Calcium (Ca²⁺) signaling plays a central role in the contractile, metabolic, and epigenetic functions of cardiac myocytes. Transient variations in cytosolic Ca²⁺ concentration ([Ca²⁺]_c) trigger cardiac myocyte contraction and relaxation. During physiological elevations of cardiac workload, the accumulation of Ca²⁺ in the mitochondrial matrix adjusts the rate of mitochondrial oxidative metabolism to the energy requirements of the cell (McCormack and Denton, 1993). In cardiac disease, Ca²⁺ mishandling can wreak havoc on cellular functions: in heart failure (HF), defective mitochondrial Ca²⁺ accumulation can cause bioenergetic mismatch and oxidative stress (Kohlhaas et al., 2017); in ischemia/reperfusion (I/R) injury, mitochondrial Ca²⁺ overload leads to a sudden increase in mitochondrial membrane permeability that is one major driver of cell death (Xuxia et al., 2020). In addition to cardiac myocyte contraction, changes in [Ca²⁺]_c modulate the activation of gene expression, a process coined excitation-transcription coupling. Here, we review the role of Ca²⁺ signaling as a regulator of mitochondrial oxidative metabolism and reactive oxygen species (ROS) emission from mitochondria, and how derangements in these processes contribute to the development of HF.

In cardiac myocytes, Ca²⁺ handling is mediated by voltage- and ligand-activated ion channels that are mainly found on the sarcolemma and on the sarcoplasmic reticulum (SR), a specialized endoplasmic reticulum that cyclically releases and removes Ca²⁺ ions from the cytosol during cardiac myocyte contraction and relaxation. During phase 2 of the action potential, opening of voltage-dependent L-type Ca²⁺ channels on the sarcolemma leads to Ca²⁺ entry into the cytosol, which in turn triggers the release of a larger amount of Ca²⁺ from the SR via type 2 ryanodine receptors (RyR2). This phenomenon, termed Ca ²⁺ -induced Ca ²⁺ release, initiates contraction as Ca²⁺ binds the regulatory protein troponin C, thereby inducing a conformational change that enables the interaction of the myosin head with actin (Bers, 2002). During diastole, Ca²⁺ is removed from the cytosol by the SR Ca²⁺ ATPase (SERCA), which takes Ca²⁺ back into the SR, and to a lesser extent by the sarcolemmal sodium (Na⁺)/Ca²⁺ exchanger (NCX), which extrudes Ca²⁺ to the extracellular space; for each Ca²⁺ ion removed, the NCX imports 3 Na⁺ ions in the cytosol.

Ca²⁺ operates as a second messenger to adjust the rate of mitochondrial oxidative metabolism according to the energy requirements of cardiac myocytes (McCormack and Denton, 1993; Brandes and Bers, 1997). Increases in heart rate and contractility as seen during physical exercise increase the rate of adenosine triphosphate (ATP) turnover by cardiac myocytes up to 5- or 6-fold the baseline rate. The increase in adenosine diphosphate delivery to mitochondria via the creatine kinase system accelerates the rate of ATP production by oxidative phosphorylation, which needs to be sustained by an increased supply of reducing equivalents that are donated to complex I and complex II of the electron transport chain (ETC) by the reduced form of nicotinamide adenine dinucleotide (NADH) and succinate, respectively, derived from the tricarboxylic acid (TCA) cycle. Ca²⁺ is pivotal to this mechano-energetic coupling by stimulating the TCA cycle dehydrogenases (Figure 1).

Furthermore, elevations in [Ca²⁺]_c promote the shuttling of reducing equivalents from cytosolic NADH to the mitochondrial NADH pool by stimulating the activity of the two aspartate/glutamate carriers (AGC1/2, also known as SLC25A12 and SLC25A13 or aralar1 and citrin, respectively) that sense Ca²⁺ levels on the outer face of the IMM via their EF-hand motifs (Palmieri et al., 2001). The efflux of aspartate from the matrix in exchange for glutamate catalyzed by AGC1/2 is one of the steps of the malate-aspartate shuttle, a shuttle system whose net effect is that NADH in the cytosol is oxidized to NAD⁺, and NAD⁺ in the matrix is reduced to NADH. AGC1 is the predominant isoform found in cardiac mitochondria, and it responds to Ca²⁺ concentrations lower than those driving Ca²⁺ accumulation in the mitochondrial matrix. On these grounds, it has been proposed that the Ca²⁺-dependent activation of AGC1 might regenerate NADH to sustain oxidative phosphorylation before matrix Ca²⁺ levels rise (Contreras et al., 2007).

The reducing equivalents derived from oxidation of substrates via the TCA cycle are also used to sustain mitochondrial antioxidant systems. In fact, the reduced form of nicotinamide adenine dinucleotide phosphate (NADPH) serves as an electron donor for the redox reactions that convert hydrogen peroxide (H₂O₂) into water. Mitochondrial oxidative metabolism normally produces a certain quantity of ROS, which also serve physiological functions as signaling molecules, but can damage cellular components when produced in exaggerate amounts or not adequately scavenged. The best-characterized sites of ROS formation in mitochondria are complexes I and III of the ETC, where the superoxide radical (^•O₂ ⁻) can be generated by the incomplete reduction of O₂ (Murphy, 2009). Superoxide is rapidly converted to H₂O₂ by the manganese-dependent superoxide dismutase (Mn-SOD). Another source of mitochondrial ROS is monoamine oxidases, a family of flavoprotein residing in the outer mitochondrial membrane that degrades endogenous monoamines using flavin adenine dinucleotide (FAD) as a cofactor and generating H₂O₂ as a product (Kaludercic et al., 2011). In turn, H₂O₂ is reduced to water by glutathione peroxidase or peroxiredoxin, which then need to be regenerated in their active (reduced) form by a cascade of redox reactions that are fueled by NADPH. Because the majority of mitochondrial NADPH is produced by reactions that use TCA cycle intermediates as substrates, the Ca²⁺-dependent stimulation of the TCA cycle is pivotal to maintain an adequate supply of reducing equivalents to both the ETC and H₂O₂-eliminating systems during elevations of cardiac workload (Figure 1).

Efficient Ca²⁺ uptake by mitochondria is made possible by the spatial proximity between mitochondria and the SR. Contact points between mitochondria and the SR create spatially localized microdomains where micromolar Ca²⁺ concentrations are transiently attained after Ca²⁺ is released from the SR. These local elevations of Ca²⁺ concentration are essential for efficient mitochondrial Ca²⁺ uptake (Giacomello et al., 2010). The regions of the SR that are tethered to mitochondria are also known as mitochondria-associated membranes (MAM), and the integrity of these contact sites is maintained by specialized proteins such as mitofusin 2 (Mfn2), a dynamin-related GTPase that is expressed on both the outer mitochondrial and SR membranes. Loss of Mfn2-mediated interorganelle tethering in the heart reduces the extent of mitochondria-SR contact sites, thereby impairing mitochondrial Ca²⁺ uptake and the Ca²⁺-dependent metabolic adaptation in response to β-adrenergic stimulation (Chen et al., 2012). Viceversa, chronic enhancement of cardiac mitochondria-SR tethering via expression of a cardiac myocyte-specific tether transgene potentiates mitochondrial-SR Ca²⁺ crosstalk, decreasing cardiac vulnerability to adrenergic stress and I/R injury (Nichtová et al., 2023).

Other proteins, such as PACS-2 (phosphofurin acidic cluster sorting protein 2), sigma-1 receptor, GRP75 (glucose-regulated protein 75), and FUNDC1 (FUN14 domain containing 1) were also identified as potential mitochondria-SR tethering proteins and shown to play a role in the communication between SR and mitochondria. Genetic deletion of these proteins results in cardiac dysfunction, ER stress, and mitochondrial fragmentation (Szabadkai et al., 2006; Hayashi and Su, 2007; Stoica et al., 2014; Wu et al., 2017). The ability of the SR to store large amounts of Ca²⁺ ions depends on the presence of Ca²⁺-binding proteins such as calsequestrin (CASQ) in its lumen. CASQ acts as a Ca²⁺ buffer that maintains the SR luminal free Ca²⁺ concentration between 100–500 μM. The expression of CASQ can be enhanced or suppressed at a gene level, resulting in an increase or decrease in SR Ca²⁺ load. In addition to acting as a major Ca²⁺ storage protein, CASQ also regulates the activity of the RyR2 channels (Fearnley et al., 2011).

Overall, there is a tight interplay between the ATP-consuming processes of excitation-contraction coupling and the ATP-producing processes of mitochondrial oxidative metabolism. Ca²⁺ plays a central role in matching cardiac energy supply to demand and sustaining mitochondrial antioxidative capacity during transitions of cardiac workload.

Mitochondrial Ca²⁺ uptake is driven by the large negative electrochemical potential (∆μ _H −180 mV) across the IMM that is maintained by the proton-pumping activity of the, ETC., complexes I, III, and IV (Vasington and Murphy, 1962). If the MCU flux was not modulated by its regulatory subunits, the negative IMM potential would drive large amounts of Ca²⁺ inside mitochondria, overwhelming the buffering capacity of inorganic phosphate contained in the mitochondrial matrix. Mitochondrial Ca²⁺ overload and uncontrolled ROS production lead to the opening of a large pore in the IMM denoted as the mitochondrial permeability transition pore (mPTP), which dissipates the proton-motive force and allows equilibration of solutes <1.5 kD, including proapoptotic factors such as cytochrome c (Rasola and Bernardi, 2011). The mPTP plays an important role in regulating cell death and mitophagy under physiological conditions (Bonora et al., 2022), and transient mPTP opening may be necessary to fine-tune mitochondrial Ca²⁺ levels, functioning like a pressure-release valve to prevent Ca²⁺ overload (Lu et al., 2016). Furthermore, mPTP opening is also a central mediator of cardiac myocyte death during ischemia-reperfusion injury (Morciano et al., 2017). Accordingly, MCU-knockout mice exhibit decreased infarct size and improved function after myocardial infarction, as defective Ca²⁺ uptake reduces cardiac myocyte death triggered by mitochondrial permeability transition (Luongo et al., 2015).

A major limitation of inhibiting the mPTP is that the molecular identity of the proteins that compose the pore is largely unknown. Multiple models were developed to explain the molecular architecture of the mPTP, and several proteins including the voltage-dependent anion channel (VDAC), the adenine nucleotide translocator (ANT), and the F₁-F_o ATP synthase (complex V) were proposed to be mPTP component (for a recent review, see (Bauer and Murphy, 2020)). Cyclosporine A inhibits permeability transition by targeting cyclophilin D, which is an activator of mPTP opening. Genetic deletion of Ppif worsens cardiac hypertrophy, fibrosis, and myocardial function in response to pressure overload. In Ppif-deficient mice, impaired mPTP opening causes an alteration in Ca²⁺ efflux, increasing mitochondrial Ca²⁺ and consequently activating Ca²⁺-dependent TCA cycle dehydrogenases. As a result, glucose oxidation is favored relative to fatty acids, reducing the metabolic flexibility of the heart during stress (Elrod et al., 2010). These findings suggest that the mPTP contributes to metabolic adaptation during changes in cardiac workload.

In the failing heart, altered Ca²⁺ handling is a major contributor to contractile dysfunction and arrhythmias (Luo and Anderson, 2013). A primary defect in HF is the reduced Ca²⁺ load of the SR (Lindner et al., 1998), which results from reduced Ca²⁺ reuptake via SERCA and Ca²⁺ release due to spontaneous RyR2 opening. Consequently, the amplitude and decay velocity of cytosolic Ca²⁺ transients are reduced, and [Ca²⁺]_c during diastole is increased, thereby lowering systolic force development and increasing wall tension during diastole (Hasenfuss et al., 1996; O’Rourke et al., 2021). Another central alteration is the increase in [Na⁺]_i, which results from increased late Na⁺ current (I _Na) (Valdivia et al., 2005), increased Na⁺/proton exchanger (NHE) activity (Baartscheer et al., 2003), and decreased Na⁺/K⁺ ATPase activity (Schwinger et al., 1999).

Alterations in cytosolic Na⁺ and Ca²⁺ handling and the resulting changes in mitochondrial Ca²⁺ signals have profound consequences on mitochondrial oxidative metabolism and antioxidant defense. In HF, reduced SR Ca²⁺ load and disruption of the tubular SR network (He et al., 2001) decrease Ca²⁺ concentration in the mitochondria-SR contact sites, thereby impairing mitochondrial Ca²⁺ uptake. In addition, mitoplasts (i.e., mitochondria stripped of the outer membrane) isolated from patients with HF exhibit decreased MCU activity compared with nonfailing controls (Michels et al., 2009), which might be explained by an increased expression of the regulatory subunits MICU1 and MICU2 (Paillard et al., 2022). Furthermore, elevated [Na⁺]_i accelerates Ca²⁺ efflux from the mitochondrial matrix via the NCLX (Maack et al., 2006). Although elevated [Na⁺]_i also favors Ca²⁺ influx in the cytosol via the reverse mode of the NCX, this is far less efficient in inducing mitochondrial Ca²⁺ uptake compared with SR Ca²⁺ release due to the slower NCX kinetics and the more remote location of the NCX flux compared with the RyR2 flux (Kohlhaas and Maack, 2010). Altogether, derangements in cellular Ca²⁺ and Na⁺ handling contribute to hamper mitochondrial Ca²⁺ signals, thereby hindering the Ca²⁺-mediated stimulation of oxidative metabolism that is required to adapt the production of reducing equivalents for the ETC to the ATP demand of cardiac myocytes (Figure 2).

At the same time, defective stimulation of the TCA cycle by mitochondrial Ca²⁺ leads to an inadequate supply of reducing equivalents required to maintain matrix NAD(P)H redox potential. In a guinea pig model of HF induced by combined chronic β-adrenergic stimulation and pressure overload, pharmacological inhibition of the NCLX could counteract the depletion of mitochondrial Ca²⁺ induced by elevated [Na⁺]_i, thereby restoring NAD(P)H redox state and reducing ROS emission from mitochondria (Kohlhaas et al., 2010). This intervention ameliorated cardiac remodeling, improved cardiac function and decreased the risk of ventricular arrhythmias (Liu et al., 2014). In an analogous guinea pig model of HF, moderate overexpression of MCU by viral gene transfer reduced oxidative stress induced by increased cardiac workload, improved systolic function and decreased the number of premature ventricular complexes (Liu et al., 2021).

The mechanisms linking restoration of mitochondrial Ca²⁺ levels by NCLX inhibition or MCU overexpression and protection from ventricular arrhythmias and sudden cardiac death in this model are not completely resolved. One hypothesis is that the antiarrhythmic effect of correcting mitochondrial Ca²⁺ levels results from decreased ROS-mediated oxidation of RyR2, which decreases spontaneous SR Ca²⁺ release events that cause delayed afterdepolarizations, a known trigger of arrhythmias (Zima and Blatter, 2006). In mice, SR Ca²⁺ leaks feed a vicious cycle whereby mitochondrial Ca²⁺ mishandling leads to mitochondrial dysfunction, ROS emission, and RyR2 oxidation, further exacerbating proarrhythmic SR Ca²⁺ leaks (Santulli et al., 2015; Hamilton et al., 2020). One additional mechanism might be represented by ROS-induced activation of Ca²⁺/calmodulin-dependent protein kinase II (CaMKII), which prolongs action potential duration by phosphorylating Na⁺ channels (Erickson et al., 2008; Wagner et al., 2011). Therefore, oxidation of mitochondrial pyridine nucleotides and mitochondrial ROS production create both a trigger and a substrate for ventricular arrhythmias, but the underlying mechanisms have not been fully elucidated.

It is important to note that this mechanistic framework is supported by studies in guinea pigs exposed to combined chronic β-adrenergic stimulation and pressure overload, which might explain the apparently conflicting results obtained in smaller rodents subjected to myocardial infarction. In the latter model, it has been reported that mitochondrial Ca²⁺ levels are increased, rather than reduced (Santulli et al., 2015), and abrogating Ca²⁺ efflux by inhibiting the mPTP or the NCLX has beneficial, rather than detrimental effects (Elrod et al., 2010; Boyman et al., 2021). Importantly, guinea pigs more closely recapitulate human excitation-contraction coupling, whereas mice exhibit substantial differences including the higher [Na⁺]_i (9–14 mmol/L vs. 4–8 mmol/L) and the smaller contractile reserve compared to humans that might partly explain the seemingly contradictory conclusions of these studies.

Pathological elevations of cardiac workload such as those induced by pressure overload also deplete mitochondrial antioxidative capacity and cause oxidative stress by reversing the reaction catalyzed by the mitochondrial nicotinamide nucleotide transhydrogenase (NNT) (Nickel et al., 2015). Under physiological conditions, the NNT transfers electrons from NADH to regenerate NADPH, thus bolstering mitochondrial antioxidative capacity (Figure 2). Pressure overload reverses the NNT reaction, thereby consuming NADPH to regenerate NADH required to sustain the ECT activity. The ensuing ROS production is a major driver of maladaptive remodeling after transverse aortic constriction in mice (Nickel et al., 2015). The yin/yang role of the NNT in the modulation of mitochondrial ROS emerged from the observation that C57BL/6J mice, an inbred mouse strain carrying a loss-of-function mutation of the Nnt gene, are protected from oxidative stress and heart failure induced by pressure overload (Nickel et al., 2015). Therefore, the use of this mouse strain in cardiovascular and/or metabolic research should be avoided.

Altogether, these studies indicate that alterations in cardiac myocyte Na⁺ and Ca²⁺ handling as seen in HF hinder cardiac metabolic adaptation and exacerbate mitochondrial ROS production by preventing Ca²⁺ accumulation in the mitochondrial matrix. In rodent models of HF, counteracting these changes with pharmacologic interventions or genetic manipulation had beneficial effects on cardiac remodeling and arrhythmias.

Despite the large body of evidence indicating that interventions aimed at ameliorating mitochondrial Ca²⁺ handling in cardiac myocytes have beneficial effects in cardiac disease, studies supporting this concept in humans remain limited. In the context of I/R injury, mitochondrial Ca²⁺ overload triggers mPTP opening, one central driver of cardiac myocyte loss. Accordingly, studies in rodent models of I/R injury demonstrated that genetic inhibition of the MCU reduced cardiac injury by inhibiting mPTP opening (Kwong et al., 2015; Wu et al., 2020). However, mPTP inhibition by cyclosporine infusion immediately prior to reperfusion did not improve cardiovascular outcomes nor attenuated maladaptive remodeling in patients with myocardial infarction in a phase 3 randomized controlled trial (Cung et al., 2015). One key translational challenge in this context is that mitochondrial Ca²⁺ overload ensues during ischemia, and mPTP opens as soon as the intracellular pH is restored upon reperfusion (Hartmann and Decking, 1999). Therefore, these detrimental processes are already at play when the patient arrives to the catheterization laboratory for percutaneous reperfusion. Viceversa, the therapeutic goal in chronic HF is to restore physiological levels of mitochondrial Ca²⁺, which can be achieved with two different approaches: boosting mitochondrial Ca²⁺ uptake with pharmacological agents or preventing mitochondrial Ca²⁺ efflux either by inhibiting the NCLX or by decreasing [Na⁺]_i. Pharmacological agents that were tested in human HF an could potentially improve mitochondrial Ca²⁺ handling are discussed below.

Alterations in Na⁺ and Ca²⁺ handling contribute to contractile dysfunction and arrhythmias in HF. In spite of our detailed mechanistic understanding of these derangements derived from animal models of HF, therapeutic strategies aimed at ameliorating cardiac myocyte ion handling and restoring mitochondrial Ca²⁺ levels have not provided clinical benefit in HF patients thus far. Further studies are needed to understand whether the cardioprotective effects of SGLT2i are mediated, at least in part, by their beneficial activity on mitochondrial Ca²⁺ and cardiac energetics.