Alteration of sarcoplasmic reticulum Ca²⁺ release termination by ryanodine receptor sensitization and in heart failure

Abstract

Many physiological processes and pharmacological agents modulate the ryanodine receptor (RyR), the primary sarcoplasmic reticulum (SR) Ca²⁺ release channel in the heart. However, how such modulations translate into functional effects during cardiac excitation–contraction coupling (ECC) is much less clear. Using a low dose (250 μm) of caffeine we sensitized the RyR and examined SR Ca²⁺ release using dynamic measurements of cytosolic Ca²⁺ ([Ca²⁺]_i) and free Ca²⁺ within the SR ([Ca²⁺]_SR). In field stimulated (1 Hz) rabbit ventricular myocytes, application of 250 μm caffeine caused an initial 33% increase in SR Ca²⁺ release, which was followed by a decrease in SR Ca²⁺ load (28%) and steady-state SR Ca²⁺ release (23%). To investigate the effects of caffeine on local SR Ca²⁺ release, we measured [Ca²⁺]_SR from individual release junctions during ECC as well as during spontaneous Ca²⁺ sparks. In intact myocytes during ECC, caffeine increased global fractional SR Ca²⁺ release by decreasing the [Ca²⁺]_SR level at which local release terminated by 21%. Similarly, in permeabilized myocytes during spontaneous Ca²⁺ sparks, caffeine decreased the [Ca²⁺]_SR level for release termination by 12%. Finally, we examined if Ca²⁺ release termination was changed in myocytes from failing hearts, where remodelling processes lead to altered RyR function. In myocytes from failing rabbit hearts, the [Ca²⁺]_SR termination level for Ca²⁺ sparks was 13% lower than that of non-failing myocytes. Collectively, these data suggest that altering the termination level for local Ca²⁺ release may represent a novel mechanism to increase SR Ca²⁺ release and contractility during ECC.

Introduction

In the heart excitation–contraction coupling (ECC) occurs via Ca²⁺-induced Ca²⁺ release (CICR), where localized influx of Ca²⁺ via L-type Ca²⁺ channels triggers further release of Ca²⁺ from the SR through ryanodine receptor (RyR) release clusters. The global intracellular Ca²⁺ increase that activates contraction during ECC is due to the summation of thousands of individual CICR events, which in ventricular myocytes occur at specialized release units where L-type Ca²⁺ channels in the T-tubule membrane closely approach the RyR release clusters of the junctional SR (for review see Bers, 2001). Elementary Ca²⁺ release events from these clusters also occur spontaneously, and can be visualized as distinct increases in cytosolic Ca²⁺, termed Ca²⁺ sparks (Cheng et al. 1993), or as the reciprocal intra-SR Ca²⁺ depletion, events called Ca²⁺ blinks (Brochet et al. 2005). During diastole, relaxation occurs as a result of Ca²⁺ being sequestered back into the SR via the sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase (SERCA), as well as by Ca²⁺ being removed from the cell, predominantly via the Na⁺/Ca²⁺ exchanger (NCX). In the steady-state, the amount of Ca²⁺ that enters the cell must be extruded, and the amount of Ca²⁺ that is released during CICR must be pumped back into the SR via SERCA (Eisner et al. 2000). Alterations in this cellular Ca²⁺ flux balance result in pathological conditions such as arrhythmogenic Ca²⁺ overload (net cellular Ca²⁺ gain) and chronic [Ca²⁺]_SR depletion and subsequent decreased contractility as seen in heart failure (net cellular Ca²⁺ loss).

Heart failure is a complex pathological condition simply defined at the organ level as a decrease in cardiac pump function. At the cellular level the reduction in contractility associates with substantial decreases in [Ca²⁺]_SR and SR Ca²⁺ release during ECC (Lindner et al. 1998; Pieske et al. 1999; Hobai & O’Rourke, 2001; Pogwizd et al. 2001). Recent evidence suggests that cardiac contraction would be further reduced if it were not for concomitant modulations to RyRs, which sensitize release clusters to release Ca²⁺. This becomes evident when SR Ca²⁺ content is experimentally matched, and Ca²⁺ spark frequency and fractional SR Ca²⁺ release during ECC are increased in heart failure myocytes compared to non-failing myocytes (Shannon et al. 2003b; Kubalova et al. 2005; Guo et al. 2007). Unfortunately, the modifications which occur in heart failure also create instability during diastole in the form of SR Ca²⁺ leak (Shannon et al. 2003b; Oda et al. 2005; Wehrens et al. 2005; Belevych et al. 2007) and spontaneous Ca²⁺ release events such as Ca²⁺ waves, making these compensatory modifications potentially arrhythmogenic (George, 2008).

Ca²⁺ release from the SR, and therefore cardiac contraction, is a finely graded process, which allows the heart to adjust its contractility to meet physiological demands. However, for SR Ca²⁺ release to be graded, robust release termination mechanisms must exist to counteract the inherent positive feedback and all-or-nothing nature of CICR. Based on global [Ca²⁺]_SR measurements during ECC (Shannon et al. 2003a), as well as of spontaneous Ca²⁺ blinks (Brochet et al. 2005; Terentyev et al. 2008b; Zima et al. 2008b), it is apparent that Ca²⁺ release does not simply terminate due to exhaustion of releasable Ca²⁺. Experimental evidence and mathematical models have suggested numerous potential termination mechanisms, including adaptation, inactivation, stochastic attrition and cluster de-activation via a luminal [Ca²⁺]_SR-dependent mechanism (Stern & Cheng, 2004). In previous work we have shown that local Ca²⁺ release terminates at a set [Ca²⁺]_SR level, irrespective of the initial [Ca²⁺]_SR or Ca²⁺ release flux (Zima et al. 2008b). These data show that Ca²⁺ release terminates via a distinct luminal [Ca²⁺]_SR-dependent mechanism, which makes termination mechanisms such as stochastic attrition (which would produce a random [Ca²⁺]_SR termination level) and [Ca²⁺]_i-dependent inactivation (higher [Ca²⁺]_SR termination level with increasing Ca²⁺ release flux) unlikely. Furthermore, by limiting release flux and maintaining [Ca²⁺]_SR above this [Ca²⁺]_SR termination level, long-lasting Ca²⁺ release events were visualized, which then likely terminated via [Ca²⁺]_SR-independent mechanisms (Zima et al. 2008a). While these data clearly point to a luminal [Ca²⁺]_SR threshold for release termination, it remains unclear if this threshold can be altered by pharmacological or physiological modulation of the RyR.

The aim of this study was to investigate the mechanisms by which RyR sensitization increases fractional SR Ca²⁺ release in rabbit ventricular myocytes. For this purpose we pharmacologically sensitized RyRs using a low dose (250 μm) of the classical RyR activator caffeine and monitored [Ca²⁺]_i and [Ca²⁺]_SR during ECC as well as during spontaneous Ca²⁺ sparks. We then used myocytes isolated from a well-characterized rabbit model of heart failure (Pogwizd, 1995) to investigate if similar effects occur in the failing heart, where endogenous modifications ‘sensitize’ RyRs and increase fractional [Ca²⁺]_SR release. We show that RyR sensitization with caffeine increases fractional SR Ca²⁺ release via several functional mechanisms. First, release clusters are sensitized to release activation, which increases release junction recruitment and synchronizes release during CICR. Furthermore, RyR sensitization also increases fractional SR Ca²⁺ release at each individual release junction by decreasing the [Ca²⁺]_SR level where release terminates. In heart failure myocytes strikingly similar effects occurred as part of the remodelling process, suggesting that these mechanisms are also relevant during physiological modulation of the RyR. Taken together, these data suggest that release activation and release termination represent two mechanisms by which SR Ca²⁺ release is regulated. Modulation of the termination threshold for Ca²⁺ release may therefore represent an important inotropic mechanism during cardiac ECC.

Preliminary accounts of this work have been published in abstract form (Zima et al. 2007; Domeier et al. 2009).

Methods

Solutions and chemicals

All chemicals and reagents were purchased from Sigma-Aldrich (St Louis, MO, USA), unless noted otherwise. Tyrode solution contained (in mm): 135 NaCl, 4 KCl, 2 CaCl₂, 1 MgCl₂, 10 d-glucose, 10 Hepes; pH 7.4 with NaOH. The internal cell solution for permeabilized myocytes contained (in mm): 100 potassium aspartate, 15 KCl, 5 KH₂PO₄, 5 MgATP, 0.35 EGTA, 0.12 CaCl₂, 0.75 MgCl₂, 10 phosphocreatine, 5 U ml⁻¹ creatine phosphokinase, 8% dextran (M_r: 40 000), 10 Hepes, 0.04 rhod-2 tripotassium salt or 0.03 fluo-4 pentapotassium salt (Molecular Probes/Invitrogen, Carlsbad, CA, USA); pH 7.2 with KOH. Free [Ca²⁺] and [Mg²⁺] of this solution were 150 nm and 1 mm, respectively.

Myocyte isolation

Ventricular myocytes were isolated from New Zealand White rabbits (25 animals, 2.5 kg, Myrtle's Rabbitry, Thompsons Station, TN, USA). Rabbits were anaesthetized with sodium pentobarbital (50 mg kg⁻¹) and hearts were excised and mounted on a Langendorff apparatus. Hearts were retrogradely perfused with nominally Ca²⁺-free Tyrode solution for 5 min, followed by minimal essential medium Eagle (MEM) solution containing 20 μm Ca²⁺ and 45 μg ml⁻¹ Liberase Blendzyme 4 (Roche Applied Science, Indianapolis, IN, USA) for 20 min at 37°C. The left ventricular free wall was removed from the heart and digested for an additional 5 min in the enzyme solution at 37°C. Digested tissue was then minced, filtered, and washed in a MEM solution containing 50 μm Ca²⁺ and 10 mg ml⁻¹ bovine serum albumin. Isolated cells were kept in MEM solution with 50 μm Ca²⁺ at room temperature (22–24°C) until used for experimentation. In addition, left ventricular myocytes were isolated from rabbits with non-ischaemic heart failure induced by combined aortic insufficiency and stenosis (6 animals) (Pogwizd, 1995). All protocols were approved by the Institutional Animal Care and Use Committee, and comply with US and UK regulations on animal experimentation (Drummond, 2009).

Ca²⁺ indicator dye loading and general procedures

To load the sarcoplasmic reticulum with the low-affinity Ca²⁺ indicator fluo-5N, ventricular myocytes were incubated with 5 μm of the membrane permeant fluo-5N/AM (Molecular Probes/Invitrogen) with 0.25% Pluronic F-127 in Tyrode solution for 2.5 h, followed by a 30 min wash, all at 37°C. For cytosolic measurements of Ca²⁺ in intact myocytes, cells were incubated at room temperature with 10 μm rhod-2/AM (Molecular Probes/Invitrogen) or 10 μm fluo-4/AM (Molecular Probes/Invitrogen) for 10 min, followed by a 20 min wash. Cells were placed on laminin-coated coverslips. Action potentials and ECC were induced by electrical field stimulation using a pair of platinum electrodes, which were connected to a Grass stimulator set at a voltage ∼50% above the threshold. To study individual Ca²⁺ release events (Ca²⁺ sparks and blinks) in a precisely controlled intracellular environment, myocytes were permeabilized with 0.005% saponin for 30 s and exposed to the internal cell solution, which contained rhod-2 tripotassium salt or fluo-4 pentapotassium salt (Molecular Probes/Invitrogen). All experiments were performed at room temperature.

Epifluorescence microscopy

Global [Ca²⁺]_SR or [Ca²⁺]_i was monitored using fluo-5N or fluo-4, respectively, with dye excitation at 480 ± 20 nm and emission collected at 535 ± 25 nm. Fluo-4 emission was background subtracted and normalized to baseline fluorescence emission (F₀), with changes of [Ca²⁺]_i presented as ΔF/F₀ (where ΔF=F−F₀). Fluo-5N fluorescence emission was background subtracted and calibrated to [Ca²⁺]_SR using [Ca²⁺]_SR=K_D× (F−F_Min)/(F_Max−F), where K_D= 400 μm (Shannon et al. 2003a). F_Min is the intra-SR fluorescence in the absence of Ca²⁺ obtained after complete emptying of [Ca²⁺]_SR with 10 mm caffeine. This value was corrected for the 15% chemical quench of fluo-5N by this concentration of caffeine. F_Max was defined as the diastolic fluorescence in the presence of 1 μm isoproterenol during 1 Hz pacing. In the presence of isoproterenol [Ca²⁺]_SR increases to levels where intra-SR fluo-5N is nearly saturated with Ca²⁺ and this therefore approximates maximal fluo-5N fluorescence (Shannon et al. 2003a).

Confocal microscopy

Confocal microscopy (Radiance 2000/MP, Bio-Rad, UK) was used to simultaneously image [Ca²⁺]_SR (fluo-5N) and [Ca²⁺]_i (rhod-2), with fluo-5N excitation at 488 nm and emission collected at 515 ± 15 nm, and rhod-2 excitation at 543 nm with emission collected at >600 nm. In some experiments [Ca²⁺]_i was monitored alone using fluo-4 with excitation at 488 nm and emission collected at >500 nm. Images were acquired in linescan mode at 3 ms per scan with a pixel size of 0.12 μm. Measurements during ECC were acquired by positioning the scan line in a central region along the short axis of the cell (transverse scan) to minimize motion artifacts. For measurements in permeabilized myocytes the scan line was positioned in the cytosol along the longitudinal axis of the cell. Rhod-2 fluorescence emission is presented as ΔF/F₀ (where ΔF=F−F₀), while fluo-5N fluorescence was calibrated to [Ca²⁺]_SR or normalized to baseline fluorescence (F/F₀) using (F−F_Min/F₀−F_Min). F₀ corresponds to diastolic fluorescence at 0.5 Hz pacing in intact myocytes (Fig. 3) or baseline resting fluorescence in permeabilized myocytes (Figs 4 and 6). To directly compare permeabilized myocytes from normal rabbits and heart failure rabbits, SR Ca²⁺ load was approximated indirectly using the amplitude of the fluo-4 [Ca²⁺]_i transient in response to 10 mm caffeine. Additionally, resting [Ca²⁺]_SR was calculated directly from the fluo-5N fluorescence signal (see above). In permeabilized cells, F_Min was the fluorescence signal obtained after application of 10 mm caffeine in a 0 Ca²⁺ solution containing 5 mm EGTA, while F_Max was obtained in the presence of 10 mm caffeine and 10 mm Ca²⁺. For F_Max calibration, 10 μm blebbistatin and 5 mm 2,3-butanedione monoxime were included to prevent cell contraction. Both approaches yielded a [Ca²⁺]_SR in heart failure myocytes that was ∼11% lower than in non-failing myocytes (see Fig. 5F). This 11% value was subsequently used to adjust fluo-5N fluorescence in heart failure myocytes to fluorescence in non-failing myocytes (see dashed line in Fig. 6).

Figure 3. During excitation–contraction coupling caffeine increases release junction recruitment and decreases the [Ca²⁺]_SR at which release terminates.

Confocal line scan images from the same cell of [Ca²⁺]_SR depletions during excitation–contraction coupling at low (A) and high (B) initial [Ca²⁺]_SR under control conditions (left) and after addition of 250 μm caffeine (right). F/F₀ profiles of fluorescence from two individual SR release junctions (denoted by black bars to left of images) are shown above (a) and below (b) images. F₀ corresponds to fluo-5N fluorescence at high load. Dashed lines denote [Ca²⁺]_SR level prior to field stimulus and at lowest level during release. Note that at low initial SR Ca²⁺ load, not all release junctions release Ca²⁺ (e.g. junction ii in Ab). Caffeine lowered the [Ca²⁺]_SR at release termination by 21 ± 2%; n= 15 events from 5 cells, P < 0.05.

Figure 4. Caffeine decreases the [Ca²⁺]_SR termination level of elementary Ca²⁺ release events.

A, cytosolic Ca²⁺ sparks (upper image and F/F₀ profile) and their corresponding intra-SR Ca²⁺ blinks (lower image and F/F₀ profile) in the absence and presence of 250 μm caffeine. F/F₀ profiles are normalized to baseline fluorescence under control conditions (F₀). Dashed lines represent initial [Ca²⁺]_SR, and [Ca²⁺]_SR at blink termination under control conditions and in the presence of caffeine. B, summary data of the relative changes (% increase versus control) in Ca²⁺ spark and Ca²⁺ blink amplitude (left), spatial width (FWHM, middle), and duration (FDHM, right) in the presence of caffeine. C, summary data showing effects of caffeine on Ca²⁺ blink termination level (black bars) at high (control, caffeine initial) and low (TG, caffeine late) [Ca²⁺]_SR. To decrease [Ca²⁺]_SR from control levels and match to SR Ca²⁺ load observed after sustained caffeine application (caffeine late), [Ca²⁺]_SR was experimentally decreased by adding the SERCA inhibitor thapsigargin (TG, 10 μm). D, summary data showing relationship between peak Ca²⁺ spark release flux and Ca²⁺ blink [Ca²⁺]_SR termination level in the absence (TG, Ctrl) and presence (Caff) of 250 μm caffeine. *P < 0.05.

Figure 6. The [Ca²⁺]_SR termination level of Ca²⁺ sparks is decreased in heart failure myocytes.

A, cytosolic Ca²⁺ sparks and intra-SR Ca²⁺ blinks from a normal myocyte (left image and F/F₀ profile) and heart failure myocyte (right image and F/F₀ profile). Resting fluorescence in heart failure myocytes was normalized to normal myocytes using the value shown in Fig. 5F. B, summary data showing the decreased Ca²⁺ blink [Ca²⁺]_SR termination level of 53.0 ± 1.9% (n= 11) in heart failure myocytes compared with 60.1 ± 0.9% (n= 82) of control myocytes. *P < 0.05 versus control.

Figure 5. Fractional SR Ca²⁺ release, Ca²⁺ spark frequency and Ca²⁺ spark amplitude are increased in heart failure myocytes.

A, example traces of [Ca²⁺]_SR during 0.5 Hz pacing followed by application of 10 mm caffeine in normal (at matched load) and heart failure myocytes. B, summary data showing the increased fractional SR Ca²⁺ release during 0.5 Hz pacing in heart failure myocytes when compared to load-matched normal myocytes. Load matching resulted in ΔF/F_Min values of 2.60 ± 0.25 (N= 8) and 2.59 ± 0.40 (N= 5) in normal and heart failure myocytes, respectively. C, example images and F/F₀ profiles of cytosolic Ca²⁺ sparks from normal (left), heart failure (middle) and load-matched normal (right) rabbit myocytes. D and E, summary data of Ca²⁺ spark frequency (D) and Ca²⁺ spark amplitude (E) from normal (N; n= 37), heart failure (HF; n= 36) and load-matched normal (N matched; n= 17) myocytes. F, SR Ca²⁺ content was 11% lower in permeabilized heart failure myocytes than in normal myocytes. *P < 0.05 versus normal myocytes, #P < 0.05 versus heart failure myocytes.

Data analysis and statistics

Global SR Ca²⁺ depletion amplitude (Δ[Ca²⁺]_SR) was calculated as [Ca²⁺]_SR,diastolic–[Ca²⁺]_SR,systolic, while SR fractional Ca²⁺ release was calculated as Δ[Ca²⁺]_SR/[Ca²⁺]_SR,diastolic. Ca²⁺ spark frequency (quantified as number of sparks per second and 100 μm of scanned distance; sparks × (100 μm)⁻¹× s⁻¹), amplitude, duration (FDHM), and spatial width (FWHM) were calculated using the SparkMaster algorithm (Picht et al. 2007). Ca²⁺ spark release flux was calculated from the first derivative of the rhod-2 cytosolic fluorescence during a Ca²⁺ spark (d(ΔF/F₀)/dt; ms⁻¹). For each detected Ca²⁺ spark, blink amplitude, duration (FDHM), and spatial width (FWHM) were calculated manually. In Fig. 2, SR Ca²⁺ load was matched between control and caffeine data groups by placing the data in 100 μm bins. For comparisons of normal and heart failure myocytes, cells paced at 0.5 Hz were load-matched by selecting normal myocytes that exhibited SR Ca²⁺ loads within the range of loads observed in heart failure myocytes. Normal permeabilized myocytes were treated with 10 μm thapsigargin to inhibit SERCA and decrease SR Ca²⁺ load to the level observed in heart failure myocytes. Statistical comparisons were made using Student's t test for paired or unpaired data, with statistical significance set at P < 0.05. Data are presented as individual observations or as the mean ±s.e.m.; n refers to the number of events (Ca²⁺ sparks and blinks), N refers to the number of cells.

Figure 2. Caffeine increases global SR Ca²⁺ release over a wide range of [Ca²⁺]_SR.

A, rest–decay protocol used to measure SR Ca²⁺ release during ECC at various [Ca²⁺]_SR. Electrical stimuli were applied at the arrows. B, example load-matched [Ca²⁺]_SR depletions in the presence (open circles in C,D) and absence (closed circles in C,D) of 250 μm caffeine. C and D, summary data showing effects of caffeine on depletion amplitude (C) and SR fractional release (D). Curves in C were fitted using exponential functions (r²: 0.93 (control), 0.84 (caffeine)). *Significant difference from the SR Ca²⁺ load-matched control value, P < 0.05.

Results

RyR sensitization with caffeine produces complex effects on SR Ca²⁺ release in paced myocytes

To characterize the mechanisms by which RyR sensitization alters SR Ca²⁺ release we loaded the SR of rabbit ventricular myocytes with the low affinity Ca²⁺ indicator fluo-5N. We used a low dose of caffeine (250 μm) to sensitize RyRs and examined its effects on ECC using simultaneous dynamic fluorescence measurements of [Ca²⁺]_i (using the high-affinity Ca²⁺ indicator rhod-2) and [Ca²⁺]_SR. As illustrated by the example traces of Fig. 1A, application of 250 μm caffeine to myocytes paced at 1 Hz produced an immediate increase in SR Ca²⁺ release, as shown by the increase in amplitude of both the [Ca²⁺]_i transient and the corresponding [Ca²⁺]_SR depletion. This was followed by a gradual decline in diastolic [Ca²⁺]_SR (Fig. 1A, lower trace) and a subsequent decrease in the amplitude of steady-state Ca²⁺ release. Upon wash out of caffeine, Ca²⁺ release was markedly depressed at the lower [Ca²⁺]_SR but then gradually recovered to the control level as the SR refilled to the original [Ca²⁺]_SR. Measurements of [Ca²⁺]_i using rhod-2 in intact myocytes (Fig. 1A) are hampered by accumulation of dye within mitochondria. Therefore, for quantitative measurements of [Ca²⁺]_i we performed separate measurements using fluo-4 (N= 8). Figure 1B,C presents summary data of [Ca²⁺]_i transients and [Ca²⁺]_SR depletions under control conditions, immediately following caffeine addition (caffeine, initial), and in the presence of caffeine after the new steady-state was achieved (caffeine, steady-state). In these experiments, intra-SR Ca²⁺ depletion amplitude (Fig. 1C) initially increased by 33%, but then subsequently decreased by 23% (versus control) as SR Ca²⁺ load decreased from 680 ± 110 μm to 492 ± 81 μm in the new steady-state with caffeine (N= 5, P < 0.05). These data indicate that sensitization of the RyR produces only transient effects on SR Ca²⁺ release due to a subsequent decline in SR Ca²⁺ load.

Figure 1. RyR sensitization with caffeine produces only transient positive inotropic effects due to unloading of [Ca²⁺]_SR.

A, simultaneous measurements of [Ca²⁺]_i (rhod-2, red image and upper trace) and [Ca²⁺]_SR (fluo-5N, green image and lower trace) before, during (grey shading), and after sensitization of RyRs with 250 μm caffeine. B and C, summary data showing initial and steady-state effects of 250 μm caffeine on [Ca²⁺]_i (B) and [Ca²⁺]_SR (C). Summary data in (B and C) are separate measurements using fluo-4 ([Ca²⁺]_i) and fluo-5N ([Ca²⁺]_SR). *Significant difference from control, P < 0.05.

Caffeine increases fractional SR Ca²⁺ release by altering release activation and termination

A single stimulus protocol was used to examine the effects of 250 μm caffeine over a wide range of initial [Ca²⁺]_SR. Rabbit ventricular myocytes exhibit the post-rest decay phenomenon. That is, SR Ca²⁺ content gradually declines after rest from steady-state pacing (Fig. 2A). We applied a single field stimulus after variable periods of rest to assess the amount of Ca²⁺ released at a given SR Ca²⁺ content (Fig. 2A). Consistent with previous reports (Guo et al. 2007), the amplitude of SR Ca²⁺ release displayed a non-linear dependence on diastolic [Ca²⁺]_SR (Fig. 2C, filled symbols). Sensitization of the RyR with caffeine resulted in more SR Ca²⁺ release at any given [Ca²⁺]_SR (Fig. 2C, open symbols), or leftward shifts in the SR Ca²⁺ release amplitude versus[Ca²⁺]_SR (Fig. 2C) and fractional SR Ca²⁺ release versus[Ca²⁺]_SR (Fig. 2D) relationships.

The increase in global fractional release (Fig. 2D) observed in the presence of caffeine could be explained by lowering the activation threshold for SR Ca²⁺ release (i.e. more release junctions fire in response to CICR), by increasing the synchronization of release junctions, or by increasing the amount of Ca²⁺ released at each individual release junction. To experimentally differentiate between these possibilities we monitored [Ca²⁺]_SR at individual release junctions during ECC in response to action potential stimulation. As shown in line-scan images of Fig. 3, individual release junctions were observed as regions of high fluo-5N fluorescence spaced ∼1.2 to 1.4 μm apart in the transverse direction of the cell. Fluorescence profiles during Ca²⁺ release from two such individual release junctions (a and b) are shown in Fig. 3 at low (A) and high (B) [Ca²⁺]_SR in the absence and presence of 250 μm caffeine. Under control conditions at low [Ca²⁺]_SR, release was inhomogeneous and [Ca²⁺]_SR depletion was not observed in all release junctions (Fig. 3A). At the same initial [Ca²⁺]_SR in the presence of 250 μm caffeine, release became homogeneous and well-synchronized (Fig. 3A, caffeine), and intra-SR depletions occurred at all release junctions. Qualitatively similar results of caffeine increasing release junction recruitment during CICR were observed in four additional cells. Strikingly, at the release junctions that fired under control conditions, the [Ca²⁺]_SR at which release terminated was lower in the presence of caffeine. At high [Ca²⁺]_SR, release was homogeneous under both control conditions and in the presence of caffeine (Fig. 3B), but caffeine again increased the amplitude of Ca²⁺ release within each release junction by lowering the [Ca²⁺]_SR level of release termination. At a given initial [Ca²⁺]_SR, there was junction-to-junction variability in the absolute [Ca²⁺]_SR level where release terminated, which can be seen by comparing profiles a and b at high [Ca²⁺]_SR (Fig. 3B). However, as [Ca²⁺]_SR was varied, the termination level for release remained constant within an individual release junction. This tendency was also observed in the presence of caffeine, albeit at an overall lower [Ca²⁺]_SR level. Summary data from 15 individual release junctions (5 cells) indicate that caffeine lowered the [Ca²⁺]_SR termination level for release during ECC by 21 ± 2% (P < 0.05). Therefore, based on these measurements it appears that RyR sensitization with caffeine increases release junction recruitment and synchronizes SR Ca²⁺ release, and also increases fractional release within the individual release junction.

Caffeine lowers the [Ca²⁺]_SR level of Ca²⁺ spark termination

Intra-SR Ca²⁺ measurements from individual release junctions during ECC must be interpreted with caution due to the signal summation from the many release junctions that fire and the rapid diffusion of Ca²⁺ within the SR network. To circumvent this issue and further examine the effects of caffeine on elementary Ca²⁺ release events from individual release junctions, we monitored spontaneous Ca²⁺ sparks and their corresponding Ca²⁺ blinks in permeabilized myocytes. This approach has the experimental advantage that single elementary Ca²⁺ release events are monitored in a precisely controlled cytosolic environment. As shown in Fig. 4A, application of 250 μm caffeine caused an immediate increase in Ca²⁺ spark frequency (by 80 ± 9.3%, N= 6) consistent with caffeine increasing the sensitivity of RyR release clusters to open at a given [Ca²⁺]_SR (Lukyanenko et al. 2001). More importantly, in contrast to Ca²⁺ blinks under control conditions which terminated at a [Ca²⁺]_SR level of 60.9 ± 2.6% (n= 16) of initial [Ca²⁺]_SR, caffeine caused the Ca²⁺ blink termination level to decrease to 53.5 ± 3.0% (n= 16, P < 0.05, Fig. 4A and C), or by 12%, similar to the effect observed in experiments during ECC in intact cells (Fig. 3). Figure 4B presents the changes in spatiotemporal properties of Ca²⁺ sparks and Ca²⁺ blinks immediately after caffeine addition at similar [Ca²⁺]_SR. Caffeine increased the amplitude of both Ca²⁺ sparks and blinks, but only increased the spatial width and duration of the Ca²⁺ sparks. We next examined if the shift in [Ca²⁺]_SR termination level was dependent on initial [Ca²⁺]_SR preceding the spark or [Ca²⁺] in the dyadic cleft during the spark. Caffeine caused a gradual decrease in resting [Ca²⁺]_SR, which allowed for monitoring of Ca²⁺ release events at various initial [Ca²⁺]_SR. To obtain [Ca²⁺]_SR-matched data points in the absence of RyR stimulation we added the SERCA inhibitor thapsigargin (TG, 10 μm), which blocked SR Ca²⁺ uptake and lowered [Ca²⁺]_SR. Ca²⁺ in the dyadic cleft can be reasonably approximated from release flux during a spark (Sipido & Wier, 1991; Song et al. 1998; Zima et al. 2008b), which progressively decreased as the gradient for Ca²⁺ release diminished at low [Ca²⁺]_SR. Figure 4C and D compares [Ca²⁺]_SR and peak release flux with Ca²⁺ blink termination level both under control conditions and in the presence of caffeine. As previously reported (Zima et al. 2008b), under control conditions and during SERCA inhibition, the Ca²⁺ blink termination level was independent of initial [Ca²⁺]_SR (Fig. 4C) or release flux (Fig. 4D). In the presence of caffeine, Ca²⁺ spark release flux was initially much greater than that of control Ca²⁺ sparks. As [Ca²⁺]_SR progressively decreased due to the increase in Ca²⁺ spark frequency, release flux also progressively declined (to near the initial control values). However, Ca²⁺ blinks still terminated at the ∼53%[Ca²⁺]_SR level (Fig. 4D). This suggests that caffeine simply resets the Ca²⁺ blink termination level to a new [Ca²⁺]_SR, an effect that is independent of initial [Ca²⁺]_SR or [Ca²⁺]_i in the dyadic cleft.

The [Ca²⁺]_SR level for Ca²⁺ spark termination is lowered during heart failure

The results using acute application of caffeine clearly show that RyR sensitization increases fractional SR Ca²⁺ release through effects on both release activation and release termination. Therefore, we hypothesized that similar effects may occur in heart failure, where myocytes exhibit an increase in fractional SR Ca²⁺ release at low pacing frequencies when SR Ca²⁺ load is experimentally matched (Shannon et al. 2003b; Guo et al. 2007). To test how SR Ca²⁺ release is altered during heart failure, we monitored [Ca²⁺]_i and [Ca²⁺]_SR in myocytes isolated from rabbits that had developed heart failure (Pogwizd, 1995). Similar to previous reports using this model (Shannon et al. 2003b; Guo et al. 2007), when diastolic [Ca²⁺]_SR was matched between normal (non-failing) and heart failure myocytes, fractional SR Ca²⁺ release was higher in the heart failure myocytes (Fig. 5A and B). We next measured Ca²⁺ sparks and blinks in permeabilized myocytes. This experimental strategy is advantageous as the L-type Ca²⁺ channel trigger for CICR may be altered during heart failure, either directly (Schroder et al. 1998) or due to changes in T-tubule architecture and coupling to SR Ca²⁺ release junctions (Gomez et al. 1997; He et al. 2001; Song et al. 2006; Lyon et al. 2009). In permeabilized heart failure myocytes, resting SR Ca²⁺ load and Ca²⁺ spark frequency were decreased compared to non-failing myocytes (Fig. 5D and F) while Ca²⁺ spark amplitude was similar between the two groups (Fig. 5E). However, it is problematic to directly compare these results due to the direct effects of [Ca²⁺]_SR on Ca²⁺ spark frequency and amplitude (Satoh et al. 1998; Lukyanenko et al. 2001). To appropriately compare Ca²⁺ spark frequency and amplitude, we matched [Ca²⁺]_SR in non-failing cells to their failing counterparts using SERCA inhibition with thapsigargin. When SR Ca²⁺ load was matched, both Ca²⁺ spark frequency and amplitude were markedly increased in failing myocytes (Fig. 5C–E). The increase in Ca²⁺ spark frequency suggests an increased sensitivity to release activation. We next examined if the increase in Ca²⁺ spark amplitude associated with a change in the [Ca²⁺]_SR level for release termination. Similar to results with caffeine, the [Ca²⁺]_SR termination level for Ca²⁺ sparks was decreased in heart failure (Fig. 6). As summarized in Fig. 6B, resting SR Ca²⁺ load was decreased in failing myocytes, and the [Ca²⁺]_SR termination level of Ca²⁺ sparks was lower in failing (53.0 ± 1.9%, n= 11) than in non-failing myocytes (60.1 ± 0.9%, n = 82, P < 0.05). This suggests that the modifications that occur in heart failure sensitize RyRs and increase fractional SR Ca²⁺ release by decreasing the [Ca²⁺]_SR level where release terminates. Collectively, these data show that RyR sensitization produces effects on both release activation and release termination, and each of these mechanisms may be altered during the progression of heart failure.

Discussion

The RyR is the principal SR Ca²⁺ release channel in the heart, and as such has become an experimental and potential therapeutic target to improve contractility in cardiac disease. However, as Eisner and colleagues have shown in rat ventricular myocytes, RyR sensitization with caffeine produces only transient positive inotropic effects, and does not lead to a sustained increase in SR Ca²⁺ release (Trafford et al. 1998). This phenomenon is explained by the initial increase in SR Ca²⁺ release shifting the cellular Ca²⁺ flux balance toward Ca²⁺ extrusion, causing eventual unloading of the SR and a decrease in the Ca²⁺ transient amplitude back to control levels. Our data from rabbit ventricular myocytes showed that SR Ca²⁺ loss was particularly pronounced, which led to a decrease in steady-state SR Ca²⁺ release (Fig. 1). The experimental differences between rat (‘autoregulation’ and return to steady-state release) and rabbit (marked SR Ca²⁺ loss and decrease in steady-state release) are likely to be due to the differences in expression of SERCA and NCX between these species (Bassani et al. 1994). Compared with rat, rabbit has less SERCA but increased NCX expression, which functionally will translate into less Ca²⁺ re-uptake into the SR via SERCA, more Ca²⁺ extrusion from the cell via NCX, and greater net cellular Ca²⁺ loss.

While RyR sensitization with caffeine did not produce a sustained increase in steady-state SR Ca²⁺ release, it became clear that when initial [Ca²⁺]_SR was matched, caffeine caused a large increase in fractional SR Ca²⁺ release (Fig. 2D). At a given [Ca²⁺]_SR, we observed two distinct functional mechanisms by which caffeine increased the amount of global Ca²⁺ release. First, RyRs were sensitized to CICR, the percentage of Ca²⁺ release units that opened during ECC was increased, and global Ca²⁺ release became highly synchronized (Fig. 3). This sensitization to release activation was also observed here, or in previous investigations, as an increase in spark frequency (Satoh et al. 1998; Lukyanenko et al. 2001) and Ca²⁺ wave occurrence (Lukyanenko et al. 1999). In addition, the increased sensitivity of RyRs in the presence of caffeine was further manifest in our experiments as an increase in Ca²⁺ spark flux at a given [Ca²⁺]_SR (Fig. 4D). This observation can be explained by several mechanisms, including the opening of more RyR channels within a release cluster or by synchronizing the opening of the RyR channels in the cluster to achieve more coordinated release.

In addition to the effect of caffeine on Ca²⁺ release activation, we found a second mechanism that explains the increase in global fractional release. At each release junction that opens during ECC, more Ca²⁺ is released due to a decrease in the [Ca²⁺]_SR termination level. This was observed at individual release junctions in response to electrical stimulation (Fig. 3) as well as during spontaneous Ca²⁺ sparks (Fig. 4). Like Ca²⁺ spark termination under control conditions, the Ca²⁺ spark [Ca²⁺]_SR termination level in the presence of caffeine was independent of Ca²⁺ spark release flux, which shows that it was independent of [Ca²⁺]_i in the dyadic cleft. These data, as well as results from other studies (Sobie et al. 2002; Terentyev et al. 2002; Hinch, 2004), suggest that [Ca²⁺]_i-dependent inactivation of the RyR is not the principal mechanism by which Ca²⁺ release terminates in cardiac myocytes. The precise mechanism by which caffeine lowers the [Ca²⁺]_SR termination level for release remains unclear, although it most likely reflects alterations in RyR channel or cluster gating by [Ca²⁺]_SR, as the set termination threshold observed under control conditions was simply ‘reset’ to a lower level in the presence of caffeine. Increasing evidence points to a luminal Ca²⁺ sensor on the RyR (Sitsapesan & Williams, 1994; Gyorke & Gyorke, 1998), and it has been suggested that caffeine exerts its effects on the RyR exclusively via altering luminal [Ca²⁺]_SR-dependent gating (Kong et al. 2008). Thus, changes in the [Ca²⁺]_SR-dependent gating mechanism would adequately explain the effects of caffeine on both the increased sensitivity for release activation and the decrease in the [Ca²⁺]_SR level for release termination.

The chronic modifications that occur in heart failure produce effects which are very similar to those caused by caffeine application. Clearly this type of comparison must be made with caution, as caffeine sensitizes RyRs acutely as a channel agonist while the RyR alterations that occur in heart failure progress over time, likely as post-translational modifications. These modifications remain highly controversial (Bers et al. 2003; Marks, 2003), with current evidence suggesting a role for protein kinase A (Marx et al. 2000), calmodulin-dependent protein kinase II (Ai et al. 2005), or reduction–oxidation processes (Terentyev et al. 2008a). Nevertheless, despite these caveats, important conceptual information can be gained by comparing the functional effects of caffeine with those observed in heart failure. Numerous Ca²⁺ transport systems are altered during the progression of heart failure, including changes in expression and/or function of the L-Type Ca²⁺ channel, NCX, SERCA, and the RyR (for review see Bers, 2006). Therefore, measurements of Ca²⁺ sparks in permeabilized myocytes are particularly useful for evaluating SR Ca²⁺ release, as the influence of the sarcolemma is eliminated, [Ca²⁺]_i and soluble components of the intracellular milieu are controlled, and [Ca²⁺]_SR is measured and controlled experimentally. SERCA activity does not contribute to the amplitude of the Ca²⁺ spark or blink (Zima et al. 2008b). Therefore, once SR Ca²⁺ content is experimentally matched, changes in Ca²⁺ sparks specifically reflect alterations in RyR release cluster function. Ca²⁺ sparks from failing myocytes exhibited alterations in both release activation and release termination, effects which were similar to those observed following RyR sensitization with caffeine. Ca²⁺ spark frequency, an index of basal sensitivity to release activation, was higher in failing than non-failing myocytes when SR Ca²⁺ load was matched (Fig. 5D) (Kubalova et al. 2005; Song et al. 2005; Lyon et al. 2009). Additionally, the [Ca²⁺]_SR level where Ca²⁺ sparks terminated was lower in heart failure myocytes (Fig. 6), an observation also seen in canine heart failure (Kubalova et al. 2005). In the intact cellular environment, the increase in Ca²⁺ spark frequency is assumed to be the primary contributing factor to the elevated diastolic SR Ca²⁺ leak in heart failure (Shannon et al. 2003b). However, at a given Ca²⁺ spark frequency, decreasing the [Ca²⁺]_SR termination level for release will further enhance SR Ca²⁺ loss. Increasing fractional release during systole will also increase total SR Ca²⁺ loss, as is clearly evident by the effects of acute caffeine application (Fig. 1). This effect will be particularly pronounced when powerful sarcolemmal Ca²⁺ extrusion mechanisms exist, such as in species with high expression of NCX (Bassani et al. 1994) or when NCX is upregulated during heart failure (Pogwizd et al. 1999). Together, these alterations explain the rather paradoxical association of increased fractional SR Ca²⁺ release with decreased total SR Ca²⁺ release in the failing heart.

Our data functionally separate the mechanisms of release activation and release termination. Implicit in the logic of this investigation is that these functional observations correspond to distinct RyR gating mechanisms, which is still unclear. If release activation and release termination are distinct mechanisms, the key remaining question is whether they can be regulated separately. Although the decrease in [Ca²⁺]_SR termination level is to some extent counterproductive as it contributes to loss of [Ca²⁺]_SR, it is undoubtedly a key means by which systolic fractional SR Ca²⁺ release is enhanced at the lower [Ca²⁺]_SR. Thus, one may speculate that decreasing the termination level for release without concomitant effects on release activation would be an ideal positive inotropic intervention, as systolic fractional SR Ca²⁺ release would be enhanced without substantial diastolic SR Ca²⁺ leak.

In summary, we find that RyR sensitization has multiple effects on global SR Ca²⁺ release. First, it alters RyR activation, increases recruitment of release junctions and synchronizes release during ECC. Furthermore, at each individual release junction, RyR sensitization lowers the [Ca²⁺]_SR level of release termination and increases the amplitude of SR Ca²⁺ release. Alteration of the [Ca²⁺]_SR termination level for release is therefore likely to serve as an additional mechanism by which ECC can be tuned to match the physiological needs of the heart.

Glossary

Abbreviations

CICR

calcium-induced calcium release

ECC

excitation–contraction coupling

FDHM

full-duration half-maximum

FWHM,

full-width half-maximum

HF

heart failure

NCX

Na⁺/Ca²⁺ exchanger

RyR

ryanodine receptor

SERCA

sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase

SR

sarcoplasmic reticulum

TG

thapsigargin

Author contributions

All authors contributed to the conception and design of the study, analysis and interpretation of data, and writing of the article. All authors have approved the version to be published. T.L.D. and A.V.Z. performed the experimental work.