Adaptation to Elevated Mitochondrial Calcium Is Distinct in the Left and Right Ventricles

Abstract

Background:

Mitochondrial ATP production, essential for cardiomyocyte function, is regulated by mitochondrial Ca²⁺ (mtCa²⁺). The primary route for mtCa²⁺ influx is the mitochondrial calcium uniporter complex (mtCU). The mtCU subunit Mitochondrial Calcium Uptake 1 (MICU1) limits mtCa²⁺ uptake, preventing mtCa²⁺ overload. Although elevated mtCa²⁺ has been observed in multiple diseases including heart failure, its effects on heart function remain elusive.

Methods:

To investigate the impact of elevated mtCa²⁺ in adult hearts, we generated a mouse model with cardiomyocyte-specific tamoxifen-inducible Micu1 deletion (Micu1^cKO). Cardiac function was assessed through echocardiography. Mitochondria, adult cardiomyocytes, and tissue extracts were isolated from the left ventricle (LV) and right ventricle (RV) for comprehensive analysis at multiple timepoints ranging from 1 to 9 weeks post-tamoxifen injection.

Results:

Acute MICU1 deficiency resulted in increased mtCa²⁺ accompanied by reduced mitochondrial respiration in both the RV and LV. Contractile function, which was diminished in both ventricles initially, remained reduced in the RV upon prolonged MICU1 deficiency. In contrast, the LV exhibited signs of recovery over time, including restored ejection fraction concurrent with normalization of mtCa²⁺ levels. This pattern was mirrored in cardiomyocyte contractility. In Micu1^cKO RV, mtCa²⁺ remained elevated, likely contributing to oxidative stress. As a potential mechanism underlying LV-specific recovery, Essential MCU Regulator (EMRE), an mtCU subunit that promotes mtCa²⁺ uptake, was found to be downregulated only in the LV. This suggested that the LV initiated a compensatory response to elevated mtCa²⁺, while the RV remained impacted. Supporting this, proteomics analysis indicated a divergent proteomic signature in Micu1^cKO RV. Follow-up experiments suggested enhanced EMRE degradation in Micu1^cKO LV mediated by m-AAA proteases through a PKA-regulated mechanism. In MICU1-deficient neonatal cardiomyocytes, pharmacological PKA inhibition was sufficient to decrease EMRE levels. Analysis of LV tissues from dilated cardiomyopathy (DCM) patients suggested this pathway may be relevant in human DCM.

Conclusions:

While elevated mtCa²⁺ disrupted cardiac function in both ventricles, it induced an LV-specific adaptive response that suppressed mtCa²⁺ intake, contributing to the recovery of mitochondrial and cardiac function. The absence of this pathway in the RV has implications for therapeutics targeting RV dysfunction, a key determinant of mortality in heart failure.

Introduction

Heart failure (HF), a life-threatening syndrome affecting more than 64 million people worldwide, is characterized by a reduced capacity to pump blood into the circulatory system¹. At the molecular level, HF is associated with energy deprivation and oxidative stress^2,3. Cardiomyocyte energy supply relies heavily on mitochondrial respiration⁴, which generates ATP for contraction but also produces reactive oxygen species (ROS)⁵. In pathological conditions, excess ROS overwhelm mitochondrial antioxidant defenses, leading to heightened oxidative stress, which contributes to cytosolic Ca²⁺ mishandling and impairments in cardiomyocyte contraction⁶. Hence, mitochondrial dysfunction is an important underlying cause of HF^7,8. Ca²⁺ is a key modulator of mitochondrial function, as the uptake of Ca²⁺ into the mitochondrial matrix activates matrix-resident dehydrogenases to meet ATP demand in cardiomyocytes^9,10. However, mitochondrial Ca²⁺ (mtCa²⁺) overload, as well as ROS, can trigger mitochondrial permeability transition, which can lead to cardiomyocyte death^11,12. Accordingly, the dysregulation of mtCa²⁺, leading to either deficiency or excess, plays a critical role in the development of HF^13–18. Notably, high levels of mtCa²⁺ have been implicated in a number of human diseases beyond HF, including muscular dystrophies^19,20 and neurodegenerative diseases²¹ such as Parkinson’s²² and Alzheimer’s disease^22,23. This commonality suggests that elevated mtCa²⁺ is a feature shared among multiple disorders.

Physiologically, mtCa²⁺ levels are tightly regulated by Ca²⁺-specific ion channels located on the inner mitochondrial membrane. The major mtCa²⁺ efflux pathway in the heart is thought to be the mitochondrial Na⁺/Ca²⁺/Li⁺ exchanger (NCLX)²⁴. NCLX protein levels correlate with mtCa²⁺ efflux rates. Rapid mitochondrial Ca²⁺ uptake is primarily facilitated by the mitochondrial calcium uniporter complex (mtCU), which consists of a tetramer of the pore-forming protein MCU^25,26 to which other protein subunits associate²⁷. These regulatory mtCU subunits include Essential MCU Regulator (EMRE)^28,29 and mitochondrial calcium uptake (MICU) 1, 2, and 3, a family of EF-hand-containing calcium-binding proteins^30–32. The relative stoichiometry of mtCU components modulates mtCa²⁺ uptake kinetics and is known to vary among tissues, such as the heart and liver^33–35.

Multiple genetically modified cellular and animal models have defined the roles of individual mtCU subunits, beginning with MCU^36–38. Loss of MICU1 induces mtCa²⁺ overload due to increased uptake at below-threshold cytosolic Ca²⁺ concentrations, indicating that MICU1 acts as a gatekeeper of the mtCU^39–41. Although the extent of MICU1-dependent gatekeeping in the heart has been questioned^42,43, a recent study provided biochemical and genetic evidence to establish that MICU1 is required to suppress mtCa²⁺ uptake in cardiomyocytes during low resting cytosolic Ca²⁺ conditions⁴⁴. Deletion of Micu1 thus represents a means to induce the elevated mtCa²⁺ that is associated with a wide spectrum of diseases to investigate the direct effects of this specific mitochondrial stress.

Absence of the uniporter subunit EMRE eliminates rapid mtCa²⁺ uptake, mimicking the loss of MCU and demonstrating that EMRE is essential for mtCU activity^45,46. Interestingly, loss of one allele of Emre in mice with global Micu1 deletion rescues Micu1^−/−-associated perinatal mortality and developmental impairment; correspondingly, this reduction in EMRE expression relieves mtCa²⁺ overload in liver mitochondria³⁹. Hence, modulation of EMRE levels appears to be a physiological strategy to tune mtCU-mediated Ca²⁺ uptake rates and restore mtCa²⁺ homeostasis.

EMRE protein levels are regulated by the matrix-facing ATPase associated with diverse cellular activities proteases (m-AAA proteases), which degrade EMRE that is unbound to the complex^47–49. These m-AAA proteases are composed of homo-oligomers of the metalloprotease AFG3L2 or hetero-oligomers of AFG3L2 and SPG7 (also known as paraplegin). To form hetero-oligomers, SPG7 needs to be processed by AFG3L2. Accordingly, the stage of SPG7 maturation reflects the activity of AFG3L2⁵⁰. In mammalian cells, protein kinase A (PKA) indirectly phosphorylates a tyrosine residue in AFG3L2, thereby negatively regulating m-AAA protease activity⁵¹. Whether EMRE degradation by this pathway occurs in the context of elevated mtCa²⁺ in the heart is unknown.

Although frequently underappreciated, there are many differences in anatomy, function, and developmental origin between the right ventricle (RV) and left ventricle (LV)^52,53. We recently showed the RV and LV also exhibit differences in physiological mtCa²⁺ handling⁵⁴. In pathological conditions such as HF, the RV is more vulnerable to oxidative stress than the LV, potentially attributable to lower levels of ROS-scavenging enzymes and higher baseline mtCa²⁺ levels⁵⁵. Several studies show mitochondrial and metabolic abnormalities in RV failure^56–58, suggesting mitochondrial dysfunction as a causal factor and a potential therapeutic target in this condition. Indeed, clinical studies demonstrate that RV dysfunction is a strong predictor of mortality in HF^52,53,59. However, the potentially distinct responses of the two ventricles to mitochondrial dysfunction, specifically elevated mtCa²⁺, are unexplored.

To examine this important knowledge gap, we used cardiomyocyte-specific tamoxifen-inducible deletion of Micu1 in adult mice to investigate the in vivo effects of elevated mtCa²⁺ in the RV and LV. This approach allowed us to distinguish between the acute and long-term effects of mtCa²⁺ dysregulation, revealing an LV-specific adaptive response to Micu1 deficiency. Our findings indicate that this adaptation, triggered by prolonged elevated mtCa²⁺ levels, is achieved through a compensatory decrease in EMRE protein levels in the LV. In contrast, the RV shows a persistent phenotype with mtCa²⁺ dysregulation and contractile dysfunction, accompanied by a divergent proteomic signature. Moreover, we find evidence in LV tissues from healthy individuals and dilated cardiomyopathy patients that the adaptation we identify in our mouse model of elevated mtCa²⁺ may have clinical relevance. In uncovering distinct responses of the RV and LV to elevated mtCa²⁺, our findings provide critical insights into ventricular differences in mitochondria-related disease progression.

Methods

Data Availability:

The data that support the findings of this study are available from the corresponding author upon reasonable request. Detailed descriptions of the methods and materials can be found in the Supplemental Material.

Data analysis

Representative images were chosen to best represent the mean of the biological replicates shown in the quantifications. Results are presented as the mean ± standard deviation (SD). After confirming the normality of data by D’Agostino-Pearson or Kolmogorov-Smirnov test as specified in the figure legends, statistical analyses were performed using ANOVA with Tukey’s multiple comparisons test. If data were not normally distributed or sample size was too small for accurate normality testing, statistical analyses were performed using nonparametric tests, including Mann-Whitney test and Kruskal-Wallis test with Dunn’s multiple comparisons test. For data normalized to the control sample from the same experiment due to batch effects, a nonparametric Wilcoxon signed rank test was used to test the null hypothesis that the difference between the true median (1) and the sample median is equal to zero. For H₂O₂ flux data from Oroboros O2k experiments, statistical significance was determined using a mixed-effects model for repeated measures followed by Tukey’s multiple comparisons test. Differences with p values below 0.05 were considered statistically significant. Only within-test corrections for multiple comparisons were made; no corrections for multiple testing were made across tests. All statistical analyses were performed using GraphPad Prism (GraphPad, CA).

Results

Acute MICU1 deficiency induced elevated mtCa²⁺ and mitochondrial dysfunction in the heart

We crossed Micu1^fl/fl mice with αMHC-MerCreMer (MCM⁺) mice to generate cardiomyocyte-specific Micu1 knockout mice (Micu1^cKO mice) in which Micu1 deletion can be induced in adult mice through tamoxifen administration. At 1-week post-tamoxifen (tmx) injection, we harvested hearts and analyzed left and right ventricles (LV and RV) separately to confirm the induction of MICU1 deficiency in individual ventricles. Both LV and RV free wall tissues exhibited about 50% reduction in MICU1 protein levels (Fig. 1B). Micu1 mRNA levels also decreased in both ventricles to comparable extents (Fig. 1C). In addition, we confirmed successful recombination of the loxP-flanked target site within the Micu1 gene using DNA isolated from the LV and RV (Fig. 1D). Next, we isolated LV and RV cardiomyocytes from adult control (MCM⁺) and Micu1^cKO mice at 1-week post-tmx and assessed mtCa²⁺ levels using Rhod-2 AM, a Ca²⁺-sensitive dye that localizes to mitochondria. Image analysis of Z-stacks collected through each cardiomyocyte was performed to quantify Rhod-2 signal exclusively within the mitochondria, which were visualized using the dye MitoView Green (Fig. S1A). MICU1 deficiency resulted in an increase in mtCa²⁺ levels in both LV and RV cardiomyocytes relative to controls, consistent with previous reports^39,41,65 (Fig. 2A). To verify these results using an independent assay, we measured mtCa²⁺ levels in LV and RV mitochondria isolated from Micu1^cKO and Micu1^fl/fl hearts (Fig. S1B–C), and found that mtCa²⁺ levels were about 40% higher in cardiac mitochondria isolated from Micu1^cKO mice compared to controls (Fig. S1C). Mitochondrial Ca²⁺ activates pyruvate dehydrogenase (PDH) phosphatase, resulting in a decrease in the levels of phosphorylated PDH (p-PDH) under conditions of high mtCa²⁺. We found that the ratio of p-PDH to total PDH was reduced in cardiac mitochondria isolated from Micu1^cKO mice (Fig. 2B), consistent with measurements indicating elevated mtCa²⁺ levels in MICU1-deficient hearts here and in another study⁴⁴.

Figure 1. Cre recombinase-mediated targeting of the Micu1 gene in the LV and RV.

A, Diagram of experimental timeline. B, Western blot of Micu1^fl/fl and Micu1^cKO LV and RV tissues (n = 4 per group) at 1 week post-tmx for MICU1 protein. GAPDH was used as a loading control. Quantifications are shown to the right. Statistical significance was determined by nonparametric Kruskal-Wallis test. C, RT-qPCR for Micu1 in mRNA isolated from Micu1^fl/fl and Micu1^cKO LV and RV tissues (n = 4 per group) at 1 week post-tmx. Statistical significance was determined by nonparametric Kruskal-Wallis test. D, DNA gel electrophoresis of the floxed and the post-Cre recombined Micu1 gene in Micu1^fl/fl and Micu1^cKO LV and RV samples. A negative control lane with no DNA is at the right. Data are shown as mean ± SD.

Figure 2. Acute Micu1 deletion induced elevated mtCa²⁺ in the LV and RV.

A, Representative 3D reconstructions from Z-stacks of Rhod-2 AM fluorescence in MCM⁺ and Micu1^cKO LV and RV adult cardiomyocytes (ACMs) at 1 week post-tmx, with quantifications to the right. Scale bars represent 10 μm. For MCM⁺ n = 36 LV and n = 33 RV ACMs and for Micu1^cKO n = 35 LV and n = 31 RV ACMs from 3 mice per group. Normality was confirmed by D’Agostino-Pearson test, and statistical significance was determined by one-way ANOVA. B, Western blot of p-PDH and total PDH in LV and RV mitochondria of Micu1^fl/fl and Micu1^cKO at 1 week post-tmx. VDAC1 was used as a loading control. Quantifications are shown to the right. Statistical significance was determined by nonparametric Kruskal-Wallis test. C, Oxygen consumption rate (OCR) measured using Oroboros O2k in LV (filled markers) and RV (open markers) mitochondria provided with substrates pyruvate/glutamate/malate from control (Micu1^fl/fl black circles, n = 3; MCM⁺ black squares, n = 3) and Micu1^cKO (red circles, n = 5) mice at 1–2 weeks post-tmx. Statistical significance was determined by nonparametric Kruskal-Wallis test. D, Hydrogen peroxide flux measured using Amplex Red in an Oroboros O2k in LV (filled markers) and RV mitochondria (open markers) from control (Micu1^fl/fl black circles, n = 3; MCM⁺ black squares, n = 3) and Micu1^cKO (red circles, n = 5) mice at 1–2 weeks post-tmx. Statistical significance was determined using a mixed-effects model for repeated measures followed by Tukey’s multiple comparisons test. Data are shown as mean ± SD.

Having confirmed that Micu1 deletion induces mtCa²⁺ elevation to similar extents in the LV and RV, we next sought to assess the effects of increased mtCa²⁺ levels observed in Micu1^cKO cardiomyocytes on mitochondrial health. By simultaneously monitoring respiration and ROS generation in freshly isolated cardiac mitochondria in a controlled environment, we found that MICU1-deficient mitochondria from both ventricles exhibited decreased oxygen consumption rates (OCR) when supplied with the Complex I-linked substrates glutamate, malate, and pyruvate (Fig. 2C). Concurrently, higher rates of H₂O₂ production were detected in RV and LV mitochondria (Fig. 2D), suggesting elevated ROS. We therefore looked for indications of oxidative stress by analyzing overall carbonylation of mitochondrial proteins. Protein carbonylation levels were similar in Micu1^cKO and Micu1^fl/fl mitochondria isolated from the LV and RV, suggesting that the elevation of mtCa²⁺ for a 1-week period may be insufficient to induce excessive oxidative damage (Fig. S1E).

MICU1 deficiency led to impaired cardiomyocyte contractility and sustained RV dysfunction

We next sought to determine whether decreased mitochondrial OCR due to short-term MICU1 deficiency disrupted cardiomyocyte function in either ventricle. In paced LV and RV adult cardiomyocytes (ACMs) isolated from Micu1^cKO and control mice at 1–2 weeks post-tmx, we measured sarcomere shortening and cytosolic Ca²⁺ transients. In both LV and RV ACMs, short-term MICU1 deficiency led to lower peak amplitudes of cytosolic Ca²⁺ (Fig. 3A–B) and compromised sarcomere fractional shortening (Fig. 3C–D). These effects in Micu1^cKO LV and RV ACMs were reversed when the ACMs were preincubated with mitoTEMPO, a mitochondrial antioxidant, suggesting oxidative stress was a causative factor (Fig. 3A–D). Further analysis revealed slight increases in the time to rise and time to 50% decay in Micu1^cKO LV and RV ACMs in both Ca²⁺ transients (Fig. S2A–B, E–F) and contraction (Fig. S2C–D, G–H). These significantly slower rates, other than the cytosolic Ca²⁺ rise times in Micu1^cKO LV ACMs which did not reach significance (Fig. S2A), were also mitigated by mitoTEMPO (Fig. S2B–H). These results indicate that elevated mtCa²⁺ disrupts mitochondrial function in isolated cardiomyocytes, potentially through increased ROS levels.

Figure 3. MICU1 deletion led to sustained functional impairment in the RV.

A–B, IonOptix traces and quantifications of cytosolic Ca²⁺ transients of paced (A) LV (filled markers) and (B) RV (open markers) ACMs from control and Micu1^cKO mice at 1–2 weeks post-tmx. C–D, IonOptix traces and quantifications of sarcomere shortening of paced (C) LV (filled markers) and D) RV (open markers) ACMs from control and Micu1^cKO mice at 1–2 weeks post-tmx. In panels (A–D), markers indicate individual ACMs as follows: control (Micu1^fl/fl black circles, MCM⁺ black squares) and Micu1^cKO (red circles). Incubation with mitoTEMPO is indicated with dashed lines in the traces and markers in quantifications as follows: control (Micu1^fl/fl black triangles, MCM⁺ black diamonds) and Micu1^cKO (red triangles). For Micu1^fl/fl n = 13 and Micu1^fl/fl with mitoTEMPO n = 10 LV and RV ACMs from 4 mice; for MCM⁺ n = 11 and MCM⁺ with mitoTEMPO n = 9 LV and RV ACMs from 4 mice; for Micu1^cKO n = 24 and Micu1^cKO with mitoTEMPO n = 19 ACMs from 5 mice. For A-D, normality was confirmed by Kolmogorov-Smirnov test, and statistical significance was determined by two-way ANOVA. Data are shown as mean ± SD. E–H, Echocardiography acquired at baseline prior to the tmx injection, 4 weeks post-tmx, and 6 weeks post-tmx was used to analyze (E) LV ejection fraction, (F) RV free wall thickness change, (G) RV stroke volume, and (H) pulmonary arterial acceleration time. Red indicates data from Micu1^cKO mice. At baseline, 4 weeks post-tmx, and 6 weeks post-tmx, respectively, n = 26, 36, 29 for MCM⁺, n = 16, 46, 24 for Micu1^fl/fl, and n = 20, 46, 31 for Micu1^cKO mice. For E-H, normality was confirmed by Kolmogorov-Smirnov test, and statistical significance was determined by one-way ANOVA within each time point. Data are shown as box and whisker plots, in which the line within the box represents the median and whiskers show the 10^th and 90^th percentiles. Data outside the 10^th to 90^th percentile are drawn as individual dots.

As MICU1 deficiency altered contractility and Ca²⁺ transients in both LV and RV cardiomyocytes, we next examined its effects on in vivo cardiac function. Using echocardiography, we measured parameters of LV and RV contractility in Micu1^cKO and control mice at 4 weeks post-tmx to avoid the effects of tmx-associated transient cardiomyopathy, which can last for up to two weeks after the injections⁶⁶. First, we confirmed that LV ejection fraction (EF) was similar in all groups at baseline prior to tmx injection (Fig 3E, baseline). At 4 weeks post-tmx, LVEF in Micu1^cKO mice was significantly reduced compared to the controls. However, no such difference was detected at 6 weeks post-tmx, suggesting that LV function recovered from mtCa²⁺ elevation (Fig. 3E). LV end systolic volume was elevated in Micu1^cKO mice relative to Micu1^fl/fl mice only at 4 weeks and not 6 weeks; otherwise, LV dimensions and heart rate did not differ between groups (Fig. S3A–C). To gain insights into RV contractile function, we compared RV free wall thickness change (Fig. 3F) and estimated RV stroke volume (Fig. 3G). These indices were significantly lower at 4 weeks post-tmx in Micu1^cKO mice relative to controls, and this impairment in RV persisted at 6 weeks post-tmx (Fig. 3F–G). To determine whether pulmonary arterial hypertension (PAH) contributed to the deterioration of RV function in Micu1^cKO mice, we measured pulmonary acceleration time in pulsed-wave doppler mode (Fig. 3H) and assessed medial thickness change in pulmonary arterioles through histological analysis of lung tissues (Fig. S3D). We did not find any changes in these parameters, implying RV dysfunction was not due to heightened afterload, and that elevated mtCa²⁺ was the driver of RV hypocontractility.

MICU1 deficiency did not result in altered LV or RV weight relative to body weight (Fig. S3E), indicating that hypertrophy was not induced. Lung weight relative to body weight was higher in Micu1^cKO mice at 1-week post-tmx (Fig. S3F) but returned to control levels at 7–9 weeks post-tmx (Fig. S3G), consistent with transiently impaired LV function. We found no evidence of increased fibrosis in either the LV or RV from Micu1^cKO mice at 7–9 weeks post-tmx (Fig. S3H). To determine the effects of prolonged Micu1 deficiency on the heart and pulmonary circulation as mice aged, we carried out similar experiments on Micu1^cKO and control mice at 11 months post-tmx. Echocardiography data indicated that LVEF in Micu1^cKO mice declined relative to Micu1^fl/fl mice at 11 months post-tmx, suggesting that LV functional recovery is not permanent and impairment still manifests over time, albeit to a modest extent nearly approaching significance (Fig. S4A). In contrast, RV function in Micu1^cKO mice at 11 months post-tmx remained impaired to a similar extent as at 6 weeks post-tmx (Fig. S4B). Furthermore, we harvested serum, lung, and heart samples at 11 months post-tmx to compare pulmonary arteriole structure, heart weight, cardiac fibrosis, and heart failure and circulatory blood markers. Although pulmonary acceleration time was slightly reduced in Micu1^cKO mice (Fig. S4C), measurements of pulmonary arterial wall thickness remained similar in Micu1^cKO and control mice at 11 months post-tmx (Fig. S4D), suggesting only slight, if any, pulmonary hypertension. Furthermore, no significant changes were observed in the serum levels of the muscle damage marker creatine kinase (CK) or systemic circulatory congestion markers (Fig. S4E–G). We confirmed that there were no overt pathological changes in heart weight to body weight ratio (Fig. S4H), indicating a continued absence of cardiac hypertrophy, and no substantial fibrosis in the LV and RV free walls (Fig. S4I). Taken together, we found that chronic Micu1 deficiency did not lead to cardiac hypertrophy, fibrosis, pulmonary hypertension, or systemic circulatory congestion. These findings strongly suggest the presence of adaptive responses to elevated mtCa²⁺, particularly in the LV given its partial recovery of contractile function from 4 to 6 weeks post-tmx.

Micu1^cKO LV mitochondria exhibited adaptive changes to restore mtCa²⁺ levels in the long term

To understand why LV function improved over time, we investigated potential molecular mechanisms underlying the divergent LV and RV responses to elevated mtCa²⁺. We isolated LV and RV mitochondria at 7–9 weeks post-tmx from Micu1^cKO and control mice to assess whether the levels of mtCa²⁺ or proteins involved in mtCa²⁺ regulation changed over time. MICU1 protein levels were still only partially depleted in Micu1^cKO LV and RV ACMs (Fig. S5A), which likely reflects incomplete recombination as Cre excision rates can be variable depending on the floxed loci⁶⁷, though high stability and long half-life of MICU1 may contribute. Although mtCa²⁺ levels were still elevated in Micu1^cKO RV cardiomyocytes, mtCa²⁺ levels in Micu1^cKO LV cardiomyocytes were not significantly different from control levels (Fig. 4A). Similar findings were obtained using cardiac mitochondria isolated from Micu1^cKO mice, in which mtCa²⁺ levels remained higher in RV mitochondria but not in LV mitochondria (Fig. S5B–D). In this assay, mtCa²⁺ levels appeared lower in the MICU1-deficient LV, consistent with a trend towards an increase in PDH phosphorylation (Fig. S5E). To investigate how mtCa²⁺ levels decreased in the LV over time despite MICU1 deficiency, we compared LV and RV protein levels of the mtCU subunits EMRE and MCU (Fig. 4B), as well as the mtCa²⁺ handling proteins LETM1 and NCLX (Fig. S6A). EMRE, a positive regulator of mtCU activity, was the only protein that showed altered expression due to MICU1 deficiency. At 7–9 weeks post-tmx, EMRE levels were significantly reduced relative to controls in Micu1^cKO mitochondria isolated from the LV but not in those from the RV (Fig. 4B). Decreased EMRE levels were not apparent at 1 week post-tmx (Fig. S6B), suggesting that EMRE downregulation took place gradually over time. At 7–9 weeks post-tmx, consistent with decreased EMRE levels, Micu1^cKO mitochondria from the LV and not the RV exhibited slower rates of mtCa²⁺ uptake compared to Micu1^fl/fl mitochondria (Fig. 4C). These findings suggest that, of the two ventricles, the LV may have a higher capacity to compensate for Micu1 deletion-induced mtCa²⁺ elevation by decreasing mtCU activity, potentially enabling a greater degree of functional recovery over time.

Figure 4. MICU1 deficiency led to persistently elevated mtCa²⁺ levels in the RV.

A, Representative 3D reconstructions from Z-stacks of Rhod-2 AM fluorescence in MCM⁺ and Micu1^cKO LV and RV adult cardiomyocytes at 7–9 weeks post-tmx, with quantifications to the right. Scale bars represent 10 μm. For MCM⁺ n = 26 LV and n = 29 RV ACMs and for Micu1^cKO n = 27 LV and n = 30 RV ACMs from 3 mice per group. B, Western blot of EMRE and MCU in Micu1^fl/fl (n = 11) and Micu1^cKO (n = 11) LV and RV mitochondria at 7–9 weeks post-tmx. VDAC1 was used as a loading control. Quantifications are shown to the right. For A and B, normality was confirmed by Kolmogorov-Smirnov test, and statistical significance was determined by one-way ANOVA. C, Representative mtCa²⁺ uptake trace from Micu1^fl/fl (black) and Micu1^cKO (red) in LV (solid) and RV (dashed) mitochondria. Quantifications of mtCa²⁺ uptake rates normalized to LV Micu1^fl/fl mitochondria at 7–9 weeks post-tmx: 0.632 ± 0.276 (Micu1^cKO LV), 0.948 ± 0.821 (Micu1^fl/fl RV), and 1.024 ± 0.580 (Micu1^cKO RV) are shown to the right. Nonparametric Wilcoxon signed rank test was used to test whether the rates normalized to LV Micu1^fl/fl mitochondria were different from 1, and * denotes that 0.0625 is the lowest attainable p-value given the sample size. D, Oxygen consumption rate (OCR) measured using Oroboros O2k in LV (filled markers) and RV (open markers) mitochondria provided with pyruvate/glutamate/malate from control (Micu1^fl/fl black circles, n = 3; MCM⁺ black squares, n = 3) and Micu1^cKO (red circles, n = 6) mice at 7–9 weeks post-tmx. Normality was confirmed by D’Agostino-Pearson test, and statistical significance was determined by ANOVA. E, Hydrogen peroxide flux measured using Amplex Red in an Oroboros O2k in LV (filled markers) and RV mitochondria (open markers) from control (Micu1^fl/fl black circles, n = 3; MCM⁺ black squares, n = 3) and Micu1^cKO (red circles, n = 6) mice at 7–9 weeks post-tmx. Statistical significance was determined using a mixed-effects model for repeated measures followed by Tukey’s multiple comparisons test. Data are shown as mean ± SD.

To determine whether the compensatory pathways that largely restored mtCa²⁺ levels in Micu1^cKO LV mitochondria also improved mitochondrial respiration, we measured OCR in LV and RV mitochondria 7–9 weeks post-tmx. We found that at this timepoint, there was no longer a statistically significant difference in OCR between Micu1^cKO and control LV mitochondria (Fig. 4D). In contrast, OCR in MICU1-deficient RV mitochondria still remained significantly depressed relative to control RV mitochondria. However, we still detected elevated mitochondrial H₂O₂ production in both ventricles, albeit to a lesser extent in the LV compared to the RV (Fig. 4E). To assess the extent of oxidative damage induced by long-term MICU1 deficiency in LV and RV mitochondria, we evaluated protein carbonylation and lipid peroxidation. At 7–9 weeks post-tmx, protein carbonylation levels in RV but not LV mitochondria from Micu1^cKO mice were elevated relative to controls (Fig. 5A). Consistent with this, the levels of BODIPY 581/591 C11, a live cell fluorescent indicator of lipid peroxidation, were significantly higher in Micu1^cKO ACMs isolated from the RV (Fig. 5B) but not in those isolated from the LV. MICU1 deficiency did not affect the protein levels of the ROS scavenging enzymes catalase and SOD2 in either ventricle (Fig. S5E, S6C). Overall, these data suggested that the sustained elevation of mtCa²⁺ in RV Micu1^cKO cardiomyocytes led to persistent impairments in mitochondrial respiration and increased ROS generation that accumulated sufficiently to result in oxidative stress, while the compensatory downregulation of EMRE in LV Micu1^cKO cardiomyocytes led to a partial protection.

Figure 5. MICU1 deficiency increased oxidative damage and affected cardiomyocyte contractility in the RV.

A, Western blot of DNPH in freshly isolated mitochondria from Micu1^fl/fl and Micu1^cKO LV and RV at 7–9 weeks post-tmx. VDAC1 was used as a loading control. Quantification of signal from entire lanes (n = 8 per group) is shown to the right. B, Representative confocal images for lipid peroxidation visualized using Bodipy C11 (green) and MitoTracker Orange imaging in live LV and RV ACMs isolated from control Micu1^fl/fl and Micu1^cKO mice. Scale bars represent 10 μm. Quantification of mean Bodipy C11 signal from 30 LV and RV ACMs from 3 Micu1^fl/fl and 3 Micu1^cKO mice is shown to the right. For A and B, normality was confirmed by Kolmogorov-Smirnov test, and statistical significance was determined by one-way ANOVA. C–D, IonOptix traces and quantifications of cytosolic Ca²⁺ transients of paced (C) LV (filled markers) and (D) RV (open markers) ACMs from control and Micu1^cKO mice at 7–9 weeks post-tmx. E–F, IonOptix traces and quantifications of sarcomere shortening of paced (E) LV (filled markers) and (F) RV (open markers) ACMs from control and Micu1^cKO mice at 7–9 weeks post-tmx. In panels (C–F), markers indicate individual ACMs as follows: control (Micu1^fl/fl black circles, MCM⁺ black squares) and Micu1^cKO (red circles). Incubation with mitoTEMPO is indicated with dashed lines in the traces and markers in quantifications as follows: control (Micu1^fl/fl black triangles, MCM⁺ black diamonds) and Micu1^cKO (red triangles). For Micu1^fl/fl n = 13 and Micu1^fl/fl with mitoTEMPO n = 10 LV and RV ACMs from 4 mice; for MCM⁺ n = 11 and MCM⁺ with mitoTEMPO n = 9 LV and RV ACMs from 6 mice; for Micu1^cKO n = 24 and Micu1^cKO with mitoTEMPO n = 19 LV and RV ACMs obtained from 5 mice. For C-F, normality was confirmed by Kolmogorov-Smirnov test, and statistical significance was determined by two-way ANOVA. Data are shown as mean ± SD.

To assess the functional consequences of long-term MICU1 deficiency in the LV and RV, we examined adult cardiomyocytes isolated at 7–9 weeks post-tamoxifen using the IonOptix system. Micu1^cKO LV ACMs no longer exhibited any discernable decrease in the peak amplitude of cytosolic Ca²⁺ transients or sarcomere shortening relative to control LV ACMs (Fig. 5C and 5E). In contrast, Ca²⁺ amplitude and sarcomere shortening remained suppressed in Micu1^cKO RV ACMs relative to control RV ACMs (Fig. 5D and 5F). The time intervals to 50% rise and 50% decay in cytosolic Ca²⁺ transients and sarcomere shortening were indistinguishable in Micu1^cKO and control LV ACMs (Fig. S7A, C, E, G). These parameters were slightly but significantly prolonged in Micu1^cKO RV ACMs relative to control RV ACMs, which was mitigated by mitoTEMPO pre-incubation (Fig. S7B, D, F, G). These results suggested that the compensatory mechanisms to counteract elevated mtCa²⁺ in LV cardiomyocytes were sufficient to restore normal function, in line with the recovery in LV EF measured by echocardiography.

Proteomics analysis revealed RV-specific alterations in Micu1^cKO mice

Next, we performed quantitative proteomics on cytosol-enriched, myofilament-depleted fractions of LV and RV cardiac tissues from Micu1^cKO and Micu1^fl/fl mice at 9 weeks post-tmx to gain unbiased insights into the ventricle-specific alterations induced by MICU1 deficiency. We quantitatively identified >3000 proteins in total. Hierarchical cluster analysis demonstrated that the Micu1^cKO RV had the most divergent proteomic signature (Fig. 6A). Then, we performed ventricle-specific evaluations and showed separation between Micu1^fl/fl and Micu1^cKO RVs using partial least squares discriminant analysis (PLSDA) (Fig. 6B). KEGG pathway analysis of the top 250 proteins defined by the variable importance in projection between Micu1^cKO RVs and Micu1^fl/fl RVs identified multiple metabolic pathways. Further analysis using WikiPathways also revealed enrichment in metabolic as well as muscle contraction pathways (Fig. 6D). Next, we analyzed the differences between the RV and LV proteomes of Micu1^cKO mice (Fig. 6E). We found clear separation between Micu1^cKO RVs compared to Micu1^cKO LVs using PLSDA (Fig. 6F). KEGG and WikiPathways analysis of the 250 proteins important for distinguishing the ventricles suggested distinct metabolic changes between the two chambers, and additionally ROS and disease-related pathways were implicated (Fig. 6G-H). Overall, this unbiased analysis indicated that MICU1 deficiency resulted in RV-specific, long-lasting changes in the heart proteome.

Figure 6. Proteomics analysis revealed RV-specific alterations in Micu1^cKO mice.

A, Hierarchical cluster analysis of proteins from the LV and RV free walls of Micu1^cKO and Micu1^fl/fl mice (n = 4 per group) at 9 weeks post-tmx. B, Partial least squares discriminant analysis (PLSDA) of Micu1^cKO RV and Micu1^fl/fl RV. C, KEGG pathway analysis of the top 250 proteins defined by the variable importance in projection in Micu1^cKO RV relative to Micu1^fl/fl RV. D, WikiPathways analysis of the top 250 proteins defined by the variable importance in projection in Micu1^cKO RV relative to Micu1^fl/fl RV. E, Hierarchical cluster analysis of proteins in Micu1^cKO RV and Micu1^cKO LV. F, PLSDA of Micu1^cKO RV and Micu1^cKO LV. G, KEGG pathway analysis of the top 250 proteins defined by the variable importance in projection in Micu1^cKO RV relative to Micu1^cKO LV. H, WikiPathways analysis of the top 250 proteins defined by the variable importance in projection in Micu1^cKO RV relative to Micu1^cKO LV.

Micu1^cKO LV mitochondria exhibited increased m-AAA protease activity

We next sought to identify potential molecular mechanisms underlying the compensatory decrease in EMRE levels observed in Micu1^cKO LV mitochondria at 7–9 weeks post-tmx. First, we asked whether Emre mRNA levels mirrored the changes in its protein levels. We found that Micu1 deficiency did not affect mRNA levels of Emre, suggesting that EMRE was downregulated post-translationally (Fig. S8A). EMRE protein stability is known to be regulated through degradation by mitochondrial m-AAA proteases^48,49. Key components of these complexes are AFG3L2 and the mature form of SPG7, which is generated through the cleavage of SPG7 precursor by AFG3L2. We did not find any significant difference in the levels of AFG3L2 between Micu1^cKO and Micu1^fl/fl mitochondria from either the LV or RV at 7–9 weeks post-tmx (Fig. S8B). However, the levels of mature SPG7 were significantly elevated in Micu1^cKO LV tissues, which was not observed in the RV (Fig. 7A). These data indicated elevated AFG3L2 activity, promoting the formation of active m-AAA protease complexes in MICU1-deficient LV. AFG3L2 activity has been reported to be indirectly suppressed by phosphorylated PKA (p-PKA)⁵¹. To investigate whether changes in PKA phosphorylation could be a factor contributing to increased AFG3L2 activity in the LV, we compared the ratio of p-PKA to total PKA in both LV and RV tissues and found that MICU1 deficiency resulted in a reduction in PKA phosphorylation only in the LV (Fig. 7B). Correspondingly, analysis of the cytosolic phospho-proteome revealed downward trends in the phosphorylation of multiple PKA targets in Micu1^cKO LV (Fig. S8C), with the notable exception of phosphorylated phospholamban, which was significantly decreased in Micu1^cKO RV, consistent with slower cytosolic Ca²⁺ transient decay (Fig. S6). Overall, these data suggest that elevated mtCa²⁺ overload induced by MICU1 deficiency leads to an adaptive decrease in EMRE protein levels in the LV but not the RV, potentially through a compensatory mechanism involving reduced PKA phosphorylation and elevated EMRE degradation by mitochondrial m-AAA proteases.

Figure 7. EMRE levels decreased as m-AAA protease activity increased in MICU1-deficient LV mitochondria.

A, Western blot of SPG7 isoforms in LV and RV tissues of Micu1^fl/fl and Micu1^cKO mice (n = 4 per group) at 7–9 weeks post-tmx. GAPDH was used as a loading control. Percentage of the mature SPG7 isoform was calculated as mature SPG divided by the total of precursor (p), intermediate (i), and mature (m) SPG7. Quantifications are shown to the right. B, Western blot of p-PKA and total PKA in LV and RV tissues of Micu1^fl/fl and Micu1^cKO mice (n = 4 per group). GAPDH was used as a loading control. Quantifications are shown to the right. For A and B, statistical significance was determined by nonparametric Kruskal-Wallis test. C, RT-qPCR of Micu1 in NMCMs from Micu1^fl/fl pups after 5 days of GFP (n = 3) or CRE adenovirus (n = 4) infection. Statistical significance was determined by nonparametric Mann-Whitney test. D, Representative confocal images for mtCa²⁺ visualized using Rhod-2 AM staining in live NMCMs (n = 66 for GFP DMSO, 88 for Cre DMSO, 56 for GFP H89, and 69 for Cre H89 cells from 3 biological replicates), with quantifications to the right. Scale bars represent 10 μm. Statistical significance was determined by nonparametric Kruskal-Wallis test. E, Western blots of p-PDH, total PDH, p-PKA, total PKA, EMRE, MCU, and precursor (p), intermediate (i), and mature (m) SPG7, after 5 days of GFP or CRE adenovirus infection in NMCMs (n = 5 biological replicates) treated or untreated with H89. GAPDH was used as a loading control. Quantifications of protein levels normalized to GFP with vehicle (no H89) control samples from the same batch are shown to the right. Nonparametric Wilcoxon signed rank test was used to test whether the normalized protein levels were different from 1, and * denotes that 0.0625 is the lowest attainable p-value given the sample size. Data are shown as mean ± SD.

Inhibition of p-PKA downregulated EMRE in Micu1-deficient neonatal cardiomyocytes

Having established a potential correlation between p-PKA downregulation and EMRE degradation in vivo, we aimed to determine whether this link might reflect a causative relation in cardiomyocytes. To this end, we utilized an in vitro neonatal mouse cardiomyocyte (NMCM) system to examine the effect of suppressed PKA phosphorylation on EMRE protein levels in both control and MICU1-deficient cells. We isolated neonatal cardiomyocytes from newborn Micu1^fl/fl pups and infected them with either GFP or Cre adenovirus to generate control and Micu1-deficient NMCMs, respectively. First, we confirmed that Micu1 mRNA expression and protein levels were reduced in Micu1 KO NMCMs 5 days after viral transduction (Fig. 7C and S8D). Similar to the result obtained in vivo in MICU1-deficient hearts shortly post-tmx, MICU1-deficient NMCMs exhibited an increase in mtCa²⁺ levels (Fig. 7D) and a reduction in the p-PDH/total PDH ratio (Fig. 7E). Of note, acute MICU1 deficiency did not induce any significant changes in EMRE levels or PKA phosphorylation, unlike the results obtained in the LV after long-term MICU1 deficiency (Fig. 7E). To directly examine the effect of reduced PKA phosphorylation, we treated NMCMs with the PKA inhibitor H89 and confirmed the H89-induced decrease in PKA phosphorylation, mimicking the change observed in the LV in response to long-term MICU1 deficiency. We found that, both in control and MICU1-deficient NMCMs, the inhibition of PKA phosphorylation enhanced the maturation of SPG7 and resulted in a decrease in EMRE protein levels (Fig. 7E). Furthermore, it abrogated the decrease in p-PDH/total PDH ratio in Micu1-deficient cells (Fig. 7E), corresponding to a partial reversal of mtCa²⁺ elevation (Fig. 7D). These data suggest that PKA inhibition may be sufficient to ameliorate MICU1 deficiency-induced mtCa²⁺ elevation in cardiomyocytes, supporting the role of reduced p-PKA in the compensatory responses observed in vivo in the LV.

EMRE was downregulated in LV tissues of patients with dilated cardiomyopathy

Mitochondrial dysfunction is associated with multiple forms of HF, including dilated cardiomyopathy (DCM)⁶⁸. To explore the clinical relevance of the LV-specific adaptation observed in our mouse model with elevated mtCa²⁺, we analyzed LV samples from individuals with non-failing (NF) hearts and patients with DCM (Fig. 8, Table 1). MICU1 protein levels and PDH phosphorylation levels were significantly lower in DCM samples, indirectly suggesting conditions of elevated mtCa²⁺. Furthermore, EMRE protein levels were decreased in DCM samples, while the levels of mature SPG7 were increased, suggesting an upregulation of m-AAA protease activity and a limitation of mtCU activity through EMRE downregulation, similar to our findings in Micu1^cKO LV tissues. Upstream, p-PKA levels in DCM samples showed a decreasing trend, which did not reach statistical significance. Unlike the phenotype observed in the LV of Micu1^cKO mice (Fig. S5D), NCLX levels were elevated in DCM LV compared to NF LV, indicating the use of an additional mechanism to limit mtCa²⁺ levels by promoting mtCa²⁺ efflux. Overall, despite some differences, we found significant similarities between the LV of DCM patients and the LV of Micu1^cKO mice, suggesting that m-AAA protease-mediated EMRE degradation might be a common mechanism to counteract mtCa²⁺ elevation in both cases. This raises the possibility that this adaptive response to restore mtCa²⁺ homeostasis may be conserved across species.

Figure 8. EMRE downregulation is evident in human LV tissues of dilated cardiomyopathy patients.

Western blot of MICU1, p-PDH, total PDH, p-PKA, total PKA, EMRE, MCU, NCLX, VDAC1, SPG7, and GAPDH in LV tissues of DCM (n = 5) and non-failing human hearts (NF) (n = 5). For p-PDH and total PDH, n = 4 per group. GAPDH was used as a loading control. Quantifications are shown in the graphs. Statistical significance was determined by nonparametric Mann-Whitney test. Data are shown as mean ± SD.

Table 1.

Patient data

Patient Study ID	Age at time of transplant	Gender	Race	Hispanic or Latino	Diagnosis	Device Therapy	Ejection Fraction (EF; %)	Milrinone (PDEI)	Dobutamine	Epinephrine	Norepinephrine	Dopamine	Vasopressin	Beta-Blocker	ACE Inhibitor
1	68	Male	White	No	Nonfailing	None	62	No	No	No	Yes	No	No	None	No
2	54	Male	White	No	Nonfailing	None	55	Unknown	Unknown	Unknown	Yes	Unknown	Yes	Unknown	Unknown
3	41	Female	White	Unknown	Nonfailing	None	75	Unknown	Unknown	Unknown	Unknown	Unknown	Yes	Unknown	Unknown
4	53	Male	White	No	Nonfailing	None	60	No	No	Yes	Yes	Yes	Yes	None	No
5	61	Male	White	No	Nonfailing	None	55	Unknown	Unknown	Yes	Yes	Unknown	Yes	Unknown	Unknown
6	68	Female	White	No	Idiopathic Dilated Cardiomyopathy	None	29.8	Unknown	Yes	Unknown	Unknown	Unknown	Unknown	Unknown	Unknown
7	38	Male	White	Yes	Idiopathic Dilated Cardiomyopathy	Uncertain	4.8	No	No	No	No	No	No	metoprolol succinate	Yes
8	65	Female	White	No	Idiopathic Dilated Cardiomyopathy	LVAD	11.5	No	No	No	No	No	No	None	Yes
9	49	Male	White	No	Idiopathic Dilated Cardiomyopathy	None	20	Unknown	Unknown	Unknown	Unknown	Unknown	Unknown	metoprolol succinate	Unknown
10	53	Male	White	Yes	Idiopathic Dilated Cardiomyopathy	LVAD	Unknown	No	No	No	No	No	No	None	No

Discussion

In this study, we investigated the physiological impact of elevated mtCa²⁺ in the heart using a mouse model with inducible cardiomyocyte-specific MICU1 deficiency. Shortly after tamoxifen treatment in adult Micu1^cKO mice, we observed similar increases in mtCa²⁺ levels in the RV and LV, leading to an initial decrease in contractility in both ventricles. Mitochondria from both ventricles exhibited decreased oxygen consumption rates and increased ROS production rates; while seemingly paradoxical, this is consistent with the complex and non-linear relationship between electron transport chain flux and ROS generation⁶⁹. In some tissues, increased respiratory rate can prevent ROS release⁷⁰, and in conditions of mitochondrial impairment and disease, a decrease in oxygen consumption rate can in fact increase ROS generation^71,72. Over time, the LV of Micu1^cKO mice showed signs of recovery while RV dysfunction persisted. Our finding that a 50% reduction in MICU1 protein was sufficient to increase mtCa²⁺ levels in both ventricles supports existing literature indicating that MICU1 plays a critical role as a gatekeeper of the mtCU in various tissues^39,41,65, including the heart⁴⁴. Across tissues, MICU1 deficiency appears to elevate mtCa²⁺; moreover, in many tissues, including the liver, skeletal muscle, and brain, this is followed by a decrease in EMRE protein expression^39–41,73. The downregulation of EMRE over time in global Micu1 knockout mice has been shown to ameliorate the pathological outcomes of elevated mtCa²⁺ by contributing to recovery of mtCa²⁺ homeostasis³⁹. Interestingly, in mice with cardiomyocyte-specific MICU1 deficiency, we observed a decline in EMRE levels only in the LV and not in the RV. Furthermore, this LV-specific EMRE downregulation was associated with a compensatory decrease in mtCU activity to counteract elevated mtCa²⁺ levels, leading to a partial restoration of LV contractile function. The RV, which appeared to lack this adaptive response, did not show improvement in its impaired contractility over time, demonstrating for the first time a link between the ability to adapt to elevated mtCa²⁺ through EMRE downregulation and physiological cardiac function. Recent clinical studies show improvements in RV contractility are associated with lower mortality rate in heart failure patients, underscoring the need to develop targeted therapeutics for the RV. Importantly, mitochondrial dysfunction is observed in RV failure and has been proposed as a potential target for enhancing RV contractility^52,53,59. Overall, our findings indicate that targeting the imbalances in mtCa²⁺ dynamics could be a therapeutic approach for cardiac diseases, including heart failure, and furthermore suggest that ventricle-specific approaches should be considered.

Several publications have demonstrated that m-AAA proteases are critical for the maintenance of physiological mtCU stoichiometry^47,48. Specifically, they can influence mtCU-mediated mtCa²⁺ influx by regulating EMRE levels through degradation. Our results show that SPG7 maturation is enhanced in the LV following prolonged MICU1 deficiency, potentially promoting the formation of heterooligomeric m-AAA proteases composed of AFG3L2 and mature SPG7. Elevated levels of mature SPG7 indicated an increase in the activity of AFG3L2, which cleaves SPG7 to its mature form. As p-PKA was found to negatively regulate SPG7 maturation⁵¹, the decreased phosphorylation of PKA in Micu1^cKO LV in our current study could lead to enhanced SPG7 maturation, leading to an increase in EMRE degradation. Indeed, we observed an increase in SPG7 maturation in MICU1-deficient LV tissues (Fig. 7A). This pathway provides a potential mechanism underlying the LV-specific adaptation to elevated mtCa²⁺ levels. Supporting this possibility, inhibiting PKA phosphorylation was sufficient to induce EMRE downregulation in freshly isolated neonatal cardiomyocytes (Fig. 7E). PKA is a well-studied excitation-contraction (E–C) coupling regulator in cardiac physiology. Although it is known to modulate cytosolic Ca²⁺ handling in response to physiological demands such as β-adrenergic stimulation⁷⁴, studies have not yet shown a role for PKA in regulating mtCa²⁺ homeostasis. Our findings suggest a novel function for PKA in mediating a compensatory response to elevated mtCa²⁺ levels.

The impact of mtCa²⁺ uptake directly on cytosolic Ca²⁺ dynamics under normal conditions remains controversial^13,75–78. Nonetheless, mitochondrial defects, such as reduced ATP production, excessive ROS generation, and permeability transition, can affect E–C coupling and cardiac contractile function in pathological conditions such as ischemia/reperfusion injury and heart failure^79,7,34,17. Our findings suggest that conditions associated with elevated mtCa²⁺ have ventricle-specific effects on E–C coupling in the cardiomyocyte and ventricular contractile function. Consistent with decreased p-PKA levels, we observed signs of decreases in the phosphorylation of multiple PKA-regulated phosphodiesterases in the LV of MICU1-deficient mice, but to varying degrees. Since some of these proteins are known to show subcellular compartmentalization⁸⁰, this raises the intriguing possibility that elevated mtCa²⁺ leads to altered localized PKA activity. For example, the decreased phosphorylation of phospholamban, an SR protein, was specific to the RV, consistent with the slower decay rates of RV cytosolic Ca²⁺ transients, but other PKA targets tended to decrease in the LV specifically. These observations, hinting at the possibility of ventricle-specific effects on local PKA signaling cascades in different cellular compartments, warrant further investigation.

To the best of our knowledge, no publication has yet reported disrupted mtCa²⁺ levels in the hearts of DCM patients. While direct measurement of mtCa²⁺ levels in patients was not possible, we were able to evaluate PDH phosphorylation, SPG7 maturation, and EMRE levels using frozen LV samples from both DCM patients and healthy individuals. This approach allowed us to investigate whether DCM LV tissues exhibited indirect signs of elevated mtCa²⁺ (such as reduced PDH phosphorylation) and adaptive responses (such as EMRE downregulation). Our data demonstrating a decrease in the p-PDH/total PDH ratio in DCM LV tissues may suggest elevated mtCa²⁺ levels (Fig. 8). In addition, we detected an increase in SPG7 maturation and a reduction in EMRE levels, indicating a compensatory response to reduce mtCa²⁺ uptake. These secondary markers indicate that mtCa²⁺ levels may also be elevated in DCM patients, similar to the phenotype observed in Micu1^cKO mice, though further research is needed. Considering the mouse data, the decrease in EMRE levels in DCM patients could be expected to counteract elevated mtCa²⁺, leading to the restoration of PDH phosphorylation. However, the p-PDH/total PDH ratio was lower in DCM patients compared to healthy controls despite EMRE downregulation (Fig. 8). Two potential explanations are possible: (i) PDH phosphorylation and activity are affected by many other metabolic factors⁸¹, so a decline in PDH phosphorylation may not necessarily correspond to elevated mtCa²⁺ levels in all conditions, or (ii) EMRE degradation alone may not be sufficient to decrease mtCa²⁺ levels in DCM, with potential mtCU-independent factors contributing to mtCa²⁺ dysregulation in this condition. Supporting the latter, we also observed an increase in the levels of the mtCa²⁺ exporter NCLX. This suggests the activation of additional pathways in human heart to reduce mtCa²⁺ levels, likely as a result of persistent mtCa²⁺ abundance. Of note, although it is conceivable that higher NCLX activity could help reverse the increase in calcium levels in the mitochondria, the role of NCLX in exacerbating or protecting against cardiac dysfunction has been controversial^82,83,16,84. Overall, the manifestation in DCM tissues of features including EMRE downregulation and enhanced SPG7 maturation observed in the MICU1-deficient LV supports the possibility that this adaptive mechanism responding to elevated mtCa²⁺ may also be relevant in humans and in different disease states.

Limitations of this study include the lack of access to MRI for further analysis of RV function. Furthermore, the mice used in the study were maintained on a C57BL6/J background, known to harbor a mutation in the Nnt gene⁸⁵, which could affect ROS levels and mitochondrial responses to increased mtCa²⁺ uptake. Future research should address these limitations and examine the impact of elevated mtCa²⁺ levels on heart function in other genetic backgrounds. MICU1 has also been implicated in the regulation of mitochondrial processes independently of mtCU-mediated mtCa²⁺ uptake^86,87. Because elevated matrix Ca²⁺, presumably due to dysregulated mtCU activity in MICU1-deficient conditions, was a primary outcome, the physiological effects due to any disruption of MICU1’s non-mtCU roles were not explored and warrant further study.

In summary, our findings indicate that elevated mtCa²⁺ levels induced in adulthood by MICU1 deficiency impairs the function of both ventricles initially. Unlike the RV, the LV can mount a compensatory response to prolonged mtCa²⁺ elevation, exhibiting improvements in mitochondrial homeostasis and ventricular function over time by reducing mtCU activity through EMRE degradation. Our findings suggest that this LV-specific response may be mediated through decreased PKA activity. The distinct adaptive responses of each ventricle contribute to differences in the recovery of mitochondrial homeostasis, affecting the degree to which ventricular contractility can be restored. Identifying effective strategies to modulate these adaptive pathways may lead to potential clinical targets for the treatment of multiple cardiac diseases, including RV failure, for which few therapeutic options currently exist.

Novelty and significance.

What is known?

• Mitochondrial Ca²⁺ overload is associated with a number of diseases, including neurodegenerative disorders, muscular dystrophies, and heart failure.

• Deletion of MICU1, a component of the mitochondrial Ca²⁺ uniporter, leads to elevated mitochondrial Ca²⁺ in multiple tissues.

• The right and left ventricles of the heart differ developmentally and structurally; more recently, differences in mitochondrial Ca²⁺ handling between the RV and LV were identified

What new information does this article contribute?

• In response to elevated mitochondrial Ca²⁺ induced by MICU1 deficiency in adult mouse hearts, over time the LV can partially recover from contractile dysfunction, whereas the RV remains impaired.

• Elevated mitochondrial Ca²⁺ leads to persistent RV dysfunction independent of pulmonary hypertension, concurrent with increased mitochondrial oxidative stress and a distinct proteomic signature.

• The LV employs a compensatory response to elevated mitochondrial Ca²⁺ levels by reducing EMRE levels through degradation, mediated by a novel PKA-dependent pathway.

Although mitochondrial Ca²⁺ (mtCa²⁺) overload is associated with multiple diseases including heart failure, how cardiomyocytes respond to this stress and recover homeostasis is poorly understood. Moreover, recent data showing mtCa²⁺ handling differences in the left and right ventricles (LV and RV) raise the question of whether the ventricles, which are developmentally and structurally distinct, respond to the detrimental effects of elevated mtCa²⁺ differently. This study shows that after elevated mtCa²⁺ is induced in both LV and RV by deletion of the mtCa²⁺ uniporter subunit Micu1 in adult mice, the time-dependent adaptive response is divergent in the two ventricles. In the LV, initial contractile dysfunction is mitigated over time by m-AAA protease-mediated degradation of the uniporter subunit EMRE. Intriguingly, despite similar initial contractile impairment, this EMRE degradation pathway appears inactive in the RV. Consequently, the RV undergoes sustained elevated mtCa²⁺ contributing to persistent dysfunction, independent of pulmonary hypertension. The data suggest that the LV-specific pathway leading to EMRE degradation is mediated by PKA and may play a role in human dilated cardiomyopathy. By identifying novel adaptive responses to elevated mtCa²⁺ in cardiomyocytes active in the LV but not RV, this study has potential translational implications in LV and RV failure.