Introduction
Diabetes mellitus is one of the major public health burdens in both developed and developing countries. The global prevalence of diabetes is predicted to be 4.4% in 2030 [
1]. Diabetic patients frequently show cardiac abnormalities and even heart failure, which can be partly explained by conventional risk factors, such as hypertension and coronary artery disease (CHD) [
2]. However, multiple epidemiological evidences revealed that there were still 2.4- to 5-fold increase of risk in developing heart failure between patients with diabetes and age/gender matched controls after conventional risk factors adjustment [
3]. Recently, it was found that diabetes can lead to cardiomyocyte injury directly without endothelium dysfunction, named diabetic cardiomyopathy (DCM) [
4]. DCM is defined as a unique form of abnormal ventricular structure and performance that occurs independently of hypertension, CHD or other potential cardiogenic etiology in diabetic patients, and its prevalence is estimated to be between 30 and 60% in preclinical and clinical stages among diabetic population [
5,
6]. DCM is initially characterized by a hidden subclinical cardiac remodeling, and followed by noticeable diastolic dysfunction, later by systolic function damage, and eventually by clinical heart failure [
7]. DCM has been confirmed as the leading cause of hospitalization and mortality in diabetic patients [
5,
8]. Despite advances in molecular etiologies, current therapy approaches failed to prevent long-term cardiac events in patients with diabetes, which stressing the importance of novel therapeutic strategies.
Although there are multifactorial mechanisms involved in DCM, it is well accepted that cardiac metabolic derangement plays a central role in its occurrence and development [
9‐
11]. With omnivorous and metabolically flexible characters, healthy adult hearts use energy from various substrates depending on the circumstances [
12]. Physiologically, fatty acid (FA) contributes around 70% of the energy supply for working heart, whereas the remainder energy source primarily relies on glucose [
12]. However, insulin resistance and hyperlipemia result in marked reduction in the ability to use glucose, and thus cardiomyocytes rely almost exclusively on FA derived energy [
13,
14]. This substrate switch is accompanied by higher rates of oxygen consumption and impaired oxidative phosphorylation that result in reactive oxygen (ROS) overproduction. Concurrently, redundant FA is diverted to cellular neutral lipid pools or converted to toxic intermediates [
15]. Subsequent injuries in mitochondria and other cell components was identified as the main triggers of cardiomyocyte apoptosis, death and consequential contractile dysfunction [
10].
microRNAs (miRNAs) are a class of highly evolutionarily conserved endogenous single-stranded small (~ 22 nt) non-coding RNAs, which participate in numerous biological processes and diseases. miRNAs typically suppress gene expression at post-transcriptional levels by directly binding to the 3′ untranslated region (UTR) of target mRNAs [
16]. Multiple miRNAs are identified to play key roles in the etiology of DCM [
17,
18]. Our previous study also showed that miR-21 protected against diabetic cardiomyopathy induced diastolic dysfunction by targeting gelsolin [
19]. miRNA based therapy for the prevention and treatment of DCM becomes promising in next-generation diabetic care [
18].
In the present study, we found that PGC-1β was upregulated in the heart of db/db mice, and contributed to the cardiac metabolic disturbances. Then, we demonstrated that PGC-1β was regulated by miR-30c in vivo and in vitro. Most importantly, we showed that exogenous miR-30c delivery by recombinant adeno-associated virus system was sufficient to attenuate cardiac metabolic derangement and cardiac dysfunction in db/db mice. Mechanistically, we found that the effects of miR-30c/PGC-1β on DCM were, at least in part, mediated by PPARα signaling pathway.
Research design and methods
Microarray data, identification of DEGs and pathway enrichment analysis
Dataset GSE4745 (Affymetrix platform GPL85) was downloaded from Gene Expression Omnibus (
https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE4745). GSE4745 contained global mRNA expression in heart tissues of control and streptozocin induced type I diabetic rats. The rats were sacrificed at 3 days (baseline), 28 days (diastolic dysfunction) and 42 days (both diastolic and systolic dysfunction) after streptozocin injection, respectively. The analyses were carried out using R project (v3.4.1) with packages from Bioconductor website.
The gene expression matrix was obtained from data of CEL format by background correction, quartile data normalization and annotation. The classical t-test was used to identify different expression genes (DEGs) with a cutoff p-value of 0.05.
KEGG is a knowledge base for systematic analysis of gene functions, linking genomic information with higher-order functional information [
20]. KEGG pathway analyses were performed using DAVID bioinformatics resources and p < 0.05 was considered statistically significant [
21].
Reagents
Cell culture medium (DMEM and OPTI-MEM) and fetal bovine serum (FBS) were purchased from GIBCO (Thermo Fisher Scientific, Waltham, MA). The reverse transcriptional and real-time PCR primers of miR-30c and U6, mimics and inhibitor of miR-30c, siRNAs targeting PGC-1β and PPARα, and their negative controls were from RiboBio (Guangzhou, China). Lipofectamine 2000 and Trizol were from Invitrogen (Thermo Fisher Scientific, Waltham, MA). The primers of mRNA real-time PCR were designed and synthesized by BGI (Shenzhen, China). Antibodies against CD36, CPT1B, PDK4, GLUT4 and β-actin were from ABclonal Biotech (Cambridge, MA). Anti-Ago2 antibody was from Novus Biologicals (Littleton, CO). Antibodies against PGC-1β and PPARα were from Abcam (Cambridge, MA). Polyvinylidene difluoride membranes were from Millipore (Merck KGaA, Darmstadt, Germany). The horseradish peroxidase-conjugated secondary antibody and enhanced chemiluminescence reagents were from Pierce Biotech (Thermo Fisher Scientific, Waltham, MA). pTK-PPREx3-Luc was a kind gift from Dr. Bruce Spiegelman (Addgene plasmid Cat# 1015) [
27]. Other reagents were from Sigma-Aldrich (Merck, St. Louis, MO) unless specified description elsewhere.
Cell culture
H9c2 and HEK293T (American Type Culture Collection, Manassas, VA) were routinely cultured in DMEM medium supplemented with 10% FBS. H9c2 cells were transfected with miRNA mimics/inhibitor (100 nM), siRNAs (100 nM) with Lipofectamine 2000 according to manufacturer’s description, respectively, and then subjected to 1 mM palmitate (bound to BSA at a ratio of 5:1) stimulation for 48 h.
Primary cardiomyocytes were prepared from neonatal rats (day 0–3) and identified by immunofluorescent staining with antibody against cardiomyocyte-specific marker ACTN2 as previously described [
28]. Primary cells were cultured with DMEM supplemented with 10% FBS and treated with 300 µM saturated FFA palmitate (16:0) for 48 h.
Dual luciferase assays
Luciferase activity was detected using the Dual Luciferase Reporter Assay System (Promega, Beijing, China) as described previously [
29].
To evaluate the binding of miRNA and target mRNA, the sequence containing predicted miR-30c binding site of the human PGC-1β 3′ UTR and corresponding mutant sequence (Additional file
1: Table S1) were synthesized and inserted into pMIR-REPORT luciferase vector (Ambion, Thermo Fisher Scientific, Waltham, MA). To determine whether miR-30c targeting 3′ UTR of PGC-1β, HEK293T cells were co-transfected with appropriate pMIR construct (1 μg/mL), pRL-TK plasmid (0.1 μg/mL) with miR-30c mimics or negative controls (100 nM).
To evaluate the effects of PGC-1β or miR-30c on PPARα transcriptional activity, H9c2 cells were first transfected with PGC-1β siRNA or miR-30c, and then transfected with pTK-PPREx3-Luc plasmid (1 μg/mL), respectively [
30]. To determine whether miR-30c regulated PPARα transcriptional activity via PGC-1β, H9c2 cells were transfected with miR-30c and then transfected pTK-PPREx3-Luc plasmid (1 μg/mL) after PGC-1β inhibition by siRNA.
Transfections were done in duplicates and repeated at least three times in independent experiments.
Preparation of rAAV system and animal study
All animal procedures were approved by the Institutional Animal Care and Use Committee of Tongji Medical College of Huazhong University of Science and Technology (HUST), and complied with ARRIVE and the Guide for the Care and Use of Laboratory Animals standards (NIH publication).
The recombinant adeno-associated virus of type-9 (rAAV9) system was a kind gift from Dr. Xiao Xiao (University of North Carolina at Chapel Hill). Synthesized fragments of miR-30c, anti-miR-30c, or miR-random (Additional file
1: Table S2) were inserted into rAAV expression plasmid, respectively. The rAAVs were prepared by triple plasmid co-transfection in human HEK293T cells as described previously [
29].
Male db/db mice with C57BL/Ks background and their control littermates were bred in Model Animal Research Center of Nanjing University and maintained in Laboratory Animal Center of HUST. Db/db mice were randomly assigned to four groups (n = 8–10 for each group). A single tail vein injection of the rAAV-miR-random, rAAV-miR-30c, or rAAV-anti-miR-30c (1 × 10
11 vector genomes in 200 μL PBS per mouse) [
31] was performed at the age of 8 weeks in db/db mice, whereas db/db control and C57BL/Ks mice were treated with equal volume of PBS in the same way. At the age of 28 weeks, echocardiographic and hemodynamic analyses were performed and then mice were sacrificed with anaesthetization by the mixture of xylazine and ketamine. Tissue samples were taken for further study.
Echocardiographic and hemodynamic measurements
Echocardiographic measurements were performed under anesthesia with 1.5% isoflurane using a high-resolution imaging system with a 30-MHz high-frequency scanhead (Vevo770, VisualSonics Inc, Toronto, Canada) as previously described [
29]. Hemodynamic measurements of left ventricle were performed under anesthesia with intraperitoneal injection of the mixture of ketamine and xylazine using a Millar Catheter System (Millar 1.4F, SPR 835, Millar Instruments Inc, Houston, TX) as described previously [
29].
Western blots analyses
Western blotting was performed using specific antibodies as described previously [
29]. The band intensities of target proteins were analyzed by using ImageJ program [
29].
ChIP–PCR
Chromatin immunoprecipitation (ChIP) assays were conducted using a kit from Beyotime (Nanjing, China), following the manufacturer’s protocol [
32]. Primers used for ChIP–PCR were listed in Additional file
1: Table S3.
Real-time PCR
Total RNA was isolated from heart tissues and H9c2 cells with Trizol reagent, and was reverse-transcribed into cDNA using RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA). Real-time PCR assays were performed using SYBR rapid quantitative PCR Kit (Kapa Biosystems, Woburn, MA) [
33]. The relative miRNA levels were determined by normalizing to U6 level.
To assess the mitochondrial biogenesis, the ratio of mitochondrial to genomic DNA was used. Real-time PCR was carried out with different primers (Additional file
1: Table S4). 18S ribosomal DNA was used to represent genomic DNA; 18S ribosomal DNA and COI was used to represent mitochondrial DNA.
Electron microscopy
Fresh specimen of mice hearts was fixed with glutaraldehyde in PBS, dehydrated, coated and then observed by electron microscopy analysis (FEI Tecnai G2, Hillsboro, OR) [
33]. Lipid droplets were counted in ten images from different fields.
TUNEL assay
To determine cell apoptosis in myocardial tissue sections, TdT-mediated dUTP nick-end labeling (TUNEL) assay was performed using a kit of R&D Systems (Minneapolis, MN) according to the manufacturer’s instructions [
33]. Pictures of slides were obtained by using a microscope (Nikon, Tokyo, Japan) equipped with digital imaging camera.
Staining of heart samples
Neutral lipids accumulating in cultured H9c2 cells was stained with BODIPY 493/503 (Invitrogen, Thermo Fisher Scientific, Waltham, MA) and then imaged by the fluorescence microscopy (Nikon, Tokyo, Japan). Myocardial lipids accumulation in mouse heart frozen sections was stained by oil Red O and then imaged by the ordinary optics microscopy as previously described [
32].
To determine intracellular ROS production in myocardial tissue sections, dihydroethidium (DHE, Beyotime, Nanjing, China) staining was performed according to the manufacturer’s instructions [
29]. Pictures of slides were obtained by using a fluorescence microscopy microscope (Nikon, Tokyo, Japan) equipped with digital imaging camera.
Flow cytometry analyses
To evaluate cell apoptosis, H9c2 cells were incubated with Annexin V-FITC/PI (BD Biosciences, Sparks, MD) [
33]. To evaluate intracellular ROS production, H9c2 cells were incubated with DCFH-DA (Beyotime, Nanjing, China) [
29]. After incubation, cells were analyzed with a FACStar Plus flow cytometer (BD Biosciences, Sparks, MD) as previously described [
8].
ATP level determination
The ATP level was determined using an Enhanced ATP Assay Kit (Beyotime, Nanjing, China) according to the manufacturer’s instructions. Briefly, cells were lysed with ATP lysis buffer on ice and centrifuged. The supernatant was then transferred to the ATP working solution and mixed quickly. A luminometer (Promega, Beijing, China) was used to measure the luminescence immediately. The luminescence data were normalized by the total protein levels of individual samples.
Detection of mitochondrial membrane potential
The mitochondrial membrane potential (ΔΨm) was assessed using a JC-1 kit from Invitrogen (Thermo Fisher Scientific, Waltham, MA) according to manufacturer’s description.
Statistical analyses
All data were shown as mean ± SEM unless otherwise noted. Differences between groups were evaluated using Student’s t-test of unpaired data or one-way analysis of variance (ANOVA) and Bonferroni post-test. p < 0.05 was considered statistically significant. All statistical tests were performed using GraphPad Prism (v5.0) and SPSS (v18.0) unless otherwise stated.
Discussion
Our present study found that miR-30c/PGC-1β/PPARα pathway involved in DCM via regulating cardiac metabolism. In the diabetic heart, the expression of miR-30c was repressed. The loss of miR-30c directly led to increased PGC-1β, which enhanced the transcriptional activity of PPARα. The upregulation of PPARα and its target genes in turn contributed to metabolic abnormalities and lipotoxicity in the heart, including enhanced FA uptake and oxidation, reduced glucose utilization, impaired ATP production, and subsequently cardiomyocyte apoptosis and cardiac dysfunction. Moreover, excessive FA uptake in turn repressed miR-30c expression, further promoted the progress of DCM. The protective roles of miR-30c/PGC-1β/PPARα in DCM via regulating cardiac metabolism were summarized in Fig.
6e.
There are indeed several different pathological manifestations of cardiac dysfunction between patients with type 1 and type 2 diabetes. For example, hyperglycemia-induced myocardial fibrosis may be predominant in type 1 diabetes, while cardiomyocyte hypertrophy is more related to the insulin resistance and dyslipidemia of type 1 diabetes [
37]. However, multiple recent studies have shown that the underlying mechanisms and phenotypes of DCM in both types of diabetes mostly overlapped, especially for the cardiac metabolic abnormalities. Both types of diabetes shared similar downstream of a switch to greater cardiomyocyte FA utilization at the expense of glucose [
38]. In animal models, both the STZ-induced and db/db cardiac dysfunction models were characterized by increased fatty acid utilization, decreased glucose utilization, impaired cardiac efficiency and function, compromised mitochondrial energetics, and increased lipid accumulation in the heart [
10]. Moreover, different animal models of type 1 and type 2 diabetes have been used interchangeably in multiple studies to understand the pathophysiological mechanisms of DCM [
39]. Therefore, we analyzed the expression dataset from STZ-induced models, because only these profiles covered three pathological stages of DCM, which may involve different molecular signals. Considering that type 2 diabetes was by far the most common type of diabetes, we investigated db/db mice model in the current study.
Multiple mechanisms are involved in the metabolic abnormalities of the diabetic heart, among which PPAR pathway was considered as an important one [
11]. Previous study showed that mice with cardiac-restricted overexpression of PPARα (MHC-PPARα) exhibited a cardiac metabolic phenotype strikingly similar to diabetic heart, which was evidenced by increased FA utilization, decreased glucose utilization, myocyte lipid accumulation and cardiac dysfunction [
40]. Whereas, excessive FA metabolism in diabetic heart was significantly relieved in PPARα knockout mice [
41]. Previously, other group, as well as ourselves, have showed that inhibition of PPARα could alleviate atherogenic dyslipidemia and DCM [
42,
43]. Consistently, our study identified that FA transport and oxidation associated genes were enriched in PPAR pathway in all three pathological stages of DCM via bioinformatics analysis.
We next investigated the key drivers of PPAR pathway and turned to PPAR co-activator PGC-1β. PGC-1β, together with PGC-1α, was identified as a master regulator of fatty acid oxidation gene expression in stress induced cardiac hypertrophy and heart failure [
44]. Although the effects of PGC-1β on metabolic relative cardiac dysfunction, especially DCM, are hardly known, the upregulation of PGC-1β in diabetic heart has been observed in several animal models [
32,
45]. PGC-1β was further determined as the top hub gene together with PPARα by transcriptome bioinformatics analysis in DCM [
45]. In addition, PGC-1α has been widely accepted as the key contributor to the metabolic disorder in DCM via coactivating PPAR [
10,
46]. PGC-1β shares similar tissue distribution pattern and significant sequence homology with PGC-1α [
47]. In line with these findings, our study revealed that cardiac PGC-1β was significantly increased in DCM, and downregulation of PGC-1β protected cardiomyocytes from glucose-FA use shift and lipotoxicity.
PGC-1β have recently been discovered as powerful regulators of cellular metabolism by interacting with several different transcription factors including estrogen-related receptors (ERRs), nuclear respiratory factors (NRFs), and the PPAR family [
47]. Our current study found that elevated PGC-1β promoted metabolic abnormalities via positively regulating PPARα transcriptional activity. However, whether other transcriptional factors, such as PPARγ and ERRγ, are involved in PGC-1β induced metabolic disorder in the diabetic heart need further study.
Most importantly, we have provided multiple lines of evidences demonstrating that PGC-1β is a target of miR-30c. Overexpression of miR-30c by rAAV system improved glucose utilization, reduced excessive ROS production and myocardial lipid accumulation, and subsequently attenuated cardiomyocyte apoptosis and cardiac dysfunction in vivo and in vitro. Our data revealed the protective roles of miR-30c in cardiac metabolism in diabetes via targeting PGC-1β. Previous studies found that both miR-30c and PGC-1β were involved in ER stress [
48,
49]. Meanwhile, many studies have shown that excessive ER stress, trigger by hyperglycemia, free fatty acids (FFAs) and inflammation, played important roles in the development of DCM [
50]. These evidences indicated that miR-30c/PGC1β may participate in the progression of DCM via ER stress.
We also found that rAAV9 mediated miR-30c delivery partly relieved hyperlipidemia and hepatic steatosis in db/db mice. Indeed, overexpression of miR-30c in liver has been demonstrated to decrease hepatic lipid biosynthesis and lipoprotein [
35]. It is possible that the systemic effects of miR-30c contributed to its protective role in heart in our mice model. In addition, cardiac-restricted overexpression of PPARα developed DCM and fatty livers, indicating that cardiac metabolic remodeling could in turn induce systemic metabolic disorders [
51]. In our current study, overexpression of miR-30c significantly reduced PPARα activity in heart of db/db mice, which might contribute to the remission of concurrent hyperlipidemia and hepatic steatosis.
Additionally, PGC-1β knockdown, as well as miR-30c overexpression, inhibited mitochondrial biogenesis but preserved mitochondrial function and ATP production in NRCMs treated with palmitate. Actually, increased mitochondrial biogenesis together with mitochondrial dysfunction has been reported as a characteristic of DCM [
52]. Mitochondrial biogenesis may be a potentially adaptive mechanism in the face of compromised mitochondrial function and cardiac energetics. Since PGC-1β is a major regulator of mitochondrial biogenesis [
53], it is possible that miR-30c/PGC-1β is involved in regulating this process. However, this has not been established definitively and need further study.
Taken together, our data demonstrated the protective role of miR-30c in cardiac metabolism in diabetes via targeting PGC-1β, and suggested that modulation of PGC-1β by miR-30c may provide a therapeutic approach for DCM.
Authors’ contributions
ZY and YZ designed the study, analysed and interpreted the data, and drafted the paper. MH, HL, JF, XN and MY contributed to data acquisition. CC and DWW designed the study and drafted the paper. All authors read and approved the final manuscript.