The relationship between arrhythmogenesis and impaired contractility in heart failure: role of altered ryanodine receptor function

Abstract

Aims

In heart failure (HF), abnormal myocyte Ca²⁺ handling has been implicated in cardiac arrhythmias and contractile dysfunction. In the present study, we investigated the relationships between Ca²⁺ handling, reduced myocyte contractility, and enhanced arrhythmogenesis during HF progression in a canine model of non-ischaemic HF.

Methods and results

Key Ca²⁺ handling parameters were determined by measuring cytosolic and intra-sarcoplasmic reticulum (SR) [Ca²⁺] in isolated ventricular myocytes at different stages of HF. The progression of HF was associated with an early and continuous increase in ryanodine receptor (RyR2)-mediated SR Ca²⁺ leak. The increase in RyR2 activity was paralleled by an increase in the frequency of diastolic spontaneous Ca²⁺ waves (SCWs) in HF myocytes under conditions of β-adrenergic stimulation. In addition to causing arrhythmogenic-delayed afterdepolarizations, SCWs decreased the amplitude of subsequent electrically evoked Ca²⁺ transients by depleting SR Ca²⁺. At late stages of HF, Ca²⁺ release oscillated essentially independent of electrical pacing. The increased propensity for the generation of SCWs in HF myocytes was attributable to reduced ability of the RyR2 channels to become refractory following Ca²⁺ release. The progressive alterations in RyR2 function and Ca²⁺ cycling in HF myocytes were associated with sequential modifications of RyR2 by CaMKII-dependent phosphorylation and thiol oxidation.

Conclusion

These findings suggest that destabilized RyR2 activity due to excessive CaMKII phopshorylation and oxidation resulting in impaired post-release refractoriness is a common mechanism involved in arrhythmogenesis and contractile dysfunction in the failing heart.

1. Introduction

Systolic heart failure (HF) is an increasingly common and often lethal disease that occurs when the cardiac muscle is too weak to maintain a sufficient cardiac output. Patients with HF typically die either due to progressive failure of cardiac mechanical function (pump failure) or ventricular arrhythmias.¹ Multiple mechanisms, including remodelling of cellular ionic currents, alterations in gap junctions, and changes in the contractile apparatus, have been implicated in the pathophysiology of HF.^2,3 In addition, abnormal myocyte Ca²⁺ handling is increasingly recognized as an important factor in the development of both ventricular arrhythmias and contractile dysfunction in the failing heart.^4–7 However, the specific mechanisms by which alterations in Ca²⁺ cycling lead to these two different clinical manifestations of HF are poorly defined.

Arrhythmogenesis in the failing heart has been associated with excessive, spontaneous sarcoplasmic reticulum (SR) Ca²⁺ release, which occurs in the form of waves of Ca²⁺-induced Ca²⁺ release (CICR; i.e. spontaneous Ca²⁺ waves, SCWs) causing oscillations in myocyte membrane potential known as early afterdepolarizations and delayed afterdepolarizations (DADs).^4,5,8,9 The enhanced predisposition of myocytes isolated from failing hearts to SCWs has been ascribed to elevated SR Ca²⁺ load secondary to the stimulation of SR Ca²⁺ ATPase (SERCA) by the increased levels of circulating catecholamines combined with enhanced β₂-adrenergic responsiveness in HF.¹⁰ Additionally, up-regulation of the sodium/calcium exchanger (NCX), which would permit a more effective translation of cytosolic Ca²⁺ elevations into DADs, has been suggested to contribute to the arrhythmogenic propensity during HF.^4,11

In contrast, the weakened mechanical performance of the failing heart has been linked to reduced SR Ca²⁺ release and weakened contractility of cardiac myocytes.¹² Increased leak of Ca²⁺ from the SR (diastolic leak) is considered to be an important factor contributing to contractile dysfunction in HF secondary to reduced SR Ca²⁺ content.^13–16 Post-translational modifications of the cardiac ryanodine receptor (RyR2) either through phosphorylation by PKA¹⁴ or CaMKII,¹⁶ or oxidation of reactive thiols,¹⁷ have been proposed as the primary mechanisms responsible for altered RyR2 function in HF. In addition to increased SR Ca²⁺ leak, enhanced NCX function and reduced SR Ca²⁺ uptake by SERCA have been shown to contribute to reduced SR Ca²⁺ content in several different forms of HF.⁶ Thus, there appears to be a contradiction regarding the mechanisms causing arrhythmogenesis and weakened contractility in HF: how can arrhythmogenesis based on increased SR Ca²⁺ content and excessive SR Ca²⁺ release coexist with reduced contractility due to decreased SR Ca²⁺ content and insufficient Ca²⁺ release?

The relationship between arrhythmogenesis and contractile dysfunction in HF is further complicated by the complex dynamics of the disease process, comprising at least two stages: an early, compensated stage followed by a late, decompensated stage.³ Although impaired SR Ca²⁺ release and diminished myocyte contractility are considered to be key features of late-stage HF,¹² their contribution to the pathology at earlier stages of HF remains to be clarified. Similarly, it is uncertain whether Ca²⁺-dependent arrhythmogenesis parallels the development of myocyte failure or has a discrete time course during the progression of HF.

In the present study, we investigated the relationship between Ca²⁺ handling and arrhythmogenesis during HF progression using a canine model of non-ischaemic HF. Our results suggest a unifying mechanism for arrhythmogenesis and contractile impairment in HF whereby a progressive destabilization of RyR2 function via phosphorylation and thiol oxidation results in impaired Ca²⁺ signalling refractoriness and irregular SR Ca²⁺ release.

2. Methods

An expanded methods section can be found in the Supplementary material online.

2.1. Canine model of non-ischaemic HF

Ventricular dysfunction was induced by right ventricular (RV) tachypacing (TP), as described previously.¹⁷ Briefly, adult hound dogs (25–28 kg) of either sex were chronically instrumented with modified Prevail 8086 pacemakers (Medtronic, Inc., Minneapolis, MN, USA) with the pacing lead (Medtronic Model 4092) placed in the RV apex. TP was performed at: 180 b.p.m. for 2 weeks, 200 b.p.m. for 6 weeks, 180 b.p.m. for 2 months, 160 b.p.m. for 6 months, and then 120 b.p.m. thereafter. The investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). All animal procedures were approved by Institutional Animal Care and Use Committee of the Ohio State University.

2.2. Ca²⁺ imaging in myocytes

Left ventricular mid-myocardial myocytes were isolated by using collagenase (Worthington type II, 0.65 mg/mL) perfusion via the left circumflex artery. Ca²⁺ currents and action potentials (APs) were recorded using patch-clamp techniques.¹³ Changes in cytosolic Ca²⁺ were monitored with Rhod-2 or Fluo-3, whereas changes in intra-SR Ca²⁺ were assessed with Fluo-5N, as described previously.¹³ SERCA activity was assessed by recording SR-trapped Fluo5N fluorescence in saponin-permeabilized myocytes: the SR Ca²⁺ store was depleted by the application of 10 mM caffeine in a Ca²⁺-free medium and SR Ca²⁺ uptake was measured upon the re-introduction of Ca²⁺ in the presence of the RyR2 inhibitor ruthenium red¹³.

The temporal dynamics in fluorescence of cytosolic dyes (Fluo-3 or Rhod-2) were expressed as F/F₀ or ▵F/F₀ ((F − F₀)/F₀), where F represents the fluorescence at time t and F₀ represents the background fluorescence. Fluorescence of Fluo-5N was usually expressed as F_norm = (F– F_min)/(F_max – F_min), where F_min represents the fluorescence after the application of 10 mM of caffeine and F_max represents the Fluo-5N fluorescence in the presence of 20 mM of [Ca²⁺]. Fluo-5N fluorescence recorded in voltage-clamped myocytes was converted to [Ca²⁺] according to the following equation: [Ca²⁺] = K_d× (F– F_min)/(F_max – F), where K_d was 400 µM.¹⁸

2.3. Phosphorylation and oxidation of RyRs

Phosphorylation of RyRs at Ser-2808 and Ser-2814 were probed with antibodies kindly provided by Dr A.R. Marks (Columbia University, NY, USA) using standard procedures.¹⁷ Reactive oxygen species (ROS) production was measured with the fluorescent indicator 5-(and-6) chloromethyl-2′,7′-dichlorodihydrofluoroscein diacetate (DCFDA).¹⁷ After subtraction of background fluorescence, the signal was normalized to maximum fluorescence achieved by the application of 10 mM H₂O₂. The content of free thiols in RyRs was determined with the monobromobimane (mBB) fluorescence method.¹⁷ mBB fluorescence was normalized to RyR2 amount determined using Coomassie Blue staining of the gels run in parallel.

2.4. Data analysis

Results are presented as mean ± SEM. Statistical significance was evaluated either by Student's t-test or by one-way ANOVA. A P-value of <0.05 was considered significant.

3. Results

3.1. Sustained functional stability of excitation–contraction during HF progression

First, to examine the effect of HF on basic aspects of myocyte Ca²⁺ handling, we measured intracellular Ca²⁺ transients in isolated intact field-stimulated cardiac myocytes from normal (control) animals and from animals at early and advanced stages of disease (1 and 4 months of TP). The amplitudes and rates of decay of Ca²⁺ transients were not altered in the 1 month heart failure (MHF) group but were significantly decreased in the 4 MHF when compared with controls (Figure 1A and B; Table 1). To gain further insights into the impact of HF on excitation–contraction (EC) coupling, we measured Ca²⁺ transients, Ca²⁺ currents (I_Ca) and SR Ca²⁺ content in patch-clamped cardiac myocytes from control and HF animals. The average density of peak Ca²⁺ currents (recorded at 0 mV) was not altered during HF progression up to 24 months of TP (Figure 1C and D). A significant decrease in the amplitude of Ca²⁺ transients was observed starting at 4 MHF (Figure 1C and D; Table 1). Additionally, the SR Ca²⁺ content was significantly reduced during 4 and 16+ MHF (Figure 1E and F). The EC coupling gain (Ca²⁺ transient amplitude/I_Ca amplitude) normalized to the SR Ca²⁺ content, a measure of the efficiency of EC coupling, was significantly increased at 1 and 4 MHF but tended to decrease at 16+ MHF (Table 1). Thus, although myocyte EC coupling demonstrated signs of remodelling even at early stages of HF (i.e. increased EC efficiency), the intrinsic ability of the SR to release Ca²⁺ was not diminished until advanced stages of disease.

Figure 1.

Time course of decrease in cytosolic Ca²⁺ transient during HF progression. (A) Representative line-scan images and corresponding profiles of cytosolic Ca²⁺ transients ([Ca²⁺]_c) evoked by electrical field stimulation at 0.3 Hz in control (0 months of TP) and HF myocytes at the indicated duration of TP. (B) Average amplitudes and time constants of exponential fit of decaying phase of [Ca²⁺]_c obtained in control and in HF myocytes. (C) Representative traces of [Ca²⁺]_c and Ca²⁺ currents (I_Ca) evoked by depolarizing steps from −50 to 0 mV in control and HF myocytes. (D) Average amplitudes of [Ca²⁺]_c and average peak density of I_Ca recorded in control and HF groups. (E) Representative traces of Ca²⁺ transients evoked by 10 mM caffeine recorded in control and HF patch-clamped myocytes. (F) Average amplitudes and time constants of decay of caffeine-induced Ca²⁺ transients ([Ca²⁺]_Caff) recorded in control and HF groups. *P < 0.05 vs. control; ^†P < 0.05 vs. 1 MHF.

Table 1.

Properties of Ca²⁺ signalling in control and during different time stages of HF

Properties of Ca2+ transients evoked by field stimulation	Control (n = 25)	1 MHF (n = 21)	4 MHF (n = 36)	16+ MHF
Amplitude (▵F/F0)	1.2 ± 0.1	1.1 ± 0.1	0.8 ± 0.1*	n.d.
Time to peak (ms)	295 ± 20	465 ± 60*	476 ± 36*	n.d.
Decay (τ, ms)	417 ± 20	467 ± 23	561 ± 25*,†	n.d.
Properties of Ca2+ currents (ICa) and Ca2+ transients in patch-clamped myocytes	Control (n = 12)	1 MHF (n = 14)	4 MHF (n = 26)	16+ MHF (n = 11)
Peak ICa (pA/pF)	−3.9 ± 0.3	−3.1 ± 0.5	−3.1 ± 0.2	−4.0 ± 0.2
Amplitude (▵F/F0)	1.9 ± 0.1	1.6 ± 0.2	1.4 ± 0.1*	1.1 ± 0.1*
Time to peak (ms)	249 ± 16	247 ± 21	291 ± 9	264 ± 18
Decay (τ, ms)	543 ± 30	540 ± 21	612 ± 24	672 ± 38†
Properties of caffeine-induced Ca2+ transients in patch-clamped myocytes	Control (n)	1 MHF (n)	4 MHF (n)	16+ MHF (n)
Amplitude (▵F/F0)	5.2 ± 0.3 (13)	4.3 ± 0.4 (8)	4.0 ± 0.3* (19)	3.4 ± 0.2* (6)
Decay (τ, s)	3.1 ± 0.2 (10)	2.2 ± 0.2* (6)	2.1 ± 0.1* (12)	2.3 ± 0.1* (6)
Gain of Ca2+-induced Ca2+ release (normalized to SR Ca2+ content)	0.10 ± 0.01 (12)	0.15 ± 0.02 (11)	0.16 ± 0.02* (18)	0.08 ± 0.01‡ (6)
Properties of the local SR Ca2+ release and uptake	Control (n)	1 MHF (n)	4 MHF (n)	16+ MHF (n)
Spark frequency (100 μm−1 s−1)	1.9 ± 0.2 (46)	4.1 ± 0.5* (63)	5.3 ± 0.5* (33)	11.2 ± 1.5*,†,‡ (30)
SR Ca2+ leak rate (10−3s−1)	1.3 ± 0.1 (44)	1.8 ± 0.2 (35)	1.9 ± 0.2* (31)	2.7 ± 0.4*,† (10)
SR Ca2+ uptake rate (10−3s−1)	79 ± 5 (24)	64 ± 0.2 (12)	90 ± 7 (14)	70 ± 6 (9)

n.d., not determined.

*P < 0.05 vs. control. ^†P < 0.05 vs. 1 month. ^‡P < 0.05 vs. 4 months (ANOVA).

NCX Ca²⁺ transport activity, estimated by measuring the rate of decay of caffeine-induced Ca²⁺ transients, increased significantly at 1 MHF but did not change further with the progression of HF (Figure 1E and F; Table 1). Thus, changes in NCX by themselves were not likely to be a significant determinant of alterations in Ca²⁺ cycling observed at late stages of HF.

3.2. RyR2-mediated SR Ca²⁺ leak arises early and increases progressively during HF development

HF-dependent changes in myocyte Ca²⁺ signalling were further studied by measuring Ca²⁺ sparks in permeabilized myocytes. The frequency of Ca²⁺ sparks increased approximately two-fold in the 1 MHF group compared with control and continued to increase further with the progression of HF (Figure 2A and D; Table 1). Consistent with Ca²⁺ spark measurements, SR Ca²⁺ leak, measured as a loss of intra-SR [Ca²⁺] on the inhibition of SERCA-mediated SR Ca²⁺ re-uptake by thapsigargin¹³, progressively increased with HF progression (Figure 2B and D; Table 1). Thus, HF development was associated with a gradual enhancement in RyR2 functional activity manifested in Ca²⁺ sparks and SR Ca²⁺ leak. At the same time, SERCA Ca²⁺ transport activity did not change in up to 24 MHF (Figure 2C and E; Table 1).

Figure 2.

Time-dependence of increase in the SR Ca²⁺ leak during HF progression (A) Representative line-scan images of Ca²⁺ sparks recorded in permeabilized control and HF myocytes at the indicated time stages of HF. (B) Time-dependent profiles of intra-SR Fluo-5N fluorescence recorded in the presence of the SERCA inhibitor thapsigargin (Tg, 10 μM) in control and HF myocytes. The decline of Fluo-5N signal in the presence of Tg was fitted to a monoexponential function. (C) Representative time-dependent profiles of Fluo-5N fluorescence used to calculate SR Ca²⁺ uptake were recorded in control and HF myocytes in the presence of the RyR2 inhibitor ruthenium red (30 μM). SR Ca²⁺ uptake was initiated by the addition of 500 nM Ca²⁺. (D) Average Ca²⁺ spark frequency and average SR Ca²⁺ leak rate, calculated from exponential time constants obtained as shown in (B). (E) Average time constants of the SR Ca²⁺ uptake recorded in control and HF myocytes. (F) Amplitudes of Ca²⁺ transients ([Ca²⁺]_c) recorded in control and HF field-stimulated myocytes are plotted against SR Ca²⁺ leak rates measured in cells isolated from the hearts with matched left ventricular fractional shortening. Data were fit to a logistic function with a leak rate of 2.42 ± 0.07 × 10⁻³ s⁻¹ corresponding to half-maximal changes in [Ca²⁺]_c amplitude. The grey area indicates the region (stability zone) where changes in SR Ca²⁺ leak rate are not associated with alterations in [Ca²⁺]_c amplitude. Each data point represents data collected from one to three hearts. *P < 0.05 vs. control; ^†P < 0.05 vs. 1 MHF; ^‡P < 0.05 vs. 4 MHF.

As seen in Figures 1 and 2, although the early increase in RyR2-mediated SR Ca²⁺ leak coincided with increased EC coupling efficiency, it preceded the decline in myocyte Ca²⁺ transient amplitude first observed after 4 MHF. To further emphasize the latter point, the initial relationship between the Ca²⁺ transient amplitude and SR Ca²⁺ leak rate is flat (i.e. ‘stability zone’) during which an increase in the leak rate is not associated with changes in Ca²⁺ transient amplitude (Figure 2F). In this region, the expected negative effect of SR Ca²⁺ leak on the Ca²⁺ transient is apparently compensated by enhanced EC coupling efficiency.

3.3. SCWs contribute to both arrhythmogenesis and contractile dysfunction in HF

The arrhythmogenic potential at early and late stages of HF (1 and +16 MHF) was examined by measuring SCWs and DADs in myocytes paced in the presence of the β-adrenergic agonist, isoproterenol (ISO), as catecholamines are known to be increased in this model of HF.¹⁹ An increase in the rate of occurrence of diastolic SCWs and DADs was observed at 1 MHF which tended to increase further at 16+ MHF (Figure 3A and B). Notably, the increase in the frequency of diastolic SCWs in ISO-stimulated HF myocytes was associated with a significant decrease in the amplitude of the systolic Ca²⁺ transient as early as 1 MHF when compared with control cells (Figure 3C). Since no such decrease in Ca²⁺ transient amplitude in 1 MHF myocytes was observed under baseline conditions when no SCW occurred, we hypothesized that the decrease in Ca²⁺ transients in HF myocytes in the presence of ISO resulted from diastolic Ca²⁺ release affecting the subsequent systolic Ca²⁺ release. Indeed, as shown in Figure 3D, the amplitude of the systolic, AP-induced Ca²⁺ transients was reduced by ∼60% when the AP-induced Ca²⁺ transients were preceded by a SCW, compared with the AP-induced Ca²⁺ transients that were not preceded by a SCW. The disrupting effect of irregular Ca²⁺ release on myocyte Ca²⁺ cycling and shortening was even more pronounced in myocytes at late stages of HF, in which cellular shortening often became effectively decoupled from pacing-induced electrical activity. An example of such severely disrupted myocyte Ca²⁺ cycling and rhythmic shortening in advanced HF is shown in Figure 3E.

Figure 3.

Diastolic Ca²⁺ waves induce membrane depolarization and affect myocyte contractility in HF. (A) Representative recordings of membrane potential with corresponding line-scan images and temporal profiles of Fluo-3 fluorescence recorded in control and HF myocytes at the indicated HF stages stimulated at 0.5 Hz in the presence of 100 nM ISO. Arrows indicate DADs. (B) Frequency of DADs were calculated in control (n = 12), in 1 MHF (n = 18), and 16+ MHF (n = 7) myocytes. (C) Average amplitudes of [Ca²⁺]_c recorded in control (n = 6) and in HF myocytes from 1 (n = 10) and 16+ (n = 7) month groups. (D) Amplitudes of [Ca²⁺]_c that were preceded with SCW in diastolic phase were normalized to those that did not display SCWs during the preceding diastolic interval. Data were recorded in HF myocytes (n = 9) as shown in (A). (E) Representative recordings of membrane potential, line-scan image and temporal profile of Fluo-3 fluorescence, and myocyte shortening (assessed from the line-scan image) obtained in an HF myocyte from the 16+ MHF group, stimulated at 1 Hz in the presence of 100 nM ISO. Part of the figure is scaled up to better illustrate the disrupting effect of SCWs (marked with white arrows) on systolic Ca²⁺ release and cellular shortening. Similar observations were made in 10% (one out of 10) myocytes from early-stage (1 month) HF and in 46% (five out of 11) myocytes from late-stage (two out of five myocytes from 4 MHF and three out of six myocytes from 16+ MHF) HF. None of control myocytes (n = 8) displayed such behaviour. Up–down arrows indicate the amplitude of Ca²⁺ transient and cellular shortening. *P < 0.05.

3.4. Enhanced SCW occurrence and reduced systolic Ca²⁺ release in HF are associated with decreased diastolic [Ca²⁺]_SR

To further explore the mechanism and role of diastolic SCW in HF myocytes, we performed simultaneous measurements of cytosolic and intra-SR Ca²⁺ levels using two different Ca²⁺ indicators (Rhod-2 and Fluo-5N, respectively). Potential effects upon Ca²⁺ handling due to HF-dependent changes in AP²⁰ were avoided by using a uniform AP clamp protocol with a ‘typical’ AP from a control cell as the voltage command. As shown in Figure 4A and B, the end-diastolic [Ca²⁺]_SR in HF myocytes was substantially lower than in control cells, consistent with increased SR Ca²⁺ leak in HF myocytes (Figure 2). Despite a lower level of diastolic [Ca²⁺]_SR in HF cells, SCWs occurred approximately six times more frequently in HF cells than in control myocytes (Figure 4C). Since spontaneous Ca²⁺ release and AP-induced Ca²⁺ release are likely to share the same Ca²⁺ pool, the occurrence of SCWs in the diastolic period could decrease the amplitude of the ensuing evoked Ca²⁺ transient through depletion of SR Ca²⁺. Consistent with this hypothesis, end-diastolic [Ca²⁺]_SR and systolic Ca²⁺ transient amplitude in HF myocytes were both decreased by ∼40% when systolic Ca²⁺ release following SCWs in the preceding diastolic interval was compared with that without SCWs in the preceding diastolic interval (Figure 4D). These results show that diastolic SCW in myocytes from failing hearts can decrease the amplitude of the systolic Ca²⁺ transients via depletion of SR Ca²⁺.

Figure 4.

Diastolic SCWs reduce end-diastolic SR Ca²⁺. (A) Representative traces of ‘typical’ control APs used as a voltage command and corresponding line-scan images and temporal profiles of Rhod-2 and Fluo-5N fluorescence recorded in control and HF myocytes. (B) Average values of end-diastolic [Ca²⁺]_SR [marked by red circles in (A)] recorded in control (n = 8) and HF (n = 9) myocytes. (C) Average frequency of SCW recorded in control (n = 12) and HF (n = 15) myocytes using the AP-clamp stimulation protocol. (D) Amplitudes of [Ca²⁺]_c and end-diastolic [Ca²⁺]_SR that were preceded with SCW in the diastolic phase were normalized to those that did not display SCW during the preceding diastolic interval. Data were recorded in HF myocytes (n = 5–9) as shown in (A). *P < 0.05.

3.5. Shortened Ca²⁺ signalling restitution contributes to increased propensity to SCWs in HF

We further studied the conditions required for the occurrence of spontaneous Ca²⁺ release in HF vs. control myocytes using a standard protocol of 20–25 control AP waveforms to load the SR with Ca²⁺. Under these conditions, SCWs regularly occurred not only in HF cells but also in control cells. However, in HF myocytes SCWs occurred following a substantially shorter time interval after cessation of the loading protocol than in control cells (Figure 5A and C). Since the occurrence of spontaneous Ca²⁺ release has been linked to the attainment of a certain ‘threshold’ SR Ca²⁺ load,²¹ one possible explanation for the shorter time to SCWs in HF could be a faster refilling of [Ca²⁺]_SR in HF than in control myocytes. This possibility was directly disproved by our [Ca²⁺]_SR measurements, indicating no substantial difference in [Ca²⁺]_SR recovery rates between HF and control myocytes (Figure 5D). Moreover, our data indicated that the attainment of a critical ‘threshold’ [Ca²⁺]_SR level was not sufficient to initiate the generation of SCWs. Indeed, the point at which [Ca²⁺]_SR reached a stable level and the onset of SCWs were always separated by a distinct time delay in both control and HF myocytes. This delay, or latency to SCWs, was significantly shorter in HF myocytes than in controls, whereas the baseline [Ca²⁺]_SR was again significantly reduced in HF cells (Figure 5B and E). While clearly showing that SCW occurrence is independent of the rate of [Ca²⁺]_SR recovery, our results do not conflict with the commonly accepted notion^21,22 that SCWs are influenced by the SR Ca²⁺ content. In fact, the observed decrease in [Ca²⁺]_SR in HF myocytes despite a higher predisposition to SCWs is consistent with the previously described increase in steady-state sensitivity of RyR2 to luminal Ca²⁺ in HF.^17,23 Therefore, to more accurately describe the behaviour of the SR Ca²⁺ store during SCWs in normal vs. HF myocytes, we introduced a refractoriness factor, R, that equals the product of latency to SCWs and the [Ca²⁺]_SR at which SCWs occur. In essence, R represents the efficacy of the Ca²⁺ release stabilization mechanisms which prevent SR Ca²⁺ release from self-activation at a given diastolic [Ca²⁺]_SR. In HF myocytes, R decreased by 3.6-fold (Figure 5F), suggesting that the intrinsic ability of the Ca²⁺ release mechanism to remain refractory following SR Ca²⁺ release is impaired, leading to the enhanced generation of SCWs.

Figure 5.

Ca²⁺ signalling refractoriness in control and HF myocytes. (A) Representative traces of ‘typical’ control APs used as voltage commands and corresponding line-scan images and temporal profiles of Rhod-2 and Fluo-5N fluorescence recorded in control and HF myocytes. (B) Average values of diastolic [Ca²⁺]_SR at the time of SCW initiation (threshold [Ca²⁺]_SR) recorded in control and HF myocytes. (C) Average time delay between maximal SR Ca²⁺ depletion during the last of the 20 AP-clamp stimuli and the onset of SCW recorded in control and HF myocytes. (D) Average rate of SR Ca²⁺ replenishing measured in control and HF myocytes. (E) Latency to SCWs (L) recorded in control and HF myocytes was measured as the time interval from the point when SR Ca²⁺ restored to 99% (five times exponential time constant) from depletion caused by the last stimulus to the point of SCW initiation. (F) Refractoriness factors (R = latency × threshold [Ca²⁺]_SR) were calculated for control and HF myocytes. Data presented in this figure were recorded in eight control and seven HF myocytes. *P < 0.05 vs. control.

3.6. Abnormal RyR2 function in HF involves both phosphorylative and redox modifications

Myocyte ROS production measured with DCFDA was not significantly altered at 1 MHF (Figure 6A and B). However, a progressive increase in ROS generation was detected at later stages of HF (at 4 and 16+ MHF). These changes in ROS production were paralleled by a similar increase in oxidation of RyR2s as determined by the mBB fluorescence method (Figure 6C and D).

Figure 6.

Oxidation and phosphorylation status of ryanodine receptors (RyR2s) in control and during HF progression. (A) Representative images of control and HF myocytes at the indicated HF stages loaded with an ROS-sensitive fluorescent indicator DCFDA. (B) Relative normalized DCFDA fluorescence from control myocytes (n = 47) and HF myocytes from 1 (n = 131), 4 (n = 24), and 16+ (n = 16) month groups. (C) Representative Coomassie-stained gels (upper panels) and mBB fluorescence intensity (lower panels) of RyR2s from control and HF hearts at 1 and 16 MHF measured under baseline conditions, in the presence of the oxidizing agent DTDP (0.2 mM), and in the presence of the reducing agent DTT (10 mM). (D) Relative free thiol content of RyR2s from control samples (n = 14) and samples from 1 (n = 9) and 16+ (n = 5) MHF groups. (E) Representative western blots showing phosphorylation of RyR2s at Ser-2808 (PKA-dependent) and Ser-2814 (CaMKII-dependent) phosphorylation sites in control and in 1 and 16 MHF measured with phosphor-specific antibodies. (F) Data pooled for Ser-2808 from six to 10 experiments and for Ser-2814 from four to nine experiments. (G) Representative line-scan images and temporal profiles of Rhod-2 fluorescence recorded in 1 MHF myocytes field-stimulated at 0.5 Hz in the presence of 100 nM ISO. Cells were pre-treated for at least 30 min with 1 µM KN 93, an inhibitor of CaMKII, or 1 µM KN 92, an inactive structural analogue of KN 93. Arrows indicate the time of electrical stimulation. (H) Average frequency of Ca²⁺ waves in 1 HF myocytes measured in the presence of 1 µM KN 92 (n= 22) and 1 µM KN 93 (n = 17). (I) Representative line-scan images and temporal profiles of Rhod-2 fluorescence recorded in 16 MHF myocytes field-stimulated at 0.3 Hz in the presence of 100 nM ISO. (J) Average frequency of Ca²⁺ waves in 16+ MHF myocytes was measured in the absence (n = 17) and in the presence (n = 11) of 1 µM KN 93. *P < 0.05 vs. control or KN 92 group; ^†P < 0.05 vs. 1 MHF; ^‡P < 0.05 vs. 4 MHF.

RyR2 phosphorylation at the putative PKA- and CaMKII-dependent sites 2808 and 2814, respectively, was assessed using phospho-specific antibodies. RyR2 phosphorylation at Ser-2814 increased approximately four-fold at 1 MHF with respect to control, and this effect was maintained at 16+ MHF (Figure 6E and F). In contrast, we observed no significant changes in RyR2 phosphorylation at Ser-2808 in either 1 or 16+ MHF group. Therefore, HF duration was associated with specific time-dependent post-translational modifications of RyR2s, i.e. early and sustained increases in RyR2 phosphorylation at Ser-2814 followed by oxidative modification of thiols. A critical role for CaMKII activity in arrhythmogenesis at the early stages of HF progression was confirmed in experiments in which pharmacological inhibition of the kinase with KN-93 abolished diastolic Ca²⁺ waves in field-stimulated HF myocytes at 1 MHF but not at 16 MHF (Figure 6G–J). In addition, the inclusion of CaMKII inhibitory peptide in the pipette solution significantly inhibited the frequency of SCW in 1 MHF patch-clamped myocytes (data not shown).

4. Discussion

4.1. Early enhancement of SR Ca²⁺ leak and sustained EC coupling functional stability during HF progression

Elevated RyR2-mediated SR Ca²⁺ leak is an important feature of myocytes from failing hearts at advanced stages of disease in both human and animal models,^13–16 which is thought to contribute to the pathophysiology of HF either directly (by weakening contractility through depletion of the SR Ca²⁺ stores and slowing relaxation via prolongation of the Ca²⁺ transient)¹³ or indirectly (by causing Ca²⁺-dependent pathological remodelling and cell death).²⁴ However, understanding the role of SR Ca²⁺ leak as a causal factor in HF has been limited by the lack of evidence as to whether alterations in leak precede, parallel, or follow the development and progression of HF. In the present study, we found that enhanced diastolic SR Ca²⁺ leak (Figure 2) was one of the earliest alterations of Ca²⁺ handling during HF development, which progressed in parallel with deterioration of in vivo contractile function. These results suggest that the level of SR Ca²⁺ leak is a sensitive indicator of HF and is consistent with a causal role of SR Ca²⁺ leak in HF development.

It has been argued that enhancing (or partially inhibiting) RyR2s can only produce temporary changes in SR Ca²⁺ release because of compensatory effects of altered SR Ca²⁺ content on SR Ca²⁺ release.^21,25 However, the significance of these self-correcting mechanisms in influencing SR Ca²⁺ cycling in RyR2-linked cardiac disease states has been uncertain as many of these conditions are characterized by persistently activated SR Ca²⁺ leak and reduced Ca²⁺ transients.^13,16,23 Our present study provides new insights into the relationships between altered RyR2 function, SR Ca²⁺ leak, and Ca²⁺ transients during HF progression. In particular, our data show that at early HF stages, enhanced mobilization of Ca²⁺ stores via sensitized RyR2s allows myocytes to maintain Ca²⁺ transients of nearly normal size, despite enhanced diastolic SR Ca²⁺ leak (Figure 2F). In addition, feedback influences of SR Ca²⁺ release on the Ca²⁺ current (via Ca²⁺-dependent inactivation) could contribute to stabilization of the Ca²⁺ transient in HF myocytes.²¹ At more advanced HF stages, however, when RyR2 dysfunction becomes more severe and the leak faster, the Ca²⁺ transients decrease in amplitude because of a massive loss of SR Ca²⁺ content.

4.2. Arrhythmogenesis and contractile impairment are intrinsically linked

Increased RyR2 activity has been shown to lead to arrhythmogenesis due to SCWs and DADs, particularly, when combined with β-adrenergic stimulation.^8,21 Accordingly, the increase in Ca²⁺ spark frequency was paralleled with an increase in propensity to Ca²⁺ waves and DADs in myocytes exposed to ISO at different stages of HF (Figure 3). Of note, the observed up-regulation of NCX would also be expected to contribute to increased arrhythmogenic propensity from early stages of HF by promoting DADs and their triggering extrasystolic APs.⁴ Besides leading to arrhythmias, SCWs have been suggested to weaken cardiac contractility by compromising systolic SR Ca²⁺ release either through depletion of SR Ca²⁺ or through inactivation of CICR. However, whether and how SCWs contribute to altered myocyte Ca²⁺ handling and contractility in HF has remained unclear. Our experiments provide direct evidence that SCWs can indeed decrease the amplitude of the systolic Ca²⁺ transient in HF myocytes (Figure 3D). Moreover, direct measurements of [Ca²⁺]_SR revealed that 40% of depression of systolic Ca²⁺ release is attributable to depletion of SR Ca²⁺ by SCWs in the preceding diastolic period (Figure 4D). SCW-dependent inhibition of systolic Ca²⁺ release was especially pronounced in myocytes at late stages of HF (Figure 3E) in which SCWs effectively resulted in cytosolic Ca²⁺ oscillating independently of the pacing-induced electrical activity in myocytes at advanced stages of HF. These results suggest a novel paradigm linking RyR2 dysfunction with both arrhythmogenesis and depressed contractility in HF. In this paradigm, spontaneous Ca²⁺ oscillations caused by altered RyR2 function, besides accounting for arrhythmogenesis, can lead to contractile failure through decoupling of Ca²⁺ release from pacing-induced electrical activity.

4.3. Ca²⁺ signalling refractoriness and spontaneous Ca²⁺ release

Spontaneous SR Ca²⁺ release that arises in the form of waves of self-propagating CICR has been commonly attributed to elevation of the SR Ca²⁺ content above a certain threshold level.^21,22 It has been suggested that in certain disease settings, including catecholaminergic polymorphic ventricular tachycardia and HF, this threshold is lowered due to sensitization of RyR2s.^26,27 Our direct measurements of [Ca²⁺]_SR revealed that myocytes indeed exhibited a certain critical [Ca²⁺]_SR level at which SCWs occurred and this [Ca²⁺]_SR was significantly lower in HF cells than in control cells (Figure 5A and B). However, importantly, SCWs did not arise immediately upon attaining the critical [Ca²⁺]_SR level as would be expected for a threshold-mediated mechanism.²² Instead, we observed a distinct time delay between the attainment of baseline [Ca²⁺]_SR and SCWs. This time delay, indicative of incomplete restitution of CICR (e.g. refractoriness of RyR2 channels), was markedly shorter in HF myocytes than in control.

On the basis of growing evidence,^26,28,29 RyR2 refractoriness is governed by luminal Ca²⁺ such that reduced [Ca²⁺]_SR results in deactivation of RyR2s, whereas elevation of [Ca²⁺]_SR restores RyR2 functional readiness. Our finding that the time delay to SCW was shortened in HF myocytes despite the reduced [Ca²⁺]_SR (Figure 5) suggests that the intrinsic ability of the Ca²⁺ release mechanism to become refractory is significantly impaired in HF myocytes. These results support the importance of store-dependent Ca²⁺ signalling in the control of SR Ca²⁺ release in normal myocytes^26,28,29 and show, for the first time, that the impairment of this mechanism contributes to arrhythmogenic disturbances of intracellular Ca²⁺ handling in HF. The concept of shortened store-dependent Ca²⁺ signalling refractoriness can also provide a framework for understanding the stimulatory role of β-adrenergic agonists on spontaneous Ca²⁺ release in HF myocytes. In this context, the stimulation of SCWs by ISO could be ascribed to acceleration of SERCA-mediated refilling of the SR Ca²⁺ stores and hence accelerated restitution from store-dependent deactivation. Additionally, phosphorylation of RyR2s either through PKA or CaMKII could further contribute to altered RyR2 function rendering them more responsive to luminal Ca²⁺.^14,16

4.4. Role of post-translational modification of RyR2 by phosphorylation and thiol oxidation

Altered RyR2 function in HF has been attributed to either hyperphosphorylation^14,16 or oxidation.¹⁷ The present study shows that both forms of post-translational modifications of RyR2 can be involved at different stages of HF: early and persistent CaMKII-mediated phosphorylation was followed by oxidation of thiols at later stages of HF (Figure 6). In agreement with these results, we recently showed using the same model that abnormal myocyte Ca²⁺ handling in late-stage HF is normalized by antioxidants.¹⁷ Future experiments will have to define the exact relationships between these two mechanisms during HF development.

4.5. Conclusions

We found that during long-term HF, a progressive loss of the ability of RyR2 activity to become refractory gives rise to an increase in arrhythmogenic potential and deterioration of systolic SR Ca²⁺ release and contractility in cardiac myocytes. These pathological changes in myocyte Ca²⁺ handling were linked to sequential modification of RyR2 channels by CaMKII phosphorylation and thiol oxidation. These results provide new insights into abnormal Ca²⁺ handling during HF and suggest molecular bases for temporally targeted mechanistically based therapeutic approaches for this disease.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Conflict of interest: none declared.

Funding

This work was supported by the American Heart Association (to A.E.B. and D.T.); National Institutes of Health grants (HL074045 and HL063043 to S.G. and HL089836 to C.A.C.). Pacemakers and pacing leads were provided by Medtronic Inc., Minneapolis, MN, USA.