Male C57BL/6, ob/ob and db/db mice, aged 10–12 weeks, were purchased from the Shanghai Laboratory Animal Company (Shanghai, China). All mice were housed and bred according to international standard conditions, with a 12-h dark/light cycle. For induction of obesity, mice aged 10 weeks were placed on a high-fat-diet (50 kcal% fat, 25 kcal% protein and 25 kcal% carbohydrate) for 12 weeks. The normal diet consisted of 10 kcal% fat, 15 kcal% protein and 75 kcal% carbohydrate. The animal protocol was reviewed and approved by the Animal Care Committee of Chongqing Ninth People's Hospital (Chongqing, China).

Human subcutaneous fat tissues from obese (BMI>25; n=25) and lean (BMI<23; n=20) adults were collected at the Department of Endocrinology at Chongqing Ninth People's Hospital. All subjects had a stable weight, defined as the absence of fluctuations of >1% of body weight for at least 2 months prior to surgery. Adipose tissue was immediately frozen in liquid nitrogen following extraction. All samples were obtained with informed consent and the study was approved by the Chongqing Ninth People's Hospital Institutional Review Board.

3T3-L1 fibroblasts were obtained from The Cell Bank of Type Culture Collection of Chinese Academy of Sciences (Shanghai, China), and cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco-BRL, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS; Gibco-BRL), 100 IU/ml penicillin and 100 µg/ml streptomycin (Gibco-BRL). Adipocyte differentiation was induced by treating the cells at 80–90% confluency with standard differentiation inducers [100 µM IBMX (Beyotime Institute of Biotechnology, Nantong, China), 100 nM dexamethasone (Beyotime Institute of Biotechnology) and 10 µg/ml insulin (Eli Lilly & Co., Indianapolis, IN, USA)] for 48 h (from day 0 to day 2). Cells were treated with DMEM supplemented with 10 µg/ml insulin and 10% FBS for another 48 h (from day 2 to day 4). Subsequently, the medium was replaced by DMEM and 10% FBS, which was changed daily, and the cells were cultured for a further 4–6 days.

Total RNA was isolated from tissues or cells using the TRIzol reagent (Invitrogen Life Technologies, Shanghai, China), and reverse transcription was performed using the RT kit from Promega Corporation (Madison, WI, USA), according to the manufacturer's instructions. Affymetrix arrays were performed using Mouse Genome 2.0 chips from Gene Tech Company Limited (Shanghai, China). In order to quantify the transcripts of the interest genes, qPCR was performed using a SYBR Green Premix Ex Taq kit (Takara Bio Inc., Dalian, China) on a Light Cycler 480 (Roche, Basel, Switzerland) with 5 mg RNA. The following primer sequences form Bioyare Biotechnology Company (Shanghai, China) were used: Mouse Grem2, forward 5′-TGTGCTGGTAAAGGTAGCTGA-3′ and reverse 5′-CCACCTCTCTGAGTTGTTGCT-3′; human Grem2, forward 5′-ATCCCCTCGCCTTACAAGGA-3′ and reverse 5′-TCTTGCACCAGTCACTCTTGA-3′; C/EBPa, forward 5′-CAAGAACAGCAACGAGTACCG-3′ and reverse 5′-GTCACTGGTCAACTCCAGCAC-3′; PPARg, forward 5′-CTCCAAGAATACCAAAGTGCGA-3′ and reverse 5′-GCCTGATGCTTTATCCCCACA-3′; LPL, forward 5′-ATGGATGGACGGTAACGGGAA-3′ and reverse 5′-CCCGATACAACCAGTCTACTACA-3′; PEPCK, forward 5′-ATGAGCCGCAAACTGGGTC-3′ and reverse 5′-AGAGCCGAACTCCACAATCTC-3′; CD36, forward 5′-AGCATTCAACGCCAGGTTC-3′ and reverse 5′-CGAGTCTGTCAGTTCAATACCAA-3′; aP2, forward 5′-AAGGTGAAGAGCATCATAACCCT-3′ and reverse 5′-TCACGCCTTTCATAACACATTCC-3′; mouse HPRT1, forward 5′-TCAGTCAACGGGGGACATAAA-3′ and reverse 5′-GGGGCTGTACTGCTTAACCAG-3′; human HPRT1, forward 5′-CCTGGCGTCGTGATTAGTGAT-3′ and reverse 5′-AGACGTTCAGTCCTGTCCATAA-3′. The PCR cycling conditions were as follows: Initial period at 94°C for 3 min, followed by a two-step PCR program consisting of 95°C for 5 sec and 60°C for 30 sec for 55 cycles. Gene expression was calculated relative to the mRNA expression of HPRT1, and relative quantitation analysis of gene expression data was performed using the 2^−ΔΔCt method.

Retroviruses containing empty vector or Grem2 cDNA were constructed by Invitrogen Life Technologies. siRNA oligos targeting Grem2 or β-catenin were purchased from Gene Pharma (Shanghai, China) and transfected into 3T3-L1 cells using Lipofectamine 2000 (Invitrogen Life Technologies), according to the manufacturer's instructions. Histological analysis was conducted using Oil Red O staining. Cells were fixed with 4% paraformaldehyde for 30 min, followed by incubation with Oil Red O (Sigma-Aldrich, St. Louis, MO, USA) for 45 min.

Cells or tissues were harvested and lysed with lysis buffer (150 mM NaCl, 50 mM Tris-HCl and 1% NP-40; pH 7.5). Proteins were quantified and separated by 10% SDS-PAGE and transferred to polyvinylidine difluoride membranes (Amersham Bioscience, Buckinghamshire, UK). After blocking with 10% non-fat milk in phosphate-buffered saline, membranes were immunoblotted with the following rabbit antibodies from Abcam (Cambridge, MA, USA): Polyclonal anti-Grem2 (ab102563; 1:1,000), monoclonal C/EBPα (ab40764; 1:1,000), polyclonal PPARγ (ab19481; 1:1,000), monoclonal P-P38 (ab178867; 1:2,000), polyclonal P-AKT (ab66138; 1:1,000), monoclonal P-CREB (ab32096; 1:2,000), monoclonal CREB (ab32515; 1:2,000), polyclonal β-catenin (ab6302; 1:1,000), polyclonal lamin B (ab16048; 1:2,000) and monoclonal GAPDH (ab181602; 1:1,000). The blots were then incubated with horseradish peroxidase-linked secondary antibodies (Cell Signaling Technology, Inc. Danvers, MA, USA). The signals were detected using the SuperSignal West Pico Chemiluminescent Substrate kit (Pierce, Rockford, IL, USA) according to manufacturer's instructions.

The results are expressed as the mean ± standard error of the mean and analyzed using Student's t-test or one way analysis of variance with GraphPad Prism software, version 5.0.1 (GraphPad Software, Inc., La Jolla, CA, USA). P<0.05 was considered to indicate a statistically significant difference.

To identify potential genes involved in adipogenesis and obesity, male C57BL/6 mice aged 10 weeks were placed on high-fat or normal diets for 12 weeks. The mice were sacrificed by cervical dislocation and gene expression arrays were performed using epididymal adipose tissues. The clustering analysis of the Affymetrix arrays revealed that 1,174 genes were upregulated and 1,572 genes were downregulated in the adipose tissues of mice fed a high-fat-diet compared with those fed a normal diet (P<0.05; data not shown). Among these genes, it was determined that Grem2 was significantly downregulated (P<0.001). Its downregulation was further confirmed by RT-qPCR and western blot analysis (Fig. 1A and B).

To determine whether the inhibition of Grem2 represents a common feature of obesity, db/db and ob/ob mice were employed as models for monogenic-induced obesity. Indeed, Grem2 mRNA and protein levels were found to be reduced in the epididymal adipose tissues of these mice, as compared with lean controls (Fig. 1C–F). Notably, Grem2 mRNA and protein levels were significantly reduced in subcutaneous fat from obese humans (Fig. 1G and H), further suggesting that the downregulation of Grem2 expression may be a conserved feature in obese mice and humans.

To explore the functions of Grem2 in adipogenesis, its gene expression was investigated at numerous time points during 3T3-L1 differentiation. The RT-qPCR results revealed that Grem2 mRNA gradually reduced from day 0 post-induction (Fig. 2A). The protein levels for Grem2 were consistent with its mRNA levels (Fig. 2B).

The role of Grem2 in adipogenesis was tested using retroviruses carrying Grem2 cDNA or empty vector, which resulted in forced overexpression of Grem2 in 3T3-L1 preadipocytes (Fig. 3A). These cells were cultured to confluence and differentiated using the standard induction cocktail for 8 days. Cell morphology indicated that cells overexpressing Grem2 accumulated significantly less lipids than the control cells (Fig. 3B). In addition, the expression levels of adipogenic regulators such as C/EBPα and PPARγ were reduced (Fig. 3C and D), which was consistent with the Oil Red O staining. Furthermore, markers of mature adipocytes LPL, PEPCK, CD36 and aP2, which are downstream targets of C/EBPα and PPARγ, were repressed by Grem2 overexpression (Fig. 3E).

To further examine the roles of Grem2, its specific siRNA was transfected into 3T3-L1 cells (Fig. 4A). As a result, knockdown of Grem2 levels showed a clear increase in the adipogenic potential of the 3T3-L1 cells compared with that of the cells treated with the negative control siRNA, as illustrated by cell morphology (Fig. 4B). Furthermore, knockdown of Grem2 increased expression levels of the adipocyte markers including C/EBPα, PPARγ, LPL, PEPCK, CD36 and aP2 (Fig. 4C and D). Together, these results indicate that Grem2 is a critical negative regulator in the adipogenesis.

Finally, the molecular basis for the observed phenotypic changes in 3T3-L1 cells with Grem2 overexpression was investigated. As shown in Fig. 5A, Wnt signaling was highly activated by Grem2 overexpression, as evidenced by nuclear accumulation of β-catenin, while other signaling molecules, including p38, AKT and cAMP/CREB, remained unaffected (Fig. 5B). Subsequently, it was investigated whether ablation of β-catenin reversed the roles of Grem2. A retrovirus targeting a β-catenin coding region (sh-catenin) was generated using pAD_BLOCK_IT_DEST vectors (Invitrogen Life Technologies, Grand Island, NY, USA) to knockdown endogenous β-catenin expression in 3T3-L1 cells (Fig. 5C). As a result, sh-catenin largely attenuated the inhibitory roles of Grem2 on the adipogenesis, as evidenced by gene expression of adipogenic markers (Fig. 5D), suggesting that the role of Grem2 in the adipogenesis, at least in part, relies on Wnt/β-catenin activation.