Research ArticleCardiology

Sarcoplasmic reticulum calcium leak contributes to arrhythmia but not to heart failure progression

See allHide authors and affiliations

Science Translational Medicine  12 Sep 2018:
Vol. 10, Issue 458, eaan0724
DOI: 10.1126/scitranslmed.aan0724

Cardiac calcium clarified

The subcellular localization of calcium within cardiomyocytes is tightly regulated during the cardiac cycle, and calcium leak from the sarcoplasmic reticulum is linked to alterations in heart rhythm and heart failure. Mohamed et al. investigated whether inhibiting calcium leak by stabilizing the sarcoplasmic calcium channel ryanodine receptor 2 corrected irregular heartbeat and prevented maladaptive myocardial remodeling and heart failure. In a mouse model of pressure overload, reduced calcium leak did not prevent heart failure; increased calcium leak in a volume overload mouse model did not exacerbate heart failure. Rather, inhibiting calcium leak corrected arrhythmias in myocytes derived from patients with heart failure and tachycardia. Although myocardial remodeling was not altered, arrhythmia was mitigated and survival was increased by ryanodine receptor stabilization in mouse models, suggesting a potential therapeutic application for inhibiting sarcoplasmic calcium leak.


Increased sarcoplasmic reticulum (SR) Ca2+ leak via the cardiac ryanodine receptor (RyR2) has been suggested to play a mechanistic role in the development of heart failure (HF) and cardiac arrhythmia. Mice treated with a selective RyR2 stabilizer, rycal S36, showed normalization of SR Ca2+ leak and improved survival in pressure overload (PO) and myocardial infarction (MI) models. The development of HF, measured by echocardiography and molecular markers, showed no difference in rycal S36– versus placebo-treated mice. Reduction of SR Ca2+ leak in the PO model by the rycal-unrelated RyR2 stabilizer dantrolene did not mitigate HF progression. Development of HF was not aggravated by increased SR Ca2+ leak due to RyR2 mutation (R2474S) in volume overload, an SR Ca2+ leak–independent HF model. Arrhythmia episodes were reduced by rycal S36 treatment in PO and MI mice in vivo and ex vivo in Langendorff-perfused hearts. Isolated cardiomyocytes from murine failing hearts and human ventricular failing and atrial nonfailing myocardium showed reductions in delayed afterdepolarizations, in spontaneous and induced Ca2+ waves, and in triggered activity in rycal S36 versus placebo cells, whereas the Ca2+ transient, SR Ca2+ load, SR Ca2+ adenosine triphosphatase function, and action potential duration were not affected. Rycal S36 treatment of human induced pluripotent stem cells isolated from a patient with catecholaminergic polymorphic ventricular tachycardia could rescue the leaky RyR2 receptor. These results suggest that SR Ca2+ leak does not primarily influence contractile HF progression, whereas rycal S36 treatment markedly reduces ventricular arrhythmias, thereby improving survival in mice.


Heart failure (HF) is a leading cause of mortality and morbidity worldwide. Patients with HF typically die because of either progressive failure of cardiac mechanical function or ventricular arrhythmias (1). Much of the impaired contractile function in HF is caused by reduced Ca2+ transient (CaT) in cardiomyocytes (CMs) that is mainly dependent on the sarcoplasmic reticulum (SR) Ca2+ content, reflecting the balance between Ca2+ uptake via SR Ca2+ adenosine triphosphatase (SERCA2a) and Ca2+ efflux via ryanodine receptor 2 (RyR2). Thus, reduced SR Ca2+ content in HF could be principally due to reduced SERCA activity or leaky RyR (2). Increasing SERCA2a expression improves Ca2+ handling and contractility and averts arrhythmias in the failing heart (3, 4). In contrast, despite agreement that SR Ca2+ leak is induced in HF (58), controversy still exists about its role in HF (816).

Not all HF models show increased SR Ca2+ leak. Previously, we reported that Ca2+ cycling in the volume overload (VO) model is not disturbed, and myocardial remodeling is more favorable, at least in its acute phase, in comparison to the pressure overload (PO) model (7, 17, 18). The pathological changes in PO are partially and early mediated by Ca2+/calmodulin-dependent kinase IIδ (CaMKIIδ) activation, leading to increased SR Ca2+ leak. In contrast, VO shows no signs of early CaMKIIδ activation and consequently no increased SR Ca2+ leak (7).

JTV-519, a 1,4-benzothiazepine, increases FKBP12.6 binding to RyR2 and reduces SR Ca2+ leak leading to inhibition of cardiac arrhythmia, sudden cardiac death (SCD), and HF development (9, 19, 20), but JTV-519 is a multichannel drug with several off-target effects (21). Here, we investigated whether long-term treatment with the RyR2-stabilizing drug rycal S36, a highly selective JTV-519 derivative (22, 23), could prevent maladaptive cardiac remodeling in wild-type (WT) mice subjected to transaortic constriction (TAC) and myocardial infarction (MI). We also tested the effects of rycal S36 in human CMs derived from failing ventricular and nonfailing atrial tissues and in human induced pluripotent stem cells (hiPSCs) isolated from a patient with catecholaminergic polymorphic ventricular tachycardia (CPVT). In a knock-in (KI) mouse model with a human CPVT-associated RyR2 mutation (R2474S) [RyR2-R2474S+/− (Ryr2RS/WT)] (22), we explored whether induced SR Ca2+ leak would exaggerate the myocardial dysfunction after aortocaval shunt–induced VO.


Diastolic SR Ca2+ leak is decreased in rycal S36–treated mouse CMs

Rycal S36 markedly reduced Ca2+ spark frequency (CaSpF) and calculated SR Ca2+ leak (Fig. 1, A to C). To assess the spark-dependent and spark-independent Ca2+ leak (24), total SR Ca2+ leak was measured using the tetracaine protocol (Fig. 1D) (25). Total SR Ca2+ leak was reduced in rycal S36–treated cells, even when normalized to SR Ca2+ load (Fig. 1, E and F). However, SR Ca2+ content, CaT, and sarcomere shortening were not altered by rycal S36 treatment (Fig. 1G and fig. S1). The rate of the decay of CaT (Ksys) depends on both SERCA and sarcolemmal Ca2+ transporters, whereas the rate of the decay of the Ca2+ release by caffeine (Kcaff) depends only on sarcolemmal transporters. Therefore, KSERCA = KsysKcaff can be used to calculate the SERCA activity (26), which was not different between placebo- and rycal S36–treated cells (fig. S1).

Fig. 1 Rycal S36 reduces SR Ca2+ leak in murine CMs.

(A) Line-scan images of Ca2+ spark in the presence of isoproterenol (ISO; 10 nM) with or without rycal S36 (1 μM). (B and C) Mean CaSpF (B) and relative SR Ca2+ leak normalized to placebo (C). (D) Representative traces of intracellular Ca2+ recordings showing measurement of SR Ca2+ leak in placebo- and rycal S36–treated CMs. (E to G) Mean total SR Ca2+ leak (E), SR Ca2+ leak–to–SR Ca2+ load ratio (F), and SR Ca2+ load (G). Data are means ± SEM, n = 5 mice per group in (A) to (C); 7 mice per group in (D) to (G). *P < 0.05 versus placebo, two-tailed unpaired Student’s t test. Numbers within columns indicate CMs/mice. AU, arbitary units.

Rycal S36 prevents SR Ca2+ leak in ventricular CMs after TAC

We and others have previously reported that SR Ca2+ leak is markedly increased in TAC-induced PO model (68). We investigated whether inhibition of SR Ca2+ leak by rycal S36 could avert PO-induced pathological remodeling. Infusion of rycal S36 at 5 mg/kg per hour resulted in drug serum concentrations of 1.48 ± 0.27 and 1.04 ± 0.11 μM at 3 and 9 weeks after operation, respectively. The pressure gradients at 2 days after TAC were comparable in placebo- and rycal S36–treated mice (fig. S2). CaSpF, spark-dependent Ca2+ leak, and spontaneous Ca2+ waves (SCaWs) were decreased in TAC–rycal S36 mice (fig. S3). Total SR Ca2+ leak and SR Ca2+ leak per SR Ca2+ load ratio were markedly reduced in TAC–rycal S36 animals (Fig. 2, A and B, and fig. S4). However, SR Ca2+ load was comparable in TAC-placebo and TAC–rycal S36 mice (Fig. 2C).

Fig. 2 Improved survival in TAC–rycal S36 mice.

(A to C) Mean total SR Ca2+ leak (A), SR Ca2+ leak per SR Ca2+ load ratio (B), and SR Ca2+ load (C) at 3 weeks after TAC measured by the tetracaine protocol in isolated murine CMs. (D) Kaplan-Meier survival curves, TAC-placebo versus TAC–rycal S36; §P < 0.05, log-rank test. (E) Echocardiographic M-mode representative images at 9 weeks after TAC. (F to I) Mean radial diastolic peak velocity (F), septum thickness (G), LV end-diastolic diameter (LVEDD) (H), and ejection fraction (EF) (I). Data are means ± SEM, n = 5 mice per group in (A) to (C); 5 mice per sham and 8 to 15 mice per TAC in (F) to (I). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 versus corresponding sham, §§§P < 0.001 versus corresponding placebo, one-way analysis of variance (ANOVA) with Bonferroni posttest. Numbers within parentheses or columns indicate CMs/mice or mice.

Rycal S36 improves survival but does not alter contractile deficit or histological and molecular signature of HF after TAC

The overall Kaplan-Meier survival was markedly improved in TAC–rycal S36 compared to TAC-placebo mice (Fig. 2D). Thus, we sought to address whether enhanced survival is due to an attenuated myocardial remodeling or decreased lethal arrhythmias. Serial echocardiography indicated no differences in the left ventricular (LV) geometry or systolic and diastolic functions between TAC-placebo and TAC–rycal S36 hearts (Fig. 2, E to I, and table S1). Nearly similar outcomes were observed in mice treated with a threefold higher dose of rycal S36 (serum concentrations of 4.19 ± 0.65 μM). The total SR Ca2+ leak and the SR Ca2+ leak–to–SR Ca2+ load ratio were markedly decreased in TAC–rycal S36 hearts (fig. S5, A and B), but the myocardial remodeling was similar in rycal S36– and placebo-treated TAC mice (fig. S5, C and D, and table S2). TAC-operated WT mice treated with dantrolene, a rycal-unrelated RyR2 stabilizer (27, 28), exhibited a nearly comparable myocardial dysfunction at 7 weeks after TAC as placebo-treated mice (table S3).

TAC-placebo and TAC–rycal S36 mice showed comparable pathological remodeling at 9 weeks after TAC. Both groups showed similar induced fibrotic response, increased myocyte cell sectional area (CSA), increased heart weight–to–tibia length (HW/TL) and lung weight–to–tibia length (LungW/TL) ratios, up-regulation of fetal cardiac genes natriuretic peptide type A (Nppa) and natriuretic peptide type B (Nppb), and down-regulation of Serca2a transcript (Fig. 3 and fig. S6).

Fig. 3 Similar pathological remodeling in placebo- and rycal S36–treated animals at 3 and 9 weeks after TAC.

(A) Cardiac cross sections of the hematoxylin and eosin (H&E; top), Masson’s Trichome (MT; middle), and wheat germ agglutinin (WGA; bottom). (B and C) Quantification of cardiac fibrosis (B) and myocyte CSA (C). (D and E) HW/TL (D) and LungW/TL ratios (E). (F) Quantitative reverse transcription polymerase chain reaction (PCR) expression analyses of Nppa, Nppb, Serca2a, and Rcan1.4 at 9 weeks after TAC. Data are means ± SEM, n = 3 to 6 mice per group in (A) to (C); 5 to 8 mice per group in (D) and (E); 3 to 5 mice per group in (F). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 versus corresponding sham, one-way ANOVA with Bonferroni posttest. Numbers within columns indicate CM/mice or mice.

TAC induces deterioration of global Ca2+ homeostasis and alters Ca2+ handling proteins’ expression independent of rycal S36 treatment

Three weeks after TAC, both TAC-placebo and TAC–rycal S36 CMs displayed minimal prolongation of CaT decline, a trend toward decreased CaT amplitude, and comparable sarcomere shortening (fig. S7). SERCA2a activity was not altered between placebo and rycal S36 mice either in sham or TAC groups at 3 weeks after surgery (fig. S7).

At 9 weeks after TAC, both rycal S36– and placebo-treated hearts exhibited a comparable reduction in the protein expression of SERCA2a, and equal increases in the protein expressions of Sorcin and sodium/calcium exchanger 1 (fig. S8) and in the mRNA expression of regulator of calcineurin 1.4 (Rcan1.4; Fig. 3F). A similar increase in the phosphorylation of CaMKIIδ, RyR2 at Ser2814, and a trend toward decreased phospholamban phosphorylation were observed in both TAC groups (fig. S8).

Rycal S36 reduces arrhythmogenicity in TAC-operated mice

A related rycal, S107, efficiently decreased burst pacing–induced atrial arrhythmias (29). We sought to determine whether rycal 36 would also abate the occurrence of TAC-triggered ventricular activities. Although rycal S36 did not exert an effect on the membrane action potential duration (APD), it markedly suppressed delayed afterdepolarizations (DADs) (Fig. 4, A and B; fig. S9A; and table S4).

Fig. 4 Rycal S36 mitigates TAC-induced arrhythmias.

(A) Mean APD (90%). (B) Average DAD incidence. Analyzed CMs/mice were 18/3 in TAC-placebo and 22/3 in TAC–rycal S36. (C) Telemetry ECGs recording at 9 weeks after TAC. PVCs (arrowheads) and polymorphic VT (arrow) are shown. (D to F) Incidence of PVCs (D), VT (E), and mean arrhythmia score (F) after TAC. (G) Representative traces of monophasic APs. (H) Ventricular arrhythmia inducibility. Data are means ± SEM, n = 2 to 3 mice per group in (A) and (B); 3 to 8 mice per group in (C) to (H). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 versus corresponding sham, §P < 0.05, §§P < 0.01, §§§P < 0.001 versus corresponding placebo, two-tailed unpaired Student’s t test (B), and one-way ANOVA with Bonferroni posttest (A and D to H). Numbers within columns indicate CMs/mice or mice.

TAC-placebo and TAC–rycal S36 mice exhibited no marked differences in electrocardiographic (ECG) parameters (table S5). At 9 weeks after TAC, the number of premature ventricular contractions (PVCs) and episodes of ventricular tachycardia (VT), which represent life-threatening events, were markedly repressed in TAC–rycal S36 hearts (Fig. 4, C to E). Consequently, arrhythmia score increased markedly in placebo animals at 9 weeks after TAC, which was blunted in rycal S36 mice (Fig. 4F). Moreover, epicardial electrical stimulation in Langendorff-perfused hearts, isolated at 9 weeks after TAC, induced ventricular arrhythmias in TAC-placebo mice, which were suppressed by rycal S36 treatment (Fig. 4, G and H).

Rycal S36 treatment reduces ventricular arrhythmias and mortality after MI

Like TAC, MI is associated with markedly increased SR Ca2+ leak (30). However, it involves replacement of dead cells with scar that may elicit radically different cellular adaptation than TAC. Therefore, we tested the effect of rycal S36 after induction of MI. Survival was markedly improved in MI–rycal S36 compared to MI-placebo mice (Fig. 5A). Serial echocardiography revealed equal deterioration of LV geometry and function in both MI groups (Fig. 5, B to D, and table S6). At 6 weeks after operation, the MI-placebo and MI–rycal S36 mice exhibited equal increases in myocyte CSA and HW/TL and LungW/TL ratios, as well as up-regulation of hypertrophic genes Nppa and Nppb (Fig. 6, A to D). Treatment with rycal S36 reduced the PVC incidence and overall arrhythmia score in MI mice (Fig. 6, E and F). Shan et al. (31) identified ventricular fibrillation as the main cause of death in RyR2-S2808D transgenic mice within the first 3 days after MI that was efficiently reduced by rycal S107 resulting in improved survival, which could be applied to our study to explain the early mortality pattern seen in MI mice.

Fig. 5 Reduced mortality in MI–rycal S36 mice.

(A) Kaplan-Meier survival curves, MI-placebo versus MI–rycal S36; §P < 0.05, log-rank test. (B) Representative echocardiographic M-mode images at 6 weeks after MI. (C) Average values of LVEDD and (D) EF. Data are means ± SEM, n = 5 mice per sham and 13 to 17 mice per MI in (B) to (D). **P < 0.01, ***P < 0.001 versus corresponding sham, one-way ANOVA with Bonferroni posttest. Numbers within parentheses or columns indicate mice.

Fig. 6 Rycal S36 abates MI-induced arrhythmias but has no beneficial effect on myocardial remodeling.

(A) Cardiac cross sections of H&E (top) and WGA (bottom). (B to D) Mean myocyte CSA (B), HW/TL and LungW/TL ratios (C), and Nppa and Nppb mRNA expressions (D) at 6 weeks after MI. (E and F) Average PVCs and arrhythmia score. Data are means ± SEM, n = 3 to 4 mice per group in (A) and (B); 7 mice per sham and 14 to 18 mice per MI in (C); 4 to 5 mice per sham and 7 to 8 mice per MI in (D); 4 to 5 mice per group in (E) and (F). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 versus corresponding sham, §P < 0.05, §§§§P < 0.0001 versus corresponding placebo, one-way ANOVA with Bonferroni posttest. Numbers within columns indicate cells/mice.

Rycal S36 has antiarrhythmic effects in human CMs

Rycal S36 treatment reduced CaSpF, kinetics, and amplitudes in human failing ventricular CMs (Fig. 7, A and B; for patient characteristics, see table S7). We also studied the effects of rycal S36 on cellular arrhythmogenic triggers by simultaneous recording of CaT and action potential (AP; Fig. 7C). Triggered DADs were suppressed by addition of rycal S36 (Fig. 7D). Rycal S36 did not affect APD, resting membrane potential (RMP), CaT, or decay time constant of CaT, indicating unaltered SERCA activity (Fig. 7, E to H).

Fig. 7 Rycal S36 averts proarrhythmic phenotype in human failing ventricular CMs.

(A) Original line-scan images of Ca2+ spark, and (B) mean CaSpF, amplitude, width [full width at half maximum (FWHM)], and duration [full duration at half maximum (FDHM)]. (C) Representative traces of intracellular CaT and AP simultaneously recorded in ventricular CMs. (D to H) Mean DADs/min (D), APD (E), RMP (F), CaT amplitude (G), and fluorescence decay time (τ) (H). Data are means ± SEM, n = 4 patients per group in (A) and (B); 4 CMs per group in (C) to (H). *P < 0.05 versus placebo, two-tailed unpaired Student’s t test. Numbers within columns indicate cells/patients.

The propensity for SR Ca2+ sparks and SCaWs generation was also assessed in atrial CMs isolated from patients in sinus rhythm (table S7). Preincubation with rycal S36 reduced CaSpF in atrial CMs treated with isoproterenol (Fig. 8, A and B). DADs could not be measured in these “healthy” atrial CMs even under isoproterenol. Elevation of extracellular Ca2+ to 5 mM (to overload the SR) induced SCaWs, but not AP, in atrial CMs, which could be successfully terminated by adding rycal S36 to the perfusion (three of three cells; Fig. 8C). Pretreatment with rycal S36 resulted in no induction of SCaWs in atrial CMs (zero of three cells; Fig. 8D). Rycal S36 treatment of atrial CMs did not alter APD or RMP (Fig. 8, E to G).

Fig. 8 Rycal S36 aborts proarrhythmic events in human nonfailing atrial CMs.

(A) Original line-scan images of Ca2+ spark and (B) mean CaSpF. (C and D) Representative traces of recorded intracellular Ca2+ and membrane voltage in atrial CMs. (E to G) Representative traces showing AP (E), mean APD (F), and RMP (G). Data are means ± SEM, n = 4 to 5 patients per group in (A); 6 to 9 CMs per group in (F) and (G). *P < 0.05, one-way ANOVA with Bonferroni posttest (A), two-tailed unpaired Student’s t test (G). Numbers within columns indicate CMs/patients.

Rycal S107 treatment of CPVT-hiPSC-CMs, harboring a missense RyR2 mutation (c.13759 A>G, p. I4587V), markedly reduced DAD incidence (32). We asked whether a different RyR2 mutation (c.12226 A>G, p. E4076K) would also profit from RyR2 stabilization. Rycal S36 treatment efficiently reduced the SR Ca2+ spark characteristics and accordingly the calculated SR Ca2+ leak in both control-derived and CPVT-hiPSC–derived CMs (fig. S10).

Genetic induction of SR Ca2+ leak in Ryr2RS/WT mice does not alter VO-induced HF

We found no difference in the CaT and sarcomere shortening between WT and Ryr2RS/WT CMs in sham or shunt groups at 3 weeks after surgery (fig. S11), indicating that the Ryr2RS/WT mutation does not alter intracellular Ca2+ cycling. In agreement with a previous report (29), we observed mild increased SR Ca2+ leak and SR Ca2+ leak–to–SR Ca2+ load ratio in sham Ryr2RS/WT versus sham WT mice. At 3 weeks after shunt, SR Ca2+ leak–to–SR Ca2+ load ratio was markedly induced in Ryr2RS/WT versus WT mice (fig. S12). Survival rate after shunt showed no difference between WT and Ryr2RS/WT mice (fig. S13A). Both genotypes of mice exhibited a comparable cardiac dilatation and contractile failure after shunt (fig. S13, B to D, and table S8). At 9 weeks after surgery, both WT and Ryr2RS/WT animals experienced similar increases in myocyte CSA and HW/TL and LungW/TL ratios, a comparable up-regulation of fetal cardiac genes, and activation of calcineurin–NFAT (nuclear factor of activated T cells) and CaMKIIδ signaling pathways (figs. S14 and S15).


The present study shows that (i) inhibition of SR Ca2+ leak by rycal S36 does not attenuate contractile dysfunction in mice after TAC and after MI; (ii) rycal S36 treatment has pronounced antiarrhythmic effects that are associated with improved survival after TAC and after MI, possibly due to blunted SCaWs and DAD-induced triggered activity; (iii) induction of increased SR Ca2+ leak in Ryr2RS/WT mice does not augment the cardiac functional deterioration after shunt; and (iv) rycal S36 quiets arrhythmogenic events in human failing, nonfailing and CPVT-hiPSC-CMs, indicating that rycal-mediated RyR2 stabilization might have therapeutic benefit with respect to SR Ca2+ leak–associated pathologies such as cardiac arrhythmogenesis in HF and CPVT patients.

Abnormal SR Ca2+ release induces SCaWs, which results in DADs and triggered activity, all of which are arrhythmogenic (33) and contribute to SCD in HF patients (1). We have previously reported that the arrhythmogenic potential of increased late Na+ current is SR Ca2+ leak–dependent (34). The convergence of several proarrhythmogenic pathways on the RyR2 makes it an interesting target for treatment of arrhythmias. Here, we showed that rycal S36–mediated inhibition of SR Ca2+ leak efficiently reduced the arrhythmia susceptibility in TAC- and MI-operated mice, whereas neither the ECG parameters nor APD was altered. Our findings are in agreement with previous studies reporting that stabilization of RyR2 with either JTV-519 or flecainide protects from lethal arrhythmias (19, 20, 35). However, both antiarrhythmic drugs act as nonselective pore blockers of transmembrane ion channels including RyR2 (21, 3537). In contrast, rycal S36 acts principally similar yet not molecularly identical to dantrolene as a specific allosteric stabilizer of the closed RyR2 channel state through binding to the cytosolic RyR domains (22, 23, 38) and has no documented off-target activity at up to 10 μM (22, 23).

Electrical and structural cardiac remodeling, although closely related, could be independently regulated in the context of HF (39). Although myocardial remodeling was comparable, the antiarrhythmic effect of rycal S36 might reflect a decrease in triggered activity rather than an effect on reentry or altered conduction. Our data do not limit the role of other proarrhythmogenic mechanisms but show that inhibition of SR Ca2+ leak either averts the induction of arrhythmias (especially PVCs) or, more likely, reduces the proarrhythmogenic milieu below a relevant threshold.

Previously, we illustrated that SR Ca2+ leak increases markedly during HF progression in humans (5). In line with clinical data, moderate HF (3 weeks of PO) resulted in mild arrhythmias (mostly single PVCs), whereas severe HF (9 weeks of PO) induced severe cardiac arrhythmias, possibly owing to massive SR Ca2+ leak in late HF stage, suggesting a direct proportional relationship between the amount of SR Ca2+ leak and arrhythmia severity.

We showed that the higher incidence of life-threatening arrhythmias and the mortality pattern are model-specific. Increased arrhythmia episodes and enhanced mortality were observed over time in the TAC model, whereas higher mortality in earlier phase was exhibited in the MI model, as is encountered in human patients, likely because of acute HF, cardiac rupture, and lethal arrhythmias. It is not surprising that the improved survival mediated by the antiarrhythmogenic effect of rycal S36 was more pronounced at the time phase corresponding to the high incidence of fatal arrhythmia, which occurs at late phase in the TAC model but at relatively early phase in the MI model.

Rycal S36 markedly reduces the Ca2+ spark width in human but not in mouse CMs, which could be due to species-specific difference in the RyR2 clustering and functions. Overall, rycal S36 treatment led to comparable reductions in SR Ca2+ leak and DADs in human and mouse CMs. Therefore, comparable antiarrhythmic effects could be expected; however, further research is needed to characterize the distinct effects of rycal in human versus mouse CMs.

Marx et al. (40) first proposed that FKBP12.6 dissociation from RyR2 enhances SR Ca2+ leak and systolic dysfunction in HF. Some reports support the negative effect of leaky RyR2 on cardiac function (811). However, several studies challenge this hypothesis (1216). Some of these supporting studies did not use an inhibitor of SR Ca2+ leak but instead used transgenic mice with a basal leaky RyR2 (810) that, together with the HF trigger (PO or MI), might induce massive SR Ca2+ leak that would activate different Ca2+-dependent signal pathways. Reduction of SR Ca2+ leak upon CaMKII inhibition improves contractility (41). However, CaMKII inhibition has effects on Na+, Ca2+ channels, and SERCA activity (42). Two groups reported increased SR Ca2+ content by reduction of SR Ca2+ leak (43, 44). In both reports, however, nonphysiological methods were used to increase SR Ca2+ leak (either caffeine or 5 mM extracellular Ca2+ concentration), and nonspecific drugs (either tetracaine or JTV519) with several off-target effects were used to avert SR Ca2+ leak. On the other hand, several reports detected no change in SR Ca2+ content, despite altered SR Ca2+ leak (45, 46). Neither dantrolene (28, 47) nor JTV519 (37) showed an influence on contractility at physiological extracellular Ca2+ concentrations. Only at supraphysiological extracellular Ca2+ concentrations of 5 mM did JTV519 show improved contractility and increased post-rest potentiation, indicating a higher SR Ca2+ content (37, 44). However, high extracellular Ca2+ would overload the SR and therefore could modify the intrinsic RyR2 gating.

Here, we used nontransgenic WT mice to closely simulate the clinical situation, and we showed that rycal-dependent and rycal-independent RyR2 stabilization, and in turn SR Ca2+ leak reduction, does not attenuate the myocardial dysfunction in experimental HF. Autoregulation of SR Ca2+ content (48) via decreased SERCA2a-mediated Ca2+ reuptake could explain the unchanged SR Ca2+ load and hence the contractility, even with a decreased SR Ca2+ leak. To our knowledge, there are only three studies showing a protective effect of RyR2 stabilizers on HF development in nontransgenic animals. Yano and colleagues (11) reported a protective effect of JTV-519–mediated RyR2 stabilization on HF using a dog model of rapid pacing. However, rapid pacing, which leads to a reversible HF phenotype, may bring about radically different cellular adaptation from that elicited by TAC and MI. Second, as discussed before, JTV-519 has several non-RyR2–associated effects (21, 37), which theoretically could mediate its protective effect. Wehrens et al. (9) showed that JTV-519 improved contractility after MI, but other parameters of cardiac remodeling like the LV end-diastolic and end-systolic diameters were not altered, indicating that the remodeling, at least regarding LV dimensions, was comparable. Decreased arrhythmias in the JTV-519–treated mice could explain the improved EF, but Wehrens et al. (9) did not analyze the rate of arrhythmias in their study. Moreover, our data do not support that a higher rate of arrhythmias impairs cardiac function. Shan et al. (31) reported that rycal S107 treatment of the leaky RyR2-S2808D mouse model resulted in improved cardiac function after MI. We can only speculate on the differences between the models. For example, rycal S107, although more specific than JTV-519, still has off-target effects in some G protein–coupled receptor and kinase activities, as reported by the same group (49). This is not the case for rycal S36 (22, 23). Differences in the infarct size could have an impact on HF progression. We and Shan et al. (31) did not measure infarct size, but because the degree of reduction in EF and the increase in LVEDD are similar, we assume that the infarct size should be comparable. It is noteworthy that Shan and colleagues (31) showed a progression of cardiac deterioration after MI that did not occur in our model. Cardiac response to hemodynamic stress is distinct among the different C57BL/6 substrains (50). Although we used C57BL/6J, Shan et al. (31) only described C57BL/6—it is not clear whether substrain J or N was used. Shan et al. (31) started the treatment 7 to 10 days after MI, whereas we started treatment immediately after MI. Although unlikely, we cannot rule out that the start of treatment has an influence on the result. Altogether, we believe that our results are robust because we could show no influence on HF progression in two different HF models (PO and MI) using two independent drugs (rycal S36 and dantrolene).

Although SR Ca2+ leak was massively induced in Ryr2RS/WT mice after shunt, both WT and Ryr2RS/WT mice experienced equal functional and structural deterioration and exhibited similar patterns of mortality. It was reported that leaky RyR2 in CPVT mouse models accelerates the development of pathological myocardial remodeling (8, 10). However, these studies used the PO model in Ryr2RQ/WT and Ryr2RC/WT KI mice, whereas we used the VO model in Ryr2RS/WT mice. In contrast to the Ryr2RS/WT mice that experience minimal SR Ca2+ leak and nearly normal Ca2+ homeostasis at baseline conditions, both Ryr2RQ/WT and Ryr2RC/WT KI mouse models show disturbed Ca2+ cycling due to massive SR Ca2+ leak that was further induced after TAC (8, 10), leading to marked loss of intracellular Ca2+ that could reduce the SR Ca2+ content and thereby influence systolic contractility. Ryr2RQ/WT mice exhibited sudden infantile death due to cardiac arrhythmias at baseline condition (51). We could study the effect of induction of nearly “normal” SR Ca2+ leak on the development of HF in the VO model that represents, at least in its early stage, a Ca2+ cycling–independent model. Because of comparable mortality after shunt, further analyses of arrhythmias were not performed. It is possible that the lack of cardiac fibrosis in shunt model prevents persistent arrhythmias despite an increased trigger by an elevated SR Ca2+ leak.

Because we could not detect an effect of SR Ca2+ leak on HF progression, we looked at other Ca2+-sensitive signaling pathways to investigate whether an autoregulation within the CMs is influenced by SR Ca2+ leak inhibition or induction (48). If SR Ca2+ leak is a detrimental component in the cardiac contractility, then compensation responsible for masking the effect of manipulating SR Ca2+ leak on contractile function of the heart would be expected. We were unable to detect any differential changes in Ca2+ regulatory proteins upon leak inhibition by rycal S36 or leak induction in shunt-Ryr2RS/WT mice, suggesting that SR Ca2+ leak could be a phenotypic consequence of the myocardial contractile dysfunction rather than a causative factor. The fact that CPVT patients and FKBP12.6 knockout mice do not exhibit HF but arrhythmias (52), combined with the results of the present study, suggests that the leaky RyR2 channel, by itself, can cause lethal cardiac arrhythmias, but not myocardial dysfunction.

Aside from SR Ca2+ leak, other mechanisms—such as deteriorated Na+ current, fibrosis-triggered reentry, and APD prolongation—cannot be ruled out as detrimental factors for severe arrhythmia susceptibility seen in our PO and MI models in late HF stage. It is possible that the effects of RyR stabilization on the excitation-contraction coupling is different in mice compared to that in human hearts that operate at a lower rate where SR Ca2+ leak may play a relevant role in the SR control and thus may still benefit from targeting SR Ca2+ leak. Negative results might have arisen because of ineffective RyR2 stabilization. However, we know that this was not the case because of the following: SR Ca2+ leak was efficiently attenuated in TAC–rycal S36, nearly to the extent of sham-placebo, indirectly indicating a successful stabilization of RyR2 channels by rycal S36; and treatment with threefold higher dose of rycal S36 led to neither more SR Ca2+ leak reduction nor beneficial effects on myocardial contractility. False-positive results in arrhythmia attenuation could have arisen because of nonspecific off-target effects of rycal S36; however, this can be ruled out because rycal S36 was reported to be a specific RyR2 stabilizer (22, 23). Because of the sporadic nature of SCD, we were not able to correlate mortality in the TAC and MI mice with ventricular arrhythmias. Although we used Fluo-4 AM, which, in contrast to Fura-2 and Indo-1, is a lower-affinity Ca2+ probe (making it suitable for detecting a wide range of intracellular Ca2+ concentrations with greater accuracy and lower saturation), we cannot rule out that changes in SR Ca2+ content might be too small to be detected with Fluo-4 AM. JTV519 improved contractility at 5 mM extracellular Ca2+ (37, 44); however, we did not measure SR Ca2+ content or contractility at this high extracellular Ca2+. Therefore, elucidation of the exact mechanism of rycal S36, especially in comparison to flecainide and other rycals, requires further analyses. Although we could show that rycal S36 efficiently blunted arrhythmogenic events in human CMs, potential off-target effects in humans cannot be excluded. Moreover, the small number of patients analyzed because of ethical concerns is a potential limitation. Rycal S36 is currently being evaluated in patients with chronic HF at risk for ventricular arrhythmia, and this merits further investigations.

Here, we showed that fixing RyR2 Ca2+ leak averts life-threatening arrhythmias but is dispensable for myocardial pump dysfunction. Stabilization of RyR2 by newly generated rycals could therefore present a promising target for treatment of fatal arrhythmias in HF and CPVT patients.


Study design

This study was designed to explore the role of SR Ca2+ leak in HF progression using multiple molecular and electrophysiological techniques in mice and human cells. First, we investigated the effect of the rycal S36–mediated SR Ca2+ leak inhibition on HF progression in nonischemic and ischemic mouse models of HF. Second, we tested whether human CMs with increased SR Ca2+ leak would also benefit from rycal S36 treatment. Finally, we addressed whether enhanced SR Ca2+ leak would aggravate pathological myocardial remodeling after shunt-induced VO in the CPVT mouse model. For in vivo experiments, mice were randomly assigned to treatment groups. The following experiments were carried out and analyzed in a blinded fashion: echocardiography, morphometry, histology, Western blotting, real-time PCR, ambulatory telemetry, and electrophysiological studies in shunt-operated mice. Numbers of tested cells and mice are outlined in each figure. Outliers were removed only if the animal was not healthy or if the treatment of the animal was not successful. Animals with welfare problems were euthanized and removed from the experiment. This investigation conforms to the Guide for the Care and Use of Laboratory Animals (National Institutes of Health publication no. 85-23, revised 1996) and was performed in accordance with the ethical standards laid down in the Declaration of Helsinki 1964. The following mice strains were used in this study: heterozygous RyR2-R2474S+/− (Ryr2RS/WT) KI mice on C57BL/6N genetic background (22) and C57BL/6J WT mice (Charles River Laboratories).

Statistical analysis

Statistical analyses were carried out using Prism software version 5.01 (GraphPad Software Inc). Two-tailed unpaired Student’s t test, one-way ANOVA with Bonferroni posttest correction, and two-way ANOVA were used as appropriate, and n = number of mice. Data are reported as means ± SEM. Kaplan-Meier survival analysis was performed, and a log-rank (Mantel-Cox) test was used to determine significance. Primary data are reported in table S9.


Materials and Methods

Fig. S1. Global intracellular CaT and sarcomere shortening in placebo- and rycal S36–treated CMs.

Fig. S2. Trans-stenotic systolic pressure gradient measurement after TAC.

Fig. S3. SR Ca2+ release events in ventricular CMs at 3 weeks after TAC.

Fig. S4. Representative traces of intracellular Ca2+ recordings illustrating measurement of total SR Ca2+ leak in CMs isolated from placebo- and rycal S36–treated mice at 3 weeks after TAC.

Fig. S5. Effect of threefold higher dose of rycal S36 (≈4 μM) on SR Ca2+ leak and myocardial remodeling at 3 weeks after TAC.

Fig. S6. Comparable perivascular fibrosis in placebo- and rycal S36–TAC mice.

Fig. S7. Unaltered intracellular Ca2+ homeostasis in placebo- and rycal S36–treated mice at 3 weeks after TAC.

Fig. S8. Western blot analyses of Ca2+ regulatory proteins at 9 weeks after TAC.

Fig. S9. Patch-clamp and telemetric analyses of the TAC-operated mice.

Fig. S10. Rycal S36 rescues SR Ca2+ leak in CPVT-E4076K CMs.

Fig. S11. Electrophysiological analyses of the WT and Ryr2RS/WT mice at 3 weeks after shunt.

Fig. S12. Increased SR Ca2+ leak in Ryr2RS/WT mice at 3 weeks after shunt.

Fig. S13. Increased SR Ca2+ leak but comparable survival and functional deficit in Ryr2RS/WT versus WT mice after shunt.

Fig. S14. Pathological myocardial remodeling is not different in WT and Ryr2RS/WT animals at 9 weeks after shunt.

Fig. S15. Western blot analyses of Ca2+ regulatory proteins in WT/WT and Ryr2RS/WT (RS/WT) mice at 9 weeks after shunt.

Table S1. Echocardiographic parameters of placebo- and rycal S36–treated mice at 3 and 9 weeks after TAC.

Table S2. Echocardiographic parameters at 3 weeks after surgery of mice treated with a threefold higher dose of rycal S36.

Table S3. Echocardiographic parameters of placebo- and dantrolene-treated mice at 7 weeks after TAC.

Table S4. Raw APD time values at 9 weeks after surgery.

Table S5. ECG parameters at 9 weeks after surgery.

Table S6. Echocardiographic parameters of placebo- and rycal S36–treated mice at 3 and 6 weeks after MI.

Table S7. Clinical data of explanted human hearts.

Table S8. Echocardiographic parameters of WT and Ryr2RS/WT mice at 20 weeks after shunt.

Table S9. Primary data (Excel file).

References (5357)


Acknowledgments: The RyR2-R2474S KI mouse strain was made available by A. R. Marks, Columbia University under a material transfer agreement with the Georg August University of Göttingen, Germany (STV file number 34906). We thank L. J. Field (Indiana University School of Medicine, USA) for reading and commenting on the manuscript. We acknowledge the excellent technical assistance of S. Gaburro, G. Nyamsuren, J. Jakubiczka-Smorag, K. Mattern, S. Koszewa, A. Kretzschmar, T. Schulte, and the service project of SFB1002 (B. Knocke, S. Zafar, R. Blume, and M. Zoremba). Funding: This work was supported by Deutsche Forschungsgemeinschaft (DFG: SFB 1002 project D04 to K.T.; C03 to S.L.; A09 to S.E.L.; A13 to N.V.; A11 to S.S.; D01 to G.H.) start-up funding from University Medical Center Göttingen (to B.A.M.), Grant ERA Net E-Rare-3 2015 Call (GU 595/3-1 to K.G.), DZHK B 15-014 Extern and funding from the Regensburg University ReForM program (to S.N.), and EKFS 2016_A20, IRTG1816 RP12, and VO 1568/3-1 (to N.V.). G.H. and S.E.L. are investigators of the DZHK. Author contributions: B.A.M. designed and performed most of the experiments, analyzed the data, and wrote the manuscript. N.H., P.T., K.S., W.L., C.R., and M.D. carried out some experiments. L.K. provided the rycal S36. J.D.S. provided the human CMs. S.N., E.M.Z., K.G., S.E.L., S.L., N.V., T.S., and S.S. helped in data analyses. G.H. designed the experiments, provided conceptual advice, and edited the manuscript. K.T. designed the experiments, analyzed the data, and wrote the manuscript. Competing interests: S.E.L. is an inventor on “patent US 20070089572A1” submitted by Columbia University that covers “Novel agents for preventing and treating disorders involving modulation of RYR receptors.” L.K. is a full-time employee of Endotherm GmbH. All other authors declare that they have no competing interests. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials.
View Abstract

Navigate This Article