MICU1 Alleviates Diabetic Cardiomyopathy Through Mitochondrial Ca²⁺–Dependent Antioxidant Response

Increased prevalence of type 2 diabetes is a major threat to human health. It is estimated that diabetes will affect nearly 366 million adults by the year 2030 (1). Cumulative evidence reveals that diabetes is associated with structural and functional abnormalities of the myocardium, leading to diabetic cardiomyopathy. However, specific strategies for the prevention and treatment of cardiomyopathy in patients with diabetes have not been clarified yet. Emerging evidence suggests that mitochondrial dysfunction may play a critical role in the pathogenesis of diabetic cardiomyopathy (2). For example, recent data have demonstrated that mitochondria isolated from atrial cardiomyocytes of patients with diabetes exhibit remarkable respiratory defects and increased generation of reactive oxygen species (ROS) when compared with those from control subjects without diabetes (3,4). Moreover, it is essential to protect mitochondria from losing their ability to generate energy and to control their own ROS emission in order to prevent diabetic cardiomyopathy (5). However, despite increasing data implicating mitochondrial pathology in diabetic cardiomyopathy, the mechanism underlying these processes still remains largely unknown.

Mitochondrial Ca²⁺ plays important roles in regulating the intrinsic functions of the organelle. One of the best-characterized functions of mitochondrial Ca²⁺ is the control of organelle metabolic activity. For example, physiological increases in mitochondrial Ca²⁺ lead to the allosteric activation of tricarboxylic acid (TCA) cycle enzymes including isocitrate dehydrogenase, α-ketogluterate dehydrogenase (α-KGDH), and pyruvate dehydrogenase (PDH). The net effect of TCA cycle activation is a boost in reduced oxidative phosphorylation substrate synthesis (NADH and flavin adenine dinucleotide), enhanced respiratory chain activity, and a subsequent increase in ATP synthesis (6). Another key role of mitochondrial Ca²⁺ is to modulate oxidative stress by misbalancing ROS generation and ROS detoxification. It has been reported that mitochondrial Ca²⁺ regulates cellular antioxidant defense systems by stimulating NADPH regeneration, which donates electrons to regenerate reduced glutathione (GSH) from oxidized glutathione (GSSG) (7,8). Interestingly, several studies have reported that reduced mitochondrial calcium uptake is observed in the hearts of diabetic rodents (9,10). However, the mechanisms by which diabetes impairs mitochondrial calcium handling in cardiomyocytes remain to be defined.

Since 2010, the discovery of the pore-forming subunit of the mitochondrial Ca²⁺ uptake channel (mitochondrial calcium uniporter [MCU]) (11,12) and its regulatory subunits, called mitochondrial calcium uptake 1 (MICU1) (13) and MCU regulator 1 (14), has opened new avenues for the study of mitochondrial Ca²⁺ homeostasis regulation and its key roles. In particular, a loss-of-function mutation in MICU1 has been reported to cause human disease through alterations in mitochondrial Ca²⁺ handling (15). A previous study reported that MICU1, but not MCU, mRNA levels were markedly downregulated in human cardiovascular disease–derived primary endothelial cells (16). To our knowledge, however, the significance of MICU1 in diabetic hearts, especially in the pathogenesis of diabetic cardiomyopathy, has never been reported.

In this study we identified downregulated MICU1 as a factor contributing to cardiomyocyte apoptosis in diabetic cardiomyopathy. Moreover, MICU1 overexpression effectively alleviated diabetic cardiomyopathy by promoting mitochondrial Ca²⁺ uptake to inhibit mitochondrial ROS (MitoROS)–mediated apoptosis. These findings suggest that upregulating MICU1 expression may be a potential therapeutic strategy in diabetic cardiomyopathy.

All experiments were performed in accordance with the National Institutes of Health Guidelines on the Use of Laboratory Animals and approved by the Fourth Military Medical University Committee on Animal Care. Leptin receptor–deficient (db/db) C57BLKS mice and wild-type (WT) C57BLKS mice were purchased from Changzhou Cavens Laboratory Animal Co. Ltd. (Jiangsu, China). The investigators were blinded to the treatment conditions during experiments.

Primary cardiomyocytes were prepared from neonatal rats, as previously described (17). For the high-glucose and high-fat (HGHF) treatments, cells were cultured with DMEM (containing 25 mmol/L glucose and 500 μmol/L of the saturated free fatty acid palmitate [16:0]) for 24 h.

MICU1-specific small interfering RNA (siRNA), Sp1-specific siRNA, and scrambled siRNA were designed and synthesized by GenePharma Company (Shanghai, China). The sequences of siRNAs are provided in Supplementary Table 1. siRNAs were transfected as previously described (17). Recombinant adenoviruses overexpressing rat MICU1 (NM_199412.1) and Sp1 (NM_012655.2) were constructed by Shanghai Genechem Co., Ltd (Shanghai, China) and HanBio Biotechnology Co., Ltd. (Shanghai, China), respectively. Recombinant adenovirus-overexpressing mouse MICU1 (NM_001291442.1) or mitochondrial-targeted parvalbumin was generated following the instructions of the ViraPower Adenoviral Expression System (Life Technologies), according to the manufacturer’s protocol. The coding sequences of MICU1 and mitochondria-targeted parvalbumin were amplified using the primers listed in Supplementary Table 2. The viral titer was determined using an Adeno-X Rapid Titer Kit (Clontech Laboratories, Mountain View, CA). Cardiomyocytes were infected with adenovirus, at a multiplicity of infection of 50, for 2 h. Subsequently, cells were cultured in serum-free DMEM for an additional 24 h and then used for further analysis. For adenovirus administration, the mice were anesthetized with 2% isoflurane, and the heart was exposed. Adenovirus (2 × 10¹⁰ infectious units/mL) was injected (using a 30-gauge needle) into the free wall of the left ventricle (LV) (10 µL at each of three sites). The transfection efficiency was evaluated using Western blotting 3 days after adenoviral injection.

RNA extraction, cDNA synthesis, quantitative RT-PCR reactions were performed as described previously (18). Primer sequences are provided in Supplementary Table 3. Mouse heart tissue or primary neonatal cardiomyocytes were processed for Western blotting, as previously described (19). Immunohistochemical staining of cardiac sections was carried out as previously described (20).

The hearts of mice were fixed in 4% paraformaldehyde (pH 7.4) overnight, embedded in paraffin, and serially sectioned (5-μm-thick slices) for histological analysis. Standard hematoxylin and eosin staining was carried out following standard procedures. Cardiac collagen content was assessed using Masson trichrome staining. Myocyte size was detected by staining with wheat germ agglutinin (red; Sigma-Aldrich).

Cell apoptosis was determined using a PE Annexin V Apoptosis Detection Kit (material no. 559763; BD Biosciences) or a Fluorescein Isothiocyanate Annexin V Apoptosis Detection Kit (BB-4101; BestBio), following the manufacturer’s instructions. For analysis of apoptosis in heart tissues, terminal deoxynucleotidyl TUNEL assay (11684795910; Roche Applied Science) was performed according to the manufacturer’s protocol. Images of TUNEL and DAPI-stained sections were obtained using a fluorescence microscope (DM5000B; Leica, Heerbrugg, Switzerland). Only TUNEL- and DAPI-positive nuclei that were located within heart tissues were counted as apoptotic nuclei.

To measure mitochondrial membrane potential, cells (1 × 10⁶/mL) were incubated with tetramethyl rhodamine methyl ester (TMRM; 10 nmol/L; Invitrogen) at 37°C for 30 min. TMRM images were captured using a confocal laser scanning microscope (FluoView FV1000; Olympus, Japan), with excitation at 530 nm and emission at 573 nm.

Mitochondria were isolated from cardiomyocytes or hearts using the Mitochondria Isolation Kit (Beyotime, Shanghai, China), according to the manufacturer’s instructions.

Mouse hearts were excised and then washed in Ca²⁺-free tyrode solution, followed by perfusion with buffer containing liberase (Roche Diagnostics) at 37°C for 20 min. The LV was subsequently dissociated into single myocytes, and extracellular Ca²⁺ was added back incrementally to 1.2 mmol/L. Cardiomyocytes were placed on glass coverslips coated with laminin (Thermo Fisher Scientific). Cells were used within 4 h after isolation.

The fluorescent dye Rhod-2/AM (Invitrogen) was used to monitor mitochondrial Ca²⁺ concentration ([Ca²⁺]) in cardiomyocytes, according to the manufacturer’s instructions. Then, cardiomyocytes were viewed with a confocal laser scanning microscope (FluoView FV1000; Olympus). For the measurement of mitochondrial Ca²⁺ uptake, adult or neonatal cardiomyocytes were stimulated by pacing (3 Hz) or histamine (10 μmol/L), respectively. Images were recorded every 5 s using the same confocal imaging system. The endogenous mitochondrial Ca²⁺ content was measured in isolated cardiac mitochondria. Briefly, isolated mitochondria were resuspended in buffers, sonicated, and then centrifuged. The supernatants were recovered and the Ca²⁺ content in the supernatants was determined using a calcium detection kit (Abcam).

MitoROS were detected using the fluorescent probe MitoSOX (Invitrogen), according to the manufacturer’s protocols. Images were captured by a laser confocal microscope (FluoView FV1000; Olympus) and ImagePro image analysis software.

The activity of PDH and α-KGDH was measured using the PDH enzyme activity microplate assay kit (Abcam) and α-KGDH activity assay kit (Biovision), according to the manufacturers’ instructions.

NADH and NADPH concentrations were measured in lysates from isolated mitochondria using NAD/NADH and NADP/NADPH assay kits (Abcam), according to the manufacturer’s instructions. The GSH-to-GSSG ratio was determined in lysates from isolated mitochondria using a GSH and GSSG assay kit (Beyotime), according to the manufacturer’s instructions.

Chromatin immunoprecipitation (ChIP) assay was performed using the SimpleChIP Plus Enzymatic Chromatin IP kit (Cell Signaling Technology, Danvers, MA), according to the manufacturer’s instructions. Briefly, after cross-linking, tissues were disaggregated into a single-cell suspension. Cells were then lysed and the chromatin was fragmented. The sheared chromatins were incubated with rabbit anti-Sp1 antibody and protein G magnetic beads. The DNA released from the precipitation was subjected to PCR analysis for detection of the presence of specific loci. The primers specific to the Sp1 binding region in the mouse MICU1 gene promoter are shown in Supplementary Table 4. Rabbit IgG and antibody against histone H3 were used as the negative and positive controls, respectively.

Noninvasive serial ultrasound examination of hearts was performed using a Vevo 2100 Microultrasound System (VisualSonics Inc., Toronto, Ontario, Canada). Mice were lightly anesthetized with isoflurane (3% for induction and 1% for maintenance). Cardiac function under physiological conditions was cursorily examined by hand-held manipulation of the ultrasound transducer to obtain two-dimensional and two-dimensional guided M-mode images. All the echocardiographic images were analyzed using Vevo 2100 software.

Invasive hemodynamic measurements in anesthetized mice were performed using a 1.4 French micromanometer (Millar Instruments, Houston, TX) inserted into the right carotid artery and guided into the LV. Heart rate, LV systolic pressure, and LV end-diastolic pressure were measured by this catheter, and data were recorded and analyzed using a PowerLab System (ADInstruments Inc., Colorado Springs, CO). These parameters, as well as maximal values of the instantaneous first derivative of LV pressure (dP/dt_max, the maximum changing speed of intraventricular pressure, as a measure of cardiac contractility) and minimum values of the instantaneous first derivative of LV pressure (dP/dt_min, the minimum changing speed of intraventricular pressure, as a measure of cardiac relaxation), were recorded.

All data were presented as the mean ± SEM of n independent experiments. For groups of three or more, the data were subjected to ANOVA, followed by a Bonferroni correction for a post hoc test. The unpaired t test was used for comparison between two groups. Correlations between measured variables were tested using Pearson correlation analysis. Probability ≤0.05 was considered statistically significant.

To investigate the alterations of the key proteins underlying mitochondrial Ca²⁺ entry into diabetic hearts, the expressions of MCU and its regulator MICU1 were determined. Western blotting (Fig. 1A) showed that MICU1 expression was significantly downregulated in db/db mouse hearts at 12 weeks of age compared with hearts of WT littermate control mice. The downregulation of MICU1 was even more obvious in hearts of 18-week-old db/db mice. By contrast, MCU expression was unchanged. A consistent pattern of mRNA expression was also observed using quantitative RT-PCR (Fig. 1B), suggesting that the expression of MICU1 in diabetic hearts is mainly regulated at the transcriptional level. Furthermore, immunohistochemical staining also demonstrated that cardiac MICU1 expression remained unchanged in wild-type mice at different ages; however, MICU1 was significantly downregulated in hearts of db/db mice at both 12 and 18 weeks of age when compared with their corresponding controls (Fig. 1C). Collectively, these findings indicate that MICU1 downregulation occurs in diabetic cardiomyopathy.

We attempted to determine whether MICU1 downregulation leads to cell apoptosis in neonatal cardiomyocytes with or without HGHF treatment. Our data indicate that the HGHF treatment resulted in significant MICU1 downregulation in cardiomyocytes, and MICU1 siRNA significantly reduced the expression of MICU1 (Fig. 2A). Importantly, the percentages of total (both early and late) apoptotic cells were significantly higher in neonatal cardiomyocytes with MICU1 knockdown or HGHF treatment than those in control cells. Furthermore, MICU1 knockdown remarkably exacerbated HGHF-induced apoptosis (Fig. 2B). Consistently, the cleavage of caspase-9 and caspase-3 (Fig. 2C), release of cytochrome c from mitochondria (Fig. 2D), and disruption of the mitochondrial membrane potential (Fig. 2E) were significantly induced by HGHF treatment, whereas all of these effects were further aggravated upon MICU1 knockdown. These findings suggest that downregulation of MICU1 induces mitochondria-dependent intrinsic apoptosis in cardiomyocytes.

We further attempted to determine whether the upregulation of MICU1 protects the heart against diabetic cardiomyopathy in db/db mice. Adenovirus encoding MICU1 was intramyocardially injected into 12-week-old db/db mice. MICU1 expression was determined 3 days after injection, and cardiac function was measured 4 weeks after injection. As shown in Fig. 3A, the infection of MICU1-expressing adenovirus resulted in an ∼1.6-fold increase in cardiac MICU1 expression compared with controls. Representative M-mode echocardiograms are shown in Fig. 3B. LV ejection fraction and LV fractional shortening were impaired in db/db mice when compared with WT mice. Interestingly, upregulated MICU1 significantly augmented LV ejection fraction and LV fractional shortening (Fig. 3C and D). Moreover, analysis with the Millar transducer showed no significant differences in heart rate and mean arterial blood pressure among all groups. However, there was a significant increase in LV end-diastolic pressure and a decrease in LV end-systolic pressure (dP/dt_max and dP/dt_min in diabetic mice), whereas the forced expression of MICU1 resulted in a global improvement of myocardial function (Fig. 3E–J). Taken together, these data indicate that MICU1 restores LV cardiac function in diabetic db/db mice.

LV hypertrophy is a common feature of cardiomyopathy. Four weeks after intramyocardial injection of adenovirus, LV hypertrophy was measured in 16-week-old db/db mice. Our data showed that db/db mice displayed symptoms of cardiac hypertrophy, as evidenced by increased heart weight, enlarged hearts, and elevated ratio of heart weight to tibia length, as well as augmented LV mass and enlarged LV inner diameter, with reduced posterior wall thickness. By contrast, the injection of MICU1-expressing adenovirus significantly attenuated these pathological changes (Fig. 4A–C and Supplementary Fig. 1). Moreover, the injection of MICU1-expressing adenovirus remarkably reduced interstitial fibrosis in the db/db hearts (Fig. 4D). In addition, the up-regulation of MICU1 decreased the cross-sectional area of cardiomyocytes in db/db hearts (Fig. 4E) and reversed the mRNA expression changes of hypertrophic genes, including the upregulation of α–myosin heavy chain and downregulation of β–myosin heavy chain, atrial natriuretic peptide, and brain natriuretic peptide (Fig. 4F).

TUNEL staining analysis showed that the forced expression of MICU1 significantly reduced the number of apoptotic cells in db/db hearts (Fig. 5A). Consistently, upregulated MICU1 remarkably decreased the cleavage of caspase-9 and caspase-3 in db/db hearts (Fig. 5B). In vitro study further confirmed that the forced expression of MICU1 remarkably inhibited cell apoptosis in HGHF-treated neonatal cardiomyocytes, as evidenced by decreased percentages of total (both early and late) apoptotic cells (Fig. 5C and Supplementary Fig. 1A), and reduced the cleavage of caspase-9 and caspase-3 (Fig. 5D). Moreover, MICU1 overexpression in cardiomyocytes significantly decreased the release of cytochrome c from mitochondria (Fig. 5E) and elevated the mitochondrial membrane potential (Fig. 5F and Supplementary Fig. 2B), indicating that MICU1 inhibits mitochondria-dependent intrinsic apoptosis in cardiomyocytes.

Previous studies have demonstrated that hearts of db/db mice exhibited excessive ROS production, which is associated with the induction of cell death (21). Our data indicated that the intramyocardial injection of MICU1-expressing adenovirus significantly reduced mitochondrial ROS production in hearts of db/db mice (Fig. 5G). Similarly, the inhibitory effect of MICU1 overexpression on MitoROS production was clearly observed in HGHF-treated neonatal cardiomyocytes in vitro (Fig. 5H). Furthermore, flow cytometry analysis showed that MICU1 overexpression remarkably blocked H₂O₂-induced apoptosis in cardiomyocytes with or without HGHF treatment (Fig. 5I and Supplementary Fig. 2C). Collectively, all our findings indicate that upregulated MICU1 inhibits cell apoptosis through reducing MitoROS in diabetic cardiomyocytes.

Since MICU1 has been identified as a critical regulator of mitochondrial Ca²⁺ uniporter, we next determined whether mitochondrial Ca²⁺ played an important role in MICU1-mediated antiapoptosis in diabetic cardiomyopathy. Endogenous mitochondrial Ca²⁺ content was measured in isolated cardiac mitochondria. As shown in Fig. 6A, MICU1 overexpression significantly increased the basal mitochondrial Ca²⁺ in db/db hearts. Consistently, the forced expression of MICU1 significantly enhanced the capacity of mitochondrial Ca²⁺ uptake following pacing (Fig. 6B). More importantly, our data demonstrated that mitochondrial Ca²⁺ buffering by mitoPV (a mitochondria-targeted Ca²⁺ binding protein parvalbumin) remarkably blocked the antioxidant effect of MICU1 overexpression in HGHF-treated cardiomyocytes (Fig. 6C and Supplementary Fig. 3). Furthermore, mitoPV treatment decreased cytochrome c release from mitochondria (Fig. 6D) and inhibited the cleavage of caspase-9 and caspase-3 (Fig. 6E). All these findings suggest that MICU1 overexpression inhibits ROS-mediated apoptosis in HGHF-treated cardiomyocytes by increasing mitochondrial Ca²⁺ uptake.

Regeneration of NADPH, a key component in cellular antioxidant systems, requires the products of the Krebs cycle, including isocitrate, malate, and NADH (22). Since the mitochondrial Ca²⁺–activated dehydrogenases of the TCA cycle are known to accelerate the regeneration of NAD(P)H (7,23), it is reasonable to hypothesize that MICU1-mediated mitochondrial Ca²⁺ uptake enhances the antioxidant system in cardiomyocytes (Fig. 7A). Therefore, we first investigated the enzyme activities of Ca²⁺-dependent dehydrogenases in neonatal cardiomyocytes. Our data indicated that HGHF treatment inhibited the enzyme activities of PDH and α-KGDH in neonatal cardiomyocytes, whereas the forced expression of MICU1 partially reversed the inhibitory effect. Furthermore, mitochondrial Ca²⁺ buffering with mitoPV remarkably blocked the effects of MICU1 (Fig. 7B and C). Then, we found that HGHF treatment reduced NADH and NADPH concentrations, whereas MICU1 overexpression remarkably reversed these effects in HGHF-treated neonatal cardiomyocytes. Similarly, mitoPV blocked MICU1-enhanced NADH and NADPH generation (Fig. 7D and E). In addition, our data showed that HGHF treatment significantly reduced the ratio of mitochondrial GSH to GSSG (an indicator of intracellular redox status), whereas MICU overexpression exhibited an opposite effect, which was significantly reversed through mitochondrial Ca²⁺ buffering by mitoPV (Fig. 7F). Collectively, these findings suggest that MICU1 overexpression and subsequent elevation of mitochondrial Ca²⁺ uptake inhibit ROS production by enhancing the NADPH-dependent antioxidant system.

We finally investigated the mechanism underlying the downregulation of MICU1 in diabetic hearts. Potential transcription factor Sp1 for MICU1 was predicted by both the Promoter 2.0 and Promoter Scan software (data not shown). Moreover, several putative Sp1 binding sites, which are conserved across multiple species, were identified in the 5′ flanking region of human MICU1 by the Jaspar program (http://jaspar.genereg.net) (Fig. 8A and Supplementary Fig. 4). We used a public microarray data set of mRNA expression (GSE26887), downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE26887), to evaluate the correlation between the mRNA expression levels of MICU1 and Sp1 in human hearts. Pearson correlation analysis indicated a significant positive correlation between the mRNA expression level of Sp1 (GSE26887 gene ID, 7955787) and MICU1 (GSE26887 gene ID, 7934255) in a panel of human heart tissues (r = 0.6003; P = 0.002). This correlation will be even more significant when we analyze only the heart tissues of patients with diabetes and heart failure from the same data set (r = 0.7923; P = 0.03) (Fig. 8B). Consistently, both mRNA and protein levels of Sp1 were significantly downregulated in db/db mouse hearts at 12 weeks of age when compared with age-matched WT mouse hearts. The downregulation of Sp1 was more obvious in hearts of 18-week-old db/db mice (Fig. 8C and D). ChIP assay demonstrated that Sp1 bound directly to the promoter region of the MICU1 gene in mouse heart tissues (Fig. 8E). Moreover, Western blotting revealed that Sp1 knockdown significantly decreased the expression of MICU1 and exacerbated the downregulation of MICU1 induced by HGHF treatment. By contrast, Sp1 overexpression exhibited the opposite effect (Fig. 8D and F). Taken together, these results demonstrate that MICU1 downregulation in diabetic cardiomyocytes is due, at least in part, to the decreased expression of transcriptional factor Sp1.

Accumulating data from experimental and clinical studies have shown that diabetes results in functional and structural changes of the myocardium (24). However, the molecular pathological mechanism of diabetic cardiomyopathy is far from clear. Here, we have made several important observations. First, we validated that MICU1 was downregulated in hearts of diabetic mice, which contributes to myocardial apoptosis in diabetes. Second, the forced expression of MICU1 inhibited the development of diabetic cardiomyopathy. Third, the antiapoptotic effect of MICU1 was mediated by promoting mitochondrial Ca²⁺ uptake and subsequently activating the antioxidant system in HGHF-treated cardiomyocytes. Last, MICU1 downregulation in diabetic cardiomyocytes was due, at least in part, to decreased Sp1 expression. Collectively, our study has established a novel mechanism by which impaired MICU1 signaling contributes to diabetic cardiomyopathy.

Mitochondrial Ca²⁺ uptake has been functionally described for decades (25). Until recently, MICU1 (the key component of the MCU complex) has been identified using integrative bioinformatics and RNA interference screening approaches, which together constitute the activity of the MCU complex (13). Antony et al. (26) reported that whole-body knockout of MICU1 on a C57BL/6J background is lethal in the first hours after birth. Another group also indicated that the absence of MICU1 (on a C57BL/6N background) induces a high rate of perinatal mortality (27). Both studies highlight the critical role of MICU1 in physiological conditions. A recent study reported that mutations in the MICU1 gene cause serious brain and muscle disorders by affecting mitochondrial calcium signaling (15). However, functional roles of MICU1 in human diseases, including diabetic cardiomyopathy, remain largely unknown. In this study we demonstrated that downregulated MICU1 in diabetic mouse hearts contributes to myocardial apoptosis in diabetes and the progression of cardiomyopathy. Consistent with our results, Hoffman et al. (16) showed that MICU1 was obviously downregulated at the mRNA level in human cardiovascular disease–derived primary endothelial cells. An in vitro study using nonexcitable cell lines also reported that MICU1 deficiency sensitizes cells to apoptotic cell death (28).

MICU1 is a Ca²⁺ binding protein that resides in the mitochondrial intermembrane space and has been proposed to be a key mediator of signal processing in the mitochondrial Ca²⁺ uniporter complex (13). However, its precise function in mitochondrial Ca²⁺ uptake remains to be debated. Mallilankaraman et al. (29) showed that decreased MICU1 expression leads to a constitutive mitochondrial Ca²⁺ overload under resting conditions, suggesting that MICU1 may act as a gatekeeper of MCU at low cytoplasm Ca²⁺ concentration ([Ca²⁺]_c) (30). However, Csordás et al. (28) showed that MICU1 loss attenuated mitochondrial Ca²⁺ accumulation during agonist-induced [Ca²⁺]_c pulses. These investigations suggest a double role for MICU1 in mitochondrial Ca²⁺ uptake, either as a gatekeeper at low [Ca²⁺]_c or an activator at high [Ca²⁺]_c, which was further confirmed by in vivo studies (26,27). In this study we demonstrated that upregulation of MICU1 significantly increased basal mitochondrial Ca²⁺ content in diabetic hearts and enhanced the capacity of mitochondrial Ca²⁺ uptake in either isolated db/db cardiomyocytes or neonatal cardiomyocytes with HGHF treatment. As indicated by previous evidence, cardiac mitochondria exposed to high [Ca²⁺]_c rapidly sequester Ca²⁺ during excitation-contraction coupling (30–32). Therefore, it is reasonable that the restoration of MICU1 has an activating effect on mitochondrial Ca²⁺ uptake in diabetic hearts.

Mitochondrial Ca²⁺ overload was detrimental, and could be a cause of ischemic heart failure (29,30). However, compared with these conditions, the effect of elevated mitochondrial Ca²⁺ might be very different in diabetic hearts. Several studies provide direct evidence that mitochondrial Ca²⁺ uptake is significantly reduced in ob/ob mouse hearts and streptozotocin-induced diabetic hearts (9,10). Therefore, the recovery of mitochondrial Ca²⁺ in diabetic hearts would promote cardiac function and the survival of cardiomyocytes. Consistent with previous studies, our data demonstrated that MICU1 overexpression recovered basic cardiac function in diabetic hearts, suggesting that impaired mitochondrial Ca²⁺ handling induced by downregulated MICU1 could be a main cause of dysfunction in diabetic hearts. However, other factors besides MICU1 defect may be responsible for impaired function of diabetic hearts, which also can be reversed by MICU1-mediated mitochondrial Ca²⁺ signaling.

ROS generated in mitochondria are widely accepted to play an important role in the development of diabetes and its complications (33). Here, we demonstrated that overexpression of MICU1 reduced cardiomyocyte apoptosis and alleviated myocardial injury in diabetes by reducing MitoROS level. Furthermore, we found that MICU1-mediated mitochondrial Ca²⁺ uptake inhibited ROS-triggered apoptosis by enhancing NAD(P)H produced during the TCA cycle. Our results were consistent with those of a previous study, which observed that blocking mitochondrial Ca²⁺ uptake potentiated H₂O₂ formation by reducing the NADPH-dependent antioxidative capacity in heart failure (7). On the contrary, overexpression of MICU1 in hearts did not benefit WT mice, but induced much higher mortality (data not shown). One of various possible explanations might be that overexpressed MICU1 induced mitochondrial Ca²⁺ overload and consequently caused heart failure. Actually, Santulli et al. (34) reported that in failing hearts, sarcoplasmic reticulum Ca²⁺ leak caused mitochondrial Ca²⁺ overload, which could trigger mitochondrial dysfunction and increase the production of ROS, finally leading to impaired cardiac function. These results clearly indicate the critical role of mitochondrial Ca²⁺ balance in maintaining cardiac function. In particular, damage caused by oxidative stress induced by Ca²⁺ overload should be treated during MICU1-based treatment for patients with diabetic cardiomyopathy.

Importantly, we found that the downregulation of MICU1 in diabetic hearts was partly the result of decreased expression of transcription factor Sp1, which is partially supported by the deregulation of Sp1 observed in various tissues from patients with type 2 diabetes or obesity (35). These results provide the first clue for understanding the regulatory mechanisms of MICU1 expression in diabetic hearts. Interestingly, Sp1 has also been found to regulate the expression of SERCA2 (a key regulator of sarcoplasmic reticulum Ca²⁺) in cardiomyocytes (36), implying that Sp1 may coordinate the regulation of Ca²⁺ signaling in cardiomyocytes.

This study has several limitations. First, the expression of MICU1 in human hearts was analyzed only at the mRNA level, based on a public microarray expression data set. Second, not all the conclusions were derived from in vivo study. Third, the causes of downregulated MICU1 in diabetic hearts still remain largely unknown, although we identified the involvement of transcription factor Sp1. Despite these limitations, we believe that our study provides important new insights into diabetic cardiomyopathy.

In summary, our study provides solid evidence that the downregulation of MICU1 promotes myocardial injury in diabetes, and the reconstitution of MICU1 alleviates diabetic cardiomyopathy via a mitochondrial Ca²⁺–dependent antioxidant pathway. This suggests a potential therapeutic target to mitigate myocardial injury in metabolic diseases.

Funding. This work was supported by the National Natural Science Foundation of China (grant nos. 81400197, 81320108021, and 81170183).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. L.J., F. Liu, and Z.J. performed most experiments, analyzed the data, and wrote the manuscript. Q.H. analyzed data and contributed to the discussion. Y.Z. participated in the in vivo study, isolated adult cardiomyocytes, and measured mitochondrial [Ca²⁺]. H.C. measured mitochondrial [Ca²⁺] and analyzed cell survival. J.L. performed hemodynamic measurements. C.Y. participated in the in vivo study. J.X. and F. Li designed the overall study, supervised the experiments, analyzed the results, and wrote the article. J.X. and F. Li are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.