Modulation of cytosolic and intra-sarcoplasmic reticulum calcium waves by calsequestrin in rat cardiac myocytes

Abstract

Waves of Ca²⁺-induced Ca²⁺ release occur in various cell types and are involved in the pathology of certain forms of cardiac arrhythmia. These arrhythmias include catecholaminergic polymorphic ventricular tachycardia (CPVT), certain cases of which are associated with mutations in the cardiac calsequestrin gene (CASQ2). To explore the mechanisms of Ca²⁺ wave generation and unravel the underlying causes of CPVT, we investigated the effects of adenoviral-mediated changes in CASQ2 protein levels on the properties of cytosolic and sarcoplasmic reticulum (SR) Ca²⁺ waves in permeabilized rat ventricular myocytes. The free [Ca²⁺] inside the sarcoplasmic reticulum ([Ca²⁺]_SR) was monitored by fluo-5N entrapped into the SR, and cytosolic Ca²⁺ was imaged using fluo-3. Overexpression of CASQ2 resulted in significant increases in the amplitude of Ca²⁺ waves and interwave intervals, whereas reduced CASQ2 levels caused drastic reductions in the amplitude and period of Ca²⁺ waves. CASQ2 abundance had no impact on resting diastolic [Ca²⁺]_SR or on the amplitude of the [Ca²⁺]_SR depletion signal during the Ca²⁺ wave. However, the recovery dynamics of [Ca²⁺]_SR following Ca²⁺ release were dramatically altered as the rate of [Ca²⁺]_SR recovery increased ∼3-fold in CASQ2-overexpressing myocytes and decreased to 30% of control in CASQ2-underexpressing myocytes. There was a direct linear relationship between Ca²⁺ wave period and the half-time of basal [Ca²⁺]_SR recovery following Ca²⁺ release. Loading the SR with the low affinity exogenous Ca²⁺ buffer citrate exerted effects quantitatively similar to those observed on overexpressing CASQ2. We conclude that free intra-SR [Ca²⁺] is a critical determinant of cardiac Ca²⁺ wave generation. Our data indicate that reduced intra-SR Ca²⁺ binding activity promotes the generation of Ca²⁺ waves by accelerating the dynamics of attaining a threshold free [Ca²⁺]_SR required for Ca²⁺ wave initiation, potentially accounting for arrythmogenesis in CPVT linked to mutations in CASQ2.

In cardiac muscle, the process of excitation–contraction (EC) coupling relies on Ca²⁺ influx from the extracellular milieu activating the SR Ca²⁺ release channels (ryanodine receptors, RyRs) in the sarcoplasmic reticulum (SR). This mechanism is known as Ca²⁺-induced Ca²⁺ release (CICR) (Fabiato, 1985). In addition to being controlled by cytosolic Ca²⁺, the RyR channels are modulated by Ca²⁺ levels inside the SR (Györke et al. 2002). The closure of Ca²⁺ release channels upon decline in intra-SR Ca²⁺ levels (luminal Ca²⁺-dependent deactivation) appears to account for the termination of CICR (Terentyev et al. 2002, 2003). Ca²⁺ released to the cytosol is taken back by the Sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA) to the SR where it is stored on the Ca²⁺-binding protein CASQ2 before it can be released again during the next activation cycle.

Due to its self-regenerating nature, CICR is an intrinsically unstable process. Under various pathological conditions, generally characterized by an increased SR Ca²⁺ content (i.e. ‘Ca²⁺ overload’), CICR can arise spontaneously, leading to self-propagating Ca²⁺ waves (Orchard et al. 1983; Wier et al. 1987; Lipp & Niggli, 1993; Cheng et al. 1996; Lukyanenko et al. 1996). Although it is clear that Ca²⁺ wave propagation involves diffusion-coupled CICR (Cheng et al. 1996; Lukyanenko et al. 1999; Izu et al. 2001), the precise roles of cytosolic and luminal Ca²⁺ in wave generation remain to be defined. The Ca²⁺ waves are arrythmogenic because they induce inward membrane currents that lead to delayed afterdepolarizations (DADs) and irregular action potentials (APs), accounting for the pathological mechanism of triggered arrhythmia (Kass & Tsien, 1982; Allen et al. 1984; Marban et al. 1986; Lakatta & Guarnieri, 1993).

Catecholaminergic polymorphic ventricular tachycardia is a form of inherited exercise-induced arrhythmia associated with mutations in CASQ2 and RyR (Coumel et al. 1978; Laitinen et al. 2001; Priori et al. 2001; Lahat et al. 2001; Postma et al. 2002). Recently, we showed that reduced CASQ2 levels or expression of a specific CPVT-associated CASQ2 mutant (D307H) leads to the generation of spontaneous Ca²⁺ transients and arrythmogenic DADs (i.e. cellular arrhythmia) in regularly paced cardiac myocytes (Terentyev et al. 2003; Viatchenko-Karpinski et al. 2004). We provided evidence to suggest that the generation of spontaneous, extrasystolic Ca²⁺ transients in these myocytes was caused by the premature recovery of the release channels from a luminal Ca²⁺-dependent refractory state. This refractory state is normally induced by the decline in [Ca²⁺]_SR following SR Ca²⁺ release (Terentyev et al. 2003). However, the specific molecular mechanisms by which genetic defects in CASQ2 lead to cellular arrhythmia remain to be elucidated.

In the present study we have directly examined the effects of altered CASQ2 levels on the properties of spontaneous Ca²⁺ waves monitored in the cytosolic and luminal compartments of myocytes permeabilized with saponin. Our results indicate that CASQ2 influences the frequency of spontaneous Ca²⁺ waves by determining the rate of recovery of [Ca²⁺]_SR after each release.

Methods

Ca²⁺ measurements in the cytosolic and SR luminal compartments

Single ventricular myocytes were obtained from adult male Sprague-Dawley rat hearts through enzymatic dissociation (Györke et al. 1997). Rats were anaesthetized with Nembutal and killed by exsanguination. The protocol for animal use was reviewed and approved by the Institutional Animal Care and Use Committee and complied with US Public Health Service policy on humane care and use of laboratory animals. The standard Tyrode solution contained (mm): 140 NaCl, 5.4 KCl, 0.5 MgCl₂, 1 CaCl₂, 10 Hepes, 0.25 NaH₂PO₄, 5.6 glucose, pH 7.3). The myocytes were permeabilized with saponin (0.01% for 45–60 s) in an ‘internal’ solution containing (mm): 120 potassium aspartate, 20 KCl, 0.81 MgCl₂, 20 Hepes, 3 MgATP, 0.5 EGTA, 0.114 CaCl₂ (free [Ca²⁺]∼100 nm), 10 phosphocreatine, 5 U ml⁻¹ creatine phosphokinase (Lukyanenko & Györke, 1999). Fluo-3 potassium salt (10 μm, TefLabs) was added to the internal solution for measurement of cytosolic [Ca²⁺] changes. The acetoxymethyl ester form of the low-affinity calcium indicator, fluo-5N (fluo-5N AM; Molecular Probes), entrapped inside the SR (Kabbara & Allen, 2001; Shannon et al. 2003) was used to assess changes in the intraluminal SR Ca²⁺ concentration. The myocytes were incubated with 10 μm fluo-5N AM and 0.05% pluronic detergent (Molecular Probes, OR, USA) for 6–8 h at 37°C to load the dye into the SR. After fluo-5N AM loading, the myocytes were permeabilized to remove the dye from the cytoplasm. Calcium waves were initiated by lowering the Ca²⁺ buffering capacity in the internal solution as a result of decreasing the concentration of EGTA from 0.5 mm to 0.1 mm (Lukyanenko & Györke, 1999) at a constant [Ca²⁺] of 100 nm. Ca²⁺ levels in the internal solutions were calculated using a computer program (WinMAXC 1.80, Stanford University, CA, USA) and verified by Ca²⁺ electrode measurements using standards from Calbiochem. In experiments involving the addition of 10 mm caffeine, the fluo-5N traces were corrected for ∼10% fluorescence quenching by this agent (determined at 100 μm Ca²⁺). In some experiments, potassium salt of citrate was added to the internal solution at a final concentration of 20 mm by replacing osmotically equivalent amounts of potassium aspartate. Myocytes were imaged with a Laser Scanning Confocal System (Bio-Rad MRC-1024ES interfaced to an Olympus IX-70 inverted microscope equipped with an Olympus ×60, 1.4 NA oil objective). Fluo-3 and fluo-5N were excited by light at 488 nm, and fluorescence was recorded at wavelengths > 515 nm in either the x–y or x–t (line scan) mode of the confocal system at a rate of 6 ms per line. All line-scan images were obtained by scanning the myocytes in the longitudinal direction. Values of relative fluorescence from line-scan images were obtained as: (F(x,t) −F_min)/(F_baseline−F_min), where F(x,t) is fluorescence at the linear coordinate x at time t, F_min is minimal fluorescence during SR depletion, and F_baseline is baseline fluorescence (Cheng et al. 1999; Lukyanenko et al. 2001). Image processing and analysis was performed using IDL software (Research Systems Inc., Boulder, CO, USA).

Adenoviral gene transfer

Recombinant adenoviruses containing the full-length CASQ2 coding sequence in the sense and antisense orientation (Ad-CASQ2 and Ad-CASQ2as, respectively) and an adenovirus containing a truncated CASQ2 coding region (1–70, Ad-control) were constructed as previously described (Terentyev et al. 2003). Isolated ventricular myocytes were plated on laminin-coated glass coverslips in serum-free medium 199 containing 25 mm NaHCO₃, 5 mm creatine, 5 mm taurine, 10 units ml⁻¹ penicillin, 10 μg ml⁻¹ streptomycin, 10 μm gentamicin and incubated at 5% CO₂–95% air at 37°C (Terentyev et al. 2003). After approximately 1 h of incubation, the medium was aspirated along with unattached myocytes and replaced with the same medium containing a particular type of adenovirus at a multiplicity of infection of 100. The myocytes were incubated with adenoviruses for 48–56 h and CASQ2 protein levels were determined by immunoblot analysis using antibodies specific for CASQ2 (1: 2500; PA-913, Affinity Bioreagents, Golden, CO, USA) (Terentyev et al. 2003).

Statistics

The results are presented as mean ±s.e.m. Statistical differences were tested using ANOVA (P < 0.05).

Results

Effects of CASQ2 levels on properties of cytosolic Ca²⁺ waves

We first investigated the effects of increased and decreased CASQ2 protein levels on the properties of cytosolically measured spontaneous Ca²⁺ waves in permeabilized myocytes. As in our previous studies (Terentyev et al. 2003; Viatchenko-Karpinski et al. 2004), we used adenoviral constructs containing the canine CASQ2 coding region in the sense (Ad-CASQ2) or antisense orientations (Ad-CASQ2as) to either increase or decrease CASQ2 expression in cardiac myocytes, respectively. Myocytes infected with a viral construct (Ad-control) containing a truncated CASQ2 coding region served as a control for the effects of viral infection (Terentyev et al. 2003). CASQ2 levels were assessed by Western blot analysis using a CASQ2-specific antibody. Consistent with our previous report (Terentyev et al. 2003), infection of myocytes with Ad-CASQ2 resulted in ∼3.5-fold increases in CASQ2 abundance, whereas infection with the Ad-CASQ2as viral construct resulted in a decrease in CASQ2 expression to 30% of control at ∼48 h after infection (not shown). As we have shown previously (Terentyev et al. 2003), these acute changes in CASQ2 abundance are not accompanied by the altered expression of other major SR Ca²⁺ handling proteins, including SERCA and phospholamban.

To promote the generation of spontaneous Ca²⁺ waves, myocytes permeabilized with saponin (Lukyanenko & Györke, 1999) were exposed to an ‘internal’ solution with low buffering strength (0.1 mm EGTA; 10 μm fluo-3, pCa 7). Under these conditions, control myocytes generated Ca²⁺ waves with a frequency of about 12 per minute. Representative line-scan images of Ca²⁺ waves along with the corresponding averaged fluo-3 fluorescence profiles measured in control, CASQ2-overexpressing and CASQ2-underexpressing myocytes are shown in Fig. 1A. The amplitude of Ca²⁺ waves and the length of interwave intervals (period) were greater in myocytes with increased CASQ2 levels (Ad-CASQ2), whereas Ca²⁺ wave amplitude and period were decreased in myocytes with reduced CASQ2 (Ad-CASQ2as) levels (Fig. 1B and C, respectively). At the same time, alterations in CASQ2 abundance did not affect the propagation velocity of Ca²⁺ waves (Fig. 1D). The relationship between the amplitude and period of Ca²⁺ waves is shown in Fig. 7A. The plot indicates that there is a direct linear correlation between wave amplitude and period. Since the amplitude of the spontaneous Ca²⁺ transients is a rough measure of the SR Ca²⁺-buffering capacity and the rate of free intra-SR [Ca²⁺] recovery after each release should be inversely proportional to the concentration of Ca²⁺-binding sites within the SR, these results suggest that Ca²⁺ wave generation is influenced by free [Ca²⁺] inside the SR luminal compartment.

Figure 1. Effects of increased and reduced CASQ2 levels on properties of cytosolic Ca²⁺ waves in permeabilized myocytes.

A, representative line-scan images of fluo-3 fluorescence and time-dependent profiles of spontaneous cytosolic Ca²⁺ waves (from the region indicated by the black bar on the right) acquired in permeabilized myocytes infected with Ad-control, Ad-CASQ2 and Ad-CASQ2as vectors (as indicated). B, C and D, pooled data on changes in the amplitude (B), period (C) and velocity (D) of cytosolic Ca²⁺ waves in myocytes infected with Ad-control, Ad-CASQ2 and Ad-CASQ2as vectors. The data are presented as means ±s.e.m., n = 4–6; * significantly different from control at P < 0.05.

Figure 7. Dependency of Ca²⁺ wave frequency on [Ca²⁺]_SR recovery dynamics.

A, Ca²⁺ wave period as a function of wave amplitude for different levels of CASQ2 and citrate present in the SR compartment. B, Ca²⁺ wave period as a function of basal [Ca²⁺]_SR recovery half-time following release for different levels of CASQ2 and citrate present in the SR compartment. The data were normalized with respect to control.

Effects of CASQ2 levels on SR luminal Ca²⁺ waves

The free intra-SR [Ca²⁺] was monitored in the same three groups of myocytes (i.e. Ad-control, Ad-CASQ2 and Ad-CASQ2as) with the low affinity Ca²⁺ indicator fluo-5N entrapped in the SR (Kabbara & Allen, 2001; Shannon et al. 2003). Loading the myocytes with fluo-5N resulted in a characteristic cross-striated staining pattern at about 2 μm that was largely eliminated by emptying the SR Ca²⁺ stores with either ryanodine (15 μm) and thapsigargin (10 μm) (Fig. 2A) or caffeine (10 mm) (not shown). In addition, line-scan image recordings from fluo-5N-stained myocytes revealed that the inhibition of Ca²⁺ leakage from the SR by tetracaine (0.2 mm), an inhibitor of RyR2, led to an increase in intra-SR fluo-5N fluorescence (8 ± 1%, n = 3). Conversely, discharging the SR Ca²⁺ stores by caffeine (10 mm) application reduced the fluorescence signal (20 ± 1%; n = 4). Representative examples of these experiments are shown in Fig. 2B. The line-scan images at the top were obtained at different times before and after exposing the cell to 0.2 mm tetracaine and 10 mm caffeine; the graph below shows the corresponding changes of average fluorescence in the course of the experiment. These results indicate that fluo-5N can be loaded into the SR to monitor luminal Ca²⁺ in permeabilized myocytes. The increase of the intra-SR fluorescence signal in response to tetracaine suggests that fluo-5N was not saturated in our experiments. These data also confirm that Ca²⁺ leakage through RyRs plays an important role in establishing the Ca²⁺ gradient across the SR (Lukyanenko et al. 2001).

Figure 2. Fluorescence signals from intra-SR fluo-5N in permeabilized myocytes.

A, images of a fluo-5N-loaded myocyte before and 10 min after application of 15 μm ryanodine and 10 μm thapsigargin (upper panels). Enlargement of the area indicated by the rectangular boxes (lower panels). B, line-scan images of fluo-5N fluorescence from a myocyte before (left) and after sequential additions of 0.2 mm tetracaine (middle) and 10 mm caffeine (right). The lower panel plots averaged fluo-5N fluorescence in the course of the experiment. C, representative line-scan images of fluo-5N fluorescence and time-dependent profiles of an intra-SR Ca²⁺ wave (left) and caffeine-induced intra-SR Ca²⁺ transient (right). The time-dependent profiles were taken from the region indicated by black bars on the right.

Spontaneous Ca²⁺ waves were marked by a transient decrease in the spatially resolved fluo-5N fluorescence along the line scan and of the averaged fluorescence signal (Fig. 2C, left). To assess the portion of total SR Ca²⁺ released by the wave, the SR Ca²⁺ stores were emptied by caffeine application. The amplitude of the luminal Ca²⁺ depletion signal was about 80% (n = 20) of that induced by caffeine (Fig. 2C), consistent with the notion that regenerative Ca²⁺ waves result in the liberation of nearly all of the Ca²⁺ available for release from the SR.

The effects of CASQ2 protein abundance on the properties of SR luminal Ca²⁺ waves are illustrated in Fig. 3A by a series of line-scan image recordings and corresponding averaged fluo-5N fluorescence signals from a representative control, CASQ2-overexpressing and CASQ2-underexpressing myocytes. CASQ2 abundance had no impact on basal (‘diastolic’) [Ca²⁺]_SR or on the amplitude of the maximum [Ca²⁺]_SR depletion signal during the wave (Fig. 4A and B, respectively). It is important to consider that the steady-state free [Ca²⁺] in a closed compartment such as the SR is determined by the transport mechanisms in the enclosing membrane. Therefore, Ca²⁺ binding to interluminal binding sites may affect the kinetics of luminal [Ca²⁺] changes but not the final steady-state [Ca²⁺] which is set by a balance between the Ca²⁺ uptake by SERCA and Ca²⁺ leak through the RyR channels. Indeed, the dynamics of [Ca²⁺]_SR recovery following release were dramatically altered by changes in CASQ2 protein levels. The SR Ca²⁺ wave signals at each pixel along the line-scan were aligned at the peak and averaged to assess the quantitative effects of alterations in CASQ2 protein levels on Ca²⁺ dynamics inside the SR during Ca²⁺ release and the subsequent refilling of the SR. Whilst the rate of [Ca²⁺]_SR decline was similar for all conditions (Figs 3C and 4C), the half-time of recovery of basal [Ca²⁺]_SR increased ∼3-fold in CASQ2-overexpressing cells and decreased to ∼25% of control levels in CASQ2-underexpressing cells (Figs 3B and 4D). These changes in intra-SR [Ca²⁺] recovery kinetics were in agreement with the observed changes in the period of cytosolically measured Ca²⁺ waves (Fig. 1A and C). Our SR luminal Ca²⁺ measurements suggest that initiation of spontaneous Ca²⁺ waves always occurs at the same free intra-SR [Ca²⁺] regardless of the amount of total Ca²⁺ in the SR, and thus provide evidence for the role of RyR luminal Ca²⁺ modulation in the initiation of Ca²⁺ waves.

Figure 3. Effects of increased and reduced CASQ2 levels on properties of intra-SR Ca²⁺ waves in permeabilized myocytes.

A, representative line-scan images of fluo-5N fluorescence and time-dependent profiles of intra-SR Ca²⁺ waves acquired in permeabilized myocytes infected with Ad-control, Ad-CASQ2 and Ad-CASQ2as vectors (as indicated). The time-dependent fluorescence profiles were taken from the region indicated by the black bars on the right. B, superimposed relative fluo-5N fluorescence signals acquired during and following SR Ca²⁺ release in myocytes infected with Ad-control, Ad-CASQ2 and A-CASQ2as vectors (as indicated). The recovery phases of the relative fluo-5N fluorescence signals were fitted with single exponential functions C, expansion of the descending phase of the relative fluorescence signals.

Figure 4. Spatio-temporal properties of intra-SR Ca²⁺ waves for different levels of CASQ2 in the SR compartment.

Pooled data on changes in the baseline (‘diastolic’) [Ca²⁺]_SR (A), amplitude of [Ca²⁺]_SR reduction (B), time to peak [Ca²⁺]_SR reduction (C) and half-time of [Ca²⁺]_SR recovery (D) in myocytes infected with Ad-control, Ad-CASQ2 and Ad-CASQ2as vectors. The data are presented as means ±s.e.m., n = 4–6; * significantly different from control at P < 0.05.

Effects of SR luminal citrate on properties of cytosolic-and luminal Ca²⁺ waves

In addition to influencing the SR Ca²⁺ release channels by modulating the free [Ca²⁺]_SR near their luminal side, CASQ2 may also affect the Ca²⁺ release mechanism more directly through protein–protein interactions with RyR, triadin and junctin (Zhang et al. 1997; Györke et al. 2004). To distinguish between these possibilities, we explored the effects of loading the SR with citrate, an exogenous low-affinity Ca²⁺ buffer (Terentyev et al. 2002). We reasoned that if the effects of CASQ2 were entirely due to its action as a buffer, loading the SR with citrate would have the same effect as increasing CASQ2 protein levels. Permeabilized myocytes were incubated in a citrate-containing internal solution for 20 min as previously described (Terentyev et al. 2002), after which citrate was removed from the cytosol, effectively leaving this Ca²⁺ chelator entrapped in the SR. Representative line-scan images of cytosolic and intra-SR Ca²⁺ waves are shown in Fig. 5A and B, respectively. Loading the SR with citrate resulted in increased amplitude and period of cytosolic Ca²⁺ waves (Fig. 6A and B, respectively) in a manner similar to that observed in CASQ2-overexpressing myocytes. Additionally, the recovery dynamics of [Ca²⁺]_SR were slowed in the presence of luminal citrate (Figs 5C and 6F), an effect similar to that seen in myocytes overexpressing CASQ2. Furthermore, in agreement with the effects of increased CASQ2 expression, luminal citrate did not alter the propagation velocity of Ca²⁺ waves (Fig. 5E). Citrate loading also did not change the rate of [Ca²⁺]_SR decline (Figs 5D and 6E) or the resting [Ca²⁺]_SR (Fig. 6C) or maximum [Ca²⁺]_SR depletion signal (Fig. 6D). Examining the relationship between the amplitude and period of cytosolic Ca²⁺ waves for control, citrate-loaded, as well as CASQ2 overexpressing and underexpressing myocytes shows that all data points fall on a single straight line (Fig. 7A). Similarly, there was a direct linear relationship between the half-time of basal [Ca²⁺]_SR recovery and Ca²⁺ wave period measured in the same four groups of cells (Fig. 7B). These results suggest that the effects of CASQ2 levels on Ca²⁺ wave properties could be attributed to its action as a Ca²⁺ buffer.

Figure 5. Effects of SR intraluminal citrate loading on properties of cytosolic and luminal Ca²⁺ waves.

A, representative line-scan images of fluo-3 fluorescence and time-dependent profiles of cytosolic Ca²⁺ waves acquired in permeabilized myocytes before and after loading their SR with 20 mm citrate (as indicated). B, representative line-scan images of fluo-5N fluorescence and time-dependent profiles of intra-SR Ca²⁺ waves acquired in permeabilized myocytes before and after SR loading with 20 mm citrate (as indicated). The time-dependent profiles were taken from the regions indicated by black bars on the right. C, superimposed relative fluo-5N fluorescence signals acquired during and following SR Ca²⁺ release before (red) and after (blue) SR loading with 20 mm citrate in myocytes infected with Ad-control, Ad-CASQ2 and Ad-CASQ2as vectors (as indicated). The recovery phases of the relative fluo-5N fluorescence signals were fitted with single exponential functions. D, expansion of the descending phase of the relative fluorescence signals. E, pooled data on changes in the velocity of Ca²⁺ waves.

Figure 6. Effects of luminal citrate on spatio-temporal properties of cytosolic and luminal Ca²⁺ waves.

A and B, pooled data on changes in the amplitude (A) and period (B) of cytosolic Ca²⁺ waves measured with fluo-3. C and D, pooled data on changes in the baseline [Ca²⁺]_SR (C) and amplitude (D) of the [Ca²⁺]_SR depletion signals measured with SR-entrapped fluo-5N. E and F, pooled data on changes in time to peak reduction (F) and half-time of recovery of [Ca²⁺]_SR (E) during and following Ca²⁺ wave-induced Ca²⁺ release in myocytes before and after SR loading with 20 mm citrate. The data are presented as means ±s.e.m., n = 3–4; * significantly different from control at P < 0.05.

Discussion

Intracellular Ca²⁺ waves produced by the release of Ca²⁺ from the SR have been implicated in various cardiac arrhythmias, including effort-induced tachycardia linked to mutations in CASQ2 (Priori & Corr, 1990; Lakatta & Guarnieri, 1993; Leenhardt et al. 1995; Terentyev et al. 2003; Viatchenko-Karpinski et al. 2004). Ca²⁺ waves are generally associated with an increased SR Ca²⁺ content. Therefore, defining the role of free luminal Ca²⁺ in the generation of Ca²⁺ waves is of considerable interest. In the present study, we directly measured free intra-SR Ca²⁺ in rat ventricular myocytes expressing different levels of CASQ2 protein to explore the underlying mechanisms of Ca²⁺ wave generation.

Role of luminal Ca²⁺ in Ca²⁺ wave generation

In principle, increased SR Ca²⁺ content (Ca²⁺ overload) could result in the generation of spontaneous Ca²⁺ waves through two different mechanisms. One possibility is that increased amounts of releasable Ca²⁺ stored in the SR stimulate the process of CICR, which is also involved in normal EC coupling, by increasing the positive feedback of released Ca²⁺ on the Ca²⁺-activated RyR channels (Stern et al. 1988). The other possibility is that elevated free [Ca²⁺] in the SR compartment enhances the functional activity of the Ca²⁺ release channels by acting at RyR-associated luminal Ca²⁺-sensing sites (Lukyanenko et al. 1999). Our direct measurements of [Ca²⁺]_SR indicate that spontaneous Ca²⁺ wave generation in cardiac myocytes consistently occurred at the same threshold free intra-SR [Ca²⁺] regardless of the amount of total Ca²⁺ available for release inside the SR and released to the cytosol. Enhancing the intra-SR Ca²⁺ buffering capacity through either overexpressing CASQ2 or loading the SR with citrate delayed the functional restitution of release sites after each release (as assessed from changes in Ca²⁺ wave period) by slowing the recovery of free intra-SR [Ca²⁺]. In contrast, reducing the intra-SR Ca²⁺-buffering capacity by decreasing CASQ2 protein levels accelerated the functional restitution or readiness of release sites by accelerating the recovery of [Ca²⁺]_SR. Thus, our results suggest that the modulation of RyR channel activity by luminal Ca²⁺ is a critical factor in the initiation of spontaneous cardiac Ca²⁺ waves.

Interestingly, CASQ2 abundance and SR luminal citrate had no discernable impact on the propagation velocity of Ca²⁺ waves despite the drastic changes in Ca²⁺ wave amplitude. Increased amounts of Ca²⁺ released to the cytosol would be expected to enhance Ca²⁺ wave propagation via classical CICR. Why was this not observed? One possibility is that Ca²⁺ wave propagation involves mechanisms other than CICR. One such model postulates that Ca²⁺ release occurs when a heavily Ca²⁺-loaded SR region reaches its limit, then extra Ca²⁺ uptake in neighbouring SR regions contributes to activation in the adjacent regions by raising [Ca²⁺]_SR. If Ca²⁺ wave propagation was due to this kind of local SR Ca²⁺ overload spreading ahead of the wave front, then increased luminal Ca²⁺ buffering should slow Ca²⁺ wave propagation by delaying the attainment of a threshold [Ca²⁺]_SR required for release activation. However, our direct measurements of [Ca²⁺]_SR did not reveal a local elevation of [Ca²⁺]_SR ahead of the wave front predicted by this model. Additionally, we have shown previously that Ca²⁺ wave propagation is not reduced but rather is enhanced when SERCA is inhibited by thapsigargin in myocytes whose SR was preloaded with Ca²⁺ (Lukyanenko et al. 1999). Thus this explanation does not seem likely. Another possibility relates to the observation that CASQ2 modulates both focal and global Ca²⁺ signals by changing the duration of their rising phases without affecting the rate of raise of these signals (Terentyev et al. 2003). These results were attributed to the ability of CASQ2 to influence Ca²⁺ release termination by buffering local free [Ca²⁺] in the vicinity of the RyR luminal Ca²⁺-sensing sites (Terentyev et al. 2003). Because local cytosolic [Ca²⁺] gradients tend to dissipate rapidly due to diffusion, CICR should be more sensitive to the rate of SR Ca²⁺ release (Ca²⁺ flux intensity) than to the overall amounts of Ca²⁺ released. This could account for the lack of effects of CASQ2 levels on Ca²⁺ wave propagation velocity. More studies, both experimental and theoretical, are needed to clearly elucidate the effects of SR luminal Ca²⁺ buffers on Ca²⁺ waves.

Molecular mechanisms of arrhythmia linked to CASQ2 mutations

Recently, we showed that reduced CASQ2 levels (or expression of a specific CPVT-related CASQ2 mutant) led to the generation of spontaneous Ca²⁺ transients and arrhythmogenic DADs in paced cardiac myocytes (Terentyev et al. 2003; Viatchenko-Karpinski et al. 2004). We hypothesized that the generation of spontaneous Ca²⁺ waves in these myocytes is due to a premature recovery of the release channels from a luminal Ca²⁺-dependent refractory state. In normal myocytes, the release channels become refractory when the Ca²⁺ level inside the SR declines as a consequence of release and stay refractory until the SR is refilled by the SERCA pump (Terentyev et al. 2002). The recovery of [Ca²⁺]_SR will be faster when the concentration of luminal Ca²⁺ binding sites is reduced, potentially resulting in premature functional recovery of the RyR channels from refractoriness in myocytes underexpressing CASQ2. In the present study we directly demonstrate that reducing CASQ2 both increased spontaneous Ca²⁺ wave frequency and accelerated the rate of attaining diastolic [Ca²⁺]_SR. Therefore these results provide compelling evidence for the proposed role of abnormal luminal [Ca²⁺] dynamics in the aetiology of CPVT.

Molecular mechanisms of CASQ2 effects

Although the precise nature of luminal Ca²⁺-sensing sites remains to be determined, it has been shown that the luminal proteins CASQ2, triadin and junctin physically associate with RyR2 in cardiac myocytes and modulate its functional activity (Zhang et al. 1997; Györke et al. 2004). Based on these studies, we proposed that CASQ2 confers on RyR the ability to sense luminal [Ca²⁺] by either dissociating from the RyR complex or by directly regulating the activity of RyRs in response to changes in [Ca²⁺] (Györke et al. 2004). In the present study, we demonstrated that the effects of increasing CASQ2 protein levels were similar to those observed in myocytes in which the SR was loaded with the exogenous low affinity Ca²⁺ chelator, citrate. These results therefore suggest that the role of CASQ2 in Ca²⁺ handling in cardiac myocytes can be attributed to its action as a Ca²⁺ buffer in the SR lumen. To reconcile this apparent discrepancy it is important to consider that CASQ2 is expressed in the SR at much higher levels than RyR, triadin or junctin (Zhang et al. 1997). Consequently, lowering CASQ2 abundance to 30% of control levels may be sufficient to observe significant differences in its buffering function, but may leave sufficient CASQ2 in the SR to preserve functions based on interactions between CASQ2 with RyR (and possibly triadin and junctin). Therefore, these results do not rule out the possibility that CASQ2 regulates RyR activity by functioning as its luminal Ca²⁺ sensor. Experiments that achieve more complete depletion of CASQ2 (e.g. by using RNA interference approaches) could clarify this issue.

Conclusions

In summary, our results show that the level of free Ca²⁺ inside the SR is a critical factor in the generation of spontaneous Ca²⁺ waves in cardiac myocytes. Ca²⁺ waves arise only when [Ca²⁺]_SR reaches a certain threshold level that sensitizes the RyR channel to CICR. This model would therefore predict that mutations in CASQ2 that act either by reducing CASQ2 expression or by impairing the Ca²⁺-binding activity of CASQ2 promote the generation of Ca²⁺ waves by accelerating the dynamics of attaining this threshold free [Ca²⁺]_SR, thereby providing a mechanism for arrhythmia in CPVT.