Human Molecular Genetics, 2011, Vol. 20, No. 15 doi:10.

1093/hmg/ddr205 Advance Access published on May 6, 2011

30223030

Seipin ablation in mice results in severe generalized lipodystrophy

Xin Cui 1,2, Yuhui Wang 1,2, Yin Tang 1,2, Yixiao Liu 1,2, Liping Zhao 4, Jingna Deng 2,3, Guoheng Xu 2,3, Xingui Peng 5, Shenghong Ju 5, George Liu 1,2, and Hongyuan Yang 6,
Institute of Cardiovascular Sciences, 2Key Laboratory of Molecular Cardiovascular Sciences, and 3Department of Physiology, Ministry of Education, Peking University Health Science Center, Beijing 100191, Peoples Republic of China, 4Transgenic Animal Center, National Institute of Biological Science, Beijing 102206, Peoples Republic of China, 5Department of Radiology, Zhong-Da Hospital, Southeast University, Jiangsu Key Laboratory of Molecular Imaging and Functional Imaging, Nanjing, Peoples Republic of China and 6School of Biotechnology and Biomolecular Sciences, the University of New South Wales, Sydney, NSW 2052, Australia
Received March 13, 2011; Revised April 27, 2011; Accepted May 3, 2011
1

Downloaded from http://hmg.oxfordjournals.org/ at Universidade Federal Fluminense on December 19, 2011

Berardinelli-Seip congenital lipodystrophy type 2 (BSCL2) is an autosomal recessive disorder characterized by an almost complete loss of adipose tissue, insulin resistance and fatty liver. Here, we create the rst murine model of BSCL2 by targeted disruption of seipin, the causative gene for BSCL2. Compared with their wild-type littermates, the seipin2/ 2 mice are viable and of normal weight but display signicantly reduced adipose tissue mass, hepatic steatosis, glucose intolerance and hyperinsulinemia. The levels of leptin and adiponectin were both signicantly decreased in seipin2/ 2 mice, so were non-esteried fatty acids upon fasting. Surprisingly, however, hypertriglyceridemia which is common in human BSCL, was not observed in seipin2/2 mice. Our ndings suggest a possible tissue-autonomous role of seipin in liver lipid storage. The availability of the seipin2/2 mice should help elucidate the molecular function of seipin and lead to a better understanding of the many metabolic consequences of human BSCL2.

INTRODUCTION
Congenital generalized lipodystrophy (CGL, also known as Berardinelli-Seip congenital lipodystrophy, or BSCL), is an autosomal recessive disorder characterized by a near total loss of adipose tissue, severe insulin resistance, hypertriglyceridemia and fatty liver (1). Genome-wide linkage analysis has identied two loci for CGL: CGL type 1 (CGL1) is caused by mutations in the 1-acylglycerol-3-phosphate-O-acyl transferase 2 (AGPAT2) gene and CGL2 by mutations in the BSCL2 gene which encodes seipin (2 4). Recently, a homozygous non-sense mutation in caveolin-1 has been discovered to cause CGL (CGL3) (5). Further, mutations in polymerase I and transcript release factor (PTRF, also known as cavin) were found in ve non-consanguineous patients with both generalized lipodystrophy (CGL4) and muscular dystrophy (6). PTRF/cavin is a caveolar-associated protein critical for the formation of caveolae and the stabilization of caveolins. Not

surprisingly, deletion of cavin/PTRF resulted in generalized lipodystrophy (7). Both AGPAT2 and caveolin-1/cavin have clear cellular functions: AGPAT2 catalyzes the formation of phosphatidic acid (PA) and appears to control adipogenesis through modulating the synthesis of phospholipids; caveolin-1/cavin has a dened role in caveola formation. In contrast, little is known about the molecular function of seipin, despite the fact that CGL2/BSCL2 represents the most severe form of human CGLs (1). Recent studies have examined the role of seipin in adipogenesis in cultured cells. One study demonstrated that knocking-down seipin in C3H10T1/2 cells impaired differentiation and caused a reduction in the expression of key genes in triacylglycerol (TAG) synthesis, but did not observe a signicant early decrease in PPARg expression (8). Another similar study showed that knocking-down seipin in C3H10T1/2 cells did not affect bone morphogenetic protein-4-induced preadipocyte commitment. Rather, the

To whom correspondence should be addressed at: Institute of Cardiovascular Research, Peking University Health Science Center, Beijing 100083, Peoples Republic of China. Email: vangeorgeliu@gmail.com (G.L.); School of Biotechnology and Biomolecular Sciences, the University of New South Wales, Sydney, NSW 2052, Australia. Email: h.rob.yang@unsw.edu.au (H.Y.)

# The Author 2011. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Human Molecular Genetics, 2011, Vol. 20, No. 15

3023

differentiation of pre-adipocyte 3T3-L1 cells was greatly impaired by seipin depletion, accompanied by suppression of PPARg expression throughout the differentiation process (9). Interestingly, addition of the PPARg agonist, pioglitazone, was able to rescue the defective differentiation caused by seipin knock-down, suggesting an intimate relationship between seipin and PPARg. The human BSCL2/seipin gene has three transcripts: 1.6, 1.8 and 2.2 kb as revealed by northern blot analyses: the 1.8 kb mRNA is exclusively expressed in brain and testis but the other two transcripts are ubiquitously expressed (10). Seipin is highly expressed in adipose tissue, and is strongly induced during adipocyte differentiation (8,9). However, the upregulation of seipin expression occurs only at late stages of pre-adipocyte differentiation (9). The basal level of seipin expression in the early stages of 3T3L1 differentiation appears to be critical to adipocyte development as depletion of seipin inhibits PPARg activation at a very early stage (9). The function of seipin in mature adipocytes remains to be elucidated, and this function may have little to do with adipogenesis. Another surprising yet exciting recent nding is that both seipin and the yeast seipin homologue, Fld1p, have been found to play a critical role in the cellular dynamics of lipid droplets (11 13). It has been suggested that Fld1p/seipin may regulate the metabolism of phospholipids/triglycerides (11,13,14). A recent study also suggests that seipin may have a structural role in the assembly of lipid droplets from the endoplasmic reticulum (ER) (15). Although progresses have been made in seipin research, the molecular function of seipin remains to be elucidated. Despite the important role of seipin in both lipid droplet formation and adipocyte differentiation, two important aspects of human lipid storage and thereby obesity, no animal model is available to examine the in vivo function of seipin. To this end, we have generated and characterized for the rst time a seipin null mouse model. Our results conrm a critical in vivo role for seipin in adipocyte development and also in hepatic lipid homeostasis.

The SKO (seipin2/2 ) pups can survive past weaning without unexpected early death. Mating between seipin+/2 mice produced offspring with the expected 1 (seipin+/+ ) : 2 (seipin+/2 ) : 1 (seipin2/2 ) Mendelian ratio. Because the epididymal fat of male mice is the appropriate adipose tissue for phenotypic analyses, male mice were chosen for this initial study. In all studies, gender-, strain- and age-matched homozygous knockout animals were compared with wild-type controls. All mice were fed standard chow diet. The body weight of the SKO mice was decreased from birth until 6 weeks old but no signicant difference was observed after 6 weeks (Fig. 2A). There is no difference in food consumption between wild-type and SKO mice but the SKO mice exhibited higher body temperature from time to time (Fig. 2B and C). Plasma lipids Total cholesterol of the SKO mice was signicantly increased only at fed state while plasma glucose was signicantly increased only during fasting (Fig. 2D and E). Hypertriglyceridemia, which is a common feature in human CGLs and has also been found in other lipodystrophic mouse models (17,18), was absent in the SKO mice. In fact, the level of triglyceride (TG) was dramatically decreased upon fasting (Fig. 2D and E). The level of non-esteried fatty acids (NEFA) was not changed at fed state but decreased in the SKO mice during starvation. This may likely be caused by a lack of adipose tissue in the SKO mice (see in what follows). Dramatic loss of adipose tissue mass in the SKO mice Previous studies found that loss-of-function mutations of seipin might be responsible for the complete loss of adipose tissue in BSCL2 patients, suggesting that seipin plays an important role in the development of adipocytes (3,4,8,9). We measured the mass of adipose depots of the SKO mice by magnetic resonance imaging (MRI) via a Bruker Pharmascan 7.0 T/16 cm spectrometer. Visual comparison of MRI images of 3-month-old mice at renal hilum of wild-type and SKO mice revealed a near absence of adipose tissue in the SKO group (Fig. 3A C). This is conrmed when little adipose was observed in the SKO mice upon dissection (Fig. 3D). The livers of the 12-week-old SKO mice were strikingly enlarged and very pale, suggesting massive deposition of fat (Fig. 3D). The weight of individually dissected tissues and fat depots from sacriced animals were measured (Fig. 3E and Table 1). All major fat depots were dramatically reduced in the SKO mice, where almost no gonadal fat was present (Table 1, Fig. 3E). There was also a 60% decrease of brown adipose tissue. The liver of the SKO mice weighed almost twice as much as that of the wild-type. Although the spleen of the SKO mice also showed signicant increase in weight, generalized organomegaly was not observed. The heart and kidney of the SKO mouse did not show any increase in weight (Table 1). Histologic analyses conrm adipose tissue loss Histologic analyses were conducted on the epididymal fat pads of 3-month-old wild-type and SKO mice. Fat pad from wildtype mice contained mature adipocytes, which were uniformly

Downloaded from http://hmg.oxfordjournals.org/ at Universidade Federal Fluminense on December 19, 2011

RESULTS
Generation of the seipin null mice We deleted the exon 3 of mouse seipin by homologous recombination using a strategy based on the Cre/loxP recombination system for generating knockout mice (Fig. 1A). Mice (C57BL/ 6) with loxP sites surrounding seipin exon 3 (E3/) were crossed with oocyte-specic Zp3-Cre transgenic mice to generate a seipin-null allele (seipin+/ ) (Fig. 1A). Seipin2/2 mice were then obtained by mating within the seipin+/2 mice. The genotype was examined by PCR using the genomic DNA, and the seipin2/2 mice (hereafter referred to as SKO mice) were identied with a single PCR product at 1.1 kb (Fig. 1B). The relative expression of seipin in wild-type and seipin2/2 (SKO) mice was examined in various tissues by real-time (RT) quantitative PCR and by regular PCR (Fig. 1C). Finally, we detected normal expression of GNG3, a gene that is located near seipin in a head-on fashion (data not shown) (16). These results conrm that the expression of seipin but not that of GNG3 has been lost in the SKO mice.

3024

Human Molecular Genetics, 2011, Vol. 20, No. 15

Downloaded from http://hmg.oxfordjournals.org/ at Universidade Federal Fluminense on December 19, 2011

Figure 1. Generation and characterization of seipin knock-out/SKO mice. (A) Components of the seipin exon 3 WT allele, the targeting vector, the targeted recombinant allele and the conditional allele in ES cells and the deleted seipin allele in Cre-recombinase transgenic mice. Open boxes represent exons and are numbered as indicated. LoxP sites and Frt sites are shown as solid and open triangles, respectively. Numbered arrows denote PCR primers for genotyping. (B) Genotyping PCR of tail clips of WT (+/+, lane 1), heterozygous (+/2, lane 2) and homozygous mice (2/2, lane 3). Multiplex PCR using forward primer 1 and reverse primer 2 and 3, yielding 300 bp product from wild-type seipin locus and 1100 bp product from deleted seipin locus. (C) Detection of seipin expression in different tissues by qRT-PCR (top). The PCR products after 35 cycles were also resolved by agarose gel electrophoresis (bottom). White adipose tissue (WAT), brown adipose tissue (BAT), brain (Brn), liver (Lvr), kidney (Kdy), skeletal muscle (soleus, SkM).

Figure 2. Characterization of the SKO mice. (A) Growth curves of control and SKO mice maintained on standard chow diet for 10 weeks from birth (n 10, P , 0.05). (B) Food consumption in 3-month-old SKO and control mice on chow diet was unaltered (n 5). (C) Mouse rectal temperature at different time points in 3-month-old SKO and control mice at room temperature (n 3, P , 0.05). Fed (D) and fasted (E) plasma total cholesterol (TC), triglyceride (TG) and glucose levels in SKO and control mice (n 8, P , 0.05). (F) Fed and fasted serum NEFA in SKO and control mice (n 8, P , 0.05).

characterized by the presence of a large, unilocular lipid droplet (Fig. 4A). In contrast, the white epididymal fat of the SKO mice consisted almost entirely of small immature adipocytes, most of which contained brightly eosinophilic cytoplasm and also a relatively small but still distinct unilocular vacuole (lipid droplet) (Fig. 4A). The subcutaneous fat of the SKO mice showed similar changes (Fig. 4B). The loss of adipose tissue in the SKO mice was further reected by

the dramatically decreased levels of plasma adiponectin and leptin, two important adipokines (Fig. 4C and D) (19,20). Abnormal lipid accumulation in the liver of SKO mice To understand the cause of the enlarged liver and its pale appearance in the SKO mice, we rst examined liver morphology and histology (Fig. 5A and B). Fixed liver sections

Human Molecular Genetics, 2011, Vol. 20, No. 15

3025

Downloaded from http://hmg.oxfordjournals.org/ at Universidade Federal Fluminense on December 19, 2011

Figure 3. The SKO mice are lipodystrophic. (A) Abdominal MRI of 3-month-old under chow diet SKO mice with decreased subcutaneous and intra-abdominal fat at right renal hilum compared with WT littermates with normal fat distribution. Fat is shown white in these MRI images. (B) MRI measures of adipose fat area at renal hilum (n 4, P , 0.05). (C) Decreasing amount of adipose tissue volume of 1 cm thick transaxial slice (average of 5 2-mm-thick slices) from inferior pole of kidney (n 4, P , 0.05). (D) Exposed ventral view of WT (left) and SKO (right) mice at 3 months of age showing reduced epididymal adipose tissue (black arrow) and enlarged liver (red arrow) in the SKO mouse. (E) Epididymal white fat/testis from a 3-month-old WT male mouse (up) and an age-matched SKO male mouse (bottom) under a chow diet.

Table 1. Phenotypic comparison of wild-type and SKO mice WT Body weight (g) Subcutaneous fat (mg) Inguinal fat (mg) Gonadal fat (mg) Retroperitoneal fat (mg) Mesenteric fat (mg) Brown adipose (mg) Heart (mg) Liver (mg) Spleen (mg) Kidney (mg) 21.9 + 0.9 208.5 + 15.8 36.6 + 2.9 298.3 + 62.6 87.8 + 12.3 193.5 + 9.6 129.6 + 21.4 123.1 + 11.2 1065.4 + 67.0 82.7 + 9.1 168.4 + 13.6 KO 21.7 + 0.7 45.8 + 2.7 13.6 + 1.7 14.7 + 2.3 72.9 + 25.3 59.2 + 5.9 126.2 + 3.1 1841.7 + 146.4 127.0 + 10.1 166.6 + 5.3 P-value 0.9 0.002 0.001 0.004 0.005 0.04 0.8 0.003 0.02 0.9

Mice (12 weeks of age) were fed a standard rodent chow and sacriced without fasting. All values are means + s.e.m. Statistical analysis was done with two-tailed unpaired Students t-test. n 4, P , 0.05, P , 0.01, P , 0.001.

were stained with H E and examined at different magnications (Fig. 5B). It appears that the SKO mouse has mixed hepatic steatosis. The steatosis is mainly macrovesicular but microvesicular steatosis can also be seen. There is no obvious inammatory cell inltration or ballooning degeneration, suggesting no steatohepatitis at this stage. Cryosections of wild-type and SKO mouse livers were stained with oil red O, and the liver of the SKO mice contained much more lipid droplets than that of wild-type mice (Fig. 5C). Interestingly, little lipid accumulation was detected in the skeletal muscle (SkM) of the SKO mice (Fig. 5C). Finally, we measured the

amount of TG in the liver and SkM of the SKO mice and wildtype littermate control. The amount of liver TG of the SKO mice was 200% higher than that of wild-type littermates (Fig. 5D). In contrast, there is little difference in SkM TG between the wild-type and SKO mice (Fig. 5D). There is no accumulation of cholesterol in the liver and SkM of the seipin mice (Fig. 5E). The increased accumulation of triglycerides in the liver could result from drastically reduced lipid storage in the adipose tissue. But a tissue autonomous role of seipin in liver lipid metabolism cannot be ruled out. The expression of some lipogenic genes such as FAS, PPARg and SCD1 was signicantly increased in the SKO liver (Fig. 5F). However, the expression of SREBP-1c, the DGATs and ACC1 did not change. These data suggest that enhanced lipogenesis alone may not account for hepatic steatosis in the SKO mice. Interestingly, the expression of microsomal triglyceride transfer protein (MTP) was signicantly reduced in the SKO mice, suggesting a possible defect in the lipidation of apolipoprotein B (ApoB). Given the known role for seipin in lipid droplet formation, a possible defect in the movement of neutral lipids from the cytoplasm to the lumen of the ER in the SKO liver would not be surprising.

SKO mice is insulin resistant We evaluated insulin sensitivity and glucose homeostasis in the SKO and wild-type littermates at 3 months of age. Glucose tolerance test (GTT) indicated that the SKO mice showed delayed glucose clearance and were therefore diabetic (Fig. 6A). Insulin tolerance test (ITT) showed that the SKO

3026

Human Molecular Genetics, 2011, Vol. 20, No. 15

Downloaded from http://hmg.oxfordjournals.org/ at Universidade Federal Fluminense on December 19, 2011

Figure 4. Impaired adipocyte development of the SKO mice. (A, B) Histological section (H&E) of epididymal (A) and subcutaneous (B) WAT from 3-monthold SKO and control mice under chow diet. Bars, 400 mm (top) and 40 mm (bottom). Serum adiponectin (C) and leptin (D) levels of WT and SKO mice (SKO n 7, WT n 10, P , 0.05, P , 0.001).

mice had impaired insulin sensitivity compared with the wildtype littermates (Fig. 6B). They also had signicantly elevated fed plasma insulin levels (Fig. 6C). To investigate the molecular basis of the insulin resistance, the mRNA levels of insulin receptor substrate (IRS1), IRS2, Glut4 and AKT2 were examined in the liver, SkM and white adipose tissue (WAT) (Fig. 6D F). The expression of all four genes was marked decreased in WAT. While the expression of IRS1, IRS2 and AKT2 but not GLUT4 was signicantly reduced in the liver, only IRS2 expression was signicantly reduced in the SkM. The expression of gluconeogenic genes such as PEPCK and G6P1 was also signicantly reduced in the liver. These early results suggest that there are impaired insulin signaling in the liver, WAT and possibly the muscle of the SKO mice.

DISCUSSION
Recent studies have revealed important roles of seipin/BSCL2 in lipid storage at both the cellular level (lipid droplets) and the whole-body level (adipose tissue development) (1,11,12). To better understand the in vivo function of seipin, we generated the rst murine model for seipin research: the seipin2/2 mouse or the SKO mice. We found that the SKO mice recapitulated many aspects of human BSLC2, such as a dramatic loss of adipose tissue, insulin resistance and hepatic steatosis. The availability of the SKO mice should lead to additional insights into the cause of metabolic disorders, e.g. insulin

resistance, and help elucidate the molecular function of the mysterious yet important protein: seipin. Both CGL1/BSCL1 (AGPAT2) and CGL2/BSCL2 (seipin) patients suffer severe fat loss. When compared with other CGL patients, CGL2/BSCL2 patients have even more severe lipodystrophy because there is a loss of both mechanical adipose tissue (found in retro-orbital, palm, sole and other areas) and metabolically active adipose tissue (found in subcutaneous, intra-abdominal, intrathoracic and other areas) (1). Interestingly, the SKO mice which clearly suffer from severe lipodystrophy, still retain a signicant portion of both WAT and BAT (Table 1), whereas the AGPAT2-decient mice has an almost complete loss of white and brown adipose tissues (21). Although not 100% identical, the genetic backgrounds of the mice used in these two studies are very similar. These data suggest that AGPAT2, but not seipin, might play a more important role in adipocyte development in mice. Hypertriglyceridemia and hepatic steatosis are common features of human CGL. Not surprisingly, both AGPAT2decient and SKO mice develop severe hepatic steatosis. However, chronic steatohepatitis was observed in the AGPAT2 mice but not in the SKO mice (21). In the case of hypertriglyceridemia, although similar rates of CGL1 and CGL2 patients (62.5% for CGL2/seipin patients and 71% for CGL1/AGPAT2) develop hypertriglyceridemia, there is a major difference between the two mouse models (22). In the SKO male mice, the plasma TG level is not elevated. In fact, the level of plasma TG dropped by over 80% upon

Human Molecular Genetics, 2011, Vol. 20, No. 15

3027

Downloaded from http://hmg.oxfordjournals.org/ at Universidade Federal Fluminense on December 19, 2011

Figure 5. Hepatic steatosis in the SKO mice. (A) Photograph of the liver of WT (left) and SKO (right) mice. (B) Histological analysis (hematoxylin/eosin staining) of liver sections with different magnication from SKO (bottom right) mice lled with vacuoles (black arrow), compared with WT littermates (bottom left). Bars, 400 mm (top) and 40 mm (bottom). (C) Images of liver and skeletal muscle (gastrocnemius) sections stained with oil red O, the red color droplets represent the lipid droplets in the liver of SKO mouse (scale bar is 50 mm). Liver and skeletal muscle TAG (D) and TC (E) contents from 3-month-old SKO mice and WT littermates (n 4, P , 0.01). (F) Expression of genes concerned with lipid synthesis and transportation in the livers of SKO mice and control littermates (n 4, P , 0.05, P , 0.01). ACC1, acetyl CoA carboxylase 1; DGAT, diacylglycerol acyl-transferase; FAS, fatty acid synthase; MTP, microsomal transfer protein; PGC-1alpha, PPAR gamma coactivator 1alpha; PPARg, peroxisome proliferator-activated receptor gamma; SCD-1, stearoyl-coenzyme A desaturase 1; SREBP1c, sterol response element binding protein 1c.

fasting. The AGPAT2-decient male mice, on the other hand, developed hypertriglyceridemia when fed with standard chow without fasting. This may be explained in part by the severity of hepatic steatosis: the TG of a 3-month-old SKO liver increased by 200%, whereas the TG of a similar aged AGPAT2-decient liver by 500%. There was also highly elevated hepatic lipogenesis in the AGPAT2-decient mice, whereas only a moderate increase in the expression of FAS, SCD-1 and PPARg was detected in the SKO mice. In fact, the expression level of ACC1, SREBP-1c and the DGATs did not change in the SKO mice (Fig. 5F). Although other possibilities cannot be ruled out, these differences imply that seipin and AGPAT2 may function through different mechanisms to regulate hepatic lipid homeostasis. The decrease in plasma TG of the SKO mice upon fasting may be caused by reduced lipidation of nascent ApoB, as the expression of MTP was signicantly reduced (Fig. 5). MTP plays an essential role in transferring the bulk of triglycerides into the lumen of the ER for very low density lipoprotein (VLDL) assembly and is required for the secretion of ApoB-100 from the liver (23). Therefore, a reduction in MTP expression may be partially responsible for the decrease of plasma TG upon fasting. There is also an intimate

relationship between cytoplamic lipid droplets and the maturation of VLDL, and a recent study demonstrated that the ER- and lipid-droplet-associated protein CideB mediates VLDL lipidation and maturation (24). Given its known role in lipid droplet formation and its localization to the ER and lipid droplets, seipin could take part in the movement of the neutral lipids across the ER and thereby directly or indirectly regulating the lipidation of nascent ApoB. Similar to the AGPAT2 decient mice but different from all other mouse models of lipodystrophy associated with hepatic steatosis, the level of NEFA was normal or low (upon fasting) in the SKO mice (21,25,26). The lack of adipose tissue is probably a major contributing factor. Likewise, VLDL may not be efciently secreted from the liver in the absence of seipin, resulting in decreased plasma NEFA upon fasting. This again suggests that seipin may have a tissue autonomous function in lipid storage and secretion in the liver. Importantly, these results also suggest that high serum NEFA levels are not required for the development of hepatic or peripheral insulin resistance, in agreement with an early study on AGPAT2 (21). In summary, we have created the rst mouse model in seipin and BSCL2 research and conducted initial

3028

Human Molecular Genetics, 2011, Vol. 20, No. 15

Downloaded from http://hmg.oxfordjournals.org/ at Universidade Federal Fluminense on December 19, 2011

Figure 6. SKO mice are insulin resistant. Glucose (A) and insulin (B) tolerance tests performed on 3-month-old male mice on regular diet (n 4, P , 0.05, P , 0.01, P , 0.001). (C) Increased serum insulin level in SKO mice compared with control littermates (SKO n 7, WT n 10, P , 0.01). Expression of selected genes in the liver (D), skeletal muscle (E) and WAT (F) (n 4, P , 0.05, P , 0.01, P , 0.001).

characterization. Although much is to be done, our ndings clearly demonstrated a crucial role of seipin in adipocyte development in mice. Future exploitation of this unique in vivo model should lead to exciting new developments in seipin, lipodystrophy and metabolic research.

MATERIALS AND METHODS

Animals Two loxP sites and a neo cassette anked by two FRT reconbination sites in seipin intron 2 and 3, respectively, were introduced into the seipin locus by homologous recombination in ES cells (Fig. 1A). The targeting construct was electroporated into 129 ES cells and G418 resistant clones were selected under standard conditions. Appropriately targeted ES cells were identied by PCR and Southern blotting and were then injected into C57BL/6 blastocysts. The blastocysts were implanted into pseudopregnant females. Male chimaeras positive for germ-line transmission were used to establish recombinant mouse lines. The neo gene was excised through breeding with Flp recombinase-expressing mice (JAX Strain #003946). Germ-line deletion of seipin exon 3 was induced by crossing mice with the loxP seipin allele to transgenic mice expressing Cre recombinase driven by a germ (oocyte) cell specic promoter (Zp3-cre, Jax #003394) (27). Progeny were screened by PCR for loss of the seipin exon 3. After the generation of the seipin-null allele (seipin+/2 ), the strain was further crossed with C57BL/6 for three generations and inbreed with heterozygotes to obtain homozygotes. The genotyping was examined by PCR using the genomic DNA obtained from the clipped tail. Primers used for the seipin gene were Seipin-1 (5 -TCTATGGCTCCTTCTACT ACTC-3 ), Seipin-3 (5 -CGAATGATATGACGACGACT-3 ), and for the wild-type allele were Seipin-1 (5 -TCTATG

GCTCCTTCTACTACTC-3 ), Seipin-2 (5 -ACTAAAAGG CAAAGGAGG-3 ). Using a mixture of these primers, PCR was performed with 35 cycles of a reaction consisting of 30 s of denaturation at 948C, 30 s of annealing at 568C and 1 min of elongation at 728C. PCR products were 1100 and 300 bp, specic for null and wild alleles, respectively. All experiments involving mice were approved by the Institutional Animal Care Research Advisory Committee of the National Institute of Biological Science and Animal Care Committee of Peking University Health Science Center. Animals were housed and allowed free access to tap water and standard laboratory chow. The Principles of Laboratory Animal Care (NIH publication no. 85 23, revised 1996) were followed. All mice were maintained on a 12 h light/dark cycle and were fed ad libitum with regular mouse chow (4% fat by weight). Mice were fasted for 4 h and blood samples were taken from the retro-orbital plexus. The liver, heart, kidney, brain and SkM were then harvested, snap-frozen in liquid nitrogen and stored at 2808C for real-time PCR. RNA isolation and quantitative real-time PCR Total RNA was extracted using Trizol reagent (Invitrogen, USA) and rst-strand cDNA was generated by using a RT kit (Invitrogen, USA). Quantitative real-time PCR was performed using primer sets shown in Table 2. Amplications were performed in 35 cycles using an opticon continuous uorescence detection system (MJ Research) with SYBR green uorescence (Molecular Probes, Eugene, USA). Each cycle consisted of heating denaturation for 30 s at 948C, annealing for 30 s at 568C and extension for 30 s at 728C. All samples were quantitated by using the comparative CT method for relative quantitation of gene expression, normalized to GAPDH (28).

Human Molecular Genetics, 2011, Vol. 20, No. 15

Table 2. List of primers used and their sequences SREBP-1c ACC1 DGAT1 DGAT2 FAS SCD1 PPARg PGC-1a Seipin forward reverse forward reverse forward reverse forward reverse forward reverse forward reverse forward reverse forward reverse forward reverse TGGAGACATCGCAAACAAG GGTAGACAACAGCCGCATC CCAGACCCTTTCTTCAGC TTGTCGTAGTGGCCGTTC ATCTGAGGTGCCATCGTC ATGCCATACTTGATAAGGTTCT TCAACCGAGACACCATAGAC CCTCAAAGATCACCTGCTT GGGTCTATGCCACGATTC GTGTCCCATGTTGGATTTG TGACCTGAAAGCCGAGAA ATGTGCCAGCGGTACTCA GACCACTCGCATTCCTTT CCACAGACTCGGCACTCA CACAAACGATGACCCTC GCATGTTGCGACTGC GGCTCCTTCTACTACTCCTACA CCGATCACGTCCACTCTT MTP IRS1 IRS2 PEPCK1 AKT2 GLUT4 G6P1 Gng3 forward reverse forward reverse forward reverse forward reverse forward reverse forward reverse forward reverse forward reverse

3029

GGAAAGCAGAGCGGAGAC AGAGCAAGGGTCAGGCAC GGATCGTCAATAGCGTAA GCTTGGCACAATGTAGAA GGGGCGAACTCTATGGGTA GCAGGCGTGGTTAGGGAAT AGTCATCATCACCCAAGAGC CCACCACATAGGGCGAGT CAGATGGTCGCCAACAGT TGCCGAGGAGTTTGAGATA ACGGATAGGGAGCAGAAA AAGGGTGAGTGAGGCATT AATCTCCTCTGGGTGGCA GCTGTAGTAGTCGGTGTCC ACGCAAGATGGTGGAACAGC GAGTAGAAGGTGCTTGGAGT

Downloaded from http://hmg.oxfordjournals.org/ at Universidade Federal Fluminense on December 19, 2011

Blood analysis Blood was obtained by retro-orbital bleed. Plasma TC, TG and glucose were determined by using enzymatic methods (Sigma kits, MO, USA). Serum insulin, leptin and adiponectin were measured by ELISA (Linco Research, St Charles, MO), and free fatty acids were measured by using a free fatty acid kit (Wako). Glucose and insulin tolerance tests For glucose and insulin tolerance tests, mice fasted for 4 h were given i.p. glucose (2 g/kg body weight; Abbott) or insulin (Humulin, 0.75 mIU/g), respectively, and blood samples were collected before (time 0) and at 15, 30, and 60 and 120 (90 for ITT) min after injection for measurement of glucose (Sigma kits, MO, USA). Histological studies Tissues were xed in 4% neutral formalin and parafn-embedded, and sections were stained with hematoxylin/eosin. The segments of liver and SkM were cryostat sectioned at a thickness of 7 mm onto poly-l-lysine slides for lipid deposition analysis by oil red O staining. Adipose tissue was prepared and subjected to hematoxylin and eosin staining. The wild-type littermates were used as controls. Analysis of liver lipids Approximately 100 mg of liver (wet weight) was weighed and homogenized in 1 ml PBS. Lipids were extracted as described by Folch et al. (29) and dissolved in 100 ml 3% Triton X-100. The determination of TG was carried out using enzymatic methods as described earlier. Magnetic resonance imaging For MRI acquisition, anesthesia was induced by inhalation of a mixture of oxygen and 5% isourane and maintained by a mixture of oxygen containing 1 2% isourane. Mice were

positioned and immobilized prone inside the tomograph with either the thoracic or the abdominal region in the center of the eld of view (FOV). All MRI experiments were carried out using a 7 T small animal magnetic resonance tomograph with Bruker Pharmascan 7.0 T/16 cm spectrometer equipped with a mini imaging gradient coil system (gradient strength, 375 mT/m) and a 1H transmit receive quatrature coil with 31 mm inner diameter. T1-weighted (T1W) images were acquired with a respiratory-gated spin echo sequence, FOV 3.5 3.5 cm, matrix size 256 256, slice thickness 2.0 mm, repetition time 500 ms and echo time 15 ms and a number of repetitions (NEX) of 4. The T1W images were used to study the distribution of fat stores. Core body temperature Core body temperature was measured using a rectal probe (Thermalet TH-5) inserted 1.0 cm deep at ambient room temperature. Food and water were provided ad libitum. Statistical analysis All data are presented as means + SEM. Statistical comparison between the two groups was performed using Students t-test or one-way ANOVA. A value of P , 0.05 was considered statistically signicant.

ACKNOWLEDGEMENTS
We thank members of the Yang and Liu laboratory for helpful discussions. Conict of Interest statement. None declared.

FUNDING
This work is supported in part by National Natural Science Foundation of the Peoples Republic of China (no. 30821001, no. 30930037), Major National Basic Research Program of the Peoples Republic of China (2011CB503900)

3030

Human Molecular Genetics, 2011, Vol. 20, No. 15

to G.L., National Natural Science Foundation of the Peoples Republic of China (no. 30971102) to Y.W. and a research grant from the National Health and Medical Research Council of Australia (#568725) to H.Y.

REFERENCES
1. Agarwal, A.K. and Garg, A. (2006) Genetic basis of lipodystrophies and management of metabolic complications. Annu. Rev. Med., 57, 297 311. 2. Agarwal, A.K., Arioglu, E., De Almeida, S., Akkoc, N., Taylor, S.I., Bowcock, A.M., Barnes, R.I. and Garg, A. (2002) AGPAT2 is mutated in congenital generalized lipodystrophy linked to chromosome 9q34. Nat. Genet., 31, 2123. 3. Magre, J., Delepine, M., Khallouf, E., Gedde-Dahl, T. Jr., Van Maldergem, L., Sobel, E., Papp, J., Meier, M., Megarbane, A., Bachy, A. et al. (2001) Identication of the gene altered in Berardinelli-Seip congenital lipodystrophy on chromosome 11q13. Nat. Genet., 28, 365370. 4. Agarwal, A.K. and Garg, A. (2004) Seipin: a mysterious protein. Trends Mol. Med., 10, 440 444. 5. Kim, C.A., Delepine, M., Boutet, E., El Mourabit, H., Le Lay, S., Meier, M., Nemani, M., Bridel, E., Leite, C.C., Bertola, D.R. et al. (2008) Association of a homozygous nonsense caveolin-1 mutation with Berardinelli-Seip congenital lipodystrophy. J. Clin. Endocrinol. Metab., 93, 1129 1134. 6. Hayashi, Y.K., Matsuda, C., Ogawa, M., Goto, K., Tominaga, K., Mitsuhashi, S., Park, Y.E., Nonaka, I., Hino-Fukuyo, N., Haginoya, K. et al. (2009) Human PTRF mutations cause secondary deciency of caveolins resulting in muscular dystrophy with generalized lipodystrophy. J. Clin. Invest., 119, 26232633. 7. Liu, L., Brown, D., McKee, M., Lebrasseur, N.K., Yang, D., Albrecht, K.H., Ravid, K. and Pilch, P.F. (2008) Deletion of Cavin/PTRF causes global loss of caveolae, dyslipidemia, and glucose intolerance. Cell Metab., 8, 310 317. 8. Payne, V.A., Grimsey, N., Tuthill, A., Virtue, S., Gray, S.L., Dalla Nora, E., Semple, R.K., ORahilly, S. and Rochford, J.J. (2008) The human lipodystrophy gene BSCL2/seipin may be essential for normal adipocyte differentiation. Diabetes, 57, 20552060. 9. Chen, W., Yechoor, V.K., Chang, B.H., Li, M.V., March, K.L. and Chan, L. (2009) The human lipodystrophy gene product BSCL2/Seipin plays a key role in adipocyte differentiation. Endocrinology. 150, 45524561. 10. Windpassinger, C., Auer-Grumbach, M., Irobi, J., Patel, H., Petek, E., Horl, G., Malli, R., Reed, J.A., Dierick, I., Verpoorten, N. et al. (2004) Heterozygous missense mutations in BSCL2 are associated with distal hereditary motor neuropathy and Silver syndrome. Nat. Genet., 36, 271 276. 11. Fei, W., Shui, G., Gaeta, B., Du, X., Kuerschner, L., Li, P., Brown, A.J., Wenk, M.R., Parton, R.G. and Yang, H. (2008) Fld1p, a functional homologue of human seipin, regulates the size of lipid droplets in yeast. J. Cell Biol., 180, 473482. 12. Szymanski, K.M., Binns, D., Bartz, R., Grishin, N.V., Li, W.P., Agarwal, A.K., Garg, A., Anderson, R.G. and Goodman, J.M. (2007) The lipodystrophy protein seipin is found at endoplasmic reticulum lipid droplet junctions and is important for droplet morphology. Proc. Natl Acad. Sci. USA, 104, 20890 20895.

13. Boutet, E., El Mourabit, H., Prot, M., Nemani, M., Khallouf, E., Colard, O., Maurice, M., Durand-Schneider, A.M., Chretien, Y., Gres, S. et al. (2009) Seipin deciency alters fatty acid Delta9 desaturation and lipid droplet formation in Berardinelli-Seip congenital lipodystrophy. Biochimie, 91, 796803. 14. Fei, W., Du, X. and Yang, H. (2011) Seipin, adipogenesis and lipid droplets. Trends Endocrinol. Metab. in press 15. Binns, D., Lee, S., Hilton, C.L., Jiang, Q.X. and Goodman, J.M. (2010) Seipin is a discrete homooligomer. Biochemistry, 49, 1074710755. 16. Downes, G.B., Copeland, N.G., Jenkins, N.A. and Gautam, N. (1998) Structure and mapping of the G protein gamma3 subunit gene and a divergently transcribed novel gene, gng3lg. Genomics, 53, 220230. 17. Savage, D.B. (2009) Mouse models of inherited lipodystrophy. Dis. Model. Mech., 2, 554562. 18. Reue, K. and Phan, J. (2006) Metabolic consequences of lipodystrophy in mouse models. Curr. Opin. Clin. Nutr. Metab. Care, 9, 436441. 19. Berg, A.H., Combs, T.P. and Scherer, P.E. (2002) ACRP30/adiponectin: an adipokine regulating glucose and lipid metabolism. Trends Endocrinol. Metab., 13, 84 89. 20. Kershaw, E.E. and Flier, J.S. (2004) Adipose tissue as an endocrine organ. J. Clin. Endocrinol. Metab., 89, 2548 2556. 21. Cortes, V.A., Curtis, D.E., Sukumaran, S., Shao, X., Parameswara, V., Rashid, S., Smith, A.R., Ren, J., Esser, V., Hammer, R.E. et al. (2009) Molecular mechanisms of hepatic steatosis and insulin resistance in the AGPAT2-decient mouse model of congenital generalized lipodystrophy. Cell Metab., 9, 165 176. 22. Agarwal, A.K., Simha, V., Oral, E.A., Moran, S.A., Gorden, P., ORahilly, S., Zaidi, Z., Gurakan, F., Arslanian, S.A., Klar, A. et al. (2003) Phenotypic and genetic heterogeneity in congenital generalized lipodystrophy. J. Clin. Endocrinol. Metab., 88, 48404847. 23. Raabe, M., Veniant, M.M., Sullivan, M.A., Zlot, C.H., Bjorkegren, J., Nielsen, L.B., Wong, J.S., Hamilton, R.L. and Young, S.G. (1999) Analysis of the role of microsomal triglyceride transfer protein in the liver of tissue-specic knockout mice. J. Clin. Invest., 103, 1287 1298. 24. Ye, J., Li, J.Z., Liu, Y., Li, X., Yang, T., Ma, X., Li, Q., Yao, Z. and Li, P. (2009) Cideb, an ER- and lipid droplet-associated protein, mediates VLDL lipidation and maturation by interacting with apolipoprotein B. Cell Metab., 9, 177 190. 25. Moitra, J., Mason, M.M., Olive, M., Krylov, D., Gavrilova, O., Marcus-Samuels, B., Feigenbaum, L., Lee, E., Aoyama, T., Eckhaus, M. et al. (1998) Life without white fat: a transgenic mouse. Genes Dev., 12, 3168 3181. 26. Shimomura, I., Hammer, R.E., Richardson, J.A., Ikemoto, S., Bashmakov, Y., Goldstein, J.L. and Brown, M.S. (1998) Insulin resistance and diabetes mellitus in transgenic mice expressing nuclear SREBP-1c in adipose tissue: model for congenital generalized lipodystrophy. Genes Dev., 12, 3182 3194. 27. Lewandoski, M., Wassarman, K.M. and Martin, G.R. (1997) Zp3-cre, a transgenic mouse line for the activation or inactivation of loxP-anked target genes specically in the female germ line. Curr. Biol., 7, 148151. 28. Fink, L., Seeger, W., Ermert, L., Hanze, J., Stahl, U., Grimminger, F., Kummer, W. and Bohle, R.M. (1998) Real-time quantitative RT-PCR after laser-assisted cell picking. Nat. Med., 4, 1329 1333. 29. Folch, J., Lees, M. and Sloane Stanley, G.H. (1957) A simple method for the isolation and purication of total lipides from animal tissues. J. Biol. Chem., 226, 497 509.

Downloaded from http://hmg.oxfordjournals.org/ at Universidade Federal Fluminense on December 19, 2011