The debate continues – What is the role of MCU and mitochondrial calcium uptake in the heart?

Abstract

Since the initial identification of the mitochondrial calcium uniporter (MCU) in 2011, several studies employing genetic models have attempted to decipher the role of mitochondrial calcium uptake in cardiac physiology. Confounding results in various mutant mouse models have led to an ongoing debate regarding the function of MCU in the heart. In this review, we evaluate and discuss the totality of evidence for mitochondrial calcium uptake in the cardiac stress response. We highlight recent reports that implicate MCU in the control of homeostatic cardiac metabolism and function. This review concludes with a discussion of current gaps in knowledge and remaining experiments to define how MCU contributes to contractile function, cell death, metabolic regulation, and heart failure progression.

1. Introduction:

The rapid uptake of calcium (Ca²⁺) into cardiac mitochondria was first reported nearly 70 years ago in early studies of oxidative phosphorylation (1). This initial observation fueled research for the next several decades into the biophysical mechanisms driving mitochondrial Ca²⁺ (_mCa²⁺) uptake, its biological consequences, and the molecular identity of the proteins necessary for the influx of Ca²⁺ into the mitochondrial matrix.

Biophysics of _mCa²⁺ uptake:

Rapid Ca²⁺ uptake into mitochondria is driven by the highly negative electrochemical potential (ΔΨ, ~−180mV) across the inner mitochondrial membrane (IMM) (2, 3), which is established by the electron transport chain (ETC) pumping protons out of the mitochondrial matrix and into the intermembrane space (IMS) (4, 5). Thus, _mCa²⁺ uptake indirectly depends upon the oxidation of carbon substrates within the TCA cycle (2), since the TCA cycle feeds electrons into the ETC. In intact cells, the biological signal that drives an increase in _mCa²⁺ uptake is an increase in local cytosolic Ca²⁺ content. Tullio Pozzan’s laboratory was the first to provide direct evidence of this phenomenon in an elegant study combining the use of the Ca²⁺sensitive dye, Fura-2, and a mitochondrial-localized Ca²⁺-sensitive photoprotein, aequorin, to simultaneously measure fluctuations in cytosolic Ca²⁺ and _mCa²⁺ in intact cells (6). This study and subsequent work showed that treatment of cells with ATP to stimulate purinergic receptors and cause an elevation in cytosolic Ca²⁺ concentration (7) elicits a coordinated, rapid increase in _mCa²⁺ concentration. Additional experiments demonstrated that inositol triphosphate (IP₃)-dependent release of Ca²⁺ from the endoplasmic reticulum (ER) generates a local elevation in Ca²⁺ concentration as much as 20-fold higher than in the bulk cytoplasm, and that microdomains of elevated Ca²⁺ levels near points of ER-mitochondrial contact are particularly effective at triggering _mCa²⁺ uptake (8–12). These results collectively hinted at the existence of a mechanism of calcium-induced _mCa²⁺ uptake, analogous to the process of calcium-induced Ca²⁺ release from the sarcoplasmic reticulum that is required for cardiomyocyte contraction.

“Classic” consequences of _mCa²⁺ uptake:

What are the consequences of acute _mCa²⁺ uptake? Initially, the matrix concentration of free, ionic Ca²⁺ increases over the range of ~0.1 – 5μM, above which any additional change in free-Ca²⁺ concentration is largely buffered by inorganic phosphate (13–15). This Ca²⁺ buffering capacity is profound, with the ratio of phosphate-bound : free calcium estimated as high as 100,000:1 (13), allowing mitochondria to act as a sink for cellular Ca²⁺ in many non-excitable cell types. Consistent with this function in buffering Ca²⁺, inorganic phosphate in the matrix has been shown to accelerate net _mCa²⁺ uptake while at the same time limiting the extent to which the matrix concentration of free, ionic Ca²⁺ rises (16). The existence of inorganic phosphate as polyphosphate polymers within the matrix is thought to minimize the precipitation of complexed Ca²⁺-phosphate under normal conditions (17).

An acute physiological increase in _mCa²⁺ concentration stimulates mitochondrial metabolism, which is especially important for the heart, a tissue with a high, dynamically changing ATP-demand. Early investigations established that the accumulation of _mCa²⁺ precedes increases in oxidative phosphorylation and ATP production (18). Subsequent work identified specific regulatory nodes at which Ca²⁺ stimulates mitochondrial metabolism. Ca²⁺ directly activates pyruvate dehydrogenase (PDH) phosphatase (19), which removes inhibitory phosphate groups from S232, S293, and S300 of the E1α regulatory subunit of PDH to increase its enzymatic activity (20, 21). This mechanism increases the oxidation of pyruvate derived from glycolysis to generate acetyl-CoA for entry into the TCA cycle, and in doing so generates NADH, a reducing equivalent for the ETC (reviewed in (add references to other reviews within this special issue) (22)). An increase in matrix Ca²⁺ also stimulates the TCA cycle enzymes isocitrate dehydrogenase and α-ketoglutarate dehydrogenase (23–26) by lowering the K_m of these enzymes for their respective substrates. This effect further promotes TCA cycle flux and the generation of additional reducing equivalents for the ETC. Finally, recent evidence indicates that increased matrix Ca²⁺ concentration can also promote ATP production via direct effects to stimulate Complexes I, III, and IV of the ETC (27). The activity of ATP synthase (Complex V) has also been reported to be stimulated by increased _mCa²⁺, leading to increased mitochondrial ATP production (28), although this notion is disputed by recent work from Lederer and colleagues (29) (reviewed in (30)). In contrast to “physiological” increases in _mCa²⁺ levels, _mCa²⁺ overload has been associated with impaired mitochondrial respiration. Recent computational and experimental studies indicate that impairments in mitochondrial ATP production may result from the accumulation of Ca²⁺-phosphate precipitates within the matrix. These precipitates appear to disrupt NADH-dependent respiration, possibly by impairing Complex I activity, interfering with the transfer of electrons from NADH to Complex I, or disrupting overall cristae integrity (31). This may be one mechanism by which excessive MCU-dependent Ca²⁺ uptake, in pathological conditions such as I/R (discussed below), contributes to mitochondrial dysfunction and reactive oxygen species (ROS) generation (32).

The second classical consequence of _mCa²⁺ overload is activation of the mitochondrial permeability transition pore (mPTP). mPTP opening in turn leads to mitochondrial swelling, outer mitochondrial membrane (OMM) rupture, and necrotic cell death (reviewed extensively in (33–35)). Such cell death is thought to contribute to much of the tissue damage resulting from acute _mCa²⁺ overload in myocardial infarction, ischemia/reperfusion injury (33, 34), and from chronic increases in cytosolic Ca²⁺ (_cCa²⁺). For example, the increase in cardiac workload during prolonged stress, such as in patients with chronic hypertension or aortic stenosis (36), drives elevations in cellular Ca²⁺ cycling that lead to the accumulation of _mCa²⁺ and activation of the mPTP. Additional studies suggest that apoptotic cell death can also contribute to loss of cardiomyocytes in hypertension- or pressure overload-induced heart failure (37, 38). Indeed, _mCa²⁺-dependent activation of the mPTP has been implicated in apoptosis, as well as necrosis (39). This observation supports the idea that increased _mCa²⁺ concentration is a central trigger for multiple forms cardiomyocyte death, and thus places the control of _mCa²⁺ homeostasis and prevention of _mCa²⁺ overload as a key determinant of cell survival and overall cardiac function.

2. MCU and the mtCU complex:

The resting _cCa²⁺ concentration of cardiomyocytes is around 100 nM, a concentration too low to activate acute mitochondrial Ca²⁺ uptake (40). When cytosolic calcium concentration rises past ~400nM, as occurs upon calcium-induced Ca²⁺ release from the sarcoplasmic reticulum following an action potential, the threshold for activation of acute _mCa²⁺ uptake is reached and Ca²⁺ rapidly flows into the mitochondrial matrix (41–43). This mode of rapid _mCa²⁺ uptake has long been described as the mitochondrial calcium uniporter current and is powered by the mitochondrial membrane potential, is inwardly rectifying, and is not directly coupled to the transport of another ion (44–47). The uniporter current is transmitted by a multiprotein complex of the IMM that is now termed the mitochondrial calcium uniporter channel complex (mtCU).

The mtCU is a large heteromeric protein complex of ~450–800kDa (48, 49). The core of the channel is formed by on average 4 subunits of the mature form of the mitochondrial calcium uniporter protein (MCU) (~31kDa in human and ~34kDa in mouse) (50–54), which was identified as the calcium channel responsible for uniporter activity in 2011 (48, 55, 56). The integral IMM protein EMRE associates with the four MCU subunits, likely in a 1:1 ratio, and is required for function of metazoan mtCUs (22, 57, 58), potentially by helping to hold the MCU subunits in an open configuration that is permissive for Ca²⁺ permeation. EMRE also associates with the regulatory protein MICU1 and helps to anchor MICU1 to the mtCU complex (49, 59). MICU1 and MICU2 heterodimerize and project into the intermembrane space (IMS), where their Ca²⁺-binding EF-hand domains allow sensing of local changes in Ca²⁺ concentration (60–63). This Ca²⁺-dependent regulation of MICU1/2 function allows them to act as mtCU gatekeepers, keeping the channel closed to prevent _mCa²⁺ uptake at low cytosolic/IMS Ca²⁺ concentrations, but promoting the cooperative activation of the channel as cytosolic and IMS Ca²⁺ concentration rises (59, 64–66). Several additional regulatory subunits including the MCU paralog MCUB and the integral IMM protein MCUR1 can also associate with the core uniporter components and modulate overall mtCU activity (67–69). The reader is referred to several excellent reviews (70, 71) for a more comprehensive discussion of mtCU structure, function, and regulation.

Function of MCU in _mCa²⁺ uptake:

Early functional studies in vitro demonstrated that knockdown of MCU in HeLa cells or mouse liver mitochondria inhibited rapid _mCa²⁺ uptake in response to a histamine-induced elevation of cytosolic Ca²⁺ concentration or a direct increase in bath Ca²⁺ (48, 55). These findings support the idea that MCU is required for acute influx of Ca²⁺ into the mitochondrial matrix. Furthermore, recombinant MCU inserted into lipid bilayers is sufficient to conduct Ca²⁺ and exhibits characteristics of the classic mitochondrial calcium uniporter current, such as inhibition by Ru360 (55). Finally, knockdown of MCU results in loss of uniporter current in HEK-293T mitoplasts (72). Together, these foundational studies confirmed MCU’s identity as the central, Ca²⁺-transporting channel within the larger mtCU complex. MCU has since been proposed to contribute to increases in free-_mCa²⁺ in response to small, physiologic elevations (~0.1–2μM) of cytosolic Ca²⁺, and to increases in the buffered pool of Ca²⁺-phosphate complexes in response to larger elevations (2–15μM) of extra-mitochondrial Ca²⁺ (14).

MCU and its homologs appear in all eukaryotic lineages, and putative MCU homologs have been detected in some species of bacteria as well (73). The high degree of MCU conservation across taxa points to a fundamental physiological role for MCU. In support of this idea is the fact that no causative mutations in MCU itself have yet been specifically linked to human disease, perhaps reflecting the possibility that complete loss of MCU function is lethal during human development. Numerous human MCU variants have been documented, but their significance for overall MCU function remains unknown. Thus, several pertinent questions remain to be answered: 1) What is the specific role of MCU and MCU-dependent _mCa²⁺ uptake in humans? and 2) What are the biological consequences for altered MCU function, in both health and disease?

3. Role of MCU in the heart – Is MCU dispensable in cardiac physiology or pathological stress?

Over the last decade, a number of laboratories have attempted to resolve this question using MCU loss-of-function mouse models. Experiments in cells derived from these animal models agree with earlier investigations in stable cell lines: namely, that MCU is strictly required for rapid, Ru360-sensitive Ca²⁺ uptake into the mitochondrial matrix (48, 55, 72, 74). However, examination of the overall whole-body and cardiac phenotypes of these models has initiated an active debate over whether MCU is required for basal heart function, or rather is restricted to coordinating the heart’s functional response to acute or chronic physiological or pathological stress. This debate over the role of MCU in the heart is particularly interesting given the potential for Ca²⁺ to act as an integrated signal that elicits both increases in contractile function (via activation of myofilament crossbridge cycling) and increases in ATP production (via stimulation of mitochondrial metabolism) that is required to fuel crossbridge cycling and control cellular ion flux. Below, we evaluate evidence from genetic mouse models of mtCU disruption in which cardiac function has been examined in both basal, unstressed conditions and following physiological or pathological stress. The reader is referred to the primary literature on mouse models of MICU1 (75–77), MICU2 (78), MCUR1 (69), and EMRE (79) perturbation, as well as recent reviews in this Special Issue (add citations), for discussion of the effects of these additional mtCU components on viability and homeostatic heart function.

3.1. Baseline role of MCU in the heart:

Basal phenotype of animal models with constitutive MCU disruption:

The first genetic mouse model with targeted disruption of MCU was described in 2013 by the Finkel and Murphy laboratories (74). This model used gene-trap technology to delete Mcu in C57BL/6 mice. However, embryonic lethality required the gene-trap line be outcrossed to the CD1 background, suggesting that hybrid vigor was required for viability (80). Even on this mixed background, MCU-null mice were still born at a frequency lower than the expected Mendelian ratio (80).The surviving outcrossed MCU-null mice displayed a reduction in overall body weight and size in addition to ablation of rapid _mCa²⁺ uptake (observed in skeletal muscle, cardiac muscle, and embryonic fibroblast mitochondria) (74). Such impaired mtCU function was associated with a ~75% reduction in the matrix Ca²⁺ content of skeletal muscle (74), indicating that mtCU-dependent Ca²⁺ uptake can exert considerable control over homeostatic matrix Ca²⁺ levels. However, some methodological issues with this measurement, such as not inhibiting MCU-dependent Ca²⁺ uptake in wild-type controls during mitochondrial isolation, may explain why this particular finding was not recapitulated in later investigations (81). Despite these changes in _mCa²⁺ handling, overall mitochondrial structure and number were normal in MCU-null mice (74). A subsequent study published by the same research group specifically examined the effects of constitutive Mcu deletion on heart function. Similar to their findings in skeletal muscle, despite a reduction in matrix calcium content, there was no discernable effect of chronic MCU deficiency on basal cardiac function in mice out to almost 2 years of age (82). A recent paper examining the consequences of functional disruption of the mtCU via global EMRE (Smdt1 gene) deletion likewise found no detrimental effect of constitutive EMRE loss on the unstressed mouse heart, despite a reduction in cardiac _mCa²⁺ content (79).

The idea that MCU has little effect on basal heart function is further supported by work from Mark Anderson’s laboratory using transgenic mice with alpha-myosin heavy chain-driven, cardiac-specific expression of a dominant negative form of MCU (DN-MCU). In this construct, critical acidic residues of the channel’s selectivity filter (53) are mutated to glutamine (DIME→QIMQ), effectively ablating MCU activity (83). As with germline Mcu deletion, transgenic DN-MCU mice are viable, and basally show very little effect of diminished mtCU activity on heart structure or function (83). The Anderson lab did note two specific traits in DN-MCU hearts that differed from control hearts under basal conditions: reduced myocardial ATP content, and reduced basal sarcoplasmic reticulum (SR) Ca²⁺ content (83). A reduction in ATP content is a surprising observation because the heart protects its ATP supply at all costs, and contractile function was not impaired in DN-MCU animals. One possible explanation for the discrepancy between these findings is that ATP content was specifically measured in atrial tissue (as were Ca²⁺ handling and SR Ca²⁺ load), whereas functional measurements of contractility were made via echocardiography of the left ventricle. The authors attributed the reduced basal SR content in sinoatrial node cells to impaired ATP-dependent pumping of Ca²⁺ into the SR by the sarcoplasmic/endoplasmic reticulum calcium ATPase (SERCA). This study also reported a reduction in basal heart rate among conscious DN-MCU mice, whereas under anesthesia, no difference in heart rate between DN-MCU and control mice was detected. The implications of this finding for MCU’s role in cardiac stress responses will be discussed further below.

Basal phenotype in MCU conditional knockout mice:

Subsequent studies used Mcu floxed mice crossed to cardiac-specific Cre drivers to investigate the role of cardiomyocyte MCU in the adult heart. This approach allowed for cell-specific MCU knockout and temporal control of MCU ablation. It should be noted that our laboratory also attempted to generate global germline null animals by crossing inbred Mcu^flfl × CMV-Cre mice, but we were unable to obtain viable animals (81). This finding recapitulates the earlier observation that loss of MCU during development is lethal to inbred mice (80). In contrast to constitutive germline knockout, acute deletion of Mcu in the adult heart was well tolerated on the inbred C57BL/6 background (81, 84). Although rapid _mCa²⁺ uptake is ablated in cardiomyocytes from constitutive Mcu^flfl × αMHC-Cre or conditional Mcu^flfl × αMHC-MerCreMer animals after tamoxifen treatment, we and Molkentin lab did not observe any effect of Mcu deletion on baseline cardiomyocyte mitochondrial Ca²⁺ content (81, 84). This finding may reflect compensatory changes that occur on an acute time scale in the adult heart in order to maintain matrix Ca²⁺ content in the absence of MCU, or difficulty in accurately quantifying matrix Ca²⁺ content in isolated mitochondria or cardiomyocytes. For instance, conditional MCU knockout hearts were found to have reduced expression and activity of NCLX, the mitochondrial sodium/calcium exchanger that mediates mitochondrial calcium efflux (84). This adaptation would in theory help to limit Ca²⁺ extrusion from the mitochondrial matrix and thus could help to normalize matrix calcium concentration in MCU-null mitochondria, particularly if alternative slower routes of _mCa²⁺ uptake are sufficient to allow some Ca²⁺ influx into the matrix despite the absence of MCU. The presence of such adaptations to normalize homeostatic matrix Ca²⁺ content may also help to explain why the Molkentin lab detected no alteration in basal cardiac phenotypes of MCU conditional knockout mice up to at least 1 year of age (84), and why our laboratory observed no effects of acute (3-week) Mcu deletion on baseline cardiac function or dimensions (81). Evidence of robust adaptation to normalize homeostatic _mCa²⁺ content is likewise observed in a mouse model of chronic cardiomyocyte NCLX deletion (Slc8b1^flfl × αMHC-Cre), where disruption of _mCa²⁺ efflux leads to a coordinated reduction in _mCa²⁺ uptake, and in mice with cardiomyocyte NCLX overexpression (TRE-NCLX × αMHC-tTA), where normal basal _mCa²⁺ content is maintained despite increased capacity for _mCa²⁺ extrusion (85).

The mouse models of Mcu deletion described above collectively support the idea that although MCU is required for rapid _mCa²⁺ uptake, loss of MCU does not impair basal cardiac function. Thus, loss of MCU was initially thought to be well tolerated in the unstressed heart, due to undefined compensatory adaptations that are sufficient to maintain cardiac ATP supply in the absence of MCU. However, recent findings from our laboratory in a mouse model with inducible, cardiomyocyte-specific overexpression of the MCU paralog, MCUB, challenge this view. Consistent with the proposed role of MCUB as a dominant negative mtCU component that reduces mtCU Ca²⁺ uptake (67), MCUB overexpression attenuated rapid _mCa²⁺ uptake (86). Examination of the cardiac phenotype of these animals 5 days after induction of MCUB expression revealed impaired fractional shortening. This phenotype was associated with increased inhibitory phosphorylation of PDH, reduced maximal respiration, and reduced respiratory reserve capacity (86). These data indicate that disruption of mtCU function is indeed sufficient to impair mitochondrial bioenergetics and consequently impair cardiac contractility. This supports the notion that some degree of cardiomyocyte _mCa²⁺ uptake is continuously occurring and is required for normal function of the homeostatic heart. This same conclusion is supported by previous experiments showing that cardiomyocyte-specific deletion of NCLX and ablation of this major route of _mCa²⁺ efflux is sufficient to cause _mCa²⁺ overload and mitochondrial permeability transition, even in the basal, unstressed condition (85). These results provide genetic evidence that continuously-active _mCa²⁺ uptake is essential for maintaining homeostatic heart function.

Earlier studies in Mcu^flfl × αMHC-MerCreMer mice (81, 84) examined animals at timepoints of 3 weeks or more after initial administration of tamoxifen to induce Mcu deletion. It is likely that the delay between the initial loss of mtCU function and cardiac phenotyping allowed sufficient time for metabolic compensation to normalize ATP production and basal contractile function despite persistent mtCU impairment. Indeed, when MCUB-transgenic mice were examined at ~1 month following transgene induction, they no longer exhibited defects in mitochondrial respiration, altered PDH phosphorylation, or impaired basal contractility, even though acute _mCa²⁺ uptake was still reduced (86). These findings suggest that although preventing acute _mCa²⁺ uptake through the mtCU is sufficient to disrupt basal cardiac energetics and function, the heart has robust ancillary mechanisms, such as _mCa²⁺-independent regulation of PDH phosphorylation state and activity, to normalize cardiac metabolism and contractility. These results suggest that compensatory rewiring of energetic pathways eventually negates their dependence on _mCa²⁺ signaling. Such a model would agree with the notion that the heart maintains its ATP production at all costs, and also would explain the apparent lack of basal phenotype initially reported in inducible cardiomyocyte MCU knockout animals. Furthermore, the existence of adaptive mechanisms to either normalize basal matrix Ca²⁺ content and/or minimize the heart’s dependence on matrix-Ca²⁺ for metabolic regulation may in part explain the early observation that constitutive MCU-null mice on the outbred CD1 background were born at frequencies about half of the expected Mendelian ratio (80). The embryos that were able to adapt most effectively to the absence of MCU survived to birth, whereas those that could not activate compensatory pathways were likely the embryos that did not survive.

If MCU is necessary to maintain metabolism and adequate pump function in the healthy, unstressed heart, how might it further affect the ability of the heart to adapt to situations where increased cardiac workload requires a coordinated increase in energy metabolism? Likewise, how might MCU impact the cardiomyocytes’ ability to withstand the energetic and Ca²⁺ stresses imposed during ischemia-reperfusion injury or chronic stress?

3.2. Role of MCU in the stressed heart:

The various genetic mouse models to perturb mtCU function have been subjected to numerous classical experimental models of physiological and pathological stress to answer these questions. This section will begin with an examination of the evidence for MCU’s role in physiological adaptations to changes in cardiac workload, such as occurs during exercise or the sympathetic fight-or-flight response. We will next discuss the experimental findings suggesting that excessive MCU-dependent _mCa²⁺ uptake is a key stimulus for mPTP opening and acute necrotic cardiomyocyte death in the context of ischemia-reperfusion injury. Finally, we will evaluate the data regarding the involvement of MCU in the heart’s responses to more chronic stresses that cause cardiac hypertrophy and non-ischemic heart failure.

The mtCU in the heart’s response to acute catecholaminergic stimulation:

The major point of agreement emerging from studies of germline MCU-null mice and mice with conditional cardiomyocyte deletion of Mcu is complete loss of rapid _mCa²⁺ accumulation in response to acute β-adrenergic stimulation (74, 84). This defect in rapid _mCa²⁺ uptake is associated with a diminished contractile response of the heart to acute β-adrenergic stimulation. Whereas control hearts respond to acute isoproterenol or dobutamine treatment with an increase in the rates of LV pressure development (dP/dT_max) and relaxation (-dP/dT_min), these responses are blunted in mice with adult cardiomyocyte Mcu deletion (81, 84). Holmstrom et al. claimed that MCU-null mice exhibited no contractile defect in response to acute isoproterenol stimulation (82). However, close examination of their data reveals that although control mice respond to acute isoproterenol treatment with a significant increase in heart rate, cardiac output, and dP/dT_max, isoproterenol did not significantly increase these parameters in MCU-null mice. Furthermore, disruption of mtCU function by acute (~5-day) transgenic expression of MCUB recapitulated the effects of acute Mcu deletion and blocked isoproterenol-induced increased in contractility and relaxation. MCUB-transgenic hearts also exhibited increased inhibitory phosphorylation of PDH following isoproterenol stimulation, consistent with diminished _mCa²⁺ uptake during this treatment (86). These data solidify the importance of Ca²⁺ as a signal that elicits both an increase in cardiac ATP consumption (contraction) and a concordant increase in cardiac ATP generation, and reveal a central role for MCU-dependent _mCa²⁺ uptake in coordinating the increase in mitochondrial metabolism needed to power an increase in cardiac work rate. Interestingly, once given sufficient time (~1 month) to adapt to MCUB overexpression, MCUB-transgenic mice no longer exhibited defective responses to catecholamine stimulation (86). Together, these mouse models of mtCU disruption thus support the overall contention that loss of _mCa²⁺ uptake disrupts the ability of the heart to appropriately increase its work rate in response to physiological stimulation, even though the specific magnitude or severity of this defect may vary in the context of acute versus constitutive mtCU disruption due to the previously discussed compensatory mechanisms that are likely in play.

Wu et al.’s observation that action potential frequency in isolated sinoatrial node cardiomyocytes treated with the MCU inhibitor Ru360 or expressing DN-MCU fails to increase in response to isoproterenol stimulation (83), provides additional evidence that MCU is required for functional cardiac responses to adrenergic signaling. In vivo experiments in intact DN-MCU transgenic mice also revealed a diminished increase in heart rate upon acute isoproterenol infusion in DN-MCU animals (83). Simultaneous tracking of the animals’ physical activity and heart rate via telemetry showed that disruption of _mCa²⁺ uptake also impaired the animals’ ability to increase heart rate in response to an increase in voluntary physical activity (83). This result strengthens the notion that MCU is critical in coordinating normal physiological responses to an increase in cardiac demand, whether due to experimental application of catecholamines or due to the endogenous activation of sympathetic signaling during exercise. Such a role of MCU in matching energy supply to energy demand, even under normal, unstressed or homeostatic workloads, could also explain why DN-MCU mice exhibited a lower heart rate than control animals when conscious and subjected to a wide range of physiological and psychological inputs, despite displaying no difference in heart rate under anesthesia when endogenous sympathetic signaling and cardiac demand are minimized (83). Along with the phenotypes of MCUB-transgenic mice (86), these data indicate that mtCU-dependent _mCa²⁺ uptake is active and necessary within the homeostatic heart, as well as for functional responses to catecholamines. A future goal will be to clarify whether MCU is specifically required for responses to sympathetic agonists, or whether it is also required to support cardiac responses to increases in pre-load or that help to maintain sufficient cardiac output in response to primary increases in afterload. Such findings would strengthen the hypothesis that Ca²⁺ uptake through the mtCU is a fundamental conduit for mitochondrial energy increases.

Why exactly is MCU required for the heart to increase its workload in response to an increase in physiological demand? The studies discussed above agree that the loss of MCU function prevents cardiomyocytes from increasing oxygen consumption and ATP production in response to a direct increase in cytosolic Ca²⁺ concentration or acute isoproterenol treatment (81, 82, 84). This defect is a thought to be a direct consequence of impaired mtCU-dependent entry of Ca²⁺ into the mitochondrial matrix. Impaired stimulus-induced _mCa²⁺ uptake in MCU-deficient cardiomyocytes subsequently impairs _mCa²⁺-dependent dephosphorylation and activation of pyruvate dehydrogenase (81), the key enzyme regulating entry of pyruvate into the TCA cycle. Impaired activation of PDH and perhaps other Ca²⁺-sensitive TCA cycle dehydrogenases, such as αKGDH and IDH (23–26), limits the generation of reducing equivalents (NADH and FADH₂) by the TCA cycle (81, 83). This overall decrease in TCA cycle flux results in diminished delivery of electrons to the electron transport chain. Thus, oxidative phosphorylation is impaired and the heart’s ability to increase ATP production to meet increased energetic demand under conditions of physiological stress is compromised. Direct evidence in support of this model comes from the Anderson laboratory’s observation that direct dialysis with ATP is sufficient to restore isoproterenol-induced increases in the action potential frequency of isolated DN-MCU or Ru360-treated sinoatrial node cells to normal (83). Additionally, treatment with ATP is effective at correcting the slowed decay phase of cytosolic Ca²⁺ transients in Ru360-treated sinoatrial cells under isoproterenol stimulation, presumably through restoration of ATP-dependent SERCA activity to pump Ca²⁺ back into the SR (83). Thus, these authors reasoned that MCU activity is required for isoproterenol-induced increases in sinoatrial node activity and heart rate because it enhances OXPHOS activity and ATP production required to fuel a concomitant acceleration of cellular Ca²⁺ cycling. These findings are consistent with the hypothesis that Ca²⁺ is the common signal that coordinates mitochondrial bioenergetic responses at the same time that cytosolic energy demand increases, acting like a feed-forward signal that serves to maintain cellular energy balance in the face of fluctuating energetic demand.

Together, the studies outlined above support a working model in which MCU is strictly required for the heart to rapidly increase its work output in response to an acute increase in cardiac demand (Figure 1). In brief, stimuli that cause an increase in contractility (i.e., sympathetic stimulation, direct experimental elevation in cytosolic Ca²⁺ transient amplitude, etc.) trigger an increase in local cytosolic Ca²⁺ concentration, e.g. at sites of SR/mitochondrial contact, that is sufficient to activate _mCa²⁺ uptake through MCU. Ca²⁺ that enters the mitochondrial matrix then stimulates TCA cycle dehydrogenases, leading to the accelerated production of reducing equivalents for the electron transport chain. This process is necessary to power an increase in ATP production to support enhanced systolic function, as well as accelerated _iCa²⁺ cycling, such as Ca²⁺ clearance via SERCA to support accelerated cardiac relaxation.

Figure 1– MCU is required for the heart to rapidly increase work output in response to an acute physiological increase in demand.

Homeostatic _mCa²⁺ cycling maintains sufficient ATP production to match basal rates of ATP consumption. Recent data suggest that the mtCU may be required for such basal _mCa²⁺ flux to support cardiac metabolism. In response to exercise or sympathetic stimulation, cytosolic Ca²⁺ concentration rises. In the wild-type heart, this increase in cytosolic Ca²⁺ concentration is rapidly communicated to the mitochondria via acute Ca²⁺ uptake through MCU, the pore-forming subunit of the mitochondrial calcium uniporter complex (mtCU). The subsequent increase in matrix Ca²⁺ concentration stimulates TCA cycle dehydrogenases and ATP synthase in order to increase OXPHOS and the rate of mitochondrial ATP production. This rapid increase in the rate of ATP production is required to fuel increased crossbridge cycling and SERCA activity to allow for an increase in heart rate and contractility. In the absence of MCU, rapid _mCa²⁺ uptake is abolished and the heart cannot quickly increase its mitochondrial metabolism in order to fuel a rapid increase in cardiac work rate.

Several additional observations suggest that this model is not yet complete and that further nuances will be elucidated. How the heart adapts to a loss of MCU function is a question of active debate. Indeed, despite mice with loss of MCU in adult cardiomyocytes having impaired treadmill running performance (84) much like mice with constitutive whole-body Mcu deletion (74), if they are given a sufficiently long exercise warm-up period, they no longer exhibit exercise deficits (84). At the cellular level, MCU-deficient cardiomyocytes are eventually able to load their mitochondria with Ca²⁺ and increase oxygen consumption upon prolonged (>30–40min) catecholamine stimulation (84). These results indicate that MCU is specifically required for rapid adaptation to increased cardiac workloads, and that alternative, slower-acting mechanisms exist that are sufficient to support more gradual adaptation to changes in workload. Additional proteins such as mitochondrial ryanodine receptors have been proposed to mediate MCU-independent _mCa²⁺ uptake (87–90), and thus may contribute to this effect.

In addition to MCU-independent mechanisms that may allow gradual _mCa²⁺ uptake, some models of chronic mtCU disruption support the existence of adaptive signaling and transcriptional mechanisms that remodel diverse cellular pathways, including cardiac energy metabolism to meet the heart’s demand for ATP. For example, the restoration of contractile responses to isoproterenol after ~1 month of MCUB expression was accompanied by restoration of PDH phosphorylation and cardiomyocyte oxygen consumption, despite a persistent reduction of acute _mCa²⁺ uptake (86). These results from Lambert et al. are perhaps best understood in the context of a surprising report from the Anderson laboratory revealing robust metabolic flexibility that, with time, may allow the heart to compensate for mtCU disruption. When examined ex vivo, DN-MCU hearts consumed more oxygen per unit mass than control hearts across a range of electrical pacing frequencies, despite attenuated ionotropic and lusitropic responses to increased pacing frequency (91). This effect of increased oxygen consumption was not conserved under conditions of low cytosolic Ca²⁺ concentration in isolated cardiomyocytes or in isolated cardiac mitochondria, suggesting it was caused by changes outside of the mitochondria and specifically related to cytosolic Ca²⁺ concentration. What’s more, chronic MCU inhibition was associated with elevated systolic and diastolic cytosolic Ca²⁺ levels (91). Rasmussen et al. reasoned that this abnormally high Ca²⁺ concentration placed an extra metabolic stress on cardiomyocytes, for example increased ATP consumption by cellular calcium pumps working to extrude the excess _iCa²⁺. Such increased ATP demand would stimulate ATP production rates and therefore could increase mitochondrial oxygen consumption, as was observed in the ex vivo DN-MCU hearts. The authors further proposed that over time, this extra metabolic stress might trigger additional compensatory changes within the heart. Fitting with this idea, they found broad transcriptional changes in a wide-range of genes including those involved in glucose and mitochondrial metabolism (91).

Together, these studies add a layer of complexity to our current understanding of the role of MCU in responses to acute physiological changes in cardiac workload. A particularly interesting question for future investigations will be to understand why acute functional responses to catecholamine stimulation eventually return in MCUB-transgenic mice, but are not similarly restored in DN-MCU mice (83, 91) or after a month or more post-tamoxifen administration in MCU^flfl × αMHC-MCM mice (81, 84). For now, we can conclude that MCU is undeniably required to mediate short-term cellular energetic adaptations to support rapid increases in cardiac work rate. However, MCU-dependent _mCa²⁺ uptake likely represents just one of many mechanisms by which the heart can tune its metabolism over longer time frames in order to meet more sustained increases in energetic demand.

MCU in cardiac arrhythmia:

In addition to increasing cardiac workload, catecholaminergic stimulation is often used experimentally to alter cardiomyocyte calcium handling to explore the susceptibility to arrhythmia. Given the role of _mCa²⁺ flux in shaping energetic responses to acute sympathetic stimulation, a relevant question is whether MCU-dependent _mCa²⁺ uptake is important in cardiac electrophysiology. This idea has not yet been specifically explored using genetic mouse models of mtCU disruption. However, growing evidence indicates that mitochondria can shape the heart’s susceptibility to arrhythmia under experimental stimulation (reviewed in (92)), and contribute to a functional decline in contractility during sustained arrhythmias.

Computational modeling suggests that mitochondria-derived ROS can alter cytosolic Ca²⁺ handling by stimulating ryanodine receptors and inhibiting SERCA, thus enhancing Ca²⁺-induced Ca²⁺ release and eliciting abnormal electrophysiological behavior (93). As mitochondrial ROS generation is influenced by _mCa²⁺ loading, some have hypothesized a link between _mCa²⁺ handling, mitochondrial ROS production, altered cellular ion flux, and arrhythmogenesis. For example, in a guinea pig heart failure model, diminished _mCa²⁺ content due to excessive _mCa²⁺ efflux through NCLX leads to increased oxidative stress associated with increased arrhythmias and sudden cardiac death (94). Pharmacologic inhibition of NCLX with CGP-37157 attenuates the incidence of premature ventricular beats and sudden cardiac death, supporting a role of _mCa²⁺-related signaling in the development of arrhythmias (94). Mitochondria may also influence arrhythmogenesis by buffering aberrant Ca²⁺ release events from other cellular compartments such as the SR. Studies in zebrafish that exhibit cardiac fibrillation due to a mutation in the plasma membrane Na⁺/Ca²⁺ exchanger 1 demonstrated that increasing _mCa²⁺ uptake via MCU overexpression is sufficient to attenuate arrhythmias and restore coordinated contractions (95). Likewise, pharmacologic activation of MCU using kaempferol can suppress cardiac arrhythmias in a ryanodine receptor 2 mutant mouse model of catecholaminergic polymorphic ventricular tachycardia (96). However, other experiments indicate that impairing MCU activity might diminish arrhythmias via alternative mechanisms. DN-MCU cardiomyocytes have fewer diastolic Ca²⁺ release events at baseline and under acute isoproterenol stimulation, which was attributed to reduced SR Ca²⁺ load secondary to impaired mitochondrial ATP production and SERCA activity (83). Finally, experiments using tachypacing in mouse atrial HL-1 cells and in Drosophila indicate that over time, MCU-dependent _mCa²⁺ uptake during arrhythmia becomes detrimenta,l leading to _mCa²⁺ overload and mitochondrial dysfunction that compromises the heart’s contractile function(97). Thus, much still remains to be determined regarding the acute and chronic consequences of _mCa²⁺ uptake in cardiac arrhythmia.

MCU and cardiomyocyte necrosis in I/R-injury:

The experimental findings discussed above reflect a broad consensus regarding the role of MCU in acutely coupling changes in mitochondrial metabolism and ATP production to changes in cardiomyocyte workload. Both mice with germline Mcu deletion and mice with tamoxifen-inducible adult cardiomyocyte Mcu deletion have been used to investigate the role of MCU in excessive _mCa²⁺ uptake that can trigger mPTP and necrotic cell death during ischemia-reperfusion (IR) injury. Constitutive versus conditional genetic mouse models have yielded disparate results regarding the role of MCU in settings of pathological stress. The initial findings from differing mouse models, and possible explanations to reconcile their discrepancies, are addressed below.

The involvement of MCU in mPTP opening and IR injury was first assessed in 2013 in outbred MCU gene-trap mice with germline Mcu deletion by Pan et al. Here, isolated Mcu^−/− cardiac mitochondria were resistant to mPTP opening and swelling induced by addition of 500μM Ca²⁺ to the bath solution (74). In this respect MCU-null mitochondria phenocopied WT mitochondria treated with the MCU inhibitor Ru360 or the mPTP/Cyclophilin-D inhibitor cyclosporine A (CsA). The authors concluded that MCU-dependent _mCa²⁺ uptake was critical for opening of the mPTP in response to an elevation in cytosolic Ca²⁺ concentration. It is interesting to note that MCU appears to be important specifically for Ca²⁺-induced mitochondrial permeability transition, as treatment with other toxic agents such as H₂O₂ and tunicamycin showed no difference in induction of cell death in wild-type versus MCU-null mouse embryonic fibroblasts (74). Despite yielding clear protection against Ca²⁺-induced mPTP opening, germline deletion of Mcu did not provide any appreciable protection from cardiac IR injury in vivo (74). Loss of MCU also prevented the ability for CsA to protect against ex vivo cardiac IR injury. Therefore, the authors concluded that Mcu^−/− animals likely exhibited compensatory upregulation of alternative cell death mechanisms that are independent of _mCa²⁺ overload and insensitive to CsA. These findings were recently recapitulated in mice with constitutive, global loss of EMRE (79). Our own investigations in constitutive Mcu^flfl × αMHC-Cre mice maintained on the C57/BL-6n background also showed no effect of chronic cardiomyocyte Mcu deletion on cardiac injury or dysfunction induced by ischemia-reperfusion injury (Figure 2), and therefore agree with the conclusion that alternative cell death pathways may be upregulated in response to chronic loss of MCU.

Figure 2– Perinatal deletion of cardiomyocyte MCU does not alter IR-injury.

A-B) There is no difference in infarct size as quantified by TTC and Evan’s blue staining (A), or in serum concentration of cardiac troponin-I (cTnI, B) between control (MCU^fl/fl) and MCU^fl/fl × αMHC-Cre mice subjected to 40 minutes of in vivo LCA ischemia and 24 hours of reperfusion. (n=9–11 mice / group). C-E) There is no difference in left ventricular end-diastolic dimension (LVEDD), end systolic dimension (LVESD), or percent fractional shortening (FS) between control and cardiac MCU-null mice as measured by echocardiography after 24 hours of reperfusion (n=11–14 mice / group). F-G) There is no difference in longitudinal or radial strain rates between control and cardiac MCU-null mice as measured by 2D echo after 24 hours of reperfusion (n=14–16 mice / group). Data are presented as mean ± S.E.M. All mouse models and experimental details for I/R are as previously reported in Luongo et al. (81). All animal experiments were approved by Temple University’s IACUC and followed AAALAC guidelines.

For instance, MCU-null hearts may be more sensitive to cell death pathways regulated by cytosolic Ca²⁺ concentration, and so preventing rapid _mCa²⁺ uptake would not have protected them from ischemic injury (80). Another possibility is that components of the mPTP itself remodel in the absence of MCU to increase the sensitivity to Ca²⁺-independent activation of the mPTP or to lower the Ca²⁺ threshold for pore opening. For example, Parks et al. recently reported that phosphorylation of Cyclophilin D (CypD S42) is increased in MCU-null hearts (98). Phosphorylation of CypD increased its association with the putative mPTP pore component, ATP synthase; amplified Ca²⁺ sensitivity of permeability transition; and increased ROS-induced cell death (98). It is unclear whether this phenotype reflects a direct sensitization of the mPTP to ROS, but if so, it could explain how MCU-null cardiomyocytes were still susceptible to mPTP opening and necrosis in the high-ROS conditions of IR, even when rapid _mCa²⁺ uptake is minimized (74). mPTP remodeling following MCU disruption may also be specific to particular cell types (i.e., cardiomyocytes), as MCU-null mouse embryonic fibroblasts (MEFs) were not sensitized to H₂O₂-induced cell death in earlier experiments (74).

Investigation of IR injury in the DN-MCU mouse, which has myocardial overexpression of the DN-MCU transgene throughout post-natal life, supports the idea that if given sufficient time to adapt to disruption of MCU function, cardiomyocytes will upregulate or remodel cell death pathways. Just like MCU gene-trap hearts, DN-MCU hearts are not protected from IR-induced cell death, even though they produce less ROS during IR than control hearts (91). More than 600 genes displayed a >2-fold difference in expression in DN-MCU hearts compared to controls, denoting the numerous compensatory changes in this model. Notably, the cell death gene Bax (99) was upregulated both at the mRNA and protein levels (91), providing one example of how DN-MCU cardiomyocytes may still be sensitized to IR-induced death even in the absence of acute _mCa²⁺ uptake. Thus, while initial investigations of the role of MCU in myocardial IR injury cast doubt onto the relevance of MCU for this process, a major complication to the interpretation of MCU’s role in pathological stress in mice with constitutive MCU disruption is their significant degree of post-natal/developmental compensatory changes. This is not unexpected given the fundamental importance of maintaining mitochondrial energetic signaling and responsiveness.

In order to evaluate the consequences of acute loss of MCU in the adult heart, while minimizing compensatory adaptations, our group and the Molkentin laboratory examined IR injury in Mcu^flfl × αMHC-MCM mice either 3 (81) or 6 (84) weeks after tamoxifen-induced deletion from adult cardiomyocytes. Both studies found that acute cardiomyocyte Mcu deletion effectively reduced cardiac infarct size, reduced cardiomyocyte cell death, and preserved contractile function of the heart following in vivo IR-injury. At the cellular level, acute Mcu deletion reduced the frequency of Ca²⁺-induced mPTP opening and mitochondrial swelling (81, 84). These results support the conclusion that MCU is a key pathway by which elevated cytosolic Ca²⁺ causes _mCa²⁺ overload, permeability transition, and cardiomyocyte death during IR injury (Figure 3).

Figure 3– MCU mediates acute _mCa²⁺ overload in the face of cytosolic Ca²⁺ overload, and triggers permeability transition and cardiomyocyte death in IR-injury.

_mCa²⁺ uptake through MCU can trigger mitochondrial permeability transition, leading to mitochondrial swelling, membrane rupture, and the initiation of necrotic cell death. This pathway is largely responsible for cardiomyocyte death in myocardial IR-injury and heart failure. Acute inhibition or genetic disruption of MCU function can prevent excessive _mCa²⁺ uptake, mPTP opening, and necrosis during IR-injury. With chronic, sustained MCU disruption, cardiomyocytes can remodel such that the protective effects against acute injury are lost. The mPTP may be sensitized to low levels of _mCa²⁺, and thus open even in the absence of rapid _mCa²⁺ uptake, perhaps in response to smaller, more gradual increases in _mCa²⁺ mediated by MCU-independent _mCa²⁺ uptake pathways. Chronic MCU disruption also causes upregulation of additional cell death pathways that may allow for cardiomyocyte death during IR-injury, even in the context of reduced _mCa²⁺ uptake.

The distinction between the results of studies using mice with chronic versus acute MCU disruption are particularly relevant for interpretation of the suitability of MCU as a therapeutic target for myocardial IR-injury moving forward. The models agree that MCU could be a good target for preventing _mCa²⁺ overload during IR. The main point of disagreement is whether mPTP opening induced specifically by an excessive increase in matrix Ca²⁺ concentration is the predominant driver of cardiomyocyte death following IR. We conclude that MCU is indispensable for cardiomyocyte death during the acute timeframe of IR injury, as would be seen clinically, and therefore MCU-dependent _mCa²⁺ uptake can be detrimental to heart function in pathological settings. This model supports the idea that transient MCU inhibition may be an attractive goal in order to mitigate reperfusion injury during clinical intervention. We further highlight that constitutive MCU mouse models engage numerous compensatory mechanisms that are sufficient to drive cell death in IR injury and related pathologies, even in the absence of acute _mCa²⁺ uptake (Figure 3).

MCU in cardiac responses to chronic changes in workload:

A number of studies have sought to understand the contribution of MCU to more chronic cardiac pathologies such as cardiac hypertrophy and non-ischemic heart failure. This is a pertinent question, as the role of _mCa²⁺ throughout the development and progression of hypertrophy and failure is debated. Some groups propose that depletion of _mCa²⁺ in end-stage heart failure leads to impaired energy metabolism that compromises contractile function (100). On the other hand, excess _mCa²⁺ uptake due to chronic elevations in diastolic Ca²⁺ levels would predispose to _mCa²⁺ overload and cardiomyocyte death that would likewise diminish contractile function. Increased MCU expression has been reported in both mouse and human hearts during physiological and pathological hypertrophy (101, 102), suggesting that MCU may have a causative role in cardiomyocyte growth in both normal development and disease. Fitting with this notion, increased _mCa²⁺ content promotes oxidation of pyruvate derived from glycolysis via activation of PDH, and in doing so can shift the balance between glucose and fatty acid metabolism in the heart (103). Cardiac hypertrophy has long been associated with a shift towards glycolysis that can provide TCA cycle-derived carbon intermediates needed to generate the molecular building blocks such as nucleotides and amino acids that are required for cell growth (reviewed in(104) and (105)). One might therefore hypothesize that prevention of MCU-dependent _mCa²⁺ uptake in situations of chronic, pathological cytosolic Ca²⁺ stress (such as pressure overload) could minimize the shift towards glucose metabolism and thereby minimize hypertrophic remodeling. Disruption of MCU-dependent _mCa²⁺ uptake – in the short term, at least – could also be predicted to prevent mPTP-dependent necrotic cardiomyocyte death and subsequent replacement fibrosis throughout the myocardium, and thus reduce the extent of pathological remodeling seen with pressure overload. Holmstrom et al. were the first to examine the specific consequences of MCU disruption on the heart’s response to chronic pressure overload, using mice with germline Mcu deletion. Their study found no effect of constitutive Mcu deletion on heart structure or function out to 8 weeks of transverse aortic constriction (TAC) in male mice 3–5 months of age (82). Kwong et al. obtained similar results in an 8-week TAC model starting 6 weeks after tamoxifen-induced cardiomyocyte Mcu deletion in adult mice. This experiment found no effect of Mcu deletion on heart weight/body weight ratio, cardiomyocyte cross-sectional area, or fractional shortening following TAC (84). Together, these findings would suggest that MCU is not required for the heart to adapt to sustained increases in hemodynamic load. One recent study attempted to investigate the role of MCU in pressure overload-induced cardiac hypertrophy and failure by administering the MCU inhibitor ruthenium red to mice subjected to 8 weeks of TAC (102). The study reported beneficial effects of ruthenium red to attenuate TAC-induced cardiac hypertrophy and contractile deficits. However, interpretation of these findings is hampered by previous reports that ruthenium red inhibits multiple Ca²⁺ channels that could impact the heart (106, 107).

Several limitations of these studies should temper any narrow interpretation of these results. First, as discussed in the context of the IR injury models, the length of time of Mcu deletion prior to TAC surgery (12–20 weeks of life with germline knockout (82) or 6 weeks in the conditional knockout (84)) may have been sufficient for activation of compensatory pathways that adapted the hearts to the lack of MCU and may have masked any immediate consequences of MCU disruption. Second, these studies evaluated cardiac structure at a single endpoint, 8 weeks post-surgery. It is therefore possible that any beneficial or detrimental effects of MCU ablation occurring earlier in the course of remodeling were simply missed in these analyses. Finally, in light of the potential involvement of slow MCU-independent _mCa²⁺ uptake pathways, it is also possible that the lack of a benefit of MCU disruption to alter the course of TAC-induced pathological cardiac remodeling could be explained by a failure of this experimental strategy to produce long-term alterations in _mCa²⁺ homeostasis. In other words, slower _mCa²⁺ uptake pathways may have been sufficient to eventually increase _mCa²⁺ content of TAC hearts even in the absence of MCU. It is worthwhile to point out here that, albeit methodologically difficult, neither of the two TAC studies (82, 84) measured _mCa²⁺ content at any stage post-TAC, so it is currently unknown whether acute or constitutive Mcu deletion was actually effective at limiting _mCa²⁺ levels throughout the course of the heart’s response to pressure overload. As such, it remains unresolved how matrix Ca²⁺ specifically affects the heart’s response to chronic stress. This is an intriguing question that is separable from the question of MCU’s particular involvement in adaptive or maladaptive functional and structural adaptations under hemodynamic stress. Yet, the limited data available indicate that MCU is in fact dispensable for this process.

MCU in non-cardiomyocyte cell populations:

Research into how MCU-dependent _mCa²⁺ uptake in non-cardiomyocyte cell populations may affect cardiac function is just beginning. This question may be important to understand the phenotypic differences in mouse models with global versus cardiomyocyte-specific disruption of mtCU function discussed above. In particular, cardiac fibroblasts influence myocardial remodeling and therefore contractile function following both ischemic injury and pressure overload. Pan et al. reported that acute _mCa²⁺ uptake is ablated in embryonic fibroblasts from MCU-null mice, which supports the functional relevance of MCU for this cell type (74). Likewise, human fibroblasts with loss-of-function mutations in the mtCU gatekeeper, MICU1, exhibit increased _mCa²⁺ uptake at low cytosolic Ca²⁺ concentrations (108), consistent with the involvement of the mtCU in fibroblast _mCa²⁺ handling. Recent studies in our lab indicate that _mCa²⁺ uptake is reduced as cardiac fibroblasts differentiate into myofibroblasts, due to upregulation of MICU1 (109). Genetic deletion of MCU from fibroblasts promoted myofibroblast differentiation, cardiac fibrosis, and subsequent contractile impairment in vivo following myocardial infarction, and also promoted cardiac fibrosis following chronic Angiotensin II infusion (109). These findings reveal that MCU function in other cell types is capable of shaping the heart’s responses to pathological stimuli. Preliminary evidence suggests that MCU-dependent _mCa²⁺ uptake can also influence the biology of endothelial cells (69) and vascular smooth muscle cells (78), but how MCU in these populations, or other specific cell types such as neurons, may influence the heart’s adaptation to physiological or pathological stress is currently unknown.

4. Conclusions and questions for future study:

Where do all of these varied findings leave our understanding of the role of MCU in the stressed heart? Considered together, they support the following working model: 1) MCU-dependent _mCa²⁺ uptake supports homeostatic cardiac metabolism and function under basal, unstressed conditions. This is perhaps not surprising given that the heart must pump continuously, even at rest, and so must have mechanisms in place to meet a constant energetic demand. The critical requirement for a continual ATP supply in the heart and need to protect energetics at all costs likely also explains why robust compensatory mechanisms exist that can bypass the effects of MCU-dependent _mCa²⁺ uptake if necessary (i.e, in cases of Mcu deletion or MCUB-transgene expression) in order to maintain sufficient ATP production that is able at least to sustain a normal, basal workload. 2) MCU is critical for rapid _mCa²⁺ uptake that allows the heart to quickly ramp up its metabolism in order to support a rapid physiological increase in cardiac demand, for instance during exercise or in sympathetic fight-or-flight responses. The heart has some capacity to partially compensate for lack of MCU (i.e, perhaps slow _mCa²⁺ uptake sufficient to increase matrix Ca²⁺ concentration after prolonged stimulation), but these mechanisms act slowly and are not sufficient to mediate rapid responses to an increase in cardiac workload. 3) MCU mediates acute, excessive _mCa²⁺ uptake that triggers mitochondrial permeability transition and necrotic cardiomyocyte death during ischemia-reperfusion injury. Thus, in the absence of other remodeling or adaptations, MCU itself is indispensable for acute ischemic tissue damage within the heart. Therefore, MCU may be an attractive therapeutic target to minimize such ischemic myocardial injury. 4) Cardiac adaptations to chronic increases in workload such as metabolic remodeling and cardiomyocyte hypertrophy can still occur independent of MCU. If these processes are indeed related to increased _mCa²⁺ content, it appears that alternative MCU-independent _mCa²⁺ uptake pathways are sufficient to support them. Future work exploring viral or transgenic MCU overexpression in the heart, as has recently been described in skeletal muscle (110), may offer an avenue to manipulate matrix Ca²⁺ homeostasis on a chronic timescale in order to further elucidate the role of matrix Ca²⁺ in chronic cardiac remodeling and non-ischemic heart failure. This approach could also provide insight into how fine-tuning net mitochondrial MCU activity—as may occur endogenously by transcriptional control or post-translational modification of MCU—may influence steady-state _mCa²⁺ content and the course of hypertrophic remodeling and/or cardiac failure.

Future studies will require new approaches to elucidate the various adaptive mechanisms that allow the heart to maintain function despite the persistent loss of acute mtCU-dependent _mCa²⁺ uptake. As the growing body of literature on the mtCU in the heart yields a clearer picture of the time course over which these functional adaptations occur following mtCU disruption, new studies will be able to compare matrix Ca²⁺ content, the mitochondrial proteome, and cellular and mitochondrial metabolism at non-adapted versus adapted time points. For example, a comparison of previously characterized models such as conditional Mcu^flfl × αMHC-MerCreMer mice (81, 84) at early time points after MCU disruption (~ 3 weeks post-tamoxifen administration) versus later time points, or inducible MCUB-transgenic mice at 1 week versus 1 month post-tamoxifen administration (86), will allow the field to correlate specific functional adaptations in the heart with corresponding molecular changes that may contribute to the adaptation. Such temporal analysis of animal models of mtCU disruption will expand upon initial studies that offered single snapshots of the alterations occurring in hearts that had already adapted to chronic loss of MCU function (91, 98).

Technological advances will facilitate future investigations into the mechanisms that adapt the heart to loss of mtCU-dependent _mCa²⁺ uptake. Three experimental goals are particularly relevant for achieving this aim. First, as noted above, reliable measurement of matrix Ca²⁺ content that accurately reflects its concentration in vivo remains challenging. Common techniques to isolate mitochondria from tissue take considerable time, during which Ca²⁺ exchange between the mitochondria and isolation buffers may occur and confound measurements of native _mCa²⁺ concentration. The use of mitochondrial-targeted genetic or virally-encoded ratiometric fluorescent Ca²⁺ reporters to measure _mCa²⁺ concentration in isolated cells or tissue slices from animals with mtCU disruption is one potential solution to this issue, but is relatively low-throughput. We also note that the existing animal studies of mtCU disruption described above have not yet been able to clearly distinguish the potential consequences of altered free ionic Ca²⁺ concentration versus altered levels of Ca²⁺-phosphate complexes on the heart. Thus, new techniques to measure both free _mCa²⁺ content as well as mitochondrial Ca²⁺-phosphate complexes are in high demand. Second, future studies should use mass spectrometry as an unbiased approach to identify changes in the mitochondrial or net cellular proteome that may account for the various functional adaptations to mtCU disruption. This method can reveal actual changes in protein expression, as compared to more traditional measurements of mRNA expression. Mass spectrometry has already been used to evaluate changes in the mitochondrial phosphoproteome of MCU-null hearts (98), and refinements in techniques for top-down mass spectrometry will facilitate global, quantitative assessment of changes in abundance or post-translational modification of the overall mitochondrial or cellular proteome. Third, greater resolution of the metabolic changes that occur following loss of acute _mCa²⁺ uptake will give better insight into the mechanisms that allow the heart to produce sufficient ATP to maintain basal contractility. Measurements of steady-state metabolite pools or the use of stable isotope resolved metabolomics to assess changes in flux through metabolic pathways in heart tissue or isolated cells are two approaches that could help achieve this goal.

Crossing the genetic mouse models of mtCU disruption described throughout this review with the recently described MITO-tag mouse model would improve the resolution of many of these proposed future measurements. The MITO-tag model allows for Cre-dependent expression of a HA-tagged EGFP localized to the outer mitochondrial membrane, which enables rapid, affinity-based isolation of mitochondria from a specific cell type of interest (111). This approach allows for specific analysis of mitochondria from cardiomyocytes, fibroblasts, or other cell populations of interest, thus minimizing confounding effects of contamination from other cell types. The increase in the speed of mitochondrial isolation using this approach would also minimize time-dependent changes in _mCa²⁺ or metabolite pools to better reflect the “endogenous” state of mitochondria. Selective, cell-permeable mtCU modulators are now being developed (112). Cell-permeable mtCU inhibitors will be used to further minimize changes in _mCa²⁺ content during mitochondrial isolations. Specific, cell-permeable mtCU activators and inhibitors would also enable acute experimental control of mtCU function to investigate the role of rapid _mCa²⁺ uptake in cardiac stress responses prior to the induction of compensatory adaptations.

Finally, it is worth noting that mice do not perfectly recapitulate all aspects of human cardiac physiology, such as ion fluxes during the action potential and excitation-contraction coupling, and the dynamic range of the heart rate response to sympathetic stimulation. As has been reviewed elsewhere (32, 113), such differences may result in humans exhibiting a greater reliance on _mCa²⁺ uptake through MCU for the regulation of cardiac metabolism during stress, as well as a greater influence of MCU on cardiac ROS generation. Additional work in large animal systems will therefore be needed to expand the conclusions drawn from studies of genetic mouse models of mtCU perturbation and determine their ultimate relevance for human biology.

In summary, MCU clearly appears to be indispensable for cardiac adaptation to acute physiological stress. Definitively delineating the involvement of MCU in the heart’s responses to more chronic and pathological pressures, and how MCU function may in turn be regulated in such conditions, will require ongoing research into the cellular consequences of altered _mCa²⁺ homeostasis and the fundamental biology of MCU itself.

Highlights.

• MCU is required to support increased cardiac contractility during sympathetic stimulation

• New data suggest MCU-dependent _mCa²⁺ uptake also supports basal cardiac function

• Acute MCU disruption protects against cardiac I/R injury

• The heart has robust compensatory mechanisms to adapt to chronic MCU disruption

• MCU-independent pathways may mediate _mCa²⁺ uptake during sustained stress

Glossary:

MCU

mitochondrial calcium uniporter

mtCU

mitochondrial calcium uniporter channel complex

Ca²⁺

calcium

_iCa²⁺

intracellular calcium

_cCa²⁺

cytosolic calcium

_mCa²⁺

mitochondrial calcium

IMM

inner mitochondrial membrane

OMM

outer mitochondrial membrane

IMS

intermembrane space

ETC

electron transport chain

TCA cycle

tricarboxylic acid cycle

OXPHOS

oxidative phosphorylation

mPTP

mitochondrial permeability transition pore

IR

ischemia-reperfusion

TAC

transverse aortic constriction