Cardiac optogenetics: a decade of enlightenment

Optogenetic sensors: GECI and GEVI

Optical tracking of changes in membrane voltage and intracellular calcium allows a comprehensive assessment of cardiac activity with very high spatiotemporal resolution. Optogenetic sensors add both cell and organelle specificity as well as the ability to perform in vitro and in vivo longitudinal studies. The cameleon proteins were the first GECIs to be developed and used calmodulin as a sensing element fused to fluorescent proteins excitable with blue–green light⁵². Various kinetic properties were optimized in the GCaMP family of GECIs (consisting of green fluorescent protein, calmodulin and Ml 3 protein), with GCaMP6 variants⁵³ being the latest member. Following the first cardiac applications of GCaMPs around 2006^7-9, these GECIs remain by far the most widely used optogenetic sensors for in vivo experiments, especially for cardiac applications. The red-shifted variants, including the R-CaMP^54,55 and the R-GECO⁵⁶ family of GECIs, were developed to be used in combination with blue-excitable optogenetic actuators. One study developed improved-sensitivity versions (JR-CaMP and jR-GECO) that were better suited for use with channelrhodopsin 2 (ChR2) actuation⁵⁷. Finally, a near-infrared (NIR) fluorescent GECI, NIR-GECO⁵⁸, was developed in 2019, with improved spectral separation from optogenetic actuators and potential voltage sensors, while also providing deeper probing owing to NIR light penetration. Some of these GECIs, such as jR-GECO, come close in performance (in terms of kinetics and quantum efficiency) to the best synthetic calcium indicators when used in vitro. For long-term, repeated probing in vitro, as well as for in vivo measurements, GECIs are extremely valuable and reasonably well tolerated. Additional advantages of GECIs include organelle-specific targeting, such as nuclear, mitochondrial and sarcoplasmic reticulum targeting, to study compartment-specific changes in calcium levels^37,59. In cardiac applications, red-shifted GECIs have been used in combination with optogenetic actuators^60-63, optogenetic voltage sensors that are blue-shifted or NIR^60,64,65 or with excellent-performance NIR synthetic dyes, such as BeRST1 or di-4ANBDQBS⁶².

Genetically encoded voltage indicators (GEVIs) have lagged behind the GECIs in development and deployment in mammalian cells, including cardiomyocytes, owing to difficulties in achieving fast kinetics of sensing as well as their potential to interfere with native electrophysiology (as many large membrane proteins would). However, great progress towards this goal has been achieved in the past decade, and some of these developments have been used in cardiac research. The first, more-widely used GEVIs were VSFP2.3^66,67 and ArcLight⁶⁸. VSFP2.3, ArcLight and the red-shifted variant FlicR1⁶⁹ (excitable at 570 nm) are all based on a phosphatase voltage-sensing domain (Ci-VSD). These GEVIs have been validated for use in cardiomyocytes in vitro^64,70-73, as well as to monitor cardiac electrical activity in transgenic mice in a cell-specific manner^74,75. VSFP2.3 and ArcLight have relatively slow kinetics compared with synthetic voltage-sensitive dyes and cannot accurately capture the upstroke and high-frequency components of the action potential; however, these indicators are useful in reporting rate responses and relative changes in action potential duration. The faster ASAP family of GEVIs^76-78 now includes the latest variant, ASAP3, with high response to two-photon excitation in vivo (in the brains of mice)⁷⁶. The GEVI Voltron, developed in 2019, uses a Halo-tag and a synthetic dye as a fluorophore instead of a fluorescent protein, thereby representing an optogenetic–chemogenetic hybrid GEVI for in vivo use⁷⁹. Voltron525 showed the highest sensitivity to voltage, although a wide spectrum of other sensors, excitable between 525 nm and 635 nm, was generated with the Voltron modular design, leveraging the spectral properties of fluorophore partners⁷⁹. ASAP and Voltron have not yet been used to record cardiac action potentials. The motivation to obtain optical sensors designed for all-optical electrophysiology (considering the shorter-wavelength excitatory and silencing opsins) led to the development of a range of NIR-absorbing GEVIs, starting with the Quasars⁸⁰ and, more recently, Archon1⁸¹, NIR-Butterfly⁸² and Voltron635⁷⁹ (Fig. 2b). The drive to generate these NIR sensors is also motivated by their expected superior performance in vivo to shorter-wavelength indicators for probing deep within the tissue by avoiding haemoglobin absorption.

CaViar⁸⁰, a combined GEVI–GECI construct, was designed and applied to study the zebrafish heart in vivo⁸³. In later studies, CaViar was used in combination with the actuator CheRiff for all-optical electrophysiology studies in human induced pluripotent stem cell (iPSC)-derived cardiomyocytes (iPSC-CMs)⁶⁰. In addition to voltage and calcium, cytosol and mitochondria redox state have been recorded using an optogenetic sensor (the glutaredoxin 1 (Grx1)–roGFP2 fusion protein) during human cardiomyocyte differentiation⁸⁴.

Unresolved challenges exist for the routine deployment of GECIs and GEVIs and other optogenetic sensors in whole hearts, especially in vivo. Robust expression of the sensors without interference with native function is a desirable target; to date, the work has been primarily done in transgenic mice. Optical measurements are confounded by the mechanical motion of the heart, the dense muscle tissue and the high absorbance of haemoglobin^21,85. For in vivo measurements, ratiometric sensors or other methods for motion correction are necessary. Continuous monitoring of in vitro cardiac function is more straightforward. The optogenetic toolbox (actuators and sensors) is ever-expanding and efforts are focused towards the development of high-throughput discovery systems with rational design of features, as exemplified by some new GEVIs^76,81.

Comprehensive reviews on optogenetic sensors (GECIs and GEVIs) offer details on the spectral and biophysical properties of these proteins and their use in generating transgenic animals for research^{24,27,34,37,38,86-91}; those with specific relevance to the heart are shown in Table. 1.

Excitatory and inhibitory optogenetic actuators

Optogenetic actuators of voltage are light-sensitive proteins that can be expressed in mammalian cells and can transform photon flux into transmembrane ion flux, thereby manipulating transmembrane potential on a millisecond scale. The discovery and cloning of ChR2, a cation-selective, light-sensitive ion channel, in 2003¹⁴ led to the development of an extensive toolkit of optogenetic actuators (opsins). This toolkit includes depolarizing (excitatory) opsins (such as ChR2¹⁴, CheRiff⁸⁰, Crimson⁹² and ReaChR⁹³) and hyperpolarizing (inhibitory) opsins (such as BLINK-1⁹⁴, PAC-K⁹⁵, GtACR1⁹⁶, archaerhodopsin T (ArchT)⁹⁷, halorhodopsin (Halo)⁹⁸ and Jaws⁹⁹) that are activatable across a wide spectrum of wavelengths (Fig. 2b). All these optical actuators are microbial opsins that use all-trans retinal as a chromophore, with which the opsins form covalent bonds.

The most widely used optogenetic actuator is a larger-photocurrent H134R mutant of ChR2¹³. A vast number of other ChR2-based, depolarizing opsins have been developed to change their spectrum, kinetics or dynamic range^14,100,101. Fig. 2b shows the blue-shifted CheRiff⁸⁰, red-shifted Crimson⁹² and red-shifted ReaChR⁹³. These optogenetic actuators are both voltage-dependent and light-dependent, which means that they are non-linear transducers of light to voltage. These actuators have voltage rectification, that is, the ChR2 current is able to limit itself after the membrane potential becomes depolarized, even if the light remains on during longer pulses. Understanding the non-linearity of the function of ChR2-based opsins (through computer modelling¹⁰²) is essential. These properties (the self-closing through voltage rectification) can be beneficial for optimizing the energy for stimulation¹⁰³ and can be leveraged for more intelligent control of actuator function. Opsins can also be modified through rational design¹⁰⁴ or metagenomic discovery¹⁰⁵ to tailor them to cardiac research.

Of note, depending on irradiance and pulse duration, excitatory opsins can act as complex modulators of cardiac function, beyond simple excitation and pacing. As discussed later, excitatory opsins can modulate the shape of the action potential^106-108, as well as cardiac rhythm even when constant light is applied^48,103,109, and the conduction properties and wave dynamics^48,110-112 of cardiac tissue. Excitatory opsins can also provide inhibition of activity during strong illumination and long-duration pulses¹⁰⁹. Unlike electrical current input, optogenetic actuation can be long-lasting and can provide feedback-controlled, complex modulation over time^113,114 and over time and space^111,112.

Fast, bidirectional, optical modulation of voltage is particularly attractive, but the hyperpolarizing (inhibitory) opsins have been less robust than ChR2 and derivatives. Cardiac research has adopted the early tools ArchT⁹⁷ (a proton pump) and Halo⁹⁸ (a chloride pump). Fig. 2b shows Jaws⁹⁹ (a red-shifted version of a chloride pump), GtARC1⁹⁶ (an anion-selective ion channel), BLINK-1⁹⁴ (a potassium channel-based inhibitory opsin) and PAC-K⁹⁵ (a newly developed, two-component potassium channel system). The search for improved hyperpolarizing opsins continues because the currently available pump-based tools lack sufficient photocurrent; in particular, the GtARC1-based tools are highly dependent on the cellular chloride reversal potential and the potassium channel opsin systems are difficult to express in cells or cannot be deployed on the same millisecond time scale as their depolarizing counterparts.

Unique applications of voltage optogenetic actuators or other optogenetic actuators (that modulate signalling pathways) include mitochondria-specific targeting of ChR2, which led to the development of optometabolic tools for cardiac research¹¹⁵. Early deployment of optogenetic tools provided insights into L-type calcium channel clustering and its effects on channel gating¹¹⁶. Optogenetic actuators for G-protein-coupled receptor (GPCR) signalling (known as opto-XRs or opto-GPCRs) have been used for precise space–time manipulation of rhythm by very low light intensities via the G_q signalling pathway with the use of genetically expressed melanopsin¹¹⁷ or via G_s signalling with the use of JellyOp (an opsin photopigment) expression¹¹⁸. Further discussion of the variety and properties of the opto-XR or opto-GPCRs classes of optogenetic actuators can be found in previous reviews^38,119. Comprehensive reviews on optogenetic actuators and their biophysical properties are listed in Table 1^{20,38,41,100,101}

A large number of cardiac studies have used all-optical electrophysiology^{1,4,60,62,63,106,110,111,120-144} by combining optogenetic actuators with optogenetic sensors, synthetic voltage-sensitive and calcium-sensitive dyes, or dye-free optical methods to image cardiac activity. Computational modelling of cardiac optogenetics has complemented experiments by providing insights into optimal design and interpretation of findings in cardiac cells and tissues^{102,103,107,108,143,145-150} (summarized in Table 1).

High-throughput drug discovery

The drug-development process is lengthy and costly and relies on high-throughput screening for target identification, target validation, compound discovery and hit validation. All-optical electrophysiology^20,80,129 could augment each of these steps in the drug-development pipeline, as reviewed in detail previously³³. Fig. 3a illustrates how bidirectional, optogenetic voltage control (with the use of depolarizing ChR2 and hyperpolarizing ArchT opsins), combined with an optogenetic voltage readout (Quasar 1) can produce similar results to the classic (electrical) voltage clamp when studying ion channel responses to drugs⁶⁵. Considering the non-linear light–opsin photocurrent relationships, this voltage control applies for certain dynamic range, depending on the opsins. The main benefit of the optogenetic approach compared with the automated planar-patch systems, which underwent rapid development in the past decade, is the non-contact nature of actuation and sensing by light, lending this technique to scalability and enabling work with a variety of cell types. Ultimately this motivates rigorous experimental design with a greater number of testing conditions.

High-throughput cardiotoxicity testing

Preclinical testing for all drugs involves a cardiotoxicity assay, which includes the detection of electrophysiological abnormalities. This assessment is currently based exclusively on compound testing in heterologous systems for potassium voltage-gated channel subfamily H member 2 (also known as hERG) channel blocking, a prime target for drugs. However, ongoing efforts are underway to consider more comprehensive cardiotoxicity assays with the use of human iPSC-CMs¹⁵². Given that drug actions are rate-dependent, it is important to capture cellular electromechanical responses under controlled, paced conditions. Electrical pacing is not easily scalable, therefore contact-free optical methods offer an attractive alternative. Similarly, optical sensing has proved to be valuable in increasing throughput as well as enabling long-term observations (Fig. 3b). The utility and limitations of all-optical electrophysiology (optogenetic actuators and optogenetic sensors) to address the needs of drug testing was investigated computationally and experimentally by Klimas and colleagues¹²⁹. An alternative implementation of optogenetic pacing without genetically modifying the cardiomyocytes is to use dedicated, light-sensitive, non-excitable cell constructs that can couple to cardiomyocytes, known as the tandem cell unit [G] approach⁴. To date, experimental studies have not reported interference with native electrophysiology or notable limitations of optogenetic tools in drug testing.

After early reports on the use of optogenetic pacing of stem-cell-derived cardiomyocytes^153,155, several groups focused on developing automated high-throughput all-optical platforms for cardiotoxicity testing. Optopatch, a fully optogenetic construct (GEVI–GECI CaViar sensor in tandem with an optogenetic actuator CheRiff) was developed by Cohen and colleagues⁸⁰ and combined with high-end macroscopic systems for cardiotoxicity testing with the use of human iPSC-CMs^60,133 (Fig. 3b). Variants of the system have been translated to an industrial setting for drug discovery¹⁴⁰. Our group developed a simpler approach for automated, all-optical, dynamic cardiac electrophysiology (OptoDyCE)^62,129, which integrates an optogenetic actuator (ChR2) with spectrally compatible, high-performance, synthetic dyes for calcium and voltage or red-shifted, optogenetic calcium sensors (Fig. 3b). The system is low-cost, using LEDs and temporal multiplexing to simultaneously record voltage and calcium responses or cell contraction with a single camera. The high-content, drug-response measurements in human iPSC-CMs have been used to construct populations of human iPSC-CM computer models for personalized medicine¹⁶⁴. Such work has been extended to include long-term, high-throughput microfluidic perfusion⁶³ and to allow for long-term pacing and recording to enable chronic drug testing, while other research groups used some of these all-optical electrophysiology tools for iPSC-CM phenotyping and drug screening^61,157. Some studies have specifically leveraged optogenetic pacing, pairing it with high-throughput microelectrode recordings to capture drug responses^158,159,161. These approaches have sparked interest from industry and the FDA and are likely to influence drug development in the near-term.

Short-term and long-term actuation and sensing in human iPSC-CMs

Human iPSC-CMs have transformed cardiac research, but concerns about the immature state of the iPSC-CMs have prevented their rapid translation into the clinic¹⁵¹. Optogenetic tools can have a crucial role in phenotyping, long-term monitoring and maturation of these cells. The main advantage of optogenetic sensors compared with synthetic voltage or calcium dyes is the opportunity to acquire repeated measurements over long periods of time. Since 2014, multiple studies have used iPSC lines with constitutively expressed GEVI ArcLight to monitor cell changes during the differentiation and maturation into cardiomyocytes and cellular responses to pharmacological agents^64,70-73,154. Both control cells and cells from patients with inherited arrhythmogenic disorders have been characterized in detail with the use of ArcLight imaging^71,72, GCaMP calcium monitoring⁷² and with a combination of ArcLight and a GECI (R-GECO)⁶⁴. Over the past 5 years, efforts involve linking phenotypes to single-cell transcriptomic signatures⁷⁰ and leveraging optogenetics to monitor redox states in engineered human tissues⁸⁴.

Optogenetic actuation has been applied in creative ways to address iPSC-CM immaturity. Fig. 4a illustrates work by Quach and colleagues to transform a classic (electrical) dynamic clamp assay into an optical dynamic clamp [G] to yield a more hyperpolarized and mature resting membrane potential in human iPSC-CMs¹¹³, to be closer to that of an adult cardiomyocyte. Viral expression of ArchT in human iPSC-CMs and computer-controlled, real-time manipulation by light (LED) was used to perform ‘optical injection’ of the inward-rectifier potassium current (I_k1) to make the resting membrane potential more negative, thereby shifting the electrophysiological phenotype of human iPSC-CMs to be more adult-like. Furthermore, a study from 2020 applied the same approach (optical dynamic clamp) to inject a synthetic, antiarrhythmic, frequency-dependent current in atrial cardiomyocytes¹¹⁴ Extending this idea to an all-optical clamp would enable multicellular and tissue-level applications. However, more work is needed to calibrate voltage responses and to implement real-time processing for multiple cells.

Dynamic stimulation (electrical or mechanical) of human iPSC-CMs has been identified as a potential approach to induce a more mature cellular phenotype. Several studies in 2019 leveraged optogenetic actuators to explore chronic optical stimulation^162,163 (Fig. 4a). One study used closed-loop control for overdrive-pacing of human iPSC-CMs during weeks in culture and quantified cell remodelling¹⁶². Another study used optogenetic pacing of engineered human tissues over 3 weeks and monitored their contractile performance¹⁶³. The stimulated tissue constructs showed electromechanical remodelling, with faster contractions, shorter action potentials and lower L-type calcium current than unstimulated samples. The optically tachypaced engineered human tissues were more susceptible to arrhythmias¹⁶³. Overall, optogenetics enables scale up and accelerates the development of more mature and physiologically-relevant human iPSC-CMs for drug screening and for applications in regenerative medicine.

Applications for disease modelling and personalized medicine

Optogenetic tools have been used in several studies on patient-specific disease modelling with human iPSC-CMs. For example, optogenetic sensors (GEVI and GECI) were used to monitor responses in Timothy syndrome, a calcium-channel genetic disorder⁶⁴. Another study used control iPSC-CMs and iPSC-CMs derived from patients with catecholaminergic polymorphic ventricular tachycardia or long QT syndrome type 2 to phenotype and test the iPSC-CMs for pharmacological interventions⁷². A 2019 study leveraged optogenetic pacing to phenotype catecholaminergic polymorphic ventricular tachycardia subtypes with the use of engineered human tissues¹⁶⁰. Fig. 4b shows an example of an all-optical electrophysiology platform used to phenotype a rare genetic disorder with cardiac abnormalities¹⁵⁶. Ogden syndrome¹⁶⁵ involves a point S37P mutation in NAA10 (encoding the catalytic subunit of N-α-acetyltransferase 10, responsible for the N-terminal acetylation of certain proteins). Patients with Ogden syndrome die young and some develop torsade de pointes, which is an extremely rare presentation of abnormal heart rhythm in children. Overall, the ability to phenotype patient-specific iPSC-CMs quickly and to quantify cellular responses to therapeutic interventions is at the core of personalized medicine. Optogenetic tools offer an enabling technology for such interventions and might help to provide better stratification during preclinical testing of drugs, taking into account patient sex, race, age and other aspects through increased throughput and improved experimental design.

Arrhythmia management

Implantable electronic pacemakers and implantable cardioverter–defibrillators are the gold-standard for arrhythmia management¹⁶⁶. Despite their established value as safe devices, they still have substantial deficiencies. In particular, high-voltage shocks damage myocardial tissue, and device battery life is limited. Batteries are drained when shocks are frequent, such as in patients with ventricular electrical storm¹⁶⁷. Shocks are also painful, causing patient anxiety and depression¹⁶⁸. Optogenetics has the potential to address these deficiencies by offering routes for non-electrical, low-energy pacing and cardioversion that can be cell-specific and painless.

Optogenetic pacing for rhythm control

Optical pacing would be an attractive alternative to electronic pacemakers if, by leveraging the advantages of optogenetics (Fig. 2a), this approach could be achieved at lower energy, thereby extending battery life^20,22, while providing the additional advantages of cell specificity and distributed actuation. In 2010–2011, the first reports demonstrated the utility of optogenetic pacing of cardiomyocytes with the use of blue-light pulses^1-4. Over the past decade, multiple other studies have used optogenetic pacing in cultured cell systems in vitro, in perfused hearts ex vivo and in anaesthetized animals^{60-63,91,102,106,124-126,128,129,132,133,135,141,143,153,156-159,161-163,169-178} (Table 1). The hearts of transgenic species have been optically paced, including zebrafish, Drosophila and mice, as well as mouse and rat hearts virally transduced to express optogenetic actuators. Although long-term optogenetic pacing with an implantable device has not yet been demonstrated in an awake animal, in this section, we summarize crucial advances towards this goal.

A study in 2010 was the first to use optogenetics to pace the mammalian heart with the use of light² (Fig. 5a). Transgenic mice expressing ChR2 within the cell membrane of cardiomyocytes enabled optical pacing of the ventricles and atria at physiologically relevant rates with the use of short (<10 ms) pulses and well-tolerated light levels. Prolonged depolarization was observed during longer pulses, both in vitro and in mouse hearts in vivo, suggesting that optical termination of arrhythmias was possible, as demonstrated later by this group and others^130,131,179. Among the pioneering studies of optogenetic pacing, Arrenberg and colleagues demonstrated cell-specific expression of opsins within the specialized conduction system of zebrafish hearts¹. They used transgenically expressed halorhodopsin and ChR2 either to optically block intrinsic activation or to optically pace zebrafish hearts at specific rates. Using patterned illumination to switch rapidly between a normal activation sequence and an arrhythmia, they demonstrated for the first time the utility of optogenetic tools for spatiotemporal control of cardiac activity. Their optogenetic interrogation of zebrafish hearts provided new insights into endogenous sinus node electrophysiology.

Clinical deployment of optogenetic pacing will involve methods beyond transgenic approaches, such as gene therapy, because viral vectors (including lentiviruses, adenoviruses and the safer adeno-associated viruses (AAVs)) provide the most efficient way to inscribe light sensitivity to the heart and are suitable for clinical use^{102,106,155,170,176}. Experiments in 2015 demonstrated how discrete regions of the adult rat heart could be enabled for optical pacing by local ChR2–AAV9 viral injections¹²⁸ (Fig. 5b). Several weeks after the injections, hearts were responsive to optical pacing in both open-chest and ex vivo studies. Multisite optical pacing further demonstrated the application of optogenetics for ventricular resynchronization therapy¹²⁸. Another study from 2015 demonstrated optogenetic transduction of the heart that did not require open-chest surgery and myocardial injections. Instead, ChR2–AAV9 was systemically delivered to mice¹⁷¹, capitalizing on the cardiac tropism of this AAV serotype and achieving robust myocardial responsiveness to optical pacing a year after transduction.

Delivery of non-excitable, light-sensitized cells could be an alternative to viral transduction for optical cardiac pacing. In 2011, Jia and colleagues termed this approach the tandem cell unit strategy^4,124 and validated it in vitro using ChR2–HEK cells coupled to either adult dog cardiomyocytes or monolayers of neonatal rat cardiomyocytes. In a hypothetical tandem cell unit strategy, patient-derived cells would be induced to express an opsin to provide robust depolarizing current upon photoactivation. After delivery to a patient’s heart, those cells would couple to cardiomyocytes to enable optical control of heart rhythm. Although exogenous delivery of light-sensitized cells has not yet been implemented, in vitro and in vivo proof-of-concept results of the tandem cell unit concept have been reported for non-excitable cells of the heart, including cardiac fibroblasts, myofibroblasts^{132,143,169,180,181}, cardiac progenitor cells¹³² and macrophages¹⁸², all of which can electrically couple to cardiomyocytes to entrain the myocardium after optical stimulation.

Low-energy optogenetic pacing could emerge as a disruptive technology if optogenetic transduction and light delivery were safe and minimally invasive^20,22,23. Strategies for translating optogenetic pacing to the clinic include obtaining a better understanding of opsin operation and function in cardiomyocytes (atrial, ventricular and Purkinje cells)^{102,103,146,175}, optimizing optogenetic actuators for cardiac cells, obtaining a deeper understanding of cell-resident cofactors (for example, all-trans retinal at the right doses considerably lowers the threshold for stimulation)¹⁷³ and selectively engaging the specialized conduction system of the heart^1,126,146. Computational modelling is an effective platform for testing these strategies and has provided new insights into the energetic benefits of longer stimulation pulses^102,103 and constant ultra-low illumination¹⁰³. Such long-duration pulses are not feasible using electrical stimulation owing to the generation of cytotoxic electrochemical by-products^103,183. Results from human heart computational models have revealed that the energy required to pace the heart is substantially reduced when ChR2 is exclusively expressed within the Purkinje system and light is delivered to the bundle of His¹⁴⁶. Cell delivery of designated, non-excitable ‘spark’ cells, as in the tandem cell unit approach, could also lower the energy required to pace the heart if the cells are clustered¹²⁴. Finally, low-energy solutions for rhythm control might also arise from optogenetic modulation of intracellular GPCR signalling, such as the G_q¹¹⁷ and G_s¹¹⁸ pathways.

Optogenetic termination of cardiac arrhythmias

The critical dependence on coordinated space–time actuation is a major challenge in arrhythmia termination. Focused actuation in either space or time alone is often inadequate for termination of complex arrhythmias. Conventional defibrillation is a high-energy termination strategy that depolarizes a large myocardial mass (>95%)¹⁸⁴ to block fibrillatory wavefronts, thereby preventing the formation of new re-entrant waves^185,186. Low-energy, electrical strategies rely on the proper timing of low-amplitude depolarizing stimuli to extinguish arrhythmic wavefronts and include anti-tachycardia pacing¹⁸⁷ and low-energy anti-fibrillation pacing¹⁸⁸. Anti-tachycardia pacing is a feature in current implantable cardioverter–defibrillators but is limited to actuation of specific locations and, therefore, is effective in terminating only non-complex, re-entrant arrhythmias. By contrast, optogenetics enables low-energy strategies that can be distributed over large areas or deliver long pulses, capturing multiple phases of the arrhythmic activity or having light that is patterned, possibly in real-time, for mechanistic termination of fibrillatory wavefronts. The light intensity could also be titrated to terminate rotors with the use of subthreshold illumination⁴⁸. By avoiding engagement of the cardiac nociceptors and preventing skeletal muscle hypercontractures, the cell-specificity of optogenetics ensures painless termination of an arrhythmia.

In an early study of optical cardioversion to terminate re-entrant arrhythmias in healthy mouse hearts, illumination of the anteroseptal epicardium with one light pulse had an 85% success rate¹⁴⁷ (Fig. 5c), and the success during acute infarction was 88%. In computer simulations of an infarcted human heart, epicardial illumination successfully terminated infarct-related ventricular tachycardia by initiating transmural depolarization to block the re-entrant wavefront¹⁴⁷. In a later study, ChR2–AAV9 constructs were used in combination with rapamycin (to suppress immune response) in rats to demonstrate ventricular cardioversion¹³⁰. Application of a single epicardial light pulse achieved 97% cardioversion for monomorphic ventricular tachycardia and 57% for polymorphic ventricular tachycardia. Another study also successfully terminated ventricular tachycardia by open-chest application of successive light pulses in rats after myocardial infarction¹⁸⁹.

An ex vivo study from 2018 found that ventricular tachycardia was terminated in transgenic ChR2-expressing mice after global illumination of the epicardial surface with a single light pulse¹³⁶. Optical mapping revealed two primary cardioversion mechanisms that were motivated by optical depolarization of the entire epicardium; the termination of a re-entrant core and subsequent refractoriness, and an unpinning and drift of the re-entrant core until collision with refractory tissue¹³⁶. Although global epicardial depolarization by light excels as a low-energy solution to cardioversion, it is not easy to implement in vivo where it is necessary to perform cardioversion with minimal area of illumination. A study in ChR2-transgenic mouse hearts addressed this issue using patterned light to annihilate arrhythmic wavefronts using mechanistic, multi-barrier illumination patterns¹³¹ (Fig. 5d). Cardioversion [G] of ventricular tachycardia required triple-barrier optical depolarization, where light was restricted to three lines oriented approximately 120° apart. This illumination pattern yielded an optical cardioversion success rate similar to that of illuminating the entire left ventricle, but used only 25% of the light energy.

Atrial cardioversion is a mid-future opportunity for translating cardiac optogenetics to the clinic. After early proof-of-principle in vitro work in monolayers and slices of rat atrial cardiomyocytes^122,175,190, followed by computational studies¹⁴⁹, two rodent studies demonstrated optogenetic termination of atrial fibrillation ^179,191. One study used mice with cardiac expression of ChR2 as well as a loss-of-function mutation in Cx40 that promoted atrial fibrillation¹⁷⁹. In open-chest experiments, the success rate of terminating atrial fibrillation was >90% using focused, low-light illumination (<0.5 mW/mm²). Optical atrial cardioversion was still possible 8 months after systemic viral delivery of ChR2, showing promise for the clinical translation of this approach. In other studies in rats, ReaChR–AAV9 with the atria-specific promoter from human NPPA was topically applied to the right atrium using the gene painting [G] technique¹⁹¹. After 4 weeks, cholinergic-mediated atrial fibrillation was induced in the anaesthetized animals and a computer-controlled, implanted LED delivered optical pulses to the right atrium when PR intervals consistent with atrial fibrillation were detected, providing a cardioversion success rate of 96%¹⁹¹ (Fig. 5e).

Light penetration into the myocardium is an important consideration for the clinical translation of optogenetic cardioversion or defibrillation, especially for the ventricles⁸⁵. Illumination with longer-wavelength NIR light, in combination with red-shifted opsins, would provide deeper myocardial penetration of optical depolarization, thereby increasing the probability of capturing a critical mass of tissue. Upconversion nanoparticles [G] administered deep within the myocardium could also be engaged by NIR light, enabling the activation of optogenetic actuators via nanoparticle-assisted light conversion¹⁹². These innovations could be optimized for the clinic using computational models^28,147-149. Interestingly, results from a recent study suggest that simple photon transport models could underestimate the depth of light penetration into the myocardium and its effect on opsin-expressing cells¹⁹³. Additional work is certainly needed to better understand light–matter interactions within the context of using distributed light for optogenetic cardioversion.

Cardioversion using well-controlled, regional excitation or inhibition is difficult to achieve using electric fields but is a unique feature of optogenetics. Indeed, several studies have used hyperpolarizing opsins to modulate cardiac activity^{95,96,169,194-196}. However, some hyperpolarizing opsins, especially anion-conducting opsins, such as ACR, have had controversial results in cardiac muscle¹⁹⁴. Furthermore, termination of complex arrhythmias using optical hyperpolarization would require global actuation (expression and illumination) to prevent enduring wavefronts from re-activating hyperpolarized regions. One study used transgenic mice expressing the hyperpolarizing opsin ArchT to optically terminate ventricular arrhythmias¹⁹⁵. Although cardioversion success was higher than in control hearts, actuation of the inhibitory opsin was inferior to actuation of an excitatory opsin, such as ChR2. A comprehensive list of studies using optogenetic tools for cardioversion or defibrillation^{48,95,96,111,112,114,121,122,128,130,131,136,137,147,149,174,179,189-191,194-199} is shown in Table 1.

Excitation wave control and feedback for rhythm management

The optogenetic approaches described above for pacing and cardioversion or defibrillation are guided by concepts used in present-day cardiac devices. Optogenetics also enables new innovations for rhythm management, including the control of action potential morphology and control of the speed and direction of targeted wavefronts, a concept known as optical wave steering [G]²⁹. These rhythm-management innovations could neither be realized using present-day devices nor be rigorously tested in living tissue before the advent of cardiac optogenetics. Indeed, multiple experiments have demonstrated that wavefronts can be modulated by light^{48,110-112,131,137-139,143,181,190,199,200} (Fig. 6a,b), among which some have implemented quasi-real-time, closed-loop control whereby the trajectory of a wavefront is sensed and then altered using patterned illumination (Fig. 6c). This approach is particularly promising for ultra-low-energy cardioversion. However, the implementation could be challenging due to the substantial spatiotemporal data that would need to be processed in quasi-real-time by an implanted device. Even so, closed-loop sensing and actuation of cellular function is a powerful feature of optogenetics that is useful not only in cardiology but also neuroscience²⁰¹.

An early experiment demonstrated optical wave steering in cardiomyocyte monolayers by combining dye-free optical imaging for wave sensing with patterned light for localized optogenetic actuation¹¹¹ (Fig. 6a). A digital micromirror device was driven by sequences of programmed images to project light for the modulation of conduction velocity, to induce unidirectional block, to initiate rotors and to instantaneously reverse the chirality of rotors. This study demonstrated how optical wave steering could provide insights into a range of excitable tissue disorders, including functional and anatomical re-entry^29,111. In other experiments, opsins were expressed in non-cardiomyocytes, including 3T3 fibroblasts¹⁸¹, cardiac myofibroblasts¹⁴³ and stellate ganglion neurons¹¹⁰, to modulate conduction speed. Excitation wave dynamics have also been studied and manipulated within geometrically patterned cultures of human iPSC-CMs and engineered excitable cells using all-optical sensing and actuation¹³⁸.

Optical wave steering using patterned light is particularly useful for investigating and implementing mechanism-based termination of re-entry. For example, mechanisms of atrial rotor termination have been revealed by illuminating ChR2-expressing rat atrial myocyte layers or slices with specific patterns of light¹⁹⁰. Stable rotors were extinguished when an optically positioned, temporary conduction block intersected both the rotor core and an unexcitable boundary. A 2018 study used optogenetics to guide spiral waves towards termination along well-defined trajectories by optically positioning functional conduction blocks with high spatiotemporal precision¹¹² (Fig. 6b). Results from these mechanistic studies and other studies using optogenetics^111,136,190 have provided new insights into guided rotor termination. Extending this approach to perfused, ChR2-expressing mouse hearts, Scardigli and colleagues designed and tested a novel system engineered for real-time, optical sensing and stimulation to control excitation waves¹³⁷ (Fig. 6c). A region-averaged fluorescence signal of electrical activity was monitored, and patterned light pulses were applied to the epicardium when a threshold was exceeded. The system reacted to non-stationary wave dynamics within 2 ms. This system could accelerate bundle-branch-mediated ventricular depolarization, restore atrioventricular conduction after atrioventricular node inhibition, and generate stable, apex-to-base re-entrant circuits, whereby sensing at the base and stimulation at the apex provided a virtual excitation pathway^137,139.

Although these innovations for rhythm control have been informed by cardiac optogenetics, their clinical implementation does not necessarily require optogenetics. Some control strategies could be deployed by distributed arrays of electrodes, but optical methods remain superior in their spatiotemporal capacity for precise wave control. Furthermore, optogenetic tools developed for research, such as the optical dynamic clamp¹¹³, have enabled experimental testing of novel in silico strategies to design synthetic antiarrhythmic ion channels that have gating properties that are modulated by the rate of electrical excitation, thereby enabling the potential for fully autonomous in-situ termination of arrhythmias¹¹⁴.

Leveraging cell-specific control of cardiac function

Cardiac function cannot be controlled in a cell-specific manner using electrical current. Pharmacological agents provide cellular specificity but have inferior resolution of spatiotemporal actuation. Optogenetics uniquely provides both the spatiotemporal resolution of cellular activation, or inhibition, as well as the cell and organelle specificity necessary to bring about new cardiac therapy paradigms (Fig. 7).

Cre–lox recombination and cell-specific promoters are used to target opsin expression within specific cell types²⁰², for example, macrophages (Fig. 7a). In a Cre–loxP transgenic approach, a Cre-responder animal would have a loxP-flanked STOP cassette upstream of the opsin gene, preventing its expression. This animal would be crossed with a Cre driver animal expressing Cre recombinase under the control of a cell-specific promoter, such as the Myh6 promoter for cardiomyocytes^126,131, the Th promoter for sympathetic neurons^127,203 or the Chat promoter for parasympathetic neurons^142,204. In the mice progeny, the opsin would then be expressed only within cells in which the promoter is active. Alternatively, with viral transduction, cell specificity is determined by the site of delivery (if localized), by tissue tropism of the viral vector and by the promoter used to drive expression. For example, for both localized and systemic vector delivery, the use of ubiquitous promoters, such as the cytomegalovirus or the CAG promoters, in combination with the cardiotropic AAV9 vector result in preferential ventricular myocyte transduction, with much lower expression in atrial myocytes and nonmyocytes. To drive more robust atrial myocyte expression, one study used local gene painting and the NPPA promoter¹⁹¹ (Fig. 5e).

A 2015 study used transgenic mice and Cre–lox recombination to express ChR2 in either cardiomyocytes (Myh6 promoter) or cells of the specialized conduction system (Cx40 promoter)¹²⁶ (Fig. 7b). This cell-specific expression provided insight into the arrhythmogenic role of the myocardium and the specialized conduction system during local ischaemia. In open-chest experiments, optical stimulation (photostimulation) of the cardiomyocytes triggered sustained arrhythmias after coronary artery ligation, but photostimulation of the specialized conduction system did not. Similar viral vector-based transduction of the Purkinje network could yield low-energy optogenetic pacing¹⁴⁶ (Fig. 7c), as discussed above. In addition, specific cell populations or myocardial regions prone to arrhythmogenic activity could be optically targeted to suppress arrhythmias.

Neurocardiology is a classic area that has benefited from the cell-specificity of optogenetics. The complexity of the intrinsic cardiac autonomic network has been revealed using Cre–lox recombination in transgenic mice or viral delivery of fluorescent protein vectors, with impressive resolution¹⁴², particularly after tissue clearing²⁰⁴. Neuron-specific expression of opsins enables precise functional mapping of cardioneural circuits, similar to what has been done for the brain²⁰⁵. Original studies in this area used transgenic mice with ChR2 expression targeted to either catecholaminergic neurons (using a Th promoter)¹²⁷ or cholinergic neurons (using a Chat promoter)¹⁴². Selective photostimulation of intrinsic catecholaminergic neurons in perfused mouse hearts initiated sudden and dramatic increases in heart rate and contractility¹²⁷ (Fig. 7d), consistent with downstream activation of β-adrenergic pathways and increases in cytosolic cAMP levels²⁰³.

Another study used a silicon-encapsulated micro LED placed on the right atrium of perfused mouse hearts to photostimulate cholinergic neurons¹⁴² (Fig. 7d). The RR interval quickly increased after illuminating the right atrium, consistent with activation of the G_αi/o-coupled cholinergic M₂ muscarinic receptors of sinus node cells²⁰⁶. LED positioning altered the responses between heart rate slowing without atrioventricular delay and atrioventricular delay without heart rate slowing, confirming a cholinergic neural network whereby the sinoatrial node and the atrioventricular node are innervated by distinct axons^207,208. Photostimulation also maintained substantially low heart rates for up to 30 min of continuously pulsed illumination¹⁴², indicating that continuous photostimulation of the cholinergic network could be an approach for in vivo heart rate control. In other studies, photostimulation of autonomic neurons in the stellate ganglion and the vagus nerve increased and decreased heart rate, respectively²⁰⁴ (Fig. 7e), supporting efficacious optical neuromodulation for the treatment of rhythm disorders²⁰⁹. Beyond rhythm control, in vitro studies with co-cultured cardiomyocytes and optogenetically enabled sympathetic neurons have revealed the effects of neuron–cardiomyocyte interactions on electrical wave dynamics¹¹⁰ as well as the effect of cardiomyocytes on neuronal maturation²¹⁰.

Optogenetic suppression of sympatho-mediated ventricular arrhythmias has been reported in canine studies of acute ischaemia²¹¹. AAV9–ArchT was injected locally in the left stellate ganglion neurons, which were subsequently illuminated by an LED for an hour after coronary artery ligation. The resulting neuronal hyperpolarization greatly reduced the incidence of ventricular tachycardia and fibrillation²¹¹, providing proof-of-concept results of optogenetic arrhythmia therapy in a large-animal model. Earlier studies expressed excitatory opsins within specific neural populations of the cardiac centres of the brain stem. Those neural populations were then photostimulated to provide myocardial protection during ischaemia–reperfusion injury in rats²¹² and to increase exercise capacity in healthy rats²¹³. Optogenetic neuromodulation to improve cardiac function and the associated devices that would deliver light to the cardiac ganglia and peripheral nerves, including the vagus nerve, are important mid-term future translational opportunities. Cell-specific control provided by optogenetics currently does not exist for electrical neuromodulation therapies, such as electrical stimulation of the vagus nerve.

Any cardiovascular cell type could likely be inscribed with light-sensitivity, especially as new Cre-expressing animal models become available and viral vectors are optimized. For example, one study used Cre–lox recombination in mice to target the expression of ChR2 to resident cardiac macrophages with the use of the Cx3cr1 promoter¹⁸² (Fig. 7a). Atrioventricular conduction was improved after macrophage photostimulation, underscoring the importance of connexin 43 coupling between macrophages and cardiomyocytes for normal atrioventricular node conduction. Other studies used targeted optogenetic sensor expression to assess the electrotonic coupling of excitable and non-excitable cells⁷⁴ and to modulate coronary artery vasomotor tone²¹⁴. A comprehensive list of studies that have applied optogenetic tools for cell-specific control of neurocardiac function^{110,127,142,203,204,210-213,215} and control of myocardial function via non-myocyte cells^{1,4,39,74,124,126,132,143,146,169,180-182,191,214,216-218} is provided in Table 1.