Mitochondrial Ca²⁺ in Heart Failure: Not Enough or Too Much?

Abstract

Ca²⁺ serves as a ubiquitous second messenger mediating a variety of cellular processes including electrical excitation, contraction, gene expression, secretion, cell death and others. The identification of the molecular components of the mitochondrial Ca²⁺ influx and efflux pathways has created a resurgent interest in the regulation of mitochondrial Ca²⁺ balance and its physiological and pathophysiological roles. While the pace of discovery has quickened with the availability of new cellular and animal models, many fundamental questions remain to be answered regarding the regulation and functional impact of mitochondrial Ca²⁺ in health and disease. This review highlights several experimental observations pertaining to key aspects of mitochondrial Ca²⁺ homeostasis that remain enigmatic, particularly whether mitochondrial Ca²⁺ signaling is depressed or excessive in heart failure, which will determine the optimal approach to therapeutic intervention.

Cellular Ion transport is maintained through ion pumps, exchangers and channels coupled to the mitochondrial proton circuit

A complete appreciation of the role of mitochondria in cellular ion homeostasis requires consideration of all of the coupled ion circuits in the cell (Figure 1)[1]. The cation gradients across the sarcolemma, mitochondrial and sarcoplasmic/endoplasmic reticulum (SR) are maintained by ATP-driven pumps, thus, they indirectly depend on the protonmotive force of the proton circuit. Cation movements across the mitochondrial inner membrane also depend on the proton gradient, as any influx of Ca²⁺ K⁺ or Na⁺ is compensated by exchange with H⁺. It is also important to understand the necessity of maintaining charge balance of cations and anions across the closed compartments of the cell. Inorganic phosphate (P_i) is the major anion transported across the mitochondrial inner membrane and, in mammalian mitochondria, the P_i/H⁺ carrier (or P_i/H₂CO₃ antiport, gene SLC25A3) accounts for the majority of Pi flux (although other modes of Pi-coupled transport are also present).

Figure 1.

Coupling of ion circuits to the mitochondrial proton circuit. Efflux of K⁺ and Na⁺ across the mitochondrial inner membrane is linked to the proton circuit by specific K⁺/H⁺ and Na⁺/H⁺ exchangers, while Ca²⁺ efflux is indirectly linked via Na⁺/Ca²⁺ exchange. Influx of K⁺ and Ca²⁺ is driven by ΔΨ_m and the concentration gradient for each ion, restricted by the need to maintain electroneutrality, which requires concomitant anion movement, principally through the inorganic phosphate (P_i) transporter[85]. Large amounts of mitochondrial Ca²⁺ can be buffered by the formation of Ca²⁺ - P_i complexes but excessive Ca²⁺ loading of the mitochondrial matrix can trigger the activation of the permeability transition pore. The mitochondrial proton circuit also indirectly controls resting transarcolemmal and sarcoplasmic reticular (SR) cation gradients through the action of ATP-dependent ion pumps (Na⁺/K⁺ ATPase, Ca²⁺ ATPases). Mitochondrial Ca²⁺ is required to match energy supply with demand through its effect on Ca²⁺ - sensitive enzymes in the tricarboxylic acid (TCA) cycle, primarily the pyruvate dehydrogenase complex, as well as by activating sites in the electron transport chain. Schema is adapted and expanded from ref. [1].

Machinery of mitochondrial Ca²⁺ influx and efflux

Ca²⁺ influx:

A watershed moment in the study of mitochondrial Ca²⁺ regulation was the discovery, by two independent groups, of the pore component of the mitochondrial Ca²⁺ uniporter (MCU) in 2011[2, 3], which came on the heels of description of MICU1 as a regulator of mitochondrial Ca²⁺ uptake in 2010[4]. These finding resolved one of the longstanding questions (>50 years) in mitochondrial biology and stimulated renewed interest in the role of mitochondrial Ca²⁺ in physiology. Since that time, other members of the uniporter complex have been identified, fulfilling roles as either assembly factors or regulatory proteins. The current view of the Ca²⁺ uptake complex includes the pore, a multimer (probably a dimer of dimers [5]) of MCU, a 35 kDa coiled-coil domain protein with 2 transmembrane helices linked via a pore-forming domain containing a conserved motif (DIME) necessary to maintain Ca²⁺ permeability; a dominant negative homolog (~50%) of MCU (MCUb), which contains 2 substitutions (R>W and E>V) proximal to the DIME domain that nullify Ca²⁺ conductance through the multimer; 2 homologous regulatory partners MICU1 and MICU2, EF-hand containing proteins which act as gatekeepers of the pore; EMRE a facilitator of complex assembly; and MCUR1, which might act as an interaction partner between EMRE and MCU[6], although an alternative view is that it is a complex IV assembly factor[7] (Figure 2).

Figure 2.

Schematic depiction of the components of the MCU complex. Ca²⁺ flux through the uniporter is thought to be modulated by conformational changes in MICU1 through binding of Ca²⁺ to its EF-hand regions. The uniporter mediates two modes of mitochondrial Ca²⁺ transport[86], i.e., a low flux, high Ca²⁺ affinity mode and a high flux, low Ca²⁺ affinity mode, with differential sensitivities to Ru360, which was also linked to MICU1 interaction with the pore[9].

MICU1, in the most widely accepted model, faces the mitochondrial intermembrane space[8] and is thought to suppress Ca²⁺ influx through MCU at low (<1μM) Ca²⁺ levels. Upon Ca²⁺ binding to the EF-hand region of MICU1 this inhibition is relieved, resulting in positive cooperativity of uptake at higher Ca²⁺ levels. Hence, MICU1 appears to control the threshold and gain properties of the MCU[8], as well as its sensitivity to ruthenium 360 (Ru360)[9]. Knockdown or knockout of MICU1 increases mitochondrial Ca²⁺ loading at resting Ca²⁺ levels. In whole animal knockouts, this results in multiple physiological abnormalities, including perinatal lethality and severe neurological and myopathic defects, which could be rescued by knockout of a single allele of EMRE[10]. In another study, MICU1 was proposed to face the matrix compartment and interact with the N-terminus coiled-coiled domains of the MCU complex, affecting channel complex oligomerization[11]. Using a FRET-based method for assessing MICU1 conformational rearrangement, a recent study reported that MICU1 multimers respond only to changes in cytosolic, not matrix, Ca²⁺, with an EC50 of 4.4 μM[12]. Despite differences in the mechanistic details, there is consensus that MICU1 suppresses MCU activity at low Ca²⁺ levels. MICU2 also acts as a gatekeeper for MCU function; however, it requires the presence of MICU1, whereas MICU1 can operate independently of MICU2. MICU2, by modifying the gain and threshold of Ca²⁺uptake, contributes to the spatial restriction of Ca²⁺ signals between the InsP₃ Receptor and the MCU[13].

Ca²⁺ efflux:

Na⁺-dependent Ca²⁺ efflux from cardiac mitochondria was first described by Carafoli in 1974[14] and, distinct from its sarcolemmal cousin, was found to be able to exchange Li⁺ for Ca²⁺ as well. The mitochondrial Na⁺ [or Li⁺]/Ca²⁺ exchanger (NCLX) was shown to be inhibited by certain benzothiazepines and benzodiazepines[15], including the clonazepam derivative CGP-37157[16], which has minimal effects on sarcolemmal NCX, the Na⁺/K⁺ ATPase or the L-type Ca²⁺ channel at low μM concentrations. Despite these distinct tools and properties, it took until 2010, just prior to the MCU findings, before the NCLX protein was identified as the gene product of SLC8B1 (aka SLC24A6)[17]. The importance of this seminal finding to cardiac physiology was validated by heart-specific knockout of the gene in mice, which can have a devastating effect on survival. When NCLX knockout was induced in adults, only 13% of these mice survived more than two weeks and severe cardiomyopathy was evident; however, if NLCX was knocked out early in development, viability was normal[18]. This implies that compensatory adaptations, including diminished Ca²⁺ uptake, occurred during the maturation of the animal to limit injury. The exacerbation of heart failure (HF) by knockout of NCLX prompted this group to further investigate the role of this exchanger in the progression of HF. Employing an inducible transgenic overexpressor of NCLX, they showed that post-myocardial infarction cardiac dysfunction was mitigated by increasing NCLX activity by 88%[18].

Physiological and pathophysiological roles of mitochondrial Ca²⁺

The ability of mitochondria to take up and store large amounts of Ca²⁺ has been a topic of great interest since the 1960’s and speculation about the physiological role of mCa²⁺ extends equally far back, proposed as an organelle mediating Ca²⁺ storage for muscle excitation-contraction coupling, biological calcification[19], and later as the signal by which Krebs cycle enzymes are activated during periods of increased energy demand[20]. The observation that mitochondrial Ca²⁺ uptake could lead to catastrophic effects on mitochondrial integrity, through the activation of a permeability transition pore (PTP), was also recognized very early[21]. To this day, defining the physiological and pathophysiological role of mitochondrial Ca²⁺ remains a subject of active investigation and controversy.

Coupling of mitochondrial Ca²⁺ to pyridine nucleotide redox balance

The heart continuously adjusts the rate of oxidative phosphorylation to changes in workload that can vary more than 5-fold[22]. Increased energy demand is accompanied by an acceleration of electron transport and proton pumping driven by NADH oxidation. In order to maintain the redox potential of the NADH/NAD⁺ redox couple, an increase in the rate of NADH production by the tricarboxylic acid (TCA; a.k.a. Krebs) cycle must occur to avoid a decrease in the protonmotive force. The matching of energy supply to demand thus involves “parallel activation” of fluxes upstream and downstream of the electron transport chain[23]. Mitochondrial Ca²⁺ uptake during increased work is required for this matching to occur. In healthy adult myocytes, during an increase in work (from rest to fast pacing with β-adrenergic stimulation) there is a transient decrease and recovery of NADH during stimulation and an NADH overshoot after cessation of stimulation (Figure 3A)[24]. Similar NADH/NAD⁺ dynamics were observed in cardiac trabeculae[25] and were shown to parallel the time course of mitochondrial Ca²⁺ increase during increased work[26]. Inhibition of the MCU with intracellular Ru360, which (at ≤ 100nM) inhibits mitochondrial Ca²⁺ uptake with minimal effects on SR Ca²⁺ release, prevented acceleration of the TCA cycle during increased energy demand, resulting in net oxidation of the mitochondrial NADH/NAD⁺ pool (Figure 3B).

Figure 3.

Mitochondrial Ca²⁺ signaling is required for redox balance in healthy adult cardiac myocytes. During increases in cardiac work, pyridine nucleotide redox balance is maintained by parallel activation of NADH oxidation (increase in electron transport chain through ADP-linked respiration) and NADH production (increased TCA cycle activity). The latter requires mitochondrial Ca²⁺ signaling. A) In cardiac myocytes, mitochondrial free Ca²⁺ ([Ca²⁺]_m) rises while NADH transiently oxidizes (undershoot) and then recovers during 4Hz pacing (in the presence of isoproterenol). Cessation of pacing results in a transient increase in NADH above baseline (overshoot) as NADH production temporarily exceeds its oxidation rate (data from ref. [24]). B) In the presence of intracellular Ru360, the rise in [Ca²⁺ ]_m during pacing is suppressed, and NADH oxidizes without recovering during pacing, or after disc ontinuing electrical stimulation (data from reference [37]).

Factors contributing to impaired mitochondrial Ca²⁺ - redox coupling in HF

Several factors known to be altered in failing hearts will influence mitochondrial Ca²⁺ - redox coupling. First, since it is a relatively low affinity process, mitochondrial Ca²⁺ uptake is facilitated by close proximity of the mitochondria to the Ca²⁺ release units of the diad (within 40nm)[27]. Local Ca²⁺ concentration can reach 10s of micromolar near the Ca²⁺ release sites at moderate sarcoplasmic reticular (SR) Ca²⁺ loads, further enhanced by positive inotropic stimulation[27]. Mitochondria are, therefore, poised to integrate the cytosolic Ca²⁺ signal in a frequency and amplitude dependent manner[24]. Frequency-dependent potentiation of the Ca²⁺ transient is markedly impaired in HF[28], as a result of depressed SR Ca²⁺ ATPase (SERCA2) function, impaired β-adrenergic signaling, and increased diastolic leak that significantly decrease SR Ca²⁺ load[29, 30]. Second, the ultrastructure of both the t-tubular and mitochondrial network is disrupted in HF. T-tubule loss[31] and fragmentation of the mitochondrial network[32] will impair close excitation-metabolic coupling. Third, a universal finding in myocytes from failing human hearts[33] and animal models is cytosolic Na⁺ overload, which may result from impaired Na⁺/K⁺ ATPase activity or increased late Na⁺ current[34]. The change in Na⁺ gradient has substantial effects on the electrochemical driving force for the sarcolemmal Na⁺/Ca²⁺ exchanger (NCX), which plays a much more dominant role in Ca²⁺ removal in HF when SERCA2 activity is decreased[35]. Increased cytosolic Na⁺ could, in theory, partially compensate for impaired excitation-metabolic coupling by offsetting the depressed SR Ca²⁺ load; however, it will have a direct negative impact on mitochondrial Ca²⁺ loading due to the presence of the mitochondrial Na⁺/Ca²⁺ exchanger (NCLX). Cardiac Na⁺-dependent mitochondrial Ca²⁺ efflux has a Km of ~8mM, which is in the middle of the range over which cytosolic Na⁺ changes during the development of HF (in a guinea pig HF model intracellular Na⁺ increased from 5 to 17 mM[36]). Impaired mitochondrial Ca²⁺ - redox coupling, similar to that observed in Ru360-treated myocytes (above), was evident in myocytes from failing hearts; increased workload was accompanied by NADH oxidation[37] (Figure 4A) and increased oxidative stress (Figure 4B), and mitochondrial Ca²⁺ loading during pacing was depressed (Figure 4C). Corroborating the in vitro depression of mitochondrial Ca²⁺ loading during EC-coupling in HF myocytes, in vivo chronic mitochondrial Ca²⁺ underloading in a guinea pig HF/SCD model is indicated by increased specific phosphorylation of regulatory sites on the pyruvate dehydrogenase (PDH) complex, which was prevented by mitochondrial antioxidant therapy[38](Figure 4C). Normally, Ca²⁺-dependent activation of a pyruvate dehydrogenase phosphatase switches on PDH complex activity to increase TCA cycle activity, so hyperphosphorylation is a biomarker of chronically decreased mitochondrial Ca²⁺ signaling. Impaired mitochondrial Ca²⁺ - redox coupling similar to the HF condition can be mimicked in normal myocytes by elevating cytosolic Na⁺ to 15 mM or by treatment with the Na⁺ pump inhibitor ouabain[39]. Interestingly, mitochondrial Ca²⁺ signaling and NADH redox homeostasis could be enhanced in HF myocytes by partial inhibition of mitochondrial Na⁺/Ca²⁺ exchange with the benzothiazepine CGP-37157[36] or by lowering cytoplasmic Na⁺[40], indicating that a large part of the defect was attributable to the effects of Na⁺ overload on mitochondrial Ca²⁺. A clear “threshold” of mitochondrial Ca²⁺ rise during pacing is required to prevent oxidation of the NADH pool, below which the degree of NADH oxidation is inversely correlated with mitochondrial Ca²⁺ loading[37]. Notably, in our guinea pig model of HF/SCD, there are no significant changes in transcript[41] or protein expression levels of MCU, MCUR1, EMRE (Smdt1), MICU1, or MICU2[38, 41], suggesting that the deficit in mitochondrial Ca²⁺ signaling in HF is likely related to changes in local ionic gradients or microdomain structure rather than a change in transcription/translation of the complex.

Figure 4.

Ventricular myocytes from failing hearts fail to maintain NAD(P)H redox potential during increased work and have blunted mitochondrial Ca²⁺ responses. A) Rapid pacing in the presence of β-adrenergic stimulation induces net NAD(P)H oxidation in cells from failing hearts (red; ACi guinea pig model) while healthy cells maintain pyridine nucelotide redox (black). Treatment with CGP-37157 (CGP), to partially inhibit mitochondrial Na⁺/Ca²⁺ - mediated Ca²⁺ efflux, restores redox balance in failing cells (green). B) Oxidation of the NAD(P)H pool was associated with cellular ROS accumulation during pacing in cells from failing hearts (red) but not healthy controls (black). CGP suppressed ROS accumulation in the failing group (green). C) The changes in redox/ROS balance and the protection by CGP could be attributed to corresponding changes in mitochondrial matrix Ca²⁺ signaling during pacing. Panels A-C are from ref. [36]. D) Hyperphosphorylation of PDHA1, a subunit of the pyruvate dehydrogenase (PDH) complex, at its regulatory sites is evident in ventricular mycocardium of failing hearts, which was prevented by in vivo treatment with a mitochondrial antioxidant (mitoTEMPO). PDH phosphorylation is a biomarker of chronic mitochondrial Ca²⁺ underloading (data from ref. [38]).

Effects of impaired mitochondrial Ca²⁺- redox coupling on ROS balance

The consequences of impaired mitochondrial Ca²⁺ - redox coupling extend beyond just maintaining the NADH/NAD⁺ redox potential for oxidative phosphorylation. Because the mitochondrial glutathione- and thioredoxin-dependent antioxidant pathways depend on NADPH redox potential, which is sustained by enzymes that also utilize TCA cycle intermediates, failure of energy supply and demand matching can also impact ROS scavenging, required to continuously detoxify ROS produced by the electron transport chain (Figure 5). When the rate of ROS production exceeds the scavenger capacity, ROS overflow from the mitochondria occurs[42]. In addition, the ability of mitochondria to act as scavengers of extramitochondrial ROS is compromised, increasing the damaging effects of other ROS sources, such as NADPH oxidase[38, 43]. Hence, impairment of mitochondrial Ca²⁺ signaling can increase oxidative stress (cf. Figure 4B), which can then impact excitation-contraction coupling, progression of decompensated HF, and sudden cardiac death from arrhythmias[36]. Interventions designed to either enhance mitochondrial Ca²⁺ signaling (e.g., CGP-37157)[36] or increase ROS scavenging using mitochondrially-targeted scavengers[38, 44, 45] can prevent HF progression and improve survival.

Figure 5.

Excitation-contraction-redox coupling. Parallel activation of NADH oxidation, due to oxidative phosphorylation, and the production of NADH and NADPH by TCA cycle intermediates is required to maintain pyridine nucleotide redox potential. Depressed SR Ca²⁺ load and elevated cytosolic Na⁺ in heart failure decrease the mitochondrial Ca²⁺ signal resulting in impaired NADPH-driven antioxidant flux to remove ROS produced as a byproduct of respiration or from external sources. The increased ROS affects various redox-sensitive targets in the mitochondria (PTP) and in the diad (SERCA pump, L-type Ca²⁺ channel and the ryanodine receptor), which exacerbates Ca²⁺ dysregulation and arrhythmias. An alternative view is that spontaneous SR Ca²⁺ leak during diastole could be the lead factor causing mitochondrial Ca²⁺ overload, metabolic dysfunction and arrhythmias in HF[87]. Cytoplasmic ROS scavenging pathways, which also depend on analagous glutathione and thioredoxin-depedent enzymes driven by NADPH (primarily generated by glucose-6-phosphate dehydrogenase), are not shown, nor are important soruces of extamitochondrial ROS (see text for discussion). NNT, nicotinamide nucleotide transhydrogenase; IDH2, isocitrate dehydrogenase; ME3, malic enzyme; GSR, glutathione reductase; GPX4, glutathione peroxidase; Txrnd2, thioredoxin reductase; PRX3, peroxiredoxin; SOD2, superoxide dismutase; MCU, mitochondiral calcium uniporter; NCLX, Na⁺ [Li⁺]/Ca²⁺ exchanger; VDAC, outer membrane voltage-dependent anion channel; FFAs, free fatty acids; Gluc, glucose; Pyr, pyruvate.

TCA cycle intermediates drive NADH and NADPH formation (Figure 5) while the rates of electron transport and antioxidant flux govern, respectively, NADH and NADPH oxidation. Recent studies have also emphasized the importance of the reversible nicotinamide nucleotide transhydrogenase (NNT) reaction as a factor contributing to the severity of HF. NNT is a transmembrane protein in the mitochondrial inner membrane that harnesses ΔΨ_m to drive the forward reaction from NADH to NADPH. Thus, in energized mitochondria, the high NADH/NAD⁺ redox potential can provide reducing equivalents to support the NADPH requiring antioxidant pathways. However, if the protonmotive force (and NADH/NAD⁺) drops or mitochondria become uncoupled, high rates of respiration may deplete NAPDH through the reversal of the NNT reaction. A recent study[46] exploited the spontaneous NNT knockout present in C57Bl/6J mice to investigate whether the presence or absence of functional NNT altered outcomes in transverse aortic constriction (TAC) - induced HF. Compared to C57Bl/6N mice, which have a functional NNT, C57Bl/6J mice showed improved cardiac function after TAC. The authors concluded that reverse NNT activity might compromise NADPH availability and increase mitochondrial ROS overload during high metabolic stress. Determining the direction of the NNT reaction in vivo is currently not possible, and it is unclear under what conditions it might switch from a protective to a destructive mode of action. When NADH and NADPH are directly coupled via NNT, a healthy TCA cycle can supply reduced pyridine nucleotides to both oxidative phosphorylation and antioxidant enzymes. On the other hand, when the TCA cycle or electron transport chain are impaired, coupling could lead to a drain in NADH by ROS scavenging or to a drain in NADPH by electron transport, depending on the direction of the reaction.

While we have focused on primary imbalances of mitochondrial Ca²⁺, ROS derived from the electron transport chain, and redox balance in this review, it is important to highlight that there are many sources of ROS and reactive nitrogen species that may contribute to myocardial and vascular dysfunction in heart failure. These include NADPH oxidases in the sarcolemma (NOX2)[47] and the mitochondria (NOX4)[48, 49], monoamine oxidases in the mitochondrial intermembrane space[50], xanthine oxidase[51], and superoxide generated by uncoupled nitric oxide synthase (NOS)[52]. While NO derived from constitutively expressed endothelial NOS (NOS3) may be protective in HF[53], inducible NOS (NOS2) can increase in chronic heart failure in response to oxidative or inflammatory stress[54] and impair contractile reserve[55]. In addition, neuronal NOS (NOS1) contributes to contractile dysfunction[57] and appears to be present in mitochondria, activated by Ca²⁺-calmodulin[56]. Peroxynitrite formation, from the reaction of superoxide with nitric oxide, can be particularly damaging, due to the high reactivity of peroxynitrite with lipids and proteins. This can further compromise mitochondrial ROS scavenging, for example, through tyrosine nitrosylation and inactivation of MnSOD[58]. The variety of documented sources of ROS and reactive nitrogen species contributing to oxidative stress in heart failure further emphasizes how important it is to maintain mitochondrial and cytosolic antioxidant fluxes. Both compartments depend on maintaining the NADPH pool in the reduced state, in the former case coupled to the TCA cycle activity (Figure 5), and in the latter, through the pentose phosphate pathway and glucose-6-phosphate dehydrogenase, with significant crosstalk between cytosolic and mitochondrial scavenging pathways [59].

Mitochondrial ROS imbalance and increased SR Ca²⁺ leak in heart failure

Oxidative stress is known to be an important component of contractile dysfunction and arrhythmogenesis in HF, and mitochondrial ROS are involved[38]. It is important to note that mitochondria are not only a source of ROS but are also efficient ROS scavengers, converting the oxidation of metabolic substrates into the reducing equivalents required for the operation of antioxidant pathways. Moreover, their close proximity to the diad suggests that mitochondria will influence the local environment of redox sensitive proteins involved in excitation-contraction coupling, such as the SR Ca²⁺ ATPase, the L-type Ca²⁺ channel, and the SR Ca²⁺ release channel, RYR2. Ryanodine receptors are particularly sensitive to their thiol redox state, with increased oxidation corresponding to increased open probability[60]. A role for mitochondria in modulating RYR activity has been demonstrated in several studies. For example, resting Ca²⁺ spark rates vary inversely, and reversibly, with ΔΨ_m and reduced glutathione (GSH) levels, attributable to mitochondrial ROS-induced ROS release[61]. Mitochondrial ROS emission was found to induce SR Ca²⁺ leak in high fat-fed obese rat skeletal muscle[62] (upstream of CaMKII activation), in ageing rabbit cardiomyocytes[63], and in response to increased energy demand[64]. In a canine heart failure model, SR Ca²⁺ leak and SR Ca²⁺ depletion were attributed to redox modification of RyR2[65], although the specific role of mitochondrial ROS was not explored. While suppression of mitochondrial ROS consistently protects against the progression of heart failure and mitigates arrhythmias[38, 44, 45], it remains challenging to determine if altered mitochondrial redox balance precedes or follows aberrant Ca²⁺ handling, as a positive feedback loop may amplify the pathophysiology, i.e., increased ROS can increase SR Ca²⁺ leak, which increases energy demand, inducing more ROS, CamKII activation, etc. In line with the concept that enhancement of mitochondrial Ca²⁺ uptake can interrupt this loop, efsevin (a VDAC modulator) or kaempferol (a natural flavanoid), reportedly suppressed arrhythmias in a mouse model of CPVT or induced pluripotent stem cell-derived myocardial cells from a CPVT patient[66]. In contrast, kaempferol or CGP-37157 exacerbated spontaneous SR Ca²⁺ release and increased oxidation of the RyR in normal or hypertrophied rat myocytes[67], once again emphasizing the lack of a definitive answer to the chicken-and-egg dilemma.

Negative impact of mitochondrial Ca²⁺ uptake during heart failure

The potential negative impact of mitochondrial Ca²⁺ to induce cell death and cause cardiac dysfunction must also be considered in the context of stressors leading to heart failure. This is most conspicuous in the face of acute ischemia-reperfusion, where mitochondrial Ca²⁺ overload during reperfusion and PTP opening is thought to be a major contributor to tissue injury. Surprisingly however, global MCU knockout in the mouse did not diminish infarct size or improve functional recovery[68], suggesting that adaptations to lifelong MCU gene ablation fundamentally changed the organism’s response to ischemia-reperfusion[69]. Similarly, diminished fight or flight responses[70] but no effect on ischemic injury, were reported for suppression of MCU activity by overexpression of a dominant-negative MCU subunit[71]. Nevertheless, cardiomyocyte specific MCU knockout prevented the activation of the mitochondrial permeability transition pore, decreased infarct size, and preserved cardiac function after ischemia-reperfusion injury[72, 73]. In these mice, β-adrenergic responses were abrogated, maximal bioenergetic reserve capacity was reduced, and the ability to activate PDH was diminished, although a recent paper using inducible cardiac specific knockout mice gave contradictory results[74]. Interestingly, neither global[75], nor conditional MCU knockout[73] protected against TAC - induced heart failure. In the latter study, it is important to note that mitochondrial Ca²⁺ signaling was still present, albeit slowed, in MCU knockout hearts, implying that alternative pathways exist that could preserve some level of Ca²⁺-dependent energy supply and demand matching in the absence of MCU[73].

A recent study implicated excessive mitochondrial Ca²⁺ loading during excitation-contraction coupling as a factor in arrhythmogenesis in a model of HF secondary to hypertension induced by unilateral nephrectomy, deoxycorticosterone acetate (DOCA) treatment, and salt loading[76]. In an uncontrolled positive feedback cycle, mitochondrial Ca²⁺ overloading contributed to afterdepolarizations and triggered arrhythmias. Prolonged action potentials with multiple early afterdepolarizations were observed that could be suppressed by inclusion of Ru360 (1μM) in the intracellular solution or by MCU knockdown. The authors proposed that MCU-mediated Ca²⁺ uptake was enhanced in the model and that Ca²⁺ taken up during systole may be spontaneously released to increase the probability of afterdepolarization. In contrast to the studies in guinea pigs described above, in which enhancement of mitochondrial Ca²⁺ uptake was therapeutic, these findings suggest that inhibition of MCU would be beneficial.

Given the disparity of conclusions regarding the role of mitochondrial Ca²⁺ in HF, that is, either no effect of MCU knockout in mouse TAC models, too much mitochondrial Ca²⁺ in the DOCA model, or too little mitochondrial Ca²⁺ in guinea pig HF, the question arises as to what is different between the various studies. Focusing on differences in species and the model of HF, there are distinct factors that could contribute to the discrepant findings. For example, the high heart rate (>600 bpm) of the mouse requires it to operate, most of the time, close to Vmax for oxidative phosphorylation, with little range for metabolic fine tuning[77]. Thus, mice might not require the same physiological mitochondrial Ca²⁺ signaling mechanisms as other species. Guinea pigs and large mammals, including humans, display excitation-contraction coupling properties that are distinct from rats and mice, including long action potential plateaus, positive force frequency relationships, and a greater contribution of NCX to Ca²⁺ removal[78]. The high resting SR Ca²⁺ load in the mouse might render it more susceptible to MCU-dependent Ca²⁺ overload and it is possible that the SR is less Ca²⁺ depleted (or possibly increased) in the DOCA HF model than in other models. As compared to acute pressure overload HF animal models, the DOCA model induces a chronic volume overload with hypertension without downregulation of SERCA2 expression[79] but with prolongation of the action potential as a consequence of transient outward K⁺ channel downregulation[80], which, in the mouse and rat, results in a prolongation of late repolarization in a negative voltage (−20 to −50 mV) range favoring early afterdepolarizations.

The outcome in a given animal HF model may hinge on some of the key questions we have raised here: Is SR Ca²⁺ load unchanged, increased, or decreased in the model? Are ultrastructural changes present that could compromise SR to mitochondrial Ca²⁺ coupling? Is cytosolic Na⁺ overload present? Is mitochondrial Ca²⁺ influx or efflux altered at the level of the isolated mitochondria? Ultimately, it will be important to determine if mitochondrial Ca²⁺ loading is depressed or excessive in myocytes from human failing hearts to guide future therapeutic interventions targeting mitochondrial Ca²⁺. Few studies have specifically addressed this point; however, a report showed decreased activity of two types of mitochondrial Ca²⁺ channels in mitoplasts from human failing hearts[81]. A significant hurdle is the lack of permeable and selective MCU modulators. The classic polycationic inhibitor ruthenium red has poor membrane permeability and promiscuously blocks a number of ion channels including the L-type Ca²⁺channel[82] and the SR Ca²⁺ release channel[83]. The colorless active component, Ru360, has been reported to be a selective inhibitor of mitochondrial Ca²⁺ uptake[84]. However, in our experience using intracellular Ru360 delivered by patch clamp, some inhibition of the cytosolic Ca²⁺ transient occurs even at 100nM Ru360[24]. This makes it difficult to draw clear conclusions from studies applying Ru360 to cells at μM concentrations, where it might be inhibiting multiple ion channels. Similarly, the specificity of mitochondrial Ca²⁺ “enhancers” is not always fully characterized in the conditions of a particular study. Hence, current pharmacological tools must be used judiciously, recognizing their limitations. In the same vein, genetically engineered mice also require consideration of adaptations that might occur in response to complete gene ablation or overexpression of mitochondrial proteins that are usually kept in exquisite balance.

Summary

Cardiac function vitally depends on precise spatiotemporal coordination of ion fluxes, energy delivery and redox reactions that can be modulated over a wide range of workloads. The free energy of ATP hydrolysis must be maintained at all costs, which depends on the controlled oxidation of metabolic substrates through coupled redox reactions — it is less clear why key proteins involved in excitation-contraction coupling operate in a regime that is so close to the threshold for maladaptive oxidative posttranslational modifications. Mitochondrial Ca²⁺ signaling appears to be essential for keeping excitation-contraction-redox coupling in balance but changes occurring during cardiac decompensation disrupt this process, and uncontrolled vicous circles involving crosstalk between the sarcolemma, sarcoplasmic reticulum and mitochondria can lead to a cascade of failures spanning from the organelle to the organ. Available evidence suggests that this loop can be broken at the level of either mitochondrial Ca²⁺ or mitochondrial ROS but there is some debate about whether the former should be increased or decreased to improve outcomes in heart failure. Differences may be related to specifics of the animal model of cardiac dysfunction and the clinical relevance is currently unknown, requiring additional studies of human myocardium.

Highlights.

• Mitochondrial Ca²⁺ plays an important physiological role but may induce cell death

• It is unclear whether mitochondrial Ca²⁺ is increased or decreased in chronic HF

• We discuss whether increasing or decreasing mitochondrial Ca²⁺ is beneficial in HF