atrial fibrillation promotion by endurance exerciseatrial fibrillation promotion by endurance...

Journal of the American College of Cardiology Vol. 62, No. 1, 2013� 2013 by the American College of Cardiology Foundation ISSN 0735-1097/$36.00Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jacc.2013.01.091

PRE-CLINICAL RESEARCH

Atrial Fibrillation Promotion by Endurance Exercise
Demonstration and Mechanistic Explorationin an Animal Model
Eduard Guasch, MD,*y Begoña Benito, MD,*yz Xiaoyan Qi, PHD,* Carlo Cifelli, PHD,xPatrice Naud, PHD,* Yanfen Shi, PHD,* Alexandra Mighiu, BSC,x Jean-Claude Tardif, MD,*

Artavazd Tadevosyan, MSC,* Yu Chen, MSC,* Marc-Antoine Gillis, MSC,* Yu-Ki Iwasaki, MD,*

Dobromir Dobrev, MD,jj Lluis Mont, MD,x{ Scott Heximer, PHD,x Stanley Nattel, MD*

Montreal, Quebec, and Toronto, Ontario, Canada; Barcelona, Catalonia, Spain; and Essen, Germany

From the *R

and Univers

University o

Hospital de

Physiology,

Research an

macology, U

Fibrillació A

cions Biomé

research was

MOP44365

Foundation,

Objectives T

esearch Center and Dep

ity of Montreal, Montrea

f Barcelona, Barcelona,

l Mar, Parc de Salut Mar

Heart and Stroke/Richard

d University of Toronto,

niversity of Duisburg-E

uricular, Hospital Clínic

diques August Pi i Sunyer

funded by the Canadian

for Dr. Nattel; MOP1066

Fondation Leducq (E

he goal of this study was to assess mechanisms underlying atrial fibrillation (AF) promotion by exercise trainingin an animal model.

Background H
igh-level exercise training promotes AF, but the underlying mechanisms are unclear.
Methods A
F susceptibility was assessed by programmed stimulation in rats after 8 (Ex8) and 16 (Ex16) weeks of daily 1-htreadmill training, along with 4 and 8 weeks after exercise cessation and time-matched sedentary (Sed) controls.Structural remodeling was evaluated by using serial echocardiography and histopathology, autonomic nervoussystem with pharmacological tools, acetylcholine-regulated potassium current (IKACh) with patch clamp recording,messenger ribonucleic acid expression with quantitative polymerase chain reaction, and regulators of G protein–signaling (RGS) 4 function in knockout mice.
Results A
F inducibility increased after 16 weeks of training (e.g., AF >30 s in 64% of Ex16 rats vs 15% of Sed rats; p < 0.01)and rapidly returned to baseline levels with detraining. Atropine restored sinus rhythm in 5 of 5 Ex rats with AFsustained >15 min. Atrial dilation and fibrosis developed after 16 weeks of training and failed to fully recover withexercise cessation. Parasympathetic tone was increased in Ex16 rats and normalized within 4 weeks of detraining.Baroreflex heart rate responses to phenylephrine-induced blood pressure elevation and IKACh sensitivity to carbacholwere enhanced in Ex16 rats, implicating both central and end-organ mechanisms in vagal enhancement. Ex ratsshowed unchanged cardiac adrenergic and cholinergic receptor and IKACh-subunit gene expression, but significantmessenger ribonucleic acid downregulation of IKACh-inhibiting RGS proteins was present at 16 weeks. RGS4knockout mice showed significantly enhanced sensitivity to AF induction in the presence of carbachol.
Conclusions C
hronic endurance exercise increased AF susceptibility in rats, with autonomic changes, atrial dilation, and fibrosisidentified as potential mechanistic contributors. Vagal promotion is particularly important and occurs via augmentedbaroreflex responsiveness and increased cardiomyocyte sensitivity to cholinergic stimulation, possibly due toRGS protein downregulation. (J Am Coll Cardiol 2013;62:68–77) ª 2013 by the American College of CardiologyFoundation
See page 78
The cardiovascular benefits of regular physical exercise arewell recognized (1). However, there is emerging evidencethat the cardiac remodeling induced by sustained high-level
artment of Medicine, Montreal Heart Institute

l, Quebec, Canada; yDepartment of Medicine,

Catalonia, Spain; zCardiology Department,

, Barcelona, Catalonia, Spain; xDepartment of

Lewar Centre of Excellence in Cardiovascular

Toronto, Ontario, Canada; jjInstitute of Phar-

ssen, Essen, Germany; and the {Unitat de

, University of Barcelona, Institut d’Investiga-

(IDIBAPS); Barcelona, Catalonia, Spain. This

Institutes of Health Research (MGP-6957 and

70, for Dr. Heximer), Quebec Heart and Stroke

uropean–North American Atrial Fibrillation

Research Alliance), and the Deutsche Forschungsgemeinschaft (Do 769/1-3). Drs.

Guasch and Benito, doctoral candidates at the University of Barcelona, were supported

by Hospital Clínic and Rio Hortega awards (CM08/00201, CM06/00189) from the

Spanish Health Ministry. Ms. Mighiu and Dr. Cifelli were supported by graduate

studentships from the Canadian Institutes of Health Research. Mr. Tadevosyan is

a recipient of a Le Fonds de la recherche en santé du Québec (FRSQ)-Le Réseau en

santé cardiovasculaire (RSCV)/Heart and Stroke Foundation of Québec (HSFQ)

FRSQ-RSCV/HSFQ doctoral scholarship. All other authors have reported that they

have no relationships relevant to the contents of this paper to disclose. The first 2

authors contributed equally to this work.

Manuscript received November 30, 2012; revised manuscript received January 16,

2013, accepted January 20, 2013.

http://dx.doi.org/10.1016/j.jacc.2013.01.091

Abbreviationsand Acronyms

CL = cycle length

DEx = discontinuation of

exercise

ECG = electrocardiography

EPS = electrophysiological

study

Ex = exercise training

IKACh = acetylcholine-

regulated potassium current

LA = left atrial

mRNA = messenger

ribonucleic acid

RA = right atrial

RAERP = right atrial

effective refractory period

RGS = regulators of

G-protein signaling

RGS4KO = G-protein–

signaling 4 knock-out

WBCL = Wenckebach

cycle length

JACC Vol. 62, No. 1, 2013 Guasch et al.July 2, 2013:68–77 AF Mechanisms With Intense Exercise Training

69

exercise training may also carry some risk. Recent reportsemphasize the development of deleterious right ventric-ular remodeling, which may present as arrhythmogeniccardiomyopathy-like manifestations, as a result of high-intensity training (2). In a rat model of chronic high-intensity endurance training, we found a substratefor ventricular arrhythmia vulnerability associated with rightventricular fibrosis, dilation, and dysfunction (3).

Emerging data also underline the importance of intenseexercise training in atrial fibrillation (AF). Marathon runners(4), elite cyclists (5), and cross-country skiers (6) are atparticularly high risk. However, the AF risk is not confined toelite athletes: a dose–response relationship has been seen inmen aged younger than 50 years and in joggers (7).

The mechanisms of AF promotion by regular high-intensity exercise are unclear. Proposed contributorsinclude atrial dilation, inflammatory changes, vagalenhancement, and structural alterations (8). Improvedunderstanding of the mechanisms of exercise-related AFmay help in the development of novel therapeuticapproaches (9) and evidence-based management guidelines(10). We designed the current studies to determine whetherendurance training increases AF vulnerability in a rat model,and if so, to assess potential underlying mechanisms.

Methods

This section contains brief summaries of principal methods.For complete details, please see the Online Appendix.Exercise training. Wistar rats were randomly assigned tomatched sedentary (Sed) or intensive exercise-training (Ex)groups as previously described (3). Ex rats underwent 1-h/day treadmill training 5 days per week. The protocolincluded an initial 1-week progressive training program,increasing to steady-state 60-min running at 28 m/min.Animals that did not readily master the running programaccording to a specific running score (Online Appendix)were excluded. Thereafter, animals were trained at this levelfor 8 (Ex8) or 16 (Ex16) weeks. This training protocol wasshown not to increase distress levels as measured witha previously validated distress score (Online Fig. 1). Anadditional group of Ex16 rats underwent discontinuation ofexercise for 4 (DEx4) or 8 (DEx8) weeks. Time-matchedparallel Sed rats were controls for each Ex and DExgroup. At study end, rats underwent an in vivo electro-physiological study (EPS), followed by cardiac excision andformalin preservation (for histology), snap-freezing in liquidnitrogen (molecular biology), or retrograde perfusion (car-diomyocyte isolation).Echocardiography. Transthoracic echocardiographic studieswere performed with a phased-array probe 10S (4.5 to11.5 MHz) in a Vivid 7 Dimension system under 2% iso-flurane anesthesia. Repeated measurements were made in2 study series: a training evolution set (baseline, 8-week, and16-week Ex) and a detraining set (16-week Ex, DEx4, andDEx8), along with corresponding Sed controls. Detailed

measurements and methods aredescribed in the Online Appendix.Recordings and analyses wereblinded to group assignment.Electrophysiological study.In vivo EPS was performed under2% isoflurane anesthesia. Thedistal dipole of a 1.9-F octapolarcatheter in the right atrium wasused for programmed stimulation.

Spontaneous (drug-free) ar-rhythmia inducibility was testedwith double and triple extra-stimuli during sinus rhythm and9-beat trains at a cycle length(CL) of 150 ms and also withburst pacing at 40- and 60-msCLs for 7.5, 15, and 30 s. AFwas defined as >1 s of irregularatrial electrograms (�800 beats/min) with irregular ventricularresponse. AF lasting 1 to 30 s wasconsidered nonsustained, �30 swas sustained AF, and >15 min
was long-lasting AF. If no sustained AF was induced,inducibility was retested 10 minutes after cumulative carba-chol doses of 25 and 50 mg/kg intraperitoneally. The rightatrial effective refractory period (RAERP) was measured inEx16 rats and corresponding Sed rats with 9-beat basicpacing-stimulus (S1) trains (CL ¼ 150 ms) followed bya premature extrastimulus (S2) with coupling interval (S1S2)reduced in 1-ms decrements, with RAERP defined as thelongest S1S2 failing to produce a propagated response.Surgical procedures. All surgical/instrumentation proce-dures were performed under 2% isoflurane anesthesia.Experiments were performed at least 48 hours postsurgery.Electrocardiography telemetry. In 15 rats (8 Ex rats,7 parallel Sed rats), repeated 24-h electrocardiography (ECG)recordings were obtained with implanted telemetry devicesat baseline, 8-week Ex, 16-week Ex, DEx4, and DEx8. RRintervals were analyzed with commercial software. Recordingswere reviewed manually to detect spontaneous arrhythmias.Autonomic tone assessment. Parasympathetic and sym-pathetic tone was evaluated based on heart rate changesdue to sequential pharmacological blockade. Experimentswere performed on separate days in awake, nonrestrained ratswith chronically implanted ECG telemetry and intravenousaccess systems. In preliminary time-course studies (OnlineFig. 2A), atropine and propranolol effects were maximalwithin 1 min and maintained for >30 min. Definitivestudies were designed based on the preliminary data. Para-sympathetic tone index was defined according to the differ-ence between heart rate after propranolol alone versuspropranolol þ atropine (Online Fig. 2B, top). Sympathetictone index was the heart rate difference between atropinealone and atropine þ propranolol (Online Fig. 2B, bottom).

Guasch et al. JACC Vol. 62, No. 1, 2013AF Mechanisms With Intense Exercise Training July 2, 2013:68–77

70

Baroreflex sensitivity was assessed in conscious, unre-strained rats. Maximum changes in blood pressure and heartrate in response to progressively increasing doses of phen-ylephrine and nitroprusside were determined at each dose(Online Fig. 3).Fibrosis quantification. Fibrosis was quantified blinded togroup assignment, with Picrosirius red staining as describedpreviously (3). Atrial photomicrographs were obtained withan Olympus BX60 (Olympus Microscopes, Richmond Hill,Ontario, Canada) microscope and QImaging QCam(Olympus Canada), and quantified with color recognitionsoftware.Acetylcholine-dependent potassium current measur-ement. Cardiomyocytes were isolated by enzymatic diges-tion. Acetylcholine-dependent potassium current (IKACh)was obtained by digital subtraction of baseline current fromcurrent in the presence of carbachol. Concentration-responsecurves were obtained with the equation: Y ¼Min þ (D/(1 þ10LogEC50-X), where Y ¼ effect at concentration X, Min ¼minimum-response constant,D¼maximum drug effect, andEC50 ¼ 50% maximal effect concentration.Messenger ribonucleic acid quantification. Ribonucleicacid was purified, quantified, and DNAse treated. Poly-merase chain reaction reactions were performed on TaqManlow-density arrays. Glyceraldehyde-3-phosphate dehydro-genase served as the reference gene.Regulator of G protein–signaling 4 knock-out mice. Theregulator of G protein–signaling (RGS) 4 knock-out(RGS4KO) mouse (Rgs4tm1Dge1/J) strain was backcrossedonto a C57Bl/6 background. In vivo EPS was conducted viaright jugular vein access with a 2-F octopolar catheter. AFinduction was attempted with burst pacing at 10- to 40-mspacing intervals and triple extrastimulus techniques at base-line (drug-free) and after 10 mg/kg intraperitoneally and30 mg/kg of carbachol.Statistical analysis. Quantitative variables are shown asmean � SEM. For details of statistical comparisons for eachanalysis, see the Online Appendix. Most comparisons wereperformed by using linear mixed-effects modeling, withmain factor and all interaction effects included. Whensignificant interactions were found, pairwise comparisonswere executed with a least significant difference test (this testdoes not offer actual correction for multiplicity).

Nonnormally distributed variables were compared byusing the Mann-Whitney U test. Categorical variables werecompared by using the Fisher exact test. Analyses wereprocessed with SPSS version 19.0 (IBM SPSS Statistics,IBM Corporation, Armonk, New York) and GraphPadversion 5.0 (GraphPad Software, Inc., La Jolla, California).A p value <0.05 was considered statistically significant.

Results

Endurance-training promotes AF vulnerability. Exercisetraining did not alter PR, QRS, or QT intervals (OnlineTable 1). Electrophysiological variables were unchanged at

8-week exercise, but 16-week exercise significantly increasedsinus rhythm CL and Wenckebach CL, consistent withenhanced vagal tone (Online Table 1). RAERP averaged34.4 � 0.7 ms in Ex16 rats compared with 37.0 � 0.9 ms incorresponding Sed controls (p < 0.05).

Figure 1A shows representative recordings during AFinduction. In themajority of Sed rats, no AF could be induced(upper left). In all Ex16 rats, AF was inducible, in some casesself-terminating (nonsustained AF, lower left), in otherssustained for considerable periods (right). Overall sponta-neous (carbachol-free) inducibility rates are shown inFigure 1B. No clear AF-promoting effects were evident with8-week exercise training, but 16-week exercise trainingstrongly increased AF inducibility (e.g., sustained AF in 9 of14 [64%] Ex rats vs. 3 of 20 [15%] Sed rats; p < 0.01). Sus-tained AF was induced by extrastimulation (Fig. 1C) in 5 of14 Ex rats versus none of the Sed rats (p < 0.01). The meanduration of sustained AF in Ex16 rats was 8.8 � 2.5 min.

Five rats in the Ex group showed long-lasting inducedAF (>15min). To assess the role of vagal tone in long-lastingAF, these rats were given intraperitoneal atropine (2 mg/kg)(Online Fig. 4), which terminated AF in all of them. In 4 ofthe 5 rats, AF could not be reinduced after atropine admin-istration. In rats without inducible AF at baseline, inducibilitywas reassessed after cholinergic stimulation with carbachol.Carbachol enhanced AF inducibility and persistence in bothSed and Ex rats, with sustained AF rates postcarbacholof 45% (9 of 20) in Sed rats and 100% (14 of 14) in Ex rats(p < 0.001 vs. Sed rats) (Online Fig. 5). No spontaneous AFor atrial tachycardia episodes occurred on 24-hour telemetry,and the frequency of premature atrial contractions did notdiffer between Sed rats and Ex rats (6� 2 vs. 4� 2 prematureatrial contractions/100,000 beats respectively; p ¼ 0.61).Autonomic tone changes. Mean hourly heart rates inconscious, unrestrained rats are shown in Figure 2. Ex ratsdisplayed slowing of heart rate within 8 weeks (Fig. 2A),which was maintained at 16 weeks (Fig. 2B). Para-sympathetic tone clearly increased at 16-week Ex (Fig. 2C),but sympathetic tone was not significantly altered by exer-cise. Intrinsic heart rate (in presence of both atropine andpropranolol) was not significantly affected by exercise(318� 8 beats/min for Sed rats vs. 338� 8 beats/min for Exrats at 16 weeks), excluding primary sinus-node alterations.Structural remodeling. Figure 3A shows the evolutionof atrial size with exercise. Left atrial (LA) dimension inatrial-diastole increased significantly at 16 weeks. Becauseexercise-related differences might be obscured by body sizediscrepancies, we examined dimensions normalized to bodyweight (Fig. 3B). Atrial dimension/body weight ratiosdecreased with aging-related growth in both groups, but by16 weeks, Ex rat LA diameters increased versus those in Sedrats by 34% and right atrial (RA) diameters by 19%. Fullechocardiographic data are provided in Online Table 2.Within 8 weeks, exercise training caused left ventricularhypertrophy and dilation, LA dilation, and impaired leftventricular diastolic function. LA dilation was accompanied by

Figure 1 AF Vulnerability With High-Intensity Exercise Training

(A) Recordings of burst pacing atrial fibrillation (AF) inductions in a sedentary, noninducible rat (Sed) (upper left) and two 16-week exercise-training (Ex) rats, 1 with inducible

nonsustained AF (lower left) and 1 with long-lasting sustained AF (right). (B) Drug-free inducibility rates for sustained AF (SAF) and nonsustained AF (NSAF). (C) Inducibility-rates

for SAF with extrastimulation techniques. ECG ¼ electrocardiography; EG ¼ electrogram. **p < 0.01, ***p < 0.001, Sed vs. Ex.


71

atrial cardiomyocyte hypertrophy, cell capacitance averaging129 � 9 pF in Ex16 rats versus 74 � 5 pF (p < 0.001) inSed rats.

Representative Picrosirius red–stained photomicrographsare shown in Figure 3C. Colorimetric image analysis indi-cated a 60% statistically significant increase of fibrous tissuecontent in both atria with exercise training (Fig. 3D).Detraining changes. Our data show that intense exercisetraining causes several changes known to promote AF (9):enhanced vagal tone, atrial dilation, and atrial fibrosis. Toobtain additional insights, we studied the time course ofchanges after cessation of 16-week exercise training. Allelectrophysiological differences between exercise-trained andSed rats, including AF indices, resolved within 4 weeks ofdetraining (Online Table 1, Fig. 4A). Similarly, diurnalheart rate differences and parasympathetic tone returned toSed values within 4 weeks of detraining (Fig. 2E). Structuralremodeling responded quite differently. Fibrous tissuecontent remained increased after 8 weeks of detraining(Fig. 4B). LA enlargement similarly failed to recover withdetraining (Fig. 4C, Online Table 3). RA enlargement did

recover at DEx8, but significant RA enlargement remainedafter 4 weeks of detraining. Overall, these results demon-strate relatively rapid reversibility of AF susceptibility, vagalenhancement, and associated heart rate changes, in contrastto delayed and/or incomplete recovery of fibrotic and otheratrial structural alterations. In conjunction with the stronglysuppressive effect of atropine on AF in exercise-trained rats,these results point to vagal enhancement as an importantcomponent of AF promotion with exercise. We thereforedirected our attention to the mechanisms by which exercisetraining enhances vagal tone.Mechanisms of enhanced vagal effects. The determinantsof vagal effects can be broadly divided into 2 components:the control of vagal nerve activity by autonomic reflex arcssuch as the baroreflex and the end-organ response. Baroreflexsensitivity data from one 16-week Sed rat and 1 Ex16 rat areplotted in Figure 5A. Although the tachycardic responses tonitroprusside-induced pressure reductions follow similarslopes, the heart rate slowing caused by phenylephrine wasgreater for any given blood pressure increase in the Ex rat,producing a steeper CL–blood pressure slope. Overall, the

Figure 2 Indices of Autonomic Tone

(A, B) Mean � SEM heart rate (HR) during 24-h ECG recordings in conscious, unrestrained rats at baseline, 8-week, and 16-week training. Shaded areas indicate dark-phase

(night) cycle. n ¼ 7 and 8 (Sed and Ex, respectively). (C) Parasympathetic tone quantification in 8-week Ex rats (n ¼ 9 and 8 for Sed and Ex rats) and in a group studied serially at

16-week Ex and discontinuation of exercise at 4 weeks (DEx4) and at 8 weeks (DEx8) (n ¼ 5/group). (D) Sympathetic tone quantification (p ¼ NS). (E) Telemetry data with

detraining (DEx4 and DEx8 time points). *p < 0.05, Sed vs. Ex. Abbreviations as in Figure 1.


72

bradycardic response to blood pressure elevation withphenylephrine, reflecting vagal enhancement, was unaffectedat 8 weeks of exercise (Fig. 5B) but was approximatelydoubled (p < 0.01) at 16 weeks. Consistent with indices ofvagal tone (sinus CL, Wenckebach CL during EPS, anddiurnal heart rates on telemetry), the bradycardic responsenormalized rapidly with detraining. The cardiac end-organresponse (IKACh amplitude produced by muscarinic cholin-ergic stimulation) to 10-nM, 100-nM, and 1-mM carbacholin 16-week Sed rat and Ex rat atrial cardiomyocytes isillustrated in Figure 6A. Mean current density voltagerelations indicate enhanced responses to carbachol at the2 lower concentrations, with no intergroup differences at thehighest concentration; these results were confirmed with10 mM of carbachol (Online Fig. 6A). Current kinetics wereunaltered (Online Fig. 6B). Figure 6B shows IKAChconcentration-response curves, which suggest enhancedsensitivity to cholinergic stimulation at 16 weeks of exercise(EC50 199 vs. 722 nmol/l, Ex vs Sed).

To explore potential underlying molecular mechanisms,we assessed the gene expression of relevant subunits. Theenhanced IKACh response cannot be attributed to increasedexpression of IKACh channel subunits, because Kir3.4 wasunchanged and Kir3.1 slightly down-regulated by exercise(Fig. 6C). We therefore examined components of the Gprotein–coupled receptor-signaling systems that mediate/modulate IKACh. There were no significant changes in the

expression of alpha-adrenergic, beta-adrenergic, or musca-rinic cholinergic receptors (Online Fig. 7A). Rats withlong-lasting AF had significantly greater Kir3.4-subunit andM3-receptor mRNA expression than rats with shorter-lasting AF, along with a trend to slower heart rates(Online Fig. 8), suggesting that cholinergic moleculardeterminants might contribute to AF maintenance.

The principal G proteins mediating cholinergic signaltransduction (Gai1-3) were unchanged with exercise (OnlineFig. 7). Gao was down-regulated but does not controlvagal responses (11). Gaq was up-regulated, possibly con-tributing to exercise-induced structural remodeling (12).RGS proteins modulate cardiac autonomic responses (13),particularly RGS4, which negatively controls cholinergicsignaling (14). We observed significant changes in manymembers of the RGS family, primarily down-regulation ofRGS3, 4, 10, 12, 14, and 16, in Ex rats versus Sed rats(Fig. 7A). To assess the potential role of RGS down-regulation in increased AF susceptibility, we evaluatedarrhythmia inducibility in RGS4 KO-mice versus wild-typelittermate controls. Both wild-type and RGS4KO-mice hadnormal ECG and intracardiac ECG traces at baseline(Fig. 7B). AF could not be induced under any of theconditions tested in wild-type mice but was readily inducedin the presence of carbachol in RGS4KO mice (Fig. 7C).Overall, RGS4 deficiency strongly enhanced carbachol-related AF susceptibility (Fig. 7D). These observations are

Figure 3 Structural Remodeling

Echocardiographic atrial dimensions: (A) raw and (B) body-weight indexed results. (C) Photomicrographs of Picrosirius red–stained samples. A larger amount of thicker,

red-stained fibrotic strands is present in trained rats. (D) Investigator-blinded Picrosirius red quantification in 16-week Ex rats and Sed controls. n ¼ 5 rats/group.

**p < 0.01, ***p < 0.001, Sed vs. Ex. 8w ¼ 8 weeks; 16w ¼ 16 weeks; Bsl ¼ baseline; LA ¼ left atrial; RA ¼ right atrial; other abbreviations as in Figure 1.


73

consistent with the notion that exercise training–inducedRGS down-regulation contributes to cholinergic enhance-ment and associated AF susceptibility.

Discussion

We developed, for the first time, an animal model of theAF-promoting substrate due to endurance exercise training,an important clinical paradigm (4–8), and explored theunderlying mechanisms. We found several potential patho-physiological determinants, including vagal enhancement,atrial dilation, and atrial fibrosis. The time-course data andresponse to atropine suggest that vagal enhancement isparticularly important. We then analyzed mechanismsunderlying vagal changes, finding both baroreflex enhance-ment and increased sensitivity of IKACh to cholinergicstimulation, and identified a potential molecular basis,down-regulation of RGS proteins, that negatively controlsthe G protein signaling mediating cholinergic activationof IKACh.Endurance training and AF-related remodeling. Ourexercise rats displayed a cardiac phenotype similar to trainedathletes, including eccentric ventricular hypertrophy andatrial dilation (15). Enhanced AF vulnerability was not

evident at 8 weeks of exercise training, requiring 16 weeks todevelop (Fig. 1). Similarly, athletes only develop AF afterprolonged high-intensity training (4–6), with shortertraining periods not appreciably increasing AF risk (16).Ongoing exercise practice is an important factor (17),consistent with our data showing relatively rapid resolutionof AF promotion with detraining.

Little is known about the mechanisms of AF promotionby high-intensity endurance training in humans, despite itsclinical importance (4–8). We noted 3 potential contributorsin our model: atrial dilation, fibrosis, and enhanced vagaltone. Atrial dilation, a typical feature of athlete’s heart (15),increases atrial substrate size, a significant determinant ofreentrant AF susceptibility (18). Atrial dilation per se is notenough to sustain AF in young competitive athletes (16).Similarly, atrial dilation alone was insufficient to promoteAF vulnerability in our model: AF vulnerability resolvedrapidly with detraining despite slow or absent resolutionof atrial dilation (Fig. 4). The limited reversibility of LAdilation in our model is consistent with clinical observationsof limited reversibility and persistent LA dilation withdetraining (19–21).

Myocardial fibrosis is a recognized pathophysiologicalfactor; both physical interruption of cardiac muscle bundle

Figure 4 AF Inducibility and Structural Remodeling With DEx4 and DEx8 Detraining After 16-Week Exercise

(A) AF inducibility. DEx4: n ¼ 20 and 8 for Sed and Ex, respectively; DEx8: n ¼ 12 and 7 for Sed and Ex. (B) Fibrosis quantification. n ¼ 3 and 5 for Ex and Sed at each time

point. (C) Atrial dimensions in 16-week exercise and detraining groups. No significant interaction but group effect for left and middle panels. Sed: n ¼ 15 for 16 weeks and

DEx4, n ¼ 10 for DEx8; Ex: n ¼ 10 for 16 weeks and DEx4, n ¼ 6 for DEx8. *p < 0.05, **p < 0.01, ***p < 0.001, Sed vs. Ex. Abbreviations as in Figures 1, 2, and 3.


74

continuity (22) and fibroblast cardiomyocyte interactions(23) may contribute. The 30% to 60% increase in fibroustissue content we observed is well below the 5- to 15-fold

Figure 5 Results of In Vivo Baroreflex Sensitivity Analysis

(A) Quantification of baroreflex responses to nitroprusside (NTP)-induced hypo-

tension (cycle length [CL] decreases) and phenylephrine (Phe)-induced hyperten-

sion (CL increases) for individual Sed (blue) and Ex (red) rats. Lines are best-fit

linear regression lines to the corresponding data points, with slope value (m)

shown. (B) Mean m values for the bradycardic and tachycardic baroreflex in

sedentary and trained groups at each time point. n ¼ 8/group for Sed and Ex 8

weeks; n ¼ 6/group for Sed and Ex 16 weeks; DEx4: n ¼ 6 for Sed, n ¼ 4 for Ex;

DEx8: n ¼ 6 for Sed, n ¼ 2 for Ex. **p < 0.01, Sed vs. Ex. BP ¼ blood pressure;

other abbreviations as in Figures 1 and 3.

increase in AF associated with heart failure in dog models(22,24) and in a mouse model of transforming growth factorbeta overactivity (25). This limited fibrous tissue accumula-tion may explain the lack of AF promotion, despite persistentfibrosis, in the DEx4 and DEx8 groups (Fig. 4). Thus, as foratrial dilation, while tissue fibrosis likely contributes toexercise-related AF promotion, it is by itself insufficient.

AF in athletes who are otherwise healthy has generallybeen considered “lone AF.” Because the presence of atrialfibrosis and dilation in trained rats indicates atrial structuralremodeling, AF cannot be considered “lone” in our model.Atrial dilation is also reported clinically in athletes(15,19,20); thus, the “lone” classification of AF in highlytrained athletes should perhaps be reconsidered. It would beinteresting to evaluate the possibility of atrial fibrosis inendurance athletes by using noninvasive methods such asdelayed-enhancement magnetic resonance imaging.

Enhanced parasympathetic tone was clearly important inAF promotion in our model. Atropine terminated long-lasting AF and suppressed AF inducibility in Ex rats, andthe time course of vagal indices matched that of AFinducibility during exercise training and detraining. Vagalactivation is well recognized to promote AF by inducingspatially heterogeneous atrial refractoriness decreases (18).Consistent with this mechanism, RAERP was shortened inour trained rats. Vagal enhancement is a well-recognizedconsequence of endurance training, and most AF episodesin athletes develop in situations of predominant vagal tone(26). Conversely, vagal enhancement alone is insufficient tofully explain exercise-related AF because autonomic tonechanges within weeks of clinical exercise training onset (27),

Figure 6 Exercise-Induced Changes in End-Organ Response to Cholinergic Stimulation

(A) Representative acetylcholine-regulated potassium current (IKACh) recordings (current in presence of carbachol minus baseline current in same cell) in Sed and Ex groups,

with mean � SEM voltage current density relationships at 10 nM (n ¼ 12/5 cells/rats for Sed, n ¼ 10/4 for Ex), 100 nM (n ¼ 9/3 cells/rats for each group) and 1 mM(n ¼ 9/4 cells/rats for each group) carbachol (CCh). (B) CCh concentration-response curves obtained from IKACh results at �120 mV. (C) Messenger ribonucleic acid (mRNA)

expression of IKACh Kir3.x ion-channel subunits. n ¼ 10 rats/group. *p < 0.05, **p < 0.01, ***p < 0.001, Sed vs Ex. Abbreviations as in Figure 1.


75

whereas AF susceptibility takes longer to develop (4–6,16).Thus, enhanced vagal tone is likely a necessary but insuffi-cient condition for exercise-induced AF promotion, withother factors such as fibrosis and atrial dilation alsocontributing.Molecular mechanisms underlying exercise-inducedatrial remodeling. Although vagal enhancement is a char-acteristic feature of highly trained athletes, the underlyingbasis is largely unknown. Exercise training enhances barore-flex sensitivity in patients with heart failure (28), consistentwith our observations. Our study is the first to examine IKAChchanges in an exercise model, and our findings suggestenhanced cardiac responsiveness. RGS protein down-regulation likely played a significant role in increasingcholinergic sensitivity, as most endogenous RGS proteins arenegative regulators of cholinergic effects (13,14,29). RGS4reduces cholinergic effects through bg5 G-protein subunitmodulation (14). RGS10 similarly accelerates IKACh deacti-vation (30). Our novel observation that RGS4 deletionenhances cholinergically mediated AF vulnerability is a proofof principle for the potential arrhythmogenic role of exercise-induced RGS dysregulation. Of note, RGS9 was the onlyRGS-protein that was up-regulated with exercise in ourstudy. In contrast to other RGS proteins, RGS9 positivelyregulates muscarinic cholinergic signaling (31). The increase

in Gaq-expression that we noted may have contributed toatrial fibrosis (12).Study limitations. Extrapolation from animal models tohumans should always be done with caution. In a previouswork with this model, we estimated that a 16-week exerciseprogram in the rat corresponds roughly to approximately10 years of exercise training in humans, at about 85% to 90%maximum oxygen uptake (3). In a young individual, runningat 7 miles/h or swimming at 75 yards/min would providecomparable exercise intensity (32).

The nature and reversibility of changes in atrial size,ventricular dimensions, mass, and function in the currentmodel correlate well with observations in humans (15,16). Ina previous study of ventricular function and ventriculartachycardia susceptibility in the same animal model (3), wenoted activation of atrial collagen genes with exercisetraining, which reversed with detraining. The persistentfibrosis with detraining noted in our study suggests thatalthough collagen synthesis may return to baseline withdetraining, the already formed collagen is not broken down.

We showed increased AF inducibility and mechanisti-cally characterized the arrhythmogenic substrate but didnot observe spontaneous arrhythmic events. Patients withexercise-related AF must also have AF-inducing triggers,which were not evident in our model. We attempted to

Figure 7 Potential Role of RGS Protein

(A) Right atrial expression levels of regulators of G protein–signaling (RGS) mRNA in Sed and Ex rats. *p < 0.05, Sed vs Ex; n ¼ 10 per group. (B) Surface ECG and intracardiac

ECG (iECG) recordings from wild-type (WT) and RGS4-knockout (RGS4KO) mice (; ¼ QRS complexes;6 ¼ P waves). (C) Intracardiac and ECG recordings on atrial burst pacing

(ATP, black bar) after 10 mg/kg of intraperitoneal CCh. Magnified views of atrial activity on iECG are shown in inset. (D) AF inducibility at baseline (drug-free) and after CCh in WT

controls and RGS4KO mice. **p < 0.01, ***p < 0.001, WT vs. RGS4KO. Other abbreviations as in Figures 1 and 6.


76

quantify RGS proteins by using western blotting withcommercially available antibodies but were unsuccessfulbecause of poor specificity. Therefore, our findings pointingto a role of RGS changes in parasympathetic enhancementare based solely on mRNA expression, which does notalways correlate with protein expression.

Conclusions

We found that sustained, high-intensity endurancetraining induces an atrial arrhythmogenic substrate thatincludes atrial dilation, myocardial fibrosis, and autonomicimbalance. Vagal enhancement was an essential contributorto AF susceptibility, and was caused by increased baroreflexsensitivity and atrial myocyte acetylcholine-hyperre-sponsiveness related to RGS-protein downregulation.

AcknowledgmentsThe authors thank France Thériault for excellent secretarialhelp and Nathalie L’Heureux, Patrick Lawler, and ChantalSt-Cyr for experimental assistance.

Reprint requests and correspondence: Dr. Stanley Nattel,Montreal Heart Institute, University of Montreal Research Center,5000 Belanger Street E., Montreal, Quebec H1T 1C8, Canada.E-mail: [email protected].

REFERENCES

1. Lawler PR, Filion KB, Eisenberg MJ. Efficacy of exercise-based cardiacrehabilitation post-myocardial infarction: a systematic review and meta-analysis of randomized controlled trials. Am Heart J 2011;162:571–84.

2. La GA, Robberecht C, Kuiperi C, et al. Lower than expecteddesmosomal gene mutation prevalence in endurance athletes withcomplex ventricular arrhythmias of right ventricular origin. Heart 2010;96:1268–74.

3. Benito B, Gay-Jordi G, Serrano-Mollar A, et al. Cardiac arrhythmo-genic remodeling in a rat model of long-term intensive exercisetraining. Circulation 2011;123:13–22.

4. Molina L, Mont L, Marrugat J, et al. Long-term endurance sportpractice increases the incidence of lone atrial fibrillation in men:a follow-up study. Europace 2008;10:618–23.

5. Baldesberger S, Bauersfeld U, Candinas R, et al. Sinus node disease andarrhythmias in the long-term follow-up of former professional cyclists.Eur Heart J 2008;29:71–8.

6. Grimsmo J, Grundvold I, Maehlum S, Arnesen H. High prevalence ofatrial fibrillation in long-term endurance cross-country skiers: echo-cardiographic findings and possible predictorsda 28-30 years follow-up study. Eur J Cardiovasc Prev Rehabil 2010;17:100–5.

7. Aizer A, Gaziano JM, Cook NR, Manson JE, Buring JE, Albert CM.Relation of vigorous exercise to risk of atrial fibrillation. Am J Cardiol2009;103:1572–7.

8. Mont L, Elosua R, Brugada J. Endurance sport practice as a risk factorfor atrial fibrillation and atrial flutter. Europace 2009;11:11–7.

9. Iwasaki YK, Nishida K, Kato T, Nattel S. Atrial fibrillationpathophysiology: implications for management. Circulation 2011;124:2264–74.

10. Nattel S. From guidelines to bench: implications of unresolved clinicalissues for basic investigations of atrial fibrillation mechanisms. Can JCardiol 2011;27:19–26.

11. Boknik P, Grote-Wessels S, Barteska G, et al. Genetic disruption of Gproteins, G(i2)alpha or G(o)alpha, does not abolish inotropic and

mailto:[email protected]


77

chronotropic effects of stimulating muscarinic cholinoceptors in atrium.Br J Pharmacol 2009;158:1557–64.

12. Nattel S. G-protein signaling and arrhythmogenic atrial remodeling:relevance to novel therapeutic targets in atrial fibrillation. HeartRhythm 2009;6:85–6.

13. Gu S, Cifelli C, Wang S, Heximer SP. RGS proteins: identifying newGAPs in the understanding of blood pressure regulation and cardio-vascular function. Clin Sci (Lond) 2009;116:391–9.

14. Cifelli C, Rose RA, Zhang H, et al. RGS4 regulates parasympatheticsignaling and heart rate control in the sinoatrial node. Circ Res 2008;103:527–35.

15. Fagard R. Athlete’s heart. Heart 2003;89:1455–61.16. Pelliccia A, Maron BJ, Di Paolo FM, et al. Prevalence and clinical

significance of left atrial remodeling in competitive athletes. J Am CollCardiol 2005;46:690–6.

17. Elosua R, Arquer A, Mont L, et al. Sport practice and the risk oflone atrial fibrillation: a case-control study. Int J Cardiol 2006;108:332–7.

18. Liu L, Nattel S. Differing sympathetic and vagal effects on atrialfibrillation in dogs: role of refractoriness heterogeneity. Am J Physiol1997;273:H805–16.

19. Hoglund C. Enlarged left atrial dimension in former enduranceathletes: an echocardiographic study. Int J Sports Med 1986;7:133–6.

20. Pelliccia A, Maron BJ, De LR, Di Paolo FM, Spataro A, Culasso F.Remodeling of left ventricular hypertrophy in elite athletes after long-term deconditioning. Circulation 2002;105:944–9.

21. Heidbuchel H, Anne W, Willems R, Adriaenssens B, Van de Werf F,Ector H. Endurance sports is a risk factor for atrial fibrillation afterablation for atrial flutter. Int J Cardiol 2006;107:67–72.

22. Burstein B, Comtois P, Michael G, et al. Changes in connexinexpression and the atrial fibrillation substrate in congestive heart failure.Circ Res 2009;105:1213–22.

23. Yue L, Xie J, Nattel S. Molecular determinants of cardiac fibroblastelectrical function and therapeutic implications for atrial fibrillation.Cardiovasc Res 2011;89:744–53.

24. Li D, Fareh S, Leung TK, Nattel S. Promotion of atrial fibrillation byheart failure in dogs: atrial remodeling of a different sort. Circulation1999;100:87–95.

25. Verheule S, Sato T, Everett T, et al. Increased vulnerability to atrialfibrillation in transgenic mice with selective atrial fibrosis caused byoverexpression of TGF-beta1. Circ Res 2004;94:1458–65.

26. Mont L, Sambola A, Brugada J, et al. Long-lasting sport practice andlone atrial fibrillation. Eur Heart J 2002;23:477–82.

27. Collier SR, Kanaley JA, Carhart R Jr., et al. Cardiac autonomicfunction and baroreflex changes following 4 weeks of resistance versusaerobic training in individuals with pre-hypertension. Acta Physiol(Oxf) 2009;195:339–48.

28. Pietila M, Malminiemi K, Vesalainen R, et al. Exercise training inchronic heart failure: beneficial effects on cardiac (11)C-hydroxyephe-drine PET, autonomic nervous control, and ventricular repolarization.J Nucl Med 2002;43:773–9.

29. Fu Y, Huang X, Zhong H, Mortensen RM, D’Alecy LG, Neubig RR.Endogenous RGS proteins and Galpha subtypes differentially controlmuscarinic and adenosine-mediated chronotropic effects. Circ Res2006;98:659–66.

30. Bender K, Nasrollahzadeh P, Timpert M, Liu B, Pott L, Kienitz MC.A role for RGS10 in beta-adrenergic modulation of G-protein-activated Kþ (GIRK) channel current in rat atrial myocytes. J Physiol2008;586:2049–60.

31. Laroche G, Giguere PM, Roth BL, Trejo J, Siderovski DP. RNAinterference screen for RGS protein specificity at muscarinic andprotease-activated receptors reveals bidirectional modulation of signal-ing. Am J Physiol Cell Physiol 2010;299:C654–64.

32. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium ofphysical activities: an update of activity codes and MET intensities.Med Sci Sports Exerc 2000;32:S498–504.

Key Words: arrhythmia mechanisms - exercise training - ion currents -

potassium channel.

APPENDIX

For supplementary materials, tables, and figures, please see the onlineversion of this article.

atrial fibrillation promotion by endurance exerciseatrial fibrillation promotion by endurance...

Documents