MITOCHONDRIAL CARDIOMYOPATHIES FEATURE INCREASED UPTAKE AND DIMINISHED EFFLUX OF MITOCHONDRIAL CALCIUM

Abstract

Calcium (Ca²⁺) influx into the mitochondrial matrix stimulates ATP synthesis. Here, we investigate whether mitochondrial Ca²⁺ transport pathways are altered in the setting of deficient mitochondrial energy synthesis, as increased matrix Ca²⁺ may provide a stimulatory boost. We focused on mitochondrial cardiomyopathies, which feature such dysfunction of oxidative phosphorylation. We study a mouse model where the main transcription factor for mitochondrial DNA (transcription factor A, mitochondrial, Tfam) has been disrupted selectively in cardiomyocytes. By the second postnatal week (10–15 day old mice), these mice have developed a dilated cardiomyopathy associated with impaired oxidative phosphorylation. We find evidence of increased mitochondrial Ca²⁺ during this period using imaging, electrophysiology, and biochemistry. The mitochondrial Ca²⁺ uniporter, the main portal for Ca²⁺ entry, displays enhanced activity, whereas the mitochondrial sodium-calcium (Na⁺-Ca²⁺) exchanger, the main portal for Ca²⁺ efflux, is inhibited. These changes in activity reflect changes in protein expression of the corresponding transporter subunits. While decreased transcription of Nclx, the gene encoding the Na⁺-Ca²⁺ exchanger, explains diminished Na⁺-Ca²⁺ exchange, the mechanism for enhanced uniporter expression appears to be post-transcriptional. Notably, such changes allow cardiac mitochondria from Tfam knockout animals to be far more sensitive to Ca²⁺-induced increases in respiration. In the absence of Ca²⁺, oxygen consumption declines to less than half of control values in these animals, but rebounds to control levels when incubated with Ca²⁺. Thus, we demonstrate a phenotype of enhanced mitochondrial Ca²⁺ in a mitochondrial cardiomyopathy model, and show that such Ca²⁺ accumulation is capable of rescuing deficits in energy synthesis capacity in vitro.

Graphical abstract

1. INTRODUCTION

Heart failure leads to a profound reprogramming of energetic production in cardiomyocytes[1]. As heart failure progresses, energetic supply fails to meet demand, and individuals with more severe supply-demand mismatch have reduced survival[2]. Amongst the pathways altered in heart failure, our interest focuses on calcium (Ca²⁺), which is central for excitation-contraction coupling, and is increasingly recognized as both an enhancer and inhibitor of mitochondrial function. Upon entering the mitochondrial matrix, Ca²⁺ can stimulate ATP production[3], but excess Ca²⁺ entry disrupts mitochondrial function. Because Ca²⁺ has bimodal effects, deciphering its contribution to energetic homeostasis during heart failure has remained difficult. Perhaps such difficulties arise from the relatively late occurrence of clinically-evident energetic depletion during heart failure[2], with mechanisms regulating mitochondrial Ca²⁺ becoming obscured by other homeostatic pathways triggered during heart failure progression. To identify clear examples of altered mitochondrial Ca²⁺ regulation, we sought a disease model that features energetic failure as an early phenotype. We expected that early onset of energetic dysfunction during heart failure would allow us to examine mitochondrial Ca²⁺ regulation in a relatively isolated manner.

A spectrum of diseases featuring energetic imbalance arise from a heterogeneous group of mutations in proteins encoded in either the nuclear or mitochondrial genome (mtDNA) that are involved in oxidative phosphorylation (OXPHOS) [4]. These disorders often feature cardiac involvement, termed the mitochondrial cardiomyopathies, and can be devastating, increasing the mortality rates of children suffering from these disorders three-fold[5]. For our studies, such deficient OXPHOS provides a unique opportunity to examine how the heart modulates mitochondrial Ca²⁺ fluxes when energy production is failing. One such mouse model was created by a cardiac-specific deletion of the main transcription factor for mtDNA (transcription factor A, mitochondrial, Tfam)[6], leading to depletion of electron transport chain (ETC) subunits encoded by mtDNA. In humans, two infants with TFAM mutations died within months of liver failure, had mitochondrial abnormalities on muscle biopsies, and in one case developed clinical heart failure[7].

In mice, cardiac Tfam deletion recapitulates the central problem caused by the heterogeneous group of mutations causing mitochondrial cardiomyopathies: mtDNA depletion and subsequent OXPHOS dysfunction. Multiple studies have revealed that these mice develop cardiac dysfunction with many of the same clinical, biochemical, and ultrastructural features found in human mitochondrial cardiomyopathies, such as an early decline in energetic function and abnormal cytoplasmic Ca²⁺ handling[8–10]. Although these mice have been otherwise characterized, regulation of their cardiac mitochondrial Ca²⁺ remains unexplored.

Studying mitochondrial Ca²⁺ regulation in this model of mitochondrial dysfunction may also provide broader clinical insight, as mitochondrial dysfunction is widespread in heart failure. OXPHOS dysfunction and mtDNA damage are prevalent in dilated cardiomyopathies, and indicate worse prognosis[11–14]. In animal models of ischemic[15], cardiotoxic[16], diabetic[17], tachycardia-induced[18], and pressure-overload[19] cardiomyopathies, reductions in TFAM, mtDNA, and OXPHOS are common. In a proteomic analysis, TFAM was the second-most downregulated protein in failing hearts[20]. Conversely, overexpressing TFAM improves function after cardiac injury[21], and is the subject of therapies targeted towards cardioprotection[22]. However, the late occurrence of clinically evident OXPHOS deficiency has made defining regulatory mechanisms difficult in common forms of heart failure, whereas it drives heart failure in mitochondrial cardiomyopathies.

Here, we study mitochondrial Ca²⁺ transport in the second postnatal week in mice with cardiac-specific Tfam deletion. As OXPHOS declines, these animals develop a dilated cardiomyopathy. In these failing hearts, we find that mitochondria become far more Ca²⁺ avid, taking up Ca²⁺ at twice the rate as controls. This occurs via increased activity of the mitochondrial Ca²⁺ uniporter, the main channel transporting Ca²⁺ from cytoplasm to mitochondrial matrix. In addition, mitochondria from these mice release matrix Ca²⁺ at half the rate as controls, indicating reduced activity of the sodium-calcium (Na⁺-Ca²⁺) exchanger. The increase in Ca²⁺ influx and reduction in Ca²⁺ efflux reflect corresponding changes in protein levels for the mitochondrial Ca²⁺ uniporter and Na⁺-Ca²⁺ exchanger. Though the decrease in Na⁺-Ca²⁺ exchanger levels occurs at the transcript level, the mechanism for increased uniporter function appears to be post-transcriptional, as mRNA levels are reduced despite higher protein amounts. Finally, we find substantial inhibition of mitochondrial respiration in these mice after Ca²⁺ depletion, whereas Ca²⁺ incubation boosts respiration to levels comparable to controls.

2. METHODS

Full description of methods is available in the Supplementary Materials.

2.1 Animal use

The floxed Tfam (a kind gift of Dr. Nils-Göran Larsson) and α-myosin heavy chain-promoter driven Cre recombinase (Myh6-Cre) mice (Jackson Labs, Bar Harbor, ME, strain #011038) have been backcrossed into a C57BL/6J background[6, 23]. All procedures were approved by the Institutional Animal Care and Use Committee in accordance with institutional guidelines. Animals used were 10–15 days old except where noted.

2.2 Mitochondrial isolation

For all assays, knockout and littermate controls were processed in parallel. Mice were euthanized with CO₂ and hearts rapidly removed. A mitochondrial fraction was subsequently isolated by differential centrifugation[24].

2.3 Mitochondrial enzyme assays

We processed 20 μg of mitochondrial fractions using the Citrate Synthase Activity Kit (Sigma, St. Louis, MO). For subsequent assays of complex I–IV activity, we determined the citrate synthase activity in a sample of Tfam heterozygous mitochondria and adjusted amounts of wild-type and knockout littermate mitochondria to match that level. Complex I–IV were analyzed as detailed previously[25], using assay rates in the presence or absence of an inhibitor to determine the complex-specific activity.

2.4 Inductively coupled plasma-mass spectrometry (ICP-MS)

Mitochondria isolated by differential centrifugation were incubated in 40 μM digitonin to permeabilize compartments other than the mitochondrial matrix, prior to final pelleting for analysis. Mitochondria were processed by the ICP-MS metals lab at the University of Utah.

2.5 Fluorescent imaging of mitochondrial inner membrane voltage gradient (ΔΨ) and Ca²⁺ transport

We performed protocols as previously described using 50 μg of the mitochondrial fractions, except for Ca²⁺-induced tetramethylrhodamine methyl ester (TMRM) depolarization, where we used 20 μg[26].

2.6 Whole-mitoplast electrophysiology

We prepared mitoplasts from mitochondrial fractions and subsequently performed voltage-clamp analysis using the Kirichok protocol[24].

2.7 Oxygen consumption analysis

Oxygen consumption was determined with a Clark electrode (Hansatech Instruments, Norfolk, UK), or with the fluorescent oxygen probe MitoXpress Xtra (Luxcel Biosciences, Cork, Ireland). For MitoXpress Xtra, we used fluorescence lifetime imaging of mineral oil-sealed wells in a 96-well microplate [27].

2.8 Statistical analysis

Microsoft Excel and R were used for data analysis. We rejected the null hypothesis for P values less than 0.05. For multiple comparison tests on samples < 15, we established an overall P-value using a Kruskal-Wallis test, and for P < 0.05 proceeded to a Dunn’s multiple comparison test. For multiple comparisons on samples > 15, we established an overall P-value using a one-way ANOVA, and for P < 0.05 proceeded to two-tailed, unequal variance, Student’s t-tests. For multiple comparison tests, final P-values between groups were Bonferroni-corrected.

3. RESULTS

3.1 The dilated cardiomyopathy produced by cardiac-specific deletion of Tfam is evident in the second postnatal week

For our studies, we crossed the previously-developed Tfam loxP mice with mice bearing a Cre recombinase gene under the control of the α-myosin heavy chain promoter (Myh6-Cre), which becomes active during embryogenesis in a cardiomyocyte-specific manner[28]. For our studies, we compared knockout animals (Tfam^loxP/loxP; +/Myh6-Cre, [Tfam KO]) to littermate Tfam heterozygotes (Tfam^loxP/+; +/Myh6-Cre, [Tfam HET]) and wild-type animals (Tfam^loxP/loxP; +/+ or Tfam^loxP/+; +/+. [Tfam WT]).

We conducted our experiments on P10–P15 animals, an age range in which most Tfam KO mice were alive, yet hearts were large enough to yield sufficient mitochondria for functional characterization of individual animals. By this age, Tfam deletion was essentially complete, reflected in the near absence of mRNA and protein in knockout hearts (Fig. 1A–C; numbers are average ± standard error fold changes compared to controls here and throughout: 0.12 ± 0.02 mRNA, 0.05 ± 0.02 protein). The Tfam KO mice experienced early mortality, with animals dying mostly from postnatal week 2 onwards (Fig. 1D). This is consistent with prior reports of cardiac Tfam deletion[6]. In one study using Tfam^loxP/loxP; +/Myh6-Cre animals, >75% of knockout animals died in the first postnatal week[29]. However, those animals were in a mixed background, and the Myh6-Cre transgene construct differed from the one used here.

Figure 1. Myh6-Cre deletes Tfam in cardiac tissue and leads to a dilated cardiomyopathy in 10–15 day old animals.

A. qPCR analysis of Tfam expression, displayed as fold change relative to wild-type animals. Myh6-Cre produces mice that are wild-type (WT), heterozygous (HET), or knockout (KO) for cardiomyocyte Tfam. Here and throughout grey points represent samples from individual mice, summary bars show mean ± standard error, and numbers listed are animals used per genotype, unless otherwise noted. (n = 9–10) B. Exemplar Western blot with each lane showing cardiac TFAM expression (top) and β-actin (loading control, bottom) from a mouse with genotype listed above. C. Summary data showing ratio of TFAM to β-actin chemiluminescence on Western, normalized to WT (n = 12). D. Mouse survival data (n = 26–71). E. Left, photograph of mice on postnatal day 10. Right, hearts extracted from the corresponding animal. Scale, 1 cm. F. Ratio of heart to body weight for individuals of listed genotype (n = 45–86). G. Exemplar transthoracic echocardiograph M-mode recordings in left ventricular short axis of mouse hearts. Calculated fractional shortening (H), left ventricular dimension in diastole (LVID, I), and calculated left ventricular (LV) mass (J) on transthoracic echocardiography (n = 5–6). P < 0.05 (*), <0.01 (**), <0.001 (***), > 0.05 (NS) here and throughout.

All Tfam KO animals assayed had a dilated cardiomyopathy within this period (Fig. 1E–F). Cardiac function on echocardiography was impaired, with severely reduced fractional shortening and dilated left ventricles (Fig. 1G–J). Tfam HET animals, despite having an intermediate reduction in TFAM, had normal cardiac function at this age.

3.2 Tfam KO cardiac mitochondria have impaired OXPHOS activity but preserved numbers and ΔΨ in the second postnatal week

Next, we examined mitochondrial function in Tfam KO animals in this timespan. First, we studied the TFAM-mediated transcription of mtDNA. Quantitative reverse transcription polymerase chain reaction (qPCR) analysis of a subset of mtDNA genes revealed minimal mRNA levels in Tfam KO compared to controls (Fig. 2A). Such disruption was replicated in enzymatic assays of ETC function. We found marked inhibition of complex I, III, and IV but not complex II activity (Fig. 2B–E; complex I: 0.23 ± 0.06, complex II: 1.18 ± 0.12, complex III: 0.72 ± 0.07, complex IV: 0.64 ± 0.07). This pattern reflects the encoding of ETC subunits in nuclear versus mtDNA, as, unlike the other complexes, complex II subunits are wholly nuclear-encoded. Additionally, the absence of heart failure in Tfam HET animals at this age, despite intermediate reductions in ETC activity, is consistent with the significant reserve capacity of cardiac mitochondria. Next, we measured respiration on complex I (glutamate/malate) or complex II substrates (rotenone/succinate), using a Clark electrode (Fig. S1A). Consistent with the qPCR and isolated complex measurements, complex I respiration was reduced in the Tfam KO (Fig. S1B; 0.6±0.1), whereas respiration on complex II was unchanged compared to controls (Fig. S1C; 1.3±0.2). Finally, we assayed ATP production. For the F₁-F_O ATP synthase, beyond the mtDNA-encoded mt-Atp6 and mt-Atp8 genes (Fig. 2A), we quantified the expression of the nuclear-encoded ATP5A1, ATP5F1, ATP5G1/2/3, and ATP5O subunits, and found that several of these were reduced in Tfam KO hearts (Fig. S1D–F, Fig. 5). We then measured ATP production using a luciferin-luciferase assay (Fig. S1G–I), integrating OXPHOS and ATP transport. Using this measure, we found that, as with respiration, Tfam KO mice had a lower rate of ATP production (0.73±0.07). Subsequent to blocking the F₁-F_O-ATP synthase with oligomycin, we found no difference between genotypes. Such deficiencies in electron transport chain activity and ATP synthesis are notable in the development of heart failure, including in the Tfam KO animals [6, 9, 30].

Figure 2. Disruption of Tfam inhibits ETC activity in KO hearts.

A. qPCR analysis of mtDNA gene expression. (n = 5). B. Spectrophotometric assay of complex I activity measured via decline in absorbance at 340 nm (ΔAbs₃₄₀/min) produced by NADH oxidation (n = 5). C. Complex II activity measured via decline in absorbance at 600 nm (ΔAbs₆₀₀/min) produced by 2,6-dichlorophenolindophenol reduction (n = 5). D. Complex III activity measured via increase in absorbance at 550 nm (ΔAbs₅₅₀/min) produced by cytochrome c reduction (n = 5). E. Complex IV activity measured via decrease in absorbance at 550 nm (ΔAbs₅₅₀/min) produced by cytochrome c oxidation (n = 5).

Figure 5. Changes in Ca²⁺ transporter protein expression underlie altered mitochondrial Ca²⁺ flux.

A. qPCR analysis of expression of mitochondrial Ca²⁺ uniporter subunits, other Ca²⁺ handling proteins, including the Na⁺-Ca²⁺ exchanger (Nclx) and Ca²⁺-H⁺ exchanger (Letm1), and other mitochondrial proteins, relative to GAPDH mRNA (n = 5). B. Summary of Western blot quantitation expressed as normalized ratio of protein to GAPDH loading control (n = 8). C. Representative samples from Western blot data. D. For the subset of animals that contributed both protein and mRNA expression data, the graph plots the protein: GAPDH ratio on Western against the relative expression level on qPCR for each animal. Grey circles, data from non-uniporter genes across genotypes, and uniporter genes for WT and HET animals. Black squares, data from core uniporter genes for Tfam KO animals.

Further analyses confirmed that the cardiomyopathy in P10–P15 animals had not produced end-stage tissue remodeling. A prior study demonstrated enhanced oxidative stress and apoptotic cell death, but not extensive fibrosis, in 2–3 week old Tfam KO animals[8]. For oxidative stress, we measured H₂O₂ production using an Amplex Red assay (Fig. S2A–C). These showed a slight increase in ADP-stimulated H₂O₂ production in Tfam KO hearts (1.23±0.03), whereas maximal antimycin A-induced H₂O₂ production was reduced (0.32±0.07), consistent with diminished complex III activity. For cell death (Fig. S2D–E), we saw an increase in terminal deoxynucleotidyl transferase dUTP nick end labeled (TUNEL) nuclei in Tfam KO hearts. Similarly, we found that in 3 of 4 Tfam KO, but no control animals, we could detect cardiomyocytes staining for activated caspase 3 (Fig. S2F). Despite these indicators of cell death, Masson’s trichrome staining did not reveal any evidence of increased fibrosis in this age range (Fig. S2G–H). These results are congruent with the prior report [8].

As dilated cardiomyopathies progress, cardiac mitochondria fission and become more abundant, with increased mitochondrial mass seen in both human disease and animal models[6, 9, 11]. We thus examined mitochondrial abundance in P10–15 Tfam KO mice using several independent methods. First, complex II activity, which does not depend on TFAM, was preserved. Second, Tfam KO mitochondrial ultrastructure under electron microscopy remained comparable to that of controls, as were mitochondrial size distributions and volume density (Fig. 3A–C). Similarly, the mitochondrial yield from each heart was not different between Tfam KO and controls (Fig. 3D). Finally, we found no difference in citrate synthase activity, a marker of mitochondrial mass (Fig. 3E).

Figure 3. Mitochondrial ultrastructure and ΔΨ are preserved.

A. Exemplar transmission electron photomicrographs showing cardiac mitochondrial ultrastructure. Scale, 2 μm. B. Distribution of mitochondrial cross-sectional area in electron microscopy images across genotypes (means: WT, 0.34 ± 0.01 μm²; HET, 0.35 ± 0.01 μm²; KO, 0.31 ± 0.02 μm², for n > 220 mitochondria per condition, with no difference between genotypes). C. Mitochondrial volume density in electron microscopy images (n = 13–24 images per condition). D. Mitochondrial yield measured as ratio of total protein in mitochondrial fraction (μg) to heart weight (mg, n = 19–21). E. Citrate synthase activity measured as increase in absorbance at 412 nm (ΔAbs₄₁₂/min) produced by formation of 5,5′-dithiobis-(2-nitrobenzoic acid) (n = 6–7). F. ΔΨ assessment via FCCP-induced depolarization, measured as the total change in fluorescence in isolated cardiac mitochondria incubated in 20 μM TMRM (ΔTMRM, arbitrary units, A.U., n = 8–10).

We then examined ΔΨ, the voltage gradient driving ATP synthesis. We were unable to enzymatically dissociate high-quality cardiomyocytes for single-cell imaging from animals <P17. In P17 Tfam KO animals, single cardiomyocytes loaded with the voltage-sensitive dye TMRM (20 nM) fluoresced less intensely than controls (0.74 ± 0.02), suggesting an abnormal, depolarized ΔΨ (Fig. S3A–B). However, since the rest of our analyses were performed on mice aged P10–P15, we sought alternative means to test ΔΨ at younger ages. After mechanically disrupting cells, mitochondrial fractions were enriched by differential centrifugation and incubated in TMRM at high concentration (20 μM), which quenched fluorescence. Adding 5 μM carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP) to this suspension uncouples mitochondria, and the total TMRM fluorescence increase associated with dequenching is a very sensitive measure of ΔΨ [31]. Using this technique, Tfam KO mitochondria had ΔΨ indistinguishable from controls at younger ages (Fig. 3F). Together, these data reveal that the P10-P15 period features disease caused by energetic failure, but has not progressed to a late phase characterized by extensive tissue remodeling, impaired mitochondrial ultrastructure, and abnormal enzymatic function.

3.3 Tfam KO cardiac mitochondria feature higher baseline Ca²⁺, increased Ca²⁺ influx, and diminished Ca²⁺ efflux

We next turned to an analysis of mitochondrial Ca²⁺. First, total mitochondrial Ca²⁺ content was measured by applying ICP-MS to mitochondrial pellets[32]. Tfam KO hearts had a 2.2 ± 0.5-fold increase over controls in mitochondrial Ca²⁺ content (Fig. 4A). This elevation was specific to Ca²⁺, as there was no difference in mitochondrial potassium content (Fig. S4A). Next, we measured electrophoretic Ca²⁺ uptake in mitochondrial fractions isolated by differential centrifugation. Mitochondrial suspensions were incubated in a Ca²⁺-sensitive dye restricted to the media[33]. A Ca²⁺ pulse produces a large fluorescence increase that declines as mitochondria take up the Ca²⁺ bolus (Fig. 4B). Notably, with this assay Tfam KO cardiac mitochondria sequestered Ca²⁺ at twice the rate as littermate controls (Fig. 4C, 2.0 ± 0.1), suggesting enhanced activity of the uniporter. Importantly, the total Ca²⁺ pulse amplitude was comparable across trials (Fig. S4B), so this faster rate was due to increased uptake by Tfam KO mitochondria. Next, we examined the predominant efflux pathway for Ca²⁺, the mitochondrial Na⁺-Ca²⁺ exchanger[34]. Following a pulse to load mitochondria with Ca²⁺ in Na⁺-free media, an injection of Ru360 and Na⁺ respectively blocks further transport through the uniporter and activates Ca²⁺ efflux. A final injection of the Na⁺-Ca²⁺ exchange blocker CGP-37157 largely abolishes further efflux, and allows measurement of the Na⁺-Ca²⁺ exchanger-specific rate (Fig. 4D). We found that efflux rates in Tfam KO mitochondria were reduced to 0.3 ± 0.1 of control rates (Fig. 4E).

Figure 4. Mitochondrial Ca²⁺ is increased in Tfam KO hearts.

A. ICP-MS analysis of mitochondrial Ca²⁺ content (n = 4–5). B. Exemplar traces display Ca²⁺ uptake of cardiac mitochondria following a 50 μM pulse (arrow). C. Summary of Ca²⁺ uptake measured as linear slope fit to 15–75 sec after Ca²⁺ pulse (n = 5–10) D. Exemplar traces display protocol for measuring Ca²⁺ efflux rates. Cardiac mitochondria are pulsed with 15 μM Ca²⁺ in Na⁺-free media, followed by addition of 5 mM Na⁺ and 1 μM uniporter blocker (Ru360), followed by addition of a blocker of Na⁺-Ca²⁺ exchange (CGP-37157, 5 μM) as indicated. Inset shows Ca²⁺ efflux traces shifted vertically to start from the same point. E. Summary of Ca²⁺ efflux measured as linear slope fit following Na⁺ addition minus residual slope after CGP-37157 addition (n = 9). F. Exemplar whole-mitoplast current density traces in 100 mM Ca²⁺, with or without ruthenium red (RuR, 0.1 μM). The inwards mitochondrial Ca²⁺ current is indicated (I_MiCa). Voltage ramp protocol is shown above the traces. G. Ca²⁺ current density at −160 mV with or without RuR. H. Summary data shows uniporter-specific Ca²⁺ current density at −160 mV, calculated by subtracting the +RuR from Ca²⁺ values (G, H: n = 12–17 mitoplasts per condition).

Ca²⁺ influx depends not only on the activity of the uniporter but other factors such as the amount of intramitochondrial buffering and the magnitude of ΔΨ driving uptake. It is unlikely that a change in ΔΨ caused the enhanced uptake, as this would require a larger ΔΨ than controls. Yet, we observed no difference at this age (Fig. 3F) and a decline in ΔΨ with disease progression (Fig. S3). Thus, we hypothesized that the increase in uptake was due to the uniporter. To confirm this, we turned to whole-mitoplast electrophysiology, the most accurate method to directly measure uniporter current[24, 35].

To measure uniporter current, we isolated cardiac mitoplasts, which are mitochondria with outer membranes stripped away. After accessing the matrix, ΔΨ was clamped across the entire inner membrane (whole-mitoplast mode). As in prior reports, we found that mouse cardiac mitochondrial currents were minuscule (Fig. 4F)[36]. To isolate the uniporter-mediated current, the total Ca²⁺ current was subtracted from the fraction left after uniporter inhibition with ruthenium red (RuR, Fig. 4G). In fact, the uniporter-mediated current was 5.0 ± 1.0-fold larger in mitochondria from Tfam KO mice compared to controls (Fig. 4H). The slight disparity in magnitude of Ca²⁺ enhancement between this measure and other assays may be due to a few outlier values present in the voltage clamp data. Without these, our results remained significant but the fold change was 3.7 ± 0.5. Thus, direct measurement of uniporter currents confirms that the enhanced mitochondrial Ca²⁺ uptake seen in these animals is mediated by increased transport through the uniporter.

Next, we examined several parameters to establish whether the change in Ca²⁺ uptake was specific or might reflect generalized mitochondrial dysfunction not captured in our prior assays (Fig. 3). First, we measured another transport pathway present in cardiac mitochondria, chloride (Cl⁻) flux through the inner membrane anion channel (IMAC)[37]. This redox-sensitive channel becomes active during periods of stress, depolarizing cardiac mitochondria[38]. Under our recording conditions, IMAC current becomes maximal at depolarized potentials, and its magnitude was unchanged between Tfam KO and control mitoplasts (Fig. S5A–B). In addition, mitoplast sizes, measured indirectly via capacitance values during electrophysiology, were comparable between Tfam KO and controls (Fig. S5C). Next, we examined if Tfam KO animals had higher levels of fibroblast mitochondria in our preparations compared to controls. Though an increased contribution of fibroblast mitochondria may partly explain differences in Ca²⁺ uptake, they are unlikely to be the main reason. Fibroblast numbers are lowest in young hearts, and the mitochondrial mass in cardiomyocytes far exceeds that of fibroblasts, demonstrated by the trivial amount of TFAM left in the KO animals (Fig. 1A–C). Nevertheless, to examine this in more detail, we quantified the levels of a fibroblast marker, vimentin, in Tfam KO versus control hearts, and found no difference (Fig. S5D). Finally, we examined whether the change in Ca²⁺ transport altered sensitivity to permeability transition, measured by pulsing mitochondria repeatedly with 10 μM Ca²⁺ boluses. In this assay, permeability transition becomes visible as a rise in the extra-mitochondrial Ca²⁺ fluorescence, which is blocked by incubating mitochondria with cyclosporin A (Fig. S6A). Tfam KO mitochondria required a similar total Ca²⁺ load to trigger permeability transition compared to controls (Fig. S6B). Thus, the enhanced Ca²⁺ transport seen in Tfam KO hearts appears to be a specific regulatory mechanism active in cardiomyocytes during energetic failure.

3.4 Altered Ca²⁺ transport in Tfam KO cardiac mitochondria is caused by changes in protein expression

Having established that Tfam KO cardiac mitochondria take up Ca²⁺ more avidly and release it more slowly than controls, we sought potential mechanisms for this effect. The most straightforward explanation for altered Ca²⁺ transport rates would be correlated changes in the expression of the uniporter and exchanger. To test this, we analyzed mRNA expression of genes encoding the uniporter, mitochondrial Na⁺-Ca²⁺ exchanger, and other mitochondrial genes (Fig. 5A). The exchanger is encoded by the Nclx gene[39]. The mitochondrial Ca²⁺ uniporter is a multi-subunit channel composed of a main pore-forming subunit (Mcu)[24, 33, 40, 41], core accessory subunits that target the channel and confer Ca²⁺ sensitivity and gating (Micu1–3, Emre, Mcub)[42–45], and other associated proteins that alter mitochondrial Ca²⁺ handling (Mcur1)[26, 46, 47].

Nclx mRNA expression was 0.4 ± 0.1 of control levels (Fig. 5A). Surprisingly, and contrary to the functional data showing enhanced Ca²⁺ uptake, transcripts of subunits forming the mitochondrial Ca²⁺ uniporter were also reduced (range: 0.49–0.65 ± 0.04–0.07). To examine if the qPCR data reflected protein levels, we measured band intensity after Western blotting, a semi-quantitative means to gauge candidate protein expression (Fig. 5B–C). We used the same loading control, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), as in our qPCR analysis. Consistent with the reduced Nclx transcripts, protein levels for the Na⁺-Ca²⁺ exchanger were also reduced in Tfam KO mitochondria (0.65 ± 0.05). For core uniporter subunits, (MCU, MICU1, EMRE, and MCUB), a 1.5–1.9 fold increase in protein levels was detected (range: 1.5–1.9 ± 0.1–0.2). The only uniporter-associated protein showing a decrease, MCUR1, appears to be a regulator rather than a core channel component[26, 46–48]. Taken together, our data suggest that the enhanced Ca²⁺ uptake reflects a post-transcriptional mechanism increasing or stabilizing the amount of uniporter present in mitochondria. This effect can be visualized in Fig. 5D, where we map cardiac protein and qPCR levels for all the genes tested across controls and Tfam KO animals. For all genes except for the uniporter subunits from Tfam KO animals, transcript versus protein levels are directly correlated (gray circles, Fig. 5D). The Tfam KO uniporter subunits (black squares, Fig. 5D) clearly form an outlier group with excess protein expression compared to their transcript levels. Thus, the enhancement in mitochondrial Ca²⁺ appears to result from coincident changes in protein levels of the mitochondrial Ca²⁺ uniporter and Na⁺-Ca²⁺ exchanger.

3.5 Ca²⁺ stimulates respiration in Tfam KO mitochondria more than control

Next, we examined the effects of mitochondrial Ca²⁺ on OXPHOS. Physiological Ca²⁺ concentrations appear not to affect ETC complex I-IV activity directly but can alter OXPHOS in two important ways[3]. Ca²⁺ entry mildly uncouples mitochondria, increasing the stimulus for electron flow through complex I–IV. Ca²⁺ also stimulates pyruvate dehydrogenase phosphatase, several dehydrogenases of the citric acid cycle, and the F₁-F_O-ATP synthase, all ultimately enhancing NADH production and ATP synthesis [3, 49].

We tested these two effects separately. First, we subjected mitochondria incubated in 20 μM TMRM (dequenching mode) to a Ca²⁺ pulse to measure changes in ΔΨ. This pulse should depolarize mitochondria as Ca²⁺ enters the matrix, visible as a rise in TMRM fluorescence with dequenching (Fig. 6A). Subsequently, enhanced ETC function should repolarize ΔΨ, visible as a reduction in fluorescence. We measured these parameters as the maximal deflection in TMRM fluorescence (Fig. 6B), an index for Ca²⁺-induced uncoupling, and the slope of TMRM fluorescence recovery (Fig. 6C), an index for the balance between ETC activity and depolarization associated with continued Ca²⁺ entry. In this test, Ca²⁺ pulses produced greater depolarization in Tfam KO mitochondria (Fig. 6B), reflecting enhanced Ca²⁺ uptake in these samples. Subsequently, Tfam KO mitochondria also repolarized faster than their control counterparts (Fig. 6C). Thus, the direct uncoupling effects of Ca²⁺, though larger in the Tfam KO mice, were largely complete within a few minutes, consistent with prior reports[49].

Figure 6. Increasing Ca²⁺ boosts mitochondrial respiration in Tfam KO hearts more than controls.

A. Exemplar trace of ΔΨ response to a 15 μM Ca²⁺ bolus, measured via TMRM dequenching as in Fig. 3F. B. Maximal change in TMRM signal following Ca²⁺ bolus (n = 8–14). C. Kinetics of ΔΨ measured as linear slope fit to 30–150 sec after Ca²⁺ pulse (n = 8–14). D. Exemplar raw traces of mitochondrial O₂ consumption stimulated by 0.5 μM FCCP, visible as an increase in τ_FL of an O₂-sensing fluorophore. After maximal consumption of O₂ in the sealed chamber, the τ_FL values plateau. E. Peak relative oxygen consumption rates (OCR) stimulated by incubation with 0.5 μM FCCP, calculated from the rising portion of τ_FL curves as in (D) (n = 9–13). F. Exemplar raw traces showing inhibition of O₂ consumption, stimulated by 0.5 mM ADP, in mitochondria incubated with 1 μM oligomycin. G. Oligomycin summary data (n = 5–6). H. Exemplar raw traces of ADP-stimulated mitochondrial O₂ consumption in the presence or absence of Ca²⁺. I. Summary of Ca²⁺ effects on ADP-stimulated OCR (n = 9–13). J. Fold change in OCR stimulated by Ca²⁺ (n = 9–13).

Next, we studied the effect of Ca²⁺ on ATP synthesis. Directly comparing synthesis rates in the presence or absence of Ca²⁺ is difficult[49], as divalent binding alters the kinetics of the luciferase reaction[50]. Therefore, as with our analysis in the absence of Ca²⁺, we normalized ATP synthesis rates to the WT level for each set of samples run in parallel. Remarkably, we found that the deficiency in Tfam KO ATP synthesis noted in Fig. S1G–H was rescued by incubating mitochondria in Ca²⁺ (Fig. S7A–C). ATP synthesis rates in Tfam KO cardiac mitochondria were comparable to WT, despite the larger transient uncoupling noted previously.

To more robustly quantify the stimulation exerted by matrix Ca²⁺ in Tfam KO hearts, we analyzed mitochondrial respiration. To avoid the effects of Ca²⁺ uncoupling, respiration was monitored over a longer period, as ΔΨ largely repolarized within 5 minutes after a single Ca²⁺ pulse (Fig. 6A) [49]. This assay minimizes the direct effects associated with increased Ca²⁺ accumulation in Tfam KO mitochondria, allowing us to focus on the steady-state effects of such uptake. To measure respiration, we incubated mitochondria in media containing an oxygen-sensitive fluorophore (MitoXpress Xtra dye), which allowed assays on smaller mitochondrial quantities (20 μg) for prolonged periods. Oxygen quenching of the dye shortens its fluorescence lifetime (τ_FL), allowing measurement of mitochondrial oxygen consumption in the sealed chamber as an increase in τ_FL. Before studying the effects of Ca²⁺, we performed several quality control assays. Maximal oxygen consumption, induced by uncoupling mitochondria in 0.5 μM FCCP, was indistinguishable between Tfam KO and controls (Fig. 6D–E). Thus, despite compromised ETC activity, cardiac mitochondria from KO animals have substantial reserve capacity. Furthermore, oxygen consumption stimulated by 0.5 mM ADP was mitochondrial, as it ceases with inhibition of the ATP synthase in 1 μM oligomycin (Fig. 6F–G).

Next, we measured ADP-stimulated respiration in mitochondrial suspensions containing either the Ca²⁺ chelator EGTA (1 mM, Ca²⁺ depleted) or 10 μM EGTA + 12 μM Ca²⁺. As expected, for WT or HET cardiac mitochondria, Ca²⁺ enhanced respiration, visible as a faster rise in τ_FL (Fig. 6H). For these controls, Ca²⁺ increased oxygen consumption 1.7 ± 0.1 fold (Fig. 6I–J), consistent with prior reports[49]. Tfam KO mitochondria had a markedly different response. After depletion of mitochondrial Ca²⁺, Tfam KO mitochondria respired at half the control rates (Fig. 6H–I, S5A; 0.45 ± 0.12), and in some instances, was almost absent (Fig. 6I, far right-most exemplar). When incubated with Ca²⁺, however, mitochondrial respiration increased 2.5 ± 0.2 fold, approaching control rates (Fig. 6I–J). Thus, the increased Ca²⁺ in Tfam KO mitochondria produces a robust activation of respiration.

4. DISCUSSION

4.1 Conclusion

We show here that a mouse model of mitochondrial cardiomyopathy features increased mitochondrial Ca²⁺ levels, using multiple independent techniques: analysis of mitochondrial Ca²⁺ content, imaging of Ca²⁺ transport, direct electrophysiological measurement of uniporter currents, and assays of Ca²⁺ transporter protein expression. Such increases can improve deficits in OXPHOS in in vitro assays. Whether such increased matrix Ca²⁺ compensates for deficient energy synthesis or ultimately leads to mitochondrial dysfunction in vivo remains to be established, and will require further studies where the genes responsible for mitochondrial Ca²⁺ transport, such as Mcu, or overload can be selectively deleted in Tfam KO animals.

4.2 Regulation of mitochondrial Ca²⁺ transport in cardiac or skeletal myopathies

Besides describing the Ca²⁺ phenotype present in these animals, we show that the cardiomyopathy features an increase in the mitochondrial Ca²⁺ uniporter. Although several prior studies have shown increased mitochondrial Ca²⁺ content in the setting of disease, typically in muscle or nervous system, such increases have not implicated enhanced uniporter expression. Much of this prior evidence comes from skeletal muscle. The prevalence of mitochondrial Ca²⁺ overload in skeletal muscle disorders may reflect the larger magnitude of uniporter-mediated transport. Uniporter current density in skeletal muscle is 30 times larger than in heart[36], and becomes active at lower cytoplasmic Ca²⁺ levels[51]. Thus, given the already-large flux, skeletal muscle mitochondria may be far more prone to pathological Ca²⁺ overload, and not achieve any benefit by enhancing the flux further. Indeed, in the analysis of skeletal muscle myopathy caused by muscle-specific Tfam deletion[52], mitochondrial Ca²⁺ overload during repetitive stimulation was attributed to the permeability transition pore. Similarly, the tendency to deleterious mitochondrial Ca²⁺ overload in skeletal muscle is seen in malignant hypothermia[32], and myopathies caused by MICU1 mutations[53].

Mitochondrial Ca²⁺ overload also appears to cause cardiac injury in heart failure[54, 55]. However, there has been a lack of evidence for regulation of the uniporter, and investigators have yet to establish if mitochondrial Ca²⁺ transport is regulated in other forms of cardiac energetic failure. In another model of cardiac mitochondrial disease, mice with complex I deficiency have normal lifespan and cardiac function, but develop accelerated heart failure with stress[56]. Mitochondria from these animals were more sensitive to Ca²⁺ overload, but the investigators did not assess whether Ca²⁺ uptake rates changed during heart failure progression. Here, we found no sensitization to Ca²⁺-induced permeability transition in the Tfam KO mice. Nevertheless, Tfam KO may still be more likely to undergo permeability transition, as enhanced baseline Ca²⁺ and Ca²⁺ influx, and diminished Ca²⁺ efflux all may be driving mitochondrial Ca²⁺ closer to activating the permeability transition.

By studying Ca²⁺ transport via whole-mitoplast voltage clamping relatively early in the cardiomyopathy, we bypassed several confounding factors associated with analyses of end-stage failure, such as an increased mass of fissioned mitochondria and reduced ΔΨ. In the late stages of heart failure, enhanced mitochondrial Ca²⁺ uptake may be obscured, especially if such feedback regulation ultimately leads to mitochondrial Ca²⁺ overload and selective injury to those very cardiomyocytes that took up the most Ca²⁺. Indeed, the effect of deleting Mcu variably affects cardiac injury depending on the exact mouse and injury model[57–59]. Moreover, in end-stage failing human hearts, mitochondrial Ca²⁺ uptake appears to be blunted[60]. Given this complexity in mitochondrial Ca²⁺ transport in cardiac injury, mitochondrial cardiomyopathies present a unique opportunity to study regulation of this pathway and these diseases may best respond to therapies targeting mitochondrial Ca²⁺ transport.

4.3 Mechanism regulating mitochondrial Ca²⁺ transport in Tfam KO hearts

The regulation of mitochondrial Ca²⁺ described here is unlikely to be due to direct transcriptional control by TFAM, as it is a transcription factor for mtDNA and not nuclear-encoded genes, such as the uniporter or exchanger subunits. In fact, recent genome-wide studies show that while TFAM coats mitochondrial DNA, it fails to bind nuclear DNA, and is essentially absent in the nucleus[61]. In addition, Tfam HET animals at this age had no observable change in mitochondrial Ca²⁺ transport despite reduced transcription of mtDNA-encoded ETC subunits, effects mediated directly by TFAM. Thus, increased mitochondrial Ca²⁺ in Tfam KO animals is not due to a direct alteration in uniporter or exchanger transcription caused by TFAM, but rather to the heart failure that develops in its absence.

The mechanism for enhanced mitochondrial Ca²⁺ in Tfam KO hearts appears to vary based on the transport pathway. The reduction in Ca²⁺ efflux appears mediated by a reduction in Nclx transcription, which leads to consequent reduction in Na⁺-Ca²⁺ exchanger protein levels. Such a reduction may ultimately predispose to sudden death via mitochondrial Ca²⁺ overload, as seen in other forms of heart failure [62]. The situation for the mitochondrial Ca²⁺ uniporter is more complicated, since we find decreased transcripts but increased protein expression of channel subunits. Such discrepancy between transcript and protein levels can occur for up to 20% of genes assayed in heart failure versus control animals[20]. However, the changes we see here likely reflect a regulatory pathway rather than heart failure epiphenomena, as we found similar magnitudes and concordant changes across all the core uniporter subunits assayed. Potential mechanisms for regulating the uniporter include altering transcript expression[63], changing channel conductance[64, 65], changing channel oligomerization state[66–68], or changing channel sensitivity to cytoplasmic Ca²⁺ [51]. Future studies will need to establish the regulatory mechanism active in Tfam KO hearts and whether it stabilizes the channel or modifies its underlying behavior.

Highlights.

• KO of mitochondrial transcription factor Tfam models mitochondrial cardiomyopathy

• Cardiac mitochondria from Tfam KO mice take up Ca²⁺ at twice the rate as controls

• Cardiac mitochondria from Tfam KO mice release Ca²⁺ at half the rate as controls

• These changes reflect increased MCU complexes and decreased NCLX protein

• Mitochondrial Ca²⁺ maintains respiratory rates despite compromised OXPHOS

ABBREVIATIONS

AA

antimycin A

CSA

cyclosporin A

ETC

Electron transport chain

IMAC

Inner membrane anion channel

OXPHOS

Oxidative phosphorylation