Obesity-Associated miR-199a/214 Cluster Inhibits Adipose Browning via PRDM16–PGC-1α Transcriptional Network

The global incidence of obesity and obesity-related disorders, including metabolic syndrome and diabetes, gives rise to a demand for effective therapeutic interventions. Brown and brown-like adipocytes (called beige adipocytes) are emerging as potential targets for the treatment of obesity and related metabolic diseases (1–4). Both brown and beige adipocytes trigger a program of mitochondrial respiration and thermogenesis through induction of UCP1 expression and dissipation of chemical energy to produce heat (5,6). Stimulation of brown and beige fat development leads to increased energy expenditure and a lean, healthy phenotype in neonatal mammals, hibernators, rodents, and adult humans (7,8). However, the mechanisms regulating thermogenic fat cells still need to be further elucidated.

The differentiation and development of brown and beige adipocytes are regulated by multiple transcriptional factors and cofactors such as PRD1-BF1-RIZ1 homologous domain-containing 16 (PRDM16) and peroxisome proliferator–activated receptor α and γ (PPARα and PPARγ) coactivator-1α (PGC-1α) (9–11). PRDM16 is a critical determinant of the brown fat lineage and a key transcriptional regulator of brown fat differentiation. It stimulates differentiation of Myf5-positive myogenic precursor cells into brown fat cells while prohibiting myogenic differentiation by robustly inducing expression of brown adipose tissue (BAT)–selective genes such as UCP1 and PGC-1α (12,13). Increased PRDM16 expression can drive the expression of BAT-selective genes in beige fat cells (14–16). PGC-1α is an important transcriptional coactivator that regulates brown fat thermogenesis (17,18). It interacts with several transcriptional factors and nuclear receptors, thus controlling the entire program of thermogenesis including the transcription of UCP1 and mitochondrial biogenesis (17–19). PGC-1α itself is regulated by a number of well-known factors related to cellular energy and mitochondrial homeostasis, including PPARα and PPARγ (20). PRDM16 positively regulates PGC-1α expression by promoting induction of PGC-1α gene transcription through PPARα (21). However, factors that enable and coordinate the program of concerted cooperation between the transcriptional factors and coregulators necessary for brown and beige adipogenesis remain largely unknown.

miRNAs are small noncoding RNAs that regulate gene expression at the posttranscriptional level. Recent studies suggest that miRNAs are important factors regulating differentiation, development, and function of brown and beige fat cells (for reviews, see refs. 4 and 22). Some miRNAs, including miR-193b/365 (23), miR-196a (24), miR-133 (25), miR-155 (26), and miR-455 (27), have been reported to modulate brown adipocyte differentiation by targeting adipogenic regulators. Additionally, miR-328 mediates the action of Dicer1 to control brown adipocyte differentiation and function (28). Very recently, we reported that miR-30 family members are required to maintain brown adipocyte function and beige fat development (29). However, it remains unclear how miRNAs modulate and integrate multiple signaling networks regulating mitochondrial biogenesis and UCP1 expression, the defining signature of brown adipogenesis.

To identify miRNAs that are important for brown adipocyte differentiation, we performed a miRNA microarray analysis and identified a group of miRNAs that are significantly up- or downregulated during brown adipocyte differentiation. In this study, we show that expression of the miR-199a/214 cluster is dramatically decreased during brown adipocyte differentiation. Reduced expression of the miR-199a/214 cluster is also observed in response to cold exposure or treatment with a β-adrenergic receptor activator. In contrast, adipose miR-199a/214 expression is positively correlated with obesity in mice and humans. Overexpression of the miR-199a/214 cluster suppresses brown adipogenesis and decreases thermogenic gene expression in brown adipocytes. Conversely, knockdown of the cluster increases the expression of Ucp1 and thermogenic genes in cultured beige adipocytes and in subcutaneous white adipose tissue (sWAT) in vivo. We further show that miR-199a/214 directly targets PRDM16 and PGC-1α, demonstrating a mechanism by which this miRNA cluster negatively regulates beiging and thermogenesis. This study highlights the inhibitory roles of the miR-199a/214 cluster in gene cascades involved in brown adipogenesis and the regulation of BAT and beige fat development, therefore uncovering a potential therapeutic target against obesity.

The isolation and culture of the stromal vascular fraction (SVF) from interscapular BAT and inguinal sWAT were performed as previously described (29,30). The BATs or sWATs from 3-week-old male C57BL/6 mice were sliced and digested in HEPES buffer containing type II collagenase (Sigma-Aldrich) and BSA (Gibco). The preadipocytes were plated, cultured, and induced to differentiate according to the procedure previously reported (18,29).

For Oil Red O staining, mature adipocytes were washed twice with PBS, fixed in 4% paraformaldehyde for 1 h, and stained with Oil Red O for 1 h. The cells were visualized and photographed using a light microscope (Olympus, Tokyo, Japan).

For transfection, the duplex oligonucleotide (mimic) or single-stain antisense (inhibitor) designed for miR-199a, miR-214, or the respective nonspecific control (NC) (GenePharma, Shanghai, China) was added to 70–80% confluent cultured cells at a final concentration of 100 nmol/L. A pCDNA3.1 vector expressing N-terminally Flag-tagged PRDM16 was a gift from Dr. J. Lin at the University of Michigan (Ann Arbor, MI). Transfections were performed using Lipofectamine 2000 according to the manufacturer’s protocol.

For treatment, the preadipocytes incubated with fresh DMEM were treated with or without CL-316243 (CL; 20 µmol/L), a selective β₃-adrenergic receptor activator; isoproterenol (8 µmol/L), a nonselective β-adrenergic receptor activator; or forskolin (20 µmol/L), a cellular cAMP inducer. Cells were collected 24 h after the treatment and stored at −80°C for further analyses.

To determine mitochondrial respiration activity, the O₂ concentration was measured using an XF24 extracellular flux analyzer (Seahorse Bioscience, Billerica, MA). Basal mitochondrial respiration was measured in untreated cells. The cells were then treated with oligomycin (OL; Sigma-Aldrich), carbonyl cyanide 4-trifluoromethoxy phenylhydrazone (FCCP; Sigma-Aldrich), or rotenone and antimycin A (Sigma-Aldrich). The oxygen consumption rate (OCR) was calculated by plotting the O₂ concentration in the medium as a function of time and protein mass (picomoles per minute per microgram protein).

To determine mtDNA content, we extracted total cellular DNA according to a previously described method (31). Quantitative PCR was performed with primers targeting 16s rRNA (mtDNA) or hexokinase 2 (a nuclear gene). Relative mtDNA levels were calculated based on the ratio of mtDNA to the nuclear gene hexokinase 2 (32).

Six-week-old male C57BL/6J mice were purchased from Shanghai Laboratory Animal Co. Ltd. (Shanghai, China) and housed in a temperature-controlled environment with a 12:12-h light/dark cycle. The mice were allowed access to food and water ad libitum. After a 1-week acclimation period, the mice were randomly distributed into weight-matched groups and fed either a normal chow diet (ND) or a high-fat diet (HFD; 60 kcal% fat) (Research Diets Inc, New Brunswick, NJ). After 16 weeks of HFD feeding, animals were sacrificed, and fat depots from interscapular BAT, inguinal sWAT, and perigonadal visceral white adipose tissues (vWATs) were rapidly removed, immediately frozen in liquid nitrogen, and stored at −80°C until further analysis.

For cold stress studies, 8-week-old C57BL/6 male mice were kept at room temperature (25°C) or cold temperature (6°C). In both groups, each mouse was maintained in a single cage on a 12:12-h light/dark cycle with free access to water and food. After 24 h or 7 days, fat depots were isolated and subjected to further analyses.

miRNA antagomirs are chemically modified, cholesterol conjugated, stable miRNA inhibitors. In vivo delivery of miRNA antagomirs is capable of specifically silencing endogenous miRNAs (29,33). To determine the role of miR-199a/214 in vivo, we used mixture of miRNA-199a and -214 antagomirs to repress the expression of miR-199a/214 in fat tissues. Briefly, the 8-week-old male ob/ob mice purchased from the National Resource Center for Mutant Mice (Nanjing, China) were inguinal sWAT injected with antagomirs miR-199a/214 (20 nmol/mice) or scrambled negative controls (Ribobio, Guangzhou, China). The injections were performed four times at 3-day intervals.

After the final injection, the mice were housed in a temperature-controlled environment with a 12:12-h light/dark cycle with access to food and water ad libitum. Body weight, food intake, and blood glucose levels were monitored weekly. Four weeks after the last injection, the body composition of mice was monitored by magnet nuclear magnetic resonance minispec (Bruker’s minispec LF50 Body Composition Analyzer; Bruker Daltonics, Hamburg, Germany), and the mice were placed in the Comprehensive Lab Animal Monitoring System (Columbus Instruments) to evaluate whole-body energy metabolism. Mice were acclimated in individual metabolic chambers with free access to food and water. After a 24-h acclimation period, the CO₂ and O₂ levels, total and wheel activities, and food intake were monitored and recorded over a period of 48 h.

We recruited 33 Chinese people (average age: 49.09 ± 2.55 years; 21 females and 12 males) from the Minimally Invasive Surgery Center, The Second Xiangya Hospital of Central South University (Hunan, China). All subjects were undergoing abdominal surgery for benign hepatobiliary conditions, such as cholecystitis or gallstones. Preoperative BMI (calculated as body weight in kilograms over squared height in meters) and blood pressure were measured, and serum levels of triglycerides, cholesterol, HDL cholesterol, LDL cholesterol, and fasting glucose were determined. The metabolic characteristics of the recruited human subjects are shown in Supplementary Table 1. During the operation, abdominal sWAT and intra-abdominal WAT (iWAT) adipose tissues (≈2 cc each) were collected, snap-frozen, and stored at −80°C before RNA extraction.

Total adipocyte miRNA was extracted using the mirVana miRNA Isolation Kit (catalog number AM1560; Applied Biosystems). Five micrograms of total miRNA was used for miRNA chip analysis using the TaqMan Array Rodent MiRNA A+B Cards Set v3.0 (catalog number 4444909; Applied Biosystems). Briefly, following reverse transcription of miRNA targets using Megaplex RT Primers (#4444746; Applied Biosystems), a TaqMan Universal PCR Master Mix (#4324018) was combined with each reaction and pipetted into each sample loading port in the TaqMan array card. The real-time PCR reaction was performed on an Applied Biosystems 7900HT system.

To analyze the data, we screened miRNAs for which threshold cycle (Ct) values in the miRNA assay were all determined at days 0, 4, and 8 of brown adipocytes differentiation and were between 20 and 30 for the quality control. To identify the expression difference of miRNAs between baseline (day 0) and days 4 or 8, we measured the log fold-change (∆∆Ct) using average ΔCt values (∆∆Ct = ∆Ct at day 4 or 8 − ∆Ct at day 0). If ∆∆Ct >0, miRNA was set as upregulated; if ∆∆Ct <0, miRNA was downregulated; and if ∆∆Ct = 0, there was no regulation.

Total RNA was extracted using the TRIzol Reagent (Invitrogen, Life Technologies, Grand Island, NY) following the manufacturer’s instructions. mRNA was reverse-transcribed, and amplification was carried out on a 7900HT Fast Real-Time PCR System (Applied Biosystems). The primer sequences for the genes are shown in Supplementary Table 2.

For miRNA analysis, cDNA was synthesized from 1 μg of RNA using the PrimeScript One Step miRNA cDNA Synthesis Kit (Takara Bio, Tokyo, Japan) and subjected to real-time PCR. The miRNA sequences are shown in Supplementary Table 3, and the relative expression levels of miRNAs were quantified by the SYBR PrimeScript miRNA RT-PCR kit (Takara Bio) and normalized to U6 expression. To estimate the absolute abundance of the miRNAs, we isolated total RNA from fat tissues, performed miRNA real-time quantitative RT-PCR (qRT-PCR), and calculated the copy numbers of miR-199a and miR-214 by comparing the Ct values of standards (Ribobio).

The primary antibodies were: anti–β-actin (A5441; Sigma-Aldrich), anti-UCP1 (U6382; Sigma-Aldrich), anti–PGC-1α (ab54481; Abcam), or anti-PRDM16 (ab106410; Abcam). Signals were detected using the ChemiDOC XRS+ and the Image Lab system (Bio-Rad).

The 3′-untranslated regions (3′-UTRs) from the mouse Prdm16 and Ppargc1a genes were PCR-amplified from cDNA using specific primers (Supplementary Table 4). The wild-type (WT) or mutant 3′-UTRs were cloned into the pmiR-RB-Report vector (Ribobio) between the XhoI and NotI sites downstream of the Renilla luciferase gene. To generate the mutant Prdm16 3′-UTR reporter, the nucleotides containing the miR-199a seed sequence (5′-CACTGGA-3′) were point-mutated (Mut) to 5′-GTGACCT-3′. For Ppargc1a, the nucleotides (1,519–1,526) of the 3′-UTR containing the miR-199a seed sequence (5′-ACACTGG-3′) were point-mutated to 5′-TGTGACC-3′. Because there are two sites containing the miR-214 seed sequence (5′-CTGCTG-3′) within the Ppargc1a 3′-UTR, we made single site mutations (Mut 1: 217–223 and Mut 2: 568–574) and double site mutations (Mut 3: 217–223/568–574) by changing the seed sequences to 5′-GACGAC-3′. The primer sequences for amplification of the WT and Mut 3′-UTRs are shown in Supplement Table 4. All mutations were generated using the Site-Directed Gene Mutagenesis Kit (Beyotime Institute of Biotechnology, Suzou, China), and both the WT and mutated constructs were verified by sequencing.

For luciferase assays, HEK293T cells were transfected with the 3′-UTR WT or mutant reporters with or without the miR-199a or miR-214 mimic using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Luciferase activity in the cell lysate was measured using the Dual-Glo Luciferase Assay System (Promega, Madison, WI).

All results were presented as the mean ± SEM. Comparisons between two groups were assessed using a Student two-tailed t test for independent samples, and multiple groups were analyzed by one-way ANOVA. The linear correlation between two variables was analyzed by Pearson correlation coefficient. Statistical analyses were carried out using SPSS statistics software (v19.0; SPSS Inc., Chicago, IL). Statistical analysis and plotting for metabolic studies was performed in the R programming language with CalR, a web-based analysis tool for indirect calorimetry experiments (34). The ANCOVA was performed to analyze differences in oxygen consumption (VO₂) and respiratory exchange ratio between the control and experimental groups while statistically controlling for the effects of covariate lean mass. Values of P < 0.05 were considered to be statistically significant.

All procedures involving animals were conducted in accordance with the guidelines set forth by the University Committee on the Care and Use of Animals of the Central South University. All recruited human participants were given written informed consent forms, and the protocol was approved by the Ethics Committee of the Second Xiangya Hospital of Central South University. A written informed consent was received from participants prior to inclusion in the study.

To uncover miRNAs that are important for brown adipocyte differentiation and maturation, we performed miRNA chip assays and compared genome-wide miRNA expression patterns of brown adipocytes collected at differentiation day 0, day 4, and day 8. Using the criteria described in research design and methods, among 768 detected miRNAs, we identified 77 miRNAs for which Ct values were all between 20 and 30 and were differentially expressed during brown adipocyte differentiation (Fig. 1A). Among these miRNAs, the miR-199a and miR-214 gene cluster was one of the most significantly downregulated (Fig. 1A). By qRT-PCR, we analyzed expression levels of miR-199a and miR-214 during BAT- and sWAT-derived SVF or beige adipogenesis. We found that the expression levels of both miR-199a and miR-214 were significantly decreased during brown and beige adipocyte differentiation (Fig. 1B). Additionally, the decreased miRNA levels were negatively correlated with the expression of Ucp1, Ppargc1a, and Prdm16 (Fig. 1B). Dnm3os, a noncoding transcript that harbors the miR-199a and miR-214 gene cluster, showed a similar expression pattern to that of the miRNAs during brown and beige adipocyte differentiation (Fig. 1B).

We analyzed expression levels of miR-199a and miR-214 in three fat depots (including BAT, sWAT, and perigonadal vWAT). Compared with those in BAT, the relative miR-199a and miR-214 levels are higher in vWAT (Fig. 1C). We also calculated the absolute abundance of miR-199a and miR-214 in the above three fat depots (Supplementary Table 5). Consistently, the absolute abundance of the miRNAs in white tissues is also higher than that in BAT.

To determine the expression of the miR-199a/214 cluster in response to cold stress, we maintained the mice at room (25°C) or cold (6°C) temperature for 24 h or 7 days. The levels of both miR-199a and miR-214 decreased significantly in BAT and sWAT, whereas only miR-199a levels decreased in vWAT in response to short- (24-h) or long-term (7-d) cold exposure (Fig. 1D). To confirm these results in vitro, we treated brown adipocytes with the selective β₃-adrenergic receptor activator CL, the nonselective β-adrenergic receptor activator isoproterenol, or the cellular cAMP inducer forskolin. Treating the cells with these chemicals significantly downregulated miR-199a and miR-214 expression (Fig. 1E), indicating the involvement of the β-adrenergic receptor signaling pathway in the regulation of miR-199a/214 expression.

To determine the potential effect of obesity on the expression of the miR-199a/214 cluster, we examined miR-199a and miR-214 levels in adipose tissue of HFD-fed C57BL/6 mice and ob/ob mice. The expression levels of both miR-199a and miR-214 were significantly increased in interscapular BAT and inguinal sWAT of both ob/ob mice (Fig. 2A) and HFD-induced obese mice (Fig. 2B) compared with their respective control mice. We also examined the expression of the miR-199a/214 cluster in sWAT and iWAT of obese and/or overweight humans. The human study group included 10 lean (BMI <25 kg/m²), 13 overweight (BMI >25 kg/m²), and 10 obese individuals (BMI >30 kg/m²). As there were no differences between male and female subjects, all data were pooled and analyzed. Expression levels of miR-199a in sWAT were significantly increased in obese people compared with the lean and overweight subjects (Fig. 2C), and there was a positive correlation between miR-199a levels and BMI (Fig. 2C). The miR-214 levels in sWAT of both overweight and obese subjects were significantly higher than those in the lean control subjects (Fig. 2D) and also significantly increased in iWAT of obese people (Fig. 2E). Similarly, miR-214 levels in both sWAT and iWAT were positively correlated with BMI of tested human subjects (Fig. 2D and E). These data strongly suggest dysregulation of the miR-199a/214 cluster in overweight or obesity.

To determine the role of the miR-199a/214 cluster in brown adipocytes, we transfected brown preadipocytes with miR-199a, miR-214 mimics, or a nonspecific oligonucleotide control (miR-ctrl). Differentiation was induced in the preadipocytes, and the effect of miR-199a and miR-214 on lipid accumulation was examined by Oil Red O staining. The results showed that transfection of miR-199a, but not miR-214, dramatically reduced the formation of lipid droplets in mature adipocytes (Fig. 3A), suggesting distinct effects of the miR-199a on lipid accumulation in brown adipocytes.

Overexpression of miR-199a and miR-214 cluster significantly decreased mRNA expression of the reported target gene Ppard (PPARδ) (35) and thermogenic genes including Ucp1, Cebpb (C/EBPβ), Cidea, Dio2, Elvol3, and Irf4 (Fig. 3B) in brown adipocytes. In addition, overexpression of the cluster significantly decreased the expression of adipogenesis-related genes including Pparg (PPARγ), Cebpa (C/EBPα), Adipoq (Adiponectin), and Fabp4 (Fig. 3B). Conversely, suppression of the miR-199a/214 cluster significantly increased the expression of thermogenic genes in brown adipocytes (Fig. 3C). These results strongly suggest an inhibitory role of the miR-199a/214 cluster on brown adipocyte differentiation and thermogenic gene expression.

To determine the effect of the cluster on cellular respiration, we treated brown adipocytes with miR-199a/214 mimics and analyzed O₂ consumption. Although there were no differences between the mimic and control groups in response to OL treatment, the basal respiration and FCCP-induced uncoupled respiration were significantly lower in brown adipocytes with miR-199a/214 overexpression compared with the control (Fig. 3D), suggesting that miR-199a/214 might impair mitochondrial electron transport capacity, but not ATP production. Conversely, suppression of the miR-199a/214 cluster significantly increased basal and FCCP-induced maximal respiration in brown adipocytes (Fig. 3E).

To determine the role of miR-199a/214 in beige adipocytes, we transfected sWAT-derived beige adipocytes with miR-199a and miR-214 antisense oligonucleotides. Suppression of the miR-199a/214 cluster significantly increased the expression of both thermogenic and beige cell genes, including Tbx1 and Tmem26 (Fig. 4A). miR-199a/214 suppression also increased the expression of genes involved in mitochondrial metabolism (Tfam, Cox7a1, and Nrf1), fatty acid oxidation (CD36 and Acadm), and lipolysis (Atgl and Hsl) (Fig. 4B). In addition, suppression of miR-199a/214 in brown and beige adipocytes significantly increased mtDNA copy number (as indicated by the ratio of the mtDNA [16s rRNA] to the nuclear gene hexokinase 2) (Fig. 4C). Consistently, the basal and chemical-induced OCRs were significantly higher in beige adipocytes in which miR-199a/214 was suppressed (Fig. 4D). Together, these data suggest that decreased expression of the miR-199a/214 cluster could enhance mitochondrial biogenesis and respiration activity and promote beige fat development.

To determine the role of miR-199a/214 in beiging effects in vivo, we delivered miR-199a/214 antagomirs or their respective NC to the sWAT of ob/ob mice through inguinal subcutaneous fat pad injection. We found that expression of the miR-199a/214 cluster, but not other miRNAs (miR-27, miR-133, and miR-155), in sWAT could be efficiently decreased by antagomir injection (Fig. 4E). We also confirmed that expression of miRNA cluster in other tissues, including BAT, vWAT, and muscle, was not changed by injection (data not shown). Four weeks after injection, although there were no significant differences in body composition, food intake, locomotive activity, oxygen consumption, and respiratory exchange ratio (Supplementary Figs. 1–3), ob/ob mice that received antagomir injection in the sWAT had a relatively lower postinjection body weight (Fig. 4F). In addition, antagomir-based knockdown of miR-199a/214 significantly increased basal OCR levels as measured by respirometry analyzer (Fig. 4G) and mRNA levels of thermogenic genes, including Ppargc1a, Pparg, Ucp1, Cidea, Dio2, and Elvol3 as well as the beige genes Tmem26 and Tbx1 in the sWAT of obese mice (Fig. 4H). The increased expression of PGC-1α and UCP1 was also confirmed by Western blotting (Fig. 4I) and UCP1 immunohistochemistry analyses (Fig. 4J). Together, these results, which were consistent with observations in cultured beige adipocytes, suggest that suppression of miR-199a/214 could promote thermogenic gene expression and mitochondrial respiration that might contribute to body weight loss in obese mice.

In silico analysis using the online programs TargetScanMouse 6.2 (http://www.targetscan.org/) and miRDB (http://www.mirdb.org/) identified Prdm16, a critical determinant of the brown fat lineage, as a predicted target of miR-199a (Fig. 5A). To determine whether miR-199a can directly target Prdm16, we cloned the Prdm16 3′-UTR segment containing the predicted miR-199a seed sites or a mutated sequence. Luciferase reporter assays showed that overexpression of miR-199a reduced the luciferase activity of the WT but not the mutant Prdm16 3′-UTR segment reporter (Fig. 5B), demonstrating that miR-199a directly interacts with the predicted target sites in the Prdm16 transcript.

To confirm these results, we transfected brown preadipocytes with miR-199a and miR-214 mimics alone or together and induced differentiation of the cells. Overexpression of miR-199a (but not miR-214) significantly decreased Prdm16 mRNA expression during brown adipogenesis (Fig. 5C). Conversely, the suppression of miR-199a expression induced Prdm16 mRNA levels (Fig. 5C).

To determine whether the effects of miR-199a on brown fat–selective genes occurred via targeting of Prdm16, we overexpressed miR-199a with or without Prdm16-expressing plasmid containing miR-199a target sequence. Our data showed that overexpression of miR-199a decreased, whereas overexpression of Prdm16 induced thermogenic gene expression in the brown adipocytes, and the suppressing effects of miR-199a on gene expression were rescued by coexpression of Prdm16 (Fig. 5D). These results suggest that the inhibitory effect of miR-199a on brown adipocyte thermogenic gene expression is mediated by directly targeting Prdm16.

Bioinformatic analysis also predicts PGC-1α (gene Ppargc1a) as a potential target of both miR-199a and miR-214. There is one seed site for miR-199a and two sites for miR-214 within the Ppargc1a 3′-UTR (Fig. 5E). Luciferase reporter assays showed that overexpression of either miR-199a or miR-214 significantly reduced the luciferase activity of the Ppargc1a 3′-UTR WT construct. This effect was enhanced by overexpression of miR-199a and miR-214 together. However, such effects were abolished when miR-199a binding sites or miR-214 double binding sites were mutated (Fig. 5F), indicating that both miR-199a and miR-214 could directly target the Ppargc1a 3′-UTR, and double targeting sites are required for miR-214 to bind with the 3′-UTR of Ppargc1a.

The qRT-PCR analyses confirmed that mRNA levels of Ppargc1a were significantly downregulated throughout brown adipocyte differentiation when either miR-199a or miR-214 was overexpressed. These effects could be enhanced by cotransfection of two miRNAs (Fig. 5G). Conversely, knockdown of miR-199a and miR-214 expression by inhibitors induced Ppargc1a mRNA levels in brown adipocytes (Fig. 5G). Consistent with the above findings (Fig. 5C and G), overexpressing or suppressing miR-199a/214 expression inhibited or induced, respectively, protein expression of PRDM16, PGC-1α, and UCP1 (Fig. 5H).

Brown and beige fat development and thermogenic function are involved in several essential transcriptional and epigenetic regulators, and among them, miRNAs play unique roles by targeting multiple regulatory factors (11,36). In this study, we identified the miR-199a/214 cluster as a key negative regulator of brown adipogenesis and brown/beige fat thermogenesis. Overexpression of the miR-199a/214 cluster inhibited thermogenic gene expression and mitochondrial respiration in brown adipocytes, whereas knockdown of the cluster increased these parameters in both brown and beige adipocytes. Antagomir-based knockdown of miR-199a/214 in sWAT promoted thermogenic gene expression and body weight loss in obese mice. We also demonstrated that the miR-199a/214 cluster suppresses brown adipogenesis and thermogenic gene expression by directly targeting PRDM16 and PGC-1α. Our study demonstrates for the first time that the obesity-associated miR-199a/214 cluster plays a key negative role in brown adipogenesis and the beige fat development.

Previous studies showed that miR-199a is one of the members of a muscle miRNA (or myomiR) family regulated by serum response factor (14). Expression of the miR-199a increases during myogenic differentiation and promotes normal myogenesis (14). Our study found that expression of the miR-199a is downregulated during the brown and beige adipocyte differentiation, and miR-199a plays a negative role in brown adipocyte differentiation. As brown adipocytes and skeletal myocytes have been indicated to originate from somite-derived lineage cells expressing Pax7 and Myf5 (13,37), from which PRDM16 controls the switch between brown and skeletal muscle cells in the lineages (10,38,39), it is reasonable to propose that miR-199a might be an essential lineage regulator of BAT and muscle cells by functioning upstream of PRDM16.

miRNAs play multiple important roles in maintaining metabolic homeostasis. Dysfunction or dysregulation of miRNAs is associated with various disorders, including obesity and diabetes (35,40). The miR-199a/214 is a hypoxia-induced cluster implicated as a mediator integrating hypoxic signaling in the heart (41,42). Induction of the cluster through hypoxic conditions reduces mitochondrial fatty acid oxidative capacity and cardiac contractility resulting in heart failure (41). We found that miR-199a/214 levels in adipose tissues were dramatically increased in obese mice and in overweight/obese humans. Although it is not clear whether obesity-associated upregulation of miR-199a/214 is caused by adipose tissue hypoxia, our data showed that upregulation of this gene cluster leads to reduced mitochondrial activity in the brown adipocytes, suggesting that dysregulation of this cluster might result in dysfunction of fat tissues contributing to development of obesity and related metabolic disorders.

Importantly, our studies demonstrate that suppression of miR-199a/214 expression strongly promotes beige fat development in vitro and in vivo. Knockdown of the cluster increased thermogenic gene expression and mitochondrial respiration in both brown and beige adipocytes. In addition to Ucp1 and other thermogenic genes, we observed increased expression of genes involved in fatty acid oxidation, lipolysis, and mitochondrial function in the miR-199a/214-suppressed adipocytes. Antagomir-based knockdown of the cluster in sWAT increased thermogenic gene expression and mitochondrial respiration and had potential to lower body weight in obese mice. In the future, genetically modified mouse models with fat tissue–specific manipulation of this gene cluster will help to identify a better therapeutic intervention for obesity in vivo.

PRDM16 is a key transcriptional regulator that determines the commitment, differentiation, and development of brown/beige adipocytes (8,15). We demonstrated that miR-199a directly targets the 3′-UTR of Prdm16 and that the inhibitory effect of miR-199a on brown adipogenesis could be recovered by coexpression of Prdm16, suggesting that Prdm16 mediates the effect of miR-199a on brown adipocyte differentiation. Previous studies showed that miR-199a affects myogenic cell proliferation and differentiation by targeting the WNT signaling pathway (9). Our study unveils a novel role of this miRNA in brown adipocytes. Several miRNAs have been reported to be involved in brown adipogenesis (22). Of these, miR-133a/b and miR-27 are known to play inhibitory roles via targeting of Prdm16 (25,43–45). These parallel findings suggest that several miRNAs could cotarget the same key regulatory molecules and coordinate to regulate biological processes, such as cold exposure–induced thermogenesis and brown and beige adipogenesis (27,46).

Our study also demonstrated that both miR-199a and miR-214 directly target the 3′-UTR of Ppargc1a. PGC-1α is a master regulator of thermogenesis and mitochondrial biogenesis (22). Increased PGC-1α expression could contribute to increased mitochondrial biogenesis, expression of genes involved in fatty acid mobilization and lipolysis, and mitochondrial respiration in beige adipocytes with suppression of the miR-199a/214 cluster. Previous studies using Ppargc1α-null mice show that PGC-1α is not necessary for brown fat cell differentiation per se (4,18). In agreement with this, we found that brown adipogenesis was impaired by miR-199a (which targets both Prdm16 and Ppargc1a) but not by miR-214 (which targets Ppargc1a only). Both miR-199a and miR-214 target the Ppargc1a 3′-UTR, resulting in an additive suppressive effect on Ppargc1a expression. As Prdm16 is a transcriptional regulator of PGC-1α, the inhibitory effects of the miR-199a/214 on PGC-1α expression could also be caused, in part, by targeting Prdm16. Thus, the miR-199a/214 cluster acts as a key negative regulator of brown and beige adipocyte development and function by simultaneously targeting a regulatory gene cascade involved in beige and brown adipogenesis and thermogenesis.

Like many miRNA clusters, the synergistic action of the miR-199a/214 cluster on PGC-1α and other genes might be related to mRNA secondary structures. Secondary structures surrounding seed sequences can affect RNA deadenylation or sequestration, both of which are important for miRNA–mRNA interactions and gene repression (25). It has been suggested that, due to sequence homology, miRNAs in a cluster act as a family and have both common and unique mRNA targets that lie within the same pathway, thereby allowing these miRNAs to play regulatory roles in several components of a cellular process (12,47).

In summary, we revealed the obesity-associated miR-199a/214 cluster as a novel gene cluster that negatively regulates brown and beige adipogenesis, thermogenesis, and mitochondrial respiration by targeting the major brown transcriptional regulators PRDM16 and PGC-1α. This study provides new insights into the mechanisms coordinating the gene cascades that regulate thermogenesis and energy metabolism and provides a potential therapeutic target against obesity.

Funding. This work was supported by grants from the National Natural Science Foundation of China (31471131 and 31871180) and the International Science and Technology Cooperation Program of China (2014DFG32490) to F.H., the National Basic Research Program of China (2014CB910501) to F.L., and the National Natural Science Foundation of China (31571368) and the Project of Innovation-Driven Plan of Central South University (2016CX031) to G.L.

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

Author Contributions. L.H. and M.T. designed the study, carried out the research, and analyzed the results. T.X., H.L., F.Z., and Y.X. conducted the experiments. W.L. recruited and conducted the human studies. G.L. conducted bioinformatics analysis. Z.Z. and F.L. reviewed and edited the manuscript. F.H. supervised experiments, analyzed data, and wrote and revised the manuscript. F.H. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.