miR-125a-3p and miR-483-5p promote adipogenesis via suppressing the RhoA/ROCK1/ERK1/2 pathway in multiple symmetric lipomatosis

Multiple symmetric lipomatosis (MSL) is a rare disease characterized by symmetric and abnormal distribution of subcutaneous adipose tissue (SAT); however, the etiology is largely unknown. We report here that miR-125a-3p and miR-483-5p are upregulated in the SAT of MSL patients, promoting adipogenesis through suppressing the RhoA/ROCK1/ERK1/2 pathway. TaqMan microRNA (miR) array analysis revealed that 18 miRs were upregulated in the SAT of MSL patients. Transfection of human adipose-derived mesenchymal stem cells (hADSCs) with the individual agomirs of these 18 miRs showed that miR-125a-3p and miR-483-5p significantly promoted adipogenesis. A dual-luciferase assay showed that RhoA and ERK1 were the targets of miR-125a-3p and miR-483-5p, respectively. Moreover, transfection of hADSCs with mimics of miR-125a-3p and miR-483-5p resulted in a pronounced decrease of ERK1/2 phosphorylation in the nucleus; conversely, transfection of hADSCs with inhibitors of miR-125a-3p and miR-483-5p led to a significant increase of ERK1/2 phosphorylation in the nucleus. Most importantly, we found that miR-125a-3p and miR-483-5p promoted de novo adipose tissue formation in nude mice. These results demonstrated that miR-125a-3p and miR-483-5p coordinately promoted adipogenesis through suppressing the RhoA/ROCK1/ERK1/2 pathway. Our findings may provide novel strategies for the management and treatment of MSL or obesity.

Accumulating evidence suggests that miR-125a plays an important role during adipogenesis 15,16 . Importantly, Ras homolog family member A (RhoA), a small GTPase that plays key roles in adipogenesis, has been reported as a target gene of miR-125a-3p 17 . In human mesenchymal stem cells (hMSCs) as well as mouse adipose-derived stromal cells (mASCs), overexpression of dominant-negative RhoA induced hMSCs or mMSCs to adipocytes; whereas constitutively active RhoA or Rho-associated kinase (ROCK), an effector of RhoA, led to osteogenesis 18,19 . Similarly, knockdown of RhoA with RNAi or pharmacological inhibition of RhoA or ROCK in preadipocytes promoted adipogenesis in mouse 3T3-L1 cells; in contrast, ectopic overexpression of RhoA or treatment with the RhoA agonist lysophosphatidic acid inhibited adipogenesis in mouse 3T3-L1 cells 20,21 . Thus, the RhoA/ROCK pathway is a "switch" not only in terms of the stage of stem cells to preadipocytes but also during the process of preadipocytes to mature adipocytes. Targeting RhoA indicates a key role of miR-125a-3p during adipogenesis.
In this study, we first systematically investigated the expression profile of miRs in SAT between MSL patients and control subjects. Next, we verified the regulation of adipogenesis by miR-125a-3p and miR-483-5p in hADSCs by overexpression or downregulation of miR-125a-3p and miR-483-5p, and analyzed RhoA and ERK1 by luciferase reporter assays. Then, we explored the interactions of miR-125a-3p and miR-483-5p on the RhoA/ROCK/ERK1/2 pathway. Finally, we observed adipogenesis of nude-mouse subcutaneous hADSCs following transfection of miR-125a-3p and miR-483-5p.

Methods
Sample selection and preparation. Three male MSL and three male control subjects were recruited in this study. Three control subjects had no diabetes, malignant tumors, acute infectious disease, and smoking history. Anthropometric and metabolic characteristics were evaluated according to our previously study 4 . The SAT was obtained from the right upper quadrant of the abdomen for all subjects. The study protocol was approved by the Human Ethical Review Committee of the Third Xiangya Hospital of Central South University, Changsha, China; and all subjects signed the informed written consent. All methods used in this study were carried out in accordance with the approved guidelines.
RNA extraction and TaqMan MicroRNA array analysis. Total RNA was isolated with a TRIzol RNA extraction kit (Life Technologies, Carlsbad, CA, USA). The miR enrichment was performed with an mirVana miRNA Isolation Kit and converted to cDNA by a TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems, Life Technologies, USA), according to the manufacturer's instructions. The reverse transcription products were used with the TaqMan Human MicroRNA Array A+ B cards set v3.0 (Applied Biosystems, Life Technologies, USA) to detect 754 human miRs. miRs expression fold changes were calculated by the 2 −ΔΔCT method. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed according to the method described by Wen et al. 28 .
hADSC isolation and differentiation. hADSCs were isolated from the abdominal SAT of two control patient. Ten gram tissue was washed three times with D-Hank's solution (Gibco, Life Technology, China), cut into 1-mm × 1-mm pieces, and digested with collagenase I (Gibco, Life Technology, China) for 1 h at 37 °C. The cells were collected by centrifugation at 150 × g for 10 min, then lysed in erythrocyte lysis buffer (0.154 M NH 4 Cl, 10 mM KHCO 3 , and 0.1 mM EDTA) for 10 min, and filtered by a nylon mesh. The supernatant was collected and centrifuged again. The pelleted cells were cultured in DMEM/ F12 medium (Gibco, Life Technology, China) supplemented with 10% fetal bovine serum (Gibco, Life Technology, Australia). After 24 h, the medium was replaced with fresh medium. The cells were cultured for 4-7 passages for experiments. To induce hADSC differentiation, over-confluent hADSCs were cultured in inducing medium (DMEM/F12 supplemented with 1 μ M dexamethasone, 10 μ M insulin, 0.5 mM isobutylmethylxanthine (IBMX), and 200 μ M indomethacin (all provided by Sigma, St. Louis, MO, USA). The medium was replaced every 2 days. For hADSC self-differentiation, cells were maintained in DMEM/F12 medium without the differentiation-inducing reagents. We repeated all experiments by two kinds of hADSCs that were isolated from two control patients.
Quantitative real-time polymerase chain reaction (PCR). The  Technologies, USA) with automatic baseline and threshold settings. The miRs were normalized to the level of U6 snRNA.
Oil red O staining. The cells were washed two times in D-Hank's solution, fixed in 4% formaldehyde for 30 min, and washed three times with water. Then, the cells were stained with Oil red O (Sigma, St. Louis, MO, USA) for 15 min. Following three washes in water, the lipid droplets were observed and photographed under a microscope (TE2000-E; Nikon, Japan).
Dual-luciferase reporter assay. The 293T cells were cultured in 96-well plates in DMEM supplemented with 10% fetal bovine serum at 37 °C with 5% CO 2 . When the cells reached 70-80% confluence, the wild-type (WT) or mutation-type (MT) RhoA and ERK1 3′ -UTR plasmids were cotransfected with miR-125a-3p or miR-483-5p mimics (100 nM) or negative control (NC) mimics with Lipofectamine 2000. After 48 h, the cells were washed twice with PBS and lysed with passive lysis buffer (PLB) before dual-luciferase reporter assay agents were added (Beyotime, Haimen, China). Firefly and renilla luciferase activities were measured with the Dual-Glo luciferase assay system (Promega, Madison, WI, USA). All experiments were performed three times.

Lentiviral construction and transfection.
To generate hsa-miR-125a and hsa-miR-483 lentiviral expression plasmids, 268-bp and 273-bp sequences containing pre-hsa-miR-125a or pre-hsa-miR-483 were synthesized and cloned into lentiviral expression vector pGC-FU (GeneChem, Shanghai, China), respectively. After sequence confirmation by DNA sequencing, the viruses were packaged in 293T cells after cotransfection with pHelper 1.0 and pHelper 2.0 Packing Plasmid (GeneChem, Shanghai, China) using Lipofectamine 2000. To obtain stable lentivirus-infected cell lines, the cells were plated at 30-40% confluence and lentiviral vector with 2 mg/ml polybrene (GeneChem, Shanghai, China) resolved in serum-free medium was transfected. After 16 h, the medium was replaced with fresh complete medium. 72 h later, the transfection efficiency was verified by fluorescence microscopy.

De novo adipogenesis in vivo.
To induce de novo adipose tissue, hADSCs were transfected with one of the following lentiviruses: negative control miR (LV-NC), hsa-miR-125a, hsa-miR-483, or hsa-miR-125a +483. A total of 1.5 × 10 7 transfected hADSCs were harvested and resuspended in 200 μ l of PBS. The cells or PBS were injected subcutaneously in the backs of 6-week-old male athymic Balb/c nude mice (Chinese Academy of Sciences, Shanghai, China). The mice were fed with a normal diet on a 12-h day/ night cycle for 5 weeks. Five mice from each group were sacrificed. The de novo adipose tissues were removed and weighed. Some samples were stained with hematoxylin and eosin. Adipoctye areas were measured and analysed by ImageJ 1.48μ digital imaging processing software. The other samples were used for real-time PCR or western blot analyses.
Scientific RepoRts | 5:11909 | DOi: 10.1038/srep11909 All animal experiments were approved by the Human Ethical Review Committee of the Third Xiangya Hospital of Central South University, Changsha, China, and performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals (1996).

Statistical analyses.
All the results are presented as means ± standard deviation. Two groups were compared by the unpaired Student's t-test, and multiple groups were analyzed by one-way analysis of variance. Statistical significance was defined as a P value < 0.05.

Results
Some miRs are upregulated in MSL patients. All three MSL patients [including one case we have reported previously 4 ] were addicted to alcohol for 15, 10, and 8 years and had gradually increased SAT accumulation for 7, 5, and 3 years, respectively. They all showed symmetrical and substantial SAT accumulation in the neck, upper arms, bilateral shoulders, upper thorax, back, and abdomen (Fig. 1A). In order to identify the differential expression of miRs in SAT between MSL patients and controls, the miR expression profiles were detected by TaqMan miR arrays. Some anthropometric and metabolic characteristics were described in Table 1. Among the 754 human miRs, 18 miRs were upregulated in the SAT of MSL patients compared to that of the control group; and no miR was significantly downregulated in the SAT of MSL patients ( Table 2). Ten of the 18 miRs were further verified by miR real-time PCR (Fig. 1B). KEGG pathway analysis revealed that some adipogenesis-related pathways including Wnt, TGF-β , actin cytoskeleton, and especially MAPK were significantly enriched (Fig. 1C). These results imply that miRs may play important roles during adipogenesis in MSL patients. miR-125a-3p and miR-483-5p significantly promote adipogenesis in hADSCs. To investigate which miRs may promote adipogenesis, all 18 miRs with agomir were respectively transfected into hAD-SCs, which were further induced to mature adipocytes or underwent self-differentiation for 12 days. We found that transfection with miR-125a-3p or miR-483-5p significantly promoted lipid droplet accumulation in hADSCs that were induced to mature adipocytes. Conversely, inhibition of miR-125a-3p or miR-483-5p with the corresponding antagomir markedly decreased lipid droplet accumulation ( Fig. 2A,B). The same results were observed in the hADSCs undergoing self-differentiation for 12 days (Fig. 2C). Next, after each group was induced to mature adipocytes for 12 days, C/EBPα , PPARγ , and FABP4 were upregulated by the agomir of miR-125a-3p or miR-483-5p; while they were downregulated by the antagomir of miR-125a-3p or miR-483-5p (Fig. 2D,E). These results demonstrated that miR-125a-3p and miR-483-5p promote adipogenesis in hADSCs.

Discussion
In this study, we discovered that 18 miRs were upregulated in the SAT of MSL patients and that miR-125a-3p or miR-483-5p significantly promoted adipogenesis via the RhoA/ROCK1/ERK1/2 pathway (Fig. 7). Importantly, we found that miR-125a-3p and miR-483-5p promoted de novo adipose tissue formation in nude mice. These results demonstrated that miRs play an important role during adipogenesis and could be used in potential novel strategies for the prevention and treatment of MSL or obesity. It has been reported that the m.8344A> G mutation in the tRNA Lys is an important etiological factor of MSL 29 , but we did not find this mutation in the blood or SAT samples of our three MSL patients (data not shown). Therefore, we explored the functions of miRs in the pathogenesis of MSL. After miR array analysis, we found that 18 miRs were significantly upregulated. Although alcohol is believed to be an important pathogenic factor in MSL, none of the upreguated 18 miRs was correlated with alcohol addiction except of miR-140-3p 30 . Some of these 18 miRs have been reported to participate in adipogenesis. For example, let-7 represses adipogenesis in hMSCs and 3T3-L1 cells 31,32 ; let-7e, let-7b, miR-28a-5p, and miR-10a are upregulated, while let-7c and miR-125b are downregulated in either preadipocytes or mature adipocytes of obese vs. lean SAT 33 . Moreover, miR-125b is directly and significantly correlated with body mass index 34 . In addition, miR-320 regulates insulin resistance in adipocytes 35 ; miR-140 promotes adipogenesis 36 and miR-483-3p inhibits 3T3-L1 adipogenesis and gradually decreases during In hADSCs, miR-125a-3p inhibits RhoA, resulting in decreased ROCK1 expression, which inhibits n-T-ERK1/2 and n-p-ERK1/2 in the nucleus. At the same time, miR-483-5p directly decreases T-ERK1/2 and T-p-ERK1/2 in the cytoplasm, and n-T-ERK1/2 and n-p-ERK1/2 in the nucleus. Subsequently, PPARγ , a downstream target of n-T-ERK1/2 and n-p-ERK1/2, which are negatively correlated with PPARγ , is increased and promotes adipogenesis. differentiation 22 . However, the underlying molecular mechanisms of the regulation of adipogenesis by most of these miRs remain to be determined.
Our miR pathway analysis predicted that these upregulated miRs are closely related to the MAPK, Wnt, TGF-β , and regulation of actin cytoskeleton signaling pathways, which have been shown to negatively regulate adipogenesis 18,21,37,38 . These data suggest that miRs may participate in regulating these pathways and adipogenesis in MSL. Indeed, the miR-125a-3p and miR-483-5p agomir apparently promoted adipogenesis, while their antigomir significantly prevented adipogenesis under either induction or self-differentiation conditions. Previous studies have shown that miR-125a-3p and miR-483-5p target RhoA and ERK1, respectively 17,26 , which was further supported by our dual-luciferase assay.
miR-125a-5p, another mature sequence of miR-125a, has been found to be downregulated in the epididymal fat pads of leptin-deficient ob/ob mice. Interestingly, ssc-miR-125a has been found to repress the differentiation of porcine preadipocytes 15 . Consistent with the previous observation in mouse and human MSCs 16 , we found that miR-125a-3p expression was gradually upregulated in adipogenesis. A recent study has reported that miR-125a-3p is significantly related with fat mass and waist circumstance and that miR-125a-3p expression in the abdominal omental tissue in males is much higher than that in females 39 . Different from previous studies, our data showed that miR-125a-3p was highly expressed in the SAT of MSL patients, suggesting that miR-125a-3p may play different roles in MSL adipose tissue. miR-483-5p is located within the second intron of its host gene insulin-like growth factor 2 (IGF2) and was found to be coexpressed with IGF2 in 3T3-L1, Hepa1-6, and HepG2 cells 40,41 . It has been well documented that IGF2 promotes adipogenesis and is gradually upregulated during adipogenesis 42,43 . Our results demonstrated that miR-483-5p was also gradually upregulated during adipogenesis and promoted adipogenesis. These data suggest that miR-125a-3p and -5p as well as miR-483-3p and -5p may play reverse roles during adipogenesis, at least in adipogenesis of MSL patients.
RhoA and ERK1 are important negative regulation factors of adipogenesis, and dysregulation of their levels affects adipogenesis either in the stem cell or preadipocyte stage 20,21,24 . Our results demonstrated that the RhoA, ROCK1, T-ERK1/2, and T-p-ERK1/2 levels were downregulated in the SATs of MSL patients suggesting that downregulation of these negative regulation factors may promote adipogenesis in mature adipose tissue. We also found that RhoA, ROCK1, and T-p-ERK1/2, but not T-ERK1/2, were gradually downregulated during adipogenesis, consistent with previous studies 26,44,45 . It is unclear why upregulation of miR-483-5p decreased T-p-ERK1/2 but not T-ERK1/2 during adipogenesis. It remains to be determined whether other mechanisms participate in regulating T-ERK1/2 expression.
Several studies have demonstrated that the RhoA/ROCK pathway positively regulates p-ERK1/2 in the nucleus. For example, RhoA was found to activate n-p-ERK1/2 via ROCK 46 . Moreover, Brians et al. found that RhoA did not affect T-p-ERK but increased the nuclear localization of p-ERK 47 . Similarly, in vascular smooth muscle cells, alteration of ROCK1 and ROCK2 did not affect the T-ERK1/2, T-p-ERK1/2, or n-T-ERK1/2 expression levels but promoted nuclear translocation of n-p-ERK1/2 27,48 . However, a study using human leukemia cells showed that activation of the RhoA/ROCK1 pathway decreased p-ERK1/2 in the nucleus, and inhibition of RhoA/ROCK1 led to accumulation of n-p-ERK1/2 49 . Our data showed that miR-125a-3p did not affect T-ERK1/2 or T-p-ERK1/2, whereas it apparently altered the levels of n-T-ERK1/2 and n-p-ERK1/2, which was different from other previous studies. To determine whether miR-483-5p regulates the RhoA/ROCK1/ERK1 pathway, we compared the expression levels of these proteins in hADSCs transfected with the miR-125a-3p inhibitor, miR-483-5p mimic, or cotransfection of the miR-125a-3p inhibitor and miR-483-5p mimic, and found that cotransfection did not change the T-ERK1/2 or T-p-ERK1/2 levels but apparently changed the n-T-ERK1/2 and n-p-ERK1/2 levels. Furthermore, cotransfected with the miR-125a-3p agomir and pCEP4L-ERK1 partially restored adipogenesis compared to pCEP4L-ERK1, and similar protein expression of PPARγ and C/EBPα were observed. These data further supported the conclusions that miR-125a-3p affects the T-ERK1/2 and p-ERK1/2 levels in the nucleus and that miR-125a-3p and miR-483-5p may jointly regulate the RhoA/ ROCK1/ERK1/2 pathway.
In summary, we found that both miR-125a-3p and miR-483-5p are significantly increased in the SATs of MSL patients and that miR-125a-3p and miR-483-5p promote adipogenesis through regulating the RhoA/ROCK1/ERK1/2 pathway. Although we were not allowed to get enough tissues to isolate hADSC from MSL patients, our findings suggest that the RhoA/ROCK1/ERK1/2 signaling pathway is a potential therapeutic target for the development of drugs that prevent or treat MSL or obesity patients.