MICU1 Alleviates Diabetic Cardiomyopathy ThroughMitochondrial Ca2+–Dependent Antioxidant ResponseLele Ji,1 Fengzhou Liu,2 Zhe Jing,3 Qichao Huang,1 Ya Zhao,1 Haiyan Cao,1 Jun Li,4 Chun Yin,1

Jinliang Xing,1 and Fei Li2

Diabetes 2017;66:1586–1600 | https://doi.org/10.2337/db16-1237

Diabetic cardiomyopathy is a major cause of mortality inpatients with diabetes, but specific strategies for pre-venting or treating diabetic cardiomyopathy have notbeen clarified yet. MICU1 is a key regulator of mito-chondrial Ca2+ uptake, which plays important roles inregulating mitochondrial oxidative phosphorylation andredox balance. To date, however, the significance ofMICU1 in diabetic hearts has not been investigated.Here, we demonstrate that MICU1 was downregulatedin db/db mouse hearts, which contributes to myocardialapoptosis in diabetes. Importantly, the reconstitution ofMICU1 in diabetic hearts significantly inhibited the de-velopment of diabetic cardiomyopathy, as evidenced byenhanced cardiac function and reduced cardiac hyper-trophy and myocardial fibrosis in db/db mice. Moreover,our in vitro data show that the reconstitution of MICU1inhibited the apoptosis of cardiomyocytes, induced byhigh glucose and high fat, through increasing mitochondri-al Ca2+ uptake and subsequently activating the antioxidantsystem. Finally, our results indicate that hyperglycemiaand hyperlipidemia induced the downregulation of MICU1by inhibiting Sp1 expression in diabetic cardiomyocytes.Collectively, our findings provide the first direct evidencethat upregulated MICU1 preserves cardiac function in di-abetic db/db mice, suggesting that increasing the expres-sion or activity of MICU1 may be a pharmacologicalapproach to ameliorate cardiomyopathy in diabetes.

Increased prevalence of type 2 diabetes is a major threatto human health. It is estimated that diabetes will affect

nearly 366 million adults by the year 2030 (1). Cumulativeevidence reveals that diabetes is associated with structuraland functional abnormalities of the myocardium, leading todiabetic cardiomyopathy. However, specific strategies for theprevention and treatment of cardiomyopathy in patientswith diabetes have not been clarified yet. Emerging evidencesuggests that mitochondrial dysfunction may play a criticalrole in the pathogenesis of diabetic cardiomyopathy (2). Forexample, recent data have demonstrated that mitochondriaisolated from atrial cardiomyocytes of patients with diabetesexhibit remarkable respiratory defects and increased gener-ation of reactive oxygen species (ROS) when compared withthose from control subjects without diabetes (3,4). More-over, it is essential to protect mitochondria from losing theirability to generate energy and to control their own ROSemission in order to prevent diabetic cardiomyopathy (5).However, despite increasing data implicating mitochondrialpathology in diabetic cardiomyopathy, the mechanism un-derlying these processes still remains largely unknown.

Mitochondrial Ca2+ plays important roles in regulatingthe intrinsic functions of the organelle. One of the best-characterized functions of mitochondrial Ca2+ is the controlof organelle metabolic activity. For example, physiological in-creases in mitochondrial Ca2+ lead to the allosteric activationof tricarboxylic acid (TCA) cycle enzymes including isocitratedehydrogenase, a-ketogluterate dehydrogenase (a-KGDH),and pyruvate dehydrogenase (PDH). The net effect of TCAcycle activation is a boost in reduced oxidative phosphoryla-tion substrate synthesis (NADH and flavin adenine dinucleo-tide), enhanced respiratory chain activity, and a subsequent

1State Key Laboratory of Cancer Biology and Experimental Teaching Center ofBasic Medicine, Fourth Military Medical University, Xi’an, China2Department of Cardiology, Xijing Hospital, Fourth Military Medical University,Xi’an, China3Department of Cardiology, General Hospital of Lanzhou Military Area Command,Lanzhou, China4Department of Physiology, Fourth Military Medical University, Xi’an, China

Corresponding author: Jinliang Xing, xingjinliang@163.com, or Fei Li, lifei2288@hotmail.com.

Received 10 November 2016 and accepted 3 March 2017.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db16-1237/-/DC1.

L.J., F. Liu, and Z.J. contributed equally to this study.

© 2017 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, and thework is not altered. More information is available at http://www.diabetesjournals.org/content/license.

1586 Diabetes Volume 66, June 2017

SIG

NALTRANSDUCTIO

N

increase in ATP synthesis (6). Another key role of mitochon-drial Ca2+ is to modulate oxidative stress by misbalancingROS generation and ROS detoxification. It has been report-ed that mitochondrial Ca2+ regulates cellular antioxidantdefense systems by stimulating NADPH regeneration,which donates electrons to regenerate reduced glutathione(GSH) from oxidized glutathione (GSSG) (7,8). Interest-ingly, several studies have reported that reduced mitochon-drial calcium uptake is observed in the hearts of diabeticrodents (9,10). However, the mechanisms by which diabetesimpairs mitochondrial calcium handling in cardiomyocytesremain to be defined.

Since 2010, the discovery of the pore-forming subunit ofthe mitochondrial Ca2+ uptake channel (mitochondrial calciumuniporter [MCU]) (11,12) and its regulatory subunits, calledmitochondrial calcium uptake 1 (MICU1) (13) and MCU reg-ulator 1 (14), has opened new avenues for the study of mito-chondrial Ca2+ homeostasis regulation and its key roles. Inparticular, a loss-of-function mutation in MICU1 has beenreported to cause human disease through alterations in mito-chondrial Ca2+ handling (15). A previous study reported thatMICU1, but not MCU, mRNA levels were markedly down-regulated in human cardiovascular disease–derived primaryendothelial cells (16). To our knowledge, however, the signif-icance of MICU1 in diabetic hearts, especially in the patho-genesis of diabetic cardiomyopathy, has never been reported.

In this study we identified downregulated MICU1 as afactor contributing to cardiomyocyte apoptosis in diabeticcardiomyopathy. Moreover, MICU1 overexpression effec-tively alleviated diabetic cardiomyopathy by promotingmitochondrial Ca2+ uptake to inhibit mitochondrial ROS(MitoROS)–mediated apoptosis. These findings suggestthat upregulating MICU1 expression may be a potentialtherapeutic strategy in diabetic cardiomyopathy.

RESEARCH DESIGN AND METHODS

AnimalsAll experiments were performed in accordance with theNational Institutes of Health Guidelines on the Use ofLaboratory Animals and approved by the Fourth MilitaryMedical University Committee on Animal Care. Leptinreceptor–deficient (db/db) C57BLKS mice and wild-type (WT)C57BLKS mice were purchased from Changzhou Cavens Lab-oratory Animal Co. Ltd. (Jiangsu, China). The investigatorswere blinded to the treatment conditions during experiments.

Neonatal Rat Cardiomyocyte CulturePrimary cardiomyocytes were prepared from neonatalrats, as previously described (17). For the high-glucoseand high-fat (HGHF) treatments, cells were cultured withDMEM (containing 25 mmol/L glucose and 500 mmol/L ofthe saturated free fatty acid palmitate [16:0]) for 24 h.

Knockdown and Forced Expression of Target GenesMICU1-specific small interfering RNA (siRNA), Sp1-specificsiRNA, and scrambled siRNA were designed and synthe-sized by GenePharma Company (Shanghai, China). The

sequences of siRNAs are provided in Supplementary Table1. siRNAs were transfected as previously described (17).Recombinant adenoviruses overexpressing rat MICU1(NM_199412.1) and Sp1 (NM_012655.2) were constructedby Shanghai Genechem Co., Ltd (Shanghai, China) andHanBio Biotechnology Co., Ltd. (Shanghai, China), respec-tively. Recombinant adenovirus-overexpressing mouseMICU1 (NM_001291442.1) or mitochondrial-targetedparvalbumin was generated following the instructions ofthe ViraPower Adenoviral Expression System (Life Tech-nologies), according to the manufacturer’s protocol. Thecoding sequences of MICU1 and mitochondria-targetedparvalbumin were amplified using the primers listed inSupplementary Table 2. The viral titer was determinedusing an Adeno-X Rapid Titer Kit (Clontech Laboratories,Mountain View, CA). Cardiomyocytes were infected withadenovirus, at a multiplicity of infection of 50, for 2 h.Subsequently, cells were cultured in serum-free DMEMfor an additional 24 h and then used for further analysis.For adenovirus administration, the mice were anesthe-tized with 2% isoflurane, and the heart was exposed.Adenovirus (2 3 1010 infectious units/mL) was injected(using a 30-gauge needle) into the free wall of the leftventricle (LV) (10 mL at each of three sites). The trans-fection efficiency was evaluated using Western blotting3 days after adenoviral injection.

Quantitative RT-PCR, Western Blotting, andImmunohistochemistryRNA extraction, cDNA synthesis, quantitative RT-PCRreactions were performed as described previously (18).Primer sequences are provided in Supplementary Table 3.Mouse heart tissue or primary neonatal cardiomyocyteswere processed for Western blotting, as previously de-scribed (19). Immunohistochemical staining of cardiacsections was carried out as previously described (20).

Histological AnalysisThe hearts of mice were fixed in 4% paraformaldehyde(pH 7.4) overnight, embedded in paraffin, and seriallysectioned (5-mm-thick slices) for histological analysis. Stan-dard hematoxylin and eosin staining was carried out fol-lowing standard procedures. Cardiac collagen content wasassessed using Masson trichrome staining. Myocyte sizewas detected by staining with wheat germ agglutinin(red; Sigma-Aldrich).

Cell Apoptosis AssayCell apoptosis was determined using a PE Annexin VApoptosis Detection Kit (material no. 559763; BD Bio-sciences) or a Fluorescein Isothiocyanate Annexin V Apo-ptosis Detection Kit (BB-4101; BestBio), following themanufacturer’s instructions. For analysis of apoptosisin heart tissues, terminal deoxynucleotidyl TUNEL assay(11684795910; Roche Applied Science) was performedaccording to the manufacturer’s protocol. Images ofTUNEL and DAPI-stained sections were obtained using

diabetes.diabetesjournals.org Ji and Associates 1587

a fluorescence microscope (DM5000B; Leica, Heerbrugg,Switzerland). Only TUNEL- and DAPI-positive nuclei thatwere located within heart tissues were counted as apoptoticnuclei.

Measurement of Mitochondrial Membrane PotentialTo measure mitochondrial membrane potential, cells (1 3106/mL) were incubated with tetramethyl rhodaminemethyl ester (TMRM; 10 nmol/L; Invitrogen) at 37°Cfor 30 min. TMRM images were captured using a confocallaser scanning microscope (FluoView FV1000; Olympus,Japan), with excitation at 530 nm and emission at 573 nm.

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

Isolation of Adult CardiomyocytesMouse hearts were excised and then washed in Ca2+-freetyrode solution, followed by perfusion with buffer con-taining liberase (Roche Diagnostics) at 37°C for 20 min.The LV was subsequently dissociated into single myo-cytes, and extracellular Ca2+ was added back incrementallyto 1.2 mmol/L. Cardiomyocytes were placed on glass cov-erslips coated with laminin (Thermo Fisher Scientific).Cells were used within 4 h after isolation.

Measurement of Mitochondrial Ca2+ ConcentrationThe fluorescent dye Rhod-2/AM (Invitrogen) was used tomonitor mitochondrial Ca2+ concentration ([Ca2+]) in car-diomyocytes, according to the manufacturer’s instruc-tions. Then, cardiomyocytes were viewed with a confocallaser scanning microscope (FluoView FV1000; Olympus).For the measurement of mitochondrial Ca2+ uptake, adultor neonatal cardiomyocytes were stimulated by pacing(3 Hz) or histamine (10 mmol/L), respectively. Imageswere recorded every 5 s using the same confocal imagingsystem. The endogenous mitochondrial Ca2+ content wasmeasured in isolated cardiac mitochondria. Briefly, iso-lated mitochondria were resuspended in buffers, soni-cated, and then centrifuged. The supernatants wererecovered and the Ca2+ content in the supernatants wasdetermined using a calcium detection kit (Abcam).

Detection of MitoROSMitoROS 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 anal-ysis software.

Activity Detection of PDH and a-KGDHThe activity of PDH and a-KGDH was measured using thePDH enzyme activity microplate assay kit (Abcam) anda-KGDH activity assay kit (Biovision), according to themanufacturers’ instructions.

Detection of Mitochondrial NADH and NADPHContents, and GSH-to-GSSG RatioNADH and NADPH concentrations were measured inlysates from isolated mitochondria using NAD/NADHand NADP/NADPH assay kits (Abcam), according to themanufacturer’s instructions. The GSH-to-GSSG ratio wasdetermined in lysates from isolated mitochondria using aGSH and GSSG assay kit (Beyotime), according to themanufacturer’s instructions.

Chromatin Immunoprecipitation AssayChromatin immunoprecipitation (ChIP) assay was per-formed using the SimpleChIP Plus Enzymatic ChromatinIP kit (Cell Signaling Technology, Danvers, MA), accordingto the manufacturer’s instructions. Briefly, after cross-linking, tissues were disaggregated into a single-cell sus-pension. Cells were then lysed and the chromatin wasfragmented. The sheared chromatins were incubated withrabbit anti-Sp1 antibody and protein G magnetic beads. TheDNA released from the precipitation was subjected to PCRanalysis for detection of the presence of specific loci. Theprimers specific to the Sp1 binding region in the mouseMICU1 gene promoter are shown in Supplementary Table4. Rabbit IgG and antibody against histone H3 were used asthe negative and positive controls, respectively.

EchocardiographyNoninvasive serial ultrasound examination of hearts wasperformed using a Vevo 2100 Microultrasound System(VisualSonics Inc., Toronto, Ontario, Canada). Mice werelightly anesthetized with isoflurane (3% for induction and1% for maintenance). Cardiac function under physiolog-ical conditions was cursorily examined by hand-heldmanipulation of the ultrasound transducer to obtain two-dimensional and two-dimensional guided M-mode images.All the echocardiographic images were analyzed using Vevo2100 software.

Hemodynamic MeasurementsInvasive hemodynamic measurements in anesthetizedmice were performed using a 1.4 French micromanometer(Millar Instruments, Houston, TX) inserted into the rightcarotid artery and guided into the LV. Heart rate, LVsystolic pressure, and LV end-diastolic pressure were mea-sured by this catheter, and data were recorded and analyzedusing a PowerLab System (ADInstruments Inc., ColoradoSprings, CO). These parameters, as well as maximal values ofthe instantaneous first derivative of LV pressure (dP/dtmax,the maximum changing speed of intraventricular pressure,as a measure of cardiac contractility) and minimum values ofthe instantaneous first derivative of LV pressure (dP/dtmin,the minimum changing speed of intraventricular pressure,as a measure of cardiac relaxation), were recorded.

Statistical AnalysisAll data were presented as the mean 6 SEM of n inde-pendent experiments. For groups of three or more, the

1588 MICU1 and Diabetic Cardiomyopathy Diabetes Volume 66, June 2017

data were subjected to ANOVA, followed by a Bonferronicorrection for a post hoc test. The unpaired t test wasused for comparison between two groups. Correlationsbetween measured variables were tested using Pearsoncorrelation analysis. Probability #0.05 was consideredstatistically significant.

RESULTS

MICU1 Expression Was Downregulated in DiabeticMouse HeartsTo investigate the alterations of the key proteins un-derlying mitochondrial Ca2+ entry into diabetic hearts, theexpressions of MCU and its regulator MICU1 were deter-mined. Western blotting (Fig. 1A) showed that MICU1expression was significantly downregulated in db/db mousehearts at 12 weeks of age compared with hearts of WTlittermate control mice. The downregulation of MICU1was even more obvious in hearts of 18-week-old db/dbmice. By contrast, MCU expression was unchanged. Aconsistent pattern of mRNA expression was also observedusing quantitative RT-PCR (Fig. 1B), suggesting that theexpression of MICU1 in diabetic hearts is mainly regu-lated at the transcriptional level. Furthermore, immuno-histochemical staining also demonstrated that cardiacMICU1 expression remained unchanged in wild-type

mice at different ages; however, MICU1 was significantlydownregulated in hearts of db/db mice at both 12 and18 weeks of age when compared with their correspondingcontrols (Fig. 1C). Collectively, these findings indicate thatMICU1 downregulation occurs in diabetic cardiomyopathy.

Downregulation of MICU1 Induced Mitochondria-Dependent Intrinsic Apoptosis in CardiomyocytesWe attempted to determine whether MICU1 downregu-lation leads to cell apoptosis in neonatal cardiomyocyteswith or without HGHF treatment. Our data indicate thatthe HGHF treatment resulted in significant MICU1 down-regulation in cardiomyocytes, and MICU1 siRNA signifi-cantly reduced the expression of MICU1 (Fig. 2A).Importantly, the percentages of total (both early andlate) apoptotic cells were significantly higher in neonatalcardiomyocytes with MICU1 knockdown or HGHF treat-ment than those in control cells. Furthermore, MICU1knockdown remarkably exacerbated HGHF-induced apo-ptosis (Fig. 2B). Consistently, the cleavage of caspase-9and caspase-3 (Fig. 2C), release of cytochrome c frommitochondria (Fig. 2D), and disruption of the mitochondrialmembrane potential (Fig. 2E) were significantly induced byHGHF treatment, whereas all of these effects were furtheraggravated upon MICU1 knockdown. These findings

Figure 1—MICU1 expression was downregulated in diabetic mouse hearts. A: Western blotting analyzing the protein expression levels ofMCU and MICU1 in hearts of db/db mice and WT littermate mice. The numbers under the blots indicate the relative levels of MCU andMICU1 after normalization against the b-actin contents. B: qRT-PCR analyses of the mRNA expression level of MICU1 in mouse hearts.**P < 0.01. C: Representative immunohistochemical stains of MICU1 in mouse hearts. Scale bars = 50 mm. n = 6 animals. wk, weeks.

diabetes.diabetesjournals.org Ji and Associates 1589

Figure 2—Downregulation of MICU1 induced mitochondria-dependent intrinsic apoptosis in cardiomyocytes. A: Western blotting analyzingthe protein expression levels of MICU1 in neonatal cardiomyocytes that did or did not receive HGHF treatment. B: Flow cytometry of apoptosisby annexin V (an indicator of apoptosis) and propidium iodide (PI) staining in neonatal cardiomyocytes with treatment as indicated (top).Quantitative analysis of apoptotic cells (bottom). C: Western blotting analyzing protein levels of MICU1, cleaved caspase-9, and cleavedcaspase-3 in neonatal cardiomyocytes with treatment as indicated. D: Western blotting analyzing protein levels of cytochrome c (CytoC) incytoplasm and mitochondria of neonatal cardiomyocytes with treatment as indicated. b-Actin and COX IV were used as loading controls forcytoplasm and mitochondria, respectively. E: Mitochondrial membrane potentials were analyzed by TMRM staining in neonatal cardiomyo-cytes with treatment as indicated. Representative confocal microscope images (top) are presented. Scale bars = 10 mm. Quantitative analysisof TMRM fluorescence (bottom). *P < 0.05; **P < 0.01. n = 6–8 wells. siMICU1, siRNA against MICU1; siCtrl, negative control siRNA.

1590 MICU1 and Diabetic Cardiomyopathy Diabetes Volume 66, June 2017

suggest that downregulation of MICU1 induces mitochondria-dependent intrinsic apoptosis in cardiomyocytes.

Upregulation of MICU1 Enhanced Cardiac Functionin Diabetic db/db MiceWe further attempted to determine whether the upregu-lation of MICU1 protects the heart against diabetic cardio-myopathy in db/db mice. Adenovirus encoding MICU1 wasintramyocardially 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-expressingadenovirus resulted in an ;1.6-fold increase in cardiacMICU1 expression compared with controls. RepresentativeM-mode echocardiograms are shown in Fig. 3B. LV ejectionfraction and LV fractional shortening were impaired indb/db mice when compared with WT mice. Interestingly,upregulated MICU1 significantly augmented LV ejectionfraction and LV fractional shortening (Fig. 3C and D).Moreover, analysis with the Millar transducer showed nosignificant differences in heart rate and mean arterial bloodpressure among all groups. However, there was a significant

Figure 3—Overexpression of MICU1 improved cardiac function in diabetic db/db mice. A: Western blotting (top) and quantitative analyses(bottom) for protein levels of MICU1 in hearts of WT mice or in hearts of db/db mice with or without MICU1 overexpression. B: Repre-sentative echocardiographic images of the LV in different mice. C and D: Echocardiographic assessment of LV ejection fraction (LVEF; C)and LV fractional shortening (LVFS; D) in different mice. E–J: Hemodynamic analyses in different mice. Ad-EV, control adenovirus;Ad-MICU1, recombinant adenovirus encoding MICU1; dP/dtmax, the maximal values of the instantaneous first derivative of left ventricularpressure; dP/dtmin, the minimum values of the instantaneous first derivative of left ventricular pressure; HR, heart rate; MABP, mean arterialblood pressure; LVSP, LV systolic pressure; LVEDP, LV end-diastolic pressure; **P < 0.01. n = 6–8 animals.

diabetes.diabetesjournals.org Ji and Associates 1591

increase in LV end-diastolic pressure and a decrease in LVend-systolic pressure (dP/dtmax and dP/dtmin in diabeticmice), whereas the forced expression of MICU1 resultedin a global improvement of myocardial function (Fig. 3E–J).Taken together, these data indicate that MICU1 restoresLV cardiac function in diabetic db/db mice.

Upregulation of MICU1 Ameliorated CardiacHypertrophy in Diabetic db/db MiceLV 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 ofcardiac hypertrophy, as evidenced by increased heartweight, enlarged hearts, and elevated ratio of heart weightto tibia length, as well as augmented LV mass and en-larged LV inner diameter, with reduced posterior wallthickness. By contrast, the injection of MICU1-expressingadenovirus significantly attenuated these pathologicalchanges (Fig. 4A–C and Supplementary Fig. 1). Moreover,the injection of MICU1-expressing adenovirus remarkablyreduced interstitial fibrosis in the db/db hearts (Fig. 4D).In addition, the up-regulation of MICU1 decreased thecross-sectional area of cardiomyocytes in db/db hearts(Fig. 4E) and reversed the mRNA expression changesof hypertrophic genes, including the upregulation ofa–myosin heavy chain and downregulation of b–myosinheavy chain, atrial natriuretic peptide, and brain natriureticpeptide (Fig. 4F).

Upregulated MICU1 Inhibited Cell Apoptosis byReducing MitoROS in Diabetic CardiomyocytesTUNEL staining analysis showed that the forced expres-sion of MICU1 significantly reduced the number ofapoptotic cells in db/db hearts (Fig. 5A). Consistently,upregulated MICU1 remarkably decreased the cleavageof caspase-9 and caspase-3 in db/db hearts (Fig. 5B).In vitro study further confirmed that the forced expres-sion of MICU1 remarkably inhibited cell apoptosis inHGHF-treated neonatal cardiomyocytes, as evidenced bydecreased percentages of total (both early and late) apo-ptotic cells (Fig. 5C and Supplementary Fig. 1A), and re-duced the cleavage of caspase-9 and caspase-3 (Fig. 5D).Moreover, MICU1 overexpression in cardiomyocytes sig-nificantly decreased the release of cytochrome c frommitochondria (Fig. 5E) and elevated the mitochondrialmembrane potential (Fig. 5F and Supplementary Fig. 2B),indicating that MICU1 inhibits mitochondria-dependentintrinsic apoptosis in cardiomyocytes.

Previous studies have demonstrated that hearts ofdb/db mice exhibited excessive ROS production, which isassociated with the induction of cell death (21). Our dataindicated that the intramyocardial injection of MICU1-expressing adenovirus significantly reduced mitochondrialROS production in hearts of db/db mice (Fig. 5G). Simi-larly, the inhibitory effect of MICU1 overexpression onMitoROS production was clearly observed in HGHF-treated

neonatal cardiomyocytes in vitro (Fig. 5H). Further-more, flow cytometry analysis showed that MICU1 over-expression remarkably blocked H2O2-induced apoptosis incardiomyocytes with or without HGHF treatment (Fig. 5Iand Supplementary Fig. 2C). Collectively, all our findingsindicate that upregulated MICU1 inhibits cell apoptosisthrough reducing MitoROS in diabetic cardiomyocytes.

MICU1 Inhibited ROS-Mediated Apoptosis inHGHF-Treated Cardiomyocytes by IncreasingMitochondrial Ca2+ UptakeSince MICU1 has been identified as a critical regulator ofmitochondrial Ca2+ uniporter, we next determined whethermitochondrial Ca2+ played an important role in MICU1-mediated antiapoptosis in diabetic cardiomyopathy. Endog-enous mitochondrial Ca2+ content was measured in isolatedcardiac mitochondria. As shown in Fig. 6A, MICU1 over-expression significantly increased the basal mitochondrialCa2+ in db/db hearts. Consistently, the forced expressionof MICU1 significantly enhanced the capacity of mitochon-drial Ca2+ uptake following pacing (Fig. 6B). More impor-tantly, our data demonstrated that mitochondrial Ca2+

buffering by mitoPV (a mitochondria-targeted Ca2+ bindingprotein parvalbumin) remarkably blocked the antioxidanteffect of MICU1 overexpression in HGHF-treated cardio-myocytes (Fig. 6C and Supplementary Fig. 3). Furthermore,mitoPV treatment decreased cytochrome c release frommitochondria (Fig. 6D) and inhibited the cleavage ofcaspase-9 and caspase-3 (Fig. 6E). All these findings sug-gest that MICU1 overexpression inhibits ROS-mediatedapoptosis in HGHF-treated cardiomyocytes by increasingmitochondrial Ca2+ uptake.

MICU1-Mediated Mitochondrial Ca2+ UptakeEnhanced the Antioxidant Effect in HGHF-TreatedCardiomyocytesRegeneration of NADPH, a key component in cellularantioxidant systems, requires the products of the Krebscycle, including isocitrate, malate, and NADH (22). Sincethe mitochondrial Ca2+–activated dehydrogenases of theTCA cycle are known to accelerate the regeneration ofNAD(P)H (7,23), it is reasonable to hypothesize thatMICU1-mediated mitochondrial Ca2+ uptake enhancesthe antioxidant system in cardiomyocytes (Fig. 7A).Therefore, we first investigated the enzyme activities ofCa2+-dependent dehydrogenases in neonatal cardiomyo-cytes. Our data indicated that HGHF treatment inhibitedthe enzyme activities of PDH and a-KGDH in neonatalcardiomyocytes, whereas the forced expression of MICU1partially reversed the inhibitory effect. Furthermore, mito-chondrial Ca2+ buffering with mitoPV remarkably blockedthe effects of MICU1 (Fig. 7B and C). Then, we found thatHGHF treatment reduced NADH and NADPH concentra-tions, whereas MICU1 overexpression remarkably reversedthese effects in HGHF-treated neonatal cardiomyocytes.Similarly, mitoPV blocked MICU1-enhanced NADH andNADPH generation (Fig. 7D and E). In addition, our data

1592 MICU1 and Diabetic Cardiomyopathy Diabetes Volume 66, June 2017

Figure 4—Overexpression of MICU1 ameliorated cardiac hypertrophy in diabetic db/db mice. A: The gross morphology of hearts from differentmice was analyzed by hematoxylin and eosin staining. Scale bar = 2 mm. B and C: The heart weight (B) and the ratio of heart weight to tibialength (C) were measured in different mice. D: Left: Representative images of Masson trichrome staining of hearts from different mice. Scalebars = 25 mm. Right: Quantification of interstitial fibrosis. E: Left: Cardiomyocyte cross-sectional area was analyzed by wheat germ agglutininstaining in hearts from different mice. Scale bars = 10 mm. Right: Quantification of cross-sectional area of cardiomyocytes. F: qRT-PCRanalyses of the mRNA expression level of cardiac hypertrophy–related genes in hearts from different mice. *P < 0.05; **P < 0.01. n = 5–6animals. Ad-EV, control adenovirus; Ad-MICU1, recombinant adenovirus encoding MICU1.

diabetes.diabetesjournals.org Ji and Associates 1593

Figure 5—Overexpression of MICU1 inhibited cell apoptosis by reducing MitoROS in diabetic cardiomyocytes. A: Left: Representativephotomicrographs of in situ detection of apoptotic myocytes by TUNEL staining in heart tissue. Green fluorescence shows TUNEL-positivenuclei; blue fluorescence shows nuclei of total cardiomyocytes (DAPI-positive). Right: Percentage of TUNEL-positive nuclei in heart tissuesections. Scale bars = 25 mm. n = 5–6 animals. B: Western blotting analyzing protein levels of MICU1, cleaved caspase-9, and cleavedcaspase-3 in mouse hearts. C: Flow cytometry analysis of apoptosis by annexin V and propidium iodide staining in neonatal mousecardiomyocytes with treatment as indicated. n = 6–8 wells. D: Western blotting analyzing protein levels of MICU1, cleaved caspase-9,and cleaved caspase-3 in neonatal mouse cardiomyocytes with treatment as indicated. E: Western blotting analyzing protein levels of

1594 MICU1 and Diabetic Cardiomyopathy Diabetes Volume 66, June 2017

showed that HGHF treatment significantly reduced the ratioof mitochondrial GSH to GSSG (an indicator of intracellularredox status), whereas MICU overexpression exhibited anopposite effect, which was significantly reversed throughmitochondrial Ca2+ buffering by mitoPV (Fig. 7F). Collec-tively, these findings suggest that MICU1 overexpressionand subsequent elevation of mitochondrial Ca2+ uptake in-hibit ROS production by enhancing the NADPH-dependentantioxidant system.

Hyperglycemia and Hyperlipidemia Induce MICU1Downregulation by Inhibiting Sp1 Expression inDiabetic CardiomyocytesWe finally investigated the mechanism underlying thedownregulation of MICU1 in diabetic hearts. Potentialtranscription factor Sp1 for MICU1 was predicted by boththe Promoter 2.0 and Promoter Scan software (data notshown). Moreover, several putative Sp1 binding sites,which are conserved across multiple species, were identi-fied in the 59 flanking region of human MICU1 by theJaspar program (http://jaspar.genereg.net) (Fig. 8A andSupplementary Fig. 4). We used a public microarraydata set of mRNA expression (GSE26887), downloadedfrom the GEO database (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE26887), to evaluate the corre-lation between the mRNA expression levels of MICU1 andSp1 in human hearts. Pearson correlation analysis indi-cated a significant positive correlation between the mRNAexpression level of Sp1 (GSE26887 gene ID, 7955787) andMICU1 (GSE26887 gene ID, 7934255) in a panel of hu-man heart tissues (r = 0.6003; P = 0.002). This correlationwill be even more significant when we analyze only theheart tissues of patients with diabetes and heart failurefrom the same data set (r = 0.7923; P = 0.03) (Fig. 8B).Consistently, both mRNA and protein levels of Sp1 weresignificantly downregulated in db/db mouse hearts at12 weeks of age when compared with age-matched WTmouse hearts. The downregulation of Sp1 was more ob-vious in hearts of 18-week-old db/db mice (Fig. 8C and D).ChIP assay demonstrated that Sp1 bound directly to thepromoter region of the MICU1 gene in mouse heart tissues(Fig. 8E). Moreover, Western blotting revealed that Sp1knockdown significantly decreased the expression ofMICU1 and exacerbated the downregulation of MICU1induced by HGHF treatment. By contrast, Sp1 overex-pression exhibited the opposite effect (Fig. 8D and F).Taken together, these results demonstrate that MICU1

downregulation in diabetic cardiomyocytes is due, at leastin part, to the decreased expression of transcriptionalfactor Sp1.

DISCUSSION

Accumulating data from experimental and clinical studieshave shown that diabetes results in functional and structuralchanges of the myocardium (24). However, the molecularpathological mechanism of diabetic cardiomyopathy is farfrom clear. Here, we have made several important obser-vations. First, we validated that MICU1 was downregulatedin hearts of diabetic mice, which contributes to myocardialapoptosis in diabetes. Second, the forced expression ofMICU1 inhibited the development of diabetic cardiomy-opathy. Third, the antiapoptotic effect of MICU1 wasmediated by promoting mitochondrial Ca2+ uptake andsubsequently activating the antioxidant system in HGHF-treated cardiomyocytes. Last, MICU1 downregulation indiabetic cardiomyocytes was due, at least in part, to de-creased Sp1 expression. Collectively, our study has estab-lished a novel mechanism by which impaired MICU1signaling contributes to diabetic cardiomyopathy.

Mitochondrial Ca2+ uptake has been functionally de-scribed for decades (25). Until recently, MICU1 (the keycomponent of the MCU complex) has been identified usingintegrative bioinformatics and RNA interference screeningapproaches, which together constitute the activity of theMCU complex (13). Antony et al. (26) reported that whole-body knockout of MICU1 on a C57BL/6J background islethal in the first hours after birth. Another group alsoindicated that the absence of MICU1 (on a C57BL/6Nbackground) induces a high rate of perinatal mortality(27). Both studies highlight the critical role of MICU1 inphysiological conditions. A recent study reported that mu-tations in the MICU1 gene cause serious brain and muscledisorders by affecting mitochondrial calcium signaling (15).However, functional roles of MICU1 in human diseases,including diabetic cardiomyopathy, remain largely un-known. In this study we demonstrated that downregulatedMICU1 in diabetic mouse hearts contributes to myocardialapoptosis in diabetes and the progression of cardiomyop-athy. Consistent with our results, Hoffman et al. (16)showed that MICU1 was obviously downregulated at themRNA level in human cardiovascular disease–derived pri-mary endothelial cells. An in vitro study using nonexcitablecell lines also reported that MICU1 deficiency sensitizescells to apoptotic cell death (28).

cytochrome c (CytoC) in cytoplasm and mitochondria of neonatal mouse cardiomyocytes with treatment as indicated. F: Mitochondrialmembrane potentials were analyzed by TMRM staining in neonatal cardiomyocytes with treatment as indicated. n = 6–8 wells. G: MitoROSlevels were analyzed by confocal microscopy after staining with MitoSOX in mouse hearts. Representative confocal microscope images(left) and fluorescence quantitation (right) are presented. Scale bars = 50 mm. n = 5 animals. H: Confocal images (left) of cells stained withMitoSOX and fluorescence quantitation (right) in neonatal cardiomyocytes with treatment as indicated. Scale bars = 10 mm. I: Flow cytometryof apoptosis by annexin V and 7-AAD staining (left) and quantification of apoptotic cells (right) in neonatal cardiomyocytes with treatment asindicated. n = 6–8 wells. *P < 0.05; **P < 0.01. Ad-EV, control adenovirus; Ad-MICU1, recombinant adenovirus encoding MICU1; PE,phycoerythrin.

diabetes.diabetesjournals.org Ji and Associates 1595

MICU1 is a Ca2+ binding protein that resides in the mi-tochondrial intermembrane space and has been proposed tobe a key mediator of signal processing in the mitochondrialCa2+ uniporter complex (13). However, its precise func-tion in mitochondrial Ca2+ uptake remains to be debated.Mallilankaraman et al. (29) showed that decreased MICU1

expression leads to a constitutive mitochondrial Ca2+ over-load under resting conditions, suggesting that MICU1 mayact as a gatekeeper of MCU at low cytoplasm Ca2+ concen-tration ([Ca2+]c) (30). However, Csordás et al. (28) showedthat MICU1 loss attenuated mitochondrial Ca2+ accumulationduring agonist-induced [Ca2+]c pulses. These investigations

Figure 6—MICU1 inhibited ROS-mediated apoptosis in HGHF-treated cardiomyocytes by enhancing mitochondrial Ca2+ uptake. A: Quan-tification of Ca2+ content in isolated cardiac mitochondria from the indicated groups of mice (n = 4–5 animals). B: Representative traces ofpacing (3 Hz)–induced mitochondrial Ca2+ uptake in isolated adult cardiomyocytes from the indicated groups of mice (n = 24–30 myocytesfrom at least 4 mice/group). C: MitoROS levels were analyzed by MitoSOX staining in neonatal cardiomyocytes with treatment as indicated.Representative confocal microscope images (left) and fluorescence quantitation (right) are presented. Scale bars = 10 mm. D: Western blottinganalyzed protein levels of cytochrome c (CytoC) in cytoplasm and mitochondria of neonatal cardiomyocytes with treatment as indicated. E:Western blotting analyzed protein levels of cleaved caspase-9 and cleaved caspase-3 in neonatal mouse cardiomyocytes with treatment asindicated. *P < 0.05; **P < 0.01. n = 6–8 wells. Ad-EV, control adenovirus; Ad-MICU1, recombinant adenovirus encoding MICU1.

1596 MICU1 and Diabetic Cardiomyopathy Diabetes Volume 66, June 2017

suggest a double role for MICU1 in mitochondrial Ca2+

uptake, either as a gatekeeper at low [Ca2+]c or an acti-vator at high [Ca2+]c, which was further confirmed byin vivo studies (26,27). In this study we demonstratedthat upregulation of MICU1 significantly increased basalmitochondrial Ca2+ content in diabetic hearts and enhancedthe capacity of mitochondrial Ca2+ uptake in either isolateddb/db cardiomyocytes or neonatal cardiomyocytes withHGHF 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 restorationof 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, com-pared with these conditions, the effect of elevated mito-chondrial Ca2+ might be very different in diabetic hearts.Several studies provide direct evidence that mitochondrialCa2+ uptake is significantly reduced in ob/ob mouse hearts

Figure 7—MICU1-mediated mitochondrial Ca2+ uptake enhanced the antioxidant system in HGHF-treated cardiomyocytes. A: Schematicdiagram of the connection among mitochondrial Ca2+, the TCA cycle, and the NADPH-mediated antioxidant system. B and C: Relative PDHactivity (B) and a-KGDH activity (C) were evaluated in neonatal cardiomyocytes with treatment as indicated. D and E: The concentration ofmitochondrial NADH (D) and mitochondrial NADPH (E) were evaluated in neonatal cardiomyocytes with treatment as indicated. F: The ratioof mitochondrial GSH to GSSG was evaluated in neonatal cardiomyocytes with treatment as indicated. **P < 0.01. n = 6–8 wells. Ad-EV,control adenovirus; Ad-MICU1, recombinant adenovirus encoding MICU1; IDH, isocitrate dehydrogenase.

diabetes.diabetesjournals.org Ji and Associates 1597

Figure 8—Hyperglycemia and hyperlipidemia induced MICU1 downregulation by inhibiting Sp1 expression in diabetic cardiomyocytes. A:In silico analysis of putative binding sites for the transcription factor in the human MICU1 promoter region. B: Correlations between themRNA expression levels of MICU1 and Sp1 were evaluated in human hearts based on a public microarray expression data set (GSE26887).Top panel: 19 hearts from patients with heart failure (HF) and 5 nonfailing control hearts. Bottom panel: Only seven hearts from patients withHF with type 2 diabetes were included. C and D: qRT-PCR (C) and Western blotting (D) analyzing the expression level of Sp1 in hearts ofdb/db mice and WT littermate mice. E: ChIP analysis of Sp1 binding to the MICU1 promoter in mouse hearts. F and G: Western blottinganalyzed protein levels of Sp1 and MICU1 in neonatal cardiomyocytes with treatment as indicated. **P < 0.01. n = 5 animals. Ad,adenovirus; wk, weeks.

1598 MICU1 and Diabetic Cardiomyopathy Diabetes Volume 66, June 2017

and streptozotocin-induced diabetic hearts (9,10). There-fore, the recovery of mitochondrial Ca2+ in diabetic heartswould promote cardiac function and the survival of car-diomyocytes. Consistent with previous studies, our datademonstrated that MICU1 overexpression recoveredbasic cardiac function in diabetic hearts, suggestingthat impaired mitochondrial Ca2+ handling induced bydownregulated MICU1 could be a main cause of dysfunc-tion in diabetic hearts. However, other factors besidesMICU1 defect may be responsible for impaired functionof diabetic hearts, which also can be reversed by MICU1-mediated mitochondrial Ca2+ signaling.

ROS generated in mitochondria are widely accepted toplay an important role in the development of diabetes andits complications (33). Here, we demonstrated that over-expression of MICU1 reduced cardiomyocyte apoptosisand alleviated myocardial injury in diabetes by reducingMitoROS level. Furthermore, we found that MICU1-mediated mitochondrial Ca2+ uptake inhibited ROS-triggeredapoptosis by enhancing NAD(P)H produced during the TCAcycle. Our results were consistent with those of a previousstudy, which observed that blocking mitochondrial Ca2+

uptake potentiated H2O2 formation by reducing theNADPH-dependent antioxidative capacity in heart failure(7). On the contrary, overexpression of MICU1 in heartsdid not benefit WT mice, but induced much higher mortal-ity (data not shown). One of various possible explanationsmight be that overexpressed MICU1 induced mitochondrialCa2+ overload and consequently caused heart failure. Actu-ally, Santulli et al. (34) reported that in failing hearts,sarcoplasmic reticulum Ca2+ leak caused mitochondrialCa2+ overload, which could trigger mitochondrial dysfunc-tion and increase the production of ROS, finally leading toimpaired cardiac function. These results clearly indicatethe critical role of mitochondrial Ca2+ balance in main-taining cardiac function. In particular, damage caused byoxidative stress induced by Ca2+ overload should be treatedduring MICU1-based treatment for patients with diabeticcardiomyopathy.

Importantly, we found that the downregulation ofMICU1 in diabetic hearts was partly the result of decreasedexpression of transcription factor Sp1, which is partiallysupported by the deregulation of Sp1 observed in varioustissues from patients with type 2 diabetes or obesity (35).These results provide the first clue for understanding theregulatory mechanisms of MICU1 expression in diabetichearts. Interestingly, Sp1 has also been found to regulatethe expression of SERCA2 (a key regulator of sarcoplasmicreticulum Ca2+) in cardiomyocytes (36), implying thatSp1 may coordinate the regulation of Ca2+ signaling incardiomyocytes.

This study has several limitations. First, the expressionof MICU1 in human hearts was analyzed only at the mRNAlevel, based on a public microarray expression data set.Second, not all the conclusions were derived from in vivostudy. Third, the causes of downregulated MICU1 in diabetichearts still remain largely unknown, although we identified

the involvement of transcription factor Sp1. Despite theselimitations, we believe that our study provides importantnew insights into diabetic cardiomyopathy.

In summary, our study provides solid evidence that thedownregulation of MICU1 promotes myocardial injury indiabetes, and the reconstitution of MICU1 alleviates diabeticcardiomyopathy via a mitochondrial Ca2+–dependent anti-oxidant pathway. This suggests a potential therapeutic tar-get to mitigate myocardial injury in metabolic diseases.

Funding. This work was supported by the National Natural ScienceFoundation of China (grant nos. 81400197, 81320108021, and 81170183).Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. L.J., F. Liu, and Z.J. performed most experi-ments, analyzed the data, and wrote the manuscript. Q.H. analyzed data andcontributed to the discussion. Y.Z. participated in the in vivo study, isolated adultcardiomyocytes, and measured mitochondrial [Ca2+]. H.C. measured mitochon-drial [Ca2+] and analyzed cell survival. J.L. performed hemodynamic measure-ments. C.Y. participated in the in vivo study. J.X. and F. Li designed the overallstudy, 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 allthe data in the study and take responsibility for the integrity of the data and theaccuracy of the data analysis.

References1. Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes:estimates for the year 2000 and projections for 2030. Diabetes Care 2004;27:1047–10532. Boudina S, Abel ED. Diabetic cardiomyopathy revisited. Circulation 2007;115:3213–32233. Anderson EJ, Kypson AP, Rodriguez E, Anderson CA, Lehr EJ, Neufer PD.Substrate-specific derangements in mitochondrial metabolism and redox balancein the atrium of the type 2 diabetic human heart. J Am Coll Cardiol 2009;54:1891–18984. Montaigne D, Marechal X, Coisne A, et al. Myocardial contractile dysfunctionis associated with impaired mitochondrial function and dynamics in type 2 di-abetic but not in obese patients. Circulation 2014;130:554–5645. Aon MA, Tocchetti CG, Bhatt N, Paolocci N, Cortassa S. Protective mech-anisms of mitochondria and heart function in diabetes. Antioxid Redox Signal2015;22:1563–15866. Jouaville LS, Pinton P, Bastianutto C, Rutter GA, Rizzuto R. Regulation ofmitochondrial ATP synthesis by calcium: evidence for a long-term metabolicpriming. Proc Natl Acad Sci U S A 1999;96:13807–138127. Kohlhaas M, Liu T, Knopp A, et al. Elevated cytosolic Na+ increases mi-tochondrial formation of reactive oxygen species in failing cardiac myocytes.Circulation 2010;121:1606–16138. Chen Y, Csordás G, Jowdy C, et al. Mitofusin 2-containing mitochondrial-reticular microdomains direct rapid cardiomyocyte bioenergetic responses viainterorganelle Ca(2+) crosstalk. Circ Res 2012;111:863–8759. Oliveira PJ, Seiça R, Coxito PM, et al. Enhanced permeability transitionexplains the reduced calcium uptake in cardiac mitochondria from streptozotocin-induced diabetic rats. FEBS Lett 2003;554:511–51410. Fauconnier J, Lanner JT, Zhang SJ, et al. Insulin and inositol 1,4,5-trisphosphate trigger abnormal cytosolic Ca2+ transients and reveal mito-chondrial Ca2+ handling defects in cardiomyocytes of ob/ob mice. Diabetes2005;54:2375–238111. Baughman JM, Perocchi F, Girgis HS, et al. Integrative genomics identifiesMCU as an essential component of the mitochondrial calcium uniporter. Nature2011;476:341–345

diabetes.diabetesjournals.org Ji and Associates 1599

12. De Stefani D, Raffaello A, Teardo E, Szabò I, Rizzuto R. A forty-kilodaltonprotein of the inner membrane is the mitochondrial calcium uniporter. Nature2011;476:336–34013. Perocchi F, Gohil VM, Girgis HS, et al. MICU1 encodes a mitochondrial EFhand protein required for Ca(2+) uptake. Nature 2010;467:291–29614. Mallilankaraman K, Cárdenas C, Doonan PJ, et al. MCUR1 is an essentialcomponent of mitochondrial Ca2+ uptake that regulates cellular metabolism. NatCell Biol 2012;14:1336–134315. Logan CV, Szabadkai G, Sharpe JA, et al.; UK10K Consortium. Loss-of-function mutations in MICU1 cause a brain and muscle disorder linked to primaryalterations in mitochondrial calcium signaling. Nat Genet 2014;46:188–19316. Hoffman NE, Chandramoorthy HC, Shamugapriya S, et al. MICU1 motifsdefine mitochondrial calcium uniporter binding and activity. Cell Rep 2013;5:1576–158817. Su H, Ji L, Xing W, et al. Acute hyperglycaemia enhances oxidative stressand aggravates myocardial ischaemia/reperfusion injury: role of thioredoxin-interacting protein. J Cell Mol Med 2013;17:181–19118. Huang Q, Zhan L, Cao H, et al. Increased mitochondrial fission promotesautophagy and hepatocellular carcinoma cell survival through the ROS-modulatedcoordinated regulation of the NFKB and TP53 pathways. Autophagy 2016;12:999–101419. Ji L, Zhang X, Liu W, et al. AMPK-regulated and Akt-dependent enhance-ment of glucose uptake is essential in ischemic preconditioning-alleviated re-perfusion injury. PLoS One 2013;8:e6991020. Huang Q, Li J, Xing J, et al. CD147 promotes reprogramming of glucosemetabolism and cell proliferation in HCC cells by inhibiting the p53-dependentsignaling pathway. J Hepatol 2014;61:859–86621. Bhatt NM, Aon MA, Tocchetti CG, et al. Restoring redox balance enhancescontractility in heart trabeculae from type 2 diabetic rats exposed to high glucose.Am J Physiol Heart Circ Physiol 2015;308:H291–H30222. Ying W. NAD+/NADH and NADP+/NADPH in cellular functions and cell death:regulation and biological consequences. Antioxid Redox Signal 2008;10:179–20623. Rizzuto R, De Stefani D, Raffaello A, Mammucari C. Mitochondria as sensorsand regulators of calcium signalling. Nat Rev Mol Cell Biol 2012;13:566–578

24. Fang ZY, Prins JB, Marwick TH. Diabetic cardiomyopathy: evidence,mechanisms, and therapeutic implications. Endocr Rev 2004;25:543–56725. Santo-Domingo J, Demaurex N. Calcium uptake mechanisms of mito-chondria. Biochim Biophys Acta 2010;1797:907–91226. Antony AN, Paillard M, Moffat C, et al. MICU1 regulation of mitochondrial Ca(2+)uptake dictates survival and tissue regeneration. Nat Commun 2016;7:1095527. Liu JC, Liu J, Holmström KM, et al. MICU1 serves as a molecular gatekeeperto prevent in vivo mitochondrial calcium overload. Cell Rep 2016;16:1561–157328. Csordás G, Golenár T, Seifert EL, et al. MICU1 controls both the thresholdand cooperative activation of the mitochondrial Ca2+ uniporter. Cell Metab 2013;17:976–98729. Mallilankaraman K, Doonan P, Cárdenas C, et al. MICU1 is an essentialgatekeeper for MCU-mediated mitochondrial Ca(2+) uptake that regulates cellsurvival. Cell 2012;151:630–64430. Williams GS, Boyman L, Chikando AC, Khairallah RJ, Lederer WJ. Mito-chondrial calcium uptake. Proc Natl Acad Sci U S A 2013;110:10479–1048631. Drago I, De Stefani D, Rizzuto R, Pozzan T. Mitochondrial Ca2+ uptakecontributes to buffering cytoplasmic Ca2+ peaks in cardiomyocytes. Proc NatlAcad Sci U S A 2012;109:12986–1299132. Maack C, Cortassa S, Aon MA, Ganesan AN, Liu T, O’Rourke B. Elevatedcytosolic Na+ decreases mitochondrial Ca2+ uptake during excitation-contractioncoupling and impairs energetic adaptation in cardiac myocytes. Circ Res 2006;99:172–18233. Giacco F, Brownlee M. Oxidative stress and diabetic complications. Circ Res2010;107:1058–107034. Santulli G, Xie W, Reiken SR, Marks AR. Mitochondrial calcium overload is akey determinant in heart failure. Proc Natl Acad Sci U S A 2015;112:11389–1139435. Chen J, Meng Y, Zhou J, et al. Identifying candidate genes for type 2 di-abetes mellitus and obesity through gene expression profiling in multiple tissuesor cells. J Diabetes Res 2013;2013:97043536. Brady M, Koban MU, Dellow KA, Yacoub M, Boheler KR, Fuller SJ. Sp1 andSp3 transcription factors are required for trans-activation of the human SERCA2promoter in cardiomyocytes. Cardiovasc Res 2003;60:347–354

1600 MICU1 and Diabetic Cardiomyopathy Diabetes Volume 66, June 2017