Exercise Training Stabilizes RyR2-Dependent Ca2+ Release in Post-infarction Heart Failure

Aim: Dysfunction of the cardiac ryanodine receptor (RyR2) is an almost ubiquitous finding in animal models of heart failure (HF) and results in abnormal Ca2+ release in cardiomyocytes that contributes to contractile impairment and arrhythmias. We tested whether exercise training (ET), as recommended by current guidelines, had the potential to stabilize RyR2-dependent Ca2+ release in rats with post-myocardial infarction HF. Materials and Methods: We subjected male Wistar rats to left coronary artery ligation or sham operations. After 1 week, animals were characterized by echocardiography and randomized to high-intensity interval ET on treadmills or to sedentary behavior (SED). Running speed was adjusted based on a weekly VO2max test. We repeated echocardiography after 5 weeks of ET and harvested left ventricular cardiomyocytes for analysis of RyR2-dependent systolic and spontaneous Ca2+ release. Phosphoproteins were analyzed by Western blotting, and beta-adrenoceptor density was quantified by radioligand binding. Results: ET increased VO2max in HF-ET rats to 127% of HF-SED (P < 0.05). This coincided with attenuated spontaneous SR Ca2+ release in left ventricular cardiomyocytes from HF-ET but also reduced Ca2+ transient amplitude and slowed Ca2+ reuptake during adrenoceptor activation. However, ventricular diameter and fractional shortening were unaffected by ET. Analysis of Ca2+ homeostasis and major proteins involved in the regulation of SR Ca2+ release and reuptake could not explain the attenuated spontaneous SR Ca2+ release or reduced Ca2+ transient amplitude. Importantly, measurements of beta-adrenoceptors showed a normalization of beta1-adrenoceptor density and beta1:beta2-adrenoceptor ratio in HF-ET. Conclusion: ET increased aerobic capacity in post-myocardial infarction HF rats and stabilized RyR2-dependent Ca2+ release. Our data show that these effects of ET can be gained without major alterations in SR Ca2+ regulatory proteins and indicate that future studies should include upstream parts of the sympathetic signaling pathway.


INTRODUCTION
Current guidelines recommend exercise training (ET) as part of rehabilitation programs after myocardial infarction (MI) and for patients with heart failure (HF) (1,2). ET has a range of beneficial cardiovascular effects (3) including improved aerobic capacity and cardiac contractile function (4)(5)(6) and decreases mortality after MI and in patients with HF (7,8). However, approximately half of all patients with HF still die of ventricular arrhythmias (9). To exploit fully the therapeutic potential of ET in HF, we need a better understanding of its effects on the mechanisms for arrhythmias in HF.
The mechanism that underlies increased risk of arrhythmias in post-MI HF is multifactorial and includes fibrosis, altered expression or function of ion channels, and perturbed Ca 2+ homeostasis (10). An almost ubiquitous finding in animal models of post-MI HF is a dysfunction of the cardiac ryanodine receptor (RyR2), i.e., the Ca 2+ release channel of the sarcoplasmic reticulum (SR) (11,12). This channel is essential in normal cellular Ca 2+ handling and excitation-contraction coupling (13). RyR2 dysfunction can lead to diastolic Ca 2+ leak from the SR, which contributes to contractile impairment and arrhythmias (14). Inhibition of SR Ca 2+ leak has been found to prevent arrhythmias in animal models of HF, as well as arrhythmogenic events in cardiac tissue taken from patients with HF, and is a potential future therapeutic strategy (15,16). However, no drugs that specifically target RyR2 are clinically available. Based on results from mouse models that show RyR2 dysfunction, ET might have the potential to stabilize RyR2 function and prevent arrhythmias (17,18). However, there is a scarcity of data that support such an effect of ET in post-MI HF, and therefore, there is a need for rigorous experimental data from clinically relevant models.

Animal Model of Post-infarction Heart Failure
A total of 52 male Wistar rats were included in the study. They either underwent sham operations or were subjected to MI by left coronary artery ligation. Left coronary artery ligation was performed through a thoracotomy under general anesthesia, which was achieved by inhalation of 65% N 2 O, 32% O 2 , and 2.5% isoflurane through an endotracheal tube, as previously described (19). Through a 2-cm incision in the skin over the sternum, the cutis on the left thorax was loosened from the underlying layer, and a left thoracotomy was performed in the fourth intercostal space. The pericardial sac was opened, and the left coronary artery was ligated ∼1 mm beneath the left atrium. In the control (sham) rats, the same operating procedure was employed, but the left coronary artery ligation was not performed. The skin was closed by sutures. Buprenorphine 0.2 mg/kg was administered subcutaneously for analgesia.
Rats with MI were evaluated after 1 week by 2D M-mode echocardiography. Based on previously established criteria, rats that exhibited a left atrial diameter of >5.0 mm were included in the HF group (15,17). Rats within the same group (sham or HF) were paired according to weight and were randomly assigned either to a 5-week high-intensity ET program or to a sedentary, i.e. non-exercising, control group. The rats were housed together in cages under a 12:12 h light:dark cycle, with free access to water and food.

Protocol for Exercise Training
ET was initiated 1 week after coronary artery ligation. Separate treadmill and metabolic chambers were used for exercise training and for weekly measurements of maximal oxygen uptake (VO 2max ) (Columbus Instruments, OH, USA), respectively. Three days before the first VO 2max -test, rats were habituated to the treadmill with daily 15-min exercise bouts at walking pace.
High-intensity exercise training has been shown to increase aerobic capacity in rats more than moderate exercise training (6). Pilot data, as well as previous publications, showed that the effect on VO 2max reached a plateau at 5 weeks (6,20). Therefore, all ET rats completed a 5-week high-intensity ET protocol of treadmill running. Each training session lasted for 1 h. It consisted of a 10-min warm-up period and five 8-min intervals at 80-90% of the running speed that had been found to produce VO 2max , interspersed by 2-min rest periods at 60% running speed. The treadmill was set at an inclination of 25 • at all times. The running speed was individualized for all rats according to a weekly VO 2max -test to keep the intensity constant throughout the ET program. SED rats ran on the treadmill at 60% of the speed that was found to produce VO 2max for 15 min, 2 days per week. All rats in the ET groups continued daily exercise until 1 day before euthanasia.

Echocardiography
Echocardiographic parameters were recorded with a Vevo2100 System (Fujifilm VisualSonics Inc., Canada) in rats that were anesthetized by a mixture of 65% N 2 O, 33% O 2 , and 2% isoflurane by mask ventilation. Echocardiography was performed 1 week after coronary ligation to enable stratification, and it was repeated at completion of the ET program (21).

Cardiomyocyte Calcium Imaging
Ventricular cardiomyocytes were isolated using constantpressure perfusion of the coronary arteries with collagenasecontaining solution (solution containing in mM 130 NaCl, 25 HEPES, 22 D-glucose, 5.4 KCl, 0.5 MgCl 2 , 0.4 NaH 2 PO 4 , adjusted to pH 7.4 with NaOH), as previously described (22). The hearts were mounted on a Langendorff setup and were perfused retrogradely through the aorta with a 37 • C solution that contained 0.08 mM Ca 2+ and 200 U/ml collagenase type 2 (Worthington Biochemical Corporation, Lakewood, NJ, USA). The hearts were perfused for 20 min before the left ventricular tissue was excised rapidly, and the infarcted area was removed.
The ventricular tissue was then cut into small pieces and was gently mixed with a cutoff Pasteur pipette for about 1 min in a buffer that contained 1% bovine serum albumin (BSA) (Sigma Aldrich) and 0.02 U/ml deoxyribonuclease 1 (Worthington Biochemical Corporation, Lakewood, NJ, USA). The solution with the ventricular tissue was then filtered through a nylon mesh (pore diameter, 200 µm). After sedimentation, the cells were resuspended three times in 1% BSA solutions with increasing Ca 2+ concentration (0.1, 0.2, and 0.5 mM, respectively). The cells were used for experiments within 10 h.
Whole-cell Ca 2+ imaging was performed with a Zeiss Axiovert 200M microscope (Carl Zeiss Microscopy, LCC, NY, USA). Leftventricular cardiomyocytes were field stimulated at 1, 2, and 4 Hz by a 3-ms symmetrical bipolar pulse, which was ∼20% above the voltage threshold for contraction of the individual cardiomyocytes. The experimental superfusate was based on modified HEPES-Tyrode's solution, and contained (in mM) HEPES 5, NaCl 140, KCl 5.4, MgCl 2 0.5, glucose 5.5, NaH 2 PO 4 0.4, and CaCl 2 1. pH was adjusted to 7.4 with NaOH. Cytosolic Ca 2+ was visualized the by use of 5 µM fluo-4 acetomethyl (AM) ester (Molecular Probes, Eugene, Oregon, USA), with 10-min loading before experiments. All experiments were performed at 37 • C. For beta-adrenoceptor activation, 10 nM isoprenaline sulfate (ISO) (NAF, Norway) was added to the modified HEPES-Tyrode's solution. After addition of ISO, the Ca 2+ transient amplitude reached steady state after ∼1 min. The cardiomyocytes were stimulated for another 30 s before measurements were made of the Ca 2+ transients. Ca 2+ waves were measured in a 10-s pause after the stimulation period. SR Ca 2+ removal rate was estimated as the average rate of decay (k = 1/tau) that was taken from the last three Ca 2+ transients before a pause. SR Ca 2+ content was measured as peak fluorescence after rapid application of 10 mM caffeine at 1 Hz in the presence and absence of ISO. SERCA2dependent Ca 2+ removal was measured as the difference between the decay rate of Ca 2+ transients and the decay rate of SR Ca 2+ release after exposure to 10 mM caffeine. Ca 2+ removal rate by the sodium-calcium exchanger (NCX) and plasma membrane Ca 2+ ATPase (PMCA) was measured as the decay rate of the caffeine-induced Ca 2+ release.
The relative increase in diastolic Ca 2+ during pacing was measured as the difference in mean F 0 between 1, 2, and 4 Hz for three Ca 2+ transients after 30 s stimulation.
Fractional release was measured as mean F of 3 Ca 2+ transients after 30 s stimulation divided with F for caffeine release.
Confocal microscopy line-scan imaging was performed using a Zeiss LSM 7 Live confocal microscope (Carl Zeiss Microscopy, LCC, NY, USA). Cardiomyocytes were stimulated at 1 and 4 Hz in a protocol that was similar to the one described for wholecell Ca 2+ imaging measurements of Ca 2+ waves, but with a 6-s post-stimulation rest period in which Ca 2+ sparks were recorded.

Beta-Adrenoceptor Quantification by Radioligand Binding
To assess beta-adrenoceptor density, a radioligand binding assay was performed on left ventricles snap frozen in liquid nitrogen. Crude cell membrane fractions were prepared as described by Krobert et al. (23). Radioligand binding was performed as described by Ramberg et al. (24), where membranes were incubated with [ 125 I]-(-)iodocyanopincolol (0.066 nM) and the indicated concentration of either CGP20712A or ICI118551 for 2 h at 23 • C. Data were fitted to a two-site binding model, and high and low binding affinities (pK i ) were calculated in GraphPad Prism 8.0.1 using a K d of 0.04 nM (affinity of [ 125 I]-(-)iodocyanopincolol was determined in the left ventricular membranes). Beta 1 -adrenoceptor density was determined as an average of high-affinity CGP20712A and low-affinity ICI118551 binding in the same heart. Similarly, beta 2 -adrenoceptor density was determined as an average of high-affinity ICI118551 and low-affinity CGP20712A binding in the same heart.

Statistical Analysis
All experiments and analyses were performed by investigators who were blinded to the phenotype and group identity of each cell and animal. Statistical tests were selected based on advice from an external statistical advisor. A paired Student's t-test was used to compare the effect of exercise training on VO 2max at baseline and after 5 weeks within each group. Unpaired nested ANOVA was used to compare ET and SED for all results from whole-cell Ca 2+ imaging. In addition, twoway ANOVA was used to compare difference between ET and SED in the presence or absence of ISO at all frequencies for Ca 2+ waves, Ca 2+ transient amplitude, and Ca 2+ decay rate. A unpaird Student's t-test was used to compare differences in Ca 2+ waves frequency across frequencies and in absense and presence of ISO for HF cardiomyocytes. Ca 2+ spark frequency was analyzed by application of a Poisson test to adjust for skewed distribution. One-way ANOVA was used when comparing betaadrenoceptors. Bonferroni, Holm-Sidak, or Tukey's correction was performed when appropriate. Normal distribution was assessed with Shapiro-Wilk test. All statistics were performed by using IBM SPSS statistics 27, Sigmaplot 12.5 or R software (version 3.0.2, The R Foundation for Statistical Computing). P < 0.05 was considered statistically significant. Results are reported as the mean ± standard error of mean (SEM).

Heart Failure Model, Exercise Training, and Aerobic Capacity
We performed left coronary artery ligation on 23 male Wistar rats to induce large left ventricular myocardial infarctions. Sham operations were performed on 29 rats. One week after surgery, all rats were examined by echocardiography before randomization to ET or standard sedentary conditions (SED) (Figure 1A).
One week after surgery, the aerobic capacity of HF-ET rats was 86% of that of the sham-ET rats (P < 0.05) but did not differ from that of the HF-SED rats ( Figure 1B). During the ET protocol, VO 2max increased in both ET groups, as expected. After 5 weeks of ET, VO 2max in the HF-ET group was 127% of that of the HF-SED group (P < 0.05), while VO 2max in the sham-ET group was 119% of that of the sham-SED group (P < 0.05, Figure 1B). However, VO 2max in the HF-ET group was only 88% of that of the sham-ET group (Figure 1B, P < 0.05).
Echocardiography was performed before randomization and repeated after completion of the ET protocol ( Figure 1C). As per definition, the left atrial diameter was larger in the HF groups than in the sham groups at the time of randomization 1 week after surgery and had increased further in both HF groups when examined at week 6 (P < 0.05, Figure 1D). One week after surgery, HF rats also showed a clear phenotype with increased left ventricular diameter and reduced fractional shortening (P < 0.05, Figures 1E,F). During 5 weeks of ET, the left ventricular diastolic diameter ( Figure 1E) and fractional shortening remained unchanged in the HF-ET group ( Figure 1F), and ET did not affect any of the echocardiographic parameters in sham-operated rats (Figures 1D-F).

Spontaneous Calcium Release in Cardiomyocytes From HF Rats
We isolated cardiomyocytes from the left ventricles of the HF-ET and HF-SED rats at week 6 for the analysis of RyR2dependent spontaneous Ca 2+ release. We first recorded cardiomyocyte-wide propagated Ca 2+ release events, i.e. Ca 2+ waves, after a train of electrical stimuli at increasing frequencies (Figures 2A,B). We performed this protocol in the absence and presence of ISO to simulate sympathetic activation with beta-adrenoceptor stimulation. Ca 2+ wave frequency increased with increasing pacing frequency and during exposure to ISO (Figure 2B). Importantly, Ca 2+ wave frequency during ISO was lower in HF-ET cardiomyocytes than in those of HF-SED rats at 1 Hz ( Figure 2B) and when considered across all frequencies with two-way ANOVA ( Figure 2C). This finding indicated that ET stabilized Ca 2+ handling in HF.
We also measured local Ca 2+ release events, i.e., Ca 2+ sparks, by confocal microscopy after a train of electrical stimuli ( Figure 2D). The frequency of Ca 2+ sparks increased with stimulation frequency (from 1 to 4 Hz), both in the presence and absence of ISO (P < 0.05), and was clearly lower in the HF-ET than in the HF-SED cardiomyocytes ( Figure 2E). This further indicated that ET stabilized SR Ca 2+ handling in HF.

Systolic Ca 2+ Release in Cardiomyocytes From HF Rats
If ET is to be used as an antiarrhythmic intervention, negative effects on systolic Ca 2+ release responsible for contraction should not outweigh the positive effects of ET on RyR2-dependent spontaneous Ca 2+ release associated with arrhythmias. Surprisingly, however, when we analyzed Ca 2+ transients across different pacing frequencies, the amplitude was lower in HF-ET than HF-SED both in the absence and presence of ISO (Figures 3A-D). Ca 2+ transient decay rate was also lower in HF-ET than HF-SED, but only during exposure to ISO (Figures 3E,F). To explain these findings, SR Ca 2+ content and Ca 2+ removal were analyzed during caffeine exposure ( Figure 3G). However, no difference in SR Ca 2+ content was observed between HF-ET and HF-SED ( Figure 3H) nor could the slowed removal be attributed to SERCA2-dependent or sarcolemmal Ca 2+ removal alone (Figures 3I,J). We also analyzed the increase in diastolic Ca 2+ that is expected in response to increased pacing frequencies but found no difference between HF-ET and HF-SED ( Figure 3K). Lastly, we calculated fractional SR Ca 2+ release but only observed a small increase in HF-ET compared HF-SED during ISO, which cannot explain stabilized RyR2-dependent SR Ca 2+ release ( Figure 3L) (25).

Phosphoproteins in Left Ventricles From HF Rats
To explore other potential explanations for the stabilization of RyR2-dependent spontaneous Ca 2+ release that was observed after ET, phosphoproteins levels were quantified in HF-ET and HF-SED (Figure 4). We focused on RyR2 and the Frontiers in Cardiovascular Medicine | www.frontiersin.org  SERCA/Plb system as the proteins that are mainly responsible for SR Ca 2+ release and reuptake. However, no differences in these proteins or their major regulatory phosphorylation sites were observed (Figures 4A-I). The Plb/SERCA2a ratio was also unaltered in HF-ET compared with that in HF-SED (P = 0.68).

Effects of Exercise Training in Sham-Operated Rats
To control for untoward effects of ET, we performed most experiments in parallel in sham-ET and sham-SED rats (Figures 5, 6). The ET protocol lowered Ca 2+ wave frequency even in sham-ET compared to sham-SED both in the presence and absence of ISO when compared across all frequencies with two-way ANOVA (Figures 5A-C). However, Ca 2+ spark frequency was similar in sham-ET and sham-SED (Figures 5D,E). ET had no effect on Ca 2+ transients amplitude (Figures 6A-C), Ca 2+ removal rate, diastolic Ca 2+ (Figures 6D-F), SR Ca 2+ content (Figures 6G,H), or RyR2 phosphorylation (Figures 6I,J) in sham-operated animals.

Effects of Exercise Training on Beta-Adrenoceptors in HF and Sham-Operated Rats
As an alternative explanation for our observations, we examined if beta-adrenoceptor density was affected by exercise. For this, we used radioligand binding on membranes from left ventricular cardiomyocytes. The beta 1 -adrenoceptor selective ligand CGP20712A displayed a pK i−high of 9.21 ± 0.02 and a pK i−low of 5.7 ± 0.03 (Figure 7A), while the beta 2 -adrenoceptor selective ligand ICI118551 displayed a pK i−high of 9.32 ± 0.04 and a pK i−low of 6.9 ± 0.04 (Figure 7B), similar to previously reported values (26). The total beta-adrenoceptor density did not differ between the four groups (HF-ET, HF-SED, sham-ET, sham-SED, Figure 7C). Importantly, however, the beta 1adrenoceptor density was significantly reduced in the HF-SED compared to sham-SED but increased back to normal levels in HF-ET (Figures 7D,E). The beta 2 -adrenoceptor density, on the other hand, was increased in HF-SED compared to sham-SED and was not affected by ET (Figure 7F). In contrast to the effects of ET on membrane beta 1 -adrenoceptor density, total abundance of the receptor measured by Western blotting was not altered by ET (Supplementary Figure 1).

DISCUSSION
We tested the hypothesis that ET could stabilize RyR2dependent SR Ca 2+ release associated with arrhythmias in post-MI HF rats. We subjected rats with HF to a 5week ET protocol that was initiated 1 week after the induction of MI by ligation of the left coronary artery. This protocol increased the aerobic capacity of HF-ET rats to 127% of that of HF-SED rats but did not prevent a further increase in atrial diameter and had no impact on left ventricular diameter or on contractile function measured as fractional shortening. ET reduced the frequency of spontaneous Ca 2+ release events in left ventricular cardiomyocytes, as indicated by a reduced frequency of Ca 2+ waves and Ca 2+ sparks in HF-ET cells compared with the HF-SED cells. In situations associated with increased adrenergic stress and betaadrenoceptor stimulation, the HF-ET group exhibited lower Ca 2+ transient amplitudes and decay rates than the HF-SED group, but no significant changes in RyR2, SERCA2 or their major regulatory proteins or phosphorylation sites. Shamoperated ET rats also exhibited a 135% increase in aerobic capacity compared to sham-SED and lower Ca 2+ wave frequency when measured across all frequencies. However, we observed no changes in echocardiographic parameters or changes in Ca 2+ measurements or SR Ca 2+ regulatory proteins that can explain this. Since the Ca 2+ regulatory proteins could not explain our main observation, we tested an alternative hypothesis and found that beta 1 -adrenoceptors were downregulated in HF-SED but normalized with ET.

Rat Model of Post-MI HF and RyR2 Dysfunction in Perspective
The pathophysiological changes that occur in post-MI HF are highly complex and progress over time (27). After a large MI, structural and mechanical alterations occur at the organ level, as well as in cardiomyocytes and the extracellular matrix (28). The progression of these alterations leads to different mechanisms being responsible for increased risk of arrhythmias at different time points after an MI (29). Altered Ca 2+ handling due to perturbed expression and function of Ca 2+ handling proteins is involved in arrhythmogenesis at all stages after MI, mainly because these alterations increase the risk of triggered activity (30,31). Dysfunctional RyR2 is key in this arrhythmia mechanism and has been observed in several animal models of established post-MI HF (32)(33)(34). Our model is in line with other models of HF, with reduced aerobic capacity, severely dilated left ventricular diameters, and increased left-atrial diameters as a sign of long-standing increase in end-diastolic pressure, as well as reduced density of beta 1adrenoceptors (35,36). Many models of HF with a similar phenotype show post-translational modifications of RyR2, especially increased phosphorylation (12). Thus, dysfunction and modification of RyR2 have received much attention as potential targets for prevention of arrhythmias in HF (37). Our model is therefore highly relevant to test the hypothesis that ET can stabilize RyR2.

Exercise Training as Antiarrhythmic Therapy in Post-MI HF
There is widespread interest in HF pathophysiology, RyR2 dysfunction, and the beneficial effects of ET in HF. Therefore, it is a surprise that data on the effects of ET on RyR2 function in post-MI HF are scarce. Bonilla et al. subjected dogs with anterior MI to 10 weeks of ET and showed a highly reduced tendency toward ventricular fibrillation compared with sedentary dogs (38). In line with our study, they showed that the frequency of Ca 2+ sparks in ventricular cardiomyocytes was reduced in dogs in the exercise group and that the abundance of RyR2 that was phosphorylated at serine 2814 was also reduced, compared with the sedentary group. However, there was no data that indicated that the dogs had HF. Kemi et al. performed a study of post-MI HF rats that showed many similarities with our study, although they employed a different rat strain (39). They too found a positive effect on Ca 2+ wave frequency, but they did not analyze phosphoproteins. In other publications, the same group has provided compelling evidence that ET can have beneficial effects on Ca 2+ handling and on T-tubule structure in normal rats (5,17,20,40). Based on these and other data of disease models that show dysfunctional RyR2 (17,18), we hypothesized that ET could have beneficial effects on RyR2 function. We focused specifically on serine 2808 and 2814 in RyR2, as there is mounting evidence that phosphorylation of these residues contributes to SR Ca 2+ leak in HF and that prevention of such phosphorylation can prevent arrhythmogenic Ca 2+ release (41,42). It is therefore interesting that we observed beneficial effects on spontaneous SR Ca 2+ release, without significant alterations in the abundance of key SR Ca 2+ handling proteins, or CaMKII-or protein kinase A (PKA)-dependent phosphorylation. Although previous studies have clearly shown that ET can exert beneficial effects on SR Ca 2+ handling through attenuated phosphorylation by CaMKII especially, our results show that other mechanisms may also contribute to a similar result. Interestingly, we observed that Ca 2+ transient amplitude and reuptake of Ca 2+ was lower in HF-ET rats during adrenergic stress. It was therefore somewhat surprising that this coincided with increased beta 1 -adrenoceptor density compared to HF-SED. Our radioligand binding experiments show that the number of beta 1 -adrenoceptors was reduced in HF rats and increased to normal levels by ET (Figure 7). Our experiments do not provide an explanation for how this is compatible with reduced reuptake or why the increased density was not reflected in the total abundance of the receptors when measured by Western blotting. We can only speculate that ET altered the relative pools of these receptors at the cell surface and receptors associated with the SR. Recent studies have shown that internalized and SR-bound beta 1 -adrenoceptors exert an important regulatory role in SR Ca 2+ release (43). These possible effects of ET should be explored further in future studies.

Limitations
We investigated Ca 2+ handling after MI and ET and focused specifically on SR Ca 2+ handling. Our studies lead us toward a role of beta-adrenoceptors in the effects of ET in HF rats, although the focus of the study did not allow us to conclude if this represents an explanation to reduced frequency of spontaneous Ca 2+ release events after ET. Still, there is a risk that we missed a difference in phosphoprotein abundance after ET, due to low sample size (type II error) in our analysis, and such effects should therefore not be excluded. The reader should be aware that a one-sided test of CaMKII-dependent phosphorylation of RyR2 would have resulted in a P < 0.05. Based on previous data, the a priori assumption that ET would decrease (not increase) CaMKII activity could be claimed, thus supporting a one-sided test. Furthermore, our analyses of phosphoproteins were only performed in the absence of sympathetic stimulation. Analysis performed after f.ex. exposure to ISO could potentially have clarified some of the non-significant observations. Another potentially contributing factor that should be included in future studies is altered cell ultrastructure. ET can effect cell size and has been shown to alter t-tubule organization (40). These aspects could be of importance but were not included in our study.
Our study is limited to cellular phenomena and does not include measurements of in vivo arrhythmias. This limits the potential for extrapolation to propensity for arrhythmias in vivo. However, other mechanisms are also important for antiarrhythmic effects of ET, such as alterations in repolarization reserve (38) or reductions in fibrosis levels (44), although the latter is debated after findings in healthy rats (45). As these other mechanisms were not the focus of our study, we would not have been able to ascertain a causal link between the cellular phenomena and arrhythmias regardless of any observed effects on arrhythmias.
Our model of HF is based on induction of large MIs, which leads to clear mechanical and structural alterations in the left ventricle at an early time point after the MI. Any extrapolation to HF that develops after smaller MIs, which involves less remodeling, is unclear. Especially, we did not separate cardiomyocytes according to relative localization or distance from the MI, although the effects of ET could vary between regions (46,47).
We chose to start high-intensity ET at an early time point after MI. High-intensity ET was chosen based on previous studies in rats (6). Moderate-intensity ET might have been more in line with what patients can perform, although high-intensity interval training is feasible even in patients with severe HF (48). Any extrapolation to correspondingly large MIs in humans is speculative, and no conclusion regarding the time at which ET should be initiated in patients can be drawn from studies in rats. The early initiation that was chosen in our study could mean that remodeling processes were still highly active, which could have affected our results. However, it could be argued that this is a reason to stress early initiation of ET in order to affect beneficially the remodeling process. We did not compare effects of exercise in HF and sham rats for all data, as this was not the focus of our study. The effects of ET could be different in diseased and normal animals, and the therapeutic potential should be tested in relevant disease models.

CONCLUSIONS
ET can stabilize RyR2-dependent Ca 2+ release in post-MI HF. Our data indicate that the protective mechanism involves regulation of both SR Ca 2+ release and reuptake, and that effects of ET at several levels of the sympathetic signaling pathways controlling SR Ca 2+ release should be explored in future studies.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The animal study was reviewed and approved by Norwegian National Committee for Animal Welfare.

AUTHOR CONTRIBUTIONS
TD, MS, RM, JA, MF, MH, KA, KH, and IS conceived and designed the analysis, collected data, contributed to data analysis, performed the analysis, and wrote the paper. FL, WL, OS, and MS conceived and designed the analysis, contributed to data analysis, and wrote the paper. All authors contributed to the article and approved the submitted version.