Chronic Fatigue Syndrome: Exercise Performance Related to Immune Dysfunction

CLINICAL SCIENCES Clinical Investigations Chronic Fatigue Syndrome: Exercise Performance Related to Immune Dysfunction JO NIJS1,2, MIRA MEEUS1,2, NEIL R. MCGREGOR3, ROMAIN MEEUSEN1, GUY DE SCHUTTER1, ELKE VAN HOOF1, and KENNY DE MEIRLEIR1 1 Department of Human Physiology, Faculty of Physical Education and Physiotherapy, Vrije Universiteit Brussel, Brussels, BELGIUM; 2Division of Musculoskeletal Physiotherapy, Department of Health Sciences, Hogeschool Antwerpen, Antwerp, BELGIUM; and 3Bio21, Institute of Biomedical Research, University of Melbourne, Parksville, Victoria, AUSTRALIA ABSTRACT NIJS, J., M. MEEUS, N. R. MCGREGOR, R. MEEUSEN, G. DE SCHUTTER, E. VAN HOOF, and K. DE MEIRLEIR. Chronic Fatigue Syndrome: Exercise Performance Related to Immune Dysfunction. Med. Sci. Sports Exerc., Vol. 37, No. 10, pp. 1647–1654, 2005. Purpose: To date, the exact cause of abnormal exercise response in chronic fatigue syndrome (CFS) remains to be revealed, but evidence addressing intracellular immune deregulation in CFS is growing. Therefore, the aim of this cross-sectional study was to examine the interactions between several intracellular immune variables and exercise performance in CFS patients. Methods: After venous blood sampling, subjects (16 CFS patients) performed a maximal exercise stress test on a bicycle ergometer with continuous monitoring of cardiorespiratory variables. The following immune variables were assessed: the ratio of 37 kDa Ribonuclease (RNase) L to the 83 kDa native RNase L (using a radiolabeled ligand/receptor assay), RNase L enzymatic activity (enzymatic assay), protein kinase R activity assay (comparison Western blot), elastase activity (enzymatic– colorimetric assay), the percent of monocytes, and nitric oxide determination (for monocytes and lymphocytes; flow cytometry, live cell assay). Results: Forward stepwise multiple regression analysis revealed 1) that elastase activity was the only factor related to the reduction in oxygen uptake at a respiratory exchange ratio (RER) of 1.0 (regression model: R2 ⫽ 0.53, F (1,14) ⫽ 15.5, P ⬍ 0.002; elastase activity P ⬍ 0.002); 2) that the protein kinase R activity was the principle factor related to the reduction in workload at RER ⫽ 1.0; and 3) that elastase activity was the principle factor related to the reduction in percent of target heart rate achieved. Conclusion: These data provide evidence for an association between intracellular immune deregulation and exercise performance in patients with CFS. To establish a causal relationship, further study of these interactions using a prospective longitudinal design is required. Key Words: EXERCISE PHYSIOLOGY, IMMUNITY, ELASTASE, PROTEIN KINASE R P revious research has shown that patients with chronic fatigue syndrome (CFS) present with an abnormal exercise response and exacerbation of symptoms after physical activity. Some of the main findings were a reduction in peak oxygen uptake (2,10), reduction in peak heart rate (10), earlier exhaustion (10), and accelerated glycolysis with increased lactate production (30). Contrary to these findings, others found that the aerobic capacity of CFS patients lies within the low normal range (23). The highly heterogeneous nature of the CFS population and the lack of uniformity in both the utilized diagnostic criteria and exercise testing protocols preclude pooling of data and hence to draw firm conclusions. Still, we conclude that at least a subgroup of CFS patients present with an abnormal response to exercise. In addition, because several exercise capacity variables (e.g., functional aerobic impairment, body weight–adjusted peak oxygen uptake, exercise duration) correlated with activity limitations/participation restrictions (20), evidence supporting the clinical importance of impairments in exercise performance fitness in CFS patients was provided (i.e., a poor exercise performance was associated with more severe activity limitations/participation restrictions). Importantly, the exacerbation of symptoms after exercise is seen only in the CFS population, and not in fatigue-associated disorders such as depression, rheumatoid arthritis, systemic lupus erythematosus, or multiple sclerosis (26). To date, the exact cause of the abnormal exercise performance in CFS remains to be elucidated. Earlier attempts revealed that in CFS patients kinesiophobia Address for correspondence: Dr Jo Nijs, SPORT KRO-1 V.U.B., Laarbeeklaan 101, B-1090 Brussels, Belgium; E-mail: Jo.Nijs@vub.ac.be. Submitted for publication January 2005. Accepted for publication May 2005. 0195-9131/05/3710-1647/0 MEDICINE & SCIENCE IN SPORTS & EXERCISE® Copyright © 2005 by the American College of Sports Medicine DOI: 10.1249/01.mss.0000181680.35503.ce 1647 (irrational fear of movement) is not related to exercise performance (18), and that an exercise challenge further enhances complement activation (26). Type I interferons trigger the 2=,5=-oligoadenylate (2-5A) synthetase/Ribonuclease (RNase) L activation and induce the expression of the double-stranded RNA dependent protein kinase R (PKR). The PKR enzyme and 2-5A synthetase/RNase L system are termed the “cellular doublestranded RNA-detecting systems” that are responsible for the translational inhibition in response to (viral) infection (11). The deregulation of the 2-5A synthetase/RNase L pathway in subsets of CFS patients has been reported at length in the scientific literature (3,27,28). Both elastases and calpain are capable of initiating high molecular weight RNase L (83 kDa) proteolysis, generating two major fragments with molecular masses of 37 (a truncated low molecular weight RNase L) and 30 kDa, respectively (5). Experimental data point to an activation of the PKR enzyme, parallel to the 83 kDa RNase L proteolysis, in subsets of CFS (8). PKR activation leads to phosphorylation of the inhibitor of NF(nuclear factor)-kB (IkB) and consequent NF-kB activation, which in turn causes inducible nitric oxide synthetase (iNOS) expression. iNOS generates increased production of NO by monocytes/macrophages. NO mediates important vital physiological functions such as neurotransmission, cell-mediated immune responses (strong antimicrobial and antitumour activities), and vasodilatation. Excessive and/or persistent production of NO, however, is detrimental to the body’s functions (21). Elevated NO has been documented in CFS patients (14). Elevated NO levels and consequent vasodilatation might limit CFS patients to increase blood flow during exercise, and might even cause and enhance postexercise hypotension (19). It is hypothesized that PKR activation and consequent elevated NO levels are related to poor exercise performance in CFS patients. Snell and colleagues (25) showed that CFS patients with evidence of a deregulated 2=,5=-oligoadenylate (2-5A) synthetase/RNase L pathway have a lower peak oxygen uptake than CFS patients without the intracellular immune deregulation, suggesting a link between immunopathology and exercise performance in CFS. As outlined previously (19), the deregulation of the 2-5A synthetase/RNase L pathway may be related to a channelopathy, capable of initiating both intracellular hypomagnesaemia in skeletal muscles and transient hypoglycemia. This might explain muscle weakness and the reduced peak oxygen uptake seen in CFS patients. Thus, it is hypothesized that various components of the 2-5A synthetase/RNase L pathway (i.e., the ratio of 37 kDa RNase L to the 83 kDa native RNase L for the assessment of 83 kDa RNase L proteolysis, RNase L enzymatic activity, and elastase activity) are related to exercise performance in CFS patients. Summarizing the research questions, this study aims at 1) examining whether PKR activation and consequent elevated NO levels predict poor exercise performance in CFS patients, and 2) examining whether exercise performance in CFS is associated with deregulation of the 2-5A synthetase/RNase L pathway (i.e., 83 kDa RNase 1648 Official Journal of the American College of Sports Medicine L proteolysis, RNase L activity, and elastase activity). It is hypothesized that in CFS patients, 1) both PRK activation and consequent elevated NO levels predict poor exercise performance during a graded exercise cycle test, and 2) deregulation of the 2-5A synthetase/RNase L pathway is associated with poor exercise performance during a graded exercise cycle test. METHODS Patient recruitment and research design. Sixteen randomly allocated untrained patients with CFS were enrolled. Patients were randomly allocated from consecutive referrals to our chronic fatigue clinic. To be included into the study, patients had to fulfill the U.S. Centers for Disease Control and Prevention criteria for CFS (9). Therefore, all patients underwent an extensive medical evaluation before study participation (see below). All patients had Dutch as their native language, and were within the age range of 18 – 65 yr. The study sample consisted of eight female and eight male CFS patients. The mean age was 38 ⫾ 10 yr (range 19 –59), and the mean illness duration was 31 ⫾ 11 months (range 12– 48). An information leaflet was handed out to all patients, and they were instructed to read it carefully and, if applicable, to ask for additional clarification. The study protocol was approved by the local ethics committee (Academical Hospital Vrije Universiteit Brussel; O.G. 016). Patients provided written informed consent and underwent venous blood sampling (arm vein; 40 mL; lying supine). The following immunological variables were assessed: the ratio of 37 kDa RNase L to the 83 kDa native RNase L, RNase L enzymatic activity, protein kinase R activity assay, elastase activity, the percent of monocytes, and nitric oxide determination. Patients were instructed not to take any medication during the 24 h before study participation, and not to smoke, or to drink coffee or tea on the testing day. Afterwards, all patients performed a maximal exercise test on a bicycle ergometer with continuous monitoring of cardiorespiratory variables. Diagnosis of CFS. All patients were diagnosed as CFS cases by the same physician (the final author). To fulfill the diagnostic criteria for CFS, clinically evaluated, unexplained, persistent, or relapsing chronic fatigue that is of new or definite onset, should result in a substantial reduction in previous levels of occupational, educational, social, or personal activities (9). Furthermore, at least four of the following symptoms must have persisted or recurred during six or more consecutive months and must not have predated the fatigue: impairment in short-term memory or concentration, tender cervical or axillary lymph nodes, muscle pain, multijoint pain, headache, unrefreshing sleep, and postexertional malaise for more than 24 h (9). Any active medical condition that may explain the presence of chronic fatigue prohibits the diagnosis of CFS (diabetes, cancer, AIDS, etc.). Hence, all patients underwent an extensive medical evaluation, consisting of a standard physical examination, medical history, exercise capacity test, and routine laboratory tests. The laboratory tests included a complete http://www.acsm-msse.org blood cell count, determination of the erythrocyte sedimentation rate, serum electrolyte panel, measures of renal, hepatic, and thyroid function, and rheumatic and viral screens. If a patient’s medical history did not exclude a psychiatric problem at the time of disease onset, then a structured psychiatric interview was performed. In a number of cases further neurological, gynecological, endocrine, cardiac, and/or gastrointestinal evaluations were performed. The medical records were also reviewed to determine whether patients suffered from organic or psychiatric illness that could explain their symptoms. If any of the laboratory/ additional analyzes revealed any active medical condition that may explain the presence of the patient’s symptoms, the patient was excluded from the sample. Exercise testing. The exercise tests were performed at a humidity of ⫾60% and at a room temperature of ⫾20°C (Klimakamer, Jaeger, Germany). The patients performed a bicycle ergometric test against a graded increase in workload until exhaustion was reached (2). The patients were asked to take a sitting position on the electromagnetically braked ergometer (Lode B.V., Excalibur, Groningen, the Netherlands), after 3–5 min of adjustment the test was started. Heart rate was monitored continuously at rest and during exercise. There was continuous recording of the 12-lead electrocardiogram using an electrocardiograph (ECG Esaote Biomedica S.P.A., Firenze, Italy). To collect pulmonary data during the test, an open-circuit spirometer (Metamax Cortex, Biophysics, Germany) with automatic printout every 30 s was used. Automatically averages were attained for V̇O2PEAK (peak oxygen uptake) and maximal carbon dioxide production during every 30-s interval for the duration of each stage of the exercise. A two-way breathing valve attached to a mask, which covered the patients’ nose and mouth, was used to collect the expired air. The air was analyzed continuously for ventilatory and metabolic variables. Before each test, the spirometer was calibrated for environmental conditions. For the assessment of blood lactate concentration during the exercise stress test, blood was drawn every 2 min from an anticubital vein using natriumheparinized capillaries (EKF Diagnostics, Germany). Twenty-microliter blood samples only for lactate determination were taken at the hyperaemized earlobe and assayed by ESAT 6660 lactate (Medingen GmbH, Germany). All patients started the test at 10 W, with an increase of 10 W䡠min⫺1 (2). To avoid early onset of fatigue in the lower extremities due to inadequate physical fitness, the duration of the exercise was kept below 15 min. Patients were instructed to bicycle at a constant speed of 70 rpm. The following variables were measured: heart rate at rest (HRREST), peak heart rate (HRPEAK), exercise duration, maximal work capacity attained, work capacity attained at a respiratory exchange ratio (RER) of 1.0, V̇O2PEAK per kilogram of body weight, body weight–adjusted oxygen uptake at RER ⫽ 1.0, resting and peak RER (RERPEAK), the percentage of target heart rate achieved, and both the resting and peak lactate concentrations. The age-predicted HRPEAK was calculated as 220 minus the patients’ age in years. The metabolic data analyzed were the means of the last 30 s from EXERCISE IMMUNOLOGY IN CFS the final stage of exercise or the highest value attained if a decline in V̇O2 occurred at the final workload (2). For estimating the peak workload, the following equation was used: (workload of the highest fully completed stage) ⫹ (number of seconds achieved during the final stage/60 ⫻ 10). Exercise performance testing is widely used for the assessment of patients with CFS, and it appears to be both reproducible and valid (15). The exercise testing protocol used in the present study was able to distinguish between female CFS patients and healthy sedentary females (2), and the exercise performance data obtained with this protocol correlated with activity limitations/participation restrictions in CFS patients (20). RNase L-ratio determination. The assay is performed by 1) preparation of a cytoplasmic extract of the patient’s peripheral mononuclear blood cells, 2) combination of this extract with a labeled probe that binds specifically to 2=-5= A binding proteins such as RNase L and the low molecular weight species, 3) sodium dodecylsulfate polyacrylamide gel electrophoresis, and 4) densitometry to determine the relative quantities of 2=-5= A binding proteins. The RNase L-ratio was counted using the following equation: RNase L-ratio ⫽ [low molecular weight RNase L]/[high molecular weight RNase L] ⫻ 10. In detail, peripheral mononuclear blood cells (PBMC) were separated from heparinized blood (30 mL) by Ficoll–Hypaque density gradient centrifugation within 4 h of phlebotomy. In addition, PBMC were stored at ⫺70°C until cytoplasmic extraction preparation. The latter was performed in the presence of protease inhibitors elastase inhibitor III (Calbiochem, San Diego, CA), aprotinin, leupeptin, pefabloc-SC, and EDTA (Roche Biochemicals, Mannheim, Germany). Protease inhibitors are required for preventing proteolytic cleavage. Standard laboratory procedures were used to separate serum from coagulated blood, and to store it at ⫺70°C until analysis. A modified Bradford assay method (Bio-Rad Laboratories, Hercules, CA) was used for quantification of total proteins in the patients’ cell extracts. The probe specifically attaches to 2=-5=A binding proteins like 80 kDa RNase L and 37 kDa RNase L. Two hundred milligrams of PBMC extract was incubated with 2=-5= A trimer radiolabeled at the 3= end with 32P-pCp, at 2– 4°C for 15 min. In addition, it was covalently attached to the binding proteins by the addition of cyanoborohydride (20 mM in 100 mM of phosphate buffer, pH 8.0). This reduction reaction was allowed to progress for 20 min at 2– 4°C. A tracking dye were added to the samples, and incubated at 95°C for 5 min followed by separation using standard SDS-PAGE with a 4% stacking and a 10% separating gel. The gel was dried and autoradiography was used to detect the radioactivity of the marked probe (BioRad Laboratories Molecular Imager® Fx, Hercules, CA). Densitometric analysis of the autoradiographs was followed by quantification of any present 2=-5= A binding proteins (using specializes software: Quantity One® Software, BioRad Laboratories, Hercules, CA). For ratio of the 37 kDa RNase L isoform over the 83 kDa RNase L, a threshold value of 0.4 for the diagnosis of CFS was found to have a sensitivity of 91% and a specificity of 71% (29). Using a Medicine & Science in Sports & Exercise姞 1649 threshold value of 0.5, the RNase L-ratio was able to distinguish between CFS patients (abnormal in 41 of 57 subjects or 72%), healthy controls (3/28 or 11%), and patients with Fibromyalgia (0/11) and Depression (0/14) (3). In another study the amount of 37 kDa RNase L was able to distinguish between CFS patients (N ⫽ 53) and healthy controls (N ⫽ 26; P ⫽ 0.007) (27). The RNase L enzymatic activity (enzymatic assay) was assessed as described previously (24). In several studies, the RNase L activity was able to distinguish between CFS patients and healthy controls, with higher activity in the patients’ group (27,28). For the PKR activity measurement, PAGE-separated proteins were transferred to a membrane and visualized by immunodetection. In detail, the membranes were prepared by cutting out membranes of appropriate size (15 ⫻ 10 cm) and the membranes were prewet in Methanol 100%, shaken gently until soaked, and then the liquid was poured off. The membranes were immersed in Towbin buffer, and were shaken gently (for a minimum of 2–3 min) until soaked. To remove the gel, the braces from the glass plates were removed from the slab gel unit from the electrophoresis system, the spacers were removed from between the glass plates using a spatula and opened up, the upper glass plate was lifted up, leaving the gel on the bottom glass plate, and the gel was adhered to the membrane. Next, the Mylar mask was placed on the bottom of the Electroblot instrument. For each gel/membrane, 2 ⫻ 4 pieces of Whatmann were cut out and soaked in Towbin buffer. To make a “sandwich,” four layers of Whatmann, PVDF- or nitrocellulose membrane, Poly Acrylamide Gel, and four layers of Whatmann were placed on top of the Mylar mask. Air bubbles were avoided, and equal contact was ascertained by rolling smoothly over the surfaces with a plastic pipette. For the electroblot, the lid was put on the electroblot instrument, the electrodes were connected to the power supply, a weight was put on the lid (less than 1 kg to avoid buffer being squeezed out of the sandwich), and a current of 0.8 mA 䡠 cm–2 was applied for 2 h (i.e., 240 mA for two membranes of 150 cm2 each). Expose the blotted membranes to air at room temperature (RT) for at least 1 h to make sure the blotted membrane(s) is (are) dry. To visualize the proteins, the total surface of blotted membranes (regular membrane ⫽ 10 ⫻ 15 cm ⫽ 150 cm2) was calculated and the membranes were prewet by soaking them in Methanol 70% (approximately 12 mL per blot), shaking until they were completely wet, and removing the Methanol by gently pouring it off. The membranes were washed two times with phosphate-buffered saline (PBS)-Tween® (5 min each at room temperature while gently shaking ⫺0.25 mL·cm⫺2 ⫽ 37.5 mL for regular membrane). Next, the membranes were blocked with 5% NF milk in PBS-Tween® for 1 h at room temperature while gently shaking or overnight at 4°C, and the tray was covered with a lid or with aluminum foil (0.25 mL·cm⫺2). For incubation with the primary antibody, a dilution (dependent on the manufacturer’s instructions) was made in PBS-Tween® (0.1 mL·cm⫺2 ⫽ 15 mL for regular membrane), incubated for 2 h at room temperature while 1650 Official Journal of the American College of Sports Medicine gently shaking or overnight at 4°C, and the tray was covered with a lid or with aluminum foil. The membranes were washed two times with PBS-Tween® (5 min each at room temperature while gently shaking ⫺0.25 mL·cm⫺2) and incubated with secondary antibody; a dilution (dependent on the manufacturer’s instructions) was made in PBS-Tween® (0.1 mL·cm⫺2), incubated for 30 min at room temperature, shaken gently, and the tray was covered with a lid or with aluminum foil. The membranes were washed two times (colorimetric) or four times (chemiluminescence) with PBSTween®, 5 min each at room temperature while gently shaking: 0.25 mL·cm⫺2. The colorimetric analysis was performed using Opti4CN® (Bio-Rad) by mixing Elix-H2O 9/10, Opti4CN®-diluent 1/10, and Opti4CN®-substrate 0.2/10 (10 mL per blot), incubating while gently shaking until color develops, washing for 10 min with Elix-H2O, and drying by blotting the membranes on Whatmann paper. The specific protein bands were quantified by density scanning. Elastase activity in PBMC was measured using an enzymatic– colorimetric assay: EnzChek® Elastase Assay Kit E-12056 (Molecular Probes). The EnzChek kit contains DQTM elastin–soluble bovine neck ligament elastin that has been labeled with BODIPY®FL dye such that the conjugate can be digested by elastase or other proteases to yield highly fluorescent fragments. The resulting increase in fluorescence was monitored with a fluorescence microplate reader. First, a PBMC pellet and a PBMC pellet extract (in absence of elastase inhibitor III) were prepared, the protein concentration was measured, and the samples (patient proteins– extracts) were placed on ice and let thawed for 20 –30 min. DQ Elastin Substrate (BODIPY®FL): the new vial was reconstituted by adding 1 mL of dH2O (final concentration 1.0 mg·mL⫺1) and by mixing thoroughly to dissolve. Five milliliters of the stock buffer (10X) were diluted and 45 mL of dH2O was added to obtain the reaction buffer. To obtain a positive control (porcine pancreatic elastase), the new vial was reconstituted by adding 0.5 mL of dH2O up to a final concentration of 100 U·mL⫺1. Then, a standard dilution curve was constructed from the positive control (porcine pancreatic elastase with a starting concentration of 100 U·mL⫺1) with concentration of 5.0, 1.0, 0.5, 0.1, 0.05, and 0.01 U·mL⫺1. The controls were pipetted in triplicate into the microplate. Samples of 100 ␮g of proteins extract (X ␮L) were used. The total reaction volume was set at 200 ␮L (200 ␮L – X ␮L ⫽ Y ␮L buffer 1X). Y ␮L of buffer 1X was added into the samples, and 50 ␮L of sample dilution/well was added. The samples were pipetted in triplicate into the microplate. For the substrate, 5 ␮L of DQ elastin/sample was added to 145 ␮L of reaction buffer/sample. Next, 150 ␮L of the substrate/well sample was added. Protected from light, everything was incubated for 2 h at 37°C. Afterwards, the fluorescence intensity was measured in a fluorescence microplate reader (Molecular phospho-imager®FX BioRad and external laser Molecular ImagerâFX BioRad). The values were extrapolated from the equation of the curve (fluorescence vs elastase concentration) and multiplied by 200 (U䡠mg⫺1). For each sample, the value derived from the no-enzyme control was subtracted to correct for background http://www.acsm-msse.org TABLE 1. The descriptive statistics of the exercise performance variables (N ⫽ 16). Variable Mean ⴞ SD* Range Exercise duration (min) HRREST† (bpm) HRPEAK (bpm) % target heart rate achieved Workload (W) Workload per body weight (W䡠kg⫺1) Workload at RER‡ ⫽ 1.0 (W) V̇O2PEAK¶ (mL䡠kg⫺1䡠min⫺1) V̇O2 at RER ⫽ 1.0 (mL䡠kg⫺1䡠min⫺1) RERREST RERPEAK Lactate concentration at rest (mmol䡠L⫺1) Peak lactate concentration (mmol䡠L⫺1) 14.8 ⫾ 6.5 78.6 ⫾ 14.7 158.6 ⫾ 22.3 87.3 ⫾ 13.4 148.3 ⫾ 65.0 2.0 ⫾ 0.7 121.8 ⫾ 52.6 31.0 ⫾ 9.7 26.0 ⫾ 7.2 0.68 ⫾ 0.10 1.1 ⫾ 0.2 0.9 ⫾ 0.4 8.5 ⫾ 4.9 [7–26] [47–110] [125–197] [62–105] [70–260] [0.9–3.2] [72–215] [16.2–50.1] [15.7–40.2] [0.42–0.91] [0.9–1.6] [0.5–1.8] [1.3–21.2] * SD, standard deviation; † HRREST, resting heart rate; ‡ RER, respiratory exchange ratio; ¶ V̇O2PEAK, peak oxygen uptake. fluorescence. According to the company supplying the assay, the elastase activity assay had been thoroughly tested before it was brought on the market, but reliability and validity data are proprietary and unpublished (personal communication). The measurement of nitric oxide level in isolated PBMC was performed using a live cell assay (12,13). The cells or PBMC were washed once with 500 ␮L PBS, and spun for 2.5 min at 2500 ⫻ g. A 15-mM DAF-FM solution (4-amino5-methylamino 2=,7=-difluorofluorescein diacetate, Molecular Probes D-23844: 3 mL of 5-mM stock in 1 mL of PBS) was prepared, and the cells were resuspended in 100 mL of this solution, left untouched for 45 min at room temperature in a dark environment, and spun for 2.5 min at 2500 ⫻ g. A solution of the CD14 staining (Becton Dickinson, BD345785) was prepared (6 ␮L CD14 in 60 ␮L PBS), and the cells were resuspended in 66 mL of this solution, left untouched for 25 min at room temperature in a dark environment, spun for 2.5 min at 2500 ⫻ g, and resuspended in 500 mL of PBS. Analysis was performed with a flow cytometer; the monocyte population was gated and the mean fluorescence (525-nm band pass) of this population was measured. Cells were analyzed quickly, and kept in the dark until processed. For a more detailed prescription of the assay, the reader is referred to references (12) and (13). Statistical analysis. Data were analyzed using Statistica version 5.1 (Statsoft, Tulsa, OK). Appropriate descriptive statistics were used: frequencies and percentages for the gender distribution; mean, standard deviation (SD), and range for illness duration; age; the exercise performance variables; and the immunological variables. To examine the associations between exercise performance and the immune variables, Pearson correlation analyzes were used. A one sample Kolmogorov–Smirnov (K–S) goodness-of-fit test was used to examine whether the variables entering a Pear- son correlation analysis were normally distributed. If a variable was not normally distributed, then the nonparametric Spearman correlation analysis was used. For interpreting correlation coefficients, they were squared to obtain the coefficient of determination. For the correlation analysis, the significance level was set at 0.01 to help protect against potential type I errors. In case of a “trend” towards a significant association (0.01 ⬍ P ⬍0.05), a power analysis was performed (22). A power of 80% was considered fair. Finally, the interactions between the exercise performance variables and the immunological variables were further assessed using forward stepwise multiple regression analysis. For the regression analysis, the significance level was set at 0.05. RESULTS The descriptive statistics of the exercise performance variables are displayed in Table 1, the descriptives of the immune variables in Table 2. The mean percentage of monocytes was 18.9 ⫾ 6.9 (range [9.1–32.2]). All subjects displayed abnormal responses for both the elastase activity and RNase L-ratio, and RNase L activity was abnormal in 15/16 CFS patients. However, the Protein Kinase R activity was within the normal range in the majority of the study sample (11/16), and approximately 50% of the subjects presented with normal NO levels in both monocytes and lymphocytes. Apart from the peak RER (K-S z ⫽ 1.49; P ⫽ 0.02), all exercise performance and immune variables were normally distributed. Thus, a Pearson correlation analysis was used for analyzing the majority of associations. For the examination of the associations between RERPEAK and the immune variables, a nonparametric Spearman correlation analysis was used. The outcome of the correlation analysis, TABLE 2. The descriptive statistics of the immunological variables (N ⫽ 16). Variable Mean ⴞ SD* Range Normal Range† Abnormal in N (%) patients Elastase activity (U䡠mg⫺1 extract) Protein kinase R activity RNase L activity RNase L ratio [NO] in monocytes [NO] in lymphocytes 388.0 ⫾ 469.5 1.9 ⫾ 1.1 160.1 ⫾ 111.1 1.8 ⫾ 1.7 38.0 ⫾ 18.8 20.2 ⫾ 8.2 [112–2080] [0.6–5.4] [46–537] [0.5–7.7] [13.2–72.9] [10.4–36.1] ⬍70 ⬍2.0 ⬍50 ⬍.5 [29.6–40.9] [11.2–18.5] 16 (100.0) 5 (31.3) 15 (94.8) 16 (100.0) 8 (50.0) 9 (56.3) * SD, standard deviation; † normal range based on laboratory data from 60 healthy volunteers (unpublished data provided by the laboratory). EXERCISE IMMUNOLOGY IN CFS Medicine & Science in Sports & Exercise姞 1651 TABLE 3. Intracellular immunity vs exercise performance in 16 CFS subjects. Variable Elastase Activity r* (p†) Exercise duration (min) HRREST (bpm) HRPEAK (bpm) % target heart rate achieved Workload (W) Workload per body weight (W䡠kg⫺1) Workload at RER ⫽ 1.0 (W) V̇O2PEAK (mL䡠kg⫺1䡠min⫺1) V̇O2 at RER ⫽ 1.0 (mL䡠kg⫺1䡠min⫺1) RERREST RERPEAK Lactate concentration at rest (mmol䡠L⫺1) Peak lactate concentration (mmol䡠L⫺1) Age ⫺0.39 (0.14) 0.10 (0.72) ⫺0.46 (0.07) ⫺0.58 (0.02) ⫺0.39 (0.14) ⫺0.38 (0.15) ⫺0.59 (0.02) ⫺0.44 (0.08) ⫺0.73 (0.001) ⫺0.67 (0.004) 0.05‡ (0.87) ⫺0.16 (0.56) ⫺0.41 (0.11) ⫺0.51 (0.04) PKR Activity r (p) ⫺0.45 (0.08) 0.27 (0.31) ⫺0.38 (0.14) ⫺0.50 (0.05) ⫺0.45 (0.08) ⫺0.41 (0.11) ⫺0.63 (0.009) ⫺0.46 (0.07) ⫺0.70 (0.002) ⫺0.69 (0.003) 0.02‡ (0.95) ⫺0.04 (0.88) ⫺0.37 (0.16) ⫺0.49 (0.06) RNase L Activity r (p) ⫺0.42 (0.11) 0.26 (0.33) ⫺0.43 (0.10) ⫺0.51 (0.04) ⫺0.42 (0.11) ⫺0.39 (0.14) ⫺0.57 (0.02) ⫺0.48 (0.06) ⫺0.70 (0.003) ⫺0.65 (0.007) ⫺0.11‡ (0.69) ⫺0.09 (0.75) ⫺0.46 (0.08) ⫺0.42 (0.11) RNase L Ratio r (p) ⫺0.31 (0.25) 0.29 (0.29) ⫺0.25 (0.35) ⫺0.36 (0.17) ⫺0.31 (0.25) ⫺0.31 (0.24) ⫺0.50 (0.05) ⫺0.37 (0.15) ⫺0.66 (0.006) ⫺0.70 (0.002) 0.27‡ (0.31) ⫺0.07 (0.81) ⫺0.33 (0.21) ⫺0.46 (0.07) NOmonocyte r (p) NOlymphocyte r (p) 0.40 (0.13) ⫺0.14 (0.61) 0.11 (0.68) 0.04 (0.87) 0.40 (0.13) 0.16 (0.57) 0.26 (0.34) 0.04 (0.88) 0.26 (0.34) 0.03 (0.93) ⫺0.00‡ (0.99) ⫺0.20 (0.45) ⫺0.16 (0.56) ⫺0.17 (0.54) 0.48 (0.06) ⫺0.20 (0.45) 0.20 (0.47) 0.14 (0.61) 0.48 (0.06) 0.27 (0.31) 0.32 (0.22) 0.17 (0.53) 0.32 (0.22) 0.18 (0.51) ⫺0.07‡ (0.81) ⫺0.28 (0.30) ⫺0.09 (0.73) ⫺0.12 (0.65) * Pearson correlation coefficient; † level of significance was set at 0.01 to help protect against potential type I errors; ‡ nonparametric Spearman’s rho because RERPEAK was not normally distributed (Kolmogorov–Smirnov test). examining the associations between the immune and exercise performance variables, is displayed in Table 3. Strong correlations (r ranged between 0.65 and 0.73) were observed between four intracellular immune variables (i.e., elastase activity, PKR activity, RNase L activity, and proteolysis) and both the resting RER and the oxygen uptake at RER ⫽ 1.0. The achieved workload at RER ⫽ 1.0 correlated with the PKR activity, and displayed at trend towards a statistical significant association (0.01⬍P ⬍ 0.05) with elastase activity (power ⫽ 81%), RNase L activity (power ⫽ 81%), and the RNase L-ratio (power ⫽ 63%). Likewise, elastase activity (power ⫽ 81%) and RNase L activity (power ⫽ 63%) displayed a trend towards a correlation with the percentage of target heart rate achieved. Neither the NO concentration in monocytes nor the lymphocytes’ NO concentration correlated with any of the exercise performance variables. Forward stepwise multiple regression analysis revealed 1) that elastase activity was the only factor related to the reduction in oxygen uptake at a RER of 1.0 (regression model: R2 ⫽ 0.53, F (1,14) ⫽ 15.5, P ⬍ 0.002; elastase activity P ⬍ 0.002); 2) that the PKR activity was the principle factor related to the reduction in workload at RER ⫽ 1.0 (regression model: R2 ⫽ 0.77, F (6,9) ⫽ 5.2, P ⬍ 0.01; PKR activity P ⬍ 0.009; monocyte NO P ⬍ 0.13; % monocytes P ⬍ 0.17; RNase L-ratio P ⬍ 0.23; elastase activity P ⬍ 0.051; RNase L activity P ⬍ 0.32); 3) that the elastase activity was the principle factor related to the reduction in % of target heart rate achieved (regression model: R2 ⫽ 0.89, F (7,8) ⫽ 9.5, P ⬍ 0.002; elastase activity P ⬍ 0.02; RNase L-ratio P ⬍ 0.02; PKR activity P ⬍ 0.02); and 4) that the level of elastase was the only factor inversely related to the increase in age (regression model: R2 ⫽ 0.52, F (1,14) ⫽ 5.3, P ⬍ 0.03; elastase activity P ⬍ 0.04). DISCUSSION These data add to the body of literature showing impairments of intracellular immunity in patients with CFS. The results provide evidence for an association between intracellular immune deregulation and exercise performance in 1652 Official Journal of the American College of Sports Medicine patients with CFS. Elastase activity and PKR activity were identified as determinants of the reduction in oxygen uptake at RER ⫽ 1.0, the reduction in workload at RER ⫽ 1.0, and the reduction in percent of target heart rate achieved. RNase L activity and proteolysis correlated strongly with both the resting RER and the oxygen uptake at RER ⫽ 1.0, whereas resting NO levels were not related to any of the exercise performance variables. The results are in accordance with an earlier report, providing preliminary evidence of an association between RNase L proteolysis (as assessed using the RNase L-ratio) and exercise performance in CFS patients (25). Still, the present study is the first to study numerous intracellular immune variables together with exercise performance in CFS patients. The role of elastase might be related to impairments of lung diffusion and impairments of oxygen delivery in tissues. Indeed, the exercise performance variables that displayed an association with the immune variables, including elastase activity, were mainly related to respiration (oxygen uptake and RER). On the other hand, intracellular elastase activity in peripheral monocytes/lymphocytes was assessed. Peripheral blood characteristics may not correspond to the alterations in the capillaries surrounding lung alveoli. In addition, it seems unlikely that intracellular elastase in peripheral blood is capable of causing lung tissue damage, as seen in other diseases like cystic fibrosis (7). In an animal model of cystic fibrosis, elastase degraded several alveolar surfactant proteins important for alveolar tension reduction and innate immune function (1). Apart from one study showing an association between immune activation and bronchial hyperresponsiveness (17), we are unaware of experimental data providing evidence for impairments in lung tissue in CFS patients. An increased number of cytotoxic T-cells, accompanied by a decreased amount of naı̈ve T-cells, was observed in CFS patients with bronchial hyperresponsiveness compared with CFS patients without bronchial hyperresponsiveness (17). T-cells release elastase to establish their cytotoxicity, linking the current with our previous observations. RNase L activity and proteolysis correlated strongly with both the resting RER and the oxygen uptake at RER ⫽ 1.0. http://www.acsm-msse.org These observations may be related to the increased elastase activity: elastase has been identified as one of the proteolytic enzymes responsible for RNase L cleavage (5). Whether RNase L proteolysis triggers a channelopathy and consequent intracellular hypomagnesaemia in skeletal muscles and transient hypoglycemia, as suggested in the introduction, requires further investigation. Furthermore, it was hypothesized that PKR activation and consequent elevated NO levels might limit exercise performance in CFS patients. The current observations do not support this hypothesis; although PKR activation appeared to be a determinant of exercise performance, no associations between NO levels and exercise performance were observed. However, further studying of the hypothesis assessing extracellular NO levels during the exercise challenge rather than resting NO levels in peripheral monocytes/lymphocytes is warranted. Addressing the study limitations, the cross-sectional nature of the study should be acknowledged. To establish a causal relationship, further study of these interactions in a larger study sample, using a prospective longitudinal design, is required. In addition, given that a number of correlations were significant at the 0.05, but not at the 0.01 level, one can argue that the sample size lacked strength. Depending on the parameter of interest, the power of the study varied between 63 and 81%, with an associated probability of Type II error of 37 and 19% respectively. Because a power of 80% was considered fair, increasing the sample size would only be appropriate to reveal statistically significant correlations between RNase L activity and the percentage of age-predicted target heart rate achieved, and between the RNase L-ratio and the workload at RER ⫽ 1.0. Increasing the sample size to approximately 22 subjects might have revealed significant correlations between these parameters. A power analysis should have determined the sample size before the study started, but our budget did not allow us to include more than 16 subjects. It has been concluded that the main findings of the present study were not biased by the sample size. Finally, the study sample may not be representative of the CFS population in general; the study participants were randomly selected from patients visiting a specialized chronic fatigue clinic, and the gender distribution (50% females) is not in accordance with the epidemiology of CFS. Women appear more likely to develop the disease than men and children (6). Because it is well established that women have a lower maximal oxygen uptake compared with men (16), a distinct gender distribution and even pooling of gender data may bias the results (23). In two earlier studies (18,20), however, reanalyzing the exercise performance data without pooling gender data did not change the outcome. With respect to the study limitations, it has been concluded that new evidence supportive of an interaction between intracellular immune dysfunction and exercise performance in CFS patients was provided. These findings may aid a variety of health care workers (physiotherapists, physicians, rehabilitation specialists, occupational therapists) in understanding exercise physiology in patients with CFS. Further study of these interactions is warranted. The authors would like to thank Freya Vaeyens (MSc) and Marc Frémont (PhD) (both working at RED Laboratories, Zellik, Belgium) for editing the methods’ description of the immunological parameters. REFERENCES 1. ALCORN, J. F., and J. R.WRIGHT. Degradation of pulmonarysurfactant protein D by pseudomonas aeruginosa elastase abrogates innate immunefunction. J. Biol. Chem. 279:30871–79, 2004. 2. DE BECKER, P., J. ROEYKENS, M. REYNDERS, N. MCGREGOR, and K. DE MEIRLEIR. Exercise capacity in chronic fatigue syndrome. Arch. Intern. Med. 160:3270–7, 2000. 3. DE MEIRLEIR, K., C. BISBAL, I. CAMPINE, et al. A 37 kDa 2-5A binding protein as a potential biochemical marker for Chronic Fatigue Syndrome. Am. J. Med. 108:99–105, 2000. 4. DE MEIRLEIR, K., P. DE BECKER, J. NIJS, et al. Chronic Fatigue Syndrome etiology, the immune system and infection. In: Chronic Fatigue Syndrome: A Biological Approach, P. Englebienne and K. De Meirleir. Boca Raton: CRC Medical Publications, 2002, pp. 201–228. 5. DEMETTRE, E., L. BASTIDE, A. D’HAESE, et al. Ribonuclease L proteolysis in peripheral blood mononuclear cells of chronic fatigue syndrome patients. J. Biol. Chem. 277:35746–35751, 2002. 6. DONNAY, A., and G. ZIEM. Prevalence and overlap of chronic fatigue syndrome and fibromyalgia syndrome among 100 new patients with multiple chemical sensitivity syndrome. J. Chronic Fatigue Syndr. 5:71–80, 1999. 7. DÖRING, G., F. FRANK, C. BOUDIER, S. HERBERT, B. FLEISCHER, and G. BELLON. Cleavage of lymphocyte surface antigens CD2, CD4, and CD8 by polymorphonuclear leukocyte elastase and cathepsin G in patients with cystif fibrosis. J. Immunol. 154: 4842–50, 1995. EXERCISE IMMUNOLOGY IN CFS 8. ENGLEBIENNE, P., M. FRÉMONT, F. VAEYENS, et al. Chronic fatigue syndrome (CFS) and multiple sclerosis (MS) as subsets of a group of cellular immunity disorders. ME/CFS: The Medical Practitioners’ Challenge. Proceedings at Sydney International Conference, pp. 35– 42, 2001. 9. FUKUDA, K., S. E. STRAUSS, I. HICKIE, et al. The Chronic Fatigue Syndrome, a comprehensive approach to its definition and study. Ann. Intern. Med. 121:953–959, 1994. 10. FULCHER, K. Y., and P. D. WHITE. Strength and physiological response to exercise in patients with chronic fatigue syndrome. J. Neurol. Neurosurg. Neuropsych. 69:302–307, 2000. 11. IORDANOV, M. S., J. WONG, J. C. BELL, and B. E. MAGUN. Activation of NF-kB by double-stranded RNA (dsRNA) in the absence of protein kinase R and RNase L demonstrates the existence of two separate dsRNA-triggered antiviral programs. Mol. Cell. Biol. 21:61–72, 2001. 12. ITOH, Y., F. H. MA, H. HOSHI, et al. Determination and bioimaging method for nitric oxide in biological specimens by diaminofluorescein fluorometry. Anal. Biochem. 287:203–209, 2000. 13. KOJIMA, H., Y. URANO, K. KIKUCHI, T. HIGUCHI, Y. HIRATA, and T. NAGANO. Fluorescent Indicators for Imaging Nitric Oxide Production. Angew Chem. Int. Ed. Engl. 38:3209–3212, 1999. 14. KURUP, R. K., and P. A. KURUP. Hypothalamic digoxin, cerebral chemical dominance and myalgic encephalomyelitis. Intern. J. Neuroscience. 113:683–701, 2003. 15. MANU, P., G. AFFLECK, H. TENNEN, P. A. MORSE, and J. I. ESCOBAR. Hypochondriasis influence quality-of-life outcomes in patients with chronic fatigue syndrome. Psychother Psychosom. 65:76–81, 1996. Medicine & Science in Sports & Exercise姞 1653 16. MCARDLE, W. D., F. I. KATCH, and V. L. KATCH. Exercise Physiology: Energy, Nutrition and Human Performance. Baltimore, Williams & Wilkins, 1996, pp. 204 –207. 17. NIJS, J., P. DE BECKER, DE MEIRLEIR, K. et al. Associations between bronchial hyperresponsiveness and immune cell parameters in patients with chronic fatigue syndrome. Chest 123:998–1007, 2003. 18. NIJS, J., K. DE MEIRLEIR, and W. DUQUET. Kinesiophobia in chronic fatigue syndrome: assessment and associations with disability. Arch. Phys. Med. Rehabil. 85:1586–92, 2004. 19. NIJS, J., K. DE MEIRLEIR, M. MEEUS, N. R. MCGREGOR, and P. ENGLEBIENNE. Chronic fatigue syndrome: intracellular immune deregulations as a possible etiology for abnormal exercise response. Med. Hypoth. 62:759–765, 2004. 20. NIJS, J., K. DE MEIRLEIR, S. WOLFS, and W. DUQUET. Disability evaluation in chronic fatigue syndrome: Associations between exercise capacity and activity limitations/participation restrictions. Clin. Rehabil. 18:139–148, 2004. 21. PALUDAN, S. R., S. ELLERMANN-ERIKSEN, and S. C. MOGENSEN. NF-␬B activation is responsible for the synergistic effect of herpes simplex virus type 2 infection on interferon-␥-induced nitric oxide production in macrophages. J. Gen. Virol. 79:2785–93, 1998. 22. PORTNEY, L. G., and M. P. WATKINS. Foundations of Clinical Research: Applications to Practice. New Jersey: Prentice Hall 2000, p. 724. 23. SARGENT, C., G. C. SCROOP, P. M. NEMETH, R. B. BURNET, and J. D. BUCKLEY. Maximal oxygen uptake and lactate metabolism are normal in chronic fatigue syndrome. Med. Sci. Sports Exerc. 34:51–6, 2002. 1654 Official Journal of the American College of Sports Medicine View publication stats 24. SHETZLINE, S. E., C. MARTINAND-MARI, N. L. REICHENBACH, et al. Structural and functional features of the 37-kDa 2-5A-Dependent RNase L in chronic fatigue syndrome. J. Interferon Cytokine Res. 22:443–456, 2002. 25. SNELL, C. R., J. M. VANNESS, D. R. STRAYER, and S. R. STEVENS. Physical performance and prediction of 2-5A Synthetase/RNase L antiviral pathway activity in patients with chronic fatigue syndrome. In Vivo 16:107–110, 2002. 26. SORENSEN, B., J. E. STREIB, M. STRAND, et al. Complement activation in a model of chronic fatigue syndrome. J. Allergy Clin. Immunol. 112:397–403, 2003. 27. SUHADOLNIK, R. J., D. L. PETERSON, P. R. CHENEY, et al. Biochemical dysregulation of the 2-5A synthetase/Rnase L antiviral defense pathway in chronic fatigue syndrome. J. Chronic Fatigue Syndr. 5:223–242, 1999. 28. SUHADOLNIK, R. J., N. L. REICHENBACH, P. HITZGES, et al. Upregulation of the 2-5A synthetase/Rnase L antiviral pathway associated with chronic fatigue syndrome. Clin. Infect. Dis. 18:S96–104, 1994. 29. TIEV, M. P., E. DEMETTRE, P. ERCOLANO, L. BASTIDE, B. LEBLEU, and J. CABANE. RNase L levels in peripheral blood mononuclear cells: 37- kilodalton/83-kilodalton isform ratio is a potential test for chronic fatigue syndrome. Clin. Diagn. Lab. Immunol. 10:315– 316, 2003. 30. WONG, R., G. LOPASCHUK, G. ZHU, et al. Skeletal muscle metabolism in the chronic fatigue syndrome: in vivo assessment by 31P nuclear magnetic resonance spectroscopy. Chest 102:1716–22, 1992. http://www.acsm-msse.org