Mitochondria are sources of metabolic sink and arrhythmias

Abstract

Mitochondria have long been recognized for their central role in energy transduction and apoptosis. More recently, extensive work in multiple laboratories around the world has significantly extended the role of cardiac mitochondria from relatively static arbitrators of cell death and survival pathways to highly dynamic organelles that form interactive functional networks across cardiomyocytes. These coupled networks were shown to strongly affect cardiomyoyte responses to oxidative stress by modulating cell signaling pathways that strongly impact physiological properties. Of particular importance is the role of mitochondria in modulating key electrophysiological and calcium cycling properties in cardiomyocytes, either directly through activation of a myriad of mitochondrial ion channels or indirectly by affecting cell signaling cascades, ATP levels, and the over-all redox state of the cardiomyocyte. This important recognition has ushered a renewed interest in understanding, at a more fundamental level, the exact role that cardiac metabolism, in general and mitochondria, in particular, play in both health and disease. In this article, we provide an overview of recent advances in our growing understanding of the fundamental role that cardiac mitochondria play in the genesis of lethal arrhythmias.

INTRODUCTION

Mitochondria are well recognized for their importance in energy production and apoptosis (Gustafsson & Gottlieb, 2008). They generate ATP through oxidative phosphorylation, driven by electron transport across the electron transport chain. In addition, mitochondria generate reactive oxygen species (ROS) which have diverse cell signaling functions (Droge, 2002; Becker, 2004). A key metric of mitochondrial function is the mitochondrial membrane potential (Δψ_m) which forms the proton-motive force used to produce ATP (O’Rourke, 2007). In normal hearts, Δψ_m is tightly regulated such that the production of ATP is maintained within a physiological range that matches energy production to demand (O’Rourke, 2007). This limits ROS generation and oxidative stress. In ischemia-reperfusion injury, Δψ_m is disrupted, altering over-all energy and redox balance within cardiac myocytes (Honda et al., 2005).

Seminal work has extended the role of mitochondria from static arbitrators of cell death and survival pathways to highly dynamic organelles that form interactive networks across cardiomyocytes. These coupled networks strongly affect cardiomyocyte responses to oxidative stress by modulating cell signaling pathways. This important recognition has ushered a renewed interest in understanding, at a more fundamental level, the exact role that cardiac mitochondria play in both health and disease (Michelakis, 2008). In this review, we focus on the role of mitochondrial dysfunction in promoting cardiac electrophysiological abnormalities at the cellular level and malignant arrhythmias at the tissue-network level.

Mitochondrial Criticality and Metabolic Oscillations

Zorov et al (Zorov et al., 2000; Zorov et al., 2006) advanced the notion of ROS-induced ROS-release (RIRR) to explain how local ROS injury within a discrete region of a cardiomyocyte can rapidly accumulate across a critical mass of the mitochondrial network to cause cellular oxidative stress. In these studies, RIRR was described as a fundamental mechanism by which cardiac mitochondria respond to elevated ROS levels by stimulating endogenous ROS production in a regenerative, autocatalytic process that ultimately results in cellular dysfunction and death (Zorov et al., 2006).

Distinct modes of RIRR have been postulated based on their dependence on various mitochondrial ion channels (Yang et al.). Specifically, Zorov et al (Zorov et al., 2000; Zorov et al., 2006) demonstrated a convincing relationship between the destabilization of Δψ_m upon mitochondrial oxidation and the induction of the mitochondrial permeability transition which causes apoptosis (Zorov et al., 2000). On the other hand, studies by Aon et al. (Aon et al., 2003) provided strong evidence in support of the inner membrane anion channel (IMAC) as a mediator of RIRR and associated electrophysiological and metabolic instabilities. In these studies, photo-induced oxidation of a discrete region within the cardiac myocyte unleashed a regenerative process of RIRR that was dependent on IMAC activation and not the mPTP. Once a threshold level of ROS was exceeded across a critical mass of the mitochondrial network (ie mitochondrial criticality), sustained Δψ_m oscillations were initiated (Aon et al., 2006; Aon et al., 2009). Similar Δψ_m oscillations are also generated in isolated myocytes subjected to oxidative stress via substrate deprivation (Romashko et al., 1998), ATP depletion (Ryu et al., 2005), diamide (Aon et al., 2007), and respiratory inhibition (Ryu et al., 2005). Recent evidence using two-photon microscopy confirmed these cellular data as reversible collapses in Δψ_m were observed in intact hearts exposed to global ischemia/reperfusion or diamide administration (Slodzinski et al., 2008). As will be discussed next, these mitochondrial oscillations can result in cellular electrophysiological oscillations via cyclical activation of sarcolemmal K-ATP (sarcK_ATP) channels providing compelling evidence of a mechanistic link between mitochondrial dynamics and cellular electrical dysfunction. In what follows, we describe the downstream ionic mediator of electrical inexcitability caused by mitochondrial dysfunction, followed by a discussion of key upstream mechanisms that regulate arrhythmias, including mitochondrial ion channels and the redox state of the cardiomyocyte.

Down-stream mediator of metabolic stress: Role of sarcolemmal K_ATP channels

SarcK_ATP channels link membrane excitability to metabolism (Nichols, 2006). They are regulated by intracellular nucleotides, membrane phospholipids, protein kinases and phosphatases (Nichols, 2006). SarcK_ATP channel activation can precede cellular ATP depletion because the open probability of these channels is increased when cofactors like ADP, pH and Mg²⁺ begin to rise. SarcK_ATP channels activate rapidly when mitochondria uncouple because the drop in Δψ_m due to increased proton leak causes the reversal of the ATP synthase, thus consuming cytoplasmic ATP and decreasing the phosphorylation potential. Tight coupling between the mitochondrial energy state and sarcK_ATP channel activation is facilitated by the high energy phosphoryl transfer reactions of the cytoplasm (Sasaki et al., 2001).

Due to their abundance in the plasma membrane, the opening of sarcK_ATP channels causes rapid action potential shortening, loss of intracellular K⁺, and reduction in myocyte excitability (Billman, 2008). In fact, increased K⁺ conductance through sarcK_ATP channels can effectively lock the resting membrane potential close to the equilibrium potential for K⁺ (Kleber, 1983). Indeed, sarcK_ATP channel activation accounts for most of the action potential shortening during ischemia, as evidenced by the ability of K_ATP channel blockers (ie, glibenclamide) to prevent the decrease in action potential duration during early ischemia (Akar et al., 2005).

The dynamic relationship between sarcK_ATP channel activation and the metabolic status of the cardiomyocyte was first observed by O’Rourke and colleagues (O’Rourke et al., 1994; O’Rourke et al., 1995). Following metabolic stress either by substrate deprivation or increased ADP levels, sarcK_ATP currents were activated in phase with NADH fluctuations. In these experiments, sustained Δψ_m oscillations occurred in phase with cellular electrophysiological (namely action potential) oscillations that were driven by ‘out-of-phase’ sarcK_ATP current activation (Aon et al., 2003).

While sarcK_ATP channel activation is thought to protect the viability of ischemic tissue by limiting calcium cycling and force generation during periods of reduced energy supply, increased potassium conductance through these channels predisposes to electrical dysfunction and arrhythmias (Billman, 1994; Billman et al., 1998; Billman, 2008). The pro-arrhythmic potential of sarcK_ATP channel activation during ischemia-reperfusion could be attributed to increased dispersion of repolarization and shortening of the effective refractory period, and therefore the cardiac wavelength, at a time when calcium mediated triggers are known to arise. Moreover, the opening of sarcK_ATP channels creates a current sink which can slow or block conduction wavefronts in local regions where the open probability of sarcK_ATP channels is high (i.e. where the energetic status of the cell is compromised), a phenomenon that we previously termed ‘metabolic sink’ (Akar et al., 2005).

This pro-arrhythmic potential of sarcK_ATP channel activation has been confirmed in multiple studies. Preventing sarcK_ATP channel activation by pharmacological blockade of the channel decreased the incidence of ventricular arrhythmias in rat (Vajda et al., 2007), rabbit (Fischbach et al., 2004), pig (Wirth et al., 1999), dog (Billman et al., 1998), and man (Cacciapuoti et al., 1991; Lomuscio et al., 1994; Aronson et al., 2003). On the other hand, sarcK_ATP channel blockade with glibenclamide failed to delay the onset of inexcitability during late ischemia or the initiation of arrhythmias upon reperfusion in the ex vivo perfused guinea pig heart. In order to understand the factors driving the opening of sarcK_ATP channels during metabolic stress, an overview of key mitochondrial ion channels and bioenergetic properties are discussed below.

Mitochondrial ion channels as root causes of mitochondrial dysfunction and arrhythmias

The mitochondrial membrane is a highly resistive structure that maintains a large voltage gradient and proton-motive force, required for electron transport and ATP production (Brown et al., 2010). Nonetheless, a rich diversity of ion channels and transporters has been discovered in the inner and outer membranes of mitochondria. Of note to arrhythmia mechanisms are various ion channels (Figure 1) that modulate Δψ_m and also promote apoptosis (mPTP), cellular inexcitability (IMAC), cardioprotection (mitoKATP), and mitochondrial calcium influx (MCU). The interested reader is referred to excellent reviews that exclusively cover mitochondrial ion channel targets in a more comprehensive manner (Peixoto et al.).

Figure 1.

Schematic of key energy sensitive ion channels that can promote cell survival, death, or arrhythmias.

Inner Membrane Anion Channel

Anion flux across the inner mitochondrial membrane was observed in early studies in which anion movement was shown to regulate mitochondrial volume (Azzi & Azzone, 1966, 1967; Brierley, 1970). Since then, the existence of IMAC has been confirmed in multiple studies demonstrating its importance in anion efflux from energized mitochondria (Garlid & Beavis, 1986; Beavis, 1992). Although the exact structure and molecular identity of IMAC remain elusive, the tight regulation of this channel by benzodiazepine compounds (Beavis, 1989) suggests a strong association between a partially anion selective pore-forming subunit in the inner membrane and a peripheral benzodiazepine receptor in the outer membrane.

The importance of IMAC in modulating Δψ_m was first noted when several distinct IMAC ligands were shown to prevent pathological Δψ_m oscillations in isolated cardiac myocytes (Aon et al., 2003). Importantly, blocking Δψ_m oscillations by targeting the IMAC also inhibited action potential oscillations and prevented myocyte inexcitability (Aon et al., 2003). This provided indirect evidence that targeting the IMAC may be an effective strategy for preventing arrhythmias, at least at the cellular level.

Indeed, IMAC blockade successfully prevented post-ischemic arrhythmias in intact myocardium (Akar et al., 2005; Brown et al., 2008b; Brown et al., 2010). Optical mapping of the epicardial surface of guinea pig hearts revealed that IMAC blockade decreased ischemia-induced action potential shortening and markedly suppressed the incidence of ventricular tachycardia/fibrillation during the early onset of reperfusion (Akar et al., 2005). Cardioprotection mediated by IMAC blockade was also observed in isolated rabbit hearts and accompanied by improved left ventricular function (Brown et al., 2008b). Of notable clinical interest, reperfusion arrhythmias in both studies were also prevented when the IMAC blocker was delivered as a bolus injection at the onset of reperfusion (Akar et al., 2005; Brown et al., 2008b).

Mitochondrial Permeability Transition Pore

The role of the mitochondrial permeability transition pore (mPTP) in ischemia/reperfusion injury has received considerable attention (Halestrap et al., 2004; Murphy & Steenbergen, 2008; Halestrap, 2009; Halestrap & Pasdois, 2009). It is clear that the opening of the mPTP plays a significant role in the generation of necrotic and apoptotic cell death, both of which are involved in the etiology of myocardial infarction (McCully et al., 2004). Administration of cyclosporin-A or sanglifehrin-A, both blockers of the mPTP, attenuates myocardial infarction (Weinbrenner et al., 1998; Minners et al., 2000; Hausenloy et al., 2002; Argaud et al., 2004), left ventricular dysfunction (Griffiths & Halestrap, 1993; Clarke et al., 2002; Hausenloy et al., 2004; Oka et al., 2008a), cardiomyocyte death (Nazareth et al., 1991; Duchen et al., 1993; Kim et al., 2006), and ischemia-reperfusion injury (Di Lisa et al., 2001; Oka et al., 2008b). The translation of these findings was recently supported in a clinical study, in which administration of cyclosporin-A immediately prior to percutaneous coronary intervention decreased the extent of short-term injury in a small clinical trial (Piot et al., 2008).

While the role of the mPTP in cell death is well established, its involvement in the generation of arrhythmias remains controversial. While some studies showed moderate protection against arrhythmias, other studies confirmed a lack of protection in rat (Dow et al., 2009), guinea pig (Akar et al., 2005), and rabbit (Brown et al., 2008b) hearts. Moreover, delivery of a cyclosporin-A bolus prior to stenting did not seem to influence the incidence of ventricular fibrillation in humans (Piot et al., 2008). Lack of protection against arrhythmias by mPTP blockade is also supported by mechanistic studies in isolated myocytes that demonstrate that Δψ_m depolarization caused by substrate deprivation or photo-induced oxidation is not prevented by cyclosporin-A (Romashko et al., 1998; Huser & Blatter, 1999; Zorov et al., 2000; Aon et al., 2003). It is very important to note, however, that by protecting against apoptosis and reducing the size of myocardial infarction following an ischemic insult, mPTP blockade may suppress scar related arrhythmias that are associated with healed myocardial infarction. Also, by inhibiting myocyte loss and improving left ventricular function, this strategy may confer an anti-arrhythmic effect through beneficial mechano-electrical feedback or by hindering the progression of adverse electrical remodeling.

MitoK_ATP Channel

Evidence for a mitochondrial ATP-sensitive potassium (mitoK_ATP) channel was first observed in rat liver mitochondria (Inoue et al., 1991), and later confirmed in heart (Paucek et al., 1992). The opening of mitoK_ATP channels may underlie the cardioprotective effects of preconditioning stimuli by partial dissipation of Δψ_m, reduction in the driving force for calcium entry into the mitochondrial matrix, inhibition of apoptosis, and overall improvement in cellular respiration ((O’Rourke, 2000; Gross & Peart, 2003).

Numerous studies have examined the role of mitoK_ATP channel activation/blockade in altering infarct size (Takashi et al., 1999; O’Rourke, 2004). In general, mitoK_ATP channel blockade with 5-hydroxydecanoate (5-HD) abolished the ability of the cardioprotective stimulus to reduce infarct size (Takashi et al., 1999). While these studies have yielded important mechanistic insights, it is noteworthy that mitoK_ATP channel opening also fails to evoke a cardioprotective response when repetitive preconditioning stimuli, such as multiple cycles of ischemia/reperfusion (Schwartz et al., 2002) or chronic exercise (Brown et al., 2005) are administered prior to the main insult, confounding the translation of this strategy to clinical use.

Few studies have examined the role of mitoK_ATP channels in the genesis of cardiac arrhythmias. A protective role for mitoK_ATP channel activation against arrhythmias has been inferred by experiments demonstrating that mitoK_ATP channel blockers consistently abolished the anti-arrhythmic phenotype provided by preconditioning stimuli, such as ischemic preconditioning (Vegh & Parratt, 2002; Rajesh et al., 2004), adenosine (Headrick et al., 2003), delta opioid agonists (Fryer et al., 2000; Fischbach et al., 2003), estrogen (Das & Sarkar, 2006), 3-nitropropionic acid (Basgut et al., 2008), nitroglycerin (Baharvand et al., 2009), noradrenaline (Imani et al., 2008), or endothelin receptor agonists (Das et al., 2007). It is important to note, however, that mitoK_ATP channel blockade during other preconditioning stimuli; namely, bradykinin (Driamov et al., 2004), low-flow ischemia (Driamov et al., 2004), peroxynitrite (Kiss et al., 2008), and estradiol (Tsai et al., 2002) failed to attenuate the anti-arrhythmic protection of these stimuli.

Studies investigating the efficacy of direct mitoK_ATP channel activation on the suppression of post-ischemic arrhythmias have yielded discrepant results (Schwartz et al., 2002; Headrick et al., 2003). One putative explanation for the discordant findings is that various pharmacological agents used to open mitoK_ATP channels are confounded by non-specific action. In fact, the non-specificity of mitoK_ATP channel openers (such as diazoxide) and blockers (such as 5-HD) has received considerable attention in recent years (Hanley et al., 2003; Suzuki et al., 2003; O’Rourke, 2004; Brown et al., 2005; Hanley et al., 2005). Moreover, mitoK_ATP channel activity is largely dependent on complex signaling cascades, including phosphorylation by protein kinase C (Ohnuma et al., 2002), which may be differentially altered in various studies.

While the preconditioning literature provides interesting mechanistic insights regarding anti-arrhythmic strategies administered before index ischemia, the clinical relevance of these strategies should be put into question. To the clinician, arrhythmia suppression must often be attempted after, not before, the onset of the ischemic insult. Targeting mitoK_ATP channels after the onset of metabolic stress seemed promising based on cellular studies, in which the administration of mitoK_ATP channel openers effectively inhibited ongoing Δψ_m oscillations that were evoked by halting respiration (Ryu et al., 2005). This strategy also improved cellular survival and mitochondrial integrity during cellular reoxygenation (Ozcan et al., 2007). Despite these encouraging cellular findings, post-ischemic administration of mitoK_ATP channel openers failed to decrease the incidence of arrhythmias (Das & Sarkar, 2005).

Mitochondrial Calcium Uniporter

Although altered intracellular calcium cycling and cytosolic calcium overload are well established sources of arrhythmia triggers and beat-to-beat repolarization abnormalities (Wilson et al., 2006), the role of mitochondrial calcium fluxes in the generation of arrhythmias remains unclear. Mitochondrial calcium homeostasis is achieved by balanced calcium influx into the matrix via the mitochondrial calcium uniporter (MCU) and efflux out of the matrix through the mitochondrial sodium–calcium exchanger. MCU blockade with ruthenium compounds has shown some promise in suppressing the incidence of arrhythmias. Specifically, pre-ischemic administration of both ruthenium red and Ru360 decreased the incidence of ventricular fibrillation upon reperfusion in rats (Garcia-Rivas Gde et al., 2006). Moreover, both compounds converted ongoing ventricular fibrillation to ventricular tachycardia when administered after the onset of arrhythmias, although neither compound led to sinus rhythm (Kawahara et al., 2003).

Mechanisms by which MCU blockade protects against arrhythmias are not well understood but may involve a decrease in the open channel probability of the mPTP by maintaining relatively low matrix calcium concentrations (Garcia-Rivas Gde et al., 2006). While this is largely expected to confer an anti-apoptotic effect, it seems unlikely to play a major role in arrhythmogenesis since blockers of the mPTP have not been particularly effective in preventing arrhythmias, as discussed previously. Indeed, these findings are supported by cellular experiments in which the reversible collapse in Δψ_m induced during RIRR was not prevented by either ruthenium red (Romashko et al., 1998) or Ru360 (Zorov et al., 2000).

The exact role of the mitochondrial calcium uniporter in arrhythmogenesis remains unclear because of major confounding effects of the ruthenium compounds on intracellular calcium fluxes (Griffiths, 2000). For example, Ruthenium red blocks calcium entry through L-type calcium channels (Vassilev et al., 1987) and release from the sarcoplasmic reticulum (Gupta et al., 1989), suggesting that the anti-arrhythmic efficacy of this compound may be related to its prevention of intracellular calcium overload and not to its primary mitochondrial target (Griffiths & Rutter, 2009). Ru360 appears to be more specific for the MCU, but whole heart experiments are confounded by permeability issues, with some investigators showing successful drug entry into myocytes (Kawahara et al., 2003) and others arguing against it (Robert et al., 2001; Bell et al., 2006). Consistent with their ability to reduce cytosolic calcium transients, both ruthenium compounds are potent negative inotropes at concentrations that protect against arrhythmias (Gupta et al., 1988; Kimura et al., 2005), an undesirable side effect when the overall purpose of administering the compound is to improve cardiac function. Future research using novel compounds that lack these pleiotropic/permeability issues will provide better insights into the role of the MCU in post-ischemic arrhythmias.

Anti-oxidant depletion as a mechanism of mitochondrial dysfunction and arrhythmias

Oxidative stress in cardiomyocytes is caused by either increased ROS production and/or reduced scavenging capacity. In fact, myocardial Glutatione (GSH), a main anti-oxidant defense system in myocytes, is a key regulator of RIRR and mitochondrial stability. Interestingly, depletion of the intracellular antioxidant GSH pool with diamide effectively triggers Δψ_m oscillations that are similar in nature to those generated by photo-induced oxidation of the myocyte (Aon et al., 2007). These observations were extended to the level of the whole heart, in which diamide treatment of ex vivo perfused hearts resulted in heterogeneous ROS production, Δψ_m depolarization (Slodzinski et al., 2004) and ventricular fibrillation (Brown et al., 2008a). Interestingly, reduced-to-oxidized glutathione ratio (GSH/GSSG) in whole heart homogenates following diamide administration was similar to that in isolated cells undergoing RIRR and Δψ_m oscillations (Aon et al., 2007). These findings are corroborated by human data, where low GSH/GSSG ratios were observed in human heart samples from patients with heart failure (Damy et al., 2009) and type 2 diabetes (Anderson et al., 2009), both important risk factors for cardiac arrhythmias and sudden death. Consistent with this notion, administration of N-acetylcycsteine significantly decreased the incidence of cardiac arrhythmias in patients following cardiac surgery (Ozaydin et al., 2008). While promising, N-acetylcycsteine itself is confounded by limited bioavailability (Holdiness, 1991) and anaphylactoid-like reactions (Holdiness, 1991). This clearly highlights the need for alternative compounds that can more effectively and safely restore GSH levels.

Finally, the redox state of the cardiomyocyte can also modulate its excitability properties through mitochondria-independent mechanisms. For example, increased oxidation has been shown to directly activate sarcK_ATP channels (Tokube et al., 1996), alter the inactivation kinetics of L-type calcium channels, decrease sodium current density (Liu et al., 2010), increase ryanodine receptor calcium ‘leak’ (Belevych et al., 2009), and modulate the activation state of mitochondrial inner membrane ion channels. Attempts to improve the redox status of the cardiomyocyte by scavenging ROS with superoxide dismutase mimetics (Konya et al., 1992) or mitochondria-targeted anti-oxidant peptides (Cho et al., 2007) were successful in decreasing the incidence of arrhythmias. Future experiments that optimize effective delivery of ROS-scavenging agents to mitochondria have clear potential in abrogating electrical abnormalities caused by metabolic dysfunction.

Spatio-temporal dynamics of mitochondrial function across the intact heart

As mentioned above, Δψ_m depolarization is triggered by opening of mitochondrial ion channels under conditions of oxidative stress (Weiss et al., 2003; O’Rourke, 2007; Brown et al., 2010). Specifically, during metabolic insults, increased mitochondrial ROS production from complex III of the electron transport chain triggers the opening of IMAC and/or mPTP (Weiss et al., 2003). This results in ROS release from mitochondria and Δψ_m depolarization. In isolated cardiomyocytes, ROS diffusion within the cytosol triggers further ROS release from neighboring mitochondria, initiating a feedback cycle of RIRR and Δψ_m depolarization (Zhou et al.,; Zorov et al., 2000; Aon et al., 2003; Brady et al., 2004).

Despite major advances in our understanding of mitochondrial biochemistry at the subcellular/molecular levels, the pathophysiological consequences of mitochondrial dysfunction at the level of the intact heart remained unclear. Since mitochondrial function of individual cells is highly influenced by network properties, it is critical to investigate mitochondrial function within the milieu of the intact heart (Weiss et al., 2006). We recently found that the metabolic substrate of the heart during the early onset of ischemia is spatially and temporally heterogeneous (Lyon et al., 2010b). These spatio-temporal heterogeneities in mitochondrial function may ultimately dictate myocardial excitability and contribute to the formation of zones of conduction block by heterogeneous activation of surface K_ATP channels, as we had previously speculated (Akar et al., 2005).

Δψ_m depolarization

A semi-quantitative approach of optical Δψ_m imaging in the ex vivo perfused heart allowed the identification of waves of Δψ_m depolarization that actively propagate across the myocardium with a mean velocity of ~20μm/sec (Figure 2), several orders of magnitude slower than myocardial action potential propagation (Lyon et al., 2010a). We further elucidated complex spatio-temporal metabolic instabilities that preceded and accompanied the formation of these organized waves (Figure 2). Furthermore, we identified at the tissue level the presence of Δψ_m ripples prior to mitochondrial collapse during ischemia. These data suggested patterns of wave behavior spreading across the myocardium ahead of the main wave of Δψ_m depolarization, with propagation reflecting the direct interaction between adjacent cells within the intact ischemic tissue. Although we did not directly image ROS levels, it is conceivable that ROS diffusion at the interface between depolarized (acting as ROS sources) and polarized (ROS sinks) regions can drive the propagation of Δψ_m collapse, in a manner that extends the notion of RIRR from a subcellular to a multi-cellular phenomenon. The amplification and propagation of Δψ_m depolarization across the electrically coupled syncytium may present novel opportunities to limit injury by potentially targeting areas of early Δψ_m collapse that form the origin of the organized propagating wavefront of mitochondrial dysfunction.

Figure 2. Spatio-temporal fluctuations of Δψ_m during global ischemia (Adapted from Figure 4, Lyon et al. J Mol Cell Cardiol. 2010, PMID: 20624394).

Successive contour maps of normalized Δψ_m (above) and its first derivative (below) acquired at 10, 40, 70, and 180 seconds following the onset of global no-flow ischemia in a representative rat heart. These data illustrate the presence of spatially and temporally discordant kinetics of Δψ_m that exist ahead of the main depolarization wave of Δψ_m collapse, which actively propagates across the heart. Color scale: a) Δψ_m contour maps: baseline (black), depolarization (red), hyperpolarization (yellow); b) δΔψ_m/δt contour maps: baseline (black), positive slopes (turquoise), negative slopes (purple).

The importance of Δψ_m kinetics at the tissue level was also highlighted in a recent study in which cardiac arrhythmias induced by GSH oxidation were effectively inhibited by preventing Δψ_m depolarization using IMAC blockade (Brown et al., 2010). Paradoxically, we also recently found that Δψ_m depolarization was completely prevented in hypertrophied hearts that were challenged with short episodes of ischemia (Jin et al., 2010). Protection against Δψ_m depolarization in this rat model of ascending aortic banding was not, however, associated with protection against arrhythmias (Jin et al., 2010).

Finally, in embryonic mouse hearts, Chen et al (Chen et al., 2007) elegantly investigated the differential effects of inhibiting glycolysis versus oxidative phosphorylation on Δψ_m depolarization and arrhythmia propensity. While inhibition of oxidative phosphorylation but not glycolysis caused a major depolarization in Δψ_m, both strategies led to comparable slowing of heart rate, shortening of the action potential duration, blunting of the intracellular calcium transients, and promotion of arrhythmias (Chen et al., 2007). Of note is the fact that the developing myocardium is more dependent on glycolysis than is the adult heart.

Δψ_m Recovery

Prompt reperfusion is required for preventing irreversible cell damage and death. Unfortunately, restoration of blood flow, in itself, results in additional cardiac damage, known as reperfusion injury, which results from large bursts of ROS (Bolli et al., 1989). ROS-mediated oxidative damage is more severe when reperfusion therapy is delayed. Effective strategies to limit or prevent reperfusion injury have proven elusive. Despite an improved understanding of the pathophysiology of this process, the vast majority of clinical trials aimed at preventing reperfusion injury have been quite disappointing. We recently demonstrated that the successful recovery of Δψ_m upon reperfusion is indeed highly dependent on the duration of the preceding ischemic episode. Despite a comparable degree of Δψ_m depolarization following 7.5 and 15 minutes of global no-flow ischemia in the rat, reperfusion led to recovery of Δψ_m only following the short (7.5 min) but not longer episodes of ischemia (Lyon et al., 2010b). Interestingly, sustained Δψ_m recovery was also predictive of post-ischemic functional and electrical recovery (Lyon et al., 2010b). These findings reinforce the notion that reperfusion is a highly complex phenomenon which could either reverse or exacerbate ischemia mediated changes in Δψ_m. In fact, additional Δψ_m depolarization upon reperfusion following long episodes of ischemia is consistent with ROS induced damage during this phase (Lyon et al., 2010b). Strategies aimed at promoting rapid recovery of Δψ_m during the early (first 5 minutes) phase of reperfusion, potentially by ischemic or pharmacologic post-conditioning strategies, may be an effective strategy for avoiding the genesis of ventricular fibrillation (Lyon et al., 2010b).

Metabolic sinks and reperfusion arrhythmias

Spatio-temporal heterogeneities in mitochondrial function may be associated with local changes in sarcK_ATP current density which could potentially create areas of depressed excitability to form conduction block through a mechanism we termed “metabolic sink” (Akar et al., 2005). The presence of metabolic sinks may promote the genesis of arrhythmias by shortening the effective refractory period and slowing myocardial conduction in the area of the sink; thereby, shortening the excitation wavelength. Moreover, presence of heterogeneous metabolic sinks is expected to promote heterogeneous action potential repolarization across the tissue. Finally, having a discrete region or dispersed loci of metabolic sinks may predispose to arrhythmias either by forming unidirectional conduction block or causing heterogeneous conduction, respectively. In support of the concept of metabolic sink, IMAC activation using agonists of the mitochondrial benzodiazepine receptor led to an accelerated shortening of the action potential and an early form of conduction failure during ischemia. In contrast, IMAC blockade delayed action potential shortening and the onset of inexcitability (Akar et al., 2005). In this guinea pig model, sustained ventricular tachyarrhythmias were readily generated upon reperfusion in ~90% of hearts (Akar et al., 2005). Remarkably, IMAC blockade, which stabilizes Δψ_m in vitro, markedly suppressed the formation of these arrhythmias. Indeed, these data suggest that mitochondrial depolarization is the primary factor driving K_ATP channel activation in ischemia and arrhythmias upon reperfusion. The protective effect of IMAC blockade on electrical and contractile post-ischemic function was further demonstrated in a rabbit model of ischemia reperfusion injury (Brown et al., 2008b). This anti-arrhythmic effect was not evident in hearts treated with the mPTP blocker, cyclosporine A, reinforcing IMAC as the primary mitochondrial mediator of post-ischemic arrhythmias. This concept of metabolic sinks is strengthened by our Δψ_m imaging studies, which revealed complex spatio-temporal dynamics of Δψ_m properties that were closely related to post-ischemic electrical and contractile recovery (Lyon et al., 2010a). Finally, the dependence of electrical dysfunction on Δψ_m was recently argued in hearts that did not undergo ischemia-reperfusion injury, but rather, were challenged with diamide-induced glutathione oxidation. Again, IMAC blockade was effective in preventing both Δψ_m depolarization and arrhythmias in this model of metabolic stress (Brown et al., 2010).

Mitochondria as therapeutic targets

Cardiac mitochondria form a compact three dimensional lattice structure that is tightly packed between myofilaments and surrounding t-tubules. This spatial organization places mitochondria in close proximity to the major sites of energy consumption (myofilaments) and excitation-contraction coupling (diads). By being the major source of ROS production, mitochondria can intricately alter the activity of multiple ion channel, Ca²⁺ handling and contractile proteins. Moreover, the generation of metabolic intermediates within mitochondria provides the reducing equivalents required to maintain the negative redox potential of cellular antioxidant pathways. As such, mitochondria clearly represent an attractive target for altering myocyte function, including electrophysiological properties.

Uncovering mechanisms by which mitochondrial dysfunction predisposes to arrhythmias will allow us to design novel strategies. Targeting root causes (ie mitochondria) rather than downstream consequences (cell surface membrane transporters, calcium cycling proteins, etc) is expected to be advantageous as mitochondria represent a main hub of myocyte function that controls energetics, cell signaling, calcium handling and electrical function.

The development of effective therapeutic strategies targeting the mitochondrial network is currently hampered by a lack of solid molecular information regarding the identity of key mitochondrial ion channels and transporters. For example, none of the proteins involved in mitochondrial Ca²⁺ homeostasis have thus far been completely resolved. Pharmacological studies point us towards promising targets such as the IMAC, mPTP, mitoK_ATP, and MCU, but actual mitochondrial structures and macromolecular complexes that mediate changes in Δψ_m remain a subject of intensive debate and active investigation. Indeed, this field of mitochondrial biology is ripe for discovery as powerful proteomic and genomic tools become more readily available. Meanwhile, integrative multi-scale investigation, involving complementary in vivo, ex vivo, in vitro, and in silico approaches is essential for understanding how metabolic failure at the level of the organelle can scale to produce arrhythmias in the whole heart.