Inhibiting Mitochondrial Na⁺/Ca²⁺ Exchange Prevents Sudden Death in a Guinea Pig Model of Heart Failure

Abstract

Rationale

In cardiomyocytes from failing hearts, insufficient mitochondrial Ca²⁺ ([Ca²⁺]_m) accumulation secondary to cytoplasmic Na⁺ overload decreases NAD(P)H/NAD(P)⁺ redox potential and increases oxidative stress when workload increases. These effects are abolished by enhancing [Ca²⁺]_m with acute treatment with CGP-37157 (CGP), an inhibitor of the mitochondrial Na⁺/Ca²⁺ exchanger.

Objective

To determine if chronic CGP treatment mitigates contractile dysfunction and arrhythmias in an animal model of heart failure (HF) and sudden cardiac death (SCD).

Methods and Results

Here, we describe a novel guinea-pig HF/SCD model employing aortic constriction combined with daily β-adrenergic receptor stimulation (ACi) and show that chronic CGP treatment (ACi+CGP) attenuates cardiac hypertrophic remodeling, pulmonary edema, and interstitial fibrosis and prevents cardiac dysfunction and SCD. In the ACi group 4 weeks after pressure-overload, fractional shortening and the rate of left ventricular pressure development decreased by 36% and 32%, respectively, compared to sham-operated controls; in contrast, cardiac function was completely preserved in the ACi+CGP group. CGP treatment also significantly reduced the incidence of premature ventricular beats and prevented fatal episodes of ventricular fibrillation, but did not prevent QT prolongation. Without CGP treatment, mortality was 61% in the ACi group within 4 weeks of aortic constriction, while the death rate in the ACi+CGP group was not different from sham-operated animals.

Conclusions

The findings demonstrate the critical role played by altered mitochondrial Ca²⁺ dynamics in the development of HF and HF-associated SCD; moreover, they reveal a novel strategy for treating SCD and cardiac decompensation in HF.

INTRODUCTION

The clinical syndrome of heart failure (HF), which can occur after acute myocardial ischemic injury or following longstanding chronic cardiovascular disease, is a progressive disease often involving a phase of compensatory hypertrophy followed by a transition to impaired contractility (decompensation) with systolic and/or diastolic dysfunction. While mortality in the later stages of the disease is most often attributed to impaired pump function, sudden cardiac death (SCD) is the leading cause of mortality in patients with mild to moderate HF, accounting for 35-50% of deaths^1,2, prompting an increase in the use of implantable cardiac defibrillators (ICDs) as prophylactic devices³. The mechanisms underlying SCD, and the transition from compensated to decompensated HF, are incompletely understood, but neurohormonal abnormalities, ion channel/transporter changes, Ca²⁺ handling alterations, the activation of intra- and extra-cellular signaling pathways, and oxidative stress, have all been implicated in the pathophysiology⁴.

A limitation of energy production and/or substrate utilization is also thought to contribute to HF^5-7 and several studies have reported HF-associated defects in mitochondrial function^8-10. Mitochondria are not only the predominant source of ATP, upon which normal cardiac function depends, but also play a critical role in other biological processes, such as redox regulation, intermediary metabolism, cell signaling, and ion homeostasis. In the context of HF, the most detrimental consequences of mitochondrial dysfunction, either as the primary or as a secondary factor in the development of HF, are impaired bioenergetics and increased oxidative stress. Reactive oxygen species (ROS) have complex effects on the development of HF, playing a key role in cardiac remodeling by activating multiple signaling pathways involved in hypertrophic growth of cardiomyocytes, apoptosis, and proliferation of fibroblasts¹¹. Increased oxidative stress also impairs cardiac function directly by modifying the proteins involved in excitation-contraction coupling^12,13. Targeting mitochondria to improve bioenergetics and to maintain redox balance is emerging as a novel therapeutic strategy for HF^14-16.

Upstream of oxidative stress, recent evidence has suggested that altered Na⁺ and Ca²⁺ gradients across the mitochondrial inner membrane may be a proximal factor in the imbalance of energy supply and demand^17,18, as well as ROS balance¹⁹, in failing heart cells. The pathway of contractile and electrical dysfunction involves a vicious cycle whereby increased cytoplasmic Na⁺([Na⁺]_c), along with impaired SR Ca²⁺ release, leads to blunted Ca²⁺ signaling to the mitochondria during increased work^17-20. Consequently, the failure to stimulate Ca²⁺-dependent dehydrogenases of Krebs cycle^21-25 results in the net oxidation of the matrix NADH pool. The decreased NADH/NAD⁺ redox potential, on the one hand, results in a deficiency of reducing equivalents that drive oxidative phosphorylation (OXPHOS) and ATP production, and, on the other hand, compromises the ability to scavenge ROS, owing to the interdependence of the NADH and NADPH redox pools²⁰ (see schema in Online Figure I). Mitochondrial NADPH is crucial because it provides the reducing power to maintain the reduced glutathione (GSH), thioredoxin (Trx(SH)2) and glutaredoxin pools, all of which are vital for ROS scavenging and the prevention of protein thiol oxidative damage²⁶.

Mitochondrial Ca²⁺ ([Ca²⁺]m) is determined by the balance of influx through the mitochondrial Ca²⁺ uniporter (MCU) and efflux through the mitochondrial Na+/Ca²⁺ exchanger (mNCE); hence, a change in either of these pathways can disrupt mitochondrial Ca²⁺ dynamics. The Km of mNCE for Na+ is ~5-10 mmol/L, falling within the range of [Na⁺]_c levels in cardiomyocytes²⁷. Therefore, the elevated [Na⁺]_C consistently observed in HF can significantly accelerate the [Ca²⁺]m efflux rate. Studies of isolated heart mitochondria have shown that an increase of extra mitochondrial Na⁺ leads to decreased [Ca²⁺]m and compromises mitochondrial energetics, including oxidation of the NADH pool, decreased OXPHOS rate and ATP levels^28,29. Similarly, elevated [Na+]_c in isolated myocytes from failing guinea pig hearts¹⁹ (or ouabain-treated myocytes³⁰) causes insufficient [Ca²⁺]m accumulation during increased work, net oxidation of NAD(P)H, and a massive increase of intracellular ROS³¹, ultimately resulting in abnormal Ca²⁺ regulation and arrhythmias³⁰. The adverse effects of high [Na+]_c are abolished by CGP-37157 (CGP), an inhibitor of mNCE^19,30,31.

In the present study, we examine whether in vivo treatment with CGP prevents the development of HF and SCD using a novel guinea pig model of HF induced by left ventricular pressure overload combined with a daily β-adrenergic challenge. We show that partial inhibition of mNCE attenuates cardiac remodeling, preserves cardiac contractile function, and protects against HF- associated arrhythmias and SCD, significantly improving overall survival. The results highlight mNCE as a novel target for HF treatment.

METHODS

Methods are described in detail in the Online Supplement.

RESULTS

Daily β-adrenergic challenge accelerates the transition from LV hypertrophy to HF

A guinea pig HF model was developed by combining ascending aortic constriction (AC) with daily administration of the β-adrenergic agonist isoproterenol via i.p. injection (ACi). The rationale was to expose the pressure-overloaded heart to a period of increased work each day to simulate β-adrenergic stress. The isoproterenol dose was chosen so that it would have minimal effects on cardiac hypertrophic remodeling in the absence of AC. There was no evidence of cardiac hypertrophic remodeling in sham-operated animals exposed to the 4-week isoproterenol challenge (SHAMi) as compared with the sham-operated, untreated group (SHAM). However, in the presence of AC, the same isoproterenol treatment significantly exacerbated the decline in cardiac fractional shortening (FS) evoked by AC alone (Online Figure II). In ACi, FS decreased from 46.5±1.0% at week 0 (pre-banding) to 30.5±1.5% at week 4, whereas with AC alone, FS was not significantly decreased by week 4 (44.2±1.0% at week 0 versus 43.4±2.2% at week 4). Significant cardiac dysfunction in the AC group was observed only after 6 weeks of pressure overload (FS 38.4±1.2%), while at week 8, FS decreased to 29.3±3.3%, which was similar to the functional decline evident at week 4 in ACi (Online Figure II). Thus, the major effect of the daily isoproterenol challenge was to accelerate the transition from LV hypertrophy (LVH) to decompensated HF.

Inhibition of mNCErestores mitochondrial Ca²⁺ accumulation and prevents energetic mismatch and oxidative stress in myocytes from ACi hearts during a rapid increase in work

Impaired mitochondrial Ca²⁺ signaling secondary to cytoplasmic Na⁺ overload is a major cause of energy/redox imbalance in myocytes from pressure-overload induced failing hearts and interventions that raise [Ca²⁺]m above a required threshold prevent cellular ROS overload^19,31. Therefore, we first examined whether the same defect was present in the ACi model and if acute CGP treatment could mitigate the problem in vitro. [Na⁺]_c overload was also observed in the ACi model. [Na⁺]_c in cardiomyocytes isolated from ACi heart was increased by 208% compared to that in myocytes from SHAMi hearts (4.8±1.2 mmol/L in Shami v.s. 14.8±1.1 mmol/L in ACi) (Online Figure III). [Ca²⁺]_m dynamics, monitored with MityCam fluorescence, showed that mitochondria take up Ca²⁺ on a beat-to-beat basis (Fig. 1A). Upon stimulation (0.1 Hz), [Ca²⁺]_m increased abruptly followed by decay during diastole. An increase of work from 0.1Hz to 1Hz stimulation resulted in a further increase in MityCam signal (1-F/F0) (Fig. 1A). Measurement of peak MityCam signal indicated that accumulation of [Ca²⁺]_m during increased work was significantly less in myocytes from ACi hearts than SHAMi hearts, whereas 1μmol/L CGP treatment fully restored [Ca²⁺]_m in the ACi myocytes (Fig. 1 A and B). Isolated cardiomyocytes were challenged with a rapid increase in work from the resting state to 4 Hz stimulation in the presence of 100 nmol/L isoproterenol. NAD(P)H autofluorescence was strongly oxidized during 4Hz-stimulation in ACi cardiomyocytes (Fig. 1A). NAD(P)H decreased from 67%±4.4 reduced at rest to 33%±5.7 at the end of the stimulation period, while in SHAMi myocytes subjected to the same protocol, NAD(P)H levels were well maintained (67%±5.1 reduced at pre-stimulation vs. 66%±4.8 at the end of stimulation). Notably, the oxidation of NAD(P)H in the ACi myocytes was prevented by treatment with 1μmol/L CGP-37157 (68%±5.5 reduced at pre-stimulation vs. 64%±9.1 at the end of stimulation) (Fig. 1A). CGP had no effect on the sarcolemmal Na⁺-Ca²⁺ exchange current at concentrations up to 10 μM (Online Figure IV A), confirming its specificity for the mitochondrial Na⁺-Ca²⁺ exchanger.

Figure 1. The effects of CGP-37157 on [Ca²⁺]_m dynamics, NAD(P)H oxidation, and ROS production in myocytes isolated from ACi hearts.

A) Representative recording of MityCam fluorescence (1-F/F0) in cardiomyocytes isolated from SHAMi (black) and ACi with (light gray) or without (gray) the presence of 1μmol/L CGP-37157 (CGP) in the perfusate. B) Mean peak MityCam fluorescence (1-F/F0) during 1Hz stimulation in SHAMi (n=30), ACi (n=29), and ACi + CGP (n=13). *, p<0.001 as compared to SHAMi; **, p<0.001 as compared to ACi. In the presence of 100nmol/L isoproterenol, NAD(P)H autofluorescence (A) and CM-DCF oxidation (B) were recorded in myocytes isolated from SHAMi hearts (n=3) or ACi hearts, with (n=3) or without (n=3) 1μmol/L CGP-37157 (CGP) in the perfusate. Box indicates duration of 4Hz field-stimulation.

The oxidation of the NAD(P)H pool during increased work in ACi cells was associated with increased oxidative stress. ROS accumulation was monitored with the H₂O₂-sensitive fluorescent probe 5-(and -6)-chloromethyl-2’,7’-dichlorodihydrofluorescein diacetate (CM-DCF). CM-DCF oxidation rate in SHAMi myocytes was low, with no significant change induced by the increased work (Fig. 1B). In contrast, ROS accumulation in ACi cells was markedly increased upon stimulation (Fig. 1B). During stimulation, the CM-DCF oxidation rate in ACi cells was ~10-fold higher than that in SHAMi cells. The enhanced intracellular ROS accumulation rate in ACi cells was abolished by CGP-37157 treatment (Fig. 1B).

Effects of chronic CGP-37157 treatment on ACi-induced cardiac remodeling and functional deterioration

Significant cardiac hypertrophy, pulmonary congestion, and interstitial fibrosis were present in the ACi group by the end of week 4. HW/TL was 53% higher in ACi (0.75±0.04 g/cm) versus SHAM (0.49±0.02 g/cm) and LW/TL was 66% higher in ACi (5.69±0.43 g/cm) versus SHAM (3.41±0.08 g/cm) (Fig. 2A). Histological analysis showed that interstitial fibrosis in ACi hearts was 12.5-fold higher than in SHAM (Fig. 2B). Isoproterenol challenge in the absence of pressure overload did not significantly increase fibrosis, although there was a trend towards more fibrosis in the SHAMi group relative to the SHAM controls (percent fibrotic tissue area was 0.53±0.2 in SHAM versus 1.1±0.5 in SHAMi, p=0.052) after 4-weeks of injections (Fig. 2B). CGP treatment protected against HF-associated cardiac remodeling: the ACi+CGP group developed less cardiac hypertrophy (HW/TL: 0.64±0.04 g/cm, p<0.05 versus ACi) and pulmonary congestion (LW/TL: 4.04±0.22 g/cm, p<0.001 versus ACi) at 4 weeks compared to the ACi group (Fig. 2A). CGP treatment also attenuated interstitial fibrosis, which was decreased by 50% in ACi+CGP group compared to that in ACi group (Fig. 2B).

Figure 2. The effects of CGP-37157 treatment on hypertrophic remodeling and cardiac function.

A) Heart weight, HW (left panel), and Lung weight, LW (right panel), were measured at 4 weeks after surgery and normalized to tibia length (TL). Isoproterenol challenge had no effects on either HW/TL or LW/TL in sham-operated animals (SHAMi vs. SHAM). Pressure overload with daily isoproterenol challenge (ACi) induced cardiac hypertrophy and lung congestion, and CGP-37157 treatment attenuated both cardiac hypertrophy and lung congestion (ACi+CGP). B) Interstitial fibrosis was analyzed with Masson’s Trichrome staining. Left panel: representative images from each experimental group; right panel: measurements of interstitial fibrous tissue showing that pressure overload, together with isoproterenol challenge, induced significant fibrosis in ACi group, which was attenuated by CGP-37157 treatment (ACi+CGP). C) Cardiac function is preserved by CGP-37157. Upper panel: representative pictures of echocardiographic study; lower panel: measurements of fractional shortening (FS) (left) and diastolic LV internal dimension (LVIDd) (right) showing reduced FS and LV dilation in ACi group but not in CGP-37157 treated group. D) Left panel: representative LV pressure-volume loops. Heart from ACi group shows increased end-diastolic volume and depressed end-systolic pressure-volume relation. Right panel: summary data of end-systolic LV pressure (upper panel) and dP/dt_ip (dP/dt normalized to pressure) (lower panel). Isoproterenol treatment in the absence of aortic constriction (SHAMi) had no effect on cardiac function.

Pressure overload with daily β-adrenergic challenge induced LVH and progressive cardiac dilation and impaired contractility within 4 weeks of surgery (Fig. 2C). Compared to SHAM, FS in the ACi group was decreased by 36% (28.9±1.6% in ACi vs. 44.9±0.9% in SHAM, p<0.001) and diastolic LV internal dimension (LVIDd) was increased by 51% (1.07±0.04 cm in ACi vs. 0.71±0.01 cm in SHAM, p<0.001) (Fig. 2C). Contractile impairment was also confirmed by in vivo hemodynamic studies: +dP/dt normalized to pressure (+dP/dt_ip) in the ACi group was decreased by 32% compared to that of SHAM controls (73.5±4.0 in SHAM vs. 49.9±2.1 in ACi, p<0.001) (Fig. 2D). In contrast, animals treated with CGP displayed well-maintained cardiac function without LV dilation. No significant differences in FS, LVIDd, or +dP/dt_ip between the SHAM and ACi+CGP groups were observed. β-adrenergic treatment alone (SHAMi vs. SHAM) had no effect on cardiac hemodynamic parameters.

Effects of β-adrenergic challenge, pressure overload and CGP-37157 on heart rate

Heart rate (HR) responses to the daily isoproterenol injection were recorded in a subset (3-6 guinea pigs) of animals from each experimental group by means of implanted biopotential telemetric transmitters. Baseline and maximal HR, determined by analysis of ECG acquired during and after isoproterenol injections, were compared at 1 and 4 weeks post-surgery (Online Figure V). Baseline HR was unaffected by chronic β-adrenergic treatment in the absence of pressure overload (SHAMi versus SHAM); however, the ACi group had higher baseline HR at weeks 1 and 4 after pressure overload compared to SHAMi controls (Online Figure V). Chronic treatment with CGP-37157 (ACi +CGP), administered continuously at dose of 0.015 mg/kg/hour by means of an implanted osmotic pump, attenuated, but did not completely eliminate, the increase of baseline HR at week 1 but abolished it at week 4 (baseline HR at week 4: 218.8±9.0 bpm in SHAMi vs. 206.2±7.5 bpm in ACi+CGP). Following injection of isoproterenol, HR rapidly increased to a maximum of ~390-400 bpm in all groups and then slowly declined back to the baseline rate in 4-5 hours (Online Figure V). The peaks and kinetics of the normalized chronotropic response to isoproterenol were not significantly different between the experimental groups.

Antiarrhythmic effects of CGP-37157

Telemetry studies revealed that animals in the ACi HF group had a high incidence of premature ventricular beats (PVB) (Fig. 3). PVB frequency was significantly decreased by CGP treatment. To evaluate the antiarrhythmic effects of CGP, PVB were counted in a 2 hour-period at baseline and a 4 hour-period following isoproterenol injection at weeks 1 (Fig. 3A) and 4 (Fig. 3B) after surgery. At baseline, occasional PVB were detected in all groups, but the incidence of PVB was very low in SHAM and SHAMi groups. Moreover, the rates did not change from week 1 to week 4 (week1: 3.3±2.0; week 4: 4.0±2.5 counts/hour). In contrast, at baseline, the ACi group had a much higher incidence of PVB (23.5±5.3 counts/hour in week 1) and the frequency increased as HF developed (51.3±10.1 count/hour in week 4). CGP treatment significantly reduced the baseline frequency of PVB as compared to ACi (ACi+CGP PVB frequency: 9.2±6.3 counts/hr in week 1 and 13.0±3.7 counts/hr in week 4). Isoproterenol increased PVB frequency in all groups. At week 1, the PVB frequency in SHAMi after isoproterenol (1 mg/kg) was 104.0±36.2 counts/hour, while in ACi, the PVB frequency increased to 388.4±35.1 counts/hour after isoproterenol injection (Figs. 3A and 3B, right panels). CGP treatment inhibited the arrhythmogenic effect of the β-adrenergic challenge: PVB frequency following isoproterenol injection in the ACi+CGP group was 219.8±31.2 count/hour, a decrease of 43% compared to the ACi group. Although, as mentioned above, the HR response to isoproterenol was similar in all of the groups, the increase in PVB frequency following the injection was much less in week 4 than in week 1 (30.6±8.1 counts/hr in SHAMi, 105.9±12.5 in ACi, and 32.5±8.0 in ACi+CGP) (Fig. 3A and 3B, right panels), even though the isoproterenol dose was higher (2mg/kg). Increased intrinsic sympathetic drive related to post-operative stress in week 1 or β-adrenergic receptor desensitization in week 4 might explain this difference.

Figure 3. The antiarrhythmic effect of CGP-37157.

A and B) Left panels: Representative telemetric ECG recordings at week 1 (A) and week 4 (B) after aortic constriction showing baseline records before isoproterenol injection, records after isoproterenol injection (post-ISO), and during recovery 4 to 5 hours after isoproterenol injection. Black arrows indicate PVBs; white arrows indicate VF. Right panels: PVB incidence at baseline and after isoproterenol injection in week 1 and week 4. *, p<0.05 compared to SHAMi group, †, p<0.05 ACi+CGP compared to ACi group. C) Left panels: representative ECG recordings at baseline for each group at weeks 1 and 4; right panels: summary measurements of QTc at week 1 and week 4 showing that long QT was induced by pressure overload. Neither CGP-37157 treatment nor isoproterenol challenge had any effect on QTc.

The antiarrhythmic effect of CGP was greater in week 4, both at baseline and after isoproterenol injection (Fig. 3A and 3B). CGP treatment (comparing ACi+CGP with ACi) suppressed baseline PVB frequency by 60% in week 1 and 75% in week 4. In the contex of acute isoproterenol injection, CGP decreased PVB frequency by 43% in week 1 and 67% in week 4. Analysis of ECG data revealed prolonged QT intervals (QTc is corrected for HR) in both ACi and ACi+CGP groups (Fig. 3C). There was no difference in QTc between ACi and ACi+CGP groups at either week 1 or week 4, indicating that CGP treatment did not reverse QT prolongation in the HF model (Fig. 3C; right panels). QTc was also unaffected by the daily β-adrenergic challenge; there was no difference in QTc interval between SHAM and SHAMi groups.

Prevention of HF-associated SCD by CGP-37157 and overall mortality benefit

The ACi model was associated with a remarkably high incidence of SCD. 61.3% of ACi animals died by the end of 4 weeks (Fig. 4) without overt HF symptoms (e.g, cyanosis, labored breathing, inactivity, piloerection or cachexia). ECG analysis showed that death was preceded by sustained polymorphic ventricular tachycardia and/or fibrillation (VT/VF) (Fig. 3A and 3B; left panels). Interestingly, SCD did not occur during the peak of the HR response after isoproterenol injection, but usually happened several hours after the injection (5/6 SCD events occurred between 2-5 hrs post injection, 1 occurred 9 hrs post-injection), when the HR had recovered almost to baseline. CGP treatment (ACi+CGP) significantly reduced the 4-week death rate to just 14.3%, which was not significantly different from that of the SHAMi control (Fig. 4). Moreover, all of the deaths in CGP-treated group and in the SHAMi group occurred in the first week after surgery, perhaps attributable to the post-operative stress making the animals more susceptible to arrhythmias. Survival analysis of the ACi group indicated that the death rate was higher in week 1 (22.4%) and week 4 (26.9%) compared to that in week 2 (15.8%) and week 3 (18.8%). Similar to the other groups, the higher death rate in week 1 in ACi may reflect the overall effects of post-operative stress, along with a higher susceptibility to isoproterenol-induced arrhythmias. Considering that the PVB frequency following isoproterenol injection in week 4 is only 27% of that in week 1, the higher death rate (26.9%) in week 4 suggests that PVBs may be more life-threatening due to HF-associated changes in the arrhythmogenic substrate.

Figure 4. Survival curves over 4 weeks after the time of surgery for animals in SHAM, SHAMi, ACi, and ACi+CGP groups.

Survival of animals with pressure overload plus isoproterenol challenge was significantly increased by treatment with CGP-37157.

Antiarrhythmic effects of CGP-37157 on the isolated failing heart

We next determined if arrhythmia susceptibility and cardiac function in ACi failing hearts could be rescued by acute treatment with 1 μM CGP-37157 in isolated perfused hearts. Langendorff-perfused SHAMi and ACi hearts, isolated at week 4, were challenged with 10μmol/L isoproterenol for 15 min. Cardiac contractility and ECG were analyzed before (pre-ISO), during (ISO), and after (post-ISO) isoproterenol challenge. Baseline contractility before isoproterenol was significantly reduced in ACi hearts (Fig. 5A); LVDP, +dP/dt, and −dP/dt of ACi hearts were only 37%, 38%, and 36%, respectively, of values in SHAMi hearts (Table 1). Administration of isoproterenol increased cardiac contractility by a similar percentage in both ACi and SHAMi groups. However, ACi hearts were more sensitive to the stress of isoproterenol challenge; contractility of the ACi hearts was depressed more following isoproterenol washout. In the post-ISO state, LVDP, +dP/dt, and –dP/dt of ACi hearts were 12%, 16%, and 16% of values in SHAMi hearts (Table 1). Treatment with CGP-37157 had no statistically significant effects on cardiac contractility during pre-ISO or ISO states, but attenuated contractile dysfunction following isoproterenol treatment (Table 1). Treatment with CGP-37157 improved LVDP, +dP/dt, and −dP/dt recovery in the post-ISO state by 205%, 135%, and 120%, as compared to ACi hearts without CGP-37157 treatment (Table 1).

Figure 5. Effects of acute CGP-37157 treatment on isolated perfused failing hearts.

A) Representative traces of LV pressure and ECG simultaneously recorded before (pre-ISO), during (ISO), and after (post-ISO) isoproterenol (10μM) exposure in hearts of shami (n=3), ACi (n=3), and ACi with CGP-37157 (1μM) treatment (ACi+CGP) (n=3). B) Representative Poincaré plots of SHAMi, ACi, and ACi+CGP hearts during pre-ISO, ISO, and post-ISO states, respectively. Ectopic activity is evident as clustered of points only in the ACi heart post-ISO (middle right panel).

Table 1.

Measurements of cardiac contractility in isolated perfused heart.

	Shami			ACi			ACi+CGP
	Pre-iso	iso	Post-iso	Pre-iso	iso	Post-iso	Pre-iso	iso	Post-iso
LVDP	66.7±11.1	89.1±14.0	46.2±12.4	24.5±11.6	33.1±13.6*	5.6±1.1*	41.5±9.1	51.2±3.8*	17.1±2.8*†
+dP/dt	1255±137	3372±218	1114±160	483±198*	988±373*	181±36*	759±122*	1514±86*	425±65*†
-dP/dt	-983±155	-2247±352	-927±354	-355±139*	-730±323*	-151±33	-563±102	-938±101*	-330±48†

^*p<0.05 when compared to contrl;

^†p<0.05 when compared to ACi.

Heart rate variability (HRV) analysis was applied to quantify changes in the ECG in the SHAMi and ACi hearts under the different experimental conditions. Treatment with 10μmol/L isoproterenol did not induce overt arrhythmias during the period of maximum LVPD in either group. There were no differences in RR interval, standard deviation of RR interval (SDRR), root mean square of successive difference (rMSSD), short term dispersion (SD1), or long term dispersion (SD2) between SHAMi, ACi, and ACi+CGP treatment groups during pre-ISO and ISO states (Table 2). However, following washout of isoproterenol, HRV of ACi hearts was dramatically increased compared to that of SHAMi hearts (clusters of points are indicative of ectopic activity in the Poincaré plots in Fig. 5B), whereas CGP treatment completely inhibited the arrhythmias in ACi hearts in the post-ISO state (Fig. 5B and Table 2). There were no differences in RR interval, SDRR, rMSSD, SD1, or SD2 between SHAMi and ACi+CGP treated hearts (Table 2).

Table 2.

HRV analysis. (SDNN, standard deviation of normal RR interval; RMSSD, root mean square of the successive differences)

	Shami	ACi	ACi+CGP
Time domain Results
Mean RR (ms):	260±18	336±25	307±36
SDNN (ms):	3.85±1.39	42.08±8.51*	2.76±1.69†
RMSSD (ms):	0.88±0.29	65.60±15.60*	1.14±0.59†
Nonlinear Results
SD1 (ms):	0.626±0.207	46.418±11.048*	0.804±0.415†
SD2 (ms):	5.394±1.968	36.542±6.167*	3.812±2.347†

^*p<0.05 when compared to contrl;

^†p<0.05 when compared to ACi.

DISCUSSION

The major findings of the present study were that: 1) the transition from compensated pressure-overload hypertrophy to HF is exacerbated by a daily β-adrenergic challenge; 2) SCD is a major contributor to mortality in the ACi HF model; 3) the mitochondrial Na+/Ca²⁺ exchange inhibitor, CGP-37157, prevents cardiac contractile decompensation, blunts hypertrophic remodeling, decreases arrhythmia incidence, and prevents SCD associated with HF.

The present study was motivated by our earlier findings that cytoplasmic [Na+]c overload, by increasing the driving force for mNCE-mediated Ca²⁺ extrusion from the mitochondrial matrix, results in insufficient [Ca²⁺]m accumulation during increased work^17,19,30. If the Ca²⁺ signal to the mitochondria falls below a critical threshold level¹⁹, Ca²⁺-dependent activation of Krebs cycle enzymes is disrupted²⁹, causing the mitochondrial pyridine nucleotide redox potential (NAD(P)H/NAD(P)⁺) to become more oxidized. The consequent decrease in reducing equivalents driving oxidative phosphorylation (NADH) and the antioxidant pathways (NADPH) causes both an impairment of energy supply and ROS overload (Fig. 1 C and D). This pathological mechanism was demonstrated in normal myocytes overloaded with [Na+]_c¹⁷, myocytes treated with the Na+/K+ inhibitor ouabain³⁰, and in myocytes isolated from failing (AC model) guinea pig hearts¹⁹. In the present study, we demonstrated that the net oxidation of NAD(P)H pool and increased oxidative stress during rapid pacing (4 Hz) in the presence of isoproterenol is also present in the ACi model. The increase in NAD(P)H oxidation, ROS accumulation, and arrhythmias associated with [Na+]_c overload, can be prevented by either lowering [Na+]_c¹⁹ or by inhibition of mNCE with CGP-37157^19,30,31 (and the present work; Fig. 1C and D).

The findings are the first to demonstrate that chronic treatment with CGP-37157 prevents the progressive decline in contractile function and SCD in vivo in an animal model of HF. The ACi model was developed to incorporate both hemodynamic stress (left ventricular pressure overload) and chronic sympathetic hyperstimulation, two factors which are known to contribute to HF and SCD in humans^32,33. Compared to pressure overload alone, the ACi model accelerated the decline in function and pathology of HF, while isoproterenol treatment in the absence of pressure overload (SHAMi) had no significant effects on cardiac function and remodeling in control animals.

A unique feature of the ACi model is the high incidence of SCD (~60% at 4 weeks), which is manifested as an increased death rate even in the first week post-aortic constriction, when function is largely preserved, and continues over the course of heart failure progression. Notably, the ACi model shows QTc prolongation even after only 1 week of pressure overload, prior to overt contractile dysfunction, as well as a large increase in spontaneous premature ventricular beats. The PVB rate and the incidence of VT/VF were both suppressed by CGP-37157 treatment; however, the QTc prolongation was not reversed, suggesting that the alterations in sarcolemmal ionic currents during hypertrophy, by themselves, are not sufficient to account for the increased automaticity and tendency towards reentry underlying SCD. Presumably, the combined effects of ion channel remodeling, oxidative stress, and the change in the myocardial substrate were all involved in setting the stage for VT/VF. In this light, it is still not clear if long QTc is a useful predictor of SCD in human HF. Breidthardt et al³⁴ reported prolonged QTc interval (≥440ms) in 72% of patients with acute destabilized heart failure, but 720-day all-cause mortality was no different in patients with prolonged versus normal QTc intervals. Patients with prolonged QRS interval, however, had a significant 2-fold increase in mortality. Similarly, QTc was determined to have no independent prognostic value on 1 year mortality in a subset of patients with congestive heart failure in the DIAMOND trial³⁵. In general, though, the overall risk of malignant arrhythmias increases with the magnitude of QTc prolongation³⁶ and more refined analyses, such as QTc variability, may have predictive value for SCD³⁷, perhaps as a better indicator of QT dispersion³⁸.

The present results suggest that prolonged QTc may provide the substrate for potentially fatal arrhythmias, but that a secondary event - we propose an oxidative stress event secondary to an energy/redox supply and demand mismatch - would be required to initiate SCD. The non-linear nature of the secondary event could therefore account for the unpredictability of SCD.

Apart from genetically engineered animals³⁹, the ACi model is, to our knowledge, one of the few animal models of HF-associated acquired long QT that displays a high incidence of spontaneous SCD. The accelerated transition between the compensated and decompensated states should be useful for investigating the mechanisms behind progressive systolic and diastolic dysfunction, as well as vulnerability to arrhythmias, in future studies.

Altered β-adrenergic receptor (β-AR) signaling is known to be a key player in the development of HF^40,41. In addition to impairment of the inotropic, chronotropic and lusitropic effects of β-AR stimulation in HF, chronic hyperactivation of the receptor results in receptor desensitization, detrimental effects on cardiac function, and structural remodeling, which can be mitigated by β-blockade. The impact of selective β-AR signal activation on the heart depends on the dose and duration of β-AR agonist administration⁴². In our study, the chronic effects of isoproterenol on cardiac remodeling were minimized by selecting an isoproterenol dose that had no independent hypertrophic effects in sham operated animals and by administrating isoproterenol with repeated injections that induced only transient activation of β-AR signaling. The β-AR challenge had detrimental functional effects (arrhythmias and impaired post-ISO contractility) on LV pressure-overloaded animals, but not on controls, suggesting that the pressure-overloaded heart is more vulnerable to β-AR activation. This is consistent other animal studies. For example, in spontaneously hypertensive rats with LVH, infusion of isoproterenol initiated a transition from compensated hypertrophy to heart failure, whereas the same infusion had no significant effects on cardiac function and remodeling in control rats⁴³. Genetically-modified mouse lines that have constitutively increased β-AR drive with normal cardiac function and structure are also more susceptible to pressure overload^44,45. Our finding that CGP-37157 treatment protects pressure overloaded hearts (both in vivo and ex vivo) from the detrimental effects of β-AR challenge supports the hypothesis that blunted [Ca²⁺]m accumulation leading to oxidative stress is a key event underlying increased vulnerability of the failing heart to β-AR-mediated cardiac dysfunction.

In addition to exacerbating contractile dysfunction, β-AR challenge also predisposed the pressure overloaded heart to VT/VF, and CGP-37157 protected ACi animals from both high PVB rate and SCD. Increased sympathetic activity during physical or emotional stress is known to provoke VT/VF in catecholaminergic polymorphic ventricular tachycardia (CPVT), and elevated diastolic Ca²⁺ and hyperactive ryanodine receptors (RyR) are thought to be critical factors leading to SCD in CPVT animal models^46,47. β-AR stimulation can contribute to arrhythmias by increasing Ca²⁺ current and increasing in sarcoplasmic reticulum Ca²⁺ ATPase (SERCA) activity, both of which increase SR Ca²⁺ load to increase vulnerability to spontaneous Ca²⁺ release. In HF, although the Ca²⁺ transient amplitude and SR load is decreased, the prolonged duration of the AP and Ca²⁺ transient, together with increased Ca²⁺ leak via the RyR, could favor spontaneous after depolarizations. For example, DeSantiago et. al.⁴⁸ demonstrated that increased SR Ca²⁺ by β2-AR activation results in enhanced arrhythmogenesis in myocytes from human and rabbit failing hearts. Interestingly, in the ACi model, the pro-arrhythmic effect of β-AR challenge was not evident during the peak of the response, but was revealed only after recovery from the increased work. We propose that CGP-37157 is protecting against damage induced by mitochondrial ROS accumulation by preserving the pyridine nucleotide/antioxidant capacity during β-AR activation. Among the potential ROS targets, the local oxidation of the RyR by mitochondrial ROS⁴⁹ could underlie spontaneous Ca²⁺ release and ectopic activity. RyR channels have multiple cysteine residues that sense the local redox state, and their oxidation is responsible for increased spontaneous SR Ca²⁺ leak in HF, which can be normalized by dithiothreitol¹³. In this light, we have previously demonstrated that CGP-37157, by increasing the mitochondrial Ca²⁺ response and preserving NAD(P)H redox potential during increased work, can prevent DADs triggered by ouabain-induced Na⁺ overload in myocytes, hearts, and intact animals³⁰. This mechanism is the most likely explanation for the CGP-mediated protection against post-ISO ectopic activity observed in the isolated perfused ACi hearts (Fig. 5).

In addition to ROS-driven dysfunction, impaired mitochondrial energetics in HF also contributes to cardiac decompensation⁵⁰. For example, limitations of ATP delivery could impair diastolic function by affecting the Ca²⁺ removal capacity via SERCA when energy demand increases during β-AR stimulation. This has been observed in studies of energetically-deficient animal models. In the creatine kinase deficient mouse, the reduced energy reserve leads to a decreased capacity of SERCA to remove cytosolic Ca^{2+ 51,52}. A peroxisome proliferator-activated receptor gamma coactivator 1-β (PGC-1β) knockout mouse also displays reduced respiration rate and ATP production⁵³; these animals showed increased susceptibility to catecholamine-induced arrhythmias⁵⁴. Cardiomyocytes from these mice have normal Ca²⁺ transients at rest, but elevated [Ca²⁺]_c and spontaneous diastolic Ca²⁺ transients were induced following isoproterenol administration⁵⁴. By improving energy supply, CGP-37157 treatment could thus help to maintain the activity of SERCA or other energy-dependent processes during increased work to preserve systolic and diastolic function.

The present findings highlight the central role of oxidative stress in both the functional deterioration and arrhythmias associated with HF progression. In addition to direct damage to the myocyte, mounting evidence indicates that ROS modulate various signaling pathways involved in hypertrophic growth, apoptotic and necrotic cell death, and proliferation of cardiac fibroblasts^11,55,56. ROS are also downstream mediators of neurohumoral stimuli implicated in HF, such as angiotensin II, endothelin 1, and tumor necrosis factor α. Increased ROS can also directly activate signaling cascades including tyrosine kinases, MAP kinases, PKC, and PI3-kinase⁵⁷, which are potent regulators of cardiac growth. The effects of CGP on these downstream effectors of ROS signaling remain to be determined, but could also contribute to the observed protection. Nevertheless, given that the hypertrophic response was only partially decreased by CGP treatment, the major beneficial effect of CGP may be to prevent the redox modification of proteins, (e.g. ion channels, myofilament proteins, and mitochondrial enzymes) that lead to contractile dysfunction and deficient energetics. Recent studies confirm the preeminent role of mitochondrial ROS in HF, for example, a mitochondrially-targeted antioxidant peptide (SS-31) protects animals from Ang II-induced cardiac hypertrophy and fibrosis¹⁴ and overexpression of mitochondrial, but not cytosolic, catalase prevents changes in the proteome of mice subjected to pressure-overload-induced heart failure¹⁵. CGP treatment intervenes in the mechanism leading to the mitochondrial ROS overload, while preserving energy supply and demand balance, which may have advantages over strategies that simply attempt to increase the ROS scavenger capacity.

In summary, CGP-37157 preserves cardiac function, attenuates remodeling and fibrosis, and prevents SCD in the guinea pig heart failure and SCD model produced by pressure overload combined with β-AR stimulation. The beneficial effects of CGP-37157 can attributed to the effects of restored [Ca²⁺]m dynamics by CGP-37157 on mitochondrial energetics and redox balance. Additional studies will be needed to characterize all of the HF-associated modifications impacted by CGP-37157 treatment, including cellular redox status, ROS balance, the electrophysiology of the heart, excitation-contraction coupling, energy metabolism, and post-translational protein modifications. More work will also be required to determine if CGP treatment can reverse established left ventricular hypertrophic remodeling to prevent decompensation and SCD in this model. In addition, given the known variations in cytoplasmic Na⁺ in different species⁵⁸, the efficacy of CGP may be different in other models of HF and in humans. Nevertheless, the findings support the hypothesis that [Ca²⁺]m dynamics plays a critical role in the development of HF and HF-associated SCD and implicate mNCE as a novel therapeutic target for HF.

Novelty and Significance.

What Is Known?

• Sudden cardiac death (SCD) accounts for up to 50% of deaths in patients with heart failure (HF) and, in many cases, can be attributed to paroxysmal ventricular arrhythmias.

• The mechanisms underlying SCD are unknown but ion channel remodeling, action potential prolongation, and abnormal cytoplasmic and mitochondrial Ca²⁺ handling have been implicated.

• Energy metabolism is also known to be altered in HF but the links between impaired energy supply, oxidative stress, and contractile and electrical dysfunction are incompletely understood.

What New Information Does This Article Contribute?

• A novel guinea-pig model of HF/SCD is described, combining pressure-overload and daily catecholamine challenge, which displays impaired contractility, long-QT, and a high incidence of SCD due to spontaneous ventricular fibrillation over a 4-8 week time span.

• Treatment with an inhibitor of the mitochondrial Na⁺/Ca²⁺ exchanger (mNCE) prevents contractile decompensation and SCD in vivo but does not prevent QT prolongation.

• mNCE inhibition prevents NAD(P)H oxidation and ROS overload during increased work in myocytes from failing hearts and decreases the incidence of arrhythmias after a catecholamine challenge in perfused failing hearts.

• Impaired mitochondrial Ca²⁺ balance, and the consequent impairment of NAD(P)H availability, plays a key role in heart failure progression and vulnerability to fatal arrhythmias by compromising energy supply and antioxidant capacity in the failing heart.

Sudden cardiac death is a major cause of death in patients with heart failure, but the mechanisms underlying this increased vulnerability are poorly understood. In cardiomyocytes from failing hearts, it was previously shown that insufficient mitochondrial Ca²⁺ ([Ca²⁺]_m) accumulation, secondary to cytoplasmic Na⁺ overload, decreases NAD(P)H/NAD(P)⁺ redox potential, increases oxidative stress, and disrupts Ca²⁺ handling when workload increases. Here, we show that chronic in vivo treatment with an inhibitor of the mitochondrial Na⁺/Ca²⁺ exchanger (CGP-37157) prevents the progression from compensated hypertrophy to heart failure and eliminates the high incidence of spontaneous SCD. The findings highlight the critical role played by mitochondrial Ca²⁺ dynamics in maintaining redox and energy balance in the failing heart and support the feasibility of targeting mitochondria for therapeutic intervention in heart failure.

Nonstandard Abbreviations and Acronyms

HF

heart failure

SCD

sudden cardiac death

ICD

implantable cardiac defibrillator

ROS

reactive oxygen species

OXPHOS

oxidative phosphorylation

GSH

reduced glutathione

Trx(SH)2

reduced thioredoxin

[Na⁺]_c

cytoplasmic Na⁺

[Ca²⁺]_m

mitochondrial Ca²⁺

MCU

mitochondrial Ca²⁺uniporter

mNCE

mitochondrial Na⁺/Ca²⁺ exchanger

CGP

CGP-37157 (7-Chloro-5-(2-chlorophenyl)-1,5-dihydro-4,1-benzothiazepin-2(3H)-one)

AC

ascending aortic constriction

FS

fractional shortening

LVH

left ventricular hypertrophy

CM-DCF

5-(and -6)-chloromethyl-2’,7’-dichlorodihydrofluorescein

LVIDd

diastolic LV internal dimention

HR

heart rate

PVB

premature ventricular beat

HRV

HR variability

SDRR

standard deviation of RR interval

rMSSD

root mean square of successive difference

SD1

short term dispersion

SD2

long term dispersion