Altered composition of the mitochondrial Ca²⁺uniporter in the failing human heart

Abstract

Heart failure (HF) is a leading cause of hospitalization and mortality worldwide. Yet, there is still limited knowledge on the underlying molecular mechanisms, because human tissue for research is scarce, and data obtained in animal models is not directly applicable to humans. Thus, studies of human heart specimen are of particular relevance. Mitochondrial Ca²⁺ handling is an emerging topic in HF progression because its regulation is central to the energy supply of the heart contractions as well as to avoiding mitochondrial Ca²⁺ overload and the ensuing cell death induction. Notably, animal studies have already linked impaired mitochondrial Ca²⁺ transport to the initiation/progression of HF. Mitochondrial Ca²⁺ uptake is mediated by the Ca²⁺uniporter (mtCU) that consists of the MCU pore under tight control by the Ca²⁺-sensing MICU1 and MICU2. The MICU1/MCU protein ratio has been validated as a predictor of the mitochondrial Ca²⁺ uptake phenotype.

We here determined for the first time the protein composition of the mtCU in the human heart. The two regulators MICU1 and MICU2, were elevated in the failing human heart versus non-failing controls, while the MCU density was unchanged. Furthermore, the MICU1/MCU ratio was significantly elevated in the failing human hearts, suggesting altered gating of the MCU by MICU1 and MICU2. Based on a small cohort of patients, the decrease in the cardiac contractile function (ejection fraction) seems to correlate with the increase in MICU1/MCU ratio. Our findings therefore indicate a possible role for adaptive/maladaptive changes in the mtCU composition in the initiation/progression of human HF in humans and point to a potential therapeutic target at the level of the MICU1-dependent regulation of the mtCU.

1. Introduction

Impaired mitochondrial Ca²⁺ exchange is linked to heart failure (HF) development and progression. Sarcoplasmic Ca²⁺ oscillations are crucial to cardiomyocyte contraction, and their propagation to the mitochondrial matrix regulates Krebs cycle flux to match ATP production to the contractile workload of the heart. However, excessive mitochondrial Ca²⁺ uptake can also cause mitochondrial dysfunction, permeability transition and cell death. Thus, mitochondrial Ca²⁺uptake across the inner mitochondrial membrane needs to be tightly controlled. Mitochondrial Ca²⁺uptake is mediated by the mitochondrial Ca²⁺ uniporter channel (mtCU), which consists of tetramers of the pore-forming subunit, MCU, its essential scaffold, EMRE and the Ca²⁺-sensing regulator, MICU1 that dimerizes with MICU2, MICU3 or itself.

MCU-knockout mice did not have an overt cardiac phenotype [1], but inducible cardiac deletion of MCU led to impaired fight-to-flight response and protected against ischemia-reperfusion injury [2,3]. Functional studies revealed that the regulation of the mtCU by MICUs is tissue-specific, and in mouse heart, MICU1 is expressed in limiting quantity [4]. In mice, MICU1 knockdown by siRNA was shown to exacerbate cardiac ischemia-reperfusion injury [5], suggesting a protective role for MICU1-dependent gating of the mtCU against mitochondrial Ca²⁺ overload. On the other hand, cardiac MICU1 overexpression triggered contractile dysfunction, suggesting a shift in the MICU1/MCU ratio might interfere with decoding of cardiomyocyte intracellular Ca²⁺ oscillations by the mitochondria [4]. In a guinea pig model of severe pressure-overload, a small increase in MCU expression was reported to have a beneficial effect by restoring the mitochondrial Ca²⁺ content and improving left ventricular function [6]. The MCU paralog, MCUb, has also been reported to be altered in animal models of myocardial infarction [7] or catecholaminergic polymorphic ventricular tachycardia [8]; suggesting yet another mechanism of direct regulation of mtCU structure and function.

The human heart operates at a much lower frequency range than the rodent heart and thus, MICU1 might have a more significant physiological role in decoding Ca²⁺ spikes and contribute differently to HF pathogenesis in humans. Other reports showed that mRNA is elevated for both MICU1 and MICU2 in failing human hearts [9,10]. Whether these mRNA changes are translated into differences in protein expression remains to be determined. Here, our objective was to determine the protein composition of the mtCU in the human failing heart.

2. Materials and methods

Patient consent, clinical data collection, sample collection and preparation were all performed according to a protocol approved by the Lewis Katz School of Medicine Institutional Review Board. Definition of HF etiology was based on the patient diagnosis as reported in the ICD codes. In general terms, non-ischemic HF suggests some suspected causes such as dilated cardiomyopathy, hypertension or something else, yet no ischemic event, while idiopathic suggests unknown cause. Resting transthoracic echocardiography was performed to calculate the left ventricular ejection fraction (LVEF) and data were acquired post-hoc from patient charts. Left ventricle biopsies were obtained after cardiectomy at the time of transplant from 7 non-failing patients (from unused donor hearts) and 15 HF patients (3 non-ischemic, 7 ischemic and 5 with idiopathic HF). The mean age of the HF patients was 54 [48;61] versus 60 [58;72] years in non-failing patients (p < 0.05, Mann-Whitney test, median [IQ25; IQ75]). 86% of non-failing patients were female versus 20% in the HF cohort.

Immunoblotting was performed on RIPA-lysed frozen cardiac tissue, using the following antibodies: MICU1 (Sigma, HPA037480), MCU (Sigma, HPA016480), MICU2 (Abcam, ab101465), EMRE (SantaCruz, sc-86,337), MCUB (sc-163,985), mitochondrial Hsp70 (ThermoFisher, MA3–028) and OxPhos cocktail (ab110413). Specificity of the antibodies was confirmed using lysates of HEK293 human cell lines deleted for each mtCU component (Supplementary Figure 1) [11-13].

Statistical analysis was performed on GraphPad and on Python, Scikit-learn.

3. Results and discussion

The protein level of MCU and EMRE were indistinguishable between non-failing and failing hearts (Fig 1A-C). No change was observed for MCUB, when corrected to complex I as a mitochondrial loading control (Supplementary Figure 2). By contrast, MICU1 and MICU2 protein levels were increased ~two-fold in cardiac samples isolated from failing hearts (Fig. 1D,E). Calculation of the expression ratio between the different mtCU components for each patient revealed a two-fold increase in the MICU1/MCU ratio in failing hearts, as compared to non-failing controls (Fig. 1F), while the MICU2/MCU and EMRE/MCU ratios were not significantly altered (data not shown). Multivariate analysis by principal component analysis (PCA) revealed a significant divergence between the non-failing and HF patients (p < 0.001, Mann-Whitney Rank Sum Test), as displayed by their scattering along the first PC (70% of variance) calculated as 0.85*MICU1 + 0.50*MICU2 + 0.16*MCU - 0.017*EMRE. This indicates that the rightward samples, mainly the failing heart biopsies, are associated with high MICU1, MICU2 and to a lesser extent, MCU (Fig. 1G). Pairwise comparison further deciphered a significant difference between idiopathic and ischemic biopsies (p = 0.026). Interestingly, the range/variance of protein expression for MICU1 and MICU2 was also greater in the HF group, which may reflect differences in the clinical status of each donor patient. Reliable clinical data were scarce for the present cohort, but the LVEF was available for 5 non-failing and 10 HF patients: the MICU1/MCU ratio increase correlated with LVEF decrease in the dataset (Fig. 1H, Spearman Rank Order Correlation: −0.699, p = 0.00326).

Fig. 1.

Evaluation of mtCU subunit protein composition in failing human hearts. A) Representative immunoblots for various mtCU components. B–E) Quantification of the relative protein level of MCU (B), EMRE (C), MICU2 (D) and MICU1 (E) in heart lysates isolated from 7 non-failing and 15 HF patients, normalized to the mitochondrial housekeeping protein, Hsp70.

F) Protein ratios of MICU1 to MCU, calculated individually for each patient. Statistical analysis was performed on GraphPad (Mann-Whitney test).

G) Principal component analysis (on Python, Scikit-learn) shows the first and second principle components of the multi-variate scattering of all patients by the variance of the mtCU protein level.

H) Dot plot between LVEF and MICU1/MCU ratio in non-failing and failing patients.

These results provide evidence that the overall abundance of the mtCU (uniporter channel density) is unchanged, whereas the channel composition or regulatory subunit stoichiometry is altered in the failing human heart, signified by an increase in the MICU1 to MCU ratio. We propose that this molecular change leads to increased gating of the mtCU by MICU, which might serve as a compensatory mechanism to prevent mitochondrial Ca²⁺ overload in the stressed/diseased heart. However, a chronic increase in the MICU1/MCU ratio may turn maladaptive leading to an altered decoding of intracellular Ca²⁺ oscillations by mitochondria, therefore contributing to impairments in energy metabolism and contractile dysfunction. Although admittedly our study included a small number of patients, the results revealed a correlation between the MICU1/MCU ratio and decrease in cardiac contractility. This observation will stimulate in-depth studies of the clinical phenotype to test the correlation between changes in the mtCU composition and clinical endpoints. In summary, these results suggest that MICU1-dependent regulation of the mtCU may be a potential therapeutic target in HF. To test this, further studies are required on fresh cardiac biopsies from HF patients to perform functional analysis of the mtCU in correlation with clinical outcomes to decipher the contribution of the mtCU in HF with reduced or preserved ejection fraction.

Data Availability

No data was used for the research described in the article.