Abstract
Berardinelli-Seip congenital lipodystrophy type 2 (BSCL2) is the most severe form of human lipodystrophy and is caused by loss-of-function mutations in the BSCL2/seipin gene. Exactly how seipin may regulate adipogenesis remains unclear. A recent study in vitro suggested that seipin may function to inhibit the activity of glycerol-3-phosphate acyltransferases (GPATs), and increased GPAT activity may be responsible for the defective adipogenesis under seipin deficiency. Here we generated Seipin−/−Gpat3−/− mice, which had mild but significant recovery of white adipose tissue mass over Seipin−/− mice. The mass of brown adipose tissue (BAT) of the Seipin−/−Gpat3−/− mice was almost completely restored to normal level. Importantly, the Seipin−/−Gpat3−/− mice showed significant improvement in liver steatosis and insulin sensitivity over Seipin−/− mice, which is attributable to the increased BAT mass and to the enhanced browning of the subcutaneous fat of the Seipin−/−Gpat3−/− mice. Together, our results establish a functional link between seipin and GPAT3 in vivo and suggest that GPAT inhibitors may have beneficial effects on BSCL2 patients.
Introduction
Congenital generalized lipodystrophy (CGL, also known as Berardinelli-Seip congenital lipodystrophy/BSCL), is an autosomal recessive disorder characterized by a near total loss of adipose tissue, severe insulin resistance and fatty liver (1, 2). To date, mutations in four genes have been linked to CGL/BSCL, including 1-acylglycerol-3-phosphate-O-acyl transferase 2 (AGPAT2)/CGL1, seipin/CGL2, caveolin/CGL3 and cavin/CGL4 (3). The most severe form of human CGL/BSCL is caused by mutations in seipin/BSCL2, which encodes an evolutionarily conserved, integral membrane protein of the endoplasmic reticulum (ER) (3–5). Interestingly, seipin and its orthologs also regulate the biogenesis and expansion of lipid droplets (LDs) in all eukaryotic cells (6–12). Therefore, seipin is a unique protein that can regulate lipid storage at both systemic (adipogenesis) and cellular (LD biogenesis and expansion) levels. Notably, recent studies have also implicated seipin in the biogenesis of peroxisomes and in the regulation of calcium homeostasis (13–16).
An integral membrane protein of the ER, seipin is conserved from yeast to man (3, 7, 8). Recent structural studies demonstrated that both fly and human seipin exist as homo-oligomers, and that the evolutionarily conserved lumenal domain forms an eight-stranded β sandwich fold which resembles lipid-binding domains, such as that of Niemann Pick C2 (17, 18). The oligomerization state of human seipin is essential to its function (17). Seipin can bind 1-palmitoyl-2-oleoyl phosphatidic acid (POPA) and phosphatidylinositol 3-phosphate (PI3P) as shown by in vitro lipid binding assays (17). A hydrophobic helical domain of seipin can bind a monolayer of phospholipids, suggesting a role for seipin to recognize initiating LDs (18). Despite these progresses, the molecular function of seipin and how exactly it may regulate LD dynamics remain obscure.
Seipin knock-out (SKO) mice suffered from severe lipodystrophy and insulin resistance (19–21), proving an essential role of seipin in adipogenesis in vivo. Deleting seipin from mature adipose tissue also led to progressive lipodystrophy, demonstrating a role for seipin in adipose tissue maintenance (22, 23). Overexpressing seipin in adipose tissue increased lipolysis and reduced adipose mass (24). Exactly how seipin regulates adipogenesis and adipose tissue homeostasis is unknown. Studies have suggested that seipin deficiency leads to enhanced cAMP/PKA–mediated lipolysis and ATGL activation, thereby causing aborted adipocyte differentiation (20, 25). However, the differentiation of preadipocytes requires a transcriptional cascade that ultimately leads to activation of the master regulator of terminal adipogenesis: peroxisome proliferator-activated receptor-γ (PPARγ) (26–28), and metabolic disturbances associated with seipin deficiency can be partially reversed by activating PPARγ (21). Seipin deficiency has been shown to accumulate certain phospholipid species, such as phosphatidic acid (PA) (29–31), which are strong PPARγ antagonists (32, 33). The accumulation of PA may impact LD formation, block PPARγ activity and cause lipodystrophy (3, 32–36). Our previous results suggested that seipin may function to inhibit the activity of ER-localized glycerol-3 phosphate acyl transferases (GPATs) (34), whose activation in seipin deficiency may lead to accumulation of PA and blunted adipogenesis. Importantly, depleting GPAT3 or applying GPAT inhibitors partially rescued the differentiation defects of seipin-deficient preadipocytes in vitro (34).
To examine the functional relationship between seipin and GPAT3 under in vivo settings, we have now generated Seipin−/-GPAT3−/− double knockout (DKO) mice. Compared with the Seipin−/− mice, the DKO mice showed significantly reduced hepatic steatosis, as well as improved insulin sensitivity. Importantly, the white adipose tissue of the DKO mice showed prominent signs of browning. Our results establish a functional link between seipin and GPAT3 in vivo and suggest that GPAT inhibitors may have beneficial effects on BSCL2 patients.
Results
Generation and analyses of the Seipin-GPAT3 double knockout mice
We generated Seipin and GPAT3 double knockout (DKO) mice using the CRISPR/Cas9 system. The sgRNA target sites (Fig. 1A) of Seipin and GPAT3 were designed to target the second exon of Seipin and GPAT3 genes, and the two sgRNAs mixed with the mRNA of Cas9 were injected into the zygote together to generate the DKO mice. A variety of mutant forms were identified in F0 mice, including Seipin mutant forms (1 nt insertion, 8 nt deletion and 13 nt deletion) and GPAT3 mutant forms (1 nt deletion, 6 nt deletion and 7 nt deletion) (Fig. 1A). For convenience of genotyping, we subsequently chose the strain carrying Seipin 13 nt deletion and GPAT3 7 nt deletion for breeding. F0 and F1 generations were sequenced to identify gene mutations. The genotype after F2 generation was identified by agarose gel electrophoresis of PCR products (Fig. 1B). A 140 bp band was amplified in Seipin+/+ mice, while a 127 bp band was amplified in the Seipin 13 nt deletion homozygotes, and 3 bands were detected from heterozygotes. Similarly, wild-type GPAT3 showed a 120 bp band, 7 nt deletion showed a 113 bp band and heterozygotes displayed 3 bands by electrophoresis (Fig. 1B). To verify the gene disruptions, we examined the Seipin and GPAT3 mRNA and protein expression in the subcutaneous (SubQ) fat and testes (Fig. 1C–F) of three-month old wild type (WT), Seipin single knockout (SKO), GPAT3 knockout (G3-KO) and Seipin-GPAT3 double knockout (DKO) mice. At the mRNA and protein levels, no seipin expression was detected in SKO and DKO mice, and no GPAT3 expression was detected in G3-KO and DKO mice (Fig. 1C–F). Because SKO mice showed dyslipidemia (19), the levels of plasma total cholesterol (TC), triglycerides (TG) and non-esterified free fatty acids (NEFA) of the four groups of mice after 16 hours of fasting were measured. Compared with WT mice, the TC of SKO mice was increased but that of the DKO mice was restored to WT level (Fig. 1G). Compared to WT, TG of SKO mice decreased severely, and TG of G3-KO mice decreased slightly; DKO mice had the lowest TG levels (Fig. 1H). The change of NEFA was similar to that of TG: SKO and G3-KO mice had lower levels of NEFA than that of the WT mice, while DKO mice had the lowest (Fig. 1I).
Generation and analyses of Seipin-GPAT3 double knockout mice. (A) sgRNA targeting sequence and mutations detected: sgRNA target sequence is marked in red and PAM motif marked in green. In: insertion type mutation, D: deletion type mutation. (B) PCR genotyping results. Seipin: WT, a 140 bp band; KO, a 127 bp band; heterozygotes (+/−), 3 bands; GPAT3: WT, a 120 bp band; KO, a 113 bp band, +/−, 3 bands. (C, D) Seipin and GPAT3 mRNA expression of subcutaneous fat (C) and epididymal fat (D) from four strains of mice as examined by qPCR. WT, wild type mice; SKO, seipin knockout mice; G3-KO, GPAT3 knockout mice; DKO, seipin and GPAT3 double knockout mice. n = 7 for each genotype. (E, F) Seipin and GPAT3 protein expression of subcutaneous fat (E) and epididymal fat (F) as examined by western blotting. NS:non-specific band. (G–I) Plasma total cholesterol (G), triglycerides (H), non-esterified free fatty acids (I) of 3-month old male mice. (WT: n = 5, SKO: n = 5, G3-KO: n = 6, DKO: n = 8. **P < 0.01, ***P < 0.001 as compared with WT; #P < 0.05, ##P < 0.01 as compared with SKO).
Characterization of adipose tissue
The white adipose tissue (WAT) of 3-months old male mice was analyzed. By general observation, the amount of subcutaneous fat and epididymal fat of SKO mice decreased significantly when compared to WT mice. Those of the DKO mice, especially subcutaneous fat, showed improvement when compared to SKO mice (Fig. 2A). The weight of subcutaneous fat and epididymal fat of DKO mice was about 2 times higher than those of SKO mice (Fig. 2B and C). HE-staining of subcutaneous fat and epididymal fat tissue sections showed that the number of LDs in subcutaneous fat and epididymal fat of SKO mice decreased significantly as compared with WT, and the matrix tissue increased (Fig. 2D and E). Compared with SKO mice, the number of LDs in the subcutaneous fat of DKO mice increased significantly, and their sizes were more uniform (Fig. 2D). There are also more and larger LDs in the epididymal fat of DKO mice than that of SKO mice (Fig. 2E). We examined mRNA expression of several genes (Fabp4, Pparg, Cebpa, Fasn, Scd1, Cd36, leptin and adiponectin) that are highly expressed in normal adipose tissue. Compared with WT mice, the expression of these genes in the subcutaneous fat and epididymal fat of SKO mice was significantly decreased (Fig. 2F and G). The expression of Fabp4, Pparg, Cebpa, Fasn, Cd36 and adiponectin of subcutaneous fat of DKO mice was significantly higher than those of SKO mice (Fig. 2F), suggesting that the function of subcutaneous fat of DKO mice improved. However, only Fasn in epididymal fat of DKO mice was significantly higher than that of SKO mice (Fig. 2G). Importantly, the plasma level of leptin increased by ~ 200% and that of adiponectin by ~ 300%, in the DKO mice than the SKO mice (Fig. 2H and I), suggesting that the increased WAT of the DKO mice is functional. Immunohistochemistry staining with antibody against F4/80 showed reduced inflammation in subcutaneous and epididymal fat of the DKO mice relative to SKO mice (Fig. 2J and K). Masson's trichrome staining revealed significant reduction of connective tissue in the subcutaneous adipose tissue of the DKO mice relative to SKO mice, with the reduction in subcutaneous adipose tissue a lot more striking (Fig. 2L and M).
Characterization of the white adipose tissue. (A) General morphology of subcutaneous fat and epididymal fat from WT, SKO, G3-KO and DKO mice. (B, C) Weight of subcutaneous fat (B) and epididymal fat (C). (WT: n = 5, SKO: n = 5, G3-KO: n = 7, DKO: n = 9. ***P < 0.001 as compared with WT, ##P < 0.01 as compared with SKO). (D, E) HE staining of subcutaneous fat (D) and epididymal fat (E). Scale bar: 100 μm. (F, G) Expression of indicated genes of subcutaneous fat (F) and epididymal fat (G) as examined by qPCR. (n = 7. *P < 0.05, **P < 0.01 compared with WT, #P < 0.05, ##P < 0.01 compared with SKO). (H, I) Plasma leptin and adiponectin level of 3-month of male mice. (n = 6. ***P < 0.001 as compared with WT, #P < 0.05, ##P < 0.01 as compared with SKO). J and K: F4/80 immunochemical staining of subcutaneous fat (J) and epididymal fat (K). Scale bar: 100 μm. (L, M) Masson staining of subcutaneous fat (L) and epididymal fat (M). Scale bar: 100 μm.
Lipid droplet (LD) formation and differentiation of MEF cells
Seipin is known to regulate both LD formation and adipogenesis. Mouse embryonic fibroblasts (MEFs) were isolated from WT and knockout animals for analyzing lipid droplet morphology and adipogenesis. After oleic acid incubation, a large number of giant LDs appeared in the MEF cells of SKO mice, but not in those from DKO mice (Fig. 3A and B). The MEFs of SKO mice almost completely lost their ability to differentiate, whereas those from DKO mice could partially differentiate (Fig. 3C).
Lipid droplet (LD) formation and differentiation of MEF cells. (A) BODIPY-stained LDs in MEF cells incubated with 200 μM oleic acid for 24 hours. Representative confocal images are shown. Scale bar: 10 μm. (B) Diameters of the largest lipid droplets in 100 of each group were measured to determine the size of LDs. (**P < 0.01 as compared with WT, ##P < 0.01 as compared with SKO). (C) Light microscopy (Upper) or Oil red O (bottom) staining of differentiated MEF cells. Scale bar: 100 μm.
Fatty liver and insulin resistance
CGL2/BSCL2 patients and SKO mice are characterized by severe fatty liver and insulin resistance. Therefore, lipid deposition in the liver and insulin sensitivity were examined in SKO and DKO mice. Compared to SKO mice, liver weight and TG content were significantly reduced in the DKO mice (Fig. 4A and B). No significant difference was observed for total, free or esterified cholesterol between SKO and DKO mice (Fig. 4C). H&E and Oil red O staining of liver tissue sections showed severe hepatocyte steatosis in SKO liver, which almost disappeared in DKO liver (Fig. 4D and E).
Fatty liver and insulin resistance. (A, B) liver weight (A) and liver triglycerides contents (B). (WT: n = 5, SKO: n = 5, G3-KO: n = 7, DKO: n = 9. ***P < 0.001 as compared with WT, ##P < 0.01, ###P < 0.001 as compared with SKO). (C) Liver total cholesterol (TC), free cholesterol (FC) and cholesterol ester (CE) contents. (n = 8). (D, E): HE (D) and Oil red O (E) staining of the liver. Scale bar: 100 μm. (F) Plasma glucose level. (WT: n = 5, SKO: n = 5, G3-KO: n = 6, DKO: n = 8. **P < 0.01 as compared with WT). (G) Plasma insulin level (WT: n = 5, SKO: n = 6, G3-KO: n = 5, DKO: n = 6. ***P < 0.001 as compared with WT, ##P < 0.01 as compared with SKO). (H, I) Glucose tolerance test (H) and quantification of figure H (I). AUC: the area under the curve. WT: n = 4 SKO: n = 5, G3-KO: n = 6, DKO: n = 8. *P < 0.05, **P < 0.01 as compared with WT, #P < 0.05, ##P < 001 compared with SKO). (J, K) Insulin tolerance test (J) and quantification of figure J (K) AUC: the area under the curve. (WT: n = 7 SKO: n = 6, G3-KO: n = 7, DKO: n = 6. **P < 0.01 as compared with WT, #P < 0.05, ##P < 001 as compared with SKO).
Plasma glucose and insulin levels of 3-month old male mice were examined. Compared to WT mice, plasma glucose of SKO mice was slightly but significantly increased, while that of DKO mice was unchanged (Fig. 4F). Plasma insulin of the SKO mice increased more than 7 fold of WT, while plasma insulin of the DKO mice was reduced to less than 70% of that of the SKO mice (Fig. 4G). Importantly, results from glucose tolerance and insulin tolerance tests in four groups of mice showed significant glucose intolerance and insulin resistance in SKO mice, which were rescued in the DKO mice (Fig. 4H–K). Thus, GPAT3 deficiency in SKO mice can significantly improve hepatic steatosis, glucose tolerance and insulin sensitivity.
Brown fat activity and WAT browning
While the extent of fat tissue recovery, especially the recovery of visceral fat, in the DKO mice over the SKO mice was fairly moderate, the improvement in liver steatosis and insulin sensitivity was rather striking in the DKO mice. Thus, besides the recovery of WAT, other beneficial changes may have also occurred in the DKO mice. We examined brown adipose tissue (BAT) and found that BAT mass and weight almost completely recovered in the DKO mice (Fig. 5A and B). Although lipid accumulation in the BAT of DKO mice did not show clear reduction as compared with that of SKO mice (Fig. 5C), UCP1 mRNA and protein level increased in the BAT of DKO mice when compared with the SKO mice (Fig. 5D–F). There is also much less inflammation in the BAT of DKO mice, compared with the SKO mice (Fig. 5G).
Characterization of the brown adipose tissue. (A) General morphology of the brown adipose tissue. (B) Weight of brown adipose tissue. (WT: n = 5, SKO: n = 5, G3-KO: n = 7, DKO: n = 9 ***P < 0.001 as compared with WT, ###P < 0.001 as compared with SKO). (C) HE staining of brown adipose tissue. Scale bar: 100 μm. (D) Expression of Ucp1 and Cidea in brown adipose tissue as examined by qPCR. (n = 7. **P < 0.01 as compared with WT, ##P < 0.01 as compared with SKO). (E) UCP1 protein expression of brown adipose tissue. (F, G) Immunohistochemical (IHC) staining of brown adipose tissue with antisera against UCP1 (F) and F4/80 (G). Scale bar: 100 μm.
We also noticed that the color of subcutaneous fat in the DKO mice appears darker than that from WT mice (Fig. 2). Indeed, BAT-specific markers were upregulated in both epididymal and subcutaneous fat of DKO mice (Fig. 6A and B). Strikingly, both epididymal and subcutaneous fat of DKO mice expressed a significant amount of UPC1 protein (Fig. 6C–F), indicating browning of WAT in the DKO mice.
White adipose tissue browning. (A, B) Expression of indicated genes in subcutaneous fat (A) and epididymal fat (B) as examined by qPCR. (n = 7. *P < 0.05, **P < 0.01 as compared with WT, #P < 0.05, ##P < 0.01 as compared with SKO). (C, D): UCP1 protein expression of subcutaneous fat (C) and epididymal fat (D) as examined by western blotting. (E, F) Immunohistochemical (IHC) staining of subcutaneous fat (E) and epididymal fat (F) with antisera against UCP1. Scale bar: 100 μm.
Discussion
Loss-of-function mutations of seipin are associated with the most severe form of human congenital lipodystrophy: CGL2/BSCL2. These patients suffer from a range of metabolic abnormalities, including extreme insulin resistance and hepatic steatosis. Currently, it is not yet understood how seipin regulates adipogenesis, and there are limited treatment strategies against CGL2/BSCL2. Metreleptin, a recombinant analog of leptin, has been shown to improve metabolic abnormalities in patients with CGL, and was approved by the US Food and Drug Administration in 2014 for treating CGL. While Metreleptin appears moderately effective, it doesn’t improve adipogenesis and avert fat loss (37). Additional treatment strategies targeting the root cause of CGL2/BSCL2 would be desirable.
We had previously demonstrated a functional relationship between seipin and ER-localized GPATs: GPAT activity was increased in seipin-deficient cells and tissues (34). Moreover, knocking down GPAT3 or application of GPAT inhibitors partially improved adipogenesis in seipin-deficient cells. To further examine the functional relationship between seipin and GPAT3 in vivo, we generated the Seipin−/−Gpat3−/− DKO mice. While WAT mass of the DKO mice was only moderately increased over that of the SKO mice (~2-fold increase), there was significantly reduced inflammation in the subcutaneous fat of the DKO mice, as compared with the SKO mice. Importantly, thermogenic markers such as UCP1 were dramatically upregulated only in the WAT of DKO mice, indicating increased WAT browning/beiging. The increased thermogenic activity of the WAT in the DKO mice may also help reduce ectopic fat accumulation.
The moderate increase of WAT mass of the DKO mice likely results from a partial recovery of the adipogenic potential of the preadipocytes. There are two ER-localized GPATs in eukaryotic cells: Gat1p and Gat2p in yeast and GPAT3 and GPAT4 in mammals (34). Although GPAT3 is upregulated by hundreds of folds during adipogenesis and account for ~ 80% GPAT activity in mature adipocytes (38), the house-keeping GPAT activity provided by GPAT4 in preadipocytes may compensate for the loss of GPAT3 during early adipogenesis. Therefore, limiting the activity of both GPAT3 and GPAT4 may be required for a full recovery of adipocyte differentiation. When we attempted to generate the Seipin−/−Gpat4−/− mice, we failed to even obtain any Gpat4−/− mice, suggesting that GPAT4 may be essential for embryonic development. It should be noted that the published Gpat4−/− mice were generated from a gene-trap cell line where there is a possibility of incomplete deletion (39). Further studies will address the essentiality of GPAT4 in embryogenesis.
The most significant finding of this work is the drastically reduced hepatic steatosis and improved insulin sensitivity of the DKO mice over the SKO mice. A few factors may contribute to this improvement. First, the fat storage capacity as judged by WAT mass increased at least two-fold in the DKO mice relative to SKO mice. Although small, such increased capacity may provide a critical buffer zone for fatty acid storage. Also, increased WAT can secrete more adipokines such as leptin and adiponectin. Second, BAT mass was almost completely restored in the DKO mice. Thus, more energy/fat may be consumed in the DKO mice. Third, the increased browning of WAT also helps consume fat and alleviate fatty liver and insulin resistance. The metabolic improvement of the DKO mice thus provides in vivo evidence for targeting GPAT3/4 enzymes in treating CGL2/BSCL2.
In summary, GPAT3 deficiency in a mouse model of BSCL2 dramatically improved hepatic steatosis and insulin resistance. Our results suggest that GPAT inhibitors may be explored as an additional therapeutic strategy against CGL2/BSCL2.
Materials and Methods
Animals
C57BL/6 J mice were purchased from Vital River Laboratories (Beijing, China). The animals were maintained on a 12 h light/12 h dark cycle at 24°C and were fed a standard laboratory chow diet with water ad libitum. The Principles of Laboratory Animal Care (NIH publication no.85Y23, revised 1996) was followed, and the experimental protocol was approved by the Animal Care Committee, Peking University Health Science Center (LA2010-059).
CRISPR/Cas system was used to generate Seipin and Gpat3 knockout mice. The sgRNA targeting sites were designed on exon 2 of Seipin and exon 2 of Gpat3 gene using Optimized CRISPR Design (http://crispr.mit.edu/). The targeting sequences of Seipin sgRNA: TCTTGGCAGGCCGTGCCCGC[AGG]; Gpat3 sgRNA: TCGGCCTTCGGATTATCCCT[GGG]. sgRNA template was amplified by PCR and then transcribed using MEGAshortscript Kit (Ambion) in vitro, followed by a purification of sgRNA using MEGAclear Kit (ambion), which was diluted into RNase-free water at a concentration of 200 ng/μl and stored at −80°C for future use. The plasmid PXT7 carrying the humanized cas9 cDNA was used as DNA template, and then linearized with XbaI digestion. Cas9 mRNA was transcribed with mMESSAGE mMACHINE T7 kit (Ambion) and purified by phenol/chloroform extraction, followed by isopropanol precipitation. Cas9 mRNA was diluted into RNase-free water at a concentration of 500 ng/μl and stored at −80°C for future use. Seipin and Gpat3 sgRNA (20 ng/μl) and cas9 mRNA (50 ng/μl) were co-injected into the cytoplasm of fertilized eggs with well-recognized pronucleus in M2 medium (Sigma). The injected zygotes were cultured in M16 medium (Sigma) for 0.5-2 hours. Then, injected embryos with normal morphology were transferred into one side oviduct (~20 embryos) through the fimbriae of pseudopregnancy female recipients that were mated with vasectomized males 1 day before.
For genotyping, genomic DNA was extracted from the toes of each mouse. Targeted fragments were amplified from extracted genomic DNA by PCR performed on thermal cycler (95°C for 5 min, 38 cycles of [95°C for 30 s, 60°C for 45 s, 72°C for 30 s], 72°C for 10 min, 12°C for ∞.) Genotype of F0 and F1 generation was identified by Sanger sequencing. Seipin Forward primer: CGCATGTGAAGCAATGACTTTGCT; Seipin Reverse primer: CAGACTGGTATTACACACGTGACCA. Gpat3 Forward primer: CTCCAGTAGAGAGTGAGCTTGCG; Gpat3 Reverse primer: GCTACTGTTCTTGACTGGTATCAGGC. The PCR product sizes of Seipin and GPAT3 were 701 and 620 bp, respectively. Genotype after F2 generation was identified by agarose gel electrophoresis of small fragment PCR products. Seipin-small Forward primer: ATCATGGTCAATGACCCACC; Seipin-small Reverse primer: ACAGACACCCAAAGCAAAAG. Gpat3-small Forward primer: TCCTGTCCACCTGGCTGACG; Gpat3-small Reverse primer: CACTCACCTCCAAGGTTTTCA. The PCR product sizes of Seipin and GPAT3 were 140 and 120 bp, respectively.
RNA isolation and quantitative real-time PCR
Total RNA from adipose and other tissues were extracted using Trizol reagent (Invitrogen, USA) and first-strand cDNA was generated by using a RT kit (Promega, USA). Quantitative real-time PCR was performed using primer sets shown in Table S1. Amplifications were performed in 39 cycles using an Applied Biosystem with SYBR green fluorescence (Promega, USA). Each cycle consisted of heating denaturation for 30 s at 94°C, annealing and extension for 60 s at 60°C. All samples were quantitated by using the comparative CT method for relative gene expression, and then normalized to Gapdh.
Western blot analysis
Adipose and testes were homogenized in RIPA buffer and the protein content of tissue lysates was determined using a bicinchoninic acid protein assay kit (Pierce, Rockford, IL). Tissue lysates were subjected to western blotting. Seipin antibody and GPAT3 antibody were prepared by Protein tech Group Inc. UCP1 antibody was purchased from Abcam Inc. GAPDH and β-actin antibody was purchased from Protein tech Group Inc.
Blood analysis
Blood was obtained by retro-orbital bleed. Plasma cholesterol, triglyceride and glucose were determined by using enzymatic methods (Biosino Bio-Technology & Science INC, Beijing, China). Plasma levels of insulin, leptin and adiponectin were measured by ELISA (insulin: Excell, China; leptin: MultiSciences, China; adiponectin: 4A Biotech, China), and the level of non-esterified fatty acids (NEFA) was measured by a colorimetric assay (Wako Chemical, Osaka, Japan).
Histological studies
Liver was cryostat-sectioned at a thickness of 7 μm for lipid deposition analysis by Oil red O staining. Paraffin-embedded subcutaneous fat, epididymal fat, brown fat and liver were sectioned at a thickness of 7 μm and stained with hematoxylin and eosin (HE) or masson for fibrosis analysis. Immunodetections were performed with F4/80 antibody (Santa Cruz Biotechnology, Dallas, Texas) to examine macrophage infiltration in adipose tissue.
Analysis of liver lipids
Approximately 100 mg of liver (wet weight) was weighed and homogenized in 1 mL PBS. Lipids were extracted by chloroform/methanol: 2/1(v/v) then dried under N2 and dissolved in 0.5 mL 3% Triton X-100. The determination of triglyceride was carried out using enzymatic methods as described earlier. Total cholesterol and free cholesterol were determined by using enzymatic methods (Applygen, China). Cholesterol ester was equal to total cholesterol minus free cholesterol.
Glucose and insulin tolerance tests
Mice were fasted for 4 h, followed by intraperitoneal injection of glucose (2 g/kg body weight) or insulin (0.75 IU/kg body weight, Humulin), respectively. Blood samples were collected before (time 0) and at 15, 30, 60 and 120 (90 for ITT) min after injection for measurement of glucose.
Mouse embryonic fibroblast isolation and differentiation
Mouse embryonic fibroblasts (MEFs) were isolated from 13.5-day-old embryos and cultured in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS; Gibco) and penicillin-streptomycin (Pen/Strep). For lipid droplet analysis, MEF cells were incubated with 200 μM oleic acid for 24 hours. Images were obtained by Leica TCS SPS Confocal microscope. LD size was quantified using Image J software. For differentiation, MEF cells were plated and maintained until 2 days after confluence (day 0). Differentiation was induced by culturing cells in adipocyte differentiation medium (dexamethasone 0.25 μM, IBMX 0.5 μM, rosiglitazone 2 μM, insulin 1 μg/ml) for 4 days, followed by regular media in the presence of insulin alone for another 4 days. Differentiated cells were either visualized using light microscopy or stained using Oil red O staining. Oleic acid, dexamethasone, IBMX, rosiglitazone and insulin were all purchased from Sigma Inc.
Statistical analysis
All data are presented as means ± SEM. Statistical comparison between groups was performed using one-way ANOVA. A value of P < 0.05 was considered statistically significant.
Acknowledgements
M.G. generated the bulk of the results, conceived and designed the experiments, and drafted the manuscript. L.L., X.W. and H.Y.M. contributed to research data. M.G., G.L. and H.Y. designed the experiments, provided advice and reagents, and wrote the manuscript. M.G. and H.Y. are the guarantors of this work and, as such, have full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. We thank members of the Gao, Liu and Yang laboratories for helpful discussions and Prof. Peng Li (Tsinghua University, Beijing) for providing reagents.
Conflict of Interest statement
All authors declare that there are no competing financial interests in relation to the work described.
Funding
National Natural Science Foundation of the People’s Republic of China to G.L. (No. 30930037 and 30821001) and M.G. (No. 31771306 and 81700386); National Health and Medical Research Council of Australia (#1057887) to H.Y.
