Regulation of sarcoplasmic reticulum Ca²⁺ leak by cytosolic Ca²⁺ in rabbit ventricular myocytes

Abstract

Non-technical summary

Contraction and relaxation of the heart strongly depend on calcium (Ca²⁺) stored in the sarcoplasmic reticulum (SR). Ca²⁺ stored within the SR is determined by the balance between Ca²⁺ uptake and Ca²⁺ leak that occurs mainly via Ca²⁺ release channels, called ryanodine receptors (RyRs). Alterations in the RyR activity can lead to enhanced SR Ca²⁺ leak and arrhythmias. Ca²⁺ tightly regulates the RyR activity from both sides of the SR (cytosolic and luminal). In this work, we studied the effects of cytosolic Ca²⁺ on SR Ca²⁺ leak in isolated ventricular myocytes. Elevation of cytosolic Ca²⁺ increased SR Ca²⁺ leak by a direct activation of RyRs. In intact myocytes, at the end of contraction and at the beginning of the relaxation phase, SR Ca²⁺ leak remains relatively constant due to the coordinated regulation of RyRs by cytosolic and luminal Ca²⁺. Thus, this dual regulation of the RyR contributes to the control of the SR Ca²⁺ content, preventing excessive loss of Ca²⁺ that could lead to pathological conditions such as cardiac arrhythmia.

Abstract

Sarcoplasmic reticulum (SR) Ca²⁺ leak determines SR Ca²⁺ content and, therefore, the amplitude of global Ca²⁺ transients in ventricular myocytes. However, it remains unresolved to what extent Ca²⁺ leak can be modulated by cytosolic [Ca²⁺] ([Ca²⁺]_i). Here, we studied the effects of [Ca²⁺]_i on SR Ca²⁺ leak in permeabilized rabbit ventricular myocytes. Using confocal microscopy we monitored SR Ca²⁺ leak as the change in [Ca²⁺]_SR (with Fluo-5N) after complete SERCA inhibition with thapsigargin (10 μm). Increasing [Ca²⁺]_i from 150 to 250 nm significantly increased SR Ca²⁺ leak over the entire range of [Ca²⁺]_SR. This increase was associated with an augmentation of both Ca²⁺ spark- and non-spark-mediated Ca²⁺ leak. Further increasing [Ca²⁺]_i to 350 nm led to rapid [Ca]²⁺]_SR depletion due to the occurrence of Ca²⁺ waves. The augmentation of SR Ca²⁺ leak by high [Ca²⁺]_i was insensitive to inhibition of Ca²⁺–calmodulin-dependent protein kinase II. In contrast, lowering [Ca²⁺]_i to 50 nm markedly decreased SR Ca²⁺ leak rate and nearly abolished Ca²⁺ sparks. When the ryanodine receptor (RyR) was completely inhibited with ruthenium red (50 μm), changes in [Ca²⁺]_i between 50 and 350 nm did not produce any significant effect on SR Ca²⁺ leak, indicating that [Ca²⁺]_i alters SR Ca²⁺ leak solely by regulating RyR activity. In summary, [Ca²⁺]_i in the range of 50–350 nm has a significant effect on SR Ca²⁺ leak rate mainly via direct regulation of RyR activity. As RyR activity depends highly on [Ca²⁺]_i and [Ca²⁺]_SR, SR Ca²⁺ leak remains relatively constant during the declining phase of the Ca²⁺ transient when [Ca²⁺]_SR and [Ca²⁺]_i change in opposite directions.

Introduction

During systole, Ca²⁺ influx via L-type Ca²⁺ channels activates ryanodine receptors (RyRs) causing global Ca²⁺ release that initiates contraction in cardiac muscle. This process is known as Ca²⁺-induced Ca²⁺ release (CICR; Fabiato, 1983). After CICR termination, a large portion of cytosolic Ca²⁺ is pumped back into the sarcoplasmic reticulum (SR) by the Ca²⁺-ATPase (SERCA) leading to cardiac muscle relaxation. However, RyRs are not continuously closed during diastole and spontaneous openings of RyRs generate substantial SR Ca²⁺ leak. By counterbalancing SR Ca²⁺ uptake, diastolic SR Ca²⁺ leak plays an important role in setting and maintaining the appropriate SR Ca²⁺ load in the healthy heart (Shannon et al. 2002; Zima et al. 2010). However, during the development of heart failure (HF) RyRs undergo post-translational modifications that ultimately lead to an increase in SR Ca²⁺ leak (Ai et al. 2005; Belevych et al. 2007; Zima et al. 2010). This excessive Ca²⁺ leak has been implicated in reduction of contractile force as well as in cardiac arrhythmias (George, 2008). Although the precise mechanisms of RyR modification during HF remain highly controversial (Bers et al. 2003; Lehnart & Marks, 2007), it has been shown that abnormal phosphorylation of RyR either by protein kinase A (PKA; Marx et al. 2000) or Ca²⁺–calmodulin-dependent kinase type II (CaMKII; Ai et al. 2005) is involved in the augmentation of SR Ca²⁺ leak. Additionally, oxidation of thiol groups of the RyR can contribute to an alteration of SR Ca²⁺ leak in HF (Terentyev et al. 2008). Failing cardiac myocytes also exhibit a slower decay of cytosolic Ca²⁺ transient during excitation–contraction coupling (ECC) (O'Rourke et al. 1999; Jiang et al. 2002) due, presumably, to down-regulation of SERCA activity and a subsequent impairment of SR Ca²⁺ reuptake (Pieske et al. 1995; Pogwizd et al. 1999). Additionally, it has been reported that in failing myocytes cytosolic [Na⁺] is significantly increased. This can contribute to further slowing of Ca²⁺ extrusion by Na⁺–Ca²⁺ exchange (NCX) (Despa et al. 2002) leading to an increase in cytosolic [Ca²⁺] ([Ca²⁺]_i). High diastolic [Ca²⁺]_i by itself can increase RyR activity and SR Ca²⁺ leak; however, the modulation of SR Ca²⁺ leak by [Ca²⁺]_i has not been characterized in detail.

The RyR (type 2) mediates most of diastolic SR Ca²⁺ leak in ventricular myocytes (Neary et al. 2002; Shannon et al. 2002; Zima et al. 2010). The majority of RyRs localize in the junctional SR where they form clusters of 10–200 channels (Franzini-Armstrong et al. 1999). Each of these subcellular microdomains constitutes an SR Ca²⁺ release unit (Cheng & Lederer, 2008). The simultaneous activation of RyRs within the release unit generates a locally restricted increase in [Ca²⁺]_i, or Ca²⁺ spark (Cheng et al. 1993; Lopez-Lopez et al. 1995). After the discovery of Ca²⁺ sparks, it has been suggested that the entire diastolic SR Ca²⁺ leak can be explained solely by these spontaneous release events (Cheng et al. 1993; Bassani & Bers, 1995). However, this concept has been recently challenged. It has been reported that in ventricular myocytes a significant portion of SR Ca²⁺ leak occurs as undetectable openings of single RyRs or spark-independent Ca²⁺ leak (Santiago et al. 2010; Zima et al. 2010; Brochet et al. 2011; Porta et al. 2011). RyR-mediated SR Ca²⁺ leak strongly depends on intra-SR free Ca²⁺ ([Ca²⁺]_SR). At low [Ca²⁺]_SR, SR Ca²⁺ leak is mainly mediated by spark-independent pathways. At higher [Ca²⁺]_SR, however, Ca²⁺ sparks become a significant contributor to SR Ca²⁺ leak (Zima et al. 2010). In certain pathological conditions associated with SR Ca²⁺ overload, SR Ca²⁺ leak becomes exacerbated. The Ca²⁺ released during a spontaneous spark diffuses to neighbouring release junctions, triggers CICR and generates arrhythmogenic spontaneous Ca²⁺ waves (Cheng et al. 1993; Diaz et al. 1997). While the importance of [Ca²⁺]_SR in the regulation of SR Ca²⁺ leak is well established, it is less clear to what extent [Ca²⁺]_i affects the RyR-mediated Ca²⁺ leak (Dibb & Eisner, 2010). [Ca²⁺]_i can regulate SR Ca²⁺ leak directly by binding to the high-affinity activation site on the cytosolic side of the RyR (Rousseau et al. 1986; Meissner & Henderson, 1987). Additionally, [Ca²⁺]_i can affect SR Ca²⁺ leak via activation of CaMKII with subsequent phosphorylation of the RyR (Guo et al. 2006; van Oort et al. 2010).

In our previous study, we investigated the effect of [Ca²⁺]_SR on SR Ca²⁺ leak (Zima et al. 2010) in conditions where [Ca²⁺]_i was held constant. In physiological conditions, however, [Ca²⁺]_i changes dynamically during the cardiac cycle. Therefore, it is important to understand to what extent [Ca²⁺]_i affects SR Ca²⁺ leak in ventricular myocytes. In this study, we used a novel approach to directly measure SR Ca²⁺ leak as changes of [Ca²⁺]_SR after complete SERCA inhibition. To measure [Ca²⁺]_SR, we used the low-affinity Ca²⁺ indicator Fluo-5N entrapped within the SR. SR Ca²⁺ leak was studied after sarcolemma permeabilization. The advantage of this approach is that an experimental solution with known [Ca²⁺] can be easily introduced into the cytosol. The experimental solution also contained the high-affinity Ca²⁺ indicator Rhod-2 to measure Ca²⁺ spark and wave properties. We found that in rabbit ventricular myocytes [Ca²⁺]_i has a pronounced effect on SR Ca²⁺ leak. The augmentation of SR Ca²⁺ leak by [Ca²⁺]_i was attributed to an increase in both spark- and non-spark-mediated Ca²⁺ leak. The effect of [Ca²⁺]_i (within the range of 50–350 nm) on SR Ca²⁺ leak was mainly mediated by a direct regulation of RyR activity, but not by CaMKII-dependent phosphorylation of the RyR. Part of this work has been published in abstract form (Bovo et al. 2010).

Methods

Myocyte isolation

Ventricular myocytes were isolated from New Zealand White rabbits (18 animals, 2–2.5 kg; Myrtle's Rabbitry, Thompsons Station, TN, USA). The procedure of cell isolation was approved by the Institutional Animal Care and Use Committee, and complies with US and UK regulations on animal experimentation (Drummond, 2009). Adult rabbits were anaesthetized with sodium pentobarbital (50 mg kg^-1 i.v.). Following thoracotomy, hearts were quickly excised, mounted on a Langendorff apparatus and retrogradely perfused with Liberase (Roche Applied Science, Indianapolis, IN, USA) Blendzyme (Roche Applied Science, Indianapolis, IN, USA)-containing solution at 37°C according to the procedure described previously (Domeier et al. 2009). Chemicals and reagents were purchased from Sigma-Aldrich (St Louis, MO, USA) unless otherwise stated. All experiments were performed at room temperature (20–24°C).

Confocal microscopy

To record [Ca²⁺]_SR and [Ca²⁺]_i we used the low-affinity Ca²⁺ indicator Fluo-5N and the high-affinity Ca²⁺ indicator Rhod-2, respectively (both indicators were obtained from Molecular Probes/Invitrogen, Carlsbad, CA, USA). To load the SR with Ca²⁺ indicator, myocytes were incubated with 5 μm Fluo-5N-AM for 2.5 h at 37°C as described previously (Zima et al. 2008b; Domeier et al. 2009).

For premeabilized cell experiments, Fluo-5N-AM-loaded myocytes were permeabilized with 0.005% saponin (Zima et al. 2008a). The experimental solution containing Rhod-2 tripotassium salt (40 μm) was composed of (in mm): potassium aspartate 100; KCl 15; KH₂PO₄ 5; MgATP 5; EGTA 0.35; CaCl₂ 0.067; MgCl₂ 0.75; phosphocreatine 10; Hepes 10; plus creatine phosphokinase 5 U ml^-1; dextran (MW: 40,000) 4%, and pH 7.2 (KOH). Free [Ca²⁺] and [Mg²⁺] of this solution were 50 nm and 1 mm, respectively. Free [Ca²⁺] in the experimental solution was adjusted to different levels (i.e. 150, 250 and 350 nm) by adding appropriate amounts of CaCl₂ (calculated using WinMAXC 2.05, Stanford University, CA, USA). In the set of experiments when effects of increased cytosolic Ca²⁺ buffer capacity on Ca²⁺ sparks and SR Ca²⁺ leak were studied, the fast Ca²⁺ buffer BAPTA (0.7 mm) was added to the experimental solution. Free [Ca²⁺] in this solution was kept the same as the control solution (verified with a Ca²⁺-sensitive electrode; Orion Research Inc.).

For intact cell experiments, Fluo-5N-AM-loaded cells were incubated at room temperature with 10 μm Rhod-2-AM for 15 min in Tyrode solution (in mm: NaCl 140; KCl 4; CaCl₂ 2; MgCl₂ 1; glucose 10; Hepes 10; pH 7.4), followed by a 20 min wash. Action potentials were induced by electrical field stimulation using a pair of platinum electrodes, which were connected to a Grass stimulator (Astro-Med. Inc., USA) set at a voltage ∼50% above the threshold for contraction. To avoid motion artifacts, the scan line was positioned along the short axis (transversal scan) in the central region of the cell where cell motion is minimal during contraction. Stimulation frequency was 0.5 Hz. Changes in [Ca²⁺]_SR were calculated by the formula (Cannell et al. 1994): [Ca²⁺]_SR = K_d×R/(K_d/[Ca²⁺]_SR,diast–R+ 1), where R was the normalized Fluo-5N fluorescence (R = [F–F_min]/[F₀–F_min]); F₀ and F_min were the fluorescence level at rest and after depletion of the SR with caffeine, respectively; K_d (Fluo-5N Ca²⁺ dissociation constant) was 390 μm based on in situ calibrations (Zima et al. 2010), and [Ca²⁺]_SR,diast (diastolic [Ca²⁺]_SR at 0.5 Hz) was 900 μm (Shannon et al. 2003). Changes in [Ca²⁺]_i were calculated according to a similar formula (Cannell et al. 1994). K_d for Rhod-2 was 1.3 μm (calibrated in situ) and [Ca²⁺]_i,diast (diastolic [Ca²⁺]_i at 0.5 Hz) was 100 nm.

Changes in [Ca²⁺]_i and [Ca²⁺]_SR were measured with laser scanning confocal microscopy (Radiance 2000 MP, Bio-Rad, UK or LSM 410, Zeiss, Germany) equipped with a ×40 oil-immersion objective lens (NA = 1.3). Fluo-5N was excited with the 488 nm line of an argon ion laser and fluorescence was measured at 515 ± 15 nm. Rhod-2 was excited with the 543 nm line of a He–Ne laser and fluorescence was measured at wavelengths >600 nm. In experiments when Ca²⁺ sparks or global cytosolic Ca²⁺ transients were recorded simultaneously with [Ca²⁺]_SR, images were acquired in line-scan mode (3 ms per scan; pixel size 0.12 μm). When [Ca²⁺]_SR was recorded alone (only Fluo-5N signal), images were collected in 2-D mode (pixel size 0.2 μm) every 15 s.

Measurements of SR Ca²⁺ leak

SR Ca²⁺ leak as a function of [Ca²⁺]_SR was measured in permeabilized myocytes according to the protocol described previously (Zima et al. 2010). Briefly, Fluo-5N was excited with minimum laser energy of an argon ion laser (to minimize dye photobleaching). To improve the signal-to-noise ratio of the low-intensity Fluo-5N signal, fluorescence was collected with an open pinhole and averaged over the entire cellular width of a line-scan or 2-D image. At the end of each experiment, minimum (F_min) and maximum (F_max) Fluo-5N fluorescence were estimated as we described previously (Zima et al. 2010). F_min was measured after depletion of the SR with 10 mm caffeine in the presence of 5 mm EGTA. F_max was measured following an increase of [Ca²⁺] to 10 mm in the presence of caffeine. Caffeine keeps RyRs open allowing [Ca²⁺] equilibration across the SR membrane (Shannon et al. 2003). The Fluo-5N signal was converted to [Ca²⁺] using the formula: [Ca²⁺]_SR = K_d× (F–F_min)/(F_max–F), where K_d was 390 μm. SR Ca²⁺ leak was measured as the changes of total [Ca²⁺]_SR ([Ca²⁺]_SRT) over time (d[Ca²⁺]_SRT/dt) after complete SERCA inhibition with thapsigargin (TG). [Ca²⁺]_SRT was calculated as: [Ca²⁺]_SRT = B_max/(1+K_d/[Ca²⁺]_SR) +[Ca²⁺]_SR; where B_max (the total concentration of SR Ca buffer) and K_d were 2700 μm and 630 μm, respectively (Shannon et al. 2000).The rate of SR Ca²⁺ leak (d[Ca²⁺]_SRT/dt) was plotted as a function of [Ca²⁺]_SR for each time point (15 s) during [Ca²⁺]_SR decline.

Measurements of Ca²⁺ sparks and waves

Ca²⁺ sparks were detected and analysed using SparkMaster (Picht et al. 2007). To exclude false-positive events, the threshold criterion for spark detection was 3.8. At this threshold no events were detected when SR Ca²⁺ was emptied after simultaneous application of caffeine (10 mm) and TG (10 μm). Analysis of Ca²⁺ sparks included spark frequency (sparks s⁻¹ (100 μm)⁻¹), amplitude (ΔF/F₀), full duration at half-maximal amplitude (FDHM; ms) and full width at half-maximal amplitude (FWHM; μm). F₀ is the initial fluorescence recorded under steady-state conditions and ΔF = F–F₀. Ca²⁺ waves were measured in line-scan mode and analysed in terms of amplitude, frequency and propagation velocity.

Statistics

Data are presented as mean ± SEM of n measured cells. Statistical comparisons between groups were performed with Student's t test. Differences were considered statistically significant at P < 0.05.

Results

The effect of cytosolic [Ca²⁺] on SR Ca²⁺ leak

Effects of different [Ca²⁺]_i on SR Ca²⁺ leak were studied in permeabilized rabbit ventricular myocytes. SR Ca²⁺ leak was measured as the rate of [Ca²⁺]_SR decline after complete SERCA inhibition with thapsigargin (TG). Figure 1A shows representative examples of [Ca²⁺]_SR recordings in control conditions and after application of TG (10 μm). The recordings were made in different cells at [Ca²⁺]_i of 50, 150, 250 and 350 nm. An increase of [Ca²⁺]_i (applied at t = 0) augmented the initial [Ca²⁺]_SR (measured before TG application). On average, initial [Ca²⁺]_SR in the presence of 50, 150, 250 and 350 nm[Ca²⁺]_i were 759 ± 28 μm (n = 11), 765 ± 25 μm (n = 21; NS compared to 50 nm); 828 ± 38 μm (n = 16; P > 0.05 compared to 50 nm) and 830 ± 44 μm (n = 8; P > 0.05 compared to 50 nm), respectively. These results suggest that higher [Ca²⁺]_i has a more significant effect on SERCA-mediated Ca²⁺ uptake than on SR Ca²⁺ leak. At 350 nm[Ca²⁺]_i, initial [Ca²⁺]_SR frequently dropped below basal level as a result of the occurrence of Ca²⁺ waves.

Figure 1. SR Ca²⁺ leak at different [Ca²⁺]; values.

A, changes of [Ca²⁺]_SR after inhibition of SERCA with thapsigargin (TG; 10 μm). Top, images of myocyte loaded with Fluo-5N in control conditions, after application of TG (10 μm; 3 and 10 min) and caffeine (10 mm). Calibration bar corresponds to 20 μm. Bottom, myocytes were exposed to different [Ca²⁺]_i values: 50 nm, blue; 150 nm, black; 250 nm, red; 350 nm, grey; and 150 nm in the presence of the RyR inhibitor Ruthenium Red (RuR; 50 μm), green. Since RyR inhibition with RuR significantly increases initial [Ca²⁺]_SR (Zima et al. 2010), for presentation purposes [Ca²⁺]_SR decline in the presence of RuR is shown from [Ca²⁺]_SR = 800 μm to make it comparable with the other recordings. Application of 10 mm caffeine at the end of the experiment (arrowheads) indicates complete depletion of the SR. The recordings were made from different cells. B, relationships between SR Ca²⁺ leak rate and [Ca²⁺]_SR measured at different [Ca²⁺]_i values. SR Ca²⁺ leak was measured after inhibition of SERCA pump with TG.

After SERCA inhibition with TG, [Ca²⁺]_SR gradually declined until full depletion (verified as a lack of further [Ca²⁺]_SR depletion after caffeine application; arrowheads in Fig. 1A). [Ca²⁺]_SR after SERCA inhibition was converted to total [Ca²⁺]_SR ([Ca²⁺]_SRT; for details see Methods). SR Ca²⁺ leak rate, which was measured as changes of [Ca²⁺]_SRT over time (d[Ca²⁺]_SRT/dt), was plotted against the corresponding free [Ca²⁺]_SR to obtain the relationship between SR Ca²⁺ leak rate and [Ca²⁺]_SR. The leak–load relationships measured in the presence of different [Ca²⁺]_i values are shown in Fig. 1B. SR Ca²⁺ leak significantly increased over the entire range of [Ca²⁺]_SR with increasing [Ca²⁺]_i. To test whether [Ca²⁺]_i affected SR Ca²⁺ leak by activating RyRs, we studied effects of different [Ca²⁺]_i values on SR Ca²⁺ leak after complete RyR inhibition with ruthenium red (RuR). Figure 1A (green symbols) shows an example of [Ca²⁺]_SR decline in the presence of RuR (50 μm) and 150 nm[Ca²⁺]_i. RuR significantly decreased SR Ca²⁺ leak, but did not prevent it (Fig. 1B). When RuR was applied at different [Ca²⁺]_i values, the rate of the RuR-insensitive Ca²⁺ leak remained the same (data not shown). Thus, these results indicate that [Ca²⁺]_i affects SR Ca²⁺ leak solely by modulating RyR activity.

Effects of cytosolic [Ca²⁺] on Ca²⁺ sparks and Ca²⁺ waves

For all [Ca²⁺]_i values studied here, SR Ca²⁺ leak increased as a function of [Ca²⁺]_SR, with a particularly steep increase at higher [Ca²⁺]_SR (Fig. 1B). We have shown previously that the increased leak rate at higher [Ca²⁺]_SR is attributed to higher Ca²⁺ spark frequency (Zima et al. 2010). Here, we studied the effects of [Ca²⁺]_i on Ca²⁺ spark properties. Since [Ca²⁺]_SR increased as a function of [Ca²⁺]_i (Fig. 1A), we analysed the effect of changing [Ca²⁺]_i between 50 and 250 nm in a subset of cells that had similar [Ca²⁺]_SR. Figure 2A shows representative line-scan images of SR Ca²⁺ release events recorded at different [Ca²⁺]_i values. The increase of [Ca²⁺]_i from 50 to 250 nm had a pronounced effect on spark frequency (Fig. 2B), which increased 5-fold. Ca²⁺ spark amplitude decreased (Fig. 2C), whereas spark width (Fig. 2D) and duration (Fig. 2E) only slightly increased at higher [Ca²⁺]_i. Increasing [Ca²⁺]_i to 350 nm produced spontaneous Ca²⁺ waves that propagated through the cell at a constant frequency of 0.9 ± 0.2 Hz (n = 6; Fig. 2A, rightmost image). These data indicate that [Ca²⁺]_i affects SR Ca²⁺ release and, therefore, spark-mediated SR Ca²⁺ leak. This occurs mainly via recruitment of the SR Ca²⁺ release units (spark frequency), but not via a profound alteration of release unit properties (spark width and duration).

Figure 2. Effects of [Ca²⁺]_i on Ca²⁺ sparks and Ca²⁺ waves.

A, example line-scan images and F/F₀ profiles of Ca²⁺ sparks and waves recorded at different [Ca²⁺]_i values. The Ca²⁺ spark and wave profiles were obtained by averaging fluorescence from the 1 μm wide region marked by the black boxes. Summary data of Ca²⁺ spark frequency (B), amplitude (C), width (measured at half-maximal amplitude, FWHM) (D), and duration (measured at half-maximal amplitude, FDHM) (E) recorded at different [Ca]²⁺_i values. *P < 0.05 vs. 50 nm[Ca²⁺]_i.

Contribution of Ca²⁺ sparks and waves to SR Ca²⁺ leak at different cytosolic [Ca²⁺] values

We have recently shown that in rabbit ventricular myocytes RyR-mediated Ca²⁺ leak is composed of two main pathways: spark mediated and non-spark mediated (Zima et al. 2010). Depending on SR Ca²⁺ load, these two pathways contribute to a different degree to the total SR Ca²⁺ leak. At low [Ca²⁺]_SR (<400 μm), Ca²⁺ leak occurred mostly as undetectable openings of RyRs. At high [Ca²⁺]_SR (>600 μm), however, Ca²⁺ sparks became the significant pathway of SR Ca²⁺ leak. Here, we studied to what degree [Ca²⁺]_i affects these two components of SR Ca²⁺ leak. First, we obtained RyR-dependent Ca²⁺ leak as a function of [Ca²⁺]_SR (Fig. 3A) by subtracting the RuR-insensitive component of Ca²⁺ leak (Fig. 1B; green points) from the total SR Ca²⁺ leak (Fig. 1B). After subtraction, the points of SR Ca²⁺ leak at low [Ca²⁺]_SR (between 50 and 450 μm) were best fitted with single Hill functions (Fig. 3B) (Zima et al. 2010). These functions were used to describe the relationship between [Ca²⁺]_SR and non-spark-mediated Ca²⁺ leak for different [Ca²⁺]_i values. Next, we estimated the spark-mediated leak by subtracting the corresponding non-spark-mediated leak (Fig. 3B) from the RyR-mediated Ca²⁺ leak (Fig. 3A). For different [Ca²⁺]_i, the obtained points were best fitted with single exponential functions shown in Fig. 3C. In the case of 350 nm[Ca²⁺]_i (grey line), SR Ca²⁺ leak was mainly mediated by Ca²⁺ waves. Figure 3D illustrates the relative contribution of spark- and non-spark-mediated Ca²⁺ leak to the RyR-dependent Ca²⁺ leak (at [Ca²⁺]_SR = 900 μm) for different [Ca²⁺]_i values. This analysis revealed that both components of RyR-mediated Ca²⁺ leak increased as a function of [Ca²⁺]_i. While non-spark-mediated Ca²⁺ leak increased linearly, spark/wave-mediated leak rose exponentially with [Ca²⁺]_i. These results indicate that cytosolic [Ca²⁺] significantly affects both components of RyR-mediated Ca²⁺ leak in ventricular myocytes.

Figure 3. Contribution of Ca²⁺ sparks and waves to SR Ca²⁺ leak.

A, RyR-mediated SR Ca²⁺ leak as a function of [Ca²⁺]_SR measured at different [Ca²⁺]_i values. The RyR-mediated Ca²⁺ leak was obtained by subtracting SR Ca²⁺ leak in the presence of RuR from the total SR Ca²⁺ leak. B, relationships between non-spark-mediated Ca²⁺ leak and [Ca²⁺]_SR obtained at different [Ca]_i values. C, relationships between spark- and wave-mediated Ca²⁺ leak and [Ca²⁺]_SR obtained at different [Ca]_i values. D, the spark-mediated Ca²⁺ leak (filled circles) and the non-spark mediated Ca²⁺ leak (open circles) at [Ca²⁺]_SR = 900 μm were plotted as a function of [Ca²⁺]_i. The spark-mediated leak was fitted with an exponential function (R² = 0.86) and the non-spark-mediated leak was fitted with a linear function (R² = 0.91).

In the next set of experiments, we used the fast Ca²⁺ buffer BAPTA to eliminate Ca²⁺ sparks by decreasing efficacy of local CICR within RyR clusters. Addition of 0.7 mm BAPTA to the experimental solution decreased Ca²⁺ spark frequency by 87% (Fig. 4A and B), amplitude by 67% and width by 40%. Free [Ca²⁺]_i of the BAPTA solution was kept the same as the control solution (250 nm). As BAPTA did not affect basal [Ca²⁺]_SR (Fig. 4C), changes in spark properties were not due to decreased SR Ca²⁺ load. Therefore, by using BAPTA we were able to significantly eliminate spark-mediated Ca²⁺ leak without affecting single RyR activity. The analysis of SR Ca²⁺ leak at 250 nm[Ca²⁺]_i in the control solution and in the presence of BAPTA (0.7 mm) (Fig. 4C) revealed that an increase of cytosolic Ca²⁺ buffer capacity with BAPTA significantly decreased SR Ca²⁺ leak, particularly at high [Ca²⁺]_SR (>400 μm; Fig. 4D) where Ca²⁺ sparks significantly contribute to SR Ca²⁺ leak (Zima et al. 2010). Furthermore, Ca²⁺ leak measured in the presence of BAPTA was similar to non-spark-mediated Ca²⁺ leak obtained by the mathematical approach (Fig. 3B and dashed line in Fig. 4D).

Figure 4. Effects of BAPTA on Ca²⁺ sparks and SR Ca²⁺ leak.

A, line-scan images and F/F₀ profiles of Ca²⁺ sparks recorded at 250 nm[Ca]_i in control and in the presence of 0.7 mm BAPTA. The Ca²⁺ spark profiles were obtained by averaging fluorescence from the 1 μm wide regions marked by the black boxes. B, effect of BAPTA on Ca²⁺ spark frequency. C, decline of [Ca²⁺]_SR during SERCA inhibition in control solution (open circles) and in the presence of 0.7 mm BAPTA (filled circles). Free [Ca²⁺]_i of both solutions was kept at 250 nm. D, relationships between SR Ca²⁺ leak and [Ca²⁺]_SR measured at 250 nm[Ca²⁺]_i in control (open circles) and BAPTA-containing solution (filled circles). For comparison, the dashed line indicates non-spark-mediated Ca²⁺ leak at 250 nm[Ca²⁺]_i obtained by the mathematical approach (data from Fig. 3B).

Role of CaMKII in regulation of SR Ca²⁺ leak at high cytosolic [Ca²⁺]

In the following experiments, we investigated whether the augmentation of SR Ca²⁺ leak observed at high cytosolic [Ca²⁺] (>150 nm) was in part mediated by CaMKII-dependent phosphorylation of the RyR. Here we studied the effect of the specific CaMKII inhibitor autocamtide 2-related inhibitory peptide (AIP) on SR Ca²⁺ leak at high [Ca²⁺]_i. Figure 5A shows a representative recording of [Ca²⁺]_SR in the presence of AIP (5 μm), after subsequently increasing [Ca²⁺]_i to 250 nm, as well as after application of TG (10 μm). AIP did not significantly affect initial [Ca²⁺]_SR (recorded before TG application). Initial [Ca²⁺]_SR was 795 ± 44 μm in control conditions and remained steady at 771 ± 66 μm (n = 10) after AIP application. In the presence of AIP, SR Ca²⁺ leak only changed from 7.2 ± 0.6 to 8.5 ± 1.1 μm s^-1 (n = 10; [Ca²⁺]_SR = 750 μm). Thus, inhibition of CaMKII did not have a significant effect on SR Ca²⁺ leak (Fig. 5B), although the results show a tendency to increase SR Ca²⁺ leak (∼16%) at 250 nm[Ca²⁺]_i. Likewise, we did not observe a significant effect of AIP on SR Ca²⁺ leak in the presence of 350 nm[Ca²⁺]_i. SR Ca²⁺ leak was 16.0 ± 0.9 μm s^-1 (n = 8) in control conditions and changed to 17.8 ± 1.4 μm s^-1 (n = 7) after AIP (5 μm) application. AIP also did not change the properties of Ca²⁺ waves (i.e. frequency, amplitude and velocity; data not shown). These results indicate that CaMKII-mediated RyR phosphorylation is not involved in the effect of [Ca²⁺]_i (in the range from 50 to 350 nm) on SR Ca²⁺ leak. Thus, [Ca²⁺]_i regulates Ca²⁺ leak mainly via direct action on the RyR.

Figure 5. Effects of CaMKII inhibition on SR Ca²⁺ leak.

A, effect of 250 nm[Ca²⁺]_i on decline of [Ca²⁺]_SR during SERCA inhibition in the presence of AIP (5 μm). For comparison, the dashed line indicates the decline of [Ca²⁺]_SR at 250 nm[Ca²⁺]_i in the absence of CaMKII inhibition (data from Fig. 1A). B, average effect of AIP (5 μm) on SR Ca²⁺ leak measured at [Ca²⁺]_SR = 750 μm.

Changes of SR Ca²⁺ leak during excitation–contraction coupling (ECC)

RyR-mediated Ca²⁺ leak is controlled by the level of Ca²⁺ on both sides of the SR membrane (Fig. 1B). During the cardiac cycle, cytosolic and luminal [Ca²⁺] dynamically change, following opposite directions ([Ca²⁺]_SR is replenished while [Ca²⁺]_i declines and vice versa). Thus, the decline of [Ca²⁺]_i during late systole and early diastole would ultimately offset the stimulatory effect of [Ca²⁺]_SR on the RyR. In the following experiments, we measured changes of [Ca²⁺]_SR and [Ca²⁺]_i during ECC and analysed how these changes would affect SR Ca²⁺ leak. [Ca²⁺]_SR and [Ca²⁺]_i were recorded simultaneously in intact ventricular myocytes at 0.5 Hz pacing frequency. Figure 6A shows line-scan images of Ca²⁺ transient (top) and corresponding [Ca²⁺]_SR depletion (bottom) evoked by electrical field stimulation. We measured [Ca²⁺]_SR during a SR replenishing phase at the same three [Ca²⁺]_i values (150, 250 and 350 nm), that were studied in permeabilized cell experiments. [Ca²⁺]_i was inversely proportional to [Ca²⁺]_SR with a slope constant of –1100 (Fig. 6B). We estimated how changes in [Ca²⁺]_i and [Ca²⁺]_SR (which occur during diastole) affect SR Ca²⁺ leak. At 350 nm[Ca²⁺]_i, [Ca²⁺]_SR was 608±45 μm (n = 9; Fig. 6B) and SR Ca²⁺ leak was 7.9 μm s^-1 (Fig. 5C, blue open circle). At 250 nm[Ca²⁺]_i, [Ca²⁺]_SR was replenished to 718 ± 27 μm (n = 9; Fig. 6B) and Ca²⁺ leak slightly decreased to 7.5 μm s^-1 (Fig. 6C, red open circle). At 150 nm[Ca²⁺]_i, [Ca²⁺]_SR reached 819 ± 11 μm (n = 9; Fig. 6B) and SR Ca²⁺ leak was further reduced to 6.9 μm s^-1 (Fig. 6C, black open circle). This analysis suggests that only small changes (∼13%) of SR Ca²⁺ leak occur during the declining phase of the Ca²⁺ transient. Since SR Ca²⁺ leak steeply depends on both [Ca²⁺]_SR and [Ca²⁺]_i (Fig. 1B), but [Ca²⁺]_SR and [Ca²⁺]_i (Fig. 6A) change in opposite directions, SR Ca²⁺ leak remains relatively constant during the SR replenishing phase of the cardiac cycle.

Figure 6. Estimation of SR Ca²⁺ leak during the declining phase of the Ca²⁺ transient.

A, simultaneously recorded [Ca²⁺]_i transient and corresponding [Ca²⁺]_SR depletion. Top, line-scan image of Rhod-2 fluorescence and profile of [Ca²⁺]_i. Bottom, line-scan image of Fluo-5N fluorescence and profile of [Ca²⁺]_SR. The profiles were obtained by averaging fluorescence from the entire cellular width of line scans. B, summary data showing changes of [Ca²⁺]_SR measured at three different [Ca²⁺]_i values (i.e. 150, 250 and 350 nm; indicated by arrows in A). C, RyR-mediated SR Ca²⁺ leak as a function of [Ca²⁺]_SR measured at different [Ca²⁺]_i values (data were taken from Fig. 3A). For each [Ca²⁺]_i studied, SR Ca²⁺ leak was extrapolated at the corresponding [Ca²⁺]_SR, calculated from the [Ca²⁺]_SR profile in A. The results suggest that during the declining phase of the Ca²⁺ transient leak remains relatively constant.

Discussion

In recent years, a significant effort has been made to understand the mechanisms that control SR Ca²⁺ leak in healthy and diseased hearts. A relatively small augmentation of SR Ca²⁺ efflux during diastole can significantly influence cardiac function by depleting the SR Ca²⁺ content and reducing contraction. The increased Ca²⁺ leak can also be pro-arrhythmic as the released Ca²⁺ is extruded by the electrogenic Na⁺–Ca²⁺ exchange (NCX), leading to delayed afterdepolarizations (Kass et al. 1978; Pogwizd et al. 2001). The bulk of SR Ca²⁺ leak is mediated by RyRs (Neary et al. 2002; Shannon et al. 2002; Zima et al. 2010), which are complexly regulated by Ca²⁺ from both cytosolic and luminal sides of the channel (Fill & Copello, 2002; Meissner, 2004). Since Fabiato's work (Fabiato, 1985a,b), [Ca²⁺]_i has been considered a crucial factor in the activation of global SR Ca²⁺ release during systole. However, it is less clear to what extent dynamic changes of [Ca²⁺]_i during diastole affects RyR activity and SR Ca²⁺ leak. Here we directly measured Ca²⁺ leak as the rate of [Ca²⁺]_SR decline after SERCA inhibition in permeabilized rabbit ventricular myocytes. By introducing different [Ca²⁺]_i levels into the cytosol, we found that: (1) [Ca²⁺]_i increases SR Ca²⁺ leak primarily by activating the RyR-mediated leak pathway; (2) with increasing [Ca²⁺]_i, Ca²⁺ sparks and waves become significant pathways of Ca²⁺ leak; (3) within the range of 50–350 nm[Ca²⁺]_i, the effect on SR Ca²⁺ leak is predominantly mediated by direct activation of the RyR and not by CaMKII-mediated phosphorylation of the channel; and (4) during late systole and early diastole (when [Ca²⁺]_SR and [Ca²⁺]_i change in opposite directions) SR Ca²⁺ leak remains relatively constant as a result of strong dependence from both [Ca²⁺]_i and [Ca²⁺]_SR.

Our results show that [Ca²⁺]_i affects SR Ca²⁺ leak entirely by regulating RyR activity. We found that in the presence of the potent RyR inhibitor RuR, the residual SR Ca²⁺ leak was insensitive to changes in [Ca²⁺]_i. High [Ca²⁺]_i can facilitate SR Ca²⁺ leak by increasing RyR open probability. However, it also decreases the SR Ca²⁺ gradient and therefore the driving force for SR Ca²⁺ leak. Our results show that despite decrease of the SR Ca²⁺ gradient (by changing [Ca²⁺]_i from 50 to 350 nm), RyR-mediated Ca²⁺ leak increased more than 5 times (measured at the same [Ca²⁺]_SR; Figs 1B and 3A). This finding suggests that SR Ca²⁺ leak is not a simple diffusion of Ca²⁺ from the SR, but rather a well-regulated process. We have previously shown that RyR-mediated Ca²⁺ leak can occur as spontaneous Ca²⁺ sparks, but also as undetectable openings of RyRs (non-spark-mediated Ca²⁺ leak; Zima et al. 2010). The augmentation of SR Ca²⁺ leak by cytosolic Ca²⁺ was attributed to an increase of both components of SR Ca²⁺ leak (Fig. 3B and C). These two pathways, however, were differently regulated by [Ca²⁺]_i. While non-spark-mediated Ca²⁺ leak linearly depended on [Ca²⁺]_i (Fig. 3D), spark-mediated Ca²⁺ leak rose steeply at higher [Ca²⁺]_i. The latter was mainly mediated by an increase of Ca²⁺ spark frequency, since other spark properties (width and duration) were only minimally affected by [Ca²⁺]_i (Fig. 2). A similar effect of [Ca²⁺]_i on spark frequency was observed in permeabilized rat ventricular myocytes (Lukyanenko & Gyorke, 1999).

The analysis of the spark properties also shows that the amplitude proportionally decreased with an increase of [Ca²⁺]_i, indicating that this spark parameter is mainly determined by SR Ca²⁺ release flux and, therefore, by the SR Ca²⁺ gradient. However, despite the Ca²⁺ flux during individual release events at 350 nm[Ca²⁺]_i being smaller than at [Ca²⁺]_i < 350 nm, it is sufficient to increase [Ca²⁺]_i next to neighbouring junctions to activate local CICR and trigger Ca²⁺ waves. When [Ca²⁺]_SR was depleted to ∼500 μm (during SERCA inhibition) Ca²⁺ waves ceased completely (data not shown), suggesting that the [Ca²⁺] on both sides of the RyR plays an important role in the generation of Ca²⁺ waves. It appears that high [Ca²⁺]_i increases the sensitivity of release units to CICR and elevated [Ca²⁺]_SR maintains SR Ca²⁺ flux strong enough to produce Ca²⁺ waves via a ‘fire–diffuse–fire’ mechanism (Keizer & Smith, 1998). We found that the transformation of SR Ca²⁺ release from localized sparks to propagating waves almost doubled RyR-mediated Ca²⁺ leak (Fig. 3C), suggesting that Ca²⁺ waves represent spontaneous SR Ca²⁺ release events in their extreme form.

It has been suggested that an increase of SR Ca²⁺ leak by high [Ca²⁺]_i is mediated partially by CaMKII-dependent phosphorylation of RyRs (Guo et al. 2006; van Oort et al. 2010). Using the selective CaMKII inhibitor AIP, we found that the augmentation of SR Ca²⁺ leak at high [Ca²⁺]_i was insensitive to CaMKII inhibition. In contrast to previous studies, AIP treatment showed some tendency to accelerate SR Ca²⁺ leak (Fig. 5). The discrepancy with previously published work can be explained by the fact that CaMKII was not activated under our experimental conditions. It has been shown that activation of endogenous CaMKII requires much higher [Ca²⁺]_i (∼500 nm), phosphatase inhibition (e.g. with okadaic acid) and calmodulin (Guo et al. 2006). In our experiments none of these conditions were applied. The small acceleration of SR Ca²⁺ leak observed in the presence of AIP can be the result of a potential direct interaction of the peptide with the RyR. Thus, in the range of 50–350 nm[Ca²⁺]_i regulates SR Ca²⁺ leak mainly via a CaMKII-independent mechanism.

Another important finding of this study is that SR Ca²⁺ leak is highly sensitive to both [Ca²⁺]_i and [Ca²⁺]_SR (Fig. 1B). It remains unresolved whether luminal and cytosolic Ca²⁺ regulate the RyR via entirely independent mechanisms or via a well coordinated process. It has been shown that the RyR can be activated independently by Ca²⁺ from the luminal or cytosolic sides of the channel (Sitsapesan & Williams, 1994; Gyorke & Gyorke, 1998; Qin et al. 2008). However, luminal [Ca²⁺] can also regulate RyR by acting on the cytosolic Ca²⁺ activation site of the channel via a ‘feed-through’ mechanism (Laver, 2007). The combination of all these regulatory mechanisms contributes to the fine control of the RyR. Furthermore, our data suggest that [Ca²⁺]_SR regulation provides an important feedback mechanism to counter the positive feedback of high [Ca²⁺]_i during the declining phase of the Ca²⁺ transient, whereas at the end of diastole, the high SR Ca²⁺ leak driving force determined by high [Ca²⁺]_SR is well equilibrated by the low [Ca²⁺]_i. The balance established by this dual regulation results in a SR Ca²⁺ leak rate that remains relatively constant during the decline of the Ca²⁺ transient when [Ca²⁺]_SR and [Ca²⁺]_i change in opposite directions. However, alterations in the activity of RyR during diastole can affect the basal Ca²⁺ leak rate and, therefore, the amount of Ca²⁺ stored in the SR. This would lead to altered systolic Ca²⁺ transients, potential generation of arrhythmias and contractile dysfunctions. There are, however, significant differences between animal species regarding the role of RyR modification and its effect on SR Ca²⁺ release and ECC. In rat ventricular myocytes, for example, RyR sensitization with low doses of caffeine produced only a transient positive inotropic effect (Trafford et al. 2000). However, studies conducted on ventricular myocytes from larger animals (e.g. dogs and rabbits) (Belevych et al. 2007; Domeier et al. 2009) showed that caffeine led to significant depletion of [Ca²⁺]_SR and a decrease of Ca²⁺ transient amplitude. The differences between rats and rabbits are probably due to the differences in expression of SERCA and NCX (Bers, 2001). In contrast to rats, rabbits have less SERCA but increased NCX expression. This translates functionally into less SR Ca²⁺ reuptake by SERCA, more Ca²⁺ extrusion from the cell via NCX, and greater net cellular Ca²⁺ loss.

The coordinated regulation of RyR by luminal and cytosolic Ca²⁺ provides a fundamental mechanism that contributes to setting the appropriate SR Ca²⁺ load during diastole and therefore preventing the onset of pathological conditions (e.g. SR Ca²⁺ overload). This is an important step towards the understanding of the RyR regulation. In intact cells, however, additional mechanisms are probably involved. For instance, our estimation did not account for a potential SR Ca²⁺ release refractoriness which can occur during ECC in intact cells. In permeabilized cells, the potential contribution of Ca²⁺-dependent inactivation (Fabiato, 1985b) in offsetting the SR Ca²⁺ leak at high [Ca²⁺]_i can be ruled out based on several previous findings. First, Ca²⁺-dependent inactivation has not been reported in permeabilized cell experiments when SR Ca²⁺ release was measured at [Ca²⁺]_i >10 μm (Stevens et al. 2009). Second, the study of cardiac RyR in lipid bilayers has shown that RyR inactivation occurs at significantly higher [Ca²⁺]_i (>0.5 mm) (Xu et al. 1996). Thus, the 13% difference between the leak rates at the different points of [Ca²⁺]_i and corresponding [Ca²⁺]_SR during the declining phase of the Ca²⁺ transient (Fig. 6C), might be explained by the fact that our estimations obtained in permeabilized cells are not accounting for all the regulatory mechanisms that play a role in setting the RyR activity in the intact cellular environment (e.g. use-dependent inactivation).

In conclusion, this study is the first that we know of to characterize the effect of [Ca²⁺]_i on SR Ca²⁺ leak in permeabilized ventricular myocytes. We found that [Ca²⁺]_i plays a critical role in controlling SR Ca²⁺ leak. The increase of [Ca²⁺]_i and RyR sensitivity to [Ca²⁺]_i in heart failure or other pathological conditions, can lead to the alteration of diastolic SR Ca²⁺ leak and to the generation of arrhythmogenic events. These findings are of great importance because they explain a new regulatory mechanism of SR Ca²⁺ leak that is critical to maintain the physiological homeostasis of Ca²⁺ in the heart.

Glossary

Abbreviation

AIP

autocamtide 2-related inhibitory peptide

[Ca²⁺]_i

cytosolic free calcium concentration

CaMKII

Ca²⁺–calmodulin-dependent kinase type II

CICR

Ca²⁺-induced Ca²⁺ release

ECC

excitation–contraction coupling

FDHM

full duration at half-maximum

FWHM

full width at half-maximum

HF

heart failure

NCX

Na⁺–Ca²⁺ exchange

PLB

phospholamban

RuR

ruthenium red

RyR

ryanodine receptor

SERCA

sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase

SR

sarcoplasmic reticulum

[Ca²⁺]_SR

sarcoplasmic reticulum free calcium concentration

TG

thapsigargin

Author contributions

E.B., S.R.M., A.V.Z. and L.A.B. contributed to the conception and design of the study, interpretation of data and writing of the manuscript. E.B., S.R.M. and A.V.Z. performed the experimental work and analysis of results. All authors have approved the version to be published.