D) under- mapping (FAM; Figure 60-1). Although the system requires the
standing of the catheters’ locations without continuous x-ray use of Biosense Webster electrode catheters, anatomic recon-
exposure. Second, by roving the mapping catheters, 3D render- struction can be facilitated, especially in the left atrium with the
ings of various cardiac chambers can be created. Third, electrical use of multielectrode circular mapping catheters. One important
information can be superimposed onto the 3D maps; two of the caveat regarding FAM is the effect of respiration on the quality
most common are activation maps that allow an appreciation of of the anatomic rendering that is created. The multiple patches
the arrhythmia circuit and voltage amplitude maps that allow an of the CARTO 3 system permit the introduction of an algorithm
appreciation of the location of scarred myocardium, such as the for respiratory gating. To track respiration, the system uses
substrate of postmyocardial infarction VTs. Third, these systems impedance readings derived from inter-patch measurements,
allow one to highlight certain important electrophysiological termed respiration indicators. The inter-patch current (from one
phenomena with various tags, such as the location of the His patch to the other) passes through the lungs, thereby recording
bundle or the response to entrainment maneuvers at different changes in impedance owing to pulmonary air volume. For the
sites. Finally, the systems allow one to catalogue the locations of algorithm to provide good respiratory gating performance, one
ablation lesions that are placed. first performs a “training” step in which the mapping catheter is
There are two major EAM systems currently used in clinical placed in the heart, touching a chamber wall for recording heart
practice: CARTO (Biosense-Webster, Diamond Bar, CA) and motion during respiration. During training, a correlation matrix
NavX (St. Jude Medical, St. Paul, MN). In addition, a third is calculated that best correlates the respiratory indicators to
3D-EAM
60
Post
erior
V Segmented 3D-CT
LIP
LSPV
LAA
Ante
rior
A B
Figure 60-1. Left atrium and pulmonary veins (LA-PV) anatomical reconstructions by electroanatomical mapping systems: A, Shown at top is an example of the respiratory
changes during mapping. Gating is applied such that spatial information is acquired only near end-expiration, as depicted by orange. At bottom is a superior view of a
three-dimensional (3D) fast anatomic mapping reconstruction of the LA-PV anatomy using CARTO. B, The 3D LA-PV geometries constructed by NavX (top) are CT-like
quality.
catheter motion. Although training is sampled in one location, it perform catheter ablation of scar-related VT. It is possible to
remains valid for the entire heart because the training is used only create high-resolution, highly accurate, multichamber maps with
to allow the algorithm to perceive the time point in the respira- superimposed voltage amplitude data to render an excellent
tory cycle. understanding of the VT substrate and circuit location.
A training graph is generated to allow the operator to select
the detection threshold (see Figure 60-1). Using a lower respira- The NavX System
tory threshold permits more gating and results in more accurate The localization capability of the NavX system depends on the
maps; however, this comes at the expense of time. When the use of a 5.6-kHz signal that is alternately applied across three
respiratory threshold is set low, data accuracy is high; when the pairs of skin patches affixed in orthogonal planes (x, y, and z) onto
threshold is higher, data addition to the map is faster, but map the patient’s skin. This current creates a voltage gradient that is
accuracy is compromised. The operator must use discretion sensed in all three axes to calculate the simultaneous 3D position
based on the patient’s clinical indication; indeed, a low threshold of up to 64 electrodes on up to 12 conventional catheters. These
can be set in one region in which spatial accuracy is less critical, electrodes can be displayed simultaneously in isolation or relative
and then changed to a higher level for another region in the same to the reconstructed 3D chambers.
chamber. The range of options with this system (e.g., different Mapping with NavX is typically a two-step process. First, by
mapping modes, mapping with either a standard quadripolar moving a catheter to trace the endocardial contour of the chamber
mapping catheter or a multielectrode catheter, various respira- of interest, a virtual 3D geometry is constructed. Subsequently,
tory gating strategies) can be somewhat bewildering, but ulti- sequential point-by-point mapping can be performed to generate
mately one must identify a workflow to achieve the desired color-coded maps of electrical information such as activation,
outcome. voltage amplitude, and propagation. Over time, the ability of the
For example, when mapping the left atrium and pulmonary system to perform these two steps has improved tremendously.
veins, it is preferable to perform FAM using a circular mapping By using a multielectrode catheter, it is possible to generate a
catheter (see Figure 60-1). First, respiratory gating training is computed tomography (CT)–like geometric rendering of the left
performed by placing the catheter in the left inferior pulmonary atrium and PVs; in some sense, this can be viewed as electroana-
vein (PV). For mapping the PVs, FAM is performed using the tomical imaging (see Figure 60-1). (This is also true for FAM
lowest respiratory threshold possible, with the FAM reconstruc- maps created using CARTO.) Furthermore, multielectrode spiral
tion resolution set at a high level because accuracy of these or penta-array mapping catheters can be used to create high-
regions is key to successful ablation. In addition, each PV and left density, accurate electrical maps in only a few minutes.
atrium (LA) appendage is acquired in separate maps. After The NavX system has proved itself to be particularly useful
mapping all veins, the respiratory threshold is increased some- for mapping and ablating complex atypical atrial flutters (Figure
what when acquiring the LA body. Finally, before initiating abla- 60-2).5 This process involves four steps. First, the multielectrode
tion, the magnetically localized ablation catheter is introduced to catheter is maneuvered with a deflectable sheath along the
regions of importance, such as the LA appendage ridge, to ensure chamber to create a high-density map—approximately 500 points
that they were mapped correctly. One important limitation to in less than 10 minutes. Second, the wavefronts are analyzed to
using a circular mapping catheter is the fact that the PVs can be identify potential critical isthmuses as areas of constrained activa-
somewhat stretched by the almost inevitable mismatch between tion (resulting from idiopathic or iatrogenic scars and anatomic
the catheter diameter and PV. barriers), often also containing fractionated electrograms. Third,
Once the anatomy is created, various types of electrical infor- entrainment pacing maneuvers are performed from the ablation
mation can be displayed onto the anatomy: activation maps, catheter at these sites to determine which of these wavefronts are
propagation movies, voltage amplitude maps, or even specialized actually “active” parts of the circuit versus “passive” bystanders.
maps such as CFAE maps. Because of this system’s high degree Finally, the area of slow conduction that is active in the circuit
of spatial resolution and excellent respiratory gating capabilities, and preferably in a constrained region is targeted for catheter
it is the system of choice for most physicians who ablation.
ECG
CS d
CS p
ABL
400 ms
200 ms
0 ms
–200 ms
–400 ms
Figure 60-2. Fast multielectrode mapping of atypical atrial flutter. This patient with a medical history of both catheter and surgical AF ablation had incessant atypical atrial
flutter. A multielectrode catheter (such as the penta-array or spiral catheter) was maneuvered around the LA-PV chamber to acquire a dense activation map by NavX in
approximately 5 minutes. An area of constrained activation between atrial scar and the anterosuperior mitral valve annulus was identified; at this location, the electrogram
was fractionated, and the postpacing interval after entrainment was equal to the flutter cycle length. A single lesion at this location permanently eliminated the flutter.
Because all catheters can be visualized nonfluoroscopically by

the NavX system, there is the possibility for truly fluoroless
mapping and ablation of cardiac arrhythmias, thereby reducing
or eliminating exposure to ionizing radiation to patients and staff Deployed Catheter
members. Using the NavX system, it is possible to introduce and
map the cardiac chambers to perform catheter ablation of not just
simple arrhythmias such as right-sided supraventricular tachycar-
dias (SVTs), but even complex arrhythmias such as atrial Undeployed Catheter
fibrillation.6-8 To accomplish this goal, first the diagnostic cath-
eters are visualized using the NavX system from the point of
entry into the femoral vessels to the heart. Second, the transseptal
puncture procedure is performed using intracardiac ultrasound
(discussed later) to visualize the guidewire being placed into the
superior vena cava (SVC), advancement of the transseptal sheath
over the guidewire, movement of the transseptal sheath and
needle assembly down to the interatrial septum, and puncture of
the septum into the left atrium. Third, a multielectrode catheter
is maneuvered to create a 3D cast of the left atrium and pulmo-
nary veins (LA-PVs). Finally, the ablation catheter is manipulated
to perform the ablation procedure. Although the clinical utility
of completely fluoroless catheter mapping and ablation in most
adult patients is debatable, it is clear that certain populations
derive unique benefits from this approach: children (who can
absorb tremendous amounts of radiation) and pregnant women Figure 60-3. The Rhythmia EAM System. The mapping catheter, shown with the
whose arrhythmias cannot otherwise be managed. Most impor- small mapping basket both deployed and not deployed, is maneuvered in the left
tantly, the fact that truly fluoroless atrial fibrillation (AF) ablation atrium to generate simultaneously both the LA-PV geometry and a high-density
is possible clearly indicates that with conscious effort, x-ray expo- activation map of this patient’s mitral isthmus–dependent atypical atrial flutter.
sure can be minimized to a much greater extent than is typical in
(Images courtesy of Rhythmia Medical, Burlington, MA.)
most electrophysiology laboratories.
Finally, it is interesting that just as CARTO has incorporated
both electrical and magnetic localization capabilities, the next-
generation NavX system also incorporates magnetic localization The Rhythmia System
(Mediguide, St. Jude Medical). This system is limited by the need The Rhythmia system uses a steerable small basket array catheter
for incorporating the electromagnets within the x-ray systems, of 64 electrodes to obtain a high-resolution electroanatomical
but a free-standing electromagnet is being developed to allow activation map, hopefully with minimal need for annotation
widespread use with NavX. As a result, it is reasonable to expect (Figure 60-3).9 The 64 electrodes are arranged on 8 splines
the same degree of spatial accuracy enjoyed by magnetic localiza- forming a spherical shape when deployed. Mapping is performed
tion along with the flexibility of electrical impedance–based by roving this fully- or partially-deployed catheter within the
localization. chamber while displaying electrogram and EAM information.
For electrical mapping, a suite of algorithms monitors the incom- rotational activity, and focal impulses are identified by centrifugal
60
ing data, aligns it according to a timing reference, and uses a activation from a point of origin. Rotors and focal sources are
rules-based approach to determine continuously which beat diagnosed only if stable for thousands of cycles and mapped in
should be acquired into the map and where the electrogram time-lapse fashion (multiple epochs) for longer than 10 minutes
should be annotated. To determine which beats should be (≥3000 cycles), to exclude transient pivot points of passive fibril-
acquired, the software considers multiple factors, such as respira- latory activity. Next, those areas are targeted for catheter abla-
tion and electrogram morphology. Rules are defined by the tion. Although still preliminary, the results of this approach are
operator and can be adjusted separately for each map. promising. This approach and the resulting clinical outcomes are
The anatomic shell is determined by aggregating all catheter discussed in detail in Chapter 43.
locations of acquired beats and fitting a 3D surface over them to
represent the endocardial boundary. Electrograms associated Noninvasive Electrocardiographic Imaging
with electrodes in acquired beats that are in close proximity to A noninvasive electrocardiographic imaging system (ECVUE,
the determined endocardial boundary are included in the electri- Cardioinsight, Cleveland, OH) uses an array of body surface
cal map, whereas those farther away are excluded. By collecting electrodes and CT-based cardiac geometry to map cardiac
multiple electrodes in each acquired beat and acquiring a large arrhythmias noninvasively with an accuracy of 6 mm. As dis-
number of beats per minute, a high-resolution EAM can theo- cussed in Chapter 70, this system has been used to localize various
retically be quickly created and, hopefully, with minimal user ventricular arrhythmias and to map ventricular activation in an
intervention. However, this EAM system is relatively new, and attempt to optimize cardiac resynchronization therapy. Interest-
there is not a great degree of clinical experience with it. Only ingly, this system has also been used recently to map atrial fibril-
time will determine how well this system is able to realize its goal lation noninvasively to identify rotors or focal sources that
of high-resolution mapping with minimal requirement for user maintain the rhythm (see Figure 60-4).11 To begin, a 252-electrode
intervention. vest is applied to the patient’s torso to record unipolar surface
potentials during AF, followed by a noncontrast CT scan to
obtain high-resolution images of the biatrial geometry with the
Panoramic Mapping Systems relative positions of the electrodes. Cardiac surface potentials and
unipolar electrograms are reconstructed using mathematical
Although the EAM systems require sequential mapping, albeit algorithms. Maps of AF are generated using specific algorithms
accelerated by the use of multielectrode catheters, there are other combining filtering, wavelet transform, and phase mapping. Ani-
systems with a goal for simultaneous full-chamber mapping. mations demonstrating multiple simultaneous wave propagation
There are three such panoramic mapping systems currently being patterns are recorded over a defined period, and beat-to-beat
explored in clinical practice—one old (the Ensite Array) and two changes in these patterns are color-coded and displayed on the
new (FIRM and noninvasive ECG Imaging). segmented 3D biatrial geometry, and then targeted for ablation.
The results are preliminary but promising.
Noncontact Ensite Array
The noncontact Ensite Array mapping system (EnSite 3000, St.
Jude Medical St. Paul, MN) uses a 9F multielectrode array (MEA) Remote Navigation Systems
catheter with an 8-mL ellipsoid balloon, surrounded by a braid
with 64 unipolar electrodes. The low-amplitude far-field poten- The conventional approach to cardiac mapping involves the use
tials detected by the array are mathematically enhanced and of deflectable catheters and, more recently, their combined use
resolved. An inverse solution to the Laplace equation is used to with deflectable sheaths. However, this approach requires a
reconstruct the signals detected by the MEA, and a boundary certain amount of experience to develop this technical skill, and
element method applies the inverse solution to resolve the matrix it can be prohibitive for less experienced operators when trying
of such signals at the endocardial boundary. A low-level 5.68-kHz to approach certain complex arrhythmias. To this end, there has
“locator” current from the MEA is used to locate a standard been significant interest in developing computerized remote
mapping catheter, which in turn is used to create the 3D chamber navigation systems with software interfaces to facilitate catheter
anatomy. The system reconstructs approximately 3000 virtual navigation, potentially allowing for programmed catheter
endocardial electrograms as detected by the 64 electrodes of the mapping and ablation of certain anatomic lesion sets, such as
MEA. Simultaneous acquisition of data from the entire chamber during PV isolation. In addition, remote navigation provides for
allows analysis of endocardial activation from a single beat of the possibility of reduced radiation exposure to the operator, as
tachycardia. The system has been used successfully to map and well as the potential for fewer orthopedic problems related to
guide ablation of both atrial and ventricular arrhythmias; its great- wearing protective lead aprons. Three remote navigation systems
est clinical utility seems to be for mapping transient or hemody- have been used in clinical practice.
namically unstable rhythms. However, with a few exceptions, use
of this system has largely faded from clinical practice in favor of Niobe Magnetic Navigation System
the improved EAM systems as described earlier. The Niobe system (Stereotaxis, St Louis, MO) uses two perma-
nent magnets positioned on either side of the patient table to
Focal Impulse and Rotor Map generate a 0.08-T magnetic field. The mapping and ablation
The focal impulse and rotor map (FIRM) involves the use of a catheter contains three inner magnets that align parallel to the
64-pole basket catheter (Constellation; Boston Scientific) to applied magnetic field. Catheter navigation is achieved by chang-
provide for relatively low-density panoramic atrial mapping to ing the orientation of the magnetic field, and remotely advancing
identify areas of quasi-stable rotors and focal sources purported and retracting the catheter using a motorized external module.
to sustain atrial fibrillation (Figure 60-4).10 The RhythmView The ablation catheters are designed with the magnetic localiza-
system (Topera, San Diego, CA) filters contact electrograms tion sensor to allow compatibility with the CARTO system. This
during AF to exclude noise and far-field signals using rate dynam- design allows for facile manipulation of the magnetic vectors
ics of human atrial action potential duration to estimate minimum on the EAM system. After the initial clinical experience with
activation time, and conduction velocity to identify physiologic SVT ablation demonstrated the system’s feasibility, it was
propagation paths. Electrograms are analyzed to construct iso- used for ablation of complex arrhythmias such as atrial fibrillation
potential movies of successive AF cycles, rotors are identified by and scar-related VTs.12-13 These clinical experiences have
Superior Mitral
Late
140
Lateral
Septal
0 ms
Early
A Inferior Mitral
B
Figure 60-4. Panoramic mapping systems. A, Using the RhythmView system, a 64-electrode basket catheter was placed within the left atrium (LA) during atrial fibrillation
(AF) to generate a focal impulse and rotor map (FIRM) of an inferoseptal left atrial counterclockwise rotor; ablation at this location terminated the AF. For orientation of the
FIRM, the LA is opened horizontally through the mitral valve plane, with the superior mitral folded up and the inferior mitral folded down. B, For the ECVUE system, the
patient wears the 252-electrode surface vest while undergoing a noncontrast computed tomographic scan. In this patient with persistent AF despite prior extensive LA
ablation, when the atrial activation map computed from these multiple surface electrocardiograms was projected onto the segmented biatrial anatomy, a clockwise rotor
driving AF was identified at the base of the RA appendage.
(A, FIRM image courtesy Sanjiv Narayan, University of California–San Diego. B, ECVUE map courtesy Michel Haissaguerre, Hôpital Cardiologique du Haut-Leveque, Bordeaux,
France.)
demonstrated feasibility, but it is still less clear that they provide catheter in a stable location despite the biologic spatial noise
equivalent clinical outcomes, both from the perspective of the related to cardiac and respiratory motion. The initial clinical
ease of performing the procedure and the efficacy of the proce- experience with this system is early but promising.
dure. It is hoped that with the most recent version of this system,
which additionally incorporates a remotely deflectable sheath, Sensei Robotic Navigation System
the procedure outcomes will be improved further. The Sensei system (Hansen Medical, Sunnyvale, CA) facilitates
catheter navigation using two steerable sheathes incorporating an
CGCI Magnetic Navigation System ablation catheter. The outer and inner sheaths are both manipu-
The CGCI system (Magnetecs, Los Angeles, CA) is a real-time, lated via a pull-wire mechanism by a sheath carrying a robotic
high-speed, closed-loop, magnetic navigation system that uses arm that is fixed to the patient table. The robotic arm obeys the
eight electromagnets to create dynamically shaped magnetic commands of the central workstation positioned in the control
fields around the patient’s torso.14 The real-time reshaping of room. Catheter navigation using a 3D joystick allows a broad
these magnetic fields produces the appropriate 3D motion or range of motion in any direction. This navigation system has
change in direction of a magnetized ablation catheter. The CGCI been used in concert with both the NavX and CARTO EAM
is integrated with the NavX 3D EAM system and allows for both systems for the ablation of a broad range of arrhythmias, includ-
joystick and automated navigation. One of the theoretical advan- ing AF and VT.15-16 This system has clearly demonstrated clinical
tages of this system is related to the fact that it uses electromag- feasibility, but it is unclear that clinical outcomes are equivalent
nets instead of fixed magnets. As a result, the magnetic field can or superior to standard manual mapping—both from the per-
be altered in a fraction of a second, thus allowing for computer- spective of the ease of performing the procedure and the proce-
controlled closed-loop software algorithms to constantly keep the dure’s safety and efficacy.
guidance within the patient-specific 3D model of the chamber.

Imaging
60
To accomplish this paradigm of image integration, the preac-
quired 3D image must be accurately registered with the EAM
The importance of imaging during catheter ablation procedures system—that is, both CT/MRI and EAM constructs must be
has grown over the past decade. The various imaging approaches aligned precisely. Malchano et al.23,24 first reported on the pre-
can be divided into those that are acquired before a procedure clinical feasibility and accuracy of image integration for VT abla-
versus those obtained in real time or near real time. tion, and then on the clinical feasibility of image integration for
AF ablation (Figure 60-5). Using custom software, they demon-
strated that the spatial accuracy for ventricular registration was
Preacquired Imaging on the order of less than 2 mm of misregistration. Furthermore,
accurate LA-PV registration was feasible, but it was critical to
Preoperative CT and magnetic resonance imaging (MRI) pro- maintain the same phase of respiration during image acquisition
vides detailed anatomic and physiological information about as during mapping (i.e., the end of quiet expiration). Since these
normal and damaged myocardial tissue. For AF ablation proce- initial experiments with custom software, both the NavX and
dures, an understanding of PV anatomy is critical; it is clear that CARTO systems have developed commercial software solutions
there can be tremendous variation in both the number of PVs, size to allow for image integration of either CT or MR images with
of the PVs and LA chamber, and the morphology of the chamber real-time EAM.
(e.g., the size and depth of the ridge between the LA appendage For AF ablation, there have been a number of studies that
and left PVs). CT also serves as another screening procedure to examined the effect of image integration; some have demon-
identify LA appendage thrombus. For VT ablation, CT can also strated improvements in clinical outcome compared with using
provide an understanding of both the geometry of the ventricle EAM alone, but others have not seen any such improvements.25-28
and the location and size of the coronary arteries—critical infor- In some studies, there were improvements in procedural param-
mation when performing epicardial mapping and ablation. Occa- eters such as fluoroscopy and procedure times. These differences
sionally, CT imaging can identify other unsuspected pathologies, in outcome are probably related to a host of factors, including
such as lung lesions, that might require management. experience with image integration and the baseline level of expe-
Unlike CT, MRI has the advantage of not exposing the patient rience with LA-PV mapping and ablation. Ultimately, each oper-
to ionizing radiation. Like CT, magnetic resonance (MR) angi- ator must choose whether the marginal time required to register
ography of the LA-PVs can also provide a volumetric image that the anatomy provides enough clinical utility to warrant its use.
can be used to define the chamber size and morphology. In addi- However, there is little doubt that image integration is particu-
tion, MRI with gadolinium-delayed enhancement (DE-MRI) can larly useful in cases with unusual anatomies, such as in congenital
identify myocardial scar tissue. For VT ablation procedures, heart disease (see Figure 60-5).
DE-MRI can provide a 3D representation of the location and For VT ablation, image integration has been used less rou-
epicardial extent of the scar. This knowledge can be particularly tinely, in part because of the technical complexity of obtaining
important in nonischemic cardiomyopathies in which the myo- MR images in patients with implantable cardioverter defibrilla-
cardial scar has a higher incidence of being either midmyocardial tors. However, image integration is being used more often during
or epicardial.17 There are early data suggesting the possibility of VT ablation as the field has gained experience with the use of
using DE-MRI to identify conducting channels of live myocar- MRI in patients with defibrillators and with the use of CT-based
dial tissue interspersed within scar (i.e., potentially visualizing the imaging modalities (Figure 60-6), including imaging with posi-
VT circuit itself); however, this work is preliminary.18 Finally, tron emission tomography–CT.29-31 Again, image integration for
DE-MRI has recently been reported to be able to identify atrial VT ablation is particularly useful when a scar is mid-myocardial
scar, and the extent of this atrial scar has been reported to cor- and thus difficult to identify with electrogram-based substrate
relate with both the stroke risk and the clinical outcome after AF mapping alone.17,31
catheter ablation.19-20
Integration of CT and MRI with Fluoroscopy Real-Time or Near–Real-Time Imaging

In addition to simply examining the preacquired CT/MR images
to gain a 3D understanding of the chamber anatomy or scar loca- Preacquired CT/MR images have the advantage of being high
tion, it is possible to integrate segmented 3D renderings of the resolution, even if they are acquired the day before the procedure;
MR/CT anatomies with most modern fluoroscopy systems. In however, image registration can be frustrated by anatomic vari-
patients undergoing AF ablation, preacquired CT scans of the ances related to differences in the volume state, cardiac rhythm,
LA-PVs can be segmented and then registered to live fluoroscopy and respiratory phase. Ideally, 3D volumetric images would be
with a simple transformation linking—for example, the superior obtained real time to minimize these errors.
vena cava and the coronary sinus from the CT model with a
catheter placed inside the coronary sinus via the superior vena Rotational Angiography
cava.21 Alternatively, preacquired MR images can be segmented Rotational angiography involves creating a 3D reconstruction of
into 3D anatomies and then registered with live fluoroscopy the LA-PVs by rapidly rotating the cardiac C-arm while perform-
using contrast injection into the cardiac chamber for integra- ing a contrast atriogram. Initially, this technique was performed
tion.22 For both, the integrated image can be projected onto the by injecting contrast using a standard pigtail catheter placed into
fluoroscopy screen such that the highly-defined 3D anatomy is the pulmonary artery, and performing the C-arm rotation along
observed as a “ghost” on the live fluoroscopy image. The advan- a circular 198-degree arc during an inspiratory breath-hold with
tage of fluoroscopy integration is the relative ease with which it the arms placed above the patient’s head (to avoid shadowing of
can be performed, because the integration software is available the humerus).32 The segmented 3D LA-PV image can then be
on most modern fluoroscopy systems. transferred and registered to fluoroscopy or an EAM system, or
both (Figure 60-7).33 At least one study has suggested that the
Integration of CT and MR Imaging with Electroanatomical integration of 3D images obtained by rotational angiography into
Mapping Systems EAM systems may be superior to integration of 3D CT/MR
The 3D MR/CT imaging data can be integrated directly into the images.34 Another randomized study comparing CARTO with
EAM system, which allows for catheter visualization and rotational angiography integrated with fluoroscopy found similar
LIPV
1.28 cm
A Z. Malchano/V. Reddy, 2003

-165ms
Left Atrium
ts
Right
Atrium
-291ms
Common
Ventricle
B 2.00 cm
Figure 60-5. Computed tomographic (CT) and magnetic resonance (MR) image integration with electroanatomical mapping (EAM) systems. A, Image integration was
performed using custom software to allow the placement of ablation lesions directly on the more realistic segmented three-dimensional (3D) CT/MR anatomy (left panel).
Since then, image integration of preacquired 3D CT/MR anatomies is commercially available with both the NavX (middle panel) and CARTO EAM systems. More recently,
multimodality image integration has become possible, shown here in an example of additional integration of intracardiac echocardiography such that the location of the
ultrasound image is shown in real time relative to the 3D CT image and CARTO (right panel). B, 3D CT/MR image integration is of particular utility for ablation of arrhythmias
in the setting of congenital heart disease. Shown are the segmented 3D geometries of the cardiac chambers derived from a preacquired CT image in a patient with atypical
atrial flutter and a history of univentricular physiology and a prior Fontan surgery with an extracardiac total cavopulmonary connection. With image integration using
CARTO, what could otherwise have been a complex procedure was instead a straightforward one of an atypical flutter potentiated by the surgical scar.
overall fluoroscopy times and procedure times in both treatment access, and it can guide accurate placement of diagnostic and
arms, thereby suggesting that the major advantage of EAM ablation catheters to allow for safe and effective titration of
systems (reduced x-ray exposure) can be replaced by using a 3D energy and early recognition of complications. In typical clinical
rotational image, albeit without the advanced mapping functions use, the phased-array ICE catheter (the most commonly used
of the EAM system.35 The major clinical issue with rotational ICE probe) is introduced via the femoral vein to the right atrium.
angiography is one of workflow because acquisition of the rota- An initial quick scan allows delineation of the overall atrial
tional image does take time at the beginning of the procedure. anatomy, interatrial septum, and PVs to identify any anomalies
To this end, there has been one important procedural improve- (or during VT ablation, a quick scan of the left ventricular
ment: direct LA injection via a transseptal approach during rapid chamber might identify the highly refractile myocardial scar,
ventricular pacing permits sufficient opacification of the LA- which can be subendocardial, midmyocardial, or epicardial). ICE
PV anatomy that a good 3D image can be obtained, even when can evaluate the LA-PVs and particularly the LA appendage to
using a 159-degree arc with both arms at the patient side (see verify the absence of thrombus; by placing the ICE probe in the
Figure 60-7). pulmonary artery, it is possible to obtain exquisite images of the
interstices of the LA appendage. Although data are insufficient
Intracardiac Echocardiography Imaging to comment on the relative sensitivity and specificity of this
Ultrasound imaging has the advantage of providing continuous approach relative to transesophageal echocardiographic imaging,
real-time imaging of the cardiac anatomy. Although transesopha- there is one recent clinical report in which ICE was able to accu-
geal echocardiography is not practical for continuous use during rately adjudicate the presence or absence of LAA thrombus in
ablation procedures, intracardiac echocardiography (ICE) has patients with a diagnosis of equivocal LAA thrombus by trans-
proved itself useful, particularly during AF ablation. Specifically, esophageal echocardiography.36
it allows facile identification of relevant anatomic structures, it is ICE can identify variants of the septum that might complicate
a complementary tool to fluoroscopy to guide safe transseptal the transseptal puncture step, including the presence of a thick
500 ms V1
II
III
V6
ABLd
450 60
aVR
aVF ABLp
aVL 450 472
RV
V1
V2
V3 7.98 mV
V4
V5
V6A
1.50 mV
1.51 mV
0.03 mV
0.03 mV
0.10 mV
1.90 c
B C
Figure 60-6. Computed tomographic (CT) image integration for ventricular tachycardia (VT) ablation. A, The substrate of this patient’s VT was hypertrophic cardiomyopathy;
shown in the CT image is a left ventricular (LV) apical aneurysm with midventricular obstruction. B, The segmented 3D LV endocardial surface and aortic root with proximal
coronary arteries are registered with a voltage amplitude CARTO map of the LV endocardium (red represents low voltage, which equals diseased or scarred myocardium).
C, Mapping the ventricular epicardial surface delineated the epicardial extent of the scar, and importantly, a site from which entrainment identified a good target for
catheter ablation. Given the absence of an epicardial coronary vessel at this location, radiofrequency energy was delivered to this site to eliminate the VT.
septum, aneurysmal septum, double membrane septum, patent and to deliver therapeutic ultrasound energy for ablation (Figure
foramen ovale, atrial septal defect, or other closure and repair 60-8). This robotically driven catheter is programmed to scan the
devices. ICE also helps in determining the exact location of the tissue in a predetermined fashion. The M-mode ultrasound
transseptal puncture, ideally in the posteroinferior aspect of the data allow the system to identify the distance from the catheter
fossa ovalis for most standard AF ablation procedures, or antero- tip to the tissue, thereby allowing a 3D reconstruction of the
inferior for transseptal left ventricular mapping. The location of chamber geometry. Next, the operator can specify the path of
the transseptal puncture can be especially critical for balloon- ablation and allow the system to create the lesion set with minimal
based AF ablation procedures. An inferior puncture is typically technical interaction by the operator. Although this system has
important to achieve good apposition with and isolation of the been used only preclinically and is just starting first-in-man clini-
right inferior pulmonary vein. As discussed earlier, ICE can even cal testing, the concept underscores the power of ultrasound as
guide the transseptal puncture without the use of fluoroscopy. an imaging technology. Indeed, future work might demonstrate
During catheter manipulation, ICE can determine (1) the that the M-mode image is of high enough resolution to deter-
position of the circular mapping catheter relative to the PV mine the approximate thickness of the tissue; this information
ostium, (2) whether the ablation catheter is actually in contact might allow for online power titration tailored to the target
with the target tissue during ablation, (3) whether there is good tissue.
contact or occlusion with the target PV during balloon-based
ablation, and (4) whether there are any complications such as Interventional Magnetic Resonance Imaging
cardiac perforation with impending cardiac tamponade. There are a number of reasons why intraprocedure MRI could
Most recently, by placing a magnetic localization sensor in the be attractive for guiding electrophysiology procedures. First,
ICE catheter to allow for precise localization and orientation of MRI offers the possibility of both substrate visualization and
the scanning transducer in space, ultrasound-based near–real- ablation lesion imaging.37-39 In addition, the ability to obtain
time reconstruction of cardiac anatomy is possible (see Figure images in arbitrary orientations opens the potential for high
60-5). Multiple scanning planes are obtained by rotating the quality visualization of catheters, anatomy, and electrode tissue
probe, and the chamber’s endocardial surface is traced manually contact. Furthermore, the position errors introduced by register-
or by edge enhancement technology. This process forms a series ing the catheter to preacquired 3D images can be largely avoided
of atrial rings that are connected to create a 3D reconstruction because both real-time MR images and 3D MR images are
of the LA-PVs. acquired in the same coordinate system, and 3D images could be
reacquired multiple times during the procedure if needed.
Collimated Ultrasound-Guided Imaging and Ablation Continuous MRI has been used to guide nonferromagnetic
One example of a potential future application of ultrasound is electrophysiology (EP) catheter positioning from an internal
demonstrated by the LICU system (Vytronus, Sunnyvale, CA). jugular vein to selected atrial and ventricular locations, as well as
This catheter mapping and ablation system uses a narrow, unfo- to create and monitor ablations in vivo. The ability to perform
cused beam of collimated ultrasound to image tissue by M-mode the transseptal puncture procedure using real-time MR guidance
EAM-CT
EAM-RotX
B C
Figure 60-7. Rotational angiography. A, Rotational angiography was performed with direct LA injection during rapid ventricular pacing to generate this three-dimensional
(3D) LA-PV anatomy. B, The segmented 3D rotational angiography image can be displayed as a ghost atop live x-ray so as to better appreciate the locations of the diagnostic
and therapeutic catheters relative to the LA-PV anatomy. C, The 3D image can also be integrated into electroanatomical mapping systems. Shown for comparison are com-
puted tomographic (CT); top) and rotational angiogram–derived (bottom) 3D LA-PV geometries registered with CARTO during an atrial fibrillation ablation procedure.
2D Image 3D Image
1–2 mm wide
M-Mode
A C
Lesion Path
B D
Figure 60-8. Collimated ultrasound-guided imaging and ablation. A, This system uses an unfocused collimated ultrasound beam for both M-mode ultrasound imaging
and therapeutic ultrasound ablation. B, The beam is driven by a robotically controlled catheter to maneuver precisely within the cardiac chamber. C, For imaging, the
chamber surface is automatically identified by M-mode imaging scan (left panel) to allow the system to generated a two-dimensional map that is color-coded such that
red-yellow represent far distances and blue-purple close distances (middle panel); a three-dimensional volume can then be reconstructed (right panel). D, For ablation, the
operator indicates the desired lesion path for the robotically driven catheter to follow. Shown is an example of the selected lesion path (left panel) and consequent linear
lesion on gross pathology (right panel) in the porcine left atrium.
60
A B C
D E F
Figure 60-9. Interventional magnetic resonance imaging (MRI). A, This MRI-compatible catheter was constructed with five coils to allow for active tracking with MRI. B,
This allows for real-time visualization of the catheter atop two-dimensional (2D) planar MRI images or segmented three-dimensional (3D) volume reconstructions (inset).
C, In this porcine model of healed myocardial infarction, a voltage amplitude map of the anteroapical left ventricle (LV) myocardial scar is projected onto the segmented
3D LV anatomy. D, During LA-PV mapping, the catheter can again be visualized on the 2D planar MRI image. E, The catheter is depicted within the 3D volume reconstruc-
tion of the LA-PVs during peri-PV radiofrequency ablation. F, On gross pathologic examination, the corresponding ablation lesions are present.
has also been demonstrated. In porcine models of chronic myo- permits direct visualization of the LA-PV junction (Figure
cardial infarction, MR tracking has been used to navigate cath- 60-10).43 As a result, an optical fiber within the central shaft of
eters to the left ventricle, to measure electrogram activity, and to the catheter can be easily manipulated circumferentially to deliver
render accurate 3D voltage maps, completely in the MRI envi- a short arc of ablative laser energy to the visualized tissue such
ronment (Figure 60-9).40 Furthermore, porcine left atrial that each lesion adequately overlaps with the previous lesion. As
mapping, radiofrequency ablation at the PV ostia, and AV node during surgery, the operator can visualize the target tissue and
ablation were demonstrated under MRI guidance (see Figure adapt the lesion strategy (including energy dosing) to the highly
60-9).41 These and other experiments have demonstrated the variable LA-PV anatomy. Consistent with preclinical porcine
potential for truly interventional MRI for electrophysiology pro- data revealing that surface overlap of contiguous ablation lesions
cedures. However, much work is still necessary before this predicts a greater than 90% rate of lesion transmurality and
imaging approach is ready for clinical use, (1) improving and durable PV isolation, in a multicenter clinical study in which
simplifying the pulse sequences, (2) making the necessary diag- patients underwent PV isolation with the visually guided laser
nostic or therapeutic catheters, sheaths, and guidewires compat- and then underwent a protocol-mandated EP study approxi-
ible with MRI and visible, and (3) making all the monitoring mately 3 months later to assess the durability of PV isolation, the
equipment necessary in an EP lab to be compatible with MRI.42 rate of durable PV isolation was 86%.44,45 This technology has
ushered in a completely novel approach to catheter ablation that
Direct Visually Guided Ablation more closely emulates cardiac surgery.
Although EAM systems allow the operator to catalogue the loca-
tion of the ablation lesions, generating a continuous contiguous
lesion set, as necessary for PV isolation, is certainly not assured
by a continuous line of lesion markers on the anatomical map. Conclusions
To this end, one approach to achieving continuous lesions is
direct visual guidance; as such, there is a visually guided laser Over the years, technology has rapidly progressed from two-
balloon ablation catheter that is equipped with an endoscope that dimensional fluoroscopy to sophisticated 3D EAM systems using
RSPV
“Static” Blood
Aiming in LSPV
Beam
LSPV
LIPV
LAA
Ablation
RSPV Lesions
B C E
Figure 60-10. Pulmonary vein (PV) isolation with a visually guided laser catheter. A, The compliant balloon of this catheter can be inflated to variable dimension. B, The
projected green arc or spot can be easily maneuvered circumferentially around the balloon to direct the location of laser ablation. C, In this in vivo image, the tip of the
catheter is within the left superior PV. The aiming beam is projected onto the blanched tissue while blood is red. The laser is projected to the same location of the aiming
beam to ablate the visualized tissue. D, In this porcine model, a contrast injection outlines the right superior PV. E, After visually guided laser ablation, this gross pathologic
specimen reveals continuous circumferential lines of ablation around the right and left superior PVs.
magnetic- and impedance-based sensors, as well as volumetric capable of completing the requisite task, the approach one ulti-
preprocedure and intraprocedural imaging modalities derived mately chooses will be based on the practicality of the workflow,
from CT, MRI, ICE, and rotational angiography. The future the attendant radiation exposure to both the patient and EP lab
might involve additional refinements in these technologies, or staff, and increasingly importantly, the cost associated with the
perhaps other real-time imaging modalities such as interventional technology. Indeed, as economic pressures continue to supervene
MRI with intraprocedural lesion imaging, robotically controlled in health care, the ability to obtain exquisite pictures will not be
collimated ultrasound for both imaging and ablation, or endo- sufficient justification for the routine use of any imaging technol-
scopically driven direct visual imaging of the target tissue. ogy during EP procedures in the absence of tangible benefits and
Although most, if not all, of these approaches will likely be demonstrated cost effectiveness.
three-dimensional mapping. Pacing Clin Electro- 13. Aryana A, d’Avila A, Heist EK, et al: Remote mag-
References physiol 30:510–518, 2007. netic navigation to guide endocardial and epicar-
7. Tuzcu V: A nonfluoroscopic approach for electro- dial catheter mapping of scar-related ventricular
1. Waldo AL: Atrial flutter: entrainment characteris- physiology and catheter ablation procedures using tachycardia. Circulation 115(10):1191–1200, 2007.
tics. J Cardiovasc Electrophysiol 8(3):337–352, a three-dimensional navigation system. Pacing 14. Gang ES, Nguyen BL, Shachar Y, et al: Dynami-
1997. Clin Electrophysiol 30:519–525, 2007. cally shaped magnetic fields: initial animal valida-
2. Stevenson WG, Sager PT, Friedman PL: Entrain- 8. Reddy VY, Morales G, Ahmed H, et al: Catheter tion of a new remote electrophysiology catheter
ment techniques for mapping atrial and ventricular ablation of atrial fibrillation without the use of guidance and control system. Circ Arrhythm Elec-
tachycardias. J Cardiovasc Electrophysiol 6(3):201– fluoroscopy. Heart Rhythm 7:1644–1653, 2010. trophysiol 4:770–777, 2011.
216, 1995. 9. Nakagawa H, Ikeda A, Sharma T, et al: Rapid high 15. Reddy VY, Neuzil P, Malchano ZJ, et al: View-
3. Cappato R, Schlüter M, Weiss C, et al: Catheter- resolution electroanatomical mapping: evaluation synchronized robotic image-guided therapy for
induced mechanical conduction block of right- of a new system in a canine atrial linear lesion atrial fibrillation ablation: experimental validation
sided accessory fibers with Mahaim-type model. Circ Arrhythm Electrophysiol 5(2):417– and clinical feasibility. Circulation 115(21):2705–
preexcitation to guide radiofrequency ablation. 424, 2012. 2714, 2007.
Circulation 90(1):282–290, 1994. 10. Narayan SM, Franz MR, Clopton P, et al: Repo- 16. Valderrábano M, Dave AS, Báez-Escudero JL,
4. Friedman PL, Dubuc M, Green MS, et al: Cath- larization alternans reveals vulnerability to human et al: Robotic catheter ablation of left ventricular
eter cryoablation of supraventricular tachycardia: atrial fibrillation. Circulation 123:2922–2930, tachycardia: initial experience. Heart Rhythm
results of the multicenter prospective “frosty” trial. 2011. 8(12):1837–1846, 2011.
Heart Rhythm 1(2):129–138, 2004. 11. Haissaguerre M, Hocini M, Shah AJ, et al: 17. Yoshida K, Yokokawa M, Desjardins B, et al: Septal
5. Patel AM, d’Avila A, Neuzil P, et al: Atrial tachy- Noninvasive panoramic mapping of human atrial involvement in patients with post-infarction ven-
cardia after ablation of persistent atrial fibrillation: fibrillation mechanisms: a feasibility report. J Car- tricular tachycardia: implications for mapping and
identification of the critical isthmus with a combi- diovasc Electrophysiol 2012; doi:10.1111/ radiofrequency ablation. J Am Coll Cardiol
nation of multielectrode activation mapping and jce.12075. 58(24):2491–2500, 2011.
targeted entrainment mapping. Circ Arrhythm 12. Miyazaki S, Shah A, Xhaet O, et al: Remote mag- 18. Perez-David E, Arenal A, Rubio-Guivernau JL,
Electrophysiol 1:14–22, 2008. netic navigation with irrigated tip catheter for abla- et al: Noninvasive identification of ventricular
6. Smith G, Clark J: Elimination of fluoroscopy use tion of paroxysmal atrial fibrillation. Circ Arrhythm tachycardia-related conducting channels using
in a pediatric electrophysiology laboratory utilizing Electrophysiol 3(6):585–589, 2010. contrast-enhanced magnetic resonance imaging in
patients with chronic myocardial infarction. 28. Caponi D, Corleto A, Scaglione M, et al: Ablation 36. Ren JF, Marchlinski FE, Supple GE, et al: Intra-
J Am Coll Cardiol 57:184–194, 2011. of atrial fibrillation: Does the addition of three- cardiac echocardiographic diagnosis of thrombus
19. Mahnkopf C, Badger TJ, Burgon NS, et al: Evalu-
ation of the left atrial substrate in patients with lone
atrial fibrillation using delayed-enhanced MRI:
dimensional magnetic resonance imaging of the
left atrium to electroanatomical mapping improve
the clinical outcome?: A randomized comparison
formation in the left atrial appendage: a comple-
mentary role to transesophageal echocardiography.
Echocardiography 30(1):72–80, 2013.
60
implications for disease progression and response of carto-merge vs. Carto-xp three-dimensional 37. Lardo AC, McVeigh ER, Jumrussirikul P, et al:
to catheter ablation. Heart Rhythm 7(10):1475– mapping ablation in patients with paroxysmal and Visualization and temporal/spatial characterization
1481, 2010. persistent atrial fibrillation. Europace 12:1098– of cardiac radiofrequency ablation lesions using
20. Daccarett M, Badger TJ, Akoum N, et al: Associa- 1104, 2010. magnetic resonance imaging. Circulation 102:698–
tion of left atrial fibrosis detected by delayed- 29. Tian J, Jeudy J, Smith MF, et al: Three-dimensional 705, 2000.
enhancement magnetic resonance imaging and the contrast-enhanced multidetector CT for anatomic, 38. Kolandaivelu A, Zviman MM, Castro V, et al:
risk of stroke in patients with atrial fibrillation. dynamic, and perfusion characterization of abnor- Noninvasive assessment of tissue heating during
J Am Coll Cardiol 57(7):831–838, 2011. mal myocardium to guide ventricular tachycardia cardiac radiofrequency ablation using mri ther-
21. Sra J, Krum D, Malloy A, et al: Registration of ablations. Circ Arrhythm Electrophysiol 3:496– mography. Circ Arrhythm Electrophysiol 3:521–
three-dimensional left atrial computed tomo- 504, 2010. 529, 2010.
graphic images with projection images obtained 30. Andreu D, Berruezo A, Ortiz-Pérez JT, et al: Inte- 39. Vergara GR, Vijayakumar S, Kholmovski EG, et al:
using fluoroscopy. Circulation 112:3763–3768, gration of 3D electroanatomical maps and mag- Real-time magnetic resonance imaging-guided
2005. netic resonance scar characterization into the radiofrequency atrial ablation and visualization of
22. Ector J, DeBuck S, Adams J, et al: Cardiac three- navigation system to guide ventricular tachycardia lesion formation at 3 tesla. Heart Rhythm 8:295–
dimensional magnetic resonance imaging and ablation. Circ Arrhythm Electrophysiol 4:674–683, 303, 2011.
fluoroscopy merging: a new approach for electro- 2011. 40. Dukkipati SR, Mallozzi R, Schmidt EJ, et al: Elec-
anatomical mapping to assist catheter ablation. 31. Dickfeld T, Tian J, Ahmad G, et al: MRI-guided troanatomical mapping of the left ventricle in a
Circulation 112:3769–3776, 2005. ventricular tachycardia ablation: integration of late porcine model of chronic myocardial infarction
23. Malchano ZJ, Neuzil P, Cury RC, et al: Integration gadolinium-enhanced 3d scar in patients with with magnetic resonance–based catheter tracking.
of cardiac CT/MR imaging with three-dimensional implantable cardioverter-defibrillators. Circ Circulation 118(8):853–862, 2008.
electroanatomical mapping to guide catheter Arrhythm Electrophysiol 4:172–184, 2011. 41. Schmidt EJ, Mallozzi RP, Thiagalingam A,
manipulation in the left atrium: implications for 32. Thiagalingam A, Manzke R, D’Avila A, et al: Intra- et al: Electroanatomical mapping and radiofre-
catheter ablation of atrial fibrillation. J Cardiovasc procedural volume imaging of the left atrium and quency ablation of porcine left atria and atrioven-
Electrophysiol 17:1221–1229, 2006. pulmonary veins with rotational x-ray angiography: tricular nodes using magnetic resonance catheter
24. Reddy VY, Malchano ZJ, Holmvang G, et al: Inte- Implications for catheter ablation of atrial fibrilla- tracking. Circ Arrhythm Electrophysiol 2:695–
gration of cardiac magnetic resonance imaging tion. J Cardiovasc Electrophysiol 19:293–300, 704, 2009.
with three-dimensional electroanatomical mapping 2008. 42. Nordbeck P, Bauer WR, Fidler F, et al: Feasibility
to guide left ventricular catheter manipulation: fea- 33. Li JH, Haim M, Movassaghi B, et al: Segmentation of real-time MRI with a novel carbon catheter for
sibility in a porcine model of healed myocardial and registration of three-dimensional rotational interventional electrophysiology. Circ Arrhythm
infarction. J Am Coll Cardiol 44(11):2202–2213, angiogram on live fluoroscopy to guide atrial fibril- Electrophysiol 2:258–267, 2009.
2004. lation ablation: A new online imaging tool. Heart 43. Reddy VY, Neuzil P, Themistoclakis S, et al:
25. Kistler PM, Rajappan K, Harris S, et al: The Rhythm 6:231–237, 2009. Visually-guided balloon catheter ablation of atrial
impact of image integration on catheter ablation of 34. Nolker G, Gutleben KJ, Marschang H, et al: fibrillation: Experimental feasibility and first-in-
atrial fibrillation using electroanatomical mapping: Three-dimensional left atrial and esophagus recon- human multicenter clinical outcome. Circulation
A prospective randomized study. Eur Heart J struction using cardiac c-arm computed tomogra- 120:12–20, 2009.
29:3029–3036, 2008. phy with image integration into fluoroscopic views 44. Dukkipati SR, Neuzil P, Skoda J, et al: Visual
26. Della Bella P, Fassini G, Cireddu M, et al: Image for ablation of atrial fibrillation: Accuracy of a balloon-guided point-by-point ablation: reliable,
integration-guided catheter ablation of atrial fibril- novel modality in comparison with multislice com- reproducible, and persistent pulmonary vein isola-
lation: A prospective randomized study. J Cardio- puted tomography. Heart Rhythm 5:1651–1657, tion. Circ Arrhythm Electrophysiol 3(3):266–273,
vasc Electrophysiol 20:258–265, 2009. 2008. 2010.
27. Bertaglia E, Bella PD, Tondo C, et al: Image inte- 35. Knecht S, Wright M, Akrivakis S, et al: Prospective 45. Dukkipati SR, Neuzil P, Kautzner J et al: The dura-
gration increases efficacy of paroxysmal atrial randomized comparison between the conventional bility of pulmonary vein isolation using the visually
fibrillation catheter ablation: Results from the car- electroanatomical system and three-dimensional guided laser balloon catheter: multicenter results
tomerge Italian registry. Europace 11:1004–1010, rotational angiography during catheter ablation for of pulmonary vein remapping studies. Heart
2009. atrial fibrillation. Heart Rhythm 7:459–465, 2010. Rhythm 9(6):919–925, 2012.
CT and MRI for Electrophysiology 61
Hiroshi Ashikaga, Aravindan Kolandaivelu,
Saman Nazarian, and Henry R. Halperin
CHAPTER OUTLINE involvement is determined by the underlying genotype. In

Diagnosis 595 patients harboring plakophilin mutations (PKP2), LV abnormal-
ity often occurs in the form of focal epicardial fat infiltration of
Prognosis 596 the LV lateral wall with preserved LV function.5 On the other
Procedure Guidance 597 hand, mutations in the desmoplakin (DSP) gene are associated
with LV functional involvement. ARVD/C is also associated with
Summary 601 RV mechanical dyssynchrony in up to 50% of cases, and with RV
remodeling.6 Minor MRI abnormalities in RV structure, espe-
cially in the RV outflow tract, have been reported in up to 60%
Cardiac computed tomography (CT) and magnetic resonance of mutation carriers and in none of the noncarriers5; however,
imaging (MRI) have become essential tools for evaluating patients the prognostic value of these findings remains unclear.
with cardiac arrhythmias. Multidetector CT (MDCT) and MRI
have complementary advantages; MDCT provides crucial struc-
tural information with high resolution within a short acquisition Cardiac Sarcoidosis
time, whereas MRI has unparalleled soft tissue contrast resolu-
tion, particularly with the use of late gadolinium enhancement Sarcoidosis is a multisystem granulomatous disease of unknown
(LGE). Of essence is the ability of these imaging modalities to origin that is characterized by the presence of noncaseating gran-
identify the regions of abnormal tissue characteristics, which ulomas in involved organs. Manifestations of cardiac sarcoidosis
often coincide with regions with abnormal electrophysiology. include conduction block, dilated cardiomyopathy (DCM), heart
This critical advantage bridges arrhythmia and imaging through failure, ventricular arrhythmia, and SCD. Diagnostic reliance of
correlation between anatomical and electrophysiological sub- the criteria on subjective assessment of LV wall thinning, wall
strates, and has accelerated the evolution of image-based electro- motion abnormalities, and dilatation of the LV on echocardiog-
physiological intervention. raphy turned out to be insensitive for diagnosis.7 For example, at
least some of the patients previously diagnosed with DCM are
re-diagnosed with cardiac sarcoidosis once myocardial tissue is
available for histologic analysis after ventricular assist device
Diagnosis implantation.8 Technical advances in cardiac MRI have improved
its capability to image the myocardial scar associated with cardiac
Arrhythmogenic Right Ventricular Dysplasia/ sarcoidosis, and cardiac MRI plays a critical role in the diagnosis
Cardiomyopathy (ARVD/C) of cardiac sarcoidosis in the 2006 updated diagnostic criteria put
forth by the Ministry of Health, Labor, and Welfare of Japan (Box
ARVD/C is a genetic disease characterized by fibrofatty infiltra- 61-2).9 Diagnostic workup of cardiac sarcoidosis involves cardiac
tion of the right ventricle (RV) and less commonly the left ven- MRI with LGE to detect myocardial damage (Figure 61-1, B),
tricle (LV). An implantable cardioverter-defibrillator (ICD) and 18F-fluorodeoxyglucose (FDG)-positron emission tomogra-
provides reasonable therapy for patients who have one or more phy (PET) to detect disease activity.10,11 MRI is sensitive for
risk factors for sudden cardiac death (SCD), including inducible cardiac involvement, and a positive MRI finding is associated
ventricular tachycardia (VT) during electrophysiological testing.1 with future adverse events including cardiac death.12
Therefore, the diagnosis of ARVD/C has a significant impact not
only on the proband with clinical events but also on the family
members. Cardiac MRI plays a critical role in the diagnosis of Myocarditis
ARVD/C in recently revised diagnostic criteria (Box 61-1).2 MRI
overcomes the limitations of conventional imaging modalities Severe myocarditis presents as DCM, heart failure, ventricular
such as echocardiography and provides multiplanar images of the arrhythmia, and SCD. Cardiac MRI provides valuable clinical
right ventricle, allowing accurate evaluation of global and regional information on abnormal tissue characteristics associated with
RV function. Further, MRI provides RV tissue characterization myocarditis,13 including (1) intracellular and interstitial edema
by depicting both fat infiltration and fibrosis (Figure 61-1, A). (T2-weighted edema imaging), (2) hyperemia and capillary
Although detecting RV fibrosis is challenging because of the thin leakage (myocardial early gadolinium enhancement [EGEr]), and
wall of the RV, quantitative evaluation of the RV ejection fraction (3) necrosis and fibrosis (LGE).14 T2-weighted edema imaging is
(EF) by cardiac MRI provides high sensitivity and specificity.3 used to evaluate initial changes in myocardial tissue during the
The sensitivity of cardiac MRI ranges from 79% to 89% for first phase of myocardial inflammation.15,16 Myocardial edema
major criteria and from 68% to 78% for minor criteria, with appears as an area of high signal intensity on T2-weighted
specificity of 96% to 100%.1 ARVD/C is associated with regional images. EGEr uses T1-weighted images that are obtained both
LV dysfunction; however, the overall LVEF is preserved.4 Inde- before and within the first minutes after di-N-methylglucamine
pendent LV involvement is rare, and the pattern of LV salt of gadopentetate (Gd-DTPA) injection.14 Gd-DTPA
595
accumulates in the myocardium, with hyperemia and capillary

leakage during the early washout period. Because enhancement Prognosis
is visually subtle in most cases, quantitative analysis of myocardial
EGEr is usually required. Several studies have confirmed the Hypertrophic Cardiomyopathy (HCM)
diagnostic value of this sequence, although it is prone to artifacts
that decrease specificity.15 LGE allows visualization of the myo- Most individuals with HCM are asymptomatic, and the first
cardium with necrosis and fibrosis. In myocarditis, LGE shows manifestation of the condition may be SCD, which is generally
two patterns of myocardial damage: (1) an intramural, rim-like related to ventricular arrhythmia triggered by factors such as
pattern in the septal wall; and (2) a patchy subepicardial distribu- ischemia, outflow obstruction, or atrial fibrillation.18 Current
tion in the free LV lateral wall.17 However, LGE cannot differ- guidelines recommend ICD implantation for patients with HCM
entiate between acute and chronic inflammation. The combination who have one or more major risk factors for SCD.1 Recent
of these different MRI techniques provides high sensitivity and advances in MRI technologies have demonstrated that myocar-
specificity.14,15 dial fibrosis detected by LGE is a relatively early manifestation
of HCM, and that the presence of fibrosis may itself represent a
high risk for SCD in patients without conventional risk factors
(Figure 61-1, C).19 In patients with HCM, myocardial fibrosis
Box 61-1 Magnetic Resonance Imaging Criteria for Global or measured by LGE is associated with increased frequency of ven-
Regional Dysfunction and Structural Alterations tricular premature contractions and nonsustained VT on Holter
monitoring.20 LGE is also an independent predictor of adverse
Major outcomes, including cardiovascular death, unplanned cardiovas-
• Regional RV akinesia or dyskinesia or dyssynchronous RV cular admission, sustained VT or ventricular fibrillation (VF), or
contraction and appropriate ICD therapy.21 These findings suggest that routine
• One of the following: use of cardiac MRI in patients with HCM may improve risk
• Ratio of RV end-diastolic volume to BSA ≥110 mL/m2 stratification.
(male) or ≥100 mL/m2 (female) or
• RV ejection fraction ≤40%
Ischemic Cardiomyopathy (ICM)
Minor
• Regional RV akinesia or dyskinesia or dyssynchronous RV Ventricular arrhythmia associated with myocardial scar in patients
contraction and with ICM is a life-threatening arrhythmia that remains one of
• One of the following: the most challenging clinical problems. Cardiac MRI is an
• Ratio of RV end-diastolic volume to BSA ≥100 to extremely powerful tool for predicting ventricular arrhythmia
<110 mL/m2 (male) or ≥90 to <100 mL/m2 (female) or and SCD in patients with ICM (Figure 61-1, D). Schmidt et al22
• RV ejection fraction >40% to ≤45% studied the relationship between the arrhythmic substrate and the
heterogeneous zone (HZ) in the infarct periphery, with interme-
diate signal intensity between the scar and the normal myocar-
2010 Revised ARVC/D Task Force Criteria.2
dium, and found that the HZ volume identified by LGE is an
Definite diagnosis: 2 major or 1 major and 2 minor criteria or 4 minor from
important clinical predictor of inducible VT at the time of an
different categories; borderline: 1 major and 1 minor or 3 minor criteria from
electrophysiological study (EPS). Other groups also found that
different categories; possible: 1 major or 2 minor criteria from different
the HZ volume is an independent predictor of appropriate ICD
categories. BSA, Body surface area; RV, right ventricular.
therapy23 and major cardiovascular events.24 These findings
A B C
D E
Figure 61-1. Late Gadolinium Enhancement (LGE) of the Myocardium (white arrows) A, Arrhythmogenic right ventricular dysplasia/cardiomyopathy
(ARVD/C). B, Cardiac sarcoidosis. C, Hypertrophic cardiomyopathy (HCM). D, Ischemic cardiomyopathy (ICM). E, Nonischemic cardiomyopathy (NICM).
suggest that the HZ identified by LGE represents the HZ at the

Box 61-2 Guidelines for the Diagnosis of Cardiac Sarcoidosis
61
tissue level, defined as a highly complex mixture of scar and
normal-appearing tissue in transition between the scar and the
Histologic Diagnosis Case
preserved normal tissues, which creates the regions of slow con-
• Positive endomyocardial biopsy demonstrating noncaseating duction and serves as the substrate for reentry circuits. Our labo-
epithelioid cell granulomas, and one of the following: ratory also demonstrated that cardiac MRI can identify the
• Positive biopsy demonstrating noncaseating epithelioid distinct differences in scar structure between the heart with and
cell granulomas in an extracardiac organ or the heart without VT.25 The heart without inducible VT is char-
• Clinical diagnosis of extracardiac sarcoidosis or acterized by smooth scar edges with very few islands of viable
• Two findings in the list below: tissue within the scar, whereas the heart with inducible VT is
characterized by complex, irregular edges with substantial viable
Laboratory Findings Suggestive of Systemic Involvement tissue intermingled with the scar. Therefore, cardiac MRI with
• Bilateral hilar lymphadenopathy evaluation of the structure and the extent of the scar provides an
• High serum ACE important additional prognostic value that identifies the patient
• Negative PPD test population at risk for ventricular arrhythmia and SCD.
• Increased 67 Ga uptake in any organ(s)
• High CD4/CD8 ratio in BAL
• High serum or urine calcium Nonischemic Cardiomyopathy (NICM)
Clinical Diagnosis Case The pathogenesis of NICM with ventricular dilatation and
• Negative endomyocardial biopsy and reduced cardiac function in the absence of flow-limiting coronary
• One finding in the list above, and one of the following: artery disease (CAD) can be genetic, inflammatory, toxic, or viral.
• Two major criteria or However, in the vast majority of cases, the origin is unclear. Our
• One major criterion and two minor criteria laboratory demonstrated for the first time that the midwall
NICM enhancement pattern is associated with inducible VT26
Major Criteria (Figure 61-1, E). The presence of the myocardial scar is observed
• Advanced atrioventricular block in up to 42% of patients with NICM.27 In patients with newly
• Basal thinning of the interventricular septum diagnosed NICM, the extent of myocardial scar as quantified by
• Abnormal 67Ga uptake in the heart LGE is independently associated with lack of response to medical
• Depressed LVEF (<50%) therapy, as well as with the combined end point of mortality and
hospitalizations.28 The transmural extent of the myocardial scar
Minor Criteria
predicts inducibility of VT at the time of EPS, and the presence
• Electrocardiography (ECG): ventricular arrhythmia (ventricular of myocardial scar predicts the composite end point of hospital-
tachycardia, multifocal or frequent PVCs), RBBB, axis ization for heart failure, appropriate ICD firing, and cardiac
deviation, or abnormal Q wave death.27
• Echocardiography: Regional wall motion abnormality or
structural abnormality (ventricular aneurysm, hypertrophy)
• Myocardial perfusion imaging (MPI): perfusion defect in 201Tl,
99m
Tc-MIBI, or 99mTc-tetrofosmin MPI
Procedure Guidance
• Cardiac MRI: late gadolinium enhancement (LGE) of the
myocardium Atrial Fibrillation (AF)
• Endomyocardial biopsy: interstitial fibrosis or monocyte
In patients undergoing catheter ablation of AF (pulmonary vein
infiltration above the moderate grade isolation [PVI]), preprocedural CT and MRI of the heart not
Notes only provides an anatomical shell of the left atrium (LA) to merge
1. Coronary angiogram should be used when CAD needs to into the electroanatomical mapping system to guide the ablation
be ruled out. procedure, but also reveals important anatomical information
2. Follow-up ECG and echocardiography on a regular basis that contributes to preprocedural planning. For example, prepro-
cedural imaging identifies the number and relative locations of
are recommended for patients with extracardiac
pulmonary vein branches, and detects pulmonary vein stenosis in
sarcoidosis, because cardiac involvement of sarcoidosis patients who have undergone PVI in the past. Cardiac CT identi-
may not be apparent at the time of diagnosis of fies the relative location of the esophagus with reference to the
extracardiac sarcoidosis. LA to minimize the risk of atrioesophageal fistula as a complica-
3. Abnormal uptake in the heart with 18F-FDG tion of ablation (Figure 61-2). Cardiac CT may also be used to
(fluorodeoxyglucose)-PET has a high diagnostic value. assess for the presence of thrombus in the LA appendage before
4. Some cases of cardiac sarcoidosis present with complete the procedure, but a head-to-head comparison with transesopha-
heart block with no minor criteria. geal echocardiography revealed that cardiac CT showed high
5. Some develop cardiac sarcoidosis immediately after interobserver variability and has only modest diagnostic accuracy
pericarditis, involving ST elevation and pericardial effusion. for the detection of LA thrombus.29
Cardiac MRI with LGE serves as a noninvasive means of
6. Diagnostic yield of endomyocardial biopsy for
assessing LA myocardial tissue in patients with AF, and may
noncaseating epithelioid cell granulomas is not high. provide insight into predicting response to AF ablation and
disease progression. The extent of the LA scar in pre-ablation
ACE, Angiotensin-converting enzyme; PPD, purified protein derivative; Ga, MRI predicts clinical outcome after AF ablation, where the
gallium; Tl, thallium; Tc, technetium; CAD, coronary artery disease; LVEF, left overall size of the scar before AF ablation is a predictor of AF
ventricular ejection fraction; PVC, premature ventricular complex; RBBB, right recurrence after successful AF ablation.30 For example, those with
bundle branch block. minimal scar (<5% relative to the LA wall volume) had a recur-
2006 Updated Diagnostic Criteria by the Ministry of Health, Labor, and Welfare rence of rate of up to 14%, whereas those with extensive scar
of Japan.9
identifies epicardial fat that can mimic scar tissue during epicar-
dial voltage mapping.36
The critical advantage of cardiac MRI in VT ablation is its
ability to identify the scar regions with abnormal electrophysiol-
ogy. Another important advantage of cardiac MRI is that it can
allow visualization of ablation lesions for prediction of clinical
outcomes37 (Figure 61-4). Moreover, cardiac MRI can quantify
the tissue temperature to assess the efficacy of ablation (Figure
61-5).38 These advantages motivated the integration of MRI scar
information into the standard electroanatomical mapping system
to facilitate target localization for VT ablation procedures,39 thus
accelerating the evolution of image-based electrophysiological
intervention (Figure 61-6). MRI scar integration has been applied
to VT ablation procedures associated with both ICM40 and
NICM.41 The value of MRI scar integration for guiding VT
ablation is particularly high in patients with NICM, for whom
A no information as to the presence or location of the scar is
obtained before ablation when a conventional ablation approach
is used.41 MRI scar integration has provided quantitative informa-
tion as to the optimal VT ablation target. A significant correla-
tion has been noted between local bipolar voltage in the standard
electroanatomical mapping system and scar transmurality as
quantified by LGE.42 Critical sites for maintenance of VT and
central common pathways for reentry are associated with >25%
and >75% scar transmurality, respectively.43 The HZ derived
from LGE, which appears to serve as a substrate for slow conduc-
tion, is found in the central common pathway within reentry VT
circuits44 (Figure 61-7). Furthermore, successful VT ablation
sites are localized within the HZ, and incomplete ablation of HZ
is associated with VT recurrence.25 These findings suggest that
MRI can identify the anatomical substrates that are critical for
maintenance of VT, thus serving as an optimal target for ablation.
B In addition, the image-electrophysiology correlation may allow
Figure 61-2. Three-Dimensional Volume-Rendered Computed prediction of the VT circuit and/or optimal ablation using MRI-
Tomography (CT) Image A, Pre-ablation image, showing a posterior view of based computational simulation. Using geometry-based simula-
the left atrium (cyan) and the left atrial appendage (light gray). Distal pulmonary tion, Ciaccio et al demonstrated that scar geometry based on
vein branches are removed. The image shows four small right pulmonary veins MRI predicts the reentry VT circuit in a dog model of myocar-
entering into the left atrium—a common anatomical variation. The manually seg- dial infarction (MI).45 Ng et al also showed the feasibility of
mented esophagus (purple) is also shown to demonstrate the relative anatomical MRI-based mathematical modeling to predict VT circuits in a
relationship with the left atrium. B, Post-ablation image, showing ablation points porcine model of MI.46 Furthermore, using sophisticated finite
(red dots). element whole heart electrophysiology modeling, our laboratory
demonstrated that computational simulation accurately predicts
successful ablation targets in human patients.47 Last real-time
(>35%) had a recurrence rate of up to 75%.30,31 LGE also allows MRI-guided electrophysiological intervention is another emerg-
noninvasive identification of myocardial scar induced by RF abla- ing application that can be used to take advantage of MRI tech-
tion following AF ablation.32 RF-induced scar appears to form by nology to guide VT ablation.48
3 months post ablation. At 24 h post ablation, LGE appears A potential limitation of cardiac MRI in guiding VT ablation
consistent with a transient inflammatory response rather than is the fact that most patients who are referred for VT ablation
with stable LA scar formation.33 AF recurrence during the first have an ICD, which is traditionally considered as a contraindica-
year is associated with a lesser degree of pulmonary vein (PV) tion for MRI. Our laboratory demonstrated the safety of MRI in
and left atrial scarring on LGE.34 Circumferential PV antral scar- patients with cardiac implantable electrical devices with appro-
ring predicts ablation success in mild LA fibrosis, and posterior priate electrophysiological monitoring during MRI.49 However,
wall and septal scarring is needed for moderate fibrosis. This may susceptibility artifacts from the device, particularly ICDs
facilitate selection of the proper candidate and strategy in AF implanted in the left infraclavicular region, could have a signifi-
ablation.30 A visual and quantitative correspondence was noted cant impact on LGE assessment of the anterior and apical LV.50
between CARTO ablation sites and the LGE scar, but for 20% The development of artifact suppression methods is essential for
of CARTO ablation sites, corresponding LGE was visible.35 improving the utility of MRI-guided VT ablation.
Ventricular Tachycardia (VT) Mechanical Dyssynchrony and Cardiac

Resynchronization Therapy (CRT)
In patients undergoing catheter ablation of scar-related VT, pre-
procedural cardiac CT and MRI provide important anatomical Cardiac CT and MRI provide essential information for candidate
information. For example, cardiac CT identifies the locations of selection and for preprocedural planning for CRT. For example,
myocardial thinning, aneurysm, and calcification, which corre- cardiac CT can define the coronary sinus anatomy, which is
spond with the myocardial scar. In addition, the multiplanar extremely variable and often lacks an appropriate branch for
reformat feature allows preprocedural identification of optimal optimal lead placement.51 On the other hand, cardiac MRI can
epicardial access location (Figure 61-3). Moreover, cardiac CT quantify myocardial scar distribution with LGE, particularly in
61
A B C
D E
Figure 61-3. Computed Tomography (CT)-Guided Preprocedural Identification of Optimal Epicardial Access Location for Ventricular Tachy-
cardia (VT) Ablation Images from a 28-year-old woman with arrhythmogenic right ventricular dysplasia/cardiomyopathy (ARVD/C) who was referred for epicardial VT
ablation. A, Coronal (anteroposterior), (B) sagittal (lateral), (C) axial, and (D) volume-rendered images of the preprocedural CT are shown. The yellow arrow indicates optimal
epicardial access from the subxiphoid approach. E shows the intraprocedural image of epicardial electroanatomical mapping.
A B C D
T1-weighted MRI T2-weighted MRI Gross pathology Histology
(Masson Trichrome)
Figure 61-4. Magnetic Resonance Imaging (MRI) Visualization of Ablation Lesion Necrotic cavities within high-energy radiofrequency (RF) ablation lesions
and corresponding pathologic specimens. Typical appearance of intralesional necrotic cavities as areas of increased signal intensity on T2-weighted magnetic resonance
images (A) and decreased signal intensity on T1-weighted magnetic resonance images (B), corresponding gross pathology (C), and histology with Masson trichrome
stain (D).
(Figure reproduced with permission from Dickfeld T, Kato R, Zviman M, et al: Characterization of acute and subacute radiofrequency ablation lesions with nonenhanced magnetic
resonance imaging. Heart Rhythm 4:208–214, 2007.)
patients with ICM. The transmurality of the myocardial scar is Cardiac MRI played a critical role in improving our under-
an important predictor of lack of response to CRT. The presence standing of the pathophysiology of mechanical dyssynchrony53
of a posterolateral scar is an independent predictor of major (Figure 61-8). A variety of MRI-derived indices have been pro-
cardiovascular events, and multi-site LV pacing may be beneficial posed to quantify mechanical dyssynchrony and to predict the
in these patients.52 response to CRT. Tissue synchronization index (TSI), which
3 Watts 10 Watts 24 Watts LGE
A B C
E
D
60 seconds cooling
F
Pathology
Figure 61-5. Magnetic Resonance (MR) Thermography A-C demonstrate an expanding thermal lesion as radiofrequency (RF) power increased, which is no longer
visible by 60 seconds after RF power is stopped (D). The heating location (C) corresponds to the lesion location on late gadolinium enhancement (LGE) (E) and pathology
(F). The transmural lesion extent by thermography was within 20% of that obtained by pathology and LGE. The red lesion border indicates 50° C, chosen to correspond to
irreversible tissue damage.
(Figure modified with permission from Kolandaivelu A, Zviman MM, Castro V, et al: Noninvasive assessment of tissue heating during cardiac radiofrequency ablation using MRI
thermography. Circ Arrhythm Electrophysiol 3:521–529, 2010.)
1 2 3 16.06mV
Bipolar
2-rv voltage > 103points
Epi Endo
4 5 High-intensity 6
artifact 0.07mV
RV SAscar
LV 1.63cm
A B Scar
LGE ICD-lead LGE RV
Figure 61-6. Scar Integration A, Short-axis late gadolinium enhancement (LGE) images of a patient with an implantable cardioverter-defibrillator (ICD). ICD results in
high-intensity artifact, affecting the anterior wall. ICD lead seen as low-intensity signal in inferior right ventricle (RV). Transmural LGE seen in inferior wall of left ventricle
(LV). B, Registration of LGE images with the CARTO electroanatomical mapping system. LV voltage map is aligned with magnetic resonance imaging (MRI)-extracted
endocardial shell (turquoise). Rotational errors are corrected by aligning the RV voltage map with the reconstructed MRI RV slice (red). After integration of the MRI-extracted
epicardial shell (blue mesh), an embedded MRI-derived transmural scar (brown) is visible within the myocardial wall. Endo, Endocardial; Epi, epicardial.
(Figure reproduced with permission from Dickfeld T, Tian J, Ahmad G, et al: MRI-guided ventricular tachycardia ablation: Integration of late gadolinium-enhanced 3D scar in
patients with implantable cardioverter-defibrillators. Circ Arrhythm Electrophysiol 4:172–184, 2011.)
represents the deviation of radial wall motion in LV short-axis
61
cine MRI from an arbitrarily fitted sine function, is a powerful
predictor of major cardiovascular events after CRT, particularly
when combined with the presence of posterolateral scar.54 The
circumferential uniformity ratio estimate (CURE), which esti-
mates circumferential mechanical dyssynchrony on the basis of
RV tagged cine MRI, predicts improved function class with 90%
accuracy, and accuracy was improved to 95% when LGE data
Lateral LV
were added (% total scar <15%).55 The regional vector of cir-
cumferential strain variance (RVV), based on tagged cine MRI,
predicts response to CRT independently of LGE and the origin
CL = 242 ms of heart failure (ICM vs. NICM).56 Cardiac CT is also used to
assess mechanical dyssynchrony on the basis of changes in wall
thickness over time, but its ability to predict CRT response
remains unclear.57
Early Late
Summary
Cardiac CT and MRI play a unique role in cardiac electrophysi-
ology. Cardiac MRI is the imaging modality of choice for evaluat-
ing patients suspected of having ARVD/C, cardiac sarcoidosis,
Figure 61-7. Epicardial Reentry Ventricular Tachycardia (VT) Circuit
and myocarditis. Cardiac MRI is also valuable for risk stratifica-
Registered With Magnetic Resonance Imaging (MRI)-Derived Scar
The central common pathway (circumscribed by a broken red line) was located within
tion of patients with HCM, ICM, and NICM. Cardiac CT and
the heterogeneous zone (HZ) in the anterior wall (see the late gadolinium enhance- MRI play an important role in guiding ablation procedures for
ment [LGE] image on the right). complex arrhythmias, including AF and VT, and in candidate
selection for cardiac resynchronization therapy. Technical
(Figure reproduced with permission from Ashikaga H, Sasano T, Dong J, et al: Mag- advances in cardiac CT and MRI will continue to expand their
netic resonance-based anatomical analysis of scar-related ventricular tachycardia: role in cardiac electrophysiology.
Implications for catheter ablation. Circ Res 101:939–947, 2007.)
10
Septal Posterior Lateral Anterior
5
Basal
0
Strain
–5
Mid
–10
Apical –15
–20
Early systole (60ms)
= Contraction = Stretch
A B Late systole (300ms)
10
5
Basal
0
Strain
–5
Mid
–10
Apical –15
–20
C D
Figure 61-8. Spatiotemporal Circumferential Strain Maps Derived from Tagged Magnetic Resonance Imaging (MRI) In the normal subject, the
progression of strain versus time (negative strain represents systole) is uniform in each of the 24 segments in each slice (A), and synchronous negative strain is seen for
each segment along the circumference of the left ventricle (B). In the subject with mechanical dyssynchrony and cardiomyopathy, strain versus time maps show variable
timing of contraction (blue arrows = negative strain) and stretch (orange arrows = positive strain) in septal versus lateral segments (C). In this subject, some segments have
positive strain (stretch) and others have negative strain (contraction) during systole (D).
sudden cardiac death in hypertrophic cardiomy- gadolinium enhancement cardiovascular magnetic

References opathy. JAMA 298:405–412, 2007. resonance. Heart Rhythm 7:489–496, 2010.
19. Ismail TF, Prasad SK, Pennell DJ: Prognostic 36. Desjardins B, Morady F, Bogun F: Effect of epicar-
1. Epstein AE, DiMarco JP, Ellenbogen KA, et al: importance of late gadolinium enhancement cardial fat on electroanatomical mapping and epicar-
ACC/AHA/HRS 2008 guidelines for device-based diovascular magnetic resonance in cardiomyopa- dial catheter ablation. J Am Coll Cardiol
therapy of cardiac rhythm abnormalities: A report thy. Heart 98:438–442, 2012. 56:1320–1327, 2010.
of the American College of Cardiology/American 20. Adabag AS, Maron BJ, Appelbaum E, et al: 37. Ranjan R, Kato R, Zviman MM, et al: Gaps in the
Heart Association Task Force on Practice Guide- Occurrence and frequency of arrhythmias in ablation line as a potential cause of recovery from
lines (writing committee to revise the ACC/AHA/ hypertrophic cardiomyopathy in relation to electrical isolation and their visualization using
NASPE 2002 guideline update for implantation of delayed enhancement on cardiovascular magnetic MRI. Circ Arrhythm Electrophysiol 4:279–286,
cardiac pacemakers and antiarrhythmia devices): resonance. J Am Coll Cardiol 51:1369–1374, 2011.
Developed in collaboration with the American 2008. 38. Kolandaivelu A, Zviman MM, Castro V, et al: Non-
Association for Thoracic Surgery and Society of 21. O’Hanlon R, Grasso A, Roughton M, et al: Prog- invasive assessment of tissue heating during cardiac
Thoracic Surgeons. Circulation 117:e350–e408, nostic significance of myocardial fibrosis in hyper- radiofrequency ablation using MRI thermography.
2008. trophic cardiomyopathy. J Am Coll Cardiol Circ Arrhythm Electrophysiol 3:521–529, 2010.
2. Marcus FI, McKenna WJ, Sherrill D, et al: Diag- 56:867–874, 2010. 39. Tian J, Ahmad G, Mesubi O, et al: Three-
nosis of arrhythmogenic right ventricular 22. Schmidt A, Azevedo CF, Cheng A, et al: Infarct dimensional delayed-enhanced cardiac MRI recon-
cardiomyopathy/dysplasia: Proposed modification tissue heterogeneity by magnetic resonance structions to guide ventricular tachycardia ablations
of the task force criteria. Circulation 121:1533– imaging identifies enhanced cardiac arrhythmia and assess ablation lesions. Circ Arrhythm Electro-
1541, 2010. susceptibility in patients with left ventricular dys- physiol 5:e31–e35, 2012.
3. Tandri H, Macedo R, Calkins H, et al: Role of function. Circulation 115:2006–2014, 2007. 40. Desjardins B, Crawford T, Good E, et al: Infarct
magnetic resonance imaging in arrhythmogenic 23. Roes SD, Borleffs CJ, van der Geest RJ, et al: architecture and characteristics on delayed
right ventricular dysplasia: Insights from the North Infarct tissue heterogeneity assessed with contrast- enhanced magnetic resonance imaging and electro-
American Arrhythmogenic Right Ventricular Dys- enhanced MRI predicts spontaneous ventricular anatomic mapping in patients with postinfarction
plasia (ARVD/C) study. Am Heart J 155:147–153, arrhythmia in patients with ischemic cardiomyopa- ventricular arrhythmia. Heart Rhythm 6:644–651,
2008. thy and implantable cardioverter-defibrillator. Circ 2009.
4. Jain A, Shehata ML, Stuber M, et al: Prevalence of Cardiovasc Imaging 2:183–190, 2009. 41. Bogun FM, Desjardins B, Good E, et al: Delayed-
left ventricular regional dysfunction in arrhythmo- 24. Heidary S, Patel H, Chung J, et al: Quantitative enhanced magnetic resonance imaging in nonisch-
genic right ventricular dysplasia: A tagged MRI tissue characterization of infarct core and border emic cardiomyopathy: Utility for identifying the
study. Circ Cardiovasc Imaging 3:290–297, 2010. zone in patients with ischemic cardiomyopathy by ventricular arrhythmia substrate. J Am Coll Cardiol
5. Dalal D, Tandri H, Judge DP, et al: Morphologic magnetic resonance is associated with future car- 53:1138–1145, 2009.
variants of familial arrhythmogenic right ventricu- diovascular events. J Am Coll Cardiol 55:2762– 42. Sasaki T, Miller CF, Hansford R, et al: Non-
lar dysplasia/cardiomyopathy: A genetics-magnetic 2768, 2010. invasive magnetic resonance based three-
resonance imaging correlation study. J Am Coll 25. Estner HL, Zviman MM, Herzka D, et al: The dimensional voltage maps for electrophysiology
Cardiol 53:1289–1299, 2009. critical isthmus sites of ischemic ventricular tachy- procedure guidance. Circulation 124:A9642, 2011.
6. Tandri H, Asimaki A, Abraham T, et al: Prolonged cardia are in zones of tissue heterogeneity, visual- 43. Sasaki T, Hansford R, Zviman MM, et al: Critical
rv endocardial activation duration: A novel marker ized by magnetic resonance imaging. Heart sites of post-infarct ventricular tachycardia on late
of arrhythmogenic right ventricular dysplasia/ Rhythm 8:1942–1949, 2011. gadolinium enhanced cardiac magnetic resonance
cardiomyopathy. Heart Rhythm 6:769–775, 2009. 26. Nazarian S, Bluemke DA, Halperin HR: Applica- in patients with ischemic cardiomyopathy. Heart
7. Hiraga H, Yuwa K, Hiroe M: Guideline for the tions of cardiac magnetic resonance in electro- Rhythm 9:S80, 2012 (AB36-AB03).
diagnosis of cardiac sarcoidosis: Study report on physiology. Circ Arrhythm Electrophysiol 2:63–71, 44. Ashikaga H, Sasano T, Dong J, et al: Magnetic
diffuse pulmonary disease [in Japanese]. Jpn Min- 2009. resonance-based anatomical analysis of scar-related
istry Health Welfare 23–24, 1993. 27. Wu KC, Weiss RG, Thiemann DR, et al: Late ventricular tachycardia: Implications for catheter
8. Roberts WC, Vowels TJ, Ko JM, et al: Cardiac gadolinium enhancement by cardiovascular mag- ablation. Circ Res 101:939–947, 2007.
transplantation for cardiac sarcoidosis with initial netic resonance heralds an adverse prognosis in 45. Ciaccio EJ, Ashikaga H, Kaba RA, et al: Model of
diagnosis by examination of the left ventricular nonischemic cardiomyopathy. J Am Coll Cardiol reentrant ventricular tachycardia based on infarct
apical “core” excised for insertion of a left ventricu- 51:2414–2421, 2008. border zone geometry predicts reentrant circuit
lar assist device for severe chronic heart failure. Am 28. Leong DP, Chakrabarty A, Shipp N, et al: features as determined by activation mapping.
J Cardiol 103:110–114, 2009. Effects of myocardial fibrosis and ventricular dys- Heart Rhythm 4:1034–1045, 2007.
9. Diagnostic standard and guidelines for sarcoidosis. synchrony on response to therapy in new- 46. Ng J, Jacobson JT, Ng JK, et al: Virtual electro-
Jpn J Sarcoid Granulom Disord 27:89–192, 2007. presentation idiopathic dilated cardiomyopathy: physiological study in a 3-dimensional cardiac
10. Tahara N, Tahara A, Nitta Y, et al: Heterogeneous Insights from cardiovascular magnetic resonance magnetic resonance imaging model of porcine
myocardial FDG uptake and the disease activity in and echocardiography. Eur Heart J 33:640–648, myocardial infarction. J Am Coll Cardiol 60:423–
cardiac sarcoidosis. J Am Coll Cardiol 3:1219– 2012. 430, 2012.
1228, 2010. 29. Gottlieb I, Pinheiro A, Brinker JA, et al: Diagnostic 47. Ashikaga H, Arevalo H, Vadakkumpadan F, et al:
11. Youssef G, Leung E, Mylonas I, et al: The use of accuracy of arterial phase 64-slice multidetector Feasibility of image-based simulation to estimate
18F-FDG PET in the diagnosis of cardiac sarcoid- CT angiography for left atrial appendage thrombus ablation target in human ventricular arrhythmia.
osis: A systematic review and metaanalysis includ- in patients undergoing atrial fibrillation ablation. Heart Rhythm 2013 (in press).
ing the Ontario experience. J Nucl Med J Cardiovasc Electrophysiol 19:247–251, 2008. 48. Nazarian S, Kolandaivelu A, Zviman MM, et al:
53:241–248, 2012. 30. Akoum N, Daccarett M, McGann C, et al: Atrial Feasibility of real-time magnetic resonance
12. Patel MR, Cawley PJ, Heitner JF, et al: Detection fibrosis helps select the appropriate patient and imaging for catheter guidance in electrophysiology
of myocardial damage in patients with sarcoidosis. strategy in catheter ablation of atrial fibrillation: A studies. Circulation 118:223–229, 2008.
Circulation 120:1969–1977, 2009. DE-MRI guided approach. J Cardiovasc Electro- 49. Nazarian S, Hansford R, Roguin A, et al: A pro-
13. Kindermann I, Barth C, Mahfoud F, et al: Update physiol 22:16–22, 2011. spective evaluation of a protocol for magnetic reso-
on myocarditis. J Am Coll Cardiol 59:779–792, 31. Oakes RS, Badger TJ, Kholmovski EG, et al: nance imaging of patients with implanted cardiac
2012. Detection and quantification of left atrial structural devices. Ann Intern Med 155:415–424, 2011.
14. Friedrich MG, Sechtem U, Schulz-Menger J, et al: remodeling with delayed-enhancement magnetic 50. Sasaki T, Hansford R, Zviman MM, et al: Quanti-
Cardiovascular magnetic resonance in myocarditis: resonance imaging in patients with atrial fibrilla- tative assessment of artifacts on cardiac magnetic
A JACC white paper. J Am Coll Cardiol 53:1475– tion. Circulation 119:1758–1767, 2009. resonance imaging of patients with pacemakers and
1487, 2009. 32. Peters DC, Wylie JV, Hauser TH, et al: Detection implantable cardioverter-defibrillators. Circ Car-
15. Abdel-Aty H, Boye P, Zagrosek A, et al: Diagnostic of pulmonary vein and left atrial scar after catheter diovasc Imaging 4:662–670, 2011.
performance of cardiovascular magnetic resonance ablation with three-dimensional navigator-gated 51. Joshi SB, Blum AR, Mansour M, Abbara S: CT
in patients with suspected acute myocarditis: Com- delayed enhancement MR imaging: Initial experi- applications in electrophysiology. Cardiol Clin
parison of different approaches. J Am Coll Cardiol ence. Radiology 243:690–695, 2007. 27:619–631, 2009.
45:1815–1822, 2005. 33. Badger TJ, Oakes RS, Daccarett M, et al: Temporal 52. Ginks MR, Duckett SG, Kapetanakis S, et al:
16. Aletras AH, Kellman P, Derbyshire JA, Arai AE: left atrial lesion formation after ablation of atrial Multi-site left ventricular pacing as a potential
Acute TSE-SSFP: A hybrid method for fibrillation. Heart Rhythm 6:161–168, 2009. treatment for patients with postero-lateral scar:
T2-weighted imaging of edema in the heart. Magn 34. Peters DC, Wylie JV, Hauser TH, et al: Recur- Insights from cardiac magnetic resonance imaging
Reson Med 59:229–235, 2008. rence of atrial fibrillation correlates with the extent and invasive haemodynamic assessment. Europace
17. Mahrholdt H, Wagner A, Deluigi CC, et al: Pre- of post-procedural late gadolinium enhancement: 214:373–379, 2012.
sentation, patterns of myocardial damage, and A pilot study. J Am Coll Cardiol 2:308–316, 53. Bilchick KC, Helm RH, Kass DA: Physiology of
clinical course of viral myocarditis. Circulation 2009. biventricular pacing. Curr Cardiol Rep 9:358–365,
114:1581–1590, 2006. 35. Taclas JE, Nezafat R, Wylie JV, et al: Relationship 2007.
18. Maron BJ, Spirito P, Shen WK, et al: Implantable between intended sites of RF ablation and post- 54. Leyva F, Foley PW, Stegemann B, et al: Develop-
cardioverter-defibrillators and prevention of procedural scar in AF patients, using late ment and validation of a clinical index to predict
survival after cardiac resynchronisation therapy. 56. Petryka J, Misko J, Przybylski A, et al: Magnetic 57. Truong QA, Singh JP, Cannon CP, et al: Quantita-
Heart 95:1619–1625, 2009. resonance imaging assessment of intraventricular tive analysis of intraventricular dyssynchrony using
55. Bilchick KC, Dimaano V, Wu KC, et al: Cardiac
magnetic resonance assessment of dyssynchrony
and myocardial scar predicts function class
dyssynchrony and delayed enhancement as predic-
tors of response to cardiac resynchronization
therapy in patients with heart failure of ischaemic
wall thickness by multidetector computed tomog-
raphy. J Am Coll Cardiol 1:772–781, 2008. 61
improvement following cardiac resynchronization and non-ischaemic etiologies. Eur J Radiol
therapy. J Am Coll Cardiol 1:561–568, 2008. 81:2639–2647, 2011.
Intracardiac Echocardiography
for Electrophysiology 62
Mathew D. Hutchinson and David J. Callans
transducer is mounted on a 9 French, nonsteerable catheter and

CHAPTER OUTLINE
emits an imaging beam at a 15° forward angle, perpendicular to
Rationale 605 the long axis of the catheter. The transducer is typically delivered
through a curved or steerable introducer sheath into the chamber
ICE Platforms 605
of interest. The transducer rotates at 1800 rpm and has a fixed,
Basic ICE Imaging Planes 605 9-MHz frequency; it produces a 360° imaging plane perpendicu-
lar to the axis of the catheter. Limited far-field resolution neces-
Sinus Node Modification 605
sitates imaging proximal to the structure of interest.
Atrial Fibrillation Ablation 606 The phased array ICE catheter contains a 64-element trans-
ducer with variable frequency ranging from 5 to 10 MHz, thereby
Ventricular Tachycardia Ablation 607
providing greater flexibility to image remote structures (up to
Visualization of Ablation Lesions 609 15 cm). The phased array transducer is capable of full spectral
and color Doppler measurements, greatly enhancing achievable
Visualization of Arrhythmia Substrate 610
physiological data. The transducer is mounted on a bidirectional
Detecting and Preventing Complications 611 or multidirectional 8 to 10 French catheter. The most commonly
used transducer is the AcuNav™ (Siemens Medical, Mountain
Summary 611
View, California) system, which can be deflected in four direc-
tions (anterior, posterior, right, and left) in addition to providing
360° axial rotation. Unless otherwise indicated, the remainder of
this chapter describes imaging with the phased array ICE system.
Rationale
The increasing complexity of contemporary catheter ablation
procedures has created an important niche for real-time imaging Basic ICE Imaging Planes
modalities in the EP laboratory. Traditional modalities such as
fluoroscopy provide limited anatomical detail and incur signifi- Although imaging with ICE requires a modest learning curve, it
cant potential patient and provider exposure risk. The trend is a logical extension for operators with basic catheter manipula-
toward anatomically based ablation targets and procedural end tion and echocardiography skills. Most new operators find the
points subsumes an intimate knowledge of the specific anatomical images somewhat disorienting when taken out of context. Only
variations of the individual patient. Integrating pre-acquired two- after an individual echo “view” is integrated mentally with the
or three-dimensional images (e.g., tomographic reconstructions) operator’s inherent knowledge of intracardiac anatomy does
can be quite useful for anatomical characterization, but can inher- obtaining and interpreting ICE images become intuitive. It is of
ently lack real-time feedback for temporal changes that occur paramount importance to realize that the infinite potential ori-
during the procedure. entations of the transducer within the cardiac chambers produce
Uniquely suited for EP procedures, intracardiac echocardiog- infinite possible imaging planes. To avoid confusion, it is useful
raphy (ICE) is capable of providing substantial real-time data to learn a few fiducial imaging planes from which an ICE survey
with relative ease. Indeed ICE provides both physiological is easily generated. The cardinal imaging plane or “home view”
and anatomical data that facilitate the following important is obtained by placing the ICE catheter in a neutral, mid–right
functions: atrial position and imaging through the tricuspid valve. Other
• Characterization of anatomy and variants relevant structures are easily viewed with gentle clockwise (CW)
• Positioning of intracardiac catheters rotation of the imaging catheter along its axis.
• Confirmation of catheter contact
• Assessment of ablation lesion creation
• Visualization of arrhythmia substrate
• Detection and prevention of complications Sinus Node Modification
This chapter provides an introduction to and comparison of
currently available ICE platforms. Core concepts regarding cath- Among the earliest reported clinical applications for ICE was
eter manipulation and image optimization are reviewed. Last, the guiding sinus node modification, during which the superior
use of ICE in specific EP procedures is reviewed. lateral crista terminalis (CT) is targeted. Localizing catheters
proximate to the CT with the use of fluoroscopic guidance alone
is often inaccurate, with mean distances of >1 cm in more than
50% of cases.1 With ICE, the CT can be precisely identified,
ICE Platforms thereby avoiding delivery of ineffective lesions. ICE can also be
used to evaluate the diameter of the junction of the superior vena
Two types of ICE transducers are commercially available: cava and the right atrium, thereby avoiding excessive narrowing
mechanical (radial) and phased array systems. The radial ICE during ablation. Additionally, the presence of echointensity
605
A
SVC
ECS
CT
CT
A B
Figure 62-1. Images are taken (A) before and (B) after sinus node modification for inappropriate sinus tachycardia. The ICE catheter is positioned within the superior vena
cava (SVC) just above its origin. The superior portion of the crista terminalis (CT) is viewed in the near-field. After ablation, increased thickening and echointensity of the
CT are noted. The appearance of an echocardiographic clear space (ECS) adjacent to the ablation catheter (A) represents transmural extension of the lesion to the epicardium
and is correlated with acute heart rate slowing and an inferior shift in sinus rhythm activation.
extending to the epicardial surface, often coupled with the have suggested that the aforementioned alternative imaging
appearance of an adjacent echodense region (representing epicar- planes allow imaging more proximate to the LAA, and thus
dial edema), correlates strongly with the achievement of acute provide enhanced tissue characterization.
heart rate slowing during ablation2 (Figure 62-1).
Three-Dimensional Image Integration
Integration of preacquired tomographic images with electroana-
tomical (EA) mapping systems is widely used to facilitate AF
Atrial Fibrillation Ablation ablation. Accurate registration of these images can be challeng-
ing, in part because of the complex topography of the LA. ICE
Pulmonary Vein (PV) Anatomy can facilitate this process by providing real-time feedback regard-
ing registration quality by allowing visualization of the position-
The most complete integration of ICE in EP is achieved with ing of intracardiac catheters at fiducial locations (e.g., the ligament
atrial fibrillation (AF) ablation. Before the left atrium (LA) is of Marshall, the PV carina) (Figure 62-3). If misalignment of the
accessed, full characterization of the PVs is easily achieved from three-dimensional (3D) geometry is noted, reregistration is easily
a right atrial transseptal view. At 60° CW rotation from the home performed to improve spatial accuracy.
view, the left atrial appendage is visualized. Continued rotation Integration of phased array 2D ICE with EA mapping allows
allows characterization of the left PVs. The right PVs are directed the creation of 3D geometries that require neither fluoroscopy
180° posteriorly from the home view. nor point-to-point mapping (CARTO Sound™, Biosense
Visualization of the right superior PV is often challenging Webster, Inc., Diamond Bar, California). With this technology,
because of its septal location; this can be overcome by (1) distend- the position of the ICE catheter is localized relative to the
ing the interatrial septum toward the LA; (2) passing the imaging skin patches; tracing the border of the chamber of interest
catheter into the LA through a transseptal defect; or (3) deflect- creates sequential 2D contours that are interpolated and rendered
ing the transducer toward the tricuspid annulus. in 3D space. This technique produces a 3D geometry that
Anatomical variations in PV anatomy are common and can be is spatially accurate compared with point-to-point mapping,
fully characterized with ICE, thereby avoiding inadvertently and can be used as a stand-alone map to guide ablation, or as a
ignoring or damaging them during ablation. The diameter and reference with which to co-register a preacquired tomographic
orientation of the veins are recorded, and baseline pulse wave image.4
Doppler flow velocities are measured (Figure 62-2). If a circular
mapping catheter is used to guide PV isolation, determination of
the PV ostial diameter with ICE is useful in selecting an appro- Transseptal Puncture
priate size.
Although transseptal puncture can be safely performed under
Left Atrial Appendage (LAA) Visualization fluoroscopic guidance, the integration of ICE has greatly
Characterization of the LAA is routinely performed before LA enhanced both the learning curve and operator comfort with the
ablation in patients with inadequate preoperative anticoagulation procedure. Whether radial or phased array ICE is used, the posi-
and/or persistent AF. The advent of LAA occlusion devices pro- tion of the transseptal needle relative to critical adjacent struc-
vides another potential niche for ICE imaging. The LAA can be tures (e.g., aortic root) is easily visualized; this is particularly
viewed with ICE from several different imaging planes: (1) from critical with increasing use of dual systemic anticoagulation for
the right atrium across the atrial septum; (2) from the left atrium; AF ablation. The fossa ovalis is visualized with the ICE catheter
(3) from the coronary sinus; or (4) from the pulmonary artery. The in a mid-RA position and gently rotated in a CW fashion
recent Intra-Cardiac Echocardiography–guided Cardioversion to (approximately 45° to 90° from the home view).
Help Interventional Procedures Study (ICE-CHIP) study pro- ICE provides important context regarding not only the posi-
spectively compared LAA imaging with transesophageal echo tion, but also the direction of needle crossing, by allowing
(TEE) versus phased array ICE; the study found incomplete LAA visualization of far-field structures (Figure 62-4). Abnormalities
imaging with ICE in 15% of patients, as well as a lower sensitivity of the interatrial septum such as lipomatous hypertrophy or
to detect LAA thrombus compared with TEE.3 The comparative aneurysm can be fully characterized to optimize the individual
image quality in ICE-CHIP was potentially biased by the exclu- crossing strategy. Patients with prior surgical patching or percu-
sive use of a right atrial imaging plane with ICE. Other reports taneous closure devices may provide specific challenges to sheath
Intracardiac Echocardiography for Electrophysiology 607
62
LA
D
LS S
LI
A B
LA
LA
RS
RI
C D
Figure 62-2. Baseline color Doppler imaging of the left common (A), right superior (C), and right inferior (D) pulmonary veins. The ICE catheter is positioned along the
right atrial aspect of the interatrial septum. Ostial diameter and pulsed wave Doppler flow velocity (B) for each vein are measured both at baseline and after ablation.
crossing; ICE allows the operator to rapidly characterize these the position of the ablation catheter in real time on the 2D echo
cases. slice. The development of probes capable of 3D imaging may
provide even greater anatomical context, although with a con-
Assessing Catheter Positioning and Contact comitant reduction in spatial resolution.5
During AF ablation, ICE is quite useful during positioning of the
ablation and circular mapping of catheters relative to the PV
orifice. Rotating the imaging plane CW from the left PVs
or counterclockwise (CCW) from the right PVs can allow Ventricular Tachycardia Ablation
identification of the course of the esophagus along the posterior
LA wall. Left Ventricular Outflow Tract Ablation
It is axiomatic that maintaining catheter contact during abla-
tion is of paramount importance in lesion creation, and the devel- Ablation of ventricular arrhythmias (VAs) arising from the left
opment of catheters with contact force–sensing capabilities has ventricular (LV) ostium presents unique challenges, in part
further underscored this concept. In addition to poor lesion because of the complex anatomical relationships.6 Traditional
quality, inadequate contact may result in undersampling during aortic root mapping using a combination of fluoroscopy and
EA mapping, which may result in inaccurate 3D geometries and intracardiac electrogram (EGM) characteristics is limited by its
tomographic co-registrations. These registration errors may be variable position and angulation. Because most of the arrhyth-
compounded with multi-electrode mapping techniques. The mias targeted from the aortic root actually originate from the LV
adequacy of intracardiac catheter contact has relied historically ostium, it is important to sample the entire surface of the adjacent
upon a combination of tactile feedback, electrogram quality, and aortic root. Incorporating ICE into these procedures allows the
fluoroscopic guidance. Despite these clues, factors such as operator to determine not only which cusp is being sampled but
complex intracardiac topography as well as cardiac and respira- also the precise aspect of the cusp.
tory motion may prevent consistent contact. ICE is capable of The LV outflow tract (OT) may be visualized from the right
displaying the position of any intracardiac catheter relative to the atrium (30° CW rotation from the “home view”) or from the
myocardium in real time by simply adjusting the angulation of right ventricle (RV) (Figure 62-5). With imaging from the RA,
the ICE transducer. Because the imaging plane produced with the aortic root is seen in long axis. This view allows the operator
currently available ICE platforms is inherently 2D, resolving the to determine the depth of sampling within the aortic root (e.g.,
position and orientation of 3D intracardiac structures inherently sinotubular junction, cusp nadir), as well as whether the ablation
requires the acquisition and integration of multiple 2D slices. catheter is located above, below, or at the level of the valve plane.
The CARTO Sound system facilitates this process by displaying The long axis view also allows the operator to determine the
LAA LS
LA RS
LI
RI
E
A B
Ao
LA
L A
LA
LS R LAA
C
Ao LS
LI
C D
Figure 62-3. Registration of pre-acquired CT or MR 3D datasets with electroanatomical (EA) mapping systems is facilitated by ICE. Fiducial locations on the EA map (A, left)
and a 3D CT dataset (A, left) are co-localized using real-time ICE imaging. Panel C demonstrates the circular mapping catheter (L) positioned at the ostium of the left
superior (LS) PV. In panel D, the ablation catheter (A) is placed along the ridge (R) separating the LS PV and the left atrial appendage (LAA). Using this technique, the
accuracy of the 3D registration is optimized. Ao, Aorta; C, carina; E, esophagus; LI, left inferior PV; RI, right inferior PV; RS, right superior PV.
T T
LA
LA T1
T2
LS
LAA LI LA
LI
A B C
Figure 62-4. With ICE, the location of transseptal needle crossing can be optimized. In panel A, the transseptal (T) needle is directed toward the left atrial appendage
(LAA). After slight clockwise rotation of the transseptal needle, the apparatus is now directed toward the left pulmonary veins. Panel C shows tenting of a second transseptal
needle (T2); the sheath from the prior crossing (T1) is located slightly inferiorly.
vertical distance of the mapping catheter from the coronary arte- coronary sinuses are appreciated. The short axis view allows the
rial ostia. With application of gentle anterior flexion of the ICE operator to easily determine whether the mapping catheter is
probe from the “home” view, the catheter can be easily passed positioned within a single coronary sinus or between two adjacent
through the tricuspid annulus into the RV. With 180° CW rota- sinuses.
tion, a short axis view of the aortic root is obtained. In this posi- Integration of 2D ICE and EA mapping with the CARTO
tion, the ICE probe is placed at the inferior aspect of Sound system enables the rapid creation of detailed reconstruc-
the right ventricular outflow tract (RVOT); the close anatomical tions of the aortic root (see Figure 62-5). This technique maxi-
relationships between the RVOT and the right and left mizes root topography with minimal fluoroscopy use, thereby
62
RCC
NCC
LCC RVOT
LA
PV A I
PA
RCC
LCC
A
NCC
Ao
0.56cm
LCC
RCC
PV
A C
LV
B
Figure 62-5. The aortic root is viewed with ICE in short axis from the right ventricular outflow tract (A) and in long axis from the right atrium (B). The close spatial relation-
ship between the right and left ventricular outflow tracts is apparent. The short axis view is used to guide mapping within and between the aortic sinuses. The long axis
view demonstrates the vertical position of the mapping catheter. In panel B, the ablation catheter is looped into the LV and is pulled back against the aortic valve plane.
Three-dimensional reconstruction of the aortic sinuses (C) using the CARTO Sound™ system clarifies its complex anatomical relationships and facilitates catheter mapping.
The ablation (A) and ICE (I) catheter positions are displayed in real time.
limiting point-to-point mapping. When creating 2D ICE con- EA mapping by providing a real-time assessment of the adequacy
tours, it is helpful to use complementary long- and short-axis of catheter contact along the papillary muscle.7 In this manner,
imaging planes while avoiding excessive overlap between indi- precise 3D reconstructions of the papillary muscle can be gener-
vidual contours. ated via point-to-point mapping. The CARTO Sound system is
capable of producing nonfluoroscopic 3D geometries from 2D
ICE contours (Figure 62-6).
Papillary Muscle Ablation
The RV and LV papillary muscles are increasingly recognized
as an important source of idiopathic VAs, and are frequently Visualization of Ablation Lesions
encountered barriers to effective ablation in patients with struc-
tural heart disease–related VT. Traditional electroanatomical Regional variability in tissue characteristics with echo imaging is
mapping of these structures can be challenging because of their principally due to differential acoustic impedance (i.e., density)
complex structure and intracavitary position. Furthermore, between tissues within the imaging sector (e.g., blood-
because each VA originates from a discrete aspect of an individual myocardium, normal muscle-scar).8 The emitted ultrasound
papillary muscle, precise positioning of the ablation catheter is beam is attenuated as it passes through tissue, making far-field
required to avoid ineffective lesions. structures appear darker for any given transducer frequency.
Full anatomical characterization of the RV and LV papillary When the ultrasound beam encounters a tissue with different
muscles is possible with ICE. As has been described, this is easily acoustic impedance, more of the ultrasound beam is reflected and
accomplished from the “home” view by gently flexing the cath- detected by the imaging transducer. Regions with increased
eter anteriorly and passing it through the tricuspid annulus. From density reflect more of the ultrasound beam and thus appear
the inferior RVOT, the moderator band and the RV papillary bright or echointense. Because tissue imaging with ICE requires
apparatus are viewed directly. Gentle CCW rotation of the trans- the detection of reflected waves from the emitted ultrasound
ducer demonstrates the insertion of the moderator band into the beam, spatial resolution is maximized by increasing the number
RV free wall. With slow CW rotation from the RV views, the LV of ultrasound beams that reach the region of interest. Practically,
anterolateral (AL) and posteromedial (PM) papillary muscles are this is achieved by increasing the number of ultrasound pulses
seen sequentially. It may be necessary to insert or retract the ICE emitted (i.e., increasing the transducer frequency) and/or imaging
catheter in the RV to view more apical or basal aspects of the structures closer to the ICE transducer (where the emitted beams
papillary apparatus (respectively). are most closely spaced). Thus, tissue characterization with ICE
As described with aortic root ablation, ICE can be used as a is optimal when structures with differential acoustic impedances
2D or a 3D imaging modality for papillary muscle ablation. In are imaged in the echo near-field with a high transducer pulse
2D mode, ICE serves as an important adjunct to fluoroscopy and frequency.
RV
LCC
LCC
AL RCC
RCC LV
AL
NCC
A
PM
PM
PM LV
LV
A B C
Figure 62-6. ICE can facilitate ablation of arrhythmias originating from intracavitary structures such as the papillary muscles. The anterolateral (AL) and posteromedial (PM)
LV papillary muscles can be traced and displayed as separate endocardial geometries using the CARTO Sound™ module; 3D reconstructions are demonstrated in the RAO
(A) and LAO (B) projections. The adequacy of catheter contact along the papillary muscle can also be confirmed in real time (C).
(Images courtesy of Fermin C. Garcia, MD.)
Abl
LA LV
LA
Abl
LS
Ao LI
A B
Figure 62-7. ICE provides real-time feedback regarding the quality of catheter (Abl) contact during mapping and ablation. Local tissue swelling and increased echointensity
(arrows) often occur during ablation and persist for a variable duration thereafter. Ao, Aorta; LA, left atrium; LI, left inferior PV; LS, left superior PV; LV, left ventricle.
Local atrial tissue changes during ablation are often visualized Data describing left atrial fibrosis with delayed enhancement
with ICE; most commonly noted are increased tissue echointen- cardiac magnetic resonance (MR) have not been replicated with
sity and wall thickness (Figure 62-7). These changes can occur ICE, likely because of the limited spatial resolution noted with
within the first 10 seconds of energy application and persist for commercially available ICE transducers. Detection of interstitial
a variable time duration after energy delivery is terminated.9 myocardial fibrosis with echocardiography has been previously
Although data correlating acute tissue changes on ICE with long- described; however, it required imaging at an extremely high
term lesion characteristics are lacking, visualizing these changes transducer frequency.11
provides the operator with important additional qualitative feed- Patients with post-infarction VT characteristically have trans-
back regarding catheter contact. Transient changes in contractile mural infarctions, often with aneurysm formation. Similar to
function adjacent to ablation lesions can be demonstrated with transthoracic echocardiography, post-infarct changes can also be
ICE. The presence of heterogeneous myocardial perfusion with visualized with ICE as echointense regions with akinesis and wall
ICE after intracoronary contrast administration has also been thinning. By using the CARTO Sound system, diseased myocar-
correlated with lesion pathology in an animal model.10 dial segments can be traced and transposed onto the 3D EA map,
thereby limiting the extent of point-to-point geometry creation.
This technique of ICE-guided scar geometry creation has been
shown to correlate well with bipolar mapping in post-infarction
Visualization of Arrhythmia Substrate patients.12
The heterogeneous, complex 3D substrate distribution
Because regions of fibrotic and normal myocardium differ in encountered in patients with nonischemic LV cardiomyopathy-
acoustic impedance, it is technically possible to distinguish them related VT presents a unique challenge to effective catheter abla-
with ICE. It is axiomatic that larger, more confluent regions of tion. Full characterization of the substrate in these patients
fibrosis will be more easily detected. Likewise, imaging of thinner is challenging because of the limited field-of-view of bipolar
structures (e.g., atria) will require a significantly higher spatial electrograms.13 In selected patients, ICE can detect confluent
resolution than is required for thicker structures (e.g., ventricles). regions of intramural or epicardial fibrosis as an echointense
RV
RA
62
S T
IAS
LV
LA
PE
A B
Figure 62-8. Procedural complications can be detected early with ICE, thereby limiting potential damage. The presence of soft thrombus adherent to vascular sheaths or
intracardiac leads can be visualized with ICE. Panel A is an ICE image obtained from the RA, demonstrating a large thrombus (T) attached to a transseptal sheath (S) that
was detected during transseptal puncture; this occurred despite an activated clotting time (ACT) of 360 seconds and an international normalized ratio (INR) of 3.2. The
thrombus was cleared before left atrial crossing with no clinical sequelae. Routine ICE screening during complex ablation procedures may detect asymptomatic pericardial
effusions (PE) before the development of hemodynamic compromise; this is best accomplished by placing the ICE catheter in the right ventricular outflow tract (B). Panel
B demonstrates a small effusion that was detected in the asymptomatic state; systemic anticoagulation was reversed, thereby avoiding the need for percutaneous
drainage.
stripe “sandwiched” by normal myocardium.14 This distinct placed in the right ventricle (see Figure 62-8). It is our practice
echocardiographic signature may provide important insight into to screen for the presence of effusion periodically during the
the presence of VT substrate deep to the endocardium, thus ablation procedure or with any change in hemodynamic param-
facilitating the decision to pursue percutaneous epicardial eters. In the absence of pericardial adhesions (e.g., from prior
mapping. cardiac surgery), most patients have a circumferential fluid col-
lection that is initially seen around the cardiac base. With early
detection and prompt reversal of anticoagulation, some effusions
will stabilize without the need for pericardiocentesis. If drainage
Detecting and Preventing Complications is required, ICE may be used to guide the optimal location for
pericardial puncture and to document complete resolution.
The contribution of ICE to enhancing the safety of EP proce-
dures cannot be understated. The real-time imaging and excel-
lent tissue resolution achieved with ICE provide maximal Pulmonary Vein Stenosis
sensitivity to detect potential complications and to minimize
damage by providing rapid, early intervention. The incidence of symptomatic PV stenosis requiring interven-
tion after PV isolation (PVI) is approximately 0.3%.17 An approx-
imate 50% reduction in the worldwide incidence of PV stenosis
Intracardiac Thrombus has been noted since 2005, most likely because of broad adoption
of a proximal or antral PV isolation strategy. Nonetheless, rare
Occult thromboembolism represents an important source of instances of PV stenosis still occur and often incur substantial
morbidity from complex ablation procedures. The importance of patient morbidity. The incorporation of ICE with antral PV
adequate systemic anticoagulation during left atrial ablation pro- isolation may further reduce this incidence of stenosis. Acute
cedures was reinforced by frequent documentation of sheath- changes in PV pulsed wave Doppler flow velocity, as well as
associated thrombus, the incidence of which was reduced during increased turbulence of color Doppler PV flow, correlate with
AF ablation by targeting a higher activated clotting time (ACT) narrowing of PV ostial diameter. Particular attention is required
before left atrial instrumentation.15 Sheath-associated thrombosis for patients undergoing repeat PVI procedures, as significant
is occasionally seen despite aggressive anticoagulation, and if baseline ostial narrowing may be present in the absence of clinical
documented with ICE, corrective measures may prevent systemic symptoms. An acute increase in PV pulsed wave Doppler flow
embolization (Figure 62-8). Mobile thrombus has been docu- velocity beyond 100 cm/s is unusually encountered with an antral
mented with ICE in 30% of patients with implanted pacemaker ablation strategy, and should prompt reevaluation of the ablation
and defibrillator leads, and may increase the risk of both acute strategy used, as well as close long-term follow-up for clinical
and chronic embolic complications.16 symptoms.
Pericardial Effusion
Summary
The diagnosis of procedure-related pericardial effusion is often
delayed until hemodynamic compromise occurs. This time lag In conclusion, ICE is a versatile imaging platform that integrates
delays both the reversal of systemic anticoagulation and the insti- seamlessly with complex EP procedures. ICE can be used as a
tution of corrective measures. Patients with impaired cardiac stand-alone 2D modality, or it can be incorporated with EA
function are at particular risk and may exhibit profound compro- mapping to generate nonfluoroscopic 3D geometries. ICE pro-
mise due to systemic hypoperfusion. The presence of intraperi- vides real-time anatomical and physiological data that enhance
cardial fluid can be easily documented with the ICE catheter the safety and efficacy of EP procedures.
References 6. Yamada T, Litovsky SH, Kay GN: The left ven-

tricular ostium: An anatomic concept relevant to
12. Khaykin Y, Skanes A, Whaley B, et al: Real-time
integration of 2D intracardiac echocardiography
idiopathic ventricular arrhythmias. Circ Arrhythm and 3D electroanatomical mapping to guide ven-
1. Marchlinski FE, Ren JF, Schwartzman D, et al:
Electrophysiol 1:396–404, 2008. tricular tachycardia ablation. Heart Rhythm
Accuracy of fluoroscopic localization of the crista
7. Seiler J, Lee JC, Roberts-Thomson KC, et al: 5:1396–1402, 2008.
terminalis documented by intracardiac echocar-
Intracardiac echocardiography guided catheter 13. Bogun FM, Desjardins B, Good E, et al: Delayed-
diography. J Interv Card Electrophysiol 4:415–421,
ablation of incessant ventricular tachycardia from enhanced magnetic resonance imaging in non
2000.
the posterior papillary muscle causing tachycardia- ischemic cardiomyopathy: Utility for identifying
2. Ren JF, Marchlinski FE, Callans DJ, et al: Echo-
mediated cardiomyopathy. Heart Rhythm 6:389– the ventricular arrhythmia substrate. J Am Coll
cardiographic lesion characteristics associated with
392, 2009. Cardiol 53:1138–1145, 2009.
successful ablation of inappropriate sinus tachycar-
8. Silvestry FE, Wiegers SE: Intracardiac Echocar- 14. Bala R, Ren JF, Hutchinson MD, et al: Assessing
dia. J Cardiovasc Electrophysiol 12:814–818, 2001.
diography. London, New York, 2006, Taylor & epicardial substrate using intracardiac echocar-
3. Saksena S, Sra J, Jordaens L, et al: Group I-CIS: A
Francis. diography during vt ablation. Circ Arrhythm Elec-
prospective comparison of cardiac imaging using
9. Weerasooriya R, Jais P, Sanders P, et al: Images in trophysiol 4:667–673, 2011.
intracardiac echocardiography with transesopha-
cardiovascular medicine: Early appearance of an 15. Ren JF, Marchlinski FE, Callans DJ, et al: Increased
geal echocardiography in patients with atrial fibril-
edematous tissue reaction during left atrial linear intensity of anticoagulation may reduce risk of
lation: The intracardiac echocardiography guided
ablation using intracardiac echo imaging. Circula- thrombus during atrial fibrillation ablation proce-
cardioversion helps interventional procedures
tion 108:e80, 2003. dures in patients with spontaneous echo contrast. J
study. Circ Arrhythm Electrophysiol 3:571–577,
10. Khoury DS, Rao L, Ding C, et al: Localizing and Cardiovasc Electrophysiol 16:474–477, 2005.
2010.
quantifying ablation lesions in the left ventricle by 16. Supple GE, Ren JF, Zado ES, et al: Mobile throm-
4. Singh SM, Heist EK, Donaldson DM, et al: Image
myocardial contrast echocardiography. J Cardio- bus on device leads in patients undergoing abla-
integration using intracardiac ultrasound to guide
vasc Electrophysiol 15:1078–1087, 2004. tion: Identification, incidence, location, and
catheter ablation of atrial fibrillation. Heart
11. Chandraratna PA, Whittaker P, Chandraratna association with increased pulmonary artery sys-
Rhythm 5:1548–1555, 2008.
PM, et al: Characterization of collagen by high- tolic pressure. Circulation 124:772–778, 2011.
5. Lee W, Griffin W, Wildes D, et al: A 10-Fr ultra-
frequency ultrasound: Evidence for different 17. Cappato R, Calkins H, Chen SA, et al: Updated
sound catheter with integrated micromotor for
acoustic properties based on collagen fiber mor- worldwide survey on the methods, efficacy, and
4-D intracardiac echocardiography. IEEE Trans
phologic characteristics. Am Heart J 133:364–368, safety of catheter ablation for human atrial fibrilla-
Ultrason Ferroelectr Freq Control 58:1478–1491,
1997. tion. Circ Arrhythm Electrophysiol 3:32–38, 2010.
2011.
Exercise-Induced Arrhythmias 63
Felix Yang, Mintu P. Turakhia, and Victor F. Froelicher
underwent exercise testing.4 Exercise treadmill–induced supra-

CHAPTER OUTLINE
ventricular arrhythmia (ETISVA) was noted in 85 subjects (6%).
Exercise-Induced Atrial Arrhythmias 613 An eightfold increase in the relative risk of developing lone atrial
fibrillation (AF) was noted in subjects with exercise-induced
Exercise-Induced Ventricular Arrhythmias 613
ETISVA. In a 5.7-year follow-up, 10% of 85 subjects with
Outflow Tract Ventricular Tachycardia 614 ETISVA developed AF or paroxysmal supraventricular tachycar-
dia (SVT). Therefore, ETISVA may be a marker for AF or
Idiopathic Left Ventricular Tachycardia 615
paroxysmal SVT during follow-up.
Arrhythmogenic Right Ventricular Dysplasia 615 The relationship between exercise and atrial fibrillation is
particularly relevant to athletes. The overall risk for AF is signifi-
Hypertrophic Cardiomyopathy 615
cantly higher in athletes than in controls (odds ratio 5.29, 95%
Long QT Syndrome 616 confidence interval [CI] 3.57 to 7.85, P = .0001).5 Endurance
sports increase preload, which increases atrial pressure and there-
Catecholaminergic Polymorphic Ventricular
fore shortens atrial refractory periods and increases the disper-
Tachycardia 616 sion of atrial refractoriness. Increased vagal tone in athletes,
Antiarrhythmics and Exercise-Induced sympathetic surges during exercise, and fluid and electrolyte
Ventricular Arrhythmias 618 changes during exercise may also contribute to the development
of atrial arrhythmias.6 Anatomically, athletes may have larger left
Conclusion 618 atrial dimensions and fibrosis secondary to chronic systemic
inflammation from excessive endurance exercise.7
In patients with no structural heart disease, treatment is gen-
Several physiological changes that occur during exercise may erally targeted at addressing the trigger for the arrhythmia. A
precipitate cardiac arrhythmias. Activation of the sympathetic reduction in exercise intensity or duration is often highly effective
nervous system results in an increase in circulating catechol- in reducing arrhythmia burden.8 However, many patients, espe-
amines.1 Increased automaticity and enhanced triggered activity cially competitive athletes, may be unwilling or unable to reduce
may increase the likelihood of arrhythmias. The presence of or refrain from exercise. In these cases, β-blockers, antiarrhyth-
premature beats during exercise can lead to initiation of reentrant mics, and catheter ablation can be considered.
supraventricular and ventricular arrhythmias. Important factors In a study of 5375 patients with known or suspected coronary
in arrhythmogenicity include electrolyte shifts, baroreceptor acti- artery disease (CAD), 24% of patients developed atrial ectopy,
vation, myocardial stretch, ischemia, and genetic predisposition. 3.4% developed SVT, and 0.8% developed AF upon treadmill
Exercise can increase potassium levels, decrease pH, and raise testing.9 Exercise treadmill–induced supraventricular arrhyth-
catecholamines.2 These catecholamines may counterbalance the mias were not predictive of any end point.
harmful cardiac effects of hyperkalemia and acidosis and improve
action potential characteristics in potassium-depolarized ven-
tricular myocytes.3 In normal myocardial tissue, hyperkalemia
decreases the incidence of norepinephrine-induced arrhythmias. Exercise-Induced Ventricular Arrhythmias
However in ischemic or infarcted tissue, hyperkalemia and cat-
echolamines may jointly potentiate arrhythmias. The heart is also Apparently Healthy Subjects
at increased risk in the post-exercise period. During this time,
plasma potassium is low and adrenergic tone is high. An abnor- Asymptomatic patients without prior evidence of CAD have
mal regulation of sympathovagal balance and electrolytes in been noted to have variable rates of ventricular ectopy. In the
recovery, compounded with ischemia, may increase the suscepti- Advisory Group for Aerospace Research and Development study
bility to arrhythmias.1 Despite the myriad of physiological of 1640 healthy aviators, the prevalence of PVCs (other than
changes that occur in exercise, in the absence of structural or single or occasional) increased with age: 6.6% for ages 20 to 29,
electrical heart disease sudden death due to arrhythmias is 7.6% for ages 30 to 39, and 13.1% for ages 40 to 53.10 The per-
extremely rate. centages of patients with three or more consecutive PVCs were
0.8% for ages 20 to 29, 1.0% for ages 30 to 39, and 3.5% for
ages 40 to 53. In another study of 597 male and 325 female
healthy adult volunteers, only 1.1% of the patients had exercise
Exercise-Induced Atrial Arrhythmias treadmill–induced ventricular arrhythmias (ETIVAs).11 These
episodes were typically asymptomatic, short, and limited to 3 to
Exercise-induced atrial arrhythmias are less common than 6 beats, usually near peak exercise. The prognostic significance
ventricular arrhythmias. In the Baltimore Longitudinal Study of ventricular ectopy in asymptomatic healthy individuals remains
of Aging, 1383 asymptomatic volunteers aged 20 to 94 years controversial.
613
Coronary Artery Disease pulmonary artery (ALCAPA) who reached adulthood, would also
be at risk for exercise-induced arrhythmias.
ETIVA appears to be more common in patients with known
CAD. The prevalence of any ventricular arrhythmias (including
simple PVCs) in patients with CAD ranges from 10% to 40%,
with most studies ranging between 20% and 30%. Excluding Outflow Tract Ventricular Tachycardia
simple PVCs, the prevalence of more complex ventricular
arrhythmias is lower. Whether or not ETIVA in patients with Outflow tract VT should be considered in patients with a struc-
known coronary artery disease is associated with a worse prog- turally normal heart and a QRS in VT that features a left bundle
nosis is unclear. branch block morphology and an inferior axis. Most outflow tract
In a population of veterans referred for exercise stress testing, VTs originate from the RVOT (80%), and the remainder origi-
the risk of mortality in patients with resting (pre-exercise) PVCs nate from the left ventricular outflow tract (LVOT).20 RVOT VT
and ETIVA was increased.12 The combination of rest PVCs has a left bundle branch QRS morphology and an inferior axis,
and ETIVA carries the highest risk.1 These variables were inde- with an R/S transition typically in V3 or V4. LVOT VT may also
pendent predictors of cardiovascular mortality after adjustment have left bundle branch QRS morphology but with small R waves
for other clinical and exercise test variables, which included in V1 and an earlier R/S transition. Lerman and collegues have
exercise-induced ischemia. However it is possible that additional determined that outflow tract VTs are the product of triggered
adjustment for left ventricular functional abnormalities or coro- activity secondary to cyclic adenosine monophosphate–mediated
nary disease burden might have mitigated the association. delayed afterdepolarizations. 21 The VT is adrenergically medi-
Regardless, patients found to have arrhythmias during exercise ated and is sensitive to perturbations that lower intracellular
testing should undergo an evaluation of their left ventricular calcium such as adenosine and verapamil. The diagnosis of idio-
function. pathic VT is one of exclusion; therefore other causes of left
bundle branch block (LBBB) pattern VT should be considered.
Clinically, outflow tract VT accounts for the vast majority of
Coronary Anomalies idiopathic VTs. The age of presentation is usually 30 to 50 years,
and patients usually have a benign clinical course.20 The most
Congenital coronary anomalies are implicated in 10% to 20% of common complaint among patients is palpitations (48% to 80%),
all deaths in young athletes.13 The right coronary artery arising followed by presyncope or light-headedness (28% to 50%).
from the left coronary sinus is more common than the left coro- Syncope is rare (<10%) and SCD is extremely rare. However,
nary artery arising from the anterior sinus, although the latter is some patients may develop cardiomyopathy from incessant repet-
a more common cause of sudden death. Of the four types of itive monomorphic VT or from a high burden of ventricular
anomalous left coronary arteries, the interarterial type is the only premature complexes (VPCs). Ablation of the focus usually nor-
type that places the patient at increased risk for sudden death. In malizes left ventricular function in a few months.
this variant, the left coronary artery arises from the right cusp The spectrum of outflow tract VT includes three clinical
and passes anteriorly between the aorta and the right ventricular subtypes.22 Patients may have repetitive monomorphic VPCs,
outflow tract (RVOT). Similarly, an interarterial course of an repetitive nonsustained monomorphic VT, or exercise-induced
anomalous right coronary artery that arises from the left cusp sustained VT. Patients may present with a predominance of one
would also put patients at risk for sudden cardiac death (SCD). type; however, significant overlap has been noted, and it is
However, the larger territory supplied by the left coronary artery believed that the subtypes share the same cellular mechanism.
results in increased risk over right coronary artery anomalies. Generally, outflow tract tachycardias are provoked by exercise,
SCD associated with or shortly after vigorous exercise is very and treadmill testing is useful in reproducing the clinical VT.
unusual after the patient is >35 years of age. Approximately 70% of patients who present with sustained VT
Angelini et al report that the incidence of anomalous coronary will have VT induced by exercise testing.22 However, in patients
arteries was 1.07% (right anomalous coronary from the left coro- who present with monomorphic nonsustained ventricular tachy-
nary sinus in 0.92%; left anomalous coronary from the right cardia (NSVT) or VPCs on monitoring, exercise testing may
coronary sinus in 0.15%).14 Davis et al. report a prevalence of induce sustained VT in only 10%. Overall, exercise testing repro-
0.17% for anomalous origins of coronary arteries among 2388 duces VT in less than 50% of patients with clinical VT. As a
children and adolescents.15 Among 1686 coronary anomalies result, exercise testing may not be a reliable indicator of β-blocker
found in 126,595 adult coronary angiograms, Yamanaka and or antiarrhythmic efficacy, and ambulatory monitoring would be
Hobbs reported an incidence of 0.17% for anomalous left coro- an appropriate adjunct.
nary arteries and 0.107% for anomalous right coronary Two responses of outflow tract VT to exercise testing have
arteries.16 been reported. In the first case, VT occurs during acceleration
The mechanism by which the coronary anomaly causes SCD of the heart rate with exercise. A progression from VPCs to salvos
is hypothesized to be a sudden occlusion of the vessel that may of NSVT to sustained VT may be observed. In contrast, patients
involve damage, thrombosis, or spasm, resulting in severe myo- with repetitive monomorphic VT may have suppression of their
cardial ischemia and ventricular tachycardia/fibrillation.17 During VT during exercise and development of VT during the recovery
exercise, increased dP/dt and stroke volume may result in phase of exercise.23 These responses indicate that a critical
increased systolic expansion of the proximal aorta and pulmonary window of heart rates is required for VT initiation. This cycle
artery, which could collapse the proximal anomalous coronary length dependence of ventricular ectopy may also be observed on
artery. Because exercise results in a greater percentage of time ambulatory monitoring. 23,24
spent in systole, which is when compression occurs, SCD most In general, patients with exercise-induced outflow tract VTs
frequently occurs with or shortly after exercise. have no structural heart disease. In contrast, patients with post-
Other coronary anomalies may also be implicated in arrhyth- infarction septal VTs will have a history of coronary disease and
mias and sudden death. Patients with a single coronary artery that myocardial infarction. Bundle branch reentry VT is usually seen
divides into all three major branches are at risk for SCD during in the setting of structural heart disease, most commonly a dilated
athletic activity.18,19 Patients with hypoplasia of portions of the cardiomyopathy.25 Patients with antidromic atrioventricular
coronary tree or with coronary fistulas, or the small minority of reciprocating tachycardia (AVRT) using an atriofascicular bypass
patients with an anomalous left coronary artery arising from the tract might also demonstrate an LBBB morphology VT, although
the axis is usually leftward.26 The triggered activity of outflow Ventricular arrhythmias can occur during exercise, and SCD
63
tract VT also differs from the reentry mechanisms of post- is thought to be due to acceleration of VT with degeneration into
infarction VT, bundle branch reentry VT, and antidromic AVRT ventricular fibrillation. The most common arrhythmia is sus-
using an atriofascicular bypass tract. Although they characteristi- tained or nonsustained monomorphic VT from the right ven-
cally have exercise-induced VT, patients with catecholaminergic tricle, therefore manifesting a left bundle branch pattern. Usually
polymorphic ventricular tachycardia (CPVT) will have polymor- the VT seen in ARVD will be of left bundle branch morphology
phic ventricular ectopy and will manifest bidirectional VT. Last, with a superior rather than inferior axis. However, preferential
outflow tract morphology VTs may be a manifestation of arrhyth- or isolated outflow tract involvement has been described. Exer-
mogenic right ventricular dysplasia (ARVD), although the mor- cise testing may demonstrate monomorphic VT in up to 50% to
phology of VT is often varied in ARVD. 60% of patients. Exercise is thought to impart increased stress
In addition to exercise, ventricular ectopy, including sustained on the right ventricle and may increase right ventricular dilata-
VT, can be provoked by emotional stress. High levels of sympa- tion and worsen the manifestations of ARVD. Accordingly,
thetic tone contribute to arrhythmogenicity. This is also illus- patients with ARVD should not engage in competitive sports or
trated in the circadian variations in episodes of VT, ventricular endurance training and should not participate in any activities
runs (2 to 4 beats), and VPCs. Hayashi et al. demonstrated that that cause palpitations, presyncope, or syncope.34
peaks for these ventricular arrhythmias occurred around 7 am and The diagnosis of ARVD consists of the fulfillment of multiple
6 pm.27 β-Blockers completely eliminated VT episodes and criteria.35 ECG findings suggestive of ARVD include the pres-
blunted ventricular runs; however rates of single VPCs were ence of right ventricular (RV) conduction delay, precordial T
similar to those before β-blockade therapy. wave inversions, and epsilon waves.35,36 The presence of notching
in the QRS complex of the LBBB VT may also suggest ARVD.37
Treadmill exercise testing may be used to elicit the characteristic
LBBB morphology of ventricular tachycardia of ARVD. Nonsus-
Idiopathic Left Ventricular Tachycardia tained or sustained VT of LBBB morphology with a superior axis
fulfills a major criterion for the diagnosis of ARVD, whereas an
Approximately 10% of idiopathic VTs originate from the fascicles inferior axis fulfills a minor criterion.35 Other corroborative data
of the left ventricle. These arrhythmias are referred to as fascicular are needed for the diagnosis of ARVD, and the usual diagnostic
VT or verapamil-sensitive VT.28 This condition should be consid- work-up for ARVD consists of a signal-averaged ECG, echocar-
ered in the differential diagnosis of right bundle branch block diogram, cardiac magnetic resonance imaging (MRI), a detailed
(RBBB) with left anterior fascicular block pattern VT. Less com- family history, and, if needed, invasive electroanatomic mapping
monly, a left posterior block pattern may be seen (5% to 10%).28 of the right ventricle.
The usual age of presentation is between 15 and 40 years, and In addition to exercise restriction, treatment for ARVD
patients usually have normal resting electrocardiogram (ECG) involves implantation of an implantable cardioverter-defibrillator
and left ventricular function.29 Although verapamil-sensitive VT (ICD) in patients deemed at high risk for arrhythmic events and
was originally described at rest, VT is sensitive to exercise or for secondary prevention of SCD.38 Medical management with
emotional stress, is frequently precipitated by exercise, and may antiarrhythmics or sotalol may be used in those who are not
be seen during or after exertion. candidates for ICD therapy and in those with frequent shocks.
This reentrant arrhythmia is believed to involve the posterior Radiofrequency ablation also plays a role in patients with recur-
Purkinje system with the left posterior fascicle as one limb and rent ventricular arrhythmias.
abnormal Purkinje tissue with slow, decremental conduction as
the other limb.30 In the small percentage of patients with RBBB
and a right inferior axis VT (left posterior block pattern), the left
anterior fascicle is thought to be the involved Purkinje system Hypertrophic Cardiomyopathy
and the site of ventricular exit.31 The circuit is sensitive to cate-
cholamines as evidenced by induction by exercise, and induction Described in further detail in Chapter 86, hypertrophic cardio-
during electrophysiological testing is facilitated by isoproterenol myopathy (HCM) is a disease of left ventricular hypertrophy with
infusion. Whereas adenosine has no effect on the arrhythmia, a spectrum of clinical manifestations and hemodynamic abnor-
administration of verapamil slows the VT rate and then termi- malities that are most often caused by mutations in one of several
nates it.32 sarcomere genes. The mechanism of SCD in HCM patients is
Patients suspected to have verapamil-sensitive VT should likely multifactorial but ultimately results from ventricular
undergo an evaluation to exclude structural heart disease. An tachyarrhythmias. The arrhythmia is thought to result from elec-
echocardiogram, a stress test, and/or cardiac catheterization may trical instability and distorted electrophysiological propagation
be indicated depending on the clinical suspicion for coronary from the disorganized arrangement of cardiac muscle cells of
disease. Medical treatment for verapamil-sensitive VT involves HCM. Findings of increased SCD during the afternoon hours in
verapamil; however patients with severe or recurrent symptoms athletes support an exercise-induced mechanism.34 This is also
can undergo catheter-based ablation with high rates of success.33 supported by findings of sinus tachycardia preceding VT/VF
events in stored ICD events, which suggest that high sympathetic
drive is proarrhythmic when a susceptible substrate is present.39
Risk stratification for SCD involves an assessment of family
Arrhythmogenic Right Ventricular Dysplasia history of SCD, a personal history of syncope, determination of
left ventricular wall thickness, assessment of the blood pressure
Patchy replacement of right ventricular myocardium by fibrofatty response and outflow tract gradient to treadmill exercise, and
tissue is the hallmark of arrhythmogenic right ventricular dyspla- NSVT on Holter monitoring.
sia (ARVD) and provides a substrate for reentrant ventricular Although NSVT during Holter monitoring has been demon-
arrhythmias. Generally, this fibrofatty replacement preferentially strated to be a useful prognostic factor, ventricular arrhythmias
affects the free wall of the right ventricle; however, the disorder induced during exercise testing have not been well established as
can affect the outflow tract, left ventricle, and septum. ARVD is a prognostic indicator of SCD. In a study of 263 patients with
more common in males and predominantly affects young adults, HCM who underwent exercise testing, 3.0% of patients had new
with a mean age of diagnosis around 30 years. nonsustained atrial arrhythmias and 4.2% had nonsustained
ventricular arrhythmias.40 In a larger series by Gimeno et al. of 3 minutes of recovery) − (QTc baseline) ≥30 ms, reliably distin-
1380 patients referred to a cardiomyopathy clinic, only 27 (2.0%) guished patients with manifest or concealed LQT-1 from those
had NSVT (defined as 3 or more consecutive beats at a rate of with LQT-2 and LQT-3.
≥120 beats/min) or VF during exercise testing.41 Although NSVT Clinically, 40% of LQT-1 patients will become symptomatic
was not frequent, researchers found that the hazard ratio for by age 10, and very few will develop symptoms for the first time
exercise-induced NSVT/VF (HR = 3.14) in predicting SCD or after the age of 20. As a result, asymptomatic adults who are
appropriate ICD discharge was higher than that of all other diagnosed in adulthood may be at low risk for events. Regardless,
predictors (HR = 2.57 for NSVT on Holter; HR = 1.79 for family β-blockers are the treatment of choice and are extremely effective
history of SCD). High rates of SCD or ICD discharge in patients in reducing symptoms and lethal events. Patients who have recur-
with exercise-induced ventricular arrhythmia suggest that tread- rences despite maximal β-blocker therapy may be considered for
mill exercise may be a useful tool for assessing SCD risk in HCM. ICD implantation or left stellate ganglion ablation.
Prevention of ventricular tachyarrhythmias and prevention of
SCD involve the use of ICDs.42 The risk of SCD may also be
reduced with septal myectomy in patients with symptomatic
LVOT obstruction, although septal ablation has not been found Catecholaminergic Polymorphic
to yield similar results.43 Pharmacologic agents such as antiar- Ventricular Tachycardia
rhythmic drugs or negative inotropes have little impact on the
incidence of SCD. Last, because of the potential risk of SCD Catecholaminergic polymorphic ventricular tachycardia (CPVT)
associated with exercise in HCM patients, activity restriction is is a rare inherited syndrome that is characterized by adrenergi-
advised.34 cally induced supraventricular and ventricular arrhythmias in
structurally normal hearts and is associated with syncope and
sudden death. It typically manifests in children and adolescents
and provokes symptoms in up to 80% of patients younger than
Long QT Syndrome 40 years of age.51 SCD is often the first manifestation of the
disease, which carries a 30% to 40% overall mortality by age 40.51
Detailed discussion of long QT syndrome (LQTS) is found in The basis for the overwhelming majority of cases of CPVT is
other chapters; however, we will discuss here some of the distin- a mutation in the cardiac ryanodine receptor RyR2, which is
guishing features of LQT-1 that pertain to exercise. LQT-4 and inherited in an autosomal dominant fashion.52 Mutations in the
LQT-7 will be discussed in the following section, given some ryanodine receptor are implicated in the uncontrolled release of
overlapping features with CPVT. The most common LQTS is calcium from the sarcoplasmic reticulum in cardiomyocytes
LQT-1, which is the type whose events are most often triggered during electrical diastole, which produces delayed afterdepolar-
by adrenergic stimuli and exercise. Events in LQT-2 patients are izations and cardiac arrhythmias.53 Genetic screening allows suc-
less often triggered by adrenergic stimulation, and events in cessful genotyping of CPVT in approximately 50% to 60% of
LQT-3 patients usually occur during rest or sleep.44 LQT-1 patients with a definitive clinical diagnosis of CPVT.
should be considered in patients who have a prolonged QTc and Exercise testing is particularly useful in the evaluation of
a history of syncope with stress. The T wave morphology in patients with CPVT. One of the diagnostic findings in CPVT is
LQT-1 is typically broad-based and high in amplitude. This the progressive increase in supraventricular and ventricular
contrasts with the precordial biphasic T wave pattern of LQT-2 arrhythmias that parallels increases in exercise workload. Ectopic
and the long isoelectrical ST-T segment of LQT-3. supraventricular beats and polymorphic PVCs usually begin at a
Arrhythmic events in LQT-1 generally occur at relatively heart rate of approximately 100 to 110 beats/min. Supraventricu-
elevated heart rates. Tan et al. observed an average sinus heart lar arrhythmias include isolated atrial ectopic beats and can prog-
rate of 98 beats/min before torsades de pointes for LQT-1 in ress to nonsustained SVT or even short runs of atrial fibrillation.
contrast to an average of 73 beats/min for LQT-2.45 The preced- Ventricular ectopic beats also progress during exercise into
ing TdP in LQT-1 is not pause dependent, unlike that in LQT-2 bigeminy, couplets, and runs of bidirectional VT. Continuation
patients.46 of exercise may result in faster and more disorganized VT, which
The defect in LQT-1 is a mutation in KCNQ1, resulting in could degenerate into VF. During the recovery phase, these
loss-of-function in IKs.47 At fast heart rates, IKs is the most impor- arrhythmias gradually diminish.
tant current that shortens the action potential.48 As a result, the The hallmark of CPVT is bidirectional ventricular tachycar-
dysfunctional IKs in LQT-1 patients results in an inability to dia, where the beat-to-beat axis rotates 180 degrees each beat (see
appropriately shorten QTc during exercise.47 This particular Figure 63-1). This contrasts with torsades de pointes, where the
response to exercise and adrenergic stimulation is of particular QRS axis gradually and chaotically rotates around the baseline.
interest in exercise stress testing and epinephrine QT stress However, a small minority of CPVT patients have been noted to
testing. show irregular polymorphic VT. The vast majority of the initiat-
Observation of a paradoxical prolongation of the uncorrected ing beat of VT has been localized to the RVOT.54 The second
QT interval by >30 ms during infusion of low-dose epinephrine most common source was the LVOT. The morphology of the
suggests a diagnosis of LQT-1.49 Epinephrine QT stress testing initiating beat of bidirectional VT is generally reproducible in
has been demonstrated to have a positive predictive value of 76% 80% of patients with multiple episodes of VT.
and a negative predictive value of 96% for LQT-1, even when In general, CPVT patients have a normal ECG at baseline
the baseline ECG has a normal QTc.49 but may demonstrate prominent U waves and slower resting
Ackerman and associates analyzed 243 treadmill tests of heart rates. A normal QT interval is useful in distinguishing
patients studied for LQTS and reported good diagnostic accu- CPVT from long QT patients. However, some long QT patients
racy of treadmill testing for the diagnosis of LQT-1.50 During exhibit borderline QT prolongation, and a couple of long QT
exercise, QTc decreases in LQT-2, LQT-3, and control patients. syndromes share CPVT-type phenotypes. Adrenergically trig-
In contrast, LQT-1 patients essentially have unchanged QTc gered bidirectional VT is also seen in long QT-4 (a mutation in
durations. Additionally, the QTc of concealed LQT-1 patients the ANK2 gene, which codes for cardiac ankyrin-B) and long
actually increased at peak exercise (Figure 63-1). Investigators QT-7, otherwise known as Andersen-Tawil syndrome (a muta-
found that an absolute QTc ≥460 ms during the recovery phase tion in KCNJ2, which codes for a potassium channel). LQT-4 is
or a paradoxical increase in QTc, which was defined as (QTc at associated with mild QTc prolongation, AF, sinus bradycardia,
CPVT LQT1
Baseline Exercise
63
I
I
V2
II
VI
V4
aVF
V5
V4
V6
V6
A
B QTc452 ms QTc590 ms
CPVT LQT1
DAD EAD
Delayed
afterdepolarizations Long APD:
early afterdepolarizations
Ca2
I Ks 1-receptor
NCX
Adenylate
cyclase
Ca2 3Na P Gs
Ca2 Ca2 P PKA
+

Ca2
yoitao
K PP1
P P P
cAMP
P
F
P
F
P
RyR
F
P
F
Ca2
calsequestrin
Ca2
C
Figure 63-1 Exercise response in catecholaminergic polymorphic ventricular tachycardia (CPVT) and long QT syndrome type 1 (LQT-1).
A, Bidirectional ventricular tachycardia (alternating QRS axis) during exercise in a patient with CPVT. B, Paradoxical QT prolongation during exercise in a patient with LQT-1.
C, Mechanisms of arrhythmogenesis. In CPVT (left), mutations (white zigzag line) in either the ryanodine receptor (RyR) or the calsequestrin gene make calcium leak. Excess
calcium is transported by the Na+/Ca2+ exchanger (NCX), which brings in three sodium ions for each calcium, thereby generating a slow depolarization (transient inward
current) that can reach threshold and generate a delayed afterdepolarization (DAD). In long QT syndrome (right), a mutated potassium channel (zigzag line) does not
enhance IKs on phosphorylation, and the action potential prolongs, which can lead to early afterdepolarizations (EADs). APD, Action potential duration; cAMP, cyclic adenosine
monophosphate; PKA, protein kinase A; PP1, protein phosphatase 1.
(A, From Francis J, Sankar V, Nair VK, Priori SG: Catecholaminergic polymorphic ventricular tachycardia. Heart Rhythm 2:550-554, 2005. B, From Takenaka K, Ai T, Shimizu W, et al:
Exercise stress test amplifies genotype-phenotype correlation in the LQT1 and LQT2 forms of the long-QT syndrome. Circulation 107:838-844, 2003.)
sinus node dysfunction, and polyphasic U waves. Andersen-Tawil isoproterenol infusion, with a family or personal history of
syndrome is characterized by QT prolongation, prominent U syncope during exercise or emotional stress.55 Infusion of isopro-
waves, facial dysmorphisms, and periodic paralysis. The bidirec- terenol can also be used to elicit the characteristic arrhythmias,
tional VT rate in Anderson-Tawil syndrome has been reported although exercise testing is preferred. Monomorphic or polymor-
to be much slower than in CPVT.55 Additional features of these phic PVCs are insufficient for a diagnosis of CPVT, and further
syndromes, along with genetic testing, help distinguish them testing or monitoring is required for a diagnosis.
from CPVT. The main and most effective therapy for CPVT is β-blockade.38
A definitive diagnosis of CPVT depends on the finding of Follow-up exercise testing is advised as a method of assessing the
bidirectional VT in exercise ECG, Holter monitoring, or adequacy of β-blocker dosage. Compelling evidence indicates
that the addition of flecainide at doses of 150 to 200 mg/day would warrant dose reduction or discontinuation of the drug.
is also effective in suppressing ventricular arrhythmias in Provocation of ventricular tachycardia, preceded by excessive
CPVT.56 Implantation of an ICD is indicated in patients who QRS widening, has been documented on treadmill exercise
have survived cardiac arrest and who have ventricular arrhyth- testing.59 Guidelines recommend that the QRS duration should
mias despite maximally tolerated β-blockade. Finally, left cervical not prolong by more than 50% from baseline and that a QRS
sympathetic denervation should be considered for patients duration of <120 ms should be maintained.60 To ensure that the
in whom β-blockers and ICDs are insufficient.57 Patients QRS duration does not exceed these parameters at higher heart
with CPVT are advised against competitive exercise, high-stress rates, treadmill testing may be employed to determine the degree
occupations, and substances that would increase sympathetic of maximal QRS prolongation during exercise after the patient
tone. is fully loaded with flecainide or propafenone.
Antiarrhythmics and Exercise-Induced Conclusion

Ventricular Arrhythmias
Multiple mechanisms have been identified for exercise-induced
Class I sodium channel blocking antiarrhythmic drugs, such as arrhythmias. We have provided a brief overview of the various
flecainide and propafenone, can cause ventricular reentry through causes. Although most of the arrhythmias encountered during
a use-dependent mechanism. QRS duration increases progres- exercise are benign, familiarity with the differential diagnosis
sively with increasing levels of exercise while taking flecainide.58 allows an assessment of prognosis and appropriate treatment.
A QRS increase of approximately 15% to 20% is consistent with Further pathophysiological and clinical details for each topic are
pharmacologic effect; however further increases in QRS duration explored in the other chapters of this book.
Incidence and clinical importance. J Am Coll 30. Aiba T, Suyama K, Aihara N, et al: The role of
References Cardiol 37:593–597, 2001. Purkinje and pre-Purkinje potentials in the reen-
16. Yamanaka O, Hobbs RE: Coronary artery anoma- trant circuit of verapamil-sensitive idiopathic LV
1. Beckerman J, Mathur A, Stahr S, et al: Exercise- lies in 126,595 patients undergoing coronary arte- tachycardia. Pacing Clin Electrophysiol 24:333–
induced ventricular arrhythmias and cardiovascular riography. Cathet Cardiovasc Diagn 21:28–40, 344, 2001.
death. Ann Noninvasive Electrocardiol 10:47–52, 1990. 31. Zipes D, Jalife J: Cardiac Electrophysiology:
2005. 17. Cheitlin MD, MacGregor J: Congenital anomalies From Cell to Bedside: Expert Consult—Online
2. Beckerman J, Wu T, Jones S, et al: Exercise test- of coronary arteries: Role in the pathogenesis of and Print, ed 5, Philadelphia, 2009, Saunders, p
induced arrhythmias. Prog Cardiovasc Dis 47:285– sudden cardiac death. Herz 34:268–279, 2009. 1184.
305, 2005. 18. Mofrad PS, Weigold G, Clavijo LC: Sudden 32. Iwai S, Lerman BB: Management of ventricular
3. Paterson DJ: Antiarrhythmic mechanisms during cardiac death in athlete with anomalous single tachycardia in patients with clinically normal
exercise. J Appl Physiol 80:1853–1862, 1996. coronary artery. Cardiovasc Revasc Med 6:89–90, hearts. Curr Cardiol Rep 2:515–521, 2000.
4. Maurer MS, Shefrin EA, Fleg JL: Prevalence and 2005. 33. Nakagawa H, Beckman KJ, McClelland JH, et al:
prognostic significance of exercise-induced supra- 19. Choi JH, Kornblum RN: Pete Maravich’s incred- Radiofrequency catheter ablation of idiopathic left
ventricular tachycardia in apparently healthy vol- ible heart. J Forensic Sci 35:981–986, 1990. ventricular tachycardia guided by a Purkinje poten-
unteers. Am J Cardiol 75:788–792, 1995. 20. Iwai S, Cantillon DJ, Kim RJ, et al: Right and left tial. Circulation 88:2607–2617, 1993.
5. Abdulla J, Nielsen JR: Is the risk of atrial fibrilla- ventricular outflow tract tachycardias: Evidence for 34. Maron BJ, Ackerman MJ, Nishimura RA, et al:
tion higher in athletes than in the general popula- a common electrophysiologic mechanism. J Car- Task Force 4: HCM and other cardiomyopathies,
tion? A systematic review and meta-analysis. diovasc Electrophysiol 17:1052–1058, 2006. mitral valve prolapse, myocarditis, and Marfan
Europace 11:1156–1159, 2009. 21. Lerman BB: Response of nonreentrant syndrome. J Am Coll Cardiol 45:1340–1345,
6. Turagam MK, Velagapudi P, Kocheril AG: Atrial catecholamine-mediated ventricular tachycardia to 2005.
fibrillation in athletes. Am J Cardiol 109:296–302, endogenous adenosine and acetylcholine: Evidence 35. Marcus FI, McKenna WJ, Sherrill D, et al: Diag-
2012. for myocardial receptor-mediated effects. Circula- nosis of arrhythmogenic right ventricular
7. Pelliccia A, Maron BJ, Di Paolo FM, et al: Preva- tion 87:382–390, 1993. cardiomyopathy/dysplasia (ARVC/D). Circulation
lence and clinical significance of left atrial remodel- 22. Kim RJ, Iwai S, Markowitz SM, et al: Clinical and 121:1533–1541, 2010.
ing in competitive athletes. J Am Coll Cardiol electrophysiological spectrum of idiopathic ven- 36. Gaita F, Giustetto C, Di Donna P, et al: Long-term
46:690–696, 2005. tricular outflow tract arrhythmias. J Am Coll follow-up of right ventricular monomorphic extra-
8. Furlanello F, Bertoldi A, Dallago M, et al: Atrial Cardiol 49:2035–2043, 2007. systoles. J Am Coll Cardiol 38:364–370, 2001.
fibrillation in elite athletes. J Cardiovasc Electro- 23. Lerman BB, Stein KM, Markowitz SM, et al: Ven- 37. Hoffmayer KS, Machado ON, Marcus GM, et al:
physiol 9(8 Suppl):S63–S68, 1998. tricular arrhythmias in normal hearts. Cardiol Clin Electrocardiographic comparison of ventricular
9. Bunch TJ, Chandrasekaran K, Gersh BJ, et al: The 18:265–291, vii, 2000. arrhythmias in patients with arrhythmogenic right
prognostic significance of exercise-induced atrial 24. Josephson ME: Clinical Cardiac Electrophysiol- ventricular cardiomyopathy and right ventricular
arrhythmias. J Am Coll Cardiol 43:1236–1240, ogy: Techniques and Interpretations, ed 4, Balti- outflow tract tachycardia. J Am Coll Cardiol
2004. more, MD, 2008, Lippincott Williams & Wilkins, 58:831–838, 2011.
10. Froelicher VF, Allen M, Lancaster MC: Maximal p 912. 38. Zipes DP, Camm AJ, Borggrefe M, et al: ACC/
treadmill testing of normal USAF aircrewmen. 25. Blanck Z, Dhala A, Deshpande S, et al: Bundle AHA/ESC 2006 guidelines for management of
Aerosp Med 45:310–315, 1974. branch reentrant ventricular tachycardia: Cumula- patients with ventricular arrhythmias and the pre-
11. Fleg JL, Lakatta EG: Prevalence and prognosis of tive experience in 48 patients. J Cardiovasc Elec- vention of sudden cardiac death. J Am Coll Cardiol
exercise-induced nonsustained ventricular tachy- trophysiol 4:253–262, 1993. 48:e247–e346, 2006.
cardia in apparently healthy volunteers. Am J 26. Bardy GH, Fedor JM, German LD, et al: Surface 39. Cha Y-M, Gersh BJ, Maron BJ, et al: Electrophysi-
Cardiol 54:762–764, 1984. electrocardiographic clues suggesting presence of a ologic manifestations of ventricular tachyarrhyth-
12. Partington S, Myers J, Cho S, et al: Prevalence and nodofascicular Mahaim fiber. J Am Coll Cardiol mias provoking appropriate defibrillator
prognostic value of exercise-induced ventricular 3:1161–1168, 1984. interventions in high-risk patients with hypertro-
arrhythmias. Am Heart J 145:139–146, 2003. 27. Hayashi H, Fujiki A, Tani M, et al: Circadian varia- phic cardiomyopathy. J Cardiovasc Electrophysiol
13. Corrado D, Basso C, Rizzoli G, et al: Does sports tion of idiopathic ventricular tachycardia originat- 18:483–487, 2007.
activity enhance the risk of sudden death in adoles- ing from right ventricular outflow tract. Am J 40. Drinko JK, Nash PJ, Lever HM, et al: Safety of
cents and young adults? J Am Coll Cardiol Cardiol 84:99–101, A8, 1999. stress testing in patients with hypertrophic cardio-
42:1959–1963, 2003. 28. Issa ZF, Miller JM, Zipes DP: Clinical arrhythmol- myopathy. Am J Cardiol 93:1443–1444, A12, 2004.
14. Angelini P, editor: Coronary Artery Anomalies: A ogy and electrophysiology, ed 1, Philadelphia, 41. Gimeno JR, Tomé-Esteban M, Lofiego C, et al:
Comprehensive Approach, Baltimore, MD, 1999, 2009, Saunders. Exercise-induced ventricular arrhythmias and risk
Lippincott Williams & Wilkins, p 200. 29. Lerman BB, Stein KM, Markowitz SM: Mecha- of sudden cardiac death in patients with hypertro-
15. Davis JA, Cecchin F, Jones TK, et al: Major coro- nisms of idiopathic left ventricular tachycardia. phic cardiomyopathy. Eur Heart J 30:2599–2605,
nary artery anomalies in a pediatric population: J Cardiovasc Electrophysiol 8:571–583, 1997. 2009.
42. Gersh BJ, Maron BJ, Bonow RO, et al: 2011 exercise stress test in long QT syndrome patients optimal therapeutic strategies to prevent sudden
ACCF/AHA guideline for the diagnosis and treat- with KvLQT1 and HERG potassium channel death. Heart 89:66–70, 2003.
ment of hypertrophic cardiomyopathy: a report of
the American College of Cardiology Foundation/
American Heart Association Task Force on Prac-
48.
defects. J Am Coll Cardiol 34:823–829, 1999.
Faber GM, Rudy Y: Action potential and contrac-
tility changes in [Na(+)](i) overloaded cardiac myo-
55. Brugada J, Brugada P: Clinical Approach to
Sudden Cardiac Death Syndromes, ed 1, New
York, 2010, Springer, p 356.
63
tice Guidelines. Developed in collaboration with cytes: A simulation study. Biophys J 78:2392–2404, 56. van der Werf C, Kannankeril PJ, Sacher F, et al:
American Association for Thoracic Surgery, Amer- 2000. Flecainide therapy reduces exercise-induced ven-
ican Society of Echocardiography, American 49. Vyas H, Hejlik J, Ackerman MJ: Epinephrine QT tricular arrhythmias in patients with catecholamin-
Society of Nuclear Cardiology, Heart Failure stress testing in the evaluation of congenital ergic polymorphic ventricular tachycardia. J Am
Society of America, Heart Rhythm Society, Society long-QT syndrome: Diagnostic accuracy of the Coll Cardiol 57:2244–2254, 2011.
for Cardiovascular Angiography and Interventions, paradoxical QT response. Circulation 113:1385– 57. Wilde AAM, Bhuiyan ZA, Crotti L, et al: Left
and Society of Thoracic Surgeons. J Am Coll 1392, 2006. cardiac sympathetic denervation for catecholamin-
Cardiol 58:e212–e260, 2011. 50. Horner JM, Horner MM, Ackerman MJ: The ergic polymorphic ventricular tachycardia. N Engl
43. McLeod CJ, Ommen SR, Ackerman MJ, et al: Sur- diagnostic utility of recovery phase QTc during J Med 358:2024–2029, 2008.
gical septal myectomy decreases the risk for appro- treadmill exercise stress testing in the evaluation of 58. Ranger S, Talajic M, Lemery R, et al: Amplification
priate implantable cardioverter defibrillator long QT syndrome. Heart Rhythm 8:1698–1704, of flecainide-induced ventricular conduction
discharge in obstructive hypertrophic cardiomy- 2011. slowing by exercise: A potentially significant clini-
opathy. Eur Heart J 28:2583–2588, 2007. 51. Mohamed U, Napolitano C, Priori SG: Molecular cal consequence of use-dependent sodium channel
44. Schwartz PJ, Priori SG, Spazzolini C, et al: and electrophysiological bases of catecholaminer- blockade. Circulation 79:1000–1006, 1989.
Genotype-phenotype correlation in the long-QT gic polymorphic ventricular tachycardia. J Cardio- 59. Falk RH: Flecainide-induced ventricular tachycar-
syndrome: Gene-specific triggers for life- vasc Electrophysiol 18:791–797, 2007. dia and fibrillation in patients treated for atrial
threatening arrhythmias. Circulation 103:89–95, 52. Laitinen PJ, Brown KM, Piippo K, et al: Mutations fibrillation. Ann Intern Med 111:107–111, 1989.
2001. of the cardiac ryanodine receptor (RyR2) gene in 60. ACC/AHA/ESC: 2006 guidelines for the manage-
45. Tan HL, Bardai A, Shimizu W, et al: Genotype- familial polymorphic ventricular tachycardia. Cir- ment of patients with atrial fibrillation: A report of
specific onset of arrhythmias in congenital culation 103:485–490, 2001. the American College of Cardiology/American
long-QT syndrome: Possible therapy implications. 53. Liu N, Colombi B, Memmi M, et al: Arrhythmo- Heart Association Task Force on Practice Guide-
Circulation 114:2096–2103, 2006. genesis in catecholaminergic polymorphic ven- lines and the European Society of Cardiology
46. Marban E, Robinson SW, Wier WG: Mechanisms tricular tachycardia: insights from a RyR2 R4496C Committee for Practice Guidelines (writing com-
of arrhythmogenic delayed and early afterdepolar- knock-in mouse model. Circ Res 99:292–298, mittee to revise the 2001 guidelines for the man-
izations in ferret ventricular muscle. J Clin Invest 2006. agement of patients with atrial fibrillation).
78:1185–1192, 1986. 54. Sumitomo N, Harada K, Nagashima M, et al: Cat- Developed in collaboration with the European
47. Swan H, Viitasalo M, Piippo K, et al: Sinus node echolaminergic polymorphic ventricular tachycar- Heart Rhythm Association and the Heart Rhythm
function and ventricular repolarization during dia: Electrocardiographic characteristics and Society. Circulation 114:e257–e354, 2006.
Cardiac Monitoring: Short-
and Long-Term Recording 64
Andrew D. Krahn, Raymond Yee, Allan C. Skanes, and George J. Klein
data are transformed into a digital format and are analyzed with
CHAPTER OUTLINE
interpretive software, with technologist and physician editing and
Short-Term Recording 621 reporting. Additional markers for patient-activated events and
time correlates are included to allow greater diagnostic accuracy.
Intermediate-Duration Monitoring 623
Continuous electrocardiographic monitoring is possible for 24 h
Prolonged Monitoring 623 to a maximum of 72 h with conventional Holter monitors. Novel
extended-duration continuous recording technologies will be dis-
Future Directions 626
cussed subsequently. Complete rhythm capture allows documen-
Conclusion 626 tation of rhythm for symptomatic and asymptomatic events.
Holter monitoring is useful if the clinical history is suggestive of
an arrhythmic origin and the symptoms are frequent enough to
Ambulatory cardiac monitoring to detect arrhythmia became be detected within the period of monitoring. Holter monitoring
practical with the development of Holter monitoring and its may also yield a rhythm profile to provide evidence of sinus node
subsequent derivatives. The clinician is currently armed with an dysfunction or ambient arrhythmias that are potentially signifi-
array of tools to provide progressively longer durations of cant and to assess rate control of atrial fibrillation.
electrocardiographic (ECG) monitoring, to obtain a rhythm Holter monitoring has several drawbacks. Patients may not
profile potentially useful in risk stratification, and to establish a experience symptoms or cardiac arrhythmias during the usual
symptom rhythm correlation in patients with infrequent symp- Holter recording periods. In patients with syncope, the likeli-
toms (Figure 64-1). hood of another syncopal episode occurring during the monitor-
Clinical trials using implantable loop recorders (ILRs) have ing period is the major limiting factor. Presyncope is a more
validated a clear role in unexplained syncope and emerging utility common event during ambulatory monitoring but is less likely
in a range of conditions including atrial fibrillation, atypical epi- to be associated with an arrhythmia.1,2 The ubiquity of presyn-
lepsy, vasovagal syncope, and post–myocardial infarction arrhyth- cope as a symptom in the community makes its usefulness as a
mia detection. Advances in design with the potential addition of surrogate for syncope relatively uncertain. The physical size of
other physiological sensors have strengthened the capability of the device may hinder the ability of patients to sleep comfortably
implantable devices, and promise to translate into tools that not or to engage in activities that precipitate symptoms. Patients are
only detect intermittent arrhythmia in syncope, but assist in further inconvenienced because the devices have to be removed
chronic disease management of common conditions such as atrial while showering. Observations on Holter monitoring must be
fibrillation and post–myocardial infarction risk stratification. correlated with clinical context in the absence of symptoms. Sig-
Syncope is the prototype condition that is ideally served by nificant variability is often noted in patient documentation of
long-term ambulatory monitoring. The periodic and unpredict- activated events such that accurate symptom–rhythm correlation
able nature of events and the high spontaneous remission rate are is undermined.
the major obstacles to diagnosis in most patients. Other forms of It is not surprising that Holter monitoring has a low diagnos-
testing in unexplained syncope usually provide a context for tic yield. In several large series using 12 h or more of ambulatory
bedside formulation of a differential diagnosis and prognosis but monitoring for investigation of syncope, only 4% had recurrence
rarely provide a specific diagnosis. Classic “provocative” testing of symptoms during monitoring.3,4 The overall diagnostic yield
with tilt and electrophysiological testing may be negative or may of ambulatory or Holter monitoring was 19%. These studies
yield abnormalities of unknown significance, which is reflected reported symptoms that were not associated with arrhythmias in
in the poor predictive value of both tests. These obvious limita- 15% of cases. The causal relationship between arrhythmia and
tions turn our attention to long-term monitoring as a step toward syncope was frequently uncertain. Uncommon asymptomatic
the gold standard of comprehensive physiological assessment arrhythmias such as prolonged sinus pauses, atrioventricular
during spontaneous symptoms. This chapter discusses the techni- block (such as Mobitz type II block), and nonsustained ventricu-
cal aspects and established utility of monitoring devices with a lar tachycardia can provide important contributions to the diag-
focus on syncope, and explores the current and emerging use of nosis, instigating further investigations to rule out structural
monitoring technologies. heart disease and other precipitating factors. Although these
observations necessitate prompt attention, it is important to
interpret the results in the clinical context of the syncopal pre-
sentation and to not unduly exclude common causes of syncope
Short-Term Recording such as neurocardiogenic syncope.
It is also important to understand that a normal Holter
Holter Monitoring monitor does not exclude an arrhythmic cause for syncope. In
fact, an arrhythmic cause is typically the case. If the pretest prob-
The Holter monitor is a portable battery-operated device that ability is high for an arrhythmic cause, further investigations such
connects to the patient with bipolar electrodes and provides as more prolonged monitoring or cardiac electrophysiological
recordings from up to 12 electrocardiographic leads. Data are studies are required. In a study that evaluated extension of ambu-
stored in the device using analog or digital storage media. The latory Holter monitoring duration to 72 h,5 an increased number
621
Figure 64-1. Ambulatory recording during palpitations and presyncope. Note the burst of atrial tachycardia, with associated brief pauses.
of asymptomatic arrhythmias were detected, but the overall diag-

nostic yield was not increased.
In our institution, we typically use Holter monitoring for
24 h. It is a noninvasive test that provides information to establish
a rhythm profile in patients and the diagnosis in those with fre-
quent symptoms. The more frequent the symptoms, the higher
the diagnostic yield. The apparent modest yield of Holter moni-
toring presumably reflects primary care use of the device in
patients with frequent symptoms, facilitating a symptom–rhythm
correlation. This leads to selection bias in the referral population,
leading to an apparent futility in referred patients who, by defini-
tion, have failed short-term monitoring.
Nonelectrode Event Recorders

Transtelephonic monitors are a form of noncontinuous ambula-
tory recording that is convenient for patient use. During symp-
tomatic episodes, the patient activates the device, which then
records electrocardiographic signals. The recorded event must be
directly transmitted by an analog telephone line to a receiving Figure 64-2. Photograph of the external loop recorder with recording electrodes
center (Figures 64-2 and 64-3). The received signal is then con- (right).
verted to an analog recording that is displayed or printed as a
single lead rhythm strip. The device has solid-state memory electrodes and battery systems, enabling a patch-based “all in
capacity, allowing recording and storage of electrocardiographic one” system to facilitate 7 to 14 d of monitoring (Figure 64-4).
signals during symptoms. Electrocardiographic signals are col- These technologies are currently being evaluated with promising
lected prospectively for 1 to 2 minutes upon patient activation. early results, but larger-scale or comparative studies have not
The major disadvantages of such devices include the need for been performed. Two prospective comparative trials are under
patient activation; missing asymptomatic arrhythmias, which way (Holter vs. patch), as is a third trial, which was undertaken
requiring that the symptoms persist long enough for the device to validate detection of asymptomatic atrial fibrillation. Publica-
to record the event; and the inability to record events that sur- tions are anticipated by the time of the release of this chapter.
round the onset of symptoms.
Wrist Recorders
Wrist and mobile phone-based recording devices show promise
New Technology as minimally invasive recording devices, with rapidly evolving
technology. A recent preliminary report indicated the potential
Patch pulse detection capability of a wrist recording device termed the
Emerging technologies such as the iRhythm device promise wriskwatch,6 which brought attention to alternate means of
minimally invasive intermediate-term monitoring without classic recording multiple physiological parameters.
Cardiac Monitoring: Short- and Long-Term Recording 623
0:01:37.54 0:01:45.66
Intermediate-Duration Monitoring
64
205.4 BPM Rate 205.4 BPM Rate
Extended Holter
Lead_CH1 Continuous ambulatory monitoring with data transmission to a
central monitoring station staffed by Health Professionals (Car-
0:01:45.66 0:01:53.79
dionet, San Diego California) has emerged in the United States
205.4 BPM Rate
as a useful resource to extend traditional Holter monitoring
beyond 48 h. This technique is typically used for 7 to 14 d, and
has shown incremental benefit over a standard external loop
recorder in diagnosing or excluding arrhythmia.7 This approach
Lead_CH1
also provides the added layer of monitoring center involvement,
0:03:31.33 0:03:39.46 which enhances responsiveness to changes seen during monitor-
139.5 BPM Rate 143.8 BPM Rate ing. Unfortunately, this substantially increases cost.
Long-term compliance can be challenging even with these
devices because of electrode and skin-related problems and
waning of patient motivation in the absence of recurrent
Lead_CH1 symptoms.
0:03:55.71 0:04:03.84
—BPM Rate Manual 132.0 BPM Rate
activation External Loop Recorders
An external loop recorder continuously records and stores a
single external modified limb lead electrogram with a 4- to
Lead_CH1
18-min memory buffer (see Figure 64-2). After the onset of
Figure 64-3. External loop recorder download from a patient with near syncope spontaneous symptoms, the patient activates the device storing
with palpitations. Wide QRS complex tachycardia is noted at 206 bpm, which sub- the previous 3 to 14 min and the following 1 to 4 min of recorded
sequently demonstrates underlying atrial flutter with much slower conduction, and information. The captured rhythm strip subsequently can be
subsequent patient activation to capture the event. No recognized intervention uploaded and analyzed, often providing critical information
led to an abrupt reduction in conduction rate. regarding onset and termination of the arrhythmia (see Figure
64-2). This system theoretically can be used indefinitely, but in
practice, use is limited to a few weeks in most individuals because
of its limitations. The recording device is connected to skin
electrodes on the patient’s chest wall that need to be removed for
bathing or showering, and can be uncomfortable during sleep.
A randomized trial has shown diagnostic and cost-effectiveness
superiority to Holter monitors (22% symptom–rhythm correla-
tion yield for Holter monitoring vs. 56% for the loop recorder;
P < .01).8 An increment in diagnostic yield is noted when auto-
matic activation is added to patient activation.9 Automatic detec-
tion includes bradycardia, tachycardia, and pauses, as well as atrial
fibrillation based on algorithms that detect RR irregularity. Data
on the utility of this technology in detecting or excluding atrial
fibrillation are limited.
Vest Technology
Multiple vendors now offer wearable monitoring systems that
have the potential to acquire and transmit multiple physiological
parameters including ECG. These include vest and shirt tech-
nologies using integrated materials for capture of ECG and respi-
ratory parameters.10 Much of this technology has not completed
the regulatory process and has not emerged in day-to-day clinical
care but promises to revolutionize monitoring technologies. The
interested reader is directed to an in-depth review of emerging
technologies prepared by Pantelopoulos et al.11
Figure 64-4. Sample patch-based recording system that allows both acquisition Prolonged Monitoring
and storage of a single-lead electrocardiogram (ECG) for 7 to 14 d. See text for
discussion.
Implantable Cardiac Monitors
The implantable loop recorder (ILR) permits prolonged moni-
toring without external electrodes. It is ideally suited to patients
who require more prolonged monitoring such as those with
infrequent recurrent symptoms such as syncope. Similar to the manufacturer platform. Devices store approximately 40 min of
external loop recorder, it is designed to detect arrhythmia and to ECG in a range of configurable memory buffers that include
specifically correlate symptoms with recorded cardiac rhythms. symptomatic and asymptomatic recordings. Data are retrieved by
The implanted device obviates surface electrodes and accompa- interrogation with a standard pacemaker programmer, and the
nying compliance issues. Medtronic system can transmit data with remote monitoring
Implanted loop recorders are manufactured by Medtronic based on the Medtronic Carelink Network (Medtronic, Minne-
(Reveal XT Model 9529, DX Model 9538, Minneapolis, Min- apolis, Minnesota).
nesota) and by St. Jude Medical (Confirm Model DM2100, Little
Canada, Minnesota; Figure 64-5). Both are smaller than a con-
ventional pacemaker generator, record a single-lead ECG without Clinical Studies
a transvenous lead, and have the potential to deliver 3 years of
battery life. The device is typically inserted into the left chest Several studies have demonstrated the feasibility of the ILR in
using local anesthetic, usually in a high left parasternal or medial establishing a symptom–rhythm correlation during long-term
pacemaker insertion location.12 Recent evidence suggests that a monitoring in patients with syncope.14-19 Pooled data from nine
simplified implant procedure halfway between the sternal notch smaller studies summarized in the European Society of Cardiol-
and the left breast area on an oblique line consistently results in ogy (ESC) Guidelines included 506 patients with unexplained
an adequate signal.13 This observation has not been extensively syncope at the end of a complete conventional investigation. A
validated but certainly simplifies the potential need to map for symptom–rhythm correlation was found in 176 patients (35%);
an ideal location. In principle, devices can be implanted anywhere of these, 56% had asystole (or bradycardia in a few cases) at the
in the chest and abdomen where an ECG can be recorded. The time of the recorded event, 11% had tachycardia, and 33% had
implant procedure is similar to fashioning a small pacemaker no arrhythmia.20
pocket. An adequate signal can be obtained anywhere in the left A recent study of a “real-world” experience included 570
thorax, without the need for cutaneous mapping. Although patients monitored for diagnosis of syncope.21 Significant physi-
mapping is advocated, it is seldom performed, and an adequate cal trauma had been experienced in association with a syncopal
signal to assess RR intervals is generally obtained. Devices have episode by 36% of patients. Average follow-up time after ILR
been implanted in the right parasternal location to optimize P implant was 10 ± 6months. Syncope recurred in 19%, 26%, and
waves, and in an inframammary or anterior axillary location for 36% after 3, 6, and 12 months, respectively. Of 218 events during
comfort or cosmetic purposes. These alternate sites typically follow-up, ILR-guided diagnosis was obtained in 170 cases
record a lower-amplitude ECG signal. The patient along with a (78%), of which 128 (75%) were cardiac. The median number of
spouse, family member, or friend is instructed in use of the activa- tests performed per patient in the total study population was 13,
tor at the time of implant. Sterile implant technique is essential. and patients saw an average of 3 consultants before device
Prophylactic antibiotics are generally recommended, although implant. This speaks to the immense and misguided industry of
efficacy in preventing infection has not been rigorously evaluation of syncope before evaluation is completed by a syncope
established. expert. A structured history followed by targeted investigations
Devices have the ability to automatically detect high and low including early use of progressive monitoring technologies will
heart rate and pause events, with irregularity algorithms demon- lead to a diagnosis in most patients, without the use of very low-
strating reasonable ability to detect atrial fibrillation. Manual yield tests such carotid Doppler or brain imaging.22
activation remains possible if the patient experiences symptoms— Several studies have demonstrated the usefulness of prolonged
typically syncope, presyncope, or palpitations. The recorded cardiac monitoring in select populations. The ISSUE investiga-
bipolar ECG signal is stored in a loop buffer, which stores the tors (International Study on Syncope of Uncertain Etiology)
recorded ECG in several programmable memory bin configura- assessed 111 patients with presumed vasovagal syncope who
tions. Data storage and retrieval are influenced by the underwent tilt testing and loop recorder implant, regardless of
tilt result.23 Syncope recurred in 34% of patients in both tilt-
positive and tilt-negative groups, with marked bradycardia or
asystole the most common recorded arrhythmia during follow-up
(46% and 62%, respectively). Tilt testing failed to predict recur-
rence of syncope. The ISSUE2 study performed an ILR implant
in 443 patients with recurrent syncope in the absence of struc-
tural heart disease, who were presumed to have vasovagal
syncope.14 Syncope recurred in 143 patients, with a symptom–
rhythm correlation in 102 patients. In those 102 patients, loop
recorder–directed therapy (predominantly pacing for bradycar-
dia) was associated with a lower risk of syncope recurrence (10%
vs. 41%; P = .0005). This was not a randomized assessment of
the benefit of pacing, and it led to the conduct of the ISSUE3
study, wherein 511 patients over age 40 with recurrent vasovagal
syncope underwent loop recorder implant. In the 17% with asys-
tolic recurrence, dual-chamber pacemaker therapy was superior
to the “pacemaker off” approach in prevention of recurrent
syncope (57% reduction; 95% confidence interval [CI], 4% to
81%).24 This suggests that prolonged monitoring with an ILR is
useful in patients with frequent vasovagal syncope who are over
age 40, if pacemaker therapy is contemplated. This also suggests
a diminishing role for tilt testing, which has limited correlation
with spontaneous episodes recorded during extended
Figure 64-5. Implanted loop recorders. The Medtronic Reveal XT (bottom left) with monitoring.
patient activator (top left). The St Jude Confirm (bottom center) with patient activator Loop recorders have also shown utility in patients
(bottom right) with base station in the background. See text for discussion. with syncope and bundle branch block with negative
electrophysiological testing.25 Negative electrophysiological or atrial fibrillation may have different implications in patients in
64
testing does not exclude a diagnosis of intermittent complete accordance with their index indication for monitoring. The
atrioventricular (AV) block, and prolonged monitoring or con- ISSUE investigators proposed a useful classification system for
sideration of permanent pacing is reasonable in this population.26 symptomatic events in loop recorders applied to patients with
Syncope in the absence of marked reduction in left ventricular syncope,30 which is summarized in Table 64-1. This is often
function provides grounds for consideration of an ILR if an helpful in assigning a probable mechanism of syncope.
implantable cardioverter-defibrillator (ICD) is not indicated and
noninvasive testing is inconclusive.27
Two prospective randomized trials compared early use of the Additional Uses of Extended
loop recorder for prolonged monitoring with conventional Monitoring Technologies
testing in patients without significant structural heart disease
undergoing a cardiac workup for unexplained syncope.15,28 Con- Long-term cardiac monitoring is best suited to patients with
ventional testing involved external loop recorders and tilt and infrequent intermittent symptoms possibly related to arrhythmia,
electrophysiological testing. Both studies showed a higher diag-
nostic yield when an ILR was used. Overall, prolonged monitor-
ing was more likely than conventional testing to result in a
diagnosis, with a symptom–rhythm correlation obtained in 55% 10:43:57
compared with a 19% diagnostic yield in RAST (the Randomized
Assessment of Syncope Trial; P = .0014).28 Both studies also
showed that up-front use of an implantable strategy was appar- 10:44:11
ently cost-effective compared with conventional testing because
of the dramatic improvement in diagnostic yield.15,29 These data A
highlight the limitations of conventional diagnostic techniques.
In patients with infrequent syncope, the ILR is the diagnostic 10:44:25
tool of choice, especially when noninvasive testing is negative and
an arrhythmia is suspected.
10:44:39
Event Classification Figure 64-6. Rhythm strip from a Medtronic implanted loop recorder from a
77-year-old man with recurrent unexplained syncope and a normal baseline elec-
Interrogation of ILRs after recurrence of syncope often demon- trocardiogram (ECG). Note that a premature ventricular contraction (PVC) induces
strates rhythm findings that require clinical correlation and con- persistent complete atrioventricular (AV) block without an escape QRS; this is fol-
siderable judgment to interpret (Figure 64-6). Nocturnal pauses lowed by sinus slowing, suggesting a secondary cardioinhibitory vagal reaction.
Table 64-1. ISSUE Classification of Detected Rhythm from the ILR
Classification Sinus Rate AV Node Comment
Asystole (RR >3 s)
1A Arrest Normal Progressive sinus bradycardia until sinus arrest, probably

vasovagal
1B Bradycardia AV block AV block with associated sinus bradycardia, probably vasovagal
1C Normal or tachycardia AV block Abrupt AV block without sinus slowing suggests intrinsic AV
conduction disease
Bradycardia
2A Decrease >30% Normal Probably vasovagal

2B HR <40 for >10 s Normal Probably vasovagal
Minimal HR Change
3A <10% variation Normal Suggests noncardiac cause: unlikely vasovagal

3B HR increase or decrease 10%-30%, Normal Suggests vasovagal
not <40 or >120 bpm
Tachycardia
4A Progressive tachycardia Normal Sinus acceleration suggests orthostatic intolerance or noncardiac

cause
4B N/A Normal Atrial fibrillation
4C N/A Normal Supraventricular tachycardia
4D N/A Normal Ventricular tachycardia
AV, Atrioventricular; HR, heart rate; ILR, implantable loop recorder; N/A, not applicable.
Adapted from Reference 17.
or to those in whom an infrequent “silent” arrhythmia is sought. other tests incremental to a thoughtful clinical evaluation. As
Although syncope is a logical first application of this technology, stated in the ESC guidelines and in the Canadian Position State-
many other clinical disease states stand to benefit from the knowl- ment on Syncope, patients with unexplained syncope with a high
edge gained from prolonged monitoring. This includes atypical likelihood of recurrence of syncope within 3 years who are at low
epilepsy, where seizures may serve as evidence of recurrent cere- risk of sudden death should be considered for an implanted
bral hypoperfusion in conjunction with syncope, mistaken as a monitor.20,34 An ILR should also be considered in high-risk
primary neurologic event.31 patients if thorough evaluation has not demonstrated a cause of
Additional potential uses of the ILR relate to the automatic syncope (supine or exertional syncope, family history of sudden
detection feature of the device in patients for risk stratification death, preexcited QRS, or repolarization abnormality suggesting
after myocardial infarction. The CARISMA study involved an inherited syndrome with risk of sudden death).
implanted ILRs in 297 patients and followed them for 1.9 ± 0.5
years.32 Predefined arrhythmias were recorded in 137 patients
(46%), 86% of whom were asymptomatic. The ILR documented
a 28% incidence of new-onset atrial fibrillation, a 13% incidence Future Directions
of nonsustained ventricular tachycardia (≥16 beats), a 10% inci-
dence of high-degree atrioventricular block, a 12% incidence of Recent advances in loop recorder design and the emergence of
significant sinus node disease, a 3% incidence of sustained ven- multiple manufacturers are indicative of rekindled interest in
tricular tachycardia, and a 3% incidence of ventricular fibrilla- long-term physiological monitoring for a variety of disease states.
tion. Regression analysis showed that high-degree atrioventricular Enhanced data capture is likely to provide a much larger sample
block was the most powerful predictor of cardiac death. The of data for analysis, revealing potentially contributing novel
implications of this study are not completely understood, and the insights into areas such as risk stratification and arrhythmia
ability to alter outcomes remains to be established. It certainly burden assessment. Improved on-board and off-line diagnostic
demonstrates the incremental detection ability of the ILR in this capabilities may provide an early warning or patient alert mecha-
population, suggesting that historic short-term studies have nism for a range of events, including recurrence of atrial fibrilla-
grossly underestimated the incidence of potentially relevant tion or nonsustained ventricular arrhythmia. The notion of
arrhythmias. Active research in this realm is ongoing. detection of catastrophic arrhythmia with a device integrated
Finally, strong interest in silent atrial fibrillation as a cause of with an emergency response system has been raised. Of greater
stroke has led to the initiation of several studies to determine the interest is the potential to integrate a broader range of physio-
presence of asymptomatic atrial fibrillation. Multiple small logical sensors into long-term monitoring devices, including but
studies have evaluated the usefulness of external monitoring tech- not limited to blood pressure, oxygen saturation, chest imped-
nologies after ablation, and several larger studies with implanted ance, and left atrial pressure. Finally, there is a clear prospect of
devices measuring both burden of disease and efficacy of an miniaturization, which would greatly simplify the implant proce-
antiarrhythmic intervention are expected to report imminently.33 dure, particularly if coupled with elimination of reliance on
Although these studies have reported a higher rate of silent atrial highly specialized programmers.
arrhythmias than was previously suspected, the implications of
these findings are speculative. A potential use of this technology
may involve detecting atrial fibrillation in patients who have
discontinued otherwise indicated antithrombotic therapy, in cases Conclusion
where the presence of atrial fibrillation would prompt resump-
tion of therapy. This is an emerging focus of study, particularly Progressively extended monitoring technologies have signifi-
in the context of the new generation of antithrombotic therapies cantly facilitated the objective of obtaining physiological data
with rapid therapeutic onset and offset, in conjunction with the during spontaneous symptoms in patients with unexplained
prospect of smaller, easier to implant monitoring devices. syncope. Long-term monitoring with the ILR has emerged as the
test of choice in patients with problematic syncope and preserved
left ventricular function. Prolonged monitoring with external and
Indications for Prolonged Monitoring implantable loop recorders has significantly enhanced our ability
to diagnose intermittent arrhythmias in a variety of clinical set-
Loop recorders are suited to earlier implementation in the diag- tings. Ongoing clinical trials will undoubtedly expand the use of
nostic cascade of syncope, in part because of the low yield of prolonged monitoring to other disease states.
24 hours enough? Arch Intern Med 150:1073– 10. Pandian PS, Mohanavelu K, Safeer KP, et al: Smart
References 1078, 1990. Vest: Wearable multi-parameter remote physiolog-
6. Rickard J, Ahmed S, Baruch M, et al: Utility of a ical monitoring system. Med Eng Phys 30:466–
1. Olson JA, Fouts AM, Padanilam BJ, et al: Utility novel watch-based pulse detection system to detect 477, 2008.
of mobile cardiac outpatient telemetry for the diag- pulselessness in human subjects. Heart Rhythm 11. Pantelopoulos A, Bourbakis N: A survey on wear-
nosis of palpitations, presyncope, syncope, and the 8:1895–1899, 2011. able biosensor systems for health monitoring. Conf
assessment of therapy efficacy. J Cardiovasc Elec- 7. Rothman SA, Laughlin JC, Seltzer J, et al: The Proc IEEE Eng Med Biol Soc 2008:4887–4890,
trophysiol 18:473–477, 2007. diagnosis of cardiac arrhythmias: A prospective 2008.
2. Krahn AD, Klein GJ, Yee R, et al: Predictive value multi-center randomized study comparing mobile 12. Krahn AD, Klein GJ, Yee R, et al: Maturation of
of presyncope in patients monitored for assessment cardiac outpatient telemetry versus standard loop the sensed electrogram amplitude over time in a
of syncope. Am Heart J 141:817–821, 2001. event monitoring. J Cardiovasc Electrophysiol new subcutaneous implantable loop recorder.
3. Linzer M, Yang EH, Estes NA, et al: Diagnosing 18:241–247, 2007. Pacing Clin Electrophysiol 20:1686–1690, 1997.
syncope. Part 1. Value of history, physical examina- 8. Sivakumaran S, Krahn AD, Klein GJ, et al: A pro- 13. Grubb BP, Welch M, Kanjwal K, et al: An
tion, and electrocardiography. Clinical Efficacy spective randomized comparison of loop recorders anatomic-based approach for the placement of
Assessment Project of the American College of versus Holter monitors in patients with syncope or implantable loop recorders. Pacing Clin Electro-
Physicians. Ann Intern Med 126:989–996, 1997. presyncope. Am J Med 115:1–5, 2003. physiol 33:1149–1152, 2010.
4. Subbiah R, Gula LJ, Klein GJ, et al: Syncope: 9. Reiffel JA, Schwarzberg R, Murry M: Comparison 14. Brignole M, Sutton R, Menozzi C, et al: Early
Review of monitoring modalities. Curr Cardiol of autotriggered memory loop recorders versus application of an implantable loop recorder allows
Rev 4:41–48, 2008. standard loop recorders versus 24-hour Holter effective specific therapy in patients with recurrent
5. Bass EB, Curtiss EI, Arena VC, et al: The duration monitors for arrhythmia detection. Am J Cardiol suspected neurally mediated syncope. Eur Heart J
of Holter monitoring in patients with syncope. Is 95:1055–1059, 2005. 27:1085–1092, 2006.
15. Farwell DJ, Freemantle N, Sulke N: The clinical adult patients hospitalized with syncope. Arch 29. Krahn AD, Klein GJ, Yee R, et al: Cost implica-
impact of implantable loop recorders in patients Intern Med 161:1889–1895, 2001. tions of testing strategy in patients with syncope:
with syncope. Eur Heart J 27:351–356, 2006.
16. Krahn AD, Klein GJ, Yee R, et al: Detection of
asymptomatic arrhythmias in unexplained syncope.
23. Moya A, Brignole M, Menozzi C, et al: Mechanism
of syncope in patients with isolated syncope and in
patients with tilt-positive syncope. Circulation
Randomized assessment of syncope trial. J Am Coll
Cardiol 42:495–501, 2003.
30. Brignole M, Moya A, Menozzi C, et al: Proposed
64
Am Heart J 148:326–332, 2004. 104:1261–1267, 2001. electrocardiographic classification of spontaneous
17. Farwell DJ, Freemantle N, Sulke AN: Use of 24. Brignole M, Menozzi C, Moya A, et al: Pacemaker syncope documented by an implantable loop
implantable loop recorders in the diagnosis and therapy in patients with neurally mediated syncope recorder. Europace 7:14–18, 2005.
management of syncope. Eur Heart J 25:1257– and documented asystole: Third International 31. Zaidi A, Clough P, Cooper P, et al: Misdiagnosis of
1263, 2004. Study on Syncope of Uncertain Etiology (ISSUE- epilepsy: Many seizure-like attacks have a cardio-
18. Krahn AD, Klein GJ, Fitzpatrick A, et al: Predict- 3): A randomized trial. Circulation 125:2566–2571, vascular cause. J Am Coll Cardiol 36:181–184,
ing the outcome of patients with unexplained 2012. 2000.
syncope undergoing prolonged monitoring. Pacing 25. Brignole M, Menozzi C, Moya A, et al: Mechanism 32. Bloch Thomsen PE, Jons C, Raatikainen MJ, et al:
Clin Electrophysiol 25:37–41, 2002. of syncope in patients with bundle branch block Long-term recording of cardiac arrhythmias with
19. Solano A, Menozzi C, Maggi R, et al: Incidence, and negative electrophysiological test. Circulation an implantable cardiac monitor in patients with
diagnostic yield and safety of the implantable loop- 104:2045–2050, 2001. reduced ejection fraction after acute myocardial
recorder to detect the mechanism of syncope in 26. Krahn AD, Morillo CA, Kus T, et al: Empiric pace- infarction: The Cardiac Arrhythmias and Risk
patients with and without structural heart disease. maker compared with a monitoring strategy in Stratification After Acute Myocardial Infarction
Eur Heart J 25:1116–1119, 2004. patients with syncope and bifascicular conduction (CARISMA) study. Circulation Sep 28 122:1258–
20. Moya A, Sutton R, Ammirati F, et al: Guidelines block—Rationale and design of the Syncope: Pacing 1264, 2010.
for the diagnosis and management of syncope or Recording in ThE Later Years (SPRITELY) 33. Arya A, Piorkowski C, Sommer P, et al: Clinical
(version 2009). Eur Heart J 30:2631–2671, study. Europace 14:1044–1048, 2012. implications of various follow up strategies after
2009. 27. Menozzi C, Brignole M, Garcia-Civera R, et al: catheter ablation of atrial fibrillation. Pacing Clin
21. Edvardsson N, Frykman V, van Mechelen R, et al: Mechanism of syncope in patients with heart Electrophysiol 30:458–462, 2007.
Use of an implantable loop recorder to increase the disease and negative electrophysiologic test. Circu- 34. Sheldon RS, Morillo CA, Krahn AD, et al: Stan-
diagnostic yield in unexplained syncope: Results lation 105:2741–2745, 2002. dardized approaches to the investigation of
from the PICTURE registry. Europace 13:262– 28. Krahn AD, Klein GJ, Yee R, et al: Randomized syncope: Canadian Cardiovascular Society position
269, 2011. assessment of syncope trial: conventional diagnos- paper. Can J Cardiol 27:246–253, 2011.
22. Pires LA, Ganji JR, Jarandila R, et al: Diagnostic tic testing versus a prolonged monitoring strategy.
patterns and temporal trends in the evaluation of Circulation 104:46–51, 2001.
Electrocardiographic
Imaging in Patients With
Acute Coronary Syndrome 65
Olle Pahlm and Galen S. Wagner
Clinical research has focused on testing these methods relative

CHAPTER OUTLINE to established cardiac imaging methods.9-11
Limitation in Use of Conventional Displays of the
12 Standard ECG Leads for Diagnosis of Acute
Coronary Syndromes 629 Limitation in Use of Conventional Displays
Challenges in Considering ECG Limb Lead of the 12 Standard ECG Leads for Diagnosis
Relationships 629 of Acute Coronary Syndromes
Reciprocal ECG Leads 630
STEMI diagnosis requires at least 0.1-mV ST elevation in at least
Challenges in Considering ECG Chest Lead two contiguous leads.1 Contiguous leads, of course, refers to the
Relationships 630 spatial locations of the leads in the body, rather than their posi-
tions on the ECG display, whether on paper or on screen.
ECG Imaging to Overcome These Challenges 630
However, the conventional way of displaying the 12-lead ECG,
2D Imaging Based on Vectorcardiography 630 based on the historical development of leads, presents a challenge
for identifying spatially contiguous limb leads. The standard
3D Imaging Using Mercator or Polar Displays 632
ECG display format provides five potential pairs of contiguous
Conclusions and Future Developments 633 chest leads (V1/V2, V2/V3, V3/V4, V4/V5, and V5/V6) but only
three potential pairs of contiguous limb leads (I/aVL, II/aVF, and
III/aVF). This difference occurs because only the six chest leads
are displayed in their orderly sequence, while the six limb leads
When a patient presents with symptoms suggesting an acute are displayed as two groups, each consisting of three leads. Leads
coronary syndrome, the standard 12-lead electrocardiogram I, II, and III are displayed as one group, and aVR, aVL, and aVF
(ECG) remains the essential diagnostic test. Immediate accuracy are displayed as a second group. The six limb leads can, however,
in interpretation is required for support of the clinical decision be integrated into one sequence, creating a similarly logical
of triage to the wide range of appropriate therapies, including display as that used routinely for the chest leads. This Cabrera
myocardial reperfusion for individuals who indeed have acute sequence has been used routinely in Sweden for many years.12
thrombotic coronary occlusion. With the development of skilled Spatial contiguity is more readily appreciated in the Cabrera
emergency medical services, prehospital ECG recording and sequence format than in the classical display format. In the clas-
continuous monitoring are increasing in prevalence and in sical format, nonspatially contiguous limb leads are displayed
importance. adjacently; therefore it is possible for “pattern-oriented” ECG
However, few of these patients have the coronary occlusion readers to erroneously consider lead pairs such as I and II as
that requires acute reperfusion therapy to salvage myocardium at contiguous.
risk for infarction. The key ECG marker is the ST segment
deviation from the TP segment baseline that meets criteria for
ST elevation myocardial infarction (STEMI). Although these
criteria have been clearly defined by expert Standards groups,1 Challenges in Considering ECG Limb
many other conditions can cause similar ST segment elevation,2 Lead Relationships
and many acute coronary occlusions produce ST depression
rather than ST elevation on standard ECG leads.3 Limb leads are strictly related mathematically; all six leads
Also, chronic cardiac conditions are often present in these can be derived from any pair of two leads (e.g., leads I and II)
patients, such as left or right ventricular hypertrophy or bundle using
branch block, which confound the ability of emergency medical Einthoven’s law (i.e., lead II = lead I + lead III at any point in
personnel to achieve the correct diagnosis.4 These challenges time) and the way that electrodes are connected to produce the
have led clinical investigators and manufacturers of electrocar- augmented leads aVR, aVL, and aVF. Specifically, the remaining
diographs used in emergency medical services to explore alterna- four limb leads can be expressed in terms of leads I and II in the
tive display methods with images of the potentially involved following way:
myocardium. These include designation of location, size, and III = −I + II
even severity of the acute process and estimation of the relative aVR = −(I + II)/2
extent of already infarcted and potentially reversibly ischemic aVL = (I − III)/2 = I − II/2
myocardium.5-8 aVF = (II + III)/2 = II − I/2
629
and aVF/III), just as there are already five potential pairs of spa-
Reciprocal ECG Leads tially contiguous chest leads, as has been noted.
A study by Perron et al.17 of patients with complete acute
It becomes necessary to introduce the term reciprocal for under- coronary occlusion caused by angioplasty balloon inflation has
standing the relationships among the appearances of ECG wave- documented that further extension of the positive and negative
forms in different leads. Truly reciprocal views are those limb and chest lead displays around full 360-degree “clockfaces”
presenting inverted or “mirror-lake” images. No pair of standard further increases the accuracy of diagnosis. This ECG image is
ECG leads is separated by 180 degrees; however, “almost presented in Figure 65-1 for a patient with acute left circumflex
reciprocal” views are available only in leads such as aVL and III, occlusion. Note that ST elevation is seen only in nonstandard
with their positive poles oriented 150 degrees apart. When leads −V1, −V2, −V3, and −V4. The ST depression in standard
ST elevation indicative of acute ischemia of the anterior left leads V1, V2, V3, and V4 fulfills STEMI equivalent criteria in at
ventricular (LV) wall is present in lead aVL, “almost reciprocal” least two adjacent leads. For patients with STEMI criteria in only
ST depression is typically seen in lead III, and when ST elevation a single lead (e.g., aVL or III), threshold ST depression in adja-
indicative of acute ischemia of the inferior LV wall is present in cent leads (e.g., −III or −aVL) would indicate positive STEMI
lead III, “almost reciprocal” ST depression is typically noted in criteria. In the study by Perron et al.17 an increase in diagnostic
lead aVL. sensitivity without a decrease in diagnostic specificity continued
As indicated in the above formulas, lead aVR is the average of to be achieved by sequential addition of ST elevation thresholds
standard leads I and II with a negative sign. The inverted or of 0.1 mV in 7 of the 12 reciprocal leads: −V1, −V2, −V3, −aVL,
“reciprocal” lead −aVR is the average of leads I and II and there- −I, +aVR, and −III.
fore fits appropriately in the orderly clockwise Cabrera sequence Each ECG lead provides only linear or one-dimensional
of aVL, I, −aVR, II, aVF, III, as has been mentioned. observations of the P wave, the QRS complex, the ST segment,
and the T wave. The imagination of the ECG reader is required
to relate the durations and amplitudes of these waveforms to their
myocardial origins.
Challenges in Considering ECG Chest Lead Nimmermark et al.8 described and validated a method for
Relationships integrating information on ST levels in all 12 ECG leads into
two graphic displays, called ST maps (Figure 65-2). The spatial
The chest leads in their typical orderly display sequence (V1 to positions of the 12 standard leads are indicated on the 360-degree
V6) present a different challenge for their optimal use for diag- clockfaces of both frontal and horizontal planes. Concentric
nosis of acute coronary syndromes. When anterior-posterior circles are placed at sequential 0.1-mV distances from the center.
views of cardiac electrical activity are added to the left-right and Points are placed on each lead line at the level of the ST deviation
superior-inferior views provided by the limb leads, the chest leads on that ECG lead. The points are connected to produce an ST
provide key diagnostic information about LV regions at risk from map. These side-by-side limb-lead views of the frontal plane and
left anterior descending and left circumflex coronary artery chest-lead views of the transverse plane of the patient presented
occlusions. in Figure 65-3 provide an ECG image of the anterior location
ECG recording of the six chest leads requires use of the and moderate extent of acute ischemia, presumably caused by
Wilson central terminal14 to provide the negative electrode for acute left anterior descending artery (LAD) occlusion.
each lead from the averaged limb lead acquisitions. The positive
electrode is provided by the additional placement of a positive
electrode over each of the six bony thoracic landmarks. However,
unfortunately, Wilson et al. considered this single electrode to 2D Imaging Based on Vectorcardiography
provide “anterior unipolar leads,” leading to the common
current mistaken use of the term “anterior leads,” which causes Vectorcardiography (VCG) is a method of developing two-
the mistaken conception that these leads provide information dimensional (2D) images of cardiac electrical activity by display-
only from the “anterior” myocardium. Also, because placement ing the spatial locations of ECG waveforms at each sequential
of the positive electrodes is determined by reference to bony time of their duration. Only three orthogonal leads, termed X,
thoracic landmarks closer to the cardiac source, they lack the Y, and Z, are required, with X and Y providing the frontal plane
“equal” inter-lead spacing provided by the Einthoven triangle image, X and Z the transverse plane, and Y and Z the sagittal
already described for the limb leads. However, it has been dem- plane. This ECG imaging method was developed in the mid-20th
onstrated by Brody et al.15 that the frontal plane triangle is not century and was used for several years as a complement to con-
really equi-angular; therefore it becomes feasible to consider that ventional 12-lead electrocardiography. When the necessary
approximately 30-degree angles are present between each of the resources for computer-based processing of ECG and VCG
six adjacent positive and their six truly reciprocal negative limb signals became available, several studies18-20 described ways of
and chest leads. mathematically deriving VCGs from 12-lead ECGs and vice
versa. These are not exact derivations; a VCG can derive only an
approximation of the 12-lead ECG, and a 12-lead ECG can derive
only an approximation of the VCG.
ECG Imaging to Overcome These Challenges Andersen et al.21 used these vectorcardiographic principles to
develop an alternative to the frontal and transverse planar dis-
Optimal use of the ECG in the clinically challenging emergency plays described by Nimmermark et al.8 ST vectors are derived
evaluation of patients with acute coronary syndromes requires a using the ECG-to-VCG transformation approaches mentioned
change in the way that ECGs are typically interpreted by even previously. Results are presented in frontal and transverse plane
skilled cardiologists. This can be accomplished by expanding the displays, but contrary to the ST map approach, the Andersen
definition of STEMI to include ST depression as “STEMI equiv- display of the “ST compass” contains only arrows that indicate
alent”16 and providing an ECG image of a change in the limb vector direction and vector magnitude in each plane.
lead display format to the orderly sequence of aVL, I, −aVR, II, Transforming ST-J measurements from the 12-lead ECG to
aVF, III. This would provide the full five potential pairs of spa- an ST map and vectorcardiographic measurements to an ST
tially contiguous limb leads (aVL/I, I/−aVR, −aVR/II, II/aVF, compass as described previously are two ways of displaying ST
Electrocardiographic Imaging in Patients With Acute Coronary Syndrome 631
65
-aVF V2
-II -III V1 V3
aVR aVL -V6 V4
-I I -V5 V5
-aVL -aVR -V4 V6
III II -V3 -V1

aVF -V2
A B
Figure 65-1. Side-by-side clockface displays of the 12 frontal plane and 12 transverse plane leads from a patient with acute left circumflex
coronary occlusion The centers of the displays are representative silhouettes of the ventricles. Note that the leads intersect at a “center of cardiac electrical activity”
within the left ventricular (LV) cavity. The 12 inverted leads are designated by minus signs. The arrows indicate the planar arcs of the 12 inverted leads as they progress
beyond the arcs of the 12 standard leads.
I (0.4) aVR (0.0) V1 (0.4) V4 (1.1)
II (–0.3) aVL (0.5) V2 (2.1) V5 (–0.1)
III (–0.7) aVF (–0.5) V3 (2.6) V6 (–0.3)
Limb Leads (unit: mm) Chest Leads (unit: mm)
aVR aVL
+ + + V6
+3.0 I +3.0
+ + V5
-3.0 –3.0
+
V4
+ + + +
+ II + V3
III V1
aVF V2
Figure 65-2. ECG recording from a patient with acute LAD occlusion, with the standard 12-lead recording above and the ST maps below ST
segment deviations in millimeters appear after each lead label above the recording. Limb leads are displayed in the traditional sequence. The ST maps of the limb and
chest leads are displayed directly beneath the standard recordings of these leads. The concentric circles represent 1, 2, and 3 mm (1 mm = 0.1 mV) of ST segment deviation,
whether positive or negative.
Figure 65-3. The standard 12-lead ECG recording of a single cardiac cycle from a patient with early acute anterior infarction is presented
above both Mercator and polar displays of the images of abnormal QRS complex and ST segment waveforms The ECG images are shown on a
basic template of the LV myocardium with a typical distribution of its coronary arteries in both Mercator (A) and polar (B) displays. The 12 segments of the Mercator display
are numbered according to the Selvester sequence,23 and the 17 segments in the polar display are numbered according to the AHA standard.24 Images of the changes in
the QRS complex in black represent the already infarcted portion of the area at risk, and images of deviation of the ST segment in blue represent the remaining ischemic
portion.
measurements in a format that is easily taught to and understood myocardial region at risk and the final portion infarcted, as shown
by health care personnel. These methods concentrate on clarify- in Figure 65-3. The LV myocardium has been divided into 12
ing for the reader the direction and magnitude of ST-J deviation segments in the Mercator projection model as described by
in 3D space. When such a capability is introduced in the ambu- Selvester et al.,23 and more recently into 17 segments as explained
lance or in the critical care unit (CCU), the map or compass will by Cerquiera et al. in the American Heart Association (AHA)
update continuously so that the medical staff may closely follow standard model.24 Both divide the LV circumferentially into
the progression or regression of ST deviations. Alarms can be quadrants, but various terms are used, including anterior, septal or
activated when a preset threshold of ST segment deviation is anteroseptal, anterior or anterosuperior, posterior, lateral or posterolat-
surpassed. eral, and inferior or diaphragmatic. The current generally accepted
terms are septal, anterior, lateral, and inferior.24
Ubachs et al.25 applied heuristic rules on the presence/absence
of ST deviation in the 12 ECG leads of 11 patients with first-time
3D Imaging Using Mercator or Polar Displays ST elevation infarction to derive a Mercator projection of “myo-
cardial at risk” in each patient. The Mercator projection was then
For several centuries various displays have been used to present transformed to a polar plot and was compared with polar plots
a 2D map of the three-dimensional earth, including a Mercator from SPECT and magnetic resonance imaging (MRI) (Figure
display of the four circumferential quadrants of the earth verti- 65-4). Good correspondence between SPECT, MRI, and ECG
cally, and polar displays of the two hemispheres of the earth was found in this small material.
horizontally. Strauss et al.26 used the aforementioned technique of synthe-
Methods for representing the LV myocardium in polar or sizing a VCG from the 12-lead ECG using data from 32 patients
“bull’s eye” displays have been used routinely for imaging of the in the STAFF III study.11 They displayed the location of the
distribution of myocardial perfusion in various regions of LV vector on a Mercator map as well as on a polar plot .The location
myocardium by single-photon emission computed tomography of the ST vector projected within the SPECT ischemic region
(SPECT).22 Both Mercator and polar displays have recently been in 100% of patients with LAD occlusion, 75% of LCx patients,
adapted for imaging of acute ischemia and infarction based on and 65% of RCA patients. The size of the vector correlated well
ECG characteristics of ST segments and QRS complexes, respec- (P = .68) with the area at risk as assessed by myocardial SPECT
tively. Indication of the typical locations of the three major coro- imaging.
nary arteries provides the framework for relating the Bacharova et al.27 made a mathematical transformation (the
ECG-indicated sites of their occlusion to both the ischemic DECARTO method) from the ST segment distribution in the
#1 #2 #3 #4 #5 #6
LAD
65
SPECT
LAD/LAD LAD/LAD LAD/LAD LAD/LAD LAD/LAD LAD/LAD
MRI wall
thickening
LAD/LAD LAD/LAD LAD/LAD LAD/LAD LAD/LAD RCA/LCX
ECG
LAD/LAD LAD/LAD LAD/LAD LAD/LAD LAD/LAD LAD/LAD
#7 #8 #9 #10 #11
RCA LCX
SPECT SPECT
RCA/RCA RCA/RCA RCA/RCA RCA/LCX RCA/LCX
MRI wall MRI wall

thickening thickening
ECG ECG

Figure 65-4. Polar displays of tetrofosmin perfusion SPECT, myocardial thickening MRI, and ST segment deviation ECG images from six
patients with acute LAD, four with RCA, and one with LCX occlusions Note the similarities in locations of the estimated areas at risk in the images from
the three modalities.
12 leads to a polar plot of left ventricular myocardium. The polar Conclusions and Future Developments
plot localizes and estimates the size of the myocardial region
affected. Electrocardiography provides a unique method of conveying to
Horacek and coworkers28 have shown that body surface poten- emergency medical personnel the status of the ischemic myocar-
tial maps (BSPMs) of ST-J deviation need not be based on dium in patients with suspected acute coronary syndrome (ACS).
large numbers of torso ECG leads. In fact, patterns of ST devia- The present situation whereby most decisions are made from the
tion seen with BSPMs based on 80 to 200 leads are very well standard 12-lead ECG with the leads in noncontiguous sequences
retained even when the BSPM is based on the 12 leads in the is certainly very far from optimal. Universal adoption of the
conventional ECG. Figure 65-5, A, shows pairs of BSPMs from Cabrera sequence would be one possible way to advance, but the
patients with acute coronary occlusion in one of the major coro- methods presented in this chapter provide further steps toward
nary arteries, an easily assimilated and understood display that reflects all myo-
The so-called inverse problem of electrocardiography cardial regions and their ischemia. The finally presented method
(deriving epicardial potential distribution from registrations of of Horacek et al. accounts for the varying distances between
thoracic potential distribution) has been solved on the basis of torso-placed electrodes and the various LV walls to provide an
the BSPM that is generated by the 12-lead ECG to produce image of acute myocardial ischemia/infarction for clinical prac-
images of epicardial potential distribution; a few examples are tice in ambulances, in emergency departments, and in coronary
seen in Figure 65-5, B. care units.
Recorded LCx Estimated Recorded RCA Estimated
Recorded LAD Estimated
1 1 1
7 7 7
2 6 2 6 2 6
8 13 8 13 8 13
12 12 12
14 17 16 14 17 16 14 17 16
9 15 11 9 15 11 9 15 11
3 5 3 5 3 5
10 10 10
4 4 4
B
Figure 65-5. Pairs of recorded body surface maps (Panel A) using 120 torso electrodes and estimated body surface maps using the 12 standard
ECG leads are shown for five patients with LCX occlusion, two with RCA occlusion, and two with LAD occlusion The dots on each map indicate
the positions of the six chest leads. Various shades of red indicate ST elevation, and various shades of green indicate ST depression. Note the similarities between actual
and estimated maps for each patient. Panel B presents polar displays using the AHA 17-segment model.24 From left to right are the calculated epicardial potential distribu-
tions at ST-J based on the actual map, the 12-lead estimated potential distribution, and the SPECT image of myocardial perfusion. Note the similar appearances of images
of the area at risk in the LCx distribution from actual and ECG estimated distributions, and from the SPECT image.
(From Refs 28 and 29.)
3. Nikus K, Pahlm O, Wagner G, et al: Electrocardio- 8. Nimmermark MO, Wang JJ, Maynard C, et al:
References graphic classification of acute coronary syndromes: ST-segment measurements in numeric and graphic
A review by a committee of the International form for assistance of cardiologists’ decisions
1. Wagner GS, Macfarlane P, Wellens H, et al; Society for Holter and Non-Invasive Electrocardi- of reperfusion therapy for patients with acute
American Heart Association Electrocardiography ology. J Electrocardiol 43:91–103, 2010. coronary occlusion. J Electrocardiol 44:502–508,
and Arrhythmias Committee, Council on Clinical 4. Sejersten M, Young D, Clemmensen P, et al: Com- 2011.
Cardiology; American College of Cardiology parison of the ability of paramedics with that of 9. Christian TF, Gibbons RJ, Clements IP, et al: Esti-
Foundation; Heart Rhythm Society: AHA/ACCF/ cardiologists in diagnosing ST-segment elevation mates of myocardium at risk and collateral flow in
HRS recommendations for the standardization and acute myocardial infarction in patients with acute acute myocardial infarction using electrocardio-
interpretation of the electrocardiogram. Part VI. chest pain. Am J Cardiol 90:995–998, 2002. graphic indices with comparison to radionuclide
Acute ischemia/infarction: A scientific statement 5. Olson CW, Warner RA, Wagner GS, et al: A and angiographic measures. J Am Coll Cardiol
from the American Heart Association Electrocar- dynamic three-dimensional display of ventricular 26:388–393, 1995.
diography and Arrhythmias Committee, Council excitation and the generation of the vector and 10. Persson E, Pettersson J, Ringborn M, et al: Com-
on Clinical Cardiology; the American College of electrocardiogram. J Electrocardiol 34(Suppl):7– parison of ST-segment deviation to scintigraphi-
Cardiology Foundation; and the Heart Rhythm 16, 2001. cally quantified myocardial ischemia during acute
Society; endorsed by the International Society for 6. Andresen A, Gasperini MD, Myers R, et al: An coronary occlusion produced by percutaneous
Computerized Electrocardiology. Circulation improved automated ECG algorithm for detecting transluminal coronary angioplasty. Am J Cardiol
119:e262–e270, 2009. acute and prior myocardial infarction. J Electrocar- 97:295–300, 2006.
2. Birnbaum Y, Bayés de Luna A, Fiol M, et al: diol 35(Suppl):105–110, 2002. 11. van Hellemond IEG, Bouwmeester S, Olson CW,
Common pitfalls in the interpretation of ECGs 7. Horacek BM, Mirmoghisi M, Warren JW, et al: et al: Consideration of QRS complex in addition to
from patients with acute coronary syndromes with Detection of myocardial ischemia by vessel-specific ST segment abnormalities in the estimation of the
narrow QRS: A consensus report. J Electrocardiol leads derived from the 12-lead ECG and its ‘risk region’ during acute anterior myocardial
45:363–375, 2012. subsets. J Electrocardiol 41:508–517, 2008. infarction. J Electrocardiol 44:170–176, 2011.
12. White T: The sequence of the extremity leads in 18. Dower GE: The ECGD: A derivation of the ECG heart: A statement for healthcare professionals
ECG. Lakartidningen 68:1352–1356, 1971. [In from VCG leads. J Electrocardiol 17:189–191, from the cardiac imaging committee of the Council
Swedish]
13. Horacek BK: Lead theory (Chapter 10, pp 348–
374). In Macfarlane PW, van Oosterom A, Pahlm
1984.
19. Macfarlane PW, Edenbrandt L: 12-Lead vector-
cardiography in ischemic heart disease. J Electro- 25.
on Clinical Cardiology of the American Heart
Association. Circulation 105:539–542, 2002.
Ubachs JFA, Engblom H, Hedström E, et al: Loca-
65
O, et al, editors: Comprehensive Electrocardiol- cardiol 24(Suppl):188–193, 1992. tion of myocardium at risk in patients with first-
ogy, New York, 2011, Springer. 20. Kors JA, van Herpen G, Sittig AC, et al: Recon- time ST-elevation infarction: Comparison among
14. Bryant JM, Johnston FD, Wilson FN: Unipolar struction of the Frank vectorcardiogram from stan- single photon emission computed tomography,
electrocardiographic leads: Effects produced by dard electrocardiographic leads: Diagnostic magnetic resonance imaging, and electrocardiogra-
eliminating the resistors between the limb elec- comparison of different methods. Eur Heart J phy. J Electrocardiol 42:198–203, 2009.
trodes and the central terminal. Am Heart J 11:1083–1092, 1990. 26. Strauss DG, Olson CW, Wu KC, et al: Vectorcar-
37:321–332, 1949. 21. Andersen MP, Terkelsen C, Struijk J, Jr: The ST diogram synthesised from the 12-lead electrocar-
15. Brody DA: The meaning of lead vectors and Compass: Spatial visualization of ST segment diogram to image ischemia. J Electrocardiol
the Burger triangle. Am Heart J 48:730–737, deviation and estimation of the ST injury vector. 42:190–197, 2009.
1954. J Electrocardiol 42:181–189, 2009. 27. Bacharova L, Mateasik A, Carnicky J, et al: The
16. Martin TN, Groenning BA, Steedman T, 22. Goris ML, Boudier S, Briandet PA: Two- DipolarElectroCARdioTOpographic(DECARTO)-
et al: ST-segment deviation analysis of the dimensional mapping of three-dimensional like method for graphic presentation of location and
admission 12-lead electrocardiogram as an aid SPECT data: A preliminary step to the quantita- extent of area at risk estimated from ST-segment
to early diagnosis of acute myocardial infarction tion of thallium myocardial perfusion single deviations in patients with acute myocardial infarc-
with a cardiac magnetic resonance imaging gold photon emission tomography. Am J Physiol tion. J Electrocardiol 42:172–180, 2009.
standard. J Am Coll Cardiol 50:1021–1028, Imaging 2:176–180, 1987. 28. Horáček BM, Sapp JL, Penney CJ, et al: Compari-
2007. 23. Selvester RH, Strauss DG, Wagner GS: Myocar- son of epicardial potential maps derived from the
17. Perron A, Lim T, Pahlm-Webb U, et al: Maximal dial infarction (Chapter 16). In Macfarlane PW, 12-lead electrocardiograms with scintigraphic
increase in sensitivity with minimal loss of specific- van Oosterom A, Pahlm O, editors: Comprehen- images during controlled myocardial ischemia.
ity for diagnosis of acute coronary occlusion sive Electrocardiology, New York, 2011, Springer, J Electrocardiol 44:707–712, 2011.
achieved by sequentially adding leads from the pp 653–746. 29. Horáček BM, Warren JW, Wang JY: Heart surface
24-lead electrocardiogram to the orderly sequenced 24. Cerqueira MD, Weissman NJ, Dilsizian V, et al: potentials estimated from 12-lead electrocardio-
12-lead electrocardiogram. J Electrocardiol 40: Standardized myocardial segmentation and grams. Proc of Computing in Cardiology 37–40,
463–469, 2007. nomenclature for tomographic imaging of the 2010.
Head-up Tilt Table Testing 66
David G. Benditt, Wayne O. Adkisson, and Richard Sutton
CHAPTER OUTLINE Although the utility of tilt table testing remains controversial,
the test has nonetheless become widely accepted as a valuable
Historical Background 637 tool for the evaluation of susceptibility to vasovagal syncope.7
Transient Loss of Consciousness and Syncope 637 This chapter summarizes the rationale for and current status of
head-up tilt table testing as principally applied to evaluation of
Pathophysiology of Loss of Consciousness 637 the vasovagal faint.
Approach to the Syncope Diagnosis 638
Physiological Impact of Upright Posture 638
Pathophysiology of Vasovagal Syncope 639 Transient Loss of Consciousness and Syncope
Pathophysiology of Orthostatic Hypotension 640 Syncope is a clinical syndrome, not a specific diagnosis; it is
Orthostatic Provocation for Assessing Susceptibility characterized by a transient and self-terminating interruption of
to Vasovagal Syncope and Orthostatic Hypotension: normal global cerebral activity with associated loss of conscious-
Rationale and Methods 641 ness and postural tone. The underlying cause is a relatively brief
period (a minute or two at the most) of inadequate delivery of
Conclusions 646 oxygen, glucose, and other nutrients to brain tissues.7,8
Although syncope is the most common basis for transient loss
of consciousness (TLOC), it is but one of many conditions that
must be considered when a patient presents with an apparent
Historical Background collapse. Consequently, before the evaluation proceeds to assess-
ment of possible underlying causes in a patient with presumed
Head-up tilt table has been used for longer than half a century syncope, careful consideration must be given to distinguishing
by physiologists and physicians to study heart rate and blood the presenting complaint from the many other forms of TLOC
pressure adaptations to changes in position, to model responses or TLOC mimics (e.g., seizures, concussion, psychogenic
to hemorrhage, to assess the characteristics of orthostatic hypo- pseudosyncope, malingering). Unfortunately, physicians often
tension, and to evaluate hemodynamic and neuroendocrine are uncertain about this distinction, with confusion exacerbated
responses in congestive heart failure, autonomic dysfunction, and by imprecise writing in the medical literature, as has been high-
hypertension. During the course of such studies, incidental lighted recently.9,10
observations noted that some test subjects experienced total or
near-total transient loss of consciousness as a result of systemic
hypotension.1-5 Furthermore, in some cases hypotension was
associated with impressive degrees of bradycardia, including pro- Pathophysiology of Loss of Consciousness
longed self-terminating periods of asystole (Figure 66-1).
However, these early studies did not incorporate a more formal Neuronal tissue has limited energy storage capability. Conse-
assessment of the diagnostic utility of this observation at the time. quently, a well-maintained flow of oxygenated blood to the brain
The concept of head-up tilt table provocation as a diagnostic is crucial; the autoregulation of cerebrovascular blood flow is
test for reflex syncope began to evolve only after the landmark crucial in this regard. In healthy young persons, cerebral blood
report by Kenny and colleagues in 1986.6 At first, prolonged flow ranges from 50 to 60 mL per 100 g of brain tissue/min,
periods (up to 2 h) of passive head-up posture with the patient representing about 12% to 15% of resting cardiac output. A flow
resting on a tilt table (at 40° to 60°) were used as the sole means of this magnitude easily meets minimum oxygen (O2) require-
of provoking vasovagal events. Subsequently, additional interven- ments to sustain consciousness (approximately 3.0 to 3.5 mL
tions were introduced in an attempt to improve the sensitivity of O2/100 g tissue/min).11,12 However, the safety factor for oxygen
the test, although at some detriment to specificity. These inter- delivery may be markedly impaired in older individuals; in indi-
ventions included pharmacologic provocation (primarily isopro- viduals with diseases such as hypertension, diabetes mellitus, or
terenol or nitroglycerin, but also adenosine, edrophonium, and heart failure; and in those in a hypoxemic state (e.g., chronic
clomipramine). Furthermore, certain physical maneuvers includ- pulmonary disease) (Figure 66-2).
ing carotid sinus massage (at times in conjunction with edropho- In general, sudden cessation of cerebral blood flow for 10
nium) were used in some laboratories. Currently, isoproterenol seconds or longer is usually sufficient to cause complete loss of
and nitroglycerin remain the most widely used pharmacologic consciousness. Furthermore, whatever the heart rate, either a
provocative agents in diagnostic tilt table testing laboratories. In decrease in systolic blood pressure to below 60 mm Hg or a fall
a recent survey, nitroglycerin was more widely used in Europe, of as little as 20% in cerebral oxygen delivery may trigger syncope
and isoproterenol was preferred in North America. (see Figure 66-2).
637
Box 66-1 Neurally Mediated Syncope Syndromes

16.3 s
• Vasovagal syncope (common or emotional faint)
• Carotid sinus syncope
• Situational syncope
• Micturition or postmicturition syncope
• Airway stimulation–induced syncope
Continuous tracing 1s • Cough or sneeze syncope
Figure 66-1. Electrocardiographic (ECG) monitor recording obtained during a • Gastrointestinal stimulation–induced syncope
spontaneous vasovagal faint, illustrating a self-terminating prolonged asystolic • Swallow syncope
pause. The patient was deemed to have cardioinhibitory syncope. Concomitant • Defecation syncope
vasodepressor (vasodilatation) also was probably present but could not be docu- • Increased intrathoracic pressure–induced syncope
mented by current ambulatory technology.
• Trumpet playing
• Weightlifting
• Glossopharyngeal and trigeminal neuralgia
CEREBRAL AUTOREGULATION • Postprandial syncope
• Miscellaneous
• Syncope associated with aortic stenosis
• Syncope accompanying onset of certain tachyarrhythmias
Cerebral blood flow
Hypertensive (atrial fibrillation, paroxysmal supraventricular tachycardia,

patient and possibly certain episodes of ventricular tachycardia)
60 mm Hg 140 mm Hg
Typical patient diagnosis of the cause of the syncope. In such cases, the avail-
ability of diagnostic tests becomes important.
Arterial pressure
Figure 66-2. Schematic graph illustrating the relation between cerebral blood flow
(ordinate) and systemic arterial pressure (abscissa) in a “typical” patient and a hyper-
tensive patient. Cerebral flow is “autoregulated” over a wide pressure range (i.e.,
Physiological Impact of Upright Posture
does not vary over this pressure range). The range is influenced by concomitant
disease conditions such as hypertension, as indicated here. Below approximately Upright posture elicits an orthostatic stress caused by the effects
60 mm Hg, blood flow falls, which may lead to loss of consciousness. of gravity on the distribution of circulating blood volume in the
body.1,2,4,6-8,11 Initially, as upright posture is achieved, the effects
of gravity result in shifting of approximately 500 to 1000 mL of
blood to the lower part of the body and particularly into the
Approach to the Syncope Diagnosis highly compliant splanchnic bed.11,14 Most of this redistribution
occurs in the first 10 seconds of upright posture. Subsequently,
Many conditions may cause physiological disturbances leading to in normal persons, an additional 700 mL of protein-free fluid is
syncope; neurally mediated reflex syncope, orthostatic hypoten- filtered into the interstitial space in the next 10 min.11 The result
sion, and tachyarrhythmias or bradyarrhythmias are among the of these two shifts of fluid within and outside of the vascular space
more frequent causes. Vasovagal syncope is the most frequent is marked reduction of venous return and stroke volume.
form of the neurally mediated reflex faints (Box 66-1), and is the Humans attempt to compensate for diminution of stroke
most common of all causes of syncope across all age groups. volume during movement to upright posture by both increasing
Accordingly, its recognition (and, if necessary, preventive treat- heart rate and constricting resistance and capacitance vessels.11
ment) is a problem often encountered by a wide range of medical Heart rate increase alone, however, is usually insufficient to
practitioners. maintain cardiac output and cerebral blood flow during these
Determining the basis of syncope in a given patient begins circumstances. Vasoconstriction of systemic blood vessels is
with both a careful medical history (including reports from wit- crucial to maintenance of arterial blood pressure. Prevention of
nesses) and a thorough physical examination, the latter incorpo- syncope requires that the compensatory cardiovascular response
rating orthostatic blood pressure measurements in all patients maintains arterial pressure (in particular, systemic pressure at the
and carotid sinus massage in older persons (usually >50 years of level of the carotid arteries) at a value at least equal to the
age). minimum value needed to guarantee adequate cerebral blood
In the case of vasovagal syncope, the diagnosis can usually be flow (approximately 60 mm Hg).12
established by the medical history without further testing, Short-term cardiovascular adjustments triggered by ortho-
although the history taking may need to include observations static stress are mediated primarily by the autonomic nervous
made by eyewitnesses.7,8 When events surrounding the faint(s) system; the main sensory receptors involved are the arterial
are “atypical,” however, even an experienced history taker may mechanoreceptors (i.e., baroreceptors responding to pressure,
not be certain of the diagnosis. Furthermore, often it is not pos- stretch, or both) located in the aortic arch and carotid sinuses.
sible to obtain an adequate history (especially in the elderly or Mechanoreceptors in the heart walls (both in the atria and in the
the very young). In some of these latter instances, absence of ventricles) and in the lungs (cardiopulmonary receptors) are
reliable facts may be due to the patient’s lack of insight, or (espe- thought to play an additional but more minor role. In the longer
cially in the elderly) to retrograde amnesia for premonitory term, neuroendocrine (e.g., renin-angiotensin system, vasopres-
events.13 Furthermore, eyewitnesses may not have been present, sin) compensatory mechanisms may be initiated by prolonged
or their reports may not be sufficiently detailed to establish a orthostatic stress, but these adjustments are of lesser concern in
Head-up Tilt Table Testing 639
terms of the abrupt pathophysiology of vasovagal or orthostatic information, the central response to the afferent signals would be
66
faints. modified.
Reflex activation of central sympathetic outflow to systemic As noted earlier, the pathophysiology of the vasovagal form
blood vessels can be reinforced by local reflex mechanisms (e.g., of neurally mediated reflex syncope remains incompletely under-
venoarteriolar reflex) and by more global mechanisms such as the stood. However, it can be considered in terms of four basic ele-
skeletal muscle pump (even in the absence of overt movement) ments: (1) the afferent limb; (2) central nervous system processing;
and the respiratory pump. Each of these may play an important (3) the efferent limb; and (4) feedback loops.18 In brief, central
adjunctive role in the maintenance of arterial pressure by pro- and peripheral mechanoreceptors (see earlier) and occasionally
moting venous return. chemoreceptors (e.g., in the presence of myocardial ischemia) are
Failure of compensatory adjustments to orthostatic stress may, presumed to initiate afferent sensory nerve traffic to the cardio-
in the absence of preexisting volume overload, result in inability vascular modulation center in the brain stem. The manner in
to maintain central volume (the heart cannot pump what it does which sensory processing results in inappropriate vasodilatation
not receive), reduction of systemic arterial pressure, and inade- and/or bradycardia is unknown, but may in part relate to lack of
quate cerebrovascular blood flow. If the deficits are sufficiently concordance among several afferent inputs, analogous to the situ-
severe, the ultimate outcome is nearly complete or complete loss ation described earlier for CSS. In any event, systemic hypoten-
of consciousness caused by systemic hypotension. The outcome, sion (ultimately leading to a vasovagal faint, if severe) is primarily
when principally due to inadequate maintenance of blood pres- the result of vasodilatation triggered by a marked reduction in
sure in the face of gravitational stress is usually termed orthostatic sympathetic vasoconstrictor outflow to blood vessels in skeletal
hypotension or orthostatic syncope. In addition, however, in suscep- muscles and substantial increase in venous capacitance, particu-
tible persons, an inappropriate set of neural reflex responses may larly in the splanchnic bed.14 Parasympathetically mediated bra-
be triggered: vasodilatation and severe or “relative” bradycardia— dycardia (severe, or “relative,” with respect to the severity of
the vasovagal response.7,11,15 hypotension) contributes to the process but plays a lesser role,
except when prolonged asystole occurs. A poorly understood
failure of baroreceptor feedback to recognize and interrupt this
process also appears to be important in facilitating development
Pathophysiology of Vasovagal Syncope of hypotension.21
Impaired cerebral autoregulatory responses also may contrib-
The vasovagal reflex in humans appears to have a long evolution- ute to the development of vasovagal responses in some patients.
ary history,16 but the basis for the reflex remains uncertain. In this It has been reported that cerebral blood flow velocity can decline
regard, several theories have developed, two of which may be before arterial pressure and cerebral vasoconstriction, and that in
based on early protective mechanisms. The first relates to protec- some cases, cerebral hypoxia may occur in the absence of systemic
tion from predators by means of “playing dead” (perhaps a par- hypotension. The role of cerebrovascular spasm as a mechanism
ticular advantage for human noncombatants such as women and for transiently inadequate cerebral perfusion has been raised, but
children)—a strategy that is presumed to buy time and set up an its frequency and importance are unclear.22
opportunity for later escape. A second, the so-called “clotting Several observations suggest that symptomatic hypotension-
hypothesis,” may have been a means of reducing severity of hem- bradycardia associated with a positive result on head-up tilt table
orrhage by reducing blood pressure and flow. Other theories testing is comparable with spontaneous neural reflex–mediated
attribute the vasovagal reflex as being (1) the human equivalent vasovagal syncope. First, both spontaneous and induced syncopal
of the “alarm” reaction in some animals that is marked by heart episodes tend to be associated with similar premonitory symp-
rate slowing and reduced blood pressure; or (2) a means of car- toms (e.g., nausea, diaphoresis) and signs (e.g., preceding tachy-
dioprotection under conditions of extreme stress.16 cardia followed by relative or marked heart rate slowing, marked
Whatever the origins of the vasovagal reflex, it is thought to pallor, loss of postural tone). Second, the temporal sequence of
be triggered at least in part by excessive stretch of central mecha- blood pressure and heart rate changes during tilt-induced
noreceptors (including those residing in the atria and possibly syncopal spells parallels that reported for spontaneous episodes
in the left ventricle) as intrathoracic circulating volume falls. (Figure 66-3). Finally, measurements of plasma catecholamines
Of note, however, although this mechanism may contribute before and during spontaneous and tilt-induced syncope exhibit
to vasovagal faints triggered by upright posture, other common important similarities. In particular, premonitory increases in
triggers (e.g., pain, emotional upset) are not readily accounted circulating catecholamines, epinephrine more than norepineph-
for by this explanation. In any event, it is the upright posture– rine (i.e., before evident systemic hypotension), appear to char-
triggered reflex that forms the basis for the use of tilt table acterize both the spontaneous vasovagal faint and tilt-induced
testing in the evaluation of patients with suspected vasovagal hypotension-bradycardia.18,23
syncope.6,7,17,18 From a clinical perspective, most vasovagal fainters exhibit
Vasovagal syncope is the most frequently occurring form of a mixed cardioinhibitory and vasodepressor responses (Box 66-
larger group of neurally mediated reflex faints that also include 2).24,25 In some patients, however, the observed response will be
carotid sinus syndrome and postmicturition syncope (see Box predominantly cardioinhibitory, with hypotension caused by a
66-1). In terms of pathophysiology, carotid sinus syndrome (CSS) prolonged asystolic pause. On rare occasions, a pure vasodepres-
is perhaps the best understood of the neurally mediated reflex sor response may be observed, although even in these cases the
syncope syndromes.19 In CSS, excessive stretch of carotid baro- concomitant tachycardia is less than that expected for the severity
receptors (perhaps initiated by abrupt neck movements or tight of hypotension.
collars) provides an increase in afferent signals to the central Classification of the various types of vasovagal responses is
nervous system. Central processing of these signals ultimately difficult, and its clinical usefulness is doubtful; as was demon-
causes an efferent neural reflex response, resulting in heart rate strated in the International Study on Syncope of Uncertain Etiol-
slowing and vascular dilatation to compensate for a perceived ogy (ISSUE), heart rate–blood pressure features recorded during
increase in central arterial pressure. In many cases, concomitant tilt table observations only infrequently coincide with those
denervation of afferent proprioceptive nerves is thought to occur, accompanying spontaneous faints in a specific patient.25 In any
thereby depriving the central nervous system of critical informa- case, the most widely accepted classification is that derived from
tion indicating that neck movement was in fact the instigating the Vasovagal Syncope International Study (VASIS) (see Box
trigger.19,20 Presumably, on the basis of the latter additional 66-2).24,25 By definition, a pure cardioinhibitory response manifests
V1
150
100
50
FA
0
10 s
Figure 66-3. Laboratory recording illustrating t

Cardiac Electrophysiology, From Cell To Bedside, 6th Edition

Uploaded by

Copyright:

Available Formats

Cardiac Electrophysiology, From Cell To Bedside, 6th Edition

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cardiac Electrophysiology, From Cell To Bedside, 6th Edition

Uploaded by

Copyright:

Available Formats

Sixth Edition

With 742 illustrations

Cardiac Electrophysiology: From Cell to Bedside ISBN: 978-1-4557-2856-5

Library of Congress Cataloging-in-Publication Data

Cardiac electrophysiology (2014)

Senior Editor: Dolores Meloni

Last digit is the print number: 9 8 7 6 5 4 3 2 1

Hugues Abriel, MD, PhD Justus M. B. Anumonwo, PhD

Wayne O. Adkisson, MD Luciana Armaganijan, MD, PhD, MHS

Oluseun Alli, MD Peter H. Backx, DVM, PhD

Donald M. Bers, PhD Pedro Brugada, MD, PhD

Günter Breithardt, MD Hugh Calkins, MD

Ernesto Carafoli, MD Stuart J. Connolly, MD, FRCPC

Lei Chen, MD Frank A. Cuoco, Jr., MD, MBA, MS

Aman Chugh, MD Dawood Darbar, MBChB, MD

Mario Delmar, MD, PhD

Michael R. Franz, MD, PhD, FHRS William J. Groh, MD, MPH

Masahiko Hoshijima, MD, PhD José Jalife, MD

Jodie Ingles, PhD Jonathan M. Kalman, MBBS, PhD

Mohamed Hani Kanj, MD Peter Kohl, MD, PhD, FHRS

Paulus Kirchhof, MD, FESC, FHRS Karl-Heinz Kuck, MD, PhD

Bruce B. Lerman, MD John C. Lopshire, MD, PhD

Ariane J. Marelli, MD, MPH

Steven M. Markowitz, MD Fred Morady, MD

Jeanne M. Nerbonne, PhD Douglas L. Packer, MD

Brian Olshansky, MD Geoffrey S. Pitt, MD, PhD

Hakan Oral, MD Sunny S. Po, MD, PhD

Jin O-Uchi, MD, PhD Silvia G. Priori, MD, PhD

Jason O. Robertson, MD, MS Frank B. Sachse, PhD

Richard B. Robinson, PhD Lindsey L. Saint, MD

Robert A. Rose, PhD José A. Sánchez-Chapula, MD, PhD

Rainer Schimpf, MD Allan C. Skanes, MD, FRCPC

Harikrishna Tandri, MD Mintu P. Turakhia, MD, MAS

Richard G. Trohman, MD Victoria L. Vetter, MD, MPH, MSHP

Zian H. Tseng, MD, MAS Sami Viskin, MD

Niels Voigt, MD Y. Joseph Woo, MD

Voltage-Gated Sodium Channels and

conduction in artificial lipid membranes and expression of the

Subunit Structure of Sodium Channels Three-Dimensional Structure of

LEFT VENTRICLE LEFT ATRIUM

I IV I IV peptide and analysis by multidimensional NMR methods.53

of CaV1.3 have also been described.8,9 These variants result in the

LTCCs are multimeric proteins consisting of an α1-subunit that

+H3N CO–2 +H3N

Surface Membrane T-Tubule Caveolae

EC Coupling Excitability Gene Regulation Nucleus

activated T-type Ca2+ channels also play important roles in

Transient Receptor Potential (TRP) Channels

B –10 –20 –10

–100 100 mV SC = 38.4 +

lack functional Nav channels, the upstroke of the action potential

repolarization, whereas the delayed rectifiers (IK) determine the

Table 3-1. Potassium Conductances Expressed in the Mammalian Myocardium

and IKr (and possibly other cardiac channels), however, remain to

is proarrhythmic.75,76 Previous studies have identified regional

the resulting channels depends partially or solely on interactions

Mechanism of Polyamine-Induced Rectification

rectification varies among the different members of the Kir sub-

LQT7/ATS1 G144A/S G146D/S display a normal QT interval in contrast to patients affected by