Subcellular Ca²⁺ signaling in the heart: the role of ryanodine receptor sensitivity

The SR includes two primary components, the jSR and the longitudinal SR (also called the free SR or the “network” SR; see Fig. 1) (Brochet et al., 2005). In addition, the SR is connected to the ER and the nuclear Ca²⁺ store (Wu and Bers, 2006), thus making the entire Ca²⁺ storage network fully interconnected. The primary Ca²⁺ release sites are located at the jSR, an extremely small pancake-shaped sub-volume in the system (each jSR contains ∼1 attoliter, ∼1 × 10⁻¹⁸ l, a size ∼4,000 times smaller than a Ca²⁺ spark). On one face, the jSR contains a paracrystalline cluster of 10–300 SR Ca²⁺ release channels, the RYR2s that span 15 nm (the “subspace”) to the surface membrane (i.e., the transverse tubule [TT] or surface sarcolemmal [SL] membranes) (Franzini-Armstrong et al., 1999; Baddeley et al., 2009). The other face of the jSR points toward the M line and is tens of nanometers away from the Z-disk end of the mitochondria and is connected through the thin nanoscopic tubules that constitute the longitudinal SR. The L-type Ca²⁺ channels (or dihydropyridine receptors) are transmembrane proteins that span the SL and TT membranes facing the subspace and thus appose the jSR. The lumen of the jSR also contains the Ca²⁺-binding protein calsequestrin type 2 (CASQ2) and other regulatory proteins, such as junctin, triadin, junctophilin, and a small amount of other proteins including the ER Ca²⁺-binding and chaperone protein calreticulin (Bers, 1991; Györke et al., 2007).

The brief opening of either an L-type Ca²⁺ channel or an RYR2 under diastolic conditions leads to a local elevation of Ca²⁺ in the subspace ([Ca²⁺]_subspace) from the normal [Ca²⁺]_i of 100 nM to ∼10 µM (Cannell and Soeller, 1997; Soeller and Cannell, 1997; Sobie et al., 2002). This elevation of [Ca²⁺]_subspace, albeit brief (<1 ms), is sufficient to activate the RYR2 cluster to produce a Ca²⁺ spark, a process termed Ca²⁺-induced Ca²⁺ release, or CICR (Fabiato, 1983; Cheng et al., 1993; Cannell et al., 1994a,b; Sobie et al., 2002; Cheng and Lederer, 2008; Györke and Terentyev, 2008; Liu et al., 2010). The spatial organization of the jSR enables reliable activation of Ca²⁺ sparks by action potentials, but the relative insensitivity of RYR2s to calcium protects the cell from instability and enables Ca²⁺ sparks during diastole to remain isolated from neighboring jSR spark sites. Thus, Ca²⁺ sparks do not normally activate other Ca²⁺ sparks (Cheng et al., 1993). High gain in the signaling pathway is thus created by the spatial organization of RYR2s in a cluster, and stability is maintained by the relative insensitivity of the RYR2s to [Ca²⁺]_i (Cheng and Lederer, 2008).

The “life cycle” of a Ca²⁺ spark provides clues to its regulation. Should [Ca²⁺]_subspace increase to ∼10 µM, there is good probability that a Ca²⁺ spark will arise through CICR as noted above. Although Ca²⁺ spark termination is seen robustly in experiments, our understanding of how this occurs was more challenging. Mathematical modeling of spark behavior suggests that if a significant fraction of the local jSR Ca²⁺ is depleted during the Ca²⁺ spark (between 50 and 90%), a robust termination of the spark will be seen (Sobie et al., 2002). Importantly, this occurs without the need for “inactivation” of the RYR2s, unlike in other models (Stern et al., 1999). In support of this model, there is no experimental evidence to date that suggests rapid “fateful” inactivation of RYR2s (Liu et al., 2010). The [Ca²⁺]_SR depletion hypothesis regarding Ca²⁺ spark termination has been supported by the observation of Ca²⁺ “blinks” (Brochet et al., 2005). A Ca²⁺ blink is the [Ca²⁺]_SR depletion signal that occurs when a Ca²⁺ spark takes place. It is the reciprocal signal of a Ca²⁺ spark; specifically, [Ca²⁺]_SR depletion is seen whenever sparks occur because the Ca²⁺ flux that underlies the Ca²⁺ spark comes from the jSR lumen. SR Ca²⁺ leak is the term of art now used to describe the non-synchronized loss of Ca²⁺ from the SR, which may occur as Ca²⁺ sparks or as very small release events that may be invisible when viewed with a confocal microscope (Sobie et al., 2006). How Ca²⁺ leak may occur, and how this leak may influence Ca²⁺ instability and cardiac arrhythmogenesis, is actively under examination (Wehrens et al., 2003; Lehnart et al., 2006, 2008; Sobie et al., 2006).

A Ca²⁺ spark is the local Ca²⁺ signal produced by an ensemble of RYR2s at a single jSR (Cheng and Lederer, 2008). Diastolic sparks can occur when a single RYR2 opens probabilistically and increases [Ca²⁺]_subspace to a high level (i.e., ∼10 µM), activating the other RYR2s facing the same subspace and in the same jSR. In a similar manner, the opening of an L-type Ca²⁺ channel at near-diastolic potentials will also increase [Ca²⁺]_subspace and activate a Ca²⁺ spark (Cannell et al., 1995; Cheng et al., 1996; Santana et al., 1996). More recently (Jiang et al., 2004), an alternative hypothesis has been put forward suggesting that sparks are triggered by “store overload–induced Ca²⁺ release,” or SOICR. Although we have not identified experimental evidence that supports the SOICR hypothesis and distinguishes it from CICR, there is strong evidence suggesting that [Ca²⁺]_SR regulates RYR2 [Ca²⁺]_i sensitivity. [Ca²⁺]_SR is certainly an integral feature of CICR because it influences the sensitivity of the RYR2 to be opened by [Ca²⁺]_i (Györke and Györke, 1998). The influence exerted by [Ca²⁺]_SR represents one factor, along with several others such as phosphorylation of the RYR2s, nitrosylation, oxidation, and channel mutations (Wehrens et al., 2003, 2004, 2005; Lehnart et al., 2004; Viatchenko-Karpinski et al., 2004; Kubalova et al., 2005; Terentyev et al., 2005, 2008, 2009; di Barletta et al., 2006; Guo et al., 2006; Györke and Terentyev, 2008), which all can increase the voltage-independent probabilistic openings of RYR2s (Table I and Fig. 1 C). The distinction between the SOICR hypothesis and the CICR hypothesis is important to make. Both support [Ca²⁺]_SR influencing Ca²⁺ sparks. SOICR, however, suggests that SR Ca²⁺ overload triggers depolarization-independent Ca²⁺ sparks (Jiang et al., 2004, 2007, 2008; Jones et al., 2008), underlies Ca²⁺ waves, and is somehow distinctive from the probabilistic opening of RYR2. Just as the opening of L-type Ca²⁺ channels apposing the jSR “triggers” Ca²⁺ sparks by CICR, so does the probabilistic opening of RYR2s in a jSR cluster. Importantly, when a spark initiates a propagating Ca²⁺ wave or contributes to the conduction of a wave, it is CICR that links the elevation of local [Ca²⁺]_i during the spark produced by one jSR to the triggering of a Ca²⁺ spark in the next jSR. For Ca²⁺ waves to arise, it has been hypothesized that an increase in RYR2 sensitivity to [Ca²⁺]_i occurs due to the underlying pathologies, which include SR Ca²⁺ overload and catecholaminergic polymorphic ventricular tachycardia (CPVT) mutations of RYR2 and CASQ2 (Marks et al., 2002; Eldar et al., 2003; di Barletta et al., 2006; Györke, 2009). By this means, the Ca²⁺ released by a spark diffuses and elevates the [Ca²⁺]_i in the subspace of a neighboring RYR2 cluster and triggers a spark at that site. This chain reaction can continue to produce the arrhythmogenic Ca²⁺ wave that activates the Na⁺/Ca²⁺ exchanger and produces the arrhythmogenic transient inward current, I_TI (Györke et al., 2007). An increase in the sensitivity of the RYR2s to [Ca²⁺]_i is sufficient to enable this sequence. Although elevated [Ca²⁺]_i and [Ca²⁺]_SR are the simplest ways to enable this sequence, changes in RYR2 sensitivity by mutations to its primary sequence or alterations in modulatory proteins (e.g., CASQ2, calstabin2, and others) may also be sufficient. That increases in [Ca²⁺]_SR may increase RYR2 opening rate is a part of CICR that has been well articulated in the literature as noted above. It would thus seem that the CICR hypothesis can fully account for the triggering of Ca²⁺ sparks and Ca²⁺ waves in health and disease.

To the extent that the mitochondria may buffer Ca²⁺ or actively take up and release Ca²⁺, they may significantly affect [Ca²⁺]_i dynamics (Xu et al., 2002; Dedkova and Blatter, 2008; Liu and O’Rourke, 2008; Andrienko et al., 2009; Kettlewell et al., 2009; O’Rourke and Blatter, 2009). The scale of this influence is potentially large because the mitochondria take up approximately one third of the intracellular volume of cardiac myocytes. The largest subset of cardiomyocyte mitochondria are the intermyofibrillar mitochondria that run from the jSR at one Z disk through the M-line structures to the jSR at the next Z disk. Each end of these sarcomere-spanning mitochondria experiences Ca²⁺ sparks near their source (both ends) and thus may influence the magnitude and kinetics of local [Ca²⁺]_i at their ends and global [Ca²⁺]_i changes in the middle. Although the potential for mitochondria to influence [Ca²⁺]_i is enormous, their actual roles in modulating Ca²⁺ signals are debated. Experimental evidence for a substantive role (Liu and O’Rourke, 2008, 2009), as well as a minimal role (Andrienko et al., 2009; Kettlewell et al., 2009), has been put forth. Cellular and molecular tools for these investigations are limited because even the primary sequence of virtually every functional structure (channel and transporters) that may influence mitochondrial Ca²⁺ dynamics remains elusive, despite significant effort. Recently, one of the key transporters was identified, the mitochondrial Na⁺/Ca²⁺ exchanger (Palty et al., 2010), which may provide a fundamental tool that is needed to jump-start the investigation and help to resolve the dynamic role that mitochondria may play in cardiac Ca²⁺ signaling.

It has been long hypothesized that the beat-to-beat change in cell length may influence local cardiac [Ca²⁺]_i signals (Allen and Blinks, 1978) and hence Ca²⁺ sparks . In keeping with this hypothesis, diverse cytoskeletal disruptions have been shown to have profound effects on [Ca²⁺]_i. For example, mutations in the cytoskeletal adaptor protein ankyrin B (AnkB or Ank2) has been shown to affect [Ca²⁺]_i through its actions on Ca²⁺ transporters and channels (Mohler et al., 2003; Mohler and Bennett, 2005). Consistent with this role, AnkB mutations in human and experimental animals have been shown to underlie long QT syndrome type 4 (Mohler et al., 2003). The investigation, however, suggested that it was not the disruption of a dynamic beat-to-beat regulator that caused the dysfunction. Instead, it was the role played by AnkB in the steady-state organization of the Ca²⁺ transport and signaling proteins that was the critical cause of the Ca²⁺ signaling change. Similarly, mutations or knockout of the intermediate filament protein desmin leads to a complex array of cardiac cellular dysfunctions, which may reflect the influence of this cytoskeleton element on diverse steady-state organizational elements such as the mitochondria (Weisleder et al., 2004; Maloyan et al., 2005). In contrast to these findings, recent evidence suggests that there may be a beat-to-beat mechano-transduction pathway for which microtubules and tubulin contribute to stretch activated local Ca²⁺ signaling (see below) (Iribe et al., 2009).

The recent discovery that physiological stretch increases the transient occurrence of Ca²⁺ sparks (Iribe et al., 2009) reveals an important line of inquiry that can now be investigated for the first time. These findings are important in any reexamination of normal cardiac Ca²⁺ signaling as well as in the investigation of the pathophysiology of diverse diseases that appear to be linked to cellular contraction and shortening. Iribe et al. (2009) report that a stretch of 8% of the cellular diastolic length leads to a transient increase in the Ca²⁺ spark rate, which can be blunted specifically by microtubule disruption, but not by inhibitors of L-type Ca²⁺ channel flux, Na⁺/Ca²⁺ exchanger flux, stretch-activated channels, nor NO synthase. Although there was earlier evidence for spatial association of microtubules with the SR and mitochondria (Cartwright and Goldstein, 1985), no functional or contact evidence was provided. In contrast, Iribe and colleagues added to the functional connection noted above with EM tomography, demonstrating co-location of microtubules and the jSR. They thus provide strong evidence for a mechano-transduction pathway that can provide a stretch signal to the jSR. If such a signal were provided to the SR, it could affect all of the RYR2s in the jSR because they are organized in a paracrystalline array. These findings lead to the hypothesis that stretching the ventricular myocyte applies a distortion to the diastolic jSR and thereby increases the sensitivity of RYR2s in the jSR to [Ca²⁺]_i (Iribe et al., 2009). After the stretch is a burst of Ca²⁺ sparks that relax over seconds. This is the time scale needed for cellular contraction and shortening to influence local Ca²⁺ signaling and the cardiac [Ca²⁺]_i transient. Note too that such stretch-dependent changes in the open probability of the RYR2 should have analogous effects during all phases of cardiac Ca²⁺ signaling, including during systole. These exciting and provocative results, still in need of much further investigation, provide us with a lens through which we can view one of the key “stretch-dependent” cardiac diseases, muscular dystrophy. We suspect that there may be important connections between specific kinds of muscular dystrophy and stretch-modulated Ca²⁺ signaling because four interrelated components are involved in dystrophic cardiomyopathy: (1) Ca²⁺ entry, (2) Ca²⁺ release, (3) stretch-dependent modulation, and (4) molecular defects associated with some dystrophies.

DMD, the most common of the muscle dystrophies (1:3,500 males), is a fatal, X-linked disease characterized by progressive muscle weakness that most commonly leads to respiratory failure (for review see Ervasti and Campbell, 1993). DMD also results in a progressive dilated cardiomyopathy, with nearly all patients exhibiting cardiac manifestations by 20 years of age. As improved treatments for skeletal muscle weakness delay respiratory failure, cardiac complications have become increasingly limiting for the survival of DMD patients (Finsterer and Stöllberger, 2000; Muntoni et al., 2003).

The genetic basis of the DMD pathology is due to a lack of expression or dysfunction of the very large (3,500 amino acids) intracellular muscle protein, dystrophin. Dystrophin associates with the transmembrane β-dystroglycan and other proteins to form the dystrophin–glycoprotein complex and binds to cytosolic actin (Ervasti et al., 1990; Ohlendieck et al., 1993). Similar to human patients, mdx animals lacking dystrophin demonstrate contractile and conductive deficits as well as cardiac dilation and fibrosis that increase with age (Quinlan et al., 2004). Several groups have recently identified local Ca²⁺ signaling abnormalities in mdx hearts that may underlie the pathology of dystrophic cardiomyopathy (Williams and Allen, 2007a,b; Jung et al., 2008; Ward et al., 2008; Ullrich et al., 2009).

Dystrophic myocytes reveal elevated Ca²⁺ spark activity and the appearance of Ca²⁺ waves in response to mechanical stress (Yasuda et al., 2005; Jung et al., 2008; Fanchaouy et al., 2009). Of note, among the compensatory adaptations in the mdx heart is an ∼1.4-fold increase in β-tubulin (Wilding et al., 2005). Based on this cytoskeletal adaptation in conjunction with the findings of Iribe et al. (2009), it is attractive to ascribe altered mechano-transduction (via microtubules) to aberrant Ca²⁺ signals in the DMD heart. In support of this notion, Prins et al. (2009) have discussed dystrophin as a giant “cytolinker” protein, one that stabilizes structures through the ability to link to actin filaments as well as to microtubules.

Work by Yasuda et al. (2005) suggests that the stretch-induced aberrant Ca²⁺ signals in mdx myocytes depends on Ca²⁺ (and possibly Na⁺) entry. Use of the surfactant and putative “membrane sealant” poloxamer 188 (with support from experiments with the extracellular dye FM-143) has led to the hypothesis that “micro-ruptures” in the SL membrane may underlie Ca²⁺ entry in stressed mdx myocytes (Yasuda et al., 2005; Fanchaouy et al., 2009). Other reports have linked stretch-activated channels to enhanced Ca²⁺ influx in the dystrophic heart (Williams and Allen, 2007a). These channels have been shown to be overexpressed in the DMD heart, and their inhibition blunts the aberrant local Ca²⁺ activity associated with stress (Fanchaouy et al., 2009).

Other mechanisms have been reported that could indirectly affect mechano-transduction pathways of altered SR Ca²⁺ release. The dystrophic heart produces more reactive oxygen species (ROS) (Williams and Allen, 2007b), and mdx myocytes exhibit aggravated ROS production in response to osmotic shock (Jung et al., 2008; Fanchaouy et al., 2009), a commonly used, albeit non-physiological form of mechanical stress. Mechanistically, oxidation sensitizes RYRs to Ca²⁺ (Pessah et al., 2002), making excessive ROS production and RYR2 sensitization an appealing explanation for a greater degree of Ca²⁺ instability in the DMD heart. Supporting this hypothesis is the apparent ability of ROS scavengers to protect mdx cells from the Ca²⁺ signaling consequences of mechanical stress (Jung et al., 2008; Ullrich et al., 2009) and the up-regulation of NADPH oxidase in the DMD heart (Williams and Allen, 2007b). Furthermore, several additional pathways have been linked to dystrophic changes in skeletal muscle, such as RYR nitrosylation and RYR hyperphosphorylation (Bellinger et al., 2009). These pathways could also serve to increase the sensitivity of RYR2 in the DMD heart to [Ca²⁺]_i. If so, this could result in an increase in the mechano-transduction–induced gain of CICR. Clearly, these provocative hypotheses must be tested experimentally.

The regulation of local Ca²⁺ signaling is critically dependent on the sensitivity of RYR2 to [Ca²⁺]_i and the subcellular anatomy of the heart cell. Physiological and pathophysiologic changes in local cardiac Ca²⁺ signaling may depend on altered sensitivity of the RYR2s to [Ca²⁺]_i. These critical signaling changes arise from many factors that influence the open probability of RYR2, including [Ca²⁺]_SR, mutations of RYR2 itself, mutations of RYR2-linked proteins, RYR2 protein phosphorylation and nitrosylation, and other molecular and cellular features. Recently, cellular stretch and the activation of mechano-transduction pathways have been shown to contribute to this process. Quantitative molecular and biophysical investigations of the processes that regulate RYR2 behavior are needed to further elucidate cardiac Ca²⁺ signaling in health and disease.

This Perspectives series includes articles by Gordon, Parker and Smith, Xie et al., Santana and Navedo, and Hill-Eubanks et al.

We would like to acknowledge support from National Heart, Lung and Blood Institute (grants P01 HL67849 and R01-HL36974); National Institutes of Health (grant RC2 NR011968); Leducq North American-European Atrial Fibrillation Research Alliance; European Union Seventh Framework Program (FP7) “Identification and therapeutic targeting of common arrhythmia trigger mechanisms”; and support from the Maryland Stem Cell Commission.