Nuclear-mitochondrial communication involving miR-181c plays an important role in cardiac dysfunction during obesity

Abstract

Aims

In cardiomyocytes, there is microRNA (miR) in the mitochondria that originates from the nuclear genome and matures in the cytoplasm before translocating into the mitochondria. Overexpression of one such miR, miR-181c, can lead to heart failure by stimulating reactive oxygen species (ROS) production and increasing mitochondrial calcium level ([Ca²⁺]_m). Mitochondrial calcium uptake 1 protein (MICU1), a regulatory protein in the mitochondrial calcium uniporter complex, plays an important role in regulating [Ca²⁺]_m. Obesity results in miR-181c overexpression and a decrease in MICU1. We hypothesize that lowering miR-181c would protect against obesity-induced cardiac dysfunction.

Methods and results

We used an in vivo mouse model of high-fat diet (HFD) for 18 weeks and induced high lipid load in H9c2 cells with oleate-conjugated bovine serum albumin in vitro. We tested the cardioprotective role of lowering miR-181c by using miR-181c/d^−/− mice (in vivo) and AntagomiR against miR-181c (in vitro). HFD significantly upregulated heart levels of miR-181c and led to cardiac hypertrophy in wild-type mice, but not in miR-181c/d^−/− mice. HFD also increased in ROS production and pyruvate dehydrogenase activity (a surrogate for [Ca²⁺]_m), but the increases were alleviated in miR-181c/d^−/− mice. Moreover, miR-181c/d^−/− mice fed a HFD had higher levels of MICU1 than did wild-type mice fed a HFD, attenuating the rise in [Ca²⁺]_m. Overexpression of miR-181c in neonatal ventricular cardiomyocytes (NMVM) caused increased ROS production, which oxidized transcription factor Sp1 and led to a loss of Sp1, thereby slowing MICU1 transcription. Hence, miR-181c increases [Ca²⁺]_m through Sp1 oxidation and downregulation of MICU1, suggesting that the cardioprotective effect of miR-181c/d^−/− results from inhibition of Sp1 oxidation.

Conclusion

This study has identified a unique nuclear-mitochondrial communication mechanism in the heart orchestrated by miR-181c. Obesity-induced overexpression of miR-181c increases [Ca²⁺]_m via downregulation of MICU1 and leads to cardiac injury. A strategy to inhibit miR-181c in cardiomyocytes can preserve cardiac function during obesity by improving mitochondrial function. Altering miR-181c expression may provide a pharmacologic approach to improve cardiomyopathy in individuals with obesity/type 2 diabetes.

1. Introduction

Obesity has become a major public health issue around the globe. Various population studies have shown epidemic tendencies for both children and adults [1]. Being obese increases the risk of morbidity and mortality from diseases such as diabetes, atherosclerosis, hypertension, certain types of cancers, and heart disease [2–6]. A recent study concluded that adults between the ages of 40 and 59 who are obese have a significantly increased risk (up to 85% higher) of developing cardiovascular disease when compared with their normal weight peers [7]. Obesity can lead to a variety of cardiac problems, including heart failure. A sub-analysis of the Framingham data revealed that the risk of developing heart failure was twice as high for patients with a body mass index (BMI) > 30 as for those with a BMI < 25. This risk may be due to molecular changes in cardiomyocytes, including dysfunctional mitochondria.

Alterations in mitochondrial function affect multiple cellular and physiologic functions, contributing to the development of various cardiovascular diseases [8]. Mitochondrial reactive oxygen species (ROS) and mitochondrial matrix calcium ([Ca²⁺]_m) play important roles in the pathogenesis of obesity-induced heart failure [8–10]. Increased [Ca²⁺]_m can alter cellular metabolic activity by activating tricarboxylic acid (TCA) cycle enzymes, such as isocitrate dehydrogenase, α-ketoglutarate dehydrogenase, and pyruvate dehydrogenase (PDH) [11]. Therefore, by activating the TCA cycle, increased [Ca²⁺]_m can enhance mitochondrial respiratory chain activity. Furthermore, [Ca²⁺]_m plays an important role in ROS detoxification by stimulating NADPH to reduce glutathione disulfide to glutathione through glutathione reductase [12, 13]. On the other hand, it has been shown that [Ca²⁺]_m accumulation can induce cardiomyocyte death by activating mitochondrial permeability transition pore opening [14]. Therefore, in cardiomyocytes, the mechanisms that connect obesity to altered mitochondrial Ca²⁺ homeostasis and cardiac dysfunction are complex and currently poorly understood.

The mitochondrial calcium uniporter (MCU), along with its regulatory components, such as MCUR1, EMRE, and mitochondrial calcium uptake 1 protein (MICU1), is the major transporter complex through which Ca²⁺ enters the mitochondria [15–21]. Moreover, MICU1 has been identified as a key regulator of [Ca²⁺]_m signaling in the pathophysiology of various diseases [19, 22, 23], including diabetes [24]. However, the mechanisms that regulate MICU1 expression, and how MICU1 regulates [Ca²⁺]_m, are mostly unknown. It has been postulated that loss of the transcription factor Sp1 (Specificity protein 1) may be responsible for reduced MICU1 expression in diabetic myocardium [24].

miRNAs are small (19–25 nt), non-coding RNAs that can inhibit the translational process by directly binding to the 3’-UTR of a mRNA [25, 26]. We have identified a subset of miRNAs (miR-181c), which are encoded in the nuclear genome, mature in the cytoplasm, and then translocate into the mitochondrial matrix of cardiomyocytes. Previously, we have shown that only miR-181c, and not other family members (miR-181a/b/d), translocates into the mitochondrial compartment of cardiomyocytes [27]. More importantly, miR-181c directly binds to the 3’-end of mt-COX1, a mitochondrial genomic transcript, and regulates mitochondrial function [28, 29]. An in vivo study shows that miR-181c overexpression can lead to heart failure [28], and mice which are lacking miR-181c are protected against ischemia/reperfusion (I/R) injury [27]. Therefore, the functional effect of miR-181c overexpression is well established, but the pathophysiological conditions that lead to miR-181c upregulation in cardiomyocytes are unknown.

A very recent study [30] highlighted the concept that the ultimate effect of miR-181c upregulation in the heart is due to increased [Ca²⁺]_m. Sp1 and MICU1 are the key downstream targets of miR-181c in cardiomyocytes [30]. Sp1 belongs to the zinc finger protein family[31] and regulates the expression of multiple genes, including MICU1 [24]. It has been shown that oxidative stress influences Sp1 activity by post-translational modification, especially at the cysteine residues present in the zinc finger region [32, 33]. Therefore, ROS can regulate Sp1 target gene expression at the transcriptional level by downregulating Sp1 level via post-translational modification. It is to be noted that even though miR-181c overexpression can cause downregulation of both Sp1 and MICU1 protein, this is not due to direct binding of the “seed sequence” of miR-181c to the 3’-UTR of Sp1 or MICU1 mRNAs. Rather, overexpression of miR-181c affects Sp1 and MICU1 levels through ROS production by directly binding to mt-COX1 [30]. In this study, we identified a microRNA in the heart, miR-181c that regulates Sp1 activity via oxidation in the context of high-fat diet (HFD)-induced obesity. This cysteine modification on Sp1 transcriptionally inhibits MICU1 expression.

2. Methods

2.1. Animals

miR-181c/d knockout (KO) mice (miR-181c/d^−/−) were generated on the BL6 background [27]. Male C57BL6/J mice (Jackson Laboratories, Bar Harbor, ME) were used as wild-type (WT) controls for male miR-181c/d^−/− mice. Additionally, we used MCU^flox/flox mice [16]. Where indicated, mice were fed normal laboratory chow (N Chow; 2018 Teklad, Harlan, Frederick, MD) or a purified HFD (60% fat, D12492, Research Diets, New Brunswick, NJ). All mice were fed the N Chow diet from birth to 7 weeks of age. Then subsets of the WT and miR-181c/d^−/− mice were switched to the HFD, while the rest remained on N Chow, generating four experimental groups (n=5–7 per group). Mice were maintained on their respective diets, with food and water provided ad libitum, and body weight was monitored for the remaining experimental period. Mice were treated humanely, and the Institutional Animal Care and Use Committee at the Johns Hopkins University approved all experimental procedures. All the protocols with animal handling in this study were used under the NIH Guide for the Care and Use of Laboratory Animals.

The animals were anesthetized with ketamine (90 mg/kg body wt) / xylazine (10 mg/kg body wt) as an intraperitoneal cocktail, and each animal was anticoagulated with heparin sodium (500 IU/kg body weight, i.v. injection) (Elkin-Sinn Inc., Cherry Hill, NJ). Mouse hearts were excised, and immediately placed in ice cold PBS to wash away the blood.

2.2. Neonatal mouse ventricular myocyte (NMVM) isolation, culture, and transfection

Heart tissues from WT (C57BL6/J) and MCU^flox/flox mouse [16] pups at postnatal day 0 to 3 (P₀–P₃) were dissociated with the Neonatal Heart Dissociation Kit (Miltenyi Biotec, Auburn, CA). NMVMs were isolated from the heart cell suspension by magnetic labeling of non-cardiomyocytes with the Neonatal Cardiomyocyte Isolation Kit (Miltenyi Biotec) according to the manufacturer’s instructions. Isolated NMVMs were cultured on fibronectin-coated coverslips, supplemented with Medium 199 containing 2% fetal bovine serum (FBS) and 100 U of penicillin G/mL, and kept in a humidified incubator at 37C with 5% CO₂. When they reached 80% confluence, plated NMVMs were transfected with the mature miR-181c using Lipofectamine RNAiMAX Transfection Reagents (P/N56531, Invitrogen by Life Technologies) for 48 h. NMVMs were isolated from MCU^flox/flox pups, and CRE was activated with adenovirus packaged with Recombinase Cre (Cat# 1700, Vector Biolabs) for 5 days. This adenovirus infection results in loss of MCU expression. During the last 48 h of the 5-day adenovirus treatment, we also transfected cells with either scrambled sequence or miR-181c mimic (Qiagen, Valencia, CA).

2.3. Lipid-load

We cultured the myoblast H9c2 cell line at early passage (p10 to p20) in Dulbecco’s modified Eagle’s medium supplemented with 10% FBS and 100 units/mL penicillin at 37°C in humidified air with 5% CO₂. We loaded lipid into the H9c2 cells by adding 100 μM oleic acid-albumin (Cat#O3008, Sigma, St. Louis, MO) into the cell culture media for 48 h. As a control, we used 100 μM bovine serum albumin (BSA). After 8 h of lipid loading, we transfected the cells for 48 h with either scrambled sequence (Scr) or AntagomiR-181c (Qiagen) using Lipofectamine RNAiMAX Transfection Reagents (Cat# P/N56531, ThermoFisher Scientific) [34].

2.4. RNA isolation

Total RNA was extracted from whole hearts, from mitochondria isolated from the heart tissue, and from NMVMs, using an miRNeasy kit (Qiagen) and RNase-free DNase kit (Qiagen) [29]. To characterize the integrity of the isolated RNA, we performed spectrophotometric evaluation using a Nanodrop (ThermoScientific, Wilmington, DE). All RNA samples with an A₂₆₀ (absorbance at 260 nm) value >0.15 were used for additional experiments. The ratio of A₂₆₀/A₂₈₀ was also used to evaluate the purity of the isolated RNA.

2.5. qRT-PCR

An miScript Reverse Transcription Kit (Qiagen) was used to reverse transcribe RNA. PCR was carried out with an miScript SYBR green PCR kit (Qiagen) for miRNA, and PowerUp SYBR Green Master Mix (ThermoFisher Scientific) for mRNA. PCR products were detected with a QuantStudio 5 Real-Time PCR System (ThermoFisher Scientific). All reactions were repeated in triplicate. miRNA primers, miR-181a, miR-181b, miR-181c, and miR-181d were purchased from Qiagen. β-Actin (Fwd: GGGCTGTATTCCCCTCCATCG and Rev: CCAGTTGGTAACAATGCCATG) was used to normalize the MICU1 (Fwd: AACAGCAAGAAGCCTGACAC and Rev: CTCATTGGGCGTTATGGAG) expression.

2.6. ROS production assay

Hydrogen peroxide production in isolated mitochondria was measured fluorometrically by quantifying oxidation of Amplex Red to fluorescent resorufin (Life Technologies, Carlsbad, CA). Isolated heart mitochondria, NMVM, or H9c2 cells were incubated in buffer containing 120 mM KCl, 1 mM EGTA, 5 mM MOPS, 5 mM K₂HPO₄ (pH 7.25), 50 μM Amplex Red, and 5 U/mL of horseradish peroxidase. The increase in fluorescence at an excitation of 544 nm and an emission of 590 nm was monitored. Standard curves were generated with known amounts of hydrogen peroxide [35].

2.7. PDH activity

The catalytic activity of PDH in NMVMs, H9c2, or mitochondria isolated from heart tissue was determined by the PDH Enzyme Activity Microplate Assay Kit (Cat# ab109902, Abcam, Cambridge, MA) according to the manufacturer’s instructions.

2.8. Measurement of protein oxidation at cysteine residues by modified biotin-switch method

We assessed cysteine oxidation in whole heart homogenates and NMVM lysates using a modified biotin switch assay as described previously [36]. Briefly, whole heart protein extracts were generated via homogenization in non-reducing buffer (HEN: 250 mmol/L HEPES, 1 mmol/L EDTA, 0.1 mmol/L Neocuproine) containing a protease and phosphatase inhibitor cocktail (Cat#11836170001, Roche, Indianapolis, IN). Protein levels were quantified with the Bradford assay. Protein samples (500 μg) were then diluted in HEN buffer with 2% sodium dodecyl sulfate (HENS) and 20 mmol/L N-ethylmaleimide (NEM, Cat#04259–5G, Sigma-Aldrich, St. Louis, MO) for 40 min at 50°C in the dark with frequent vortexing to block unmodified cysteine residues from modification. Excess NEM was removed via acetone precipitation for 20 min. The protein pellet was re-suspended in HENS buffer containing 10 mM DTT for 45 min in the dark to reduce oxidized cysteine residues. Freshly prepared biotin polyethyleneoxide iodoacetamide (BIAM, Cat#B2059, Sigma-Aldrich; 2 mmol/L) was then added and incubated for 2 hrs in the dark at room temperature to label DTT-reduced cysteine residues. BIAM-labeled proteins were then enriched by adding NeutrAvidin agarose resin (Cat#29201, Thermo Fisher Scientific) and mixing in a rotating device for 2 hrs at 4°C. Captured proteins were then washed with HEN buffer. The proteins were separated from the beads by boiling the samples at 95°C for 5 min with NuPage sample loading buffer (Cat#NP0007, ThermoFisher Scientific) containing 10 mol/L urea. The labeled protein samples were separated by their molecular weight in NuPage gels for western blot analysis.

2.9. Western blot analysis

Mitochondrial, nuclear, and cytosolic fractions of mouse heart were separated by differential centrifugation [35]. Whole heart homogenate, subcellular fractions, or NMVM lysates were lysed with RIPA buffer containing protease and phosphatase inhibitor cocktail, and protein content was measured with a Bradford assay. Protein samples and molecular weight standards were separated by 1D gel electrophoresis. Proteins were transferred to polyvinylidene difluoride membranes, which were blocked for 1 hrs in 5% nonfat dry milk dissolved in Tris-buffered saline (pH 7.4) with 1% TWEEN 20 (TBS-T). Membranes were incubated in TBS-T with 5% BSA or nonfat dry milk at 4°C overnight with antibodies to MCU (Cat# ab121499), PDH (Cat#ab110330), and α-tubulin (Cat#ab7291) from Abcam (Cambridge, MA); Sp1 (Cat#AV37192 from Sigma-Aldrich); MICU1 (Cat#12524S), Histone H3 (Cat#9715), and voltage-dependent anion channel (VDAC, Cat#4866S) from Cell Signaling Technologies (Danvers, MA); and BNP (Cat#PA5–96084), ANP (Cat# 702539), MYH7/β-MHC (Cat# MA532986) from Thermo Fisher Scientific. Membranes were incubated with secondary antibody with the appropriate horseradish peroxidase–conjugated IgG in TBS-T with 5% nonfat dry milk for 2 hrs at room temperature. Immunoreactive proteins were visualized using an enhanced chemiluminescence reagent kit (Cat#34577, Super Signal West pico PLUS, ThermoFisher Scientific). The signals emitted for chemiluminescence were detected with the iBright™ FL100 Imaging system (ThermoFisher Scientific), and band density was analyzed with iBright software.

2.10. Statistics

The results are presented as mean and standard error of the mean (mean ± SEM). Statistical significance was determined by Student’s t test for comparison of two groups or one-way ANOVA for comparison of more than two groups. p < 0.05 was defined as statistically significant.

3. Results

3.1. Loss of miR-181c protects the heart from consequences of a HFD

To determine whether obesity causes miR-181c upregulation in the myocardium and might play a pathophysiologic role, we placed randomly selected C57BL/6 mice on a HFD. We observed a significant upregulation of miR-181c in the heart after 18 weeks of HFD (Fig. 1A). Interestingly, HFD hearts exhibited no changes in the expression of other miR-181 family members (miR-181a, miR-181b, or miR-181d) when compared to those of mice fed normal chow (Fig. 1A). Additionally, there is a 4-fold increase in miR-181c expression in the mitochondrial fraction after 18 weeks of the HFD compared to normal chow (N Chow) (Fig. 1B). This further validates our previous finding that miR-181c mainly localizes to the mitochondrial compartment of cardiomyocytes [29]. To better understand the extent to which miR-181c influences obesity-induced cardiac dysfunction, we fed the HFD to miR-181c/d^−/− mice. It has been reported previously that it takes 16 months to appreciate cardiac dysfunction in the HFD model [37]. Indeed, we observed no changes in cardiac function by echocardiography of WT or miR-181c/d^/- mice, regardless of diet. Additionally, there were no cardiac functional differences between WT and miR-181c/d^−/− with or without the HFD (Suppl Fig. S1). However, we have observed a significant increase in cardiac hypertrophy (as measured by the ratio of heart weight to tibia length) with 18 weeks of the HFD (Fig. 2A). We also corroborated the induction of molecular hypertrophic cardiac markers, such as B-type natriuretic peptide (BNP), atrial natriuretic peptide (ANP), or β-myosin heavy chain (β -MHC). There is no difference in BNP (Fig. 2B), ANP (Fig. 2C) or β -MHC (Fig. 2D) levels in miR-181c/d^−/− (c/d KO) animals with/without HFD. However, all three hypertrophic markers (BNP, ANP and β -MHC) are significantly higher in WT mice fed the HFD compared to WT mice fed normal chow (Fig. 2B, 2C, and 2D). Others have shown that in rodents, cardiac hypertrophy occurs early with a HFD [38]. Interestingly, despite similar weight gain by miR-181c/d^−/− and WT mice during the HFD (Suppl Fig. S2), there was no evidence of cardiac hypertrophy in miR-181c/d^−/− mice fed a HFD as compared to WT mice fed a normal diet, suggesting that mice lacking miR-181c were protected from the cardiovascular complications of the HFD.

Figure 1. Effect of high fat diet on cardiac mitochondrial miR-181c expression.

Quantitative PCR (SYBR) was used to assess (A) miR-181 expression in heart tissues, and (B) miR-181c expression in the mitochondrial fraction of heart tissues, from mice after 18 weeks of being fed normal chow (N Chow) or the 60% high fat diet (HFD). SNO-RD61 was used to normalize miR-181-a/b/c/d expression from the whole heart homogenate, and mitochondrial gene-encoded 12S rRNA was used to normalize the miR-181c expression from isolated mitochondrial fraction. Finally, miRNA expressions were normalized to corresponding N Chow groups (n=6–7).

Data were analyzed by 2-way ANOVA with repeated measures. *p<0.05 vs. N Chow, and **p<0.001 vs. N. Chow. (n=5).

Figure 2. miR-181c plays an important role in obesity-induced cardiac hypertrophy.

(A) Hypertrophy was measured as the ratio between the wet weight of the whole heart and the corresponding tibia length of each mouse (n=8–9). Western blot analysis of hypertrophic marker (B) B-type natriuretic peptide (BNP) (upper bands), (C) atrial natriuretic peptide (ANP) (upper bands), and (D) β-myosin heavy chain (β -MHC) (upper bands) expression in the total heart homogenate from WT and c/d KO mice fed a normal chow or HFD for 18 weeks. α-tubulin (lower bands in all three panels) was used to normalize BNP, ANP and β-MHC expression (n = 5–7).

Data were analyzed by 2-way ANOVA *p<0.05 vs. WT-N Chow, †p<0.001 vs. WT-HFD.

Numerous studies have shown that mitochondria produce ROS as obesity progresses, consistent with a role for mitochondrial ROS in the pathogenesis of type 2 diabetes [39–41]. As shown in Fig. 3A, both WT and miR-181c/d^−/− mice exhibited significantly elevated cardiac ROS generation under HFD conditions, but after 18 weeks, miR-181c/d^−/− hearts produced markedly less ROS than did WT hearts.

Figure 3. Lack of miR-181c can protect cardiac mitochondria from consequences of diet-induced obesity.

(A) Amplex red assay was performed to measure mitochondrial ROS production (n=5–7). (B) pyruvate dehydrogenase (PDH) enzyme activity was measured in the mitochondrial fractions of WT and c/d KO mice after the 18-week diet regimen (n=5–7). (C) Western blot analysis of PDH expression in the total heart homogenate from WT and c/d KO mice fed a normal or HFD for 18 weeks. PDH (upper bands) expression was normalized to α-tubulin (lower bands). A.U., arbitrary units (n = 5–7).

Data were analyzed by both 2-way ANOVA, and two-tailed unpaired t-test. *p<0.05 vs. WT-N Chow; **p<0.001 vs. WT-N Chow; †p<0.05 vs. WT-HFD.

Increased [Ca²⁺]_m is another hallmark of the obese heart [42]. We have shown that miR-181c overexpression in the heart, to an extent similar to that observed in Fig. 1A, leads to increased [Ca²⁺]_m [27, 28]. Therefore, we measured PDH activity in mitochondrial pellets of each of the four groups as a surrogate index of [Ca²⁺]_m. In both WT and miR-181c/d^−/− mice, PDH activity was significantly higher after the HFD than after the normal diet (Fig. 3B). Interestingly, HFD-fed miR-181c/d^−/− mice had markedly less PDH activity than did the WT HFD-fed group (Fig. 3B) despite similar PDH expression (Fig. 3C). These data suggest that lack of miR-181c can substantially protect the heart from the obesity-induced rise in [Ca²⁺]_m.

3.2. Molecular pathway by which lowering miR-181c protects the heart from the adverse effects of obesity

Herein, we will be studying the molecular consequences of miR-181c upregulation due to lipid-load and finally the mechanisms by which upregulated miR-181c alters cardiomyocyte function. We chose the H9c2 cell line (myoblast from early passage number) and NMVMs for these studies because these cells can be cultured several days, allowing time to manipulate each of the target genes. After altering target gene expression, we can examine the downstream consequences.

As shown previously, miR-181c overexpression can lead to mitochondrial dysfunction by causing excessive ROS production, which ultimately leads to heart failure [28, 29]. We wanted to determine the precise mechanism by which miR-181c influences cardiac dysfunction during obesity. A HFD can have multiple effects in vivo, including immune cell activation, which might influence cardiac function. To investigate the role of miR-181c in myoblast cells under simulated obesity conditions, we utilized an in vitro model wherein oleate was added to the media of H9c2 cells (lipid-load). Figure 4 shows qPCR data indicating that increasing the lipid load upregulated miR-181c by 30 hrs (Fig. 4A), and downregulated MICU1 expression after approximately 38 hrs (Fig. 4B). These data suggest that miR-181c downregulates MICU1 indirectly through one or more signaling pathways.

Figure 4. In vitro obese condition induces miR-181c expression in the heart.

Quantitative PCR (SYBR) was performed with the RNA fraction of H9c2 cells after 24, 30, 38, and 48 hrs of lipid-load for (A) miR-181c expression and (B) MICU1 expression. SNORD61 and β-actin were used to normalize miR-181c and MICU1 expression, respectively.

Data were analyzed by 2-way ANOVA with repeated measures, n = 9–12/group, *p<0.05 vs. BSA-treated group.

Based on our finding that miR-181c overexpression can stimulate mitochondrial ROS production and increase [Ca²⁺]_m, we wanted to determine the sequence of events during lipid-load conditions. Amplex red data showed that ROS production was significantly elevated beginning at 30 hrs of high lipid-load (Fig. 5A), as soon as miR-181c expression was significantly elevated (Fig. 4A); whereas, PDH activity was significantly elevated only after 48 hrs of lipid-load (Fig. 5B), when MICU1 expression is significantly downregulated (Fig. 4B). Taken together, these data suggest that miR-181c overexpression first stimulates mitochondrial ROS production but that the increase in [Ca²⁺]_m follows MICU1 downregulation. In order to demonstrate that miR-181c increases [Ca²⁺]_m through a ROS-dependent pathway, we treated with AntagomiR-181c, which specifically scavenges miR-181c, in vehicle control and lipid-load conditions (Suppl Fig. S3). This prevented lipid-load–induced ROS production in H9c2 cells and the rise in [Ca²⁺]_m at both time points (Fig. 5). Hence, these sequential changes indicate that lipid loading first induces overexpression of miR-181c, which then triggers the mitochondrial ROS production. Excessive ROS production transcriptionally inhibits MICU1 expression. Downregulation of MICU1 is the key event in the miR-181c-induced increase in [Ca²⁺]_m. To validate that miR-181c triggers ROS production by downregulating cytochrome c oxidase subunit 1 (mt-COX1) and that ROS facilitates mitochondrial Ca²⁺ influx by targeting the MCU complex, we isolated NMVMs from MCU^flox/flox mouse pups and infected the NMVM with Cre adenovirus for 5 days to knock down MCU. Western blot confirmed substantial reduction in MCU expression under Cre adenovirus treatment (Fig. 6A). As expected, in MCU deficient NMVMs, there were no differences in [Ca²⁺]_m between miR-181c overexpression and Scr, and neither of these differ from MCU containing control as determined by PDH activity, an indirect measure of [Ca²⁺]_m (Fig. 6B). However, 48 hrs after miR-181c overexpression, ROS production in MCU^−/− NMVMs was significantly higher than that in WT NMVMs, suggesting that miR-181c triggers ROS production independent of any change in [Ca²⁺]_m (Fig. 6C). This finding is consistent with our data showing that the increase in ROS precedes the rise in [Ca²⁺]_m.

Figure 5. Loss of miR-181c protects H9c2 cells from lipid-load through a mitochondrial pathway.

(A) The Amplex red assay was performed to measure the ROS production rate in H9c2 cells cultured with a high lipid load or bovine serum albumin (BSA, control) at 30 and 48 hrs after transfection with either scramble sequence or AntagomiR-181c (n=6–8/group). (B) The pyruvate dehydrogenase (PDH) activity assay was performed as a measure of [Ca²⁺]_m in H9c2 cells grown under conditions of lipid-load or BSA at 30 and 48 hrs after transfection with either scramble sequence or AntagomiR-181c (n=5–6/group).

Data were analyzed by 2-way ANOVA with repeated measures, *p<0.05 vs. BSA; #p<0.05 vs. BSA-treated Scr; †p<0.05 vs. AntagomiR-181c.

Figure 6. miR-181c generates [Ca²⁺]_m overload through mitochondrial ROS production.

NMVMs were isolated from MCU^flox/flox mouse pups. (A) Western blot analysis was performed to compare the expression of MCU in MCU^flox/flox-NMVMs infected with the Cre adenovirus (MCU^−/−) to that in non-infected controls (MCU^fl/fl). Next, NMVMs from the MCU^flox/flox Cre-infected group were transfected with either scramble sequence (Scr) or miR-181c (miR-181c overexpression [OE]) for 48 hrs. (B) Pyruvate dehydrogenase (PDH) activity (surrogate measure of [Ca²⁺]_m) was measured in the transfected NMVMs. Data presented in panel B were analyzed by both 2-way ANOVA, and two-tailed unpaired t-test, n=5–6/group. (C) Amplex red assay was used to measure mitochondrial ROS in the transfected NMVMs.

Data were analyzed by 2-way ANOVA with repeated measures, n=7–8/group. **p<0.001 vs. WT and control (not transfected; MCU^fl/fl); #p<0.0001 vs. Scr.

3.3. Increased ROS production after miR-181c overexpression increases Sp1 oxidation

Western blot analysis showed that Sp1 is present in the mitochondrial fraction of heart tissue (Fig. 7A). Studies have suggested the transcription factors can be redox regulated [43–45]. Therefore, we hypothesized that increased ROS from miR-181c overexpression might reduce Sp1 activity by oxidizing cysteines. Figure 7B shows that oxidation of Sp1 cysteine was significantly increased when miR-181c was overexpressed in NMVM cells. This effect was blocked by treating the miR-181c-overexpressing NMVMs with a mitochondrial-targeted antioxidant, Mito-TEMPO. These results provide a potential mechanism for the transcriptional inhibition of MICU1 that has been reported with miR-181c overexpression [46].

Figure 7. miR-181c overexpression-induced ROS production leads to oxidation of mitochondrial Sp1 at cysteine residues.

(A) Nuclear (Nuc), mitochondrial (Mito), and cytoplasmic (Cyto) proteins were isolated from C57BL6 mouse hearts by subcellular fractionation. Western blots were probed with anti-Sp1, anti-histone 3 (His3; nuclear marker), anti-voltage-dependent anion channel (VDAC; mitochondrial marker), and anti-α-tubulin (cytosolic marker). (B) Sp1 oxidation at the cysteine residue (Ox-Sp1) was measured in NMVMs that were isolated from C57BL6 pups, transfected with either scramble sequence (Scr) or miR-181c (miR-181c overexpression [OE]), and treated with or without Mito-TEMPO (25 μM for 48 h). Total Sp1 expression was used to normalize Ox-Sp1. miR-181c overexpression increased Sp1 oxidation, which was significantly diminished with Mito-TEMPO treatment.

Data were analyzed by 2-way ANOVA with repeated measures, n=4–5/group, *p<0.05 vs. Scr without Mito-TEMPO; †p<0.05 vs. miR-181c OE without Mito-TEMPO treatment.

3.4. miR-181c overexpression-induced Sp1 oxidation increases [Ca²⁺]_m levels through MICU1 downregulation in obesity

Next, we probed the underlying mechanism by which miR-181c downregulates MICU1 during obesity. We hypothesized that obesity-induced miR-181c overexpression causes ROS production, which oxidizes mitochondrial Sp1 at cysteine residues. This oxidation might inhibit Sp1 activity and thereby transcriptionally inhibit MICU1 expression. In accordance with previous findings (Fig. 7B), a HFD significantly increased Sp1 oxidation at cysteine residues in both WT and miR-181c/d^−/− mice (Fig. 8A). This Sp1 oxidation was significantly less in HFD miR-181c/d^−/− animals than in HFD WT animals (Fig. 8A), despite the fact that miR-181c/d^−/− mice had significantly higher levels of Sp1 protein (Fig. 8B). The association between Sp1 oxidation and lower levels of Sp1 protein suggests that oxidation enhances degradation, which reduces Sp1 activity. Western blot analysis also showed that obesity significantly lowered MICU1 expression in the heart but that MICU1 expression in the HFD miR-181c/d^−/− group was significantly higher than that of the HFD WT group, and not significantly different from WT control (Fig. 8C).

Figure 8. Mitochondrial Sp1 oxidation plays an important role in obesity-induced [Ca²⁺]_m overload.

Two groups of animals, C57BL6 (WT) and miR-181c/d^−/− (c/d KO), were fed either normal chow (N Chow) or a 60% high-fat diet (HFD) for 18 weeks. (A) Total heart homogenate was used to measure Sp1 oxidation at the cysteine residue (Ox-Sp1) by Western blot. Total Sp1 expression was used to normalize Ox-Sp1. (B, C) Western blot analysis of Sp1 and MICU1. α-Tubulin was used to normalize Sp1 and MICU1 expression (n=5–7/group).

Data were analyzed by 2-way ANOVA with repeated measures, *p<0.05 vs. WT-N Chow; †p<0.001 vs. WT-HFD

4. Discussion

A large number of studies have shown a role for [Ca²⁺]_m in the pathophysiology of diabetic cardiomyopathy [24, 47]. In one recent study, heart biopsies from diabetic patients had less MICU1 expression than did healthy controls [24]. MICU1 has been shown to change MCU to the closed conformation at low [Ca²⁺]_cyto [48–50] and contribute to the open conformation at high [Ca²⁺]_cyto [49]. Alterations in the ratio of MICU1 to MCU lead to changes in calcium uptake threshold and cooperativity [51]. MICU1 was the first regulatory MCU complex member identified and was shown to play an important role in [Ca²⁺]_m responses [52]. Several studies have focused on MCU and MCU regulatory proteins in the cardiovascular field [16, 19, 49, 50, 53–57]. However, the mechanisms by which MICU1 expression is regulated in the heart under physiologic and pathophysiologic conditions are not fully understood. In this study, we report several key findings in the pathophysiology of obesity-induced cardiac dysfunction. Previously we had demonstrated that miR-181c overexpression leads to cardiac dysfunction [27–29]. Here, we show that with obesity and type 2 diabetes, cardiomyocytes upregulate miR-181c expression, and this miR-181c accumulates in the mitochondrial compartment (Fig. 1A-B). We also demonstrated upregulation of cardiac miR-181c, which increases [Ca²⁺]_m through MICU1 downregulation during obesity. More importantly, this study highlights the underlying mechanism by which upregulation of miR-181c causes downregulation of Sp1 in the heart by post-translational cysteine modification. Lower Sp1 in the heart in response to obesity is responsible for MICU1 downregulation (Obesity→miR-181c↑→mt-COX1↓→ROS↑→Sp1Ox↑→Sp1↓→MICU1↓→[Ca²⁺]_m↑).

Previous studies showed that miR-181c overexpression created a “bottle-neck” in electron transport by directly binding to the 3’-end of mt-COX1 (a complex IV subunit) and reducing COX1 expression [27–29]. Inhibition of the electron transport chain at complex IV causes electrons to accumulate upstream, and that might explain how miR-181c overexpression can stimulate ROS production. Consistent with this concept, we identified a significant increase in mitochondrial ROS with HFD in both mouse genotypes, but ROS formation was significantly less in the mitochondria of miR-181c/d^−/− hearts than in those of WT hearts. Our study suggests that lower mitochondrial ROS and decreased [Ca²⁺]_m in the miR-181c/d^−/− animals protects them from the HFD stress.

ROS derived from mitochondria play an important role in the pathogenesis of heart disease [8–10], including cardiomyopathy induced by obesity/diabetes [39–41]. As mitochondrial ROS levels change, multiple cellular and physiologic functions are affected, contributing to the development of common obesity-related cardiac complications [41]. Protein oxidation is one major consequence, as it can alter cellular function by oxidative stress [58, 59]. Several side chain residues, such as cysteine, methionine, and tyrosine, are oxidized in response to cellular oxidative stress. Therefore, protein oxidation can serve as a potent biomarker of oxidative stress induced by prediabetes and diabetic conditions [60].

Transcription factor Sp1 is a DNA-binding protein with three zinc-fingers (Cys2His2–type zinc finger) that are required for recognizing GC-rich promoter sequences [61, 62]. Some reports highlight that ROS can affect the binding of Sp1 to DNA by modifying cysteine residues that are crucial for Sp1 transcriptional activity [32, 33, 63, 64]. It has been shown that Sp1 can regulate MICU1 expression by directly binding to the promoter region [24]. Here, we showed that Sp1 is present in the mitochondrial compartment of cardiomyocytes. Therefore, we hypothesized that Sp1 can be oxidized by ROS generated from miR-181c overexpression. Indeed, we found that ROS generated as a result of miR-181c overexpression in NMVMs is responsible for Sp1 oxidation at cysteine residues. Additionally, treatment with a mitochondrial-targeted antioxidant (Mito-TEMPO) in the setting of miR-181c overexpression reduced Sp1 oxidation, suggesting the potential role of ROS in Sp1 oxidation (Fig. 7B). Using miR-181c/d^−/− mice with/without the HFD, we confirmed that miR-181c induces Sp1 oxidation through oxidative stress and plays an important role in increasing [Ca²⁺]_m under obese conditions. Sp1 is present in all cellular components of cardiomyocytes (Fig. 7A). Whether Sp1 is oxidized in the mitochondria and then exits to the cytosol, or whether ROS generated in the mitochondria is released into the cytosol where it reacts with Sp1 (or a combination of both) the net effect is Sp1 oxidation, which affects MICU1 expression.

Nuclear-to-mitochondrial communication is a well-studied process whereby products of nuclear-encoded genes translocate into the mitochondrial compartment and alter mitochondrial function. However, there is another arm to this communication: mitochondria-to-nucleus retrograde signaling. This retrograde signaling plays an important role in the maintenance of cellular homeostasis and is poorly studied [65]. Previously, it has been highlighted that a miRNA in cancer cells can control mitochondria-to-nucleus retrograde signaling to promote tumor growth [66]. Multiple transcription factors, including Sp1, are regulated by altered mitochondrial bioenergetics. These transcription factors then move back to the nucleus. Thus, altered mitochondrial bioenergetics can influence nuclear gene expression. Some of the altered nuclear-encoded genes are responsible for the coding of mitochondrial proteins. Our study is a prime example of the importance of cellular bidirectional nuclear-mitochondrial communication. Stress from the HFD influences miR-181c expression in the cardiomyocytes. Mature miR-181c translocates into the mitochondrial compartment [28, 29] as part of the nucleus-to-mitochondria anterograde signal. Upregulation of miR-181c alters mitochondrial function by enhancing mitochondrial ROS production. This increased ROS is responsible for Sp1 oxidation-induced degradation, which, in turn, leads to downregulation of MICU1 expression. This pathway explains how damped mitochondria-to-nucleus retrograde signaling inhibits MICU1 transcription. Ultimately, MICU1 is not available to regulate mitochondrial Ca²⁺ entry.

In conclusion, this study has identified a unique nuclear-mitochondrial communication mechanism in the heart orchestrated by miR-181c, which is important in obesity-induced diabetes. During obesity, overexpression of miR-181c increases ROS production by targeting the mitochondrial gene mt-COX1. The excess ROS oxidizes cysteine residues on Sp1, which regulates MICU1 expression in the heart. MICU1 is a key regulator of [Ca²⁺]_m during obesity/type 2 diabetes stress. We confirmed the role of miR-181c in this critical signaling pathway by downregulating miR-181c in vitro, using AntagomiR-181c, and in vivo, using miR-181c/d^−/− mice. In both models, our data showed significant cardioprotection during obesity when miR-181c expression was reduced. Together, our data provide strong evidence that lowering miR-181c in the heart may have therapeutic potential for obesity-induced cardiac pathologies.

Study Limitations

We have utilized several approaches to test our hypothesis in this study, including miR-181c overexpression, lipid-load into myoblast cell line (H9c2), neonatal mouse ventricular cardiomyocytes (NMVM), miR-181c/d^−/− mice, high-fat diet (HFD) model for WT and miR-181c/d^−/− mice, and an antagomiR-181c in an in vitro model. Each model has its own limitations, such as in H9c2 the lower number of mitochondria compared to adult cardiomyocytes or the metabolic profiles of NMVM is so different than that of adult cardiomyocytes; however, using all of these models in this study collectively helped us to validate this novel mechanism. Together, it allowed us to fully understand the mechanisms by which diet-induced obesity can influence mitochondrial Ca²⁺ overload, indirectly measured using PDH activity, by upregulating miR-181c in the heart.

• ➢Diet-induced obesity upregulates miR-181c expression in the heart.
• ➢A high fat diet leads to upregulated miR-181c, which translocates into the mitochondria.
• ➢miR-181c activates Sp1 degradation by post-translational modification at cystine residue(s).
• ➢Upregulation of miR-181c causes transcriptional inhibition of the MICU1 gene.

Non Standard Abbreviations and Acronyms

miRNA:

microRNA

ROS:

Reactive oxygen species

miR-181c/d^−/− mice:

miR-181c and d knock-out mice

HFD:

High-fat diet

[Ca²⁺]_m:

Mitochondrial matrix calcium

mt-COX1:

Cytochrome c oxidase subunit 1

Sp1:

Specificity protein 1

MCU:

Mitochondrial calcium uniporter

MICU1:

Mitochondrial calcium uptake 1 protein

NMVM:

Neonatal mouse ventricular myocytes

PDH:

Pyruvate dehydrogenase