Caffeine Induces Ca²⁺ Release by Reducing The Threshold for Luminal Ca²⁺ Activation of the Ryanodine Receptor

Abstract

Caffeine has long been used as a pharmacological probe for studying ryanodine receptor (RyR)-mediated Ca²⁺ release and cardiac arrhythmias. However, the precise mechanism by which caffeine activates RyRs is elusive. Here we investigated the effects of caffeine on spontaneous Ca²⁺ release and on the response of single cardiac RyR (RyR2) channels to luminal or cytosolic Ca²⁺. We found that HEK293 cells expressing RyR2 displayed partial or “quantal” Ca²⁺ release in response to repetitive additions of submaximal concentrations of caffeine. This quantal Ca²⁺ release was abolished by ryanodine. Monitoring of endoplasmic reticulum luminal Ca²⁺ revealed that caffeine reduced the luminal Ca²⁺ threshold at which spontaneous Ca²⁺ release occurs. Interestingly, spontaneous Ca²⁺ release in the form of Ca²⁺ oscillations persisted in the presence of 10 mM caffeine, and was diminished by ryanodine, demonstrating that unlike ryanodine, caffeine, even at high concentrations, does not hold the channel open. At the single channel level, caffeine markedly reduced the threshold for luminal Ca²⁺ activation, but had little effect on the threshold for cytosolic Ca²⁺ activation, indicating that the major action of caffeine is to reduce the luminal, but not the cytosolic, Ca²⁺ activation threshold. Furthermore, as with caffeine, the clinically relevant, pro-arrhythmic methylxanthines aminophylline and theophylline potentiated luminal Ca²⁺ activation of RyR2, and increased the propensity for spontaneous Ca²⁺ release, mimicking the effects of diseased-linked RyR2 mutations. Collectively, our results demonstrate that caffeine triggers Ca²⁺ release by reducing the threshold for luminal Ca²⁺ activation of RyR2, and suggest that disease-linked RyR2 mutations and RyR2-interacting pro-arrhythmic agents may share the same arrhythmogenic mechanism.

INTRODUCTION

A number of naturally-occurring mutations in the cardiac Ca²⁺ release channel/ryanodine receptor (RyR2) have been linked to at least two forms of cardiac arrhythmias: catecholaminergic polymorphic ventricular tachycardia (CPVT) and arrhythmogenic right ventricular displaysia type 2 (ARVD2), but their causal mechanisms have not been completely defined [1]. We have recently shown that disease-causing RyR2 mutations enhance the sensitivity of the channel to activation by luminal Ca²⁺ and reduce the threshold for spontaneous Ca²⁺ release, also known as store-overload-induced Ca²⁺ release (SOICR)[2,3]. It is well known that spontaneous Ca²⁺ release can alter membrane potential by generating delayed afterdepolarizations (DADs), which in turn can lead to triggered arrhythmias [4]. Alternatively, it has also been proposed that disease-linked RyR2 mutations alter protein-protein or inter-domain interactions that are important for stabilizing the closed state of the channel, thus rendering the channel hyperactive and leaky [5,6].

In addition to RyR2 mutations, RyR2-interacting drugs, including caffeine and other methylxanthines (aminophylline and theophylline), have been shown to promote catecholamine-induced arrhythmias, but by itself caffeine is not arrhythmogenic and does not have a sustained impact on stimulated Ca²⁺ release [7–14]. This pro-arrhythmic characteristic of caffeine resembles that of the RyR2 CPVT mutations, which predispose patients to exercise or stress-induced cardiac arrhythmias, but cause no obvious structural or functional cardiac defects at rest [1]. These observations suggest that caffeine and CPVT mutations may affect the RyR2 channel in a similar manner. Consistent with this hypothesis, caffeine has been shown to reduce the threshold for spontaneous Ca²⁺ release [2,7,15]. However, exactly how caffeine reduces the threshold for spontaneous Ca²⁺ release is not well understood.

Caffeine has commonly been used as an RyR agonist for inducing Ca²⁺ release from intracellular Ca²⁺ stores [16–20]. A unique feature of caffeine-induced Ca²⁺ release from RyR-gated Ca²⁺ stores is its lack of desensitization. Multiple additions of caffeine at submaximal concentrations can each induce a partial and transient Ca²⁺ release from intracellular Ca²⁺ stores in cells expressing RyRs or from sarcoplasmic reticulum membrane vesicles [17–20], a phenomenon known as “quantal” Ca²⁺ release. The partial or quantal Ca²⁺ release induced by incremental concentrations of caffeine was thought to result from the sequential activation of different populations of RyRs expressed in the same cell with different sensitivities to caffeine in an all-or-none fashion [19,20]. However, it has also been shown that the ability of caffeine to induce Ca²⁺ release is dependent on the store Ca²⁺ content [17,21,22]. When the store Ca²⁺ level is below a threshold level, caffeine is no longer able to induce Ca²⁺ release despite its continued presence. Hence, the partial or quantal Ca²⁺ release is believed to result from store-dependent negative feedback regulation of caffeine activation of the channel [17,22]. Similar to the phenomenon of quantal Ca²⁺ release, caffeine at low concentrations has also been shown to only transiently potentiate stimulated Ca²⁺ release in cardiac cells [15]. This transient effect of caffeine is believed to be due to the auto-regulation of SR Ca²⁺ release by the SR Ca²⁺ content [23]. Furthermore, Ca²⁺ release studies using SR membrane vesicles have also shown that a certain level of store Ca²⁺ content must be present before caffeine-induced Ca²⁺ release can occur [24–26]. Together, these observations clearly indicate that SR luminal Ca²⁺ plays an important role in the action of caffeine, but the molecular basis of this luminal Ca²⁺ dependence is unclear.

Caffeine is commonly thought to sensitize the RyR2 channel to activation by cytosolic Ca²⁺, leading to an increase in the open probability (Po) of the channel [27,28]. An enhanced Po of RyR2 would result in a decreased SR Ca²⁺ content, which would, in turn, reduce the Po of RyR2. As a result of this counteractive reduction in luminal Ca²⁺, caffeine, despite its continued presence, only causes a transient effect on SR Ca²⁺ release [15]. However, recent studies revealed that single RyR2 channels are rather insensitive to caffeine in the absence of luminal Ca²⁺ [29]. Alternatively, since RyR2 is also regulated by luminal Ca²⁺ [30–32], caffeine may alter the response of the channel to luminal Ca²⁺. Based on our observation that CPVT RyR2 mutations reduce the threshold for spontaneous Ca²⁺ release by increasing the sensitivity of the channel to luminal Ca²⁺ activation, we reasoned that caffeine, which also reduces the threshold for spontaneous Ca²⁺ release, might sensitize the channel to luminal Ca²⁺ activation. To test this hypothesis, in the present study we investigated the impact of caffeine on the sensitivity of single RyR2 channels to activation by cytosolic or luminal Ca²⁺. We found that caffeine preferentially potentiated the luminal Ca²⁺ activation of RyR2 at low cytosolic Ca²⁺ concentrations. Similar effects were also observed with two other methylxanthines, aminophylline and theophylline. These observations suggest that the pro-arrhythmic action of clinically relevant methylxanthines likely results from their luminal Ca²⁺ activating properties.

EXPERIMENTAL PROCEEDURES

Materials

Soybean phosphatidylcholine, heart phosphatidylethanolamine, and brain phosphatidylserine were obtained from Avanti Polar Lipids, Inc (Alabaster, AL). [³H]ryanodine was from PerkinElmer Life Sciences. CHAPS and other reagents were purchased from Sigma.

Ca²⁺ release measurements

The free cytosolic Ca²⁺ concentration in transfected HEK293 cells was measured using the fluorescent Ca²⁺ indicator dye fluo-3-AMas described previously [33].

Generation of stable, inducible HEK293 cells

HEK293 cells expressing RyR2 (wt) were generated and characterized previously [2].

Single cell Ca²⁺ imaging (luminal Ca²⁺)

Luminal Ca²⁺ transients in HEK293cells expressing RyR2 were measured using single-cell Ca²⁺ imaging and the Ca²⁺ sensitive FRET-based cameleon protein D1ER [34]. Stable, inducible HEK293 cells expressing RyR2 were used, but were additionally transfected, using the Ca²⁺ phosphate precipitation method, with D1ER cDNA 24 hr before RyR2 expression was induced. The cells were continuously perfused with Krebs-Ringer- Hepes (KRH) buffer (125 mM NaCl, 5 mM KCl, 1.2 mM KH2PO4, 6 mM glucose, 1.2 mM MgCl2, 25 mM Hepes, pH 7.4) containing 2 mM CaCl₂, various concentrations of caffeine (0, 0.3, 1, 10 mM) and in the presence or absence of 20 μM ryanodine at room temperature. Images were captured with Compix Inc. Simple PCI 6 software at 470 nm and 535 nm emission, with excitation at 430 nm, every 2 s using an inverted microscope (Nikon TE2000-S) equipped with an S-Fluor 20x/0.75 objective. The amount of FRET was determined from the ratio of the emissions at 535 and 470 nm.

Single channel recordings

Single-channel analyses were carried out as described previously [3]. Briefly, RyR2 proteins were partially purified from cell lysate by sucrose density gradient centrifugation. Heart phosphatidylethanolamine and brain phosphatidylserine (Avanti Polar Lipid), dissolved in chloroform, were combined in a 1:1 ratio (w/w), dried under nitrogen gas and suspended in 30 μl of n-decane at a concentration of 12 mg lipid/ml. Bilayers were formed across a 250-μm hole in a Delrin partition separating two chambers. The trans chamber (800 μl) was connected to the head stage input of an Axopatch 200A amplifier (Axon Instruments Inc.). The cis chamber (1.2 ml) was held at virtual ground. A symmetrical solution containing 250 mM KCl and 25 mM Hepes (pH 7.4), was used for all recordings, unless indicated otherwise. A 4 μl aliquot (~1 μg of protein) of the sucrose density gradient-purified RyR2 protein was added to the cis chamber. Spontaneous channel activity was always tested for sensitivity to EGTA and Ca²⁺. Only those single channels that are EGTA sensitive and display a stable open probability were used for analyses. The chamber to which the addition of EGTA inhibited the activity of the incorporated channel was presumed to correspond to the cytosolic side of the channel. The direction of single channel currents was always measured from the luminal to the cytosolic side of the channel, unless mentioned otherwise. Recordings were filtered at 2,500 Hz. Free Ca²⁺ concentrations were calculated using the computer program of Fabiato and Fabiato [35]. Data analyses were carried out using the pClamp 8.1 software (Axon Instruments Inc.).

Single cell Ca²⁺ imaging of HEK293 cells

Intracellular Ca²⁺ transients in stable, inducible HEK293 cells expressing RyR2 were measured using single-cell Ca²⁺ imaging and the fluorescent Ca²⁺ indicator dye fura-2 acetoxymethyl ester (fura-2 AM) as described previously [2]. Cells grown on glass coverslips for 24 hr after induction by 1 μg/ml tetracycline were loaded with 5 μM fura-2 AM in Krebs- Ringer- Hepes (KRH) buffer (125 mM NaCl, 5 mM KCl, 1.2 mM KH2PO4, 6 mM glucose, 1.2 mM, MgCl₂, 25 mM Hepes, pH 7.4) plus 0.02% pluronic F-127 (Molecular Probes) and 0.1 mg/ml BSA for 20 min at room temperature. The coverslips were then mounted in a perfusion chamber (Warner Instruments, Hamden, CT) on an inverted microscope (Nikon TE2000-S) equipped with an S-Fluor 20x/0.75 objective. The cells were continuously perfused with KRH buffer containing 0, 0.1, 0.2, 0.3, 0.5, 1 mM CaCl₂ and 0.3 mM caffeine, aminophylline or theophylline at room temperature. 10 mM caffeine was applied at the end of each experiment. Time-lapse images (0.25 frames s⁻¹) were captured and analyzed with the Compix Inc. Simple PCI 6 software. Fluorescent intensities were measured from regions of interest centered on individual cells. Only those cells that responded to caffeine were used in analysis (60–80%).

Isolation of adult rat ventricular myocytes

All studies with rats were approved by the Animal Care Committee of the University of Calgary and complied with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication no. 85-23, revised 1996). Single rat ventricular myocytes were isolated as described previously [36]. Isolated cells were stored at room temperature in a solution containing 20 mM taurine, 5 mg/ml albumin, and 0.5 mM CaCl_2, until used for single cell Ca²⁺ imaging studies.

Single cell Ca²⁺ imaging of rat ventricular myocytes

Freshly isolated rat ventricular myocytes were placed on glass coverslips coated with 0.02% (w/v) gelatin and 10 μg/ml fibronectin, and loaded with 5 μM fluo-4-AM Ca²⁺ (Molecular Probes) plus 0.02% pluronic F-127 in Krebs-Ringer-Hepes (KRH) buffer (125 mM NaCl, 5 mM KCl, 6 mM glucose, 1.2 mM MgCl₂ and 25 mM Hepes, pH 7.4) (without KH₂PO₄) in the presence of 1.0 mM Ca²⁺ for 20 min at room temperature. The coverslips were mounted in a perfusion chamber on an inverted microscope (Nikon TE2000-S) equipped with an S-Fluor 20x/0.75 objective. The [Ca²⁺] was then stepped to 5 mM for 5 min before further increasing it to 10 mM. The cells were then continuously perfused with KRH buffer containing 10 mM CaCl₂ at room temperature in the absence and presence of 0.5 mM caffeine, aminophylline, or theophylline. Time-lapse images were captured every ~1.3 sec, during the excitation periods, and analyzed using Compix Inc. Simple PCI 6 software.

[³H]Ryanodine binding

Equilibrium [³H]ryanodine (NEN Life Science) binding to cell lysate was performed as described previously [33]. Briefly, a binding mixture (300 μl) containing 30 μl of cell lysate (3–5 mg/ml), 25 mM Tris/50 mM Hepes (pH 7.4), 5 nM [³H]ryanodine, a protease inhibitor mix, and various concentrations of CaCl₂, KCl and 2.5 mM caffeine, aminophylline or theophylline as indicated, was incubated at 37°C for 2.5–3.5 hr. The binding mixture was diluted with 5 ml of ice-cold washing buffer containing 25 mM Tris (pH 8.0), and 250 mM KCl, and immediately filtered through Whatman GF/B filters presoaked with 1% polyethylenimine. The filters were washed four times with 5 ml of ice-cold washing buffer and the radioactivity associated with the filters was determined by liquid scintillation counting. Nonspecific binding was determined by measuring [³H]ryanodine binding in the presence of 50 μM unlabeled ryanodine. All binding assays were done in duplicate. Data shown are mean ± SEM for n experiments. Statistical significance was evaluated using the unpaired Student’s t test. A P value of 0.05 is considered to be statistically significant.

RESULTS

Caffeine induces “quantal” Ca²⁺ release in HEK293 cells expressing RyR2

Partial or quantal Ca²⁺ release in response to incremental concentrations of caffeine has been observed with native cells expressing RyRs and SR vesicles isolated from skeletal muscle [17–20]. To determine whether partial or quantal Ca²⁺ release occurs in a heterologous expression system, we assessed the response of HEK293 cells transfected with mouse RyR2 cDNA to multiple additions of caffeine. As shown in Fig. 1A, the addition of 0.2 mM caffeine induced a transient Ca²⁺ release in HEK293 cells expressing RyR2. In the continued presence of caffeine, a second addition of 0.2 mM caffeine was able to trigger another transient Ca²⁺ release in these cells. This partial Ca²⁺ release was clearly observed even after the 7^th consecutive addition of 0.2 mM caffeine, although the amplitude of each Ca²⁺ release was progressively reduced. Hence, partial or quantal Ca²⁺ release also occurs in a heterologous expression system, suggesting that the quantal nature of caffeine-induced Ca²⁺ release reflects an intrinsic property of its activation of RyR2.

Fig. 1. Caffeine-induced “quantal” Ca²⁺ release in HEK293 cells expressing RyR2.

HEK293 cells expressing RyR2 were loaded with 5 μM fluo-3 AM. (A) The fluorescent intensity of the cells was monitored continuously before and after repeated additions of 0.2 mM caffeine. (B) The fluorescent signals from the fluo-3-loaded cells were measured before and after the addition of ryanodine (100 μM) followed by the repeated additions of caffeine (0.2 mM). Traces shown are from representative experiments that have been repeated three times with similar results.

The partial Ca²⁺ release induced by a submaximal concentration of caffeine (Fig. 1A) could be due to the opening of a subpopulation of the RyR2 channels. To test this hypothesis, we pretreated HEK293 cells expressing RyR2 with 100μM ryanodine before multiple additions of caffeine. Since ryanodine only binds to the open channel and the binding of ryanodine converts the channel to a fully open state [37,38], the ryanodine-modified channel is no longer sensitive to caffeine. If the first addition of 0.2 mM caffeine only activates a subpopulation of RyR2, one would expect that the ryanodine pretreatment could only modify that subpopulation of RyR2 that was opened by the first addition of caffeine, and that HEK293 cells pretreated with ryanodine would still respond to multiple additions of caffeine, as each addition of caffeine would activate a new subpopulation of channels. In contrast to this prediction, we found that cells pretreated with ryanodine only responded to the first addition of caffeine.

As shown in Fig. 1B, in the absence of caffeine, the addition of ryanodine caused a slow release of Ca²⁺. This is likely due to the binding of ryanodine to a small population of RyR2 channels that are open under these conditions and consequently an increase in Po of these channels. This slow release of Ca²⁺ would be equilibrated at some point with Ca²⁺ uptake into the endoplasmic reticulum (ER) or Ca²⁺ extrusion into the extracellular space, leading to a steady-state cytosolic Ca²⁺ level corresponding to the plateau in the fluorescent signal. The subsequent addition of 0.2 mM caffeine activated the remaining ryanodine-unmodified RyR2 channels and caused a large Ca²⁺ release. The caffeine-activated channels would then be modified by ryanodine into a fully activated state, leading to a depletion of intracellular Ca²⁺ store. The released Ca²⁺ would be extruded into the extracellular space, resulting in a transient Ca²⁺ release. Importantly, unlike those seen in Fig. 1A, 6 subsequent additions of caffeine yielded little or no Ca²⁺ release. The overall decline of fluorescent signals throughout the recordings is due to quenching of the fluo-3 fluorescent dye by caffeine. This caffeine-dependent quenching can clearly be seen in Fig. 1B, where every addition of caffeine caused an immediate drop in the fluorescent signal, while the fluorescent signals between two additions of caffeine are relatively constant. Furthermore, due to the difference in the release kinetics, the amplitudes of ryanodine-induced Ca²⁺ release and caffeine-induced Ca²⁺ release under these different conditions may not be directly comparable. The difference in the decay kinetics of the caffeine-induced Ca²⁺ transients in the presence and absence of ryanodine is likely the result of the modification of the RyR2 channel by ryanodine. These observations suggest that the first addition of 0.2 mM caffeine was able to open nearly all the RyR2 channels, which were subsequently converted by ryanodine into a fully open state and thus became unresponsive to further additions of caffeine. Hence, the partial or quantal Ca²⁺ release in HEK293 cells transfected with a single class of RyR2 cDNA is unlikely to be due to the existence of different populations of RyR2 with various caffeine sensitivities.

Fig. 6. Methylxanthines enhance spontaneous Ca²⁺ release in rat cardiomyocytes.

Rat cardiac myocytes attached to gelatin-fibronectin pre-treated glass coverslips were loaded with 5 μM fluo-4 AM. The cells were then continuously perfused with 10 mM [Ca²⁺]_o plus 0.5 mM caffeine (A), aminophylline (B) or theophylline (C). Fluo-4 fluorescent intensities of representative myocytes are shown. (D) The frequency (white) and amplitude (grey) of spontaneous Ca²⁺ waves (%, mean ± SEM) in rat cardiac myocytes in the absence of methylxanthines (control) or presence of caffeine, aminophylline or theophylline are shown. The frequency and amplitude of spontaneous Ca²⁺ release observed for the control was normalized to 100%. The total numbers of cardiac myocytes analyzed for spontaneous Ca²⁺ release were 9 for caffeine, 10 for aminophylline and 16 for theophylline from 3 separate experiments.

Caffeine reduces the threshold at which spontaneous Ca²⁺-release occurs

It has been shown that caffeine-induced Ca²⁺ release is dependent on the ER/SR luminal Ca²⁺ concentration [17,21,22]. To further investigate the luminal Ca²⁺ dependence of caffeine activation, we directly monitored the ER luminal Ca²⁺ dynamics in HEK293 cells expressing RyR2 before and after the additions of various concentrations of caffeine using a luminal Ca²⁺ indicator protein (D1ER). HEK293 cells expressing RyR2 were transfected with D1ER, a soluble fluorescence resonance energy transfer (FRET) based Ca²⁺ indicator protein that is expressed within the lumen of the ER due to a KDEL retention motif [34]. As seen in Fig. 2A, in the absence of caffeine and the presence of 2 mM external Ca²⁺, HEK293 cells expressing RyR2 displayed spontaneous Ca²⁺ release, which is reflected by the transient downward deflections in the FRET signal, similar to results reported previously [39]. It is worth noting that spontaneous Ca²⁺ release occurs when the ER Ca²⁺ reaches a certain level (represented by a dash-line at 0 mM caffeine). We referred to this luminal Ca²⁺ level as the luminal Ca²⁺ threshold at which spontaneous Ca²⁺ release occurs. In the presence of 0.3 mM caffeine, the luminal Ca²⁺ level (represented by a dash-line at 0.3 mM caffeine) at which spontaneous Ca²⁺ release occurred was reduced to 86.3 ± 0.9% (n = 34, P < 0.001) of that in the absence of caffeine (Fig. 2B). Similarly, after perfusing the cells with 1 mM caffeine, the luminal Ca²⁺ threshold (represented by a dash-line at 1mM caffeine) at which spontaneous Ca²⁺ release occurred was further reduced to 61.1 ± 1.5% (n = 34, P < 0.001) of that in the absence of caffeine (Fig. 2B). Interestingly, even in the presence of 10 mM caffeine, spontaneous Ca²⁺ release in the form Ca²⁺ oscillations still persisted despite a markedly reduced luminal Ca²⁺ threshold (21.4 ± 1.1%, n = 34, P < 0.001) (Fig. 2B). This observation indicates that the RyR2 channel is activated only when the luminal Ca²⁺ reaches a threshold level, even in the presence of 10 mM caffeine. On the other hand, the addition of 20μM ryanodine abolished Ca²⁺ oscillations. Ryanodine is known to dramatically sensitize the channel to cytosolic Ca²⁺ activation and convert the channel to a persistent activated state [37,38]. As a result, the ER Ca²⁺ store would be completely depleted in the presence of 10 mM caffeine and 20 μM ryanodine. Taken together, these observations demonstrate that unlike ryanodine, caffeine, even at high concentrations, does not always open the RyR2 channel, and that the action of caffeine is to reduce the luminal Ca²⁺ threshold at which spontaneous Ca²⁺ release occurs. The fact that the amplitude of spontaneous Ca²⁺ oscillations is also reduced as the caffeine concentration is increased further supports this view. This is because a reduced threshold at which spontaneous Ca²⁺ release occurs will reduce the maximal SR Ca²⁺ loading as Ca²⁺ will be released from the SR when it reaches the threshold level. A reduced level of SR Ca²⁺ loading will, in turn, decrease the amount of Ca²⁺ release and thus the amplitude of Ca²⁺ oscillations.

Fig. 2. Caffeine reduces the luminal Ca²⁺ threshold level at which spontaneous Ca²⁺ release occurs.

HEK293 cells expressing RyR2 were grown on glass coverslips. Cells were transfected with D1ER cDNA 48 hr before imaging and RyR2 expression was induced 24 hr before imaging. The cells were perfused with KRH buffer containing 2 mM Ca²⁺ and 0, 0.3, 1 or 10 mM caffeine with or without 20μM ryanodine. (A) A representative trace captured using single cell imaging. The dash-lines illustrate the relative luminal Ca²⁺ threshold for spontaneous Ca²⁺ release at each concentration of caffeine. (B) The relative luminal Ca²⁺ threshold for spontaneous Ca²⁺ release at various caffeine concentrations. The threshold for spontaneous Ca²⁺ release is expressed as a percentage of the threshold in the absence of caffeine. Data represents the mean ± SEM of 34 cells from 3 separate experiments.

Caffeine preferentially sensitizes the luminal Ca²⁺ activation of RyR2 at low cytosolic Ca²⁺ concentrations

To understand how caffeine reduces the threshold for spontaneous Ca²⁺ release, we assessed the impact of caffeine on single RyR2 channels. As shown in Fig. 3A, a single RyR2 channel exhibited little activity at low luminal (45 nM) and cytosolic (45 nM) Ca²⁺ concentrations. Increasing the luminal Ca²⁺ to 300 μM slightly activated the channel (Fig. 3B). A subsequent addition of 2 mM caffeine to the cytosolic side of the channel markedly increased the channel activity (Fig. 3C). The average Po after the addition of caffeine was 0.094 ± 0.025 in the presence of 300 μM luminal Ca²⁺, which was significantly greater than that before the addition of caffeine (0.006 ± 0.001) (n = 4) (P < 0.02). Importantly, this caffeine activation was dependent on luminal Ca²⁺. Reducing the luminal Ca²⁺ from 300 μM to ~45 nM decreased the activity of the channel to the basal level with Po of 0.002 ± 0.001 (n = 4) (P < 0.05) (Fig. 3D). These data indicate that caffeine preferentially potentiates the luminal Ca²⁺ response of RyR2 at low cytosolic Ca²⁺ levels.

Fig. 3. Caffeine enhances the response of single RyR2 channels to luminal Ca²⁺ activation.

Single-channel activities of RyR2 were recorded in a symmetrical recording solution containing 250 mM KCl and 25mM HEPES (pH 7.4) at a holding potential of −20 mV. EGTA was added to either the cis or trans chamber to determine the orientation of the incorporated channel. The side of the channel to which an addition of EGTA inhibited the activity of the incorporated channel presumably corresponds to the cytosolic face. The Ca²⁺ concentration on both the cytosolic and luminal sides of the incorporated channel was first adjusted to ~46 nM (A). The channel was activated by 300 μM luminal Ca²⁺ (B). Caffeine was then added to the cytosolic side of the channel in the presence of 300 μM luminal Ca²⁺ (C), followed by a decrease of luminal Ca²⁺ to ~46 nM (D). Openings are downward. Open probability (Po), arithmetic mean open time (To), and arithmetic mean closed time (Tc) are indicated at the top of each panel. A short line to the right of each current trace indicates the baseline. A continuous recording is shown. The average recording time for each condition shown in panels A–D from 4 channels is 103 s. The relationship between Po and luminal Ca²⁺ concentration is shown in E, and the relationship between Po and cytosolic Ca²⁺ concentration is shown in F. Data points shown in E are means ± SEM from 5 RyR2 channels in the presence of 2 mM caffeine (solid circles) and 8 RyR2 channels in the absence of caffeine (open circles), and those shown in F are individual measurements obtained from 7 RyR2 channels in the presence of 2 mM caffeine (solid circles) and 5 RyR2 channels in the absence of caffeine. The average recording time is 107 s for E and 93 s for F.

To further characterize the luminal and cytosolic Ca²⁺ dependence of caffeine activation, we determined the effect of caffeine on the sensitivity of single RyR2 channels to activation by luminal or cytosolic Ca²⁺. As shown in Fig. 3E, in the absence of caffeine, single RyR2 channels were activated by luminal Ca²⁺ with a threshold of ~3 mM (n =8), similar to that shown previously [40]. In the presence of 2 mM caffeine, single RyR2 channels were much more sensitive to activation by luminal Ca²⁺. Caffeine markedly reduced the threshold for luminal Ca²⁺ activation to ~0.1 mM (n =5) (Fig. 3E). On the other hand, caffeine (2 mM) only slightly reduced the EC₅₀ for activation of the RyR2 channel by cytosolic Ca²⁺ from 0.31 μM (n=5) to 0.17 μM (n = 7) (Fig. 3F). It should be noted that caffeine has little effect on the threshold for activation by cytosolic Ca²⁺ (~100 nM in the presence and absence of caffeine) (Fig. 3F). These data suggest that at low cytosolic and high luminal Ca²⁺ concentrations, a condition resembling that seen in resting cells, caffeine preferentially sensitizes the RyR2 channel to activation by luminal Ca²⁺.

Other methylxanthines also potentiate the response of RyR2 to luminal Ca²⁺

Caffeine is a member of the methylxanthine family of compounds. To determine whether other methylxanthines also preferentially sensitize the luminal Ca²⁺ activation of RyR2, we assessed the effect of the clinically relevant methylxanthines aminophylline and theophylline on the luminal Ca²⁺ response of single RyR2 channels. As shown in Fig. 4A, aminophylline (2 mM) markedly activated the RyR2 channel in the presence of 45 nM cytosolic Ca²⁺ and 300 μM luminal Ca²⁺. The average Po was significantly increased after the addition of aminophylline, increasing from 0.015 ± 0.008 to 0.102 ± 0.026 (n = 6, P < 0.02). As seen with caffeine, this aminophylline-induced enhancement of channel activity depends on luminal Ca²⁺. Reducing the luminal Ca²⁺ concentration to 45 nM abolished the effect of aminophylline by decreasing the Po to 0.002 ± 0.0003 (n = 6, P < 0.02) (Fig. 4Ad). Fig. 4B shows the impact of theophylline. Again and as with caffeine and aminophylline, theophylline activated single RyR2 channels in a luminal Ca²⁺ dependent manner. The average Po was significantly augmented by theophylline from 0.009 ± 0.003 to 0.241 ± 0.03 (n = 5, P < 0.01), and was reduced to the basal level (Po = 0.005 ± 0.003, n = 5, P < 0.01) when luminal Ca²⁺ was removed. Therefore, like caffeine, aminophylline and theophylline also preferentially sensitize the RyR2 channel to luminal Ca²⁺ activation at low cytosolic Ca²⁺ concentrations.

Fig. 4. Aminophylline and theophylline enhance the luminal Ca²⁺ activation of RyR2.

The effects of aminophylline (A) and theophylline (B) on single-channel activities of RyR2 were recorded in a symmetrical recording solution as described in Fig. 3. The Ca²⁺ concentration on both the cytosolic and luminal sides of the incorporated channel was first adjusted to ~46 nM (Aa, Ba). The channel was activated by 300 μM luminal Ca²⁺ (Ab, Bb). Aminophylline (Ac) or theophylline (Bc) was then added to the cytosolic side the channel in the presence of 300 μM luminal Ca²⁺ followed by a decrease of luminal Ca²⁺ to ~46 nM (Ad, Bd). A similar luminal Ca²⁺ dependence of activation by aminophylline or theophylline was observed in 5–6 RyR2 channels. A continuous recording is shown for each condition. The average recording time is 134 s for panel A and 96 s for panel B.

Methylxanthines increase the propensity for spontaneous Ca²⁺ release in HEK293 cells expressing RyR2

Considering their enhancement of luminal Ca²⁺ activation, which is closely linked to spontaneous Ca²⁺ release, it follows that methylxanthines should likewise increase spontaneous Ca²⁺ release. To test this possibility, we determined the impact of caffeine, aminophylline, and theophylline on spontaneous Ca²⁺ release in HEK293 cells expressing RyR2. Spontaneous Ca²⁺ release was induced in these cells by elevating the external Ca²⁺ concentrations in the absence or presence of caffeine (0.3 mM), aminophylline (0.3 mM), or theophylline (0.3 mM), and was monitored using single cell Ca²⁺ imaging and the fluorescent Ca²⁺ indicator, fura-2 AM. Analyzing the fraction of cells that displayed spontaneous Ca²⁺ release in the form of Ca²⁺ oscillations at each external Ca²⁺ concentration showed that all three methylxanthines increased the propensity for spontaneous Ca²⁺ release (Fig. 5A). For instance, the fraction of oscillating cells was 49.1 ± 4.4% (mean ± SEM) in the presence of 0.3 mM caffeine, 44.6 ± 4.6% in 0.3 mM aminophylline, or 49.3 ± 2.6% in 0.3 mM theophylline, significantly higher than that in the absence of methylxanthines (control, 14.9 ± 5.6%) (P < 0.003). The frequency of spontaneous Ca²⁺ oscillations was increased to 143.9 ± 4.1% by caffeine (P < 0.001), 145.1 ± 4.4% by aminophylline (P < 0.005), and 166.7 ± 10.2% by theophylline (P < 0.001). The store level was reduced to 74.7 ± 2.0% by caffeine (P < 0.001), 78.0 ± 7.0% by aminophylline (P < 0.05), and 69.9 ± 3.0% by theophylline (P < 0.001) (Fig. 5B). The number of HEK293 cells used for analyses was 377 for the control, 507 for caffeine, 417 for aminophylline, and 370 for theophylline. These results are consistent with the notion that methylxanthines reduce the threshold for spontaneous Ca²⁺ release.

Fig. 5. Methylxanthines increase the propensity for spontaneous Ca²⁺ release in HEK293 cells.

(A) The fraction (%, mean ± SEM) of RyR2-expressing cells that display Ca²⁺ oscillations at various [Ca²⁺]_o is shown in the absence of methylxanthines (control, open circles) and in the presence of 0.3 mM caffeine (filled circles), aminophylline (open squares) or theophylline (filled squares). The total numbers of cells analyzed for Ca²⁺ oscillations were 377 for the control, 507 for caffeine, 417 for aminophylline, and 370 for theophylline from 4–6 separate experiments. (B) The frequency (white) of spontaneous Ca²⁺ oscillations and the store Ca²⁺ level (grey) (%, mean ± SEM) in HEK293 cells in the absence of methylxanthines (control) or presence of caffeine, aminophylline or theophylline are shown. Both the frequency of Ca²⁺ oscillations and the store Ca²⁺ content were determined at 1 mM [Ca²⁺]_o. The store Ca²⁺ levels were estimated by measuring the peak of Ca²⁺ release induced by 10 mM caffeine. The average frequency and store level observed in the presence of methylxanthines were normalized to the control values (100%).

Methylxanthines enhance the propensity for spontaneous Ca²⁺ release in isolated rat cardiac myocytes

Methylxanthines have been shown to promote cardiac arrhythmias under some conditions [7–14]. It is possible that their pro-arrhythmic nature is related to their enhancing effect on spontaneous Ca²⁺ release. To test this possibility, we determined whether methylxanthines are able to enhance spontaneous Ca²⁺ release in cardiac myocytes. As with HEK293 cells, spontaneous Ca²⁺ release was induced in rat cardiac cells by increasing the external Ca²⁺ concentration and was monitored using single cell Ca²⁺ imaging and the Ca²⁺ indicator, fluo-4-AM. As shown in Fig. 6, spontaneous Ca²⁺ release in the form of Ca²⁺ waves was observed in cardiac cells in the presence of 10 mM external Ca²⁺. The addition of caffeine (0.5 mM) (Fig. 6A), aminophylline (0.5 mM) (Fig. 6B), or theophylline (0.5 mM) (Fig. 6C) increased the frequency and decreased the amplitude of spontaneous Ca²⁺ waves. The average frequency was 126.9 ± 2.6% of control (P < 0.0001) for caffeine, 146.1 ± 15.5% (P < 0.01) for aminophylline, and 108.2 ± 6.4 (P < 0.01) for theophylline. The average amplitude was 59.7 ± 6.0% of control (P < 0.0001) for caffeine, 69.8 ± 6.1% (P < 0.0001) for aminophylline, and 66.9 ± 5.3% (P < 0.001) for theophylline. The number of cardiac cells used for analyses was 9 for caffeine, 10 for aminophylline, and 16 for theophylline. These data indicate that, as with HEK293 cells, methylxanthines increase the frequency and decrease the amplitude of Ca²⁺ waves, which is consistent with the hypothesis that methylxanthines reduce the threshold for spontaneous Ca²⁺ release in cardiomyocytes.

Methylxanthines increase the basal activity of RyR2, resembling the effect of disease-linked RyR2 mutations

We have shown that disease-linked RyR2 mutations enhance the sensitivity of the channel to luminal Ca²⁺ activation and reduce the threshold for spontaneous Ca²⁺ release. These are the same properties shared by methylxanthines. We have also shown that a number of RyR2 mutations linked to cardiac arrhythmias display an increased activity at high KCl concentrations in the near absence of Ca²⁺ [2,3]. This basal activity likely reflects the stability of the closed state of the channel. To determine whether methylxanthines have any effects on the basal activity of the channel, we performed [³H]ryanodine binding assays. Fig. 7 shows that all three methylxanthines significantly increase the basal activity of the channel. For instance, the basal level of [³H]ryanodine binding at 800 mM KCl was significantly increased from 7.9 ± 1.0 % (control) to 20.2 ± 1.5 % (n = 3, P < 0.05) by caffeine (2.5 mM), 22.5 ± 0.9 % (n = 3, p < 0.001) by aminophylline (2.5 mM), or 23.9 ± 0.4 % (n = 3, P < 0.001) by theophylline (2.5 mM). These data suggest that methylxanthines destabilize the closed state of the channel in a manner similar to those RyR2 mutations known to cause cardiac arrhythmias.

Fig. 7. Effects of methylxanthines on the basal activity of [³H]ryanodine binding.

[³H]ryanodine binding to cell lysate prepared from HEK293 cells expressing RyR2 was carried out at ~3 nM Ca²⁺ (pCa 8.49), various concentrations of KCl (50–1000 mM), 2.5 mM caffeine (filled circles), aminophylline (open squares) or theophylline (filled squares), and 5 nM [³H]ryanodine. The channel activity in the absence of methylxanthines (control) is also shown (open circles). The amount of [³H]ryanodine binding at various Ca²⁺ concentrations was normalized to the maximal binding at 1000 mM KCl and pCa 4. Data points shown are mean ± SEM from 3 separate experiments.

DISCUSSION

Caffeine has widely been used as a probe to study the mechanism of RyR2-associated catecholamine-induced cardiac arrhythmias, but the molecular basis of caffeine activation of RyR2 is unclear. Based on our recent finding that disease-linked RyR2 mutations enhance the luminal Ca²⁺ activation of RyR2 and reduce the threshold for spontaneous Ca²⁺ release or store-overload-induced Ca²⁺ release (SOICR) [2,3], we propose that caffeine promotes catecholamine-induced arrhythmias by sensitizing the RyR2 channel to activation by luminal Ca²⁺. In support of this hypothesis, we have found that caffeine reduces the threshold for luminal, but not cytosolic, Ca²⁺ activation and the threshold for spontaneous Ca²⁺ release. In addition, we have found that as with caffeine, two other methylxanthine compounds, aminophylline and theophylline, also potentiate the channel to luminal Ca²⁺ activation and increase the propensity for spontaneous Ca²⁺ release. Our results suggest that altered luminal Ca²⁺ activation of RyR2 underlies a common arrhythmogenic mechanism of inherited and drug-induced arrhythmias associated with RyR2.

How does caffeine trigger Ca²⁺ release: cytosolic Ca²⁺ activation vs luminal Ca²⁺ activation?

Although caffeine has been widely used as an agonist of RyRs to induce Ca²⁺ release from intracellular stores in various muscle and non-muscle cells, it is not clear how caffeine triggers the opening of RyRs and consequently Ca²⁺ release. It is commonly believed that caffeine triggers Ca²⁺ release by sensitizing the channel to cytosolic Ca²⁺ activation [27,28]. In other words, the RyR channels are activated by the resting cytosolic Ca²⁺ upon the addition of caffeine. However, if the activation of RyRs by caffeine were mediated by the resting cytosolic Ca²⁺, one would expect that in a steady state the activation of RyRs would be maintained in the continuing presence of caffeine, as the resting cytosolic Ca²⁺ after the addition of caffeine would be similar to or greater than that before caffeine stimulation. Such a sustained activation of RyRs by cytosolic Ca²⁺ in the presence of caffeine would lead to Ca²⁺ release in an all-or-none fashion and deplete the intracellular Ca²⁺ stores. In contrast to this prediction, caffeine at submaximal concentrations is able to repetitively trigger partial Ca²⁺ release in a number of cell types, a phenomenon known as “quantal” Ca²⁺ release [16–20]. Consistent with these observations, we found that HEK293 cells expressing recombinant RyR2 also displayed partial or quantal Ca²⁺ release in response to repetitive additions of caffeine (0.2 mM) (Fig. 1A). On the other hand, when these RyR2-expressing HEK293 cells were pretreated with ryanodine, which is known to drastically (>1,000 fold) sensitize the RyR2 channel to activation by cytosolic Ca²⁺ [37], they only responded to the first addition of caffeine, but not to subsequent additions, in an all-or-none manner (Fig. 1B). It is difficult to reconcile these observations with the idea that caffeine-induced Ca²⁺ release is mediated via the activation of the RyR channel by cytosolic Ca²⁺.

However, the RyR channel can also be activated by luminal Ca²⁺. So if, alternatively, caffeine induces intracellular Ca²⁺ release by sensitizing the channel to activation by luminal Ca²⁺, this apparent paradox would be resolved. In this scheme, upon the addition of caffeine the RyR channel is opened by the store luminal Ca²⁺. Therefore, one would expect that caffeine-induced Ca²⁺ release would be partial and dependent on luminal Ca²⁺. This is because a certain concentration of caffeine will sensitize the channel to activation by a certain level of luminal Ca²⁺. As a result of Ca²⁺ release, the store luminal Ca²⁺ level will decrease. The activation of the channel by luminal Ca²⁺ and thus Ca²⁺ release would cease when the luminal Ca²⁺ concentration falls below a threshold level. However, upon increasing the caffeine concentration by a subsequent addition of caffeine, the channel will be further sensitized and again activated by the luminal Ca²⁺ until the luminal Ca²⁺ level decreases to a new steady state. Indeed, it has been shown that caffeine decreased the ER luminal Ca²⁺ level in a concentration dependent manner. The steady state luminal Ca²⁺ level was progressively decreased with increased concentrations of caffeine [17]. Interestingly, caffeine failed to trigger Ca²⁺ release if the luminal Ca²⁺ concentration was lower than the steady state level corresponding to that concentration of caffeine [17]. Similarly, caffeine was found to be unable to trigger Ca²⁺ release from SR membrane vesicles that were loaded with Ca²⁺ below a threshold level [24–26]. The failure of caffeine to trigger Ca²⁺ release in the presence of a normal resting cytosolic Ca²⁺, but a reduced luminal Ca²⁺ level, further indicates that luminal Ca²⁺, but not cytosolic Ca²⁺, is the major mediator of caffeine induced-Ca²⁺ release.

We have previously shown that when the luminal Ca²⁺ level reaches a threshold level, spontaneous Ca²⁺ release occurs in HEK293 cells expressing RyR2 [2,3,39]. In the present study, we investigated the impact of caffeine on the threshold for spontaneous Ca²⁺ release. We found that in the presence of 0.3 mM caffeine, spontaneous Ca²⁺ release occurred at a lower luminal Ca²⁺ level compared to that in the absence of caffeine (Fig. 2). The threshold for spontaneous Ca²⁺ release was further reduced after the addition of 1 mM caffeine. Interestingly, despite the markedly reduced threshold and amplitude, spontaneous Ca²⁺ release in the form of Ca²⁺ oscillations persisted even in the presence of 10 mM caffeine, and was only abolished by the addition of ryanodine. These observations indicate that, unlike ryanodine, caffeine, even at high concentrations, does not cause a sustained activation of the RyR2 channel. Caffeine only activates the channel when the luminal Ca²⁺ reaches a certain threshold. Hence, the continued presence of caffeine is not always associated with an increased Po of RyR2. Collectively, the action of caffeine is to reduce the threshold for luminal Ca²⁺ activation of RyR2, but not necessarily to increase the Po of RyR2.

How does caffeine reduce the threshold for spontaneous Ca²⁺ release?

Early studies on the effect of caffeine on the cytosolic Ca²⁺ dependent activation of single RyR2 channels were performed in planar lipid bilayers using Ca²⁺ as the charge carrier. These studies demonstrated that caffeine markedly enhanced the Po of single RyR2 channels in the presence of submicromolar concentrations of cytosolic Ca²⁺ and millimolar concentrations of luminal Ca²⁺ [27,28]. Since both cytosolic and luminal Ca²⁺ were present, it is not clear whether the activation of RyR2 by caffeine under these conditions resulted from the sensitization of the channel to cytosolic Ca²⁺ or luminal Ca²⁺ or both. To distinguish these possibilities, we determined the impact of caffeine on the cytosolic Ca²⁺ dependence of activation in the near absence of luminal Ca²⁺ or on the luminal Ca²⁺ dependence of activation in the near absence of cytosolic Ca²⁺. We found that at low concentrations of cytosolic Ca²⁺, caffeine markedly reduced the threshold for luminal Ca²⁺ activation, whereas, at low concentrations of luminal Ca²⁺, caffeine had little effect on the threshold for cytosolic Ca²⁺ activation (Fig. 3). These data indicate that at submicromolar levels of cytosolic Ca²⁺ and millimolar levels of luminal Ca²⁺, caffeine preferentially sensitizes the channel to luminal Ca²⁺ activation. Consistent with this view, it has recently been shown that in the presence of 100 nM cytosolic Ca²⁺ caffeine readily activated single RyR2 channels using Ca²⁺ as the charge carrier, but had little effect on single RyR2 channel when Ba²⁺ was used as the charge carrier [29]. These observations demonstrate that the activation of single RyR2 channels by caffeine at submicromolar levels of cytosolic Ca²⁺ is dependent on the presence of luminal Ca²⁺. Taken together, these single channel studies indicate that caffeine reduces the threshold for spontaneous Ca²⁺ release by decreasing the threshold for luminal Ca²⁺ activation of the RyR2 channel.

Caffeine mimics the actions of disease-linked RyR2 mutations

We have recently demonstrated that augmented luminal, but not cytosolic, Ca²⁺ activation of RyR2 is a common feature of a number of disease-linked RyR2 mutations [2,3,39]. Our observation that caffeine reduces the threshold for luminal, but not cytosolic, Ca²⁺ activation of single RyR2 channels indicates that caffeine mimics the effect of disease-linked RyR2 mutations. Indeed, as with disease-linked RyR2 mutations, caffeine at low concentrations reduces the threshold for spontaneous Ca²⁺ release, but has no sustained effect on Ca²⁺-induced Ca²⁺ release (CICR) [2,7,15,23]. Patients with CPVT RyR2 mutations show no structural or functional cardiac abnormalities at rest, but are predisposed to catecholamine-induced cardiac arrhythmias [1]. Similarly, caffeine alone does not induce spontaneous Ca²⁺ release, DADs, or cardiac arrhythmia, but promotes catecholamine-induced triggered activities [7–10]. Furthermore, as with disease-linked RyR2 mutations, caffeine increases the basal level of [³H]ryanodine binding (Fig. 7). Hence, caffeine and disease-linked RyR2 mutations alter the properties of the channel in the same manner.

Other methylxanthine compounds aminophylline and theophylline have been used clinically for the treatment of pulmonary diseases. However, their use has recently been limited due largely to their pro-arrhythmic properties [13,14,41,42]. We have found that, like caffeine, both aminophylline and theophylline preferentially potentiate luminal Ca²⁺ activation of RyR2, reduce the threshold for spontaneous Ca²⁺ release, and increase the basal activity of RyR2 (Figs. 4–7). These effects likely underlie the arrhythmogenic mechanism of these methylxanthines.

Luminal Ca²⁺ activation of RyR, a common target for regulation

It has been proposed that under normal SR Ca²⁺ loading the sensitivity of RyR2 to cytosolic Ca²⁺ activation is extremely low at resting cytosolic Ca²⁺ [43]. However, during SR Ca²⁺ overload, RyR2 becomes much more sensitive to activation [44]. This observation suggests that RyR2 is readily regulated by luminal Ca²⁺. We have recently shown that the activation of the channel by luminal Ca²⁺ is distinct from its activation by cytosolic Ca²⁺ [45]. An increased body of evidence indicates that luminal Ca²⁺ activation of RyR2 is an important target for regulation by endogenous and exogenous effectors [2,45–47]. In addition to methylxanthines, a number of drugs, such as sulmazole, thymol, doxorubicin, ethanol, and shingosyl phosphorylcholine have been found to induce partial or quantal Ca²⁺ release from RyR-gated intracellular Ca²⁺ stores [18]. It is possible that these drugs also induce quantal Ca²⁺ release by sensitizing the channel to activation by luminal Ca²⁺. Hence, modulating the sensitivity of the channel to luminal Ca²⁺ activation may be a common mechanism of regulation of RyRs.

Mechanisms underlying spontaneous Ca²⁺ release

The phenomenon of spontaneous SR Ca²⁺ release in cardiac cells has been known for decades. However, the exact mechanism underlying this process has not been completely defined. In early studies using skinned cardiac cells, Fabiato demonstrated that there are two kinds of Ca²⁺-induced release of Ca²⁺ from the SR [48,49]. One is termed Ca²⁺-induced Ca²⁺ release (CICR), which has a time and Ca²⁺ dependence of activation and inactivation by cytosolic Ca²⁺. The other is known as spontaneous SR Ca²⁺ release, which has no time dependence of activation and is not inactivated by high concentrations of cytosolic Ca²⁺, but requires SR Ca²⁺ overload. A key feature of the activation of SR Ca²⁺ release by cytosolic Ca²⁺ or CICR is its dependence on the rate of increase in the cytosolic Ca²⁺ concentration. A high rate of increase in the cytosolic Ca²⁺ concentration triggers CICR, whereas a low rate of increase in the cytosolic Ca²⁺ concentration inhibits CICR and causes SR Ca²⁺ accumulation. Importantly, when the SR Ca²⁺ content has accumulated to a critical level, spontaneous SR Ca²⁺ release occurs [48,49].

Consistent with Fabiato’s early observations in skinned cardiac cells, Eisner and his colleagues have shown in intact cardiac myocytes that elevated external Ca²⁺ concentrations cause a slight increase in the cytosolic Ca²⁺ level and lead to SR Ca²⁺ accumulation [50]. Similarly, they found that when the SR Ca²⁺ content reaches a threshold level, spontaneous SR Ca²⁺ release occurs in the absence of membrane depolarization. Spontaneous SR Ca²⁺ release was not observed when the SR Ca²⁺ content was below this threshold level. After spontaneous SR Ca²⁺ release occurred, further elevation of external Ca²⁺ concentration increased the frequency of spontaneous Ca²⁺ release, but had little effect on its amplitude. In other words, the spontaneous SR Ca²⁺ release induced by elevated external Ca²⁺ concentrations occurs only when the SR Ca²⁺ content reaches a threshold level. Although the cytosolic Ca²⁺ concentration also increases slightly during external Ca²⁺ elevation, the rate of increase in the cytosolic Ca²⁺ concentration may be too slow to trigger CICR. Hence, spontaneous SR Ca²⁺ release induced by elevated external Ca²⁺ concentrations is likely the result of SR Ca²⁺ overload, rather than the consequence of cytosolic Ca²⁺ activation or CICR.

We have previously demonstrated that elevating the external Ca²⁺ concentration also increases the store Ca²⁺ content in HEK293 cells expressing RyR2 [2,3], similar to those observed in cardiac cells. More importantly, and as with cardiac cells, when the store Ca²⁺ reaches a threshold level, spontaneous Ca²⁺ oscillations occur in these RyR2-expressing HEK293 cells, but not in RyR2-non-expressing cells. Recently, using an ER Ca²⁺ sensor, D1ER, we were able to directly show that spontaneous Ca²⁺ release or Ca²⁺ oscillations occur in HEK293 cells when the ER Ca²⁺ reaches a threshold level [39]. Therefore, as with cardiac cells, the spontaneous Ca²⁺ release observed in HEK293 cells is likely the result of store Ca²⁺ overload.

What then is the role of cytosolic Ca²⁺ activation or CICR in spontaneous Ca²⁺ release induced by elevated external Ca²⁺? We have recently demonstrated that a disease-associated RyR2 mutation, A4860G, abolishes the luminal Ca²⁺ activation of RyR2, but has little effect on the sensitivity of the channel to activation by cytosolic Ca²⁺ [45]. Importantly, this A4860G mutation also abolishes spontaneous Ca²⁺ oscillations in HEK293 cells, despite its normal sensitivity to cytosolic Ca²⁺ activation. These observations indicate that spontaneous Ca²⁺ release is closely linked to the luminal, but not the cytosolic, Ca²⁺ activation of the RyR2 channel, which is consistent with the fact that spontaneous Ca²⁺ release occurs only when the SR Ca²⁺ content reaches a threshold level. Based on these recent observations and those of previous studies, it is likely that spontaneous Ca²⁺ release is initiated by the luminal Ca²⁺ activation of the RyR2 channel. However, since CICR is known to be involved in the propagation of Ca²⁺ waves, spontaneous SR Ca²⁺ release in the form of propagating Ca²⁺ waves is likely the combined product of spontaneous Ca²⁺ release and CICR.

Molecular basis of luminal Ca²⁺ regulation of RyR2

It has been proposed that luminal Ca²⁺ activates RyRs by passing through the open channel and acting on the cytosolic Ca²⁺ activation site (a “feed-through” hypothesis) [31,51]. However, Gyorke et al. and Ching et al. found that RyR2 could still be activated by luminal Ca²⁺ in the absence of luminal-to-cytosolic Ca²⁺ flux [52,53]. Furthermore, the application of trypsin to the luminal side of the RyR2 channel diminishes luminal Ca²⁺ activation, but not Ca²⁺ fluxes, arguing against the “feed-through” mechanism and suggesting the existence of a luminal Ca²⁺ activation site distinct from the cytosolic Ca²⁺ activation site [53]. Recently, a third model incorporating both the feed-through and true luminal Ca²⁺ activation mechanisms, called the luminal-triggered Ca²⁺ feed-through mechanism, has been proposed. In this model, luminal-to-cytosolic Ca²⁺ flux is required for a full activation of the channel by luminal Ca²⁺ [54]. However, we have recently shown that elevating the luminal Ca²⁺ concentration to 50 mM did not activate the A4860G mutant channel, despite the presence of luminal-to-cytosolic Ca²⁺ flux and the normal activation of the channel by cytosolic Ca²⁺. These observations indicate that luminal-to-cytosolic Ca²⁺ flux does not activate the RyR2 channel, and suggest that the activation of RyR2 by luminal Ca²⁺ is mediated by a luminal Ca²⁺ sensor.

The identity of this putative luminal Ca²⁺ sensor is also controversial. It has been proposed that calsequestrin, a low affinity, high capacity SR Ca²⁺ binding protein, acts as a luminal Ca²⁺ sensor and is responsible for the activation of RyR2 by luminal Ca²⁺ [55]. According to this theory, the complex of calsequestrin, triadin and junctin confers the sensitivity of RyR2 to luminal Ca²⁺. At low concentrations of SR luminal Ca²⁺, calsequestrin binds to the triadin/junctin/RyR2 complex in a Ca²⁺-sensitive manner and suppresses the stimulatory effect of triadin/junctin on the RyR2 channel. At high SR luminal Ca²⁺ concentrations, calsequestrin dissociates from triadin/junctin/RyR2, so that triadin/junctin activates RyR2 in the absence of calsequestrin [55]. Hence, calsequestrin, by virtue of its Ca²⁺ dependent association with and dissociation from the triadin/junctin/RyR2 complex, servers as a luminal Ca²⁺ sensor for the luminal Ca²⁺ regulation of RyR2.

Recently, Qin et al. have proposed that RyR2 is regulated by luminal Ca²⁺ through a calsequestrin-independent and a calsequestrin-dependent mechanism [56]. Different from the mechanism proposed by Gyorke et al., the calsequestrin-dependent mechanism proposed by Qin et al. does not involve the association or dissociation of calsequestrin. Instead, the Ca²⁺-sensitivity of the interaction between a calsequestrin monomer and the triadin/junctin/RyR2 complex is the key in conferring the responsiveness of RyR2 to luminal Ca²⁺ activation. The removal of calsequestrin from the triadin/junctin/RyR2 complex completely abolishes the luminal Ca²⁺ response of RyR2, but does not lead to the activation of RyR2 by luminal Ca²⁺, as would be expected based on the mechanism proposed by Gyorke et al. [55]. Thus, exactly how calsequestrin is involved in the regulation of RyR2 by luminal Ca²⁺ is unclear.

The view that calsequestrin serves as the luminal Ca²⁺ sensor for RyR2 is also apparently inconsistent with the observation that purified native RyRs remain sensitive to luminal Ca²⁺ activation [31,40,45,57]. Moreover, recent studies have shown that SR Ca²⁺ release in cardiac myoctyes isolated from calsequestrin knock-out mice remains steeply nonlinear with increasing SR Ca²⁺ content, indicating that the RyR2 channel can sense luminal Ca²⁺ in the absence of calsequestrin [58]. Another important observation is that calsequestrin knockout cardiac myocytes display largely unaltered SR Ca²⁺ release and SR Ca²⁺ content under basal conditions, suggesting that calsequestrin does not play an essential role in modulating the gating of RyR2 and SR Ca²⁺ leak at rest or at low SR Ca²⁺ concentrations [58]. These observations have led to the conclusion that calsequestrin, although it may modulate SR Ca²⁺ release, is not required for luminal Ca²⁺ sensing [58]. Consistent with this finding, we found that recombinant RyR2 expressed in HEK293 cells, which lack calsequestrin, is activated by luminal Ca²⁺ [40,45]. The reasons for these apparently controversial findings regarding the role of calsequestrin in the luminal Ca²⁺ regulation of RyR2 from different groups are not clear and further studies are needed.

Summary

In summary, our data demonstrate that caffeine triggers Ca²⁺ release by reducing the threshold for luminal, but not cytosolic, Ca²⁺ activation of the RyR2 channel. Unlike ryanodine, which induces a full activation of the channel, caffeine, even at high concentrations, does not always hold the channel in the open state. Rather, the action of caffeine is to reduce the luminal Ca²⁺ threshold at which spontaneous Ca²⁺ release occurs. As with caffeine, the clinically relevant methylxanthines aminophylline and theophylline preferentially potentiate luminal Ca²⁺ activation, reduce the threshold for spontaneous Ca²⁺ release, and increase the basal channel activity, mimicking the actions of disease-linked RyR2 mutations.

ABREVIATIONS USED

RyR2

Cardiac Ryanodine Receptor

SOICR

Store Overload Induced Ca²⁺ Release

SR

Sarcoplasmic Reticulum

DAD

Delayed Afterdeplolarization

CPVT

Catecholaminergic Polymorphic Ventricular Tachycardia

ARVD2

Arrhythmogenic Right Ventricular Cardiomyopathy type 2

FRET

Fluorescence resonance energy transfer

HEK293

Human Embryonic Kidney Cells