MicroRNA-140 Promotes Adipocyte Lineage Commitment of C3H10T1/2 Pluripotent Stem Cells via Targeting Osteopetrosis-associated Transmembrane Protein 1*

Background: BMP4 treatment induces adipocyte lineage commitment of C3H10T1/2 pluripotent stem cells. Results: Expression of miR-140 increases significantly during adipocyte lineage commitment. Conclusion: miR-140 promotes adipocyte lineage commitment through down-regulating Ostm1. Significance: Which miRNA and how it functions in adipocyte lineage commitment are clarified. BMP4 has been shown to induce C3H10T1/2 pluripotent stem cells to commit to adipocyte lineage. In addition to several proteins identified, microRNAs also play a critical role in the process. In this study, we identified microRNA-140 (miR-140) as a direct downstream component of the BMP4 signaling pathway during the commitment of C3H10T1/2 cells to adipocyte lineage. Overexpression of miR-140 in C3H10T1/2 cells promoted commitment, whereas knockdown of its expression led to impairment. Additional studies indicated that Ostm1 is a bona fide target of miR-140, which is significantly decreased during commitment, and Ostm1 was also demonstrated to function as an anti-adipogenic factor.

BMP4 has been shown to induce C3H10T1/2 pluripotent stem cells to commit to adipocyte lineage. In addition to several proteins identified, microRNAs also play a critical role in the process. In this study, we identified microRNA-140 (miR-140) as a direct downstream component of the BMP4 signaling pathway during the commitment of C3H10T1/2 cells to adipocyte lineage. Overexpression of miR-140 in C3H10T1/2 cells promoted commitment, whereas knockdown of its expression led to impairment. Additional studies indicated that Ostm1 is a bona fide target of miR-140, which is significantly decreased during commitment, and Ostm1 was also demonstrated to function as an anti-adipogenic factor.
Obesity has become an escalating global epidemic since people have changed their diet and lifestyle in the 20th century; it not only affects appearance but also presents an array of serious disorders to health, including insulin resistance, type 2 diabetes, hypertension, and atherosclerosis (1,2). Elucidating the mechanisms underlying obesity is critical for understanding obesity occurrence and progression. It has been shown that adipocyte development includes two progressive stages: lineage-re-stricted preadipocytes committed from pluripotent mesenchymal stem cells (3) and lipid-laden adipocytes differentiated from growth-arrested preadipocytes (4). Researchers have applied 3T3-L1 cells, a widely used preadipocyte cell line, to clearly understand the differentiation process (5,6). However, there is still far more to learn about the commitment process (7).
The generally accepted mesenchymal stem cell line C3H10T1/2 has been used to elucidate the mechanisms of adipocyte lineage commitment (4). C3H10T1/2 cells are multipotent stem cells that can differentiate into various lineages, including osteocytes, chondrocytes, myocytes, and adipocytes (8,9). BMP4 (bone morphogenetic protein 4) treatment can efficiently induce C3H10T1/2 cells to commit to adipocyte lineage through two downstream signaling pathways, SMAD and p38 MAPK (10). The commitment of C3H10T1/2 cells is accompanied by dramatic changes in cell shape, and several cytoskeleton-associated proteins (i.e. lysyl oxidase, TPT1 (tumor protein, translationally controlled 1), and ␣B-crystallin) have been demonstrated to be up-regulated by BMP4 during commitment (11).
MicroRNAs (miRNAs) 3 function at the post-transcriptional level by negatively regulating mRNA stability or translation, and they participate in almost every physiological and pathological process (12)(13)(14)(15). Numerous miRNAs have been shown to be involved in terminal adipocyte differentiation (16). For example, microRNA (miR)-143, a well known miRNA that enhances adipogenesis, increases after induction of differentiation and targets pleiotrophin to promote differentiation of 3T3-L1 preadipocytes (17), whereas pleiotrophin plays a negative role during adipogenesis through the pleiotrophin/PI3K/ Akt/GSK3␤/␤-catenin signaling pathway. Stable transfection of 3T3-L1 cells with the miR-17-92 cluster results in accelerated differentiation by negatively regulating the tumor suppressor protein Rb2/p130, which participates in a fundamental step in mitotic clonal expansion (18). miR-375 enhances 3T3-L1 adipocyte differentiation by suppressing the phosphorylation levels of ERK1/2 (19). On the other hand, the miR-27 gene family, including miR-27a and miR-27b, is down-regulated during 3T3-L1 adipocyte differentiation, and overexpression of miR-27a and miR-27b inhibits adipocyte differentiation of 3T3-L1 preadipocytes (20). The miR-27 family has also been shown to be elevated in obese mice and to contribute to LPS-mediated inflammation by targeting peroxisome proliferator-activated receptor ␥ (21). However, there is little information regarding the roles of miRNAs during adipocyte lineage commitment.
OSTM1 is a type I transmembrane protein that localizes in intracellular vesicles, is highly expressed in cartilage, and is generally conserved in a wide range of species, including zebrafish, mice, and humans (25,26). Previous studies elucidated three biological functions of OSTM1: it serves as a ␤-subunit of ClC-7 to support bone resorption and lysosomal function, it works as an E3 ubiquitin ligase to induce proteasome-dependent degradation of G␣ i3 , and it promotes ␤-catenin/Lef1 interaction (22)(23)(24). The above findings suggest that OSTM1 has an important role in bone development. As mesenchymal stem cells can differentiate into both osteocytes and adipocytes, OSTM1 might influence cell fate determination between these cells.
In this study, we found that BMP4 treatment dramatically increased miR-140 expression, which promoted the commitment of C3H10T1/2 cells to adipocyte lineage. Furthermore, we identified Ostm1 as a direct target of miR-140 and show that it functions as an anti-adipogenic factor.

EXPERIMENTAL PROCEDURES
Cell Culture and Induction of Commitment and Differentiation-C3H10T1/2 mesenchymal stem cells and 3T3-L1 preadipocytes were propagated and differentiated as described (4).
Microarray and Data Analysis-Total RNA was extracted using the TRIzol method (Invitrogen) according to the manufacturer's protocol. miRNA microarray was performed using Agilent mouse miRNA microarrays in triplicate.
Quantitative RT-PCR Analysis-Total RNA (including total miRNA) was harvested from C3H10T1/2 cells using TRIzol. The miRNA was reverse-transcribed using the TaqMan miRNA reverse transcription kit (Applied Biosystems) and miRNA-specific primers (Applied Biosystems). miRNA expression levels were then analyzed using the appropriate TaqMan miRNA assay (Applied Biosystems) according to the manufacturer's instructions. Quantitation of the ubiquitously expressed miRNA (U6) was performed as an endogenous control.
Oil Red O Staining-Cells were washed three times with PBS and fixed for 15 min with 3.7% formaldehyde. Oil Red O (0.5% in isopropyl alcohol) was diluted with water (3:2), filtered through a 0.45-m filter, and incubated with the fixed cells for 1 h at room temperature. Cells were then washed with water, and stained lipid droplets in the cells were visualized by light microscopy and photographed.
Luciferase Assay-293T cells were transfected with 10 ng of psiCHECK2-Ostm1 3Ј-UTR vector at 24 h after plating and with psiCHECK2-mutant Ostm1 3Ј-UTR vector and miR-140 sponge vector for 6 h in reduced serum and antibiotic-free Opti-MEM I with Lipofectamine 2000. Cells were cotransfected with the miR-140 mimic or a negative control (Genepharm) at 100 nM. Firefly and Renilla luciferases were measured in cell lysates using the Dual-Luciferase reporter assay system (Promega). Firefly luciferase activity was used for normalization and as an internal control for transfection efficiency.
Transfection Assay-Transfection experiments were performed with Lipofectamine TM 2000 and Lipofectamine TM RNAiMAX (Invitrogen) following the manufacturer's instruction.
RNA Interference-Synthetic siRNA oligonucleotides specific for Smad4 (GenBank TM accession number NM_ 008540.2), p38 MAPK (GenBank TM accession U10871.1), and Ostm1 mRNAs were designed and synthesized by Invitrogen Stealth TM RNAi. The sequences were as follows: Smad4 Isolation of Stromal Vascular Fraction-Epididymal and inguinal fat pads from male C57BL/6 mice and wild-type and BMP4 transgenic mice in which BMP4 was specifically overexpressed in adipose tissue were excised and minced in PBS with 0.5% BSA. Collagenase (Sigma-Aldrich) was added to 1 mg/ml before incubation at 37°C for 2 h with shaking. Suspensions were centrifuged at 1500 ϫ g for 5 min to remove cellular debris and oil; precipitations were resuspended with RBC lysis buffer and then recentrifuged at 4000 ϫ g for 10 min. Suspensions were washed twice with PBS, centrifuged, and resuspended with cell lysis buffer. Samples were heated at 110°C for 15 min and were then ready for Western blotting.

RESULTS
miR-140 and miR-140* Are Induced during Adipocyte Lineage Commitment of C3H10T1/2 Cells by BMP4 Treatment-As shown in Fig. 1 (A and B), we confirmed that BMP4 can induce commitment of C3H10T1/2 cells from pluripotent stem cells to adipocyte lineage (4), and activation of BMP4 signaling, e.g. phosphorylation of p38 MAPK and SMAD1/5/8, was significantly induced (Fig. 1D). To determine whether miRNA is involved in this process, microarray analysis was performed,

miR-140 Promotes Adipocyte Lineage Commitment
and the results revealed that numerous miRNAs were positively or negatively regulated by BMP4 treatment (supplemental Fig.  S1). In addition to many slightly altered miRNAs, miR-140 and miR-140* increased by Ͼ2-fold in BMP4-treated C3H10T1/2 cells compared with control cells and were selected for further studies. Because miR-140 and miR-140* originated from two strands of one single pre-miRNA, their expression levels were determined by TaqMan real-time PCR assay (Fig. 1E). As illustrated in Fig. 1F, the expression of miR-140 and miR-140* began to increase gradually during commitment, peaked on day 6 after continuous BMP4 treatment, and decreased during terminal adipocyte differentiation, which implied that they might function in adipocyte lineage commitment. In contrast, the miRNA expression level changed slightly in control C3H10T1/2 cells compared BMP4-treated cells.
miR-140 Overexpression Promotes Adipocyte Lineage Commitment of C3H10T1/2 Cells-Because miR-140 and miR-140* were specifically enriched during BMP4-induced C3H10T1/2 cell commitment, we hypothesized that their introduction into C3H10T1/2 cells might bias cells toward a preadipocyte line- age. Because miR-140 and miR-140* come from the same precursor, their overexpression could not be separated. However, because miR-140* is usually degraded merely as a carrier strand, and miR-140 has been reported to function during mesoderm development, we focused our research mainly on miR-140. We constructed the miRNA-overexpressing plasmid MSCV-miR-140/140* by cloning the miR-140/140* precursor into the MSCV vector, followed by infection in C3H10T1/2 cells. Realtime PCR confirmed that the overexpression was effective ( Fig.  2A). miR-140/140*-overexpressing C3H10T1/2 cells could partially differentiate into mature adipocytes even without BMP4 treatment. Furthermore, treatment with BMP4 before confluence significantly boosted their adipogenic capacity as indicated by Oil Red O staining of fat droplet accumulation and Western blot analysis of adipocyte-specific markers (Fig. 2, B and C). These results, combined with the miRNA expression pattern showed in Fig. 1E, indicate that miRNAs play a critical role during the adipocyte lineage commitment of C3H10T1/2 cells. Thereafter, the miR-140 sponge vector was used to decrease miR-140 expression in miRNA-overexpressing C3H10T1/2 cells after BMP4 treatment (27). The effect of the sponge is shown in Fig. 3E. Consistent with our assumption, the increase in adipogenesis upon miRNA overexpression was significantly decreased after the introduction of the miR-140 sponge (Fig. 2, E and F). This confirms that miR-140 plays a very important role during commitment.

Identification of OSTM1 as a Target Protein of miR-140 during Adipocyte Lineage Commitment of C3H10T1/2 Cells-Pic-
Tar, TargetScan, and Miranda were used to identify putative target proteins of miR-140 that overlapped among these three algorithms (28). We found that most of the genes in the list matched those from the Dual-Luciferase experiments (supplemental Fig. S2); however, only a few of them could be eventually demonstrated to be coordinately regulated during the same physiological process. Accordingly, Ostm1 was markedly down-regulated at both the protein and mRNA levels when miR-140 was overexpressed in C3H10T1/2 cells (Fig. 3, A and  B). Furthermore, the repressive function of BMP4 or miR-140 on the expression of Ostm1 could be predominantly reversed by the miR-140 sponge, which could bind to miR-140 and diminish its repression of the target protein (Fig. 3C).
According to the sequence analysis, Ostm1 contains two 7-nucleotide sites within its 3Ј-UTR, which matched the seed region of miR-140. These two putative miRNA-binding sites were cloned into the psiCHECK2 vector, respectively, and only one site bound miR-140 (Fig. 3D). Compared with the negative control miRNA mimic, cotransfection of the Ostm1 3Ј-UTRluciferase reporter with the miR-140 mimic resulted in significantly repressed luciferase activity. Correspondingly, miR-140 had no effects on the mutant Ostm1 3Ј-UTR-reporters. Furthermore, when the miR-140 sponge vector was used to neutralize the effects of the miR-140 mimic, no difference in luciferase activity could be observed between control mimic-and miR-140 mimic-transfected cells (Fig. 3E). These results indicate that Ostm1 is a direct target of miR-140.
Ostm1 Is Identified as an Anti-adipogenic Factor, and Its Expression Is Decreased by BMP4 through miR-140 during Adipocyte Lineage Commitment of C3H10T1/2 Cells-Western blotting was used to detect OSTM1 expression during the commitment stage in C3H10T1/2 cells. Compared with control C3H10T1/2 cells, OSTM1 expression was significantly decreased in BMP4-treated cells (Fig. 4A). In previous studies, SMAD and p38 MAPK were downstream of the BMP4 signaling pathway in the commitment of C3H10T1/2 stem cells to adipocyte lineage. Specific Stealth TM RNAi was used to define regulation of the BMP signaling pathway upon expression of miR-140 (10). miR-140 was observed to be significantly repressed by p38 MAPK and Smad4 RNAi, respectively (Fig.  4B). In addition, p38 MAPK and Smad4 RNAi successfully res-  A and B, Western blotting and qRT-PCR were performed to detect Ostm1 expression during the commitment of C3H10T1/2 cells. Each column represents the mean Ϯ S.D. of three independent experiments. *, p Ͻ 0.05. C, MSCV-miR-140/140*-infected C3H10T1/2 cells were transfected with empty pEGFP or pEGFP-miR-140 sponge vector and treated with BMP4 or not until post-confluent. Cells were then subjected to Western blotting to detect OSTM1 expression. D, schematic of the miR-140 putative target site in the Ostm1 3Ј-UTR. mmu, Mus musculus. E, 293T cells were transfected with 10 ng of empty psiCHECK2, psiCHECK2-Ostm1, or psiCHECK2mutant Ostm1 vector and 0.8 g of empty pEGFP or pEGFP-miR-140 sponge vector. Cells were cotransfected with 100 nM miR-140 mimic or a negative control (NC). Firefly and Renilla luciferases were measured in cell lysates, and values were normalized to the psiCHECK2 vector and are presented as -fold change. Each bar represents the mean Ϯ S.D. of three independent experiments. **, p Ͻ 0.01.

miR-140 Promotes Adipocyte Lineage Commitment
cued BMP4-induced Ostm1 repression compared with control RNAi (Fig. 4C), and their relationship was further confirmed in aP2-BMP4 transgenic mice, in which BMP4 is overexpressed specifically in adipocytes (Fig. 4D). The stromal vascular fraction was isolated from the adipose tissue of wild-type and transgenic mice. Ostm1 expression was dramatically decreased in BMP4-overexpressing mice in correlation with miR-140 upregulation in the stromal vascular fraction (Fig. 4, E and F).
These results indicate that Ostm1 is decreased by BMP4 through miR-140 during adipocyte lineage commitment.
On the basis of the above results, we assumed that OSTM1 might function as an anti-adipogenic factor. As shown in Fig. 5  (C and D), adipocyte differentiation was significantly impaired in Ostm1-overexpressing C3H10T1/2 cells as indicated by Oil Red O staining and Western blot analysis of adipocyte-specific markers. Moreover, when Ostm1 expression was knocked down by siRNA (Fig. 5F), an apparent increase in adipogenic ability was observed. A half-dosage of BMP4 resulted in Ͼ70% adipogenesis when Ostm1 was knocked down compared with Ͻ10% adipogenesis in the control (Fig. 5, G and H). To further confirm that targeting Ostm1 is important for the pro-adipogenic function of miR-140, knockdown of Ostm1 was performed in miR-140-overexpressing C3H10T1/2 cells. In the presence of BMP4 treatment, there was no significant difference between the control group and the Ostm1 knockdown group (supplemental Fig. S3), possibly because BMP4 treatment of miR-140-overexpressing cells would lead to nearly total adipogenesis. However, in the absence of BMP4 treatment, knockdown of Ostm1 enhanced the pro-adipogenic effect of miR-140 (supplemental Fig. S3), indicating that Ostm1 is an important target of miR-140 during adipocyte lineage commitment. These results demonstrate that Ostm1 functions as an anti-adipogenic factor and is an important target gene of miR-140 in BMP4-induced commitment of C3H10T1/2 stem cells.

DISCUSSION
Mesenchymal stem cells are multipotent stem cells that can differentiate into a variety of cell types, including osteocytes, chondrocytes, myocytes, and adipocytes (8,29). Adipogenesis . BMP4 decreases Ostm1 expression through miR-140. A, cell lysates were collected each day during commitment and subjected to Western blotting. d, day. B and C, C3H10T1/2 stem cells were plated at 30% confluence and transfected with Smad4 or p38 MAPK Stealth TM RNAi; and 24 h later, they were treated with BMP4 or not until post-confluent. Expression of miR-140 and miR-140* was determined by qRT-PCR, and expression of p38 MAPK, SMAD4, and OSTM1 was determined by Western blotting. Each bar represents the mean Ϯ S.D. of three independent experiments. D, transgenic (TG) mice overexpress BMP4 in adipose tissue. Western blotting was performed to detect BMP4 expression levels in white adipose tissue (subcutaneous) from wild-type and aP2-BMP4 transgenic mice. E and F, the stromal vascular fraction was isolated from wild-type and transgenic mice to determine miRNA and protein levels. *, p Ͻ 0.05. MARCH 22, 2013 • VOLUME 288 • NUMBER 12

JOURNAL OF BIOLOGICAL CHEMISTRY 8227
includes two sequential processes: lineage commitment into preadipocytes and terminal differentiation into mature adipocytes (3,4). Compared with the second process, many questions are unresolved concerning commitment. Previously, we showed that C3H10T1/2 cells can commit to adipocyte lineage upon BMP4 treatment (4). In the present study, miR-140, which was previously recognized as a chondrocyte-specific miRNA (25-27, 30 -33), was identified as a downstream component of the BMP4 signaling pathway during C3H10T1/2 cell commitment. We consider that it might function as one of the commitment factors. Its overexpression in C3H10T1/2 cells validates our hypothesis. Exogenously expressed miR-140 in C3H10T1/2 cells significantly increases adipocyte differentiation even without BMP4, whereas miR-140 knockdown signif- Each bar represents the mean Ϯ S.D. of three independent experiments. C and D, upon reaching post-confluence, cells were induced to differentiate with the standard adipocyte differentiation protocol, and on day 7, Oil Red O staining and Western blotting were performed to confirm adipogenesis. E, the relative intensity of the Western blots in D was determined in three independent experiments. F, C3H10T1/2 cells were transfected with Ostm1 Stealth TM RNAi at 30% confluence; and after 24 h, they were cultured with a half-dosage of BMP4 (10 ng/ml) until post-confluent and then induced to differentiate with the standard differentiation protocol. E, relative intensity of western blotting in D was performed through three independent experiments. *, p Ͻ 0.05; **, p Ͻ 0.01. G and H, the accumulation of cytoplasmic triglyceride was detected by Oil Red O staining on day 7, at which point, the cells were photographed. The expression of adipocyte markers (422/aP2 and peroxisome proliferator-activated receptor ␥ (PPAR␥)) was also detected with cell extract on day 7. I, the relative intensity of the Western blots in H was determined in three independent experiments. I, relative intensity of western blotting in H was performed through three independent experiments. *, p Ͻ 0.05; **, p Ͻ 0.01.

miR-140 Promotes Adipocyte Lineage Commitment
icantly decreases adipocyte differentiation. This is the first time that a particular miRNA directly involved in adipocyte lineage commitment has been identified.
Previous studies have shown that miR-140 is prevalently expressed in normal cartilage and regulates cartilage development and homeostasis. The cartilage master regulator SOX9 promotes miR-140 expression. In addition, Postlethwait and co-workers (30) have shown that miR-140 regulates palatogenesis in zebrafish via suppressing PDGF receptor ␣. Kobayashi and co-workers (33) used miR-140-null mice to demonstrate that miR-140 is essential for normal endochondral bone development and that the miR-140 target Dnpep reduces BMP signaling to function in skeletal defects in the mouse model. These results, combined with our research, indicate that miR-140 is essential for mesodermal tissue development and that there is a positive relationship between miR-140 and BMP signaling.
We have identified a direct miR-140 target protein, OSTM1, which is significantly decreased after BMP4 treatment. When we overexpressed Ostm1 in C3H10T1/2 cells, adipogenic differentiation was significantly decreased, and knockdown led to the opposite result. Therefore, we consider OSTM1 function to be an anti-adipogenic factor in our system. A previous investigation indicates that OSTM1 promotes ␤-catenin/Lef1 interaction (24). If is well known that the Wnt/␤-catenin signaling pathway affects multiple cellular functions, and inhibition of this pathway would favor adipogenesis by mesenchymal stem cells (34,35). Accordingly, we predict that Ostm1 repression may lead to a decrease in the Wnt/␤-catenin signaling pathway and favor adipogenic differentiation. The crosstalk between the BMP and Wnt pathways has long been recognized and is considered to be an essential growth factor crosstalk that occurs during the entire life of an animal (36). These two signals interact at multiple levels, and our work presents the possibility that OSTM1 may serve as a new linker molecule between the BMP and Wnt pathways.
Nonetheless, a physical process must include combinatorial growth or differentiation signaling repertoires. Other undisclosed miR-140 targets may contribute to adipocyte lineage commitment. Meanwhile, other miRNAs may also function as promotion or suppression factors in adipocyte lineage commitment. More studies are needed to better explain the whole physiological process.
In conclusion, we have demonstrated that miR-140 functions as a positive regulator of adipocyte lineage commitment in response to BMP4 treatment, which in turn decreases Ostm1 at both the mRNA and protein levels and leads to adipogenesis. This is the first time report that a specific miRNA participates in BMP4-induced C3H10T1/2 adipocyte lineage commitment. Therefore, our results offer new insight into the mechanisms of lineage commitment.