Overexpression of SREBP-1a in Mouse Adipose Tissue Produces Adipocyte Hypertrophy, Increased Fatty Acid Secretion, and Fatty Liver

Transcription

1 JBC Papers in Press. Published on July 10, 2003 as Manuscript M Overexpression of SREBP-1a in Mouse Adipose Tissue Produces Adipocyte Hypertrophy, Increased Fatty Acid Secretion, and Fatty Liver Jay D. Horton 1,2, Iichiro Shimomura 1*, Shinji Ikemoto 1*#, Yuriy Bashmakov 1, and Robert E. Hammer 3,4 1 Departments of Molecular Genetics, 2 Internal Medicine and 3 Biochemistry, 4 Howard Hughes Medical Institute, The University of Texas Southwestern Medical Center at Dallas, Dallas, Texas USA Corresponding authors: Robert E. Hammer Department of Biochemistry and Howard Hughes Medical Institute, The University of Texas Southwestern Medical Center at Dallas, Dallas, Texas USA Tel: , fax: Robert.hammer@utsouthwestern.edu Jay D. Horton Departments of Molecular Genetics and Internal Medicine The University of Texas Southwestern Medical Center at Dallas, Dallas, Texas USA Tel: , fax: Jay.horton@utsouthwestern.edu Copyright 2003 by The American Society for Biochemistry and Molecular Biology, Inc.

2 2 *Current address: Department of Medicine and Pathophysiology Osaka University Graduate School of Medicine 2-2 Yamadaoka Suita Osaka JAPAN #Current address: Shinji Ikemoto Division of Clinical Nutrition The National Institute of Health and Nutrition , Toyama-cho Shinjuku-ku Tokyo JAPAN Running Title: Fat-specific transgenic nsrebp-1a mice [Key words: SREBP, transgenic, adipogenesis, fatty acids, hepatic steatosis]

3 3 ABSTRACT Sterol regulatory element-binding proteins (SREBPs) are a family of membranebound transcription factors that regulate cholesterol and fatty acid homeostasis. In mammals, three SREBP isoforms have been identified that are designated SREBP- 1a, SREBP-1c, and SREBP-2. SREBP-1a and SREBP-1c are derived from the same gene by virtue of alternatively spliced first exons. SREBP-1a has a longer transcription-activating domain and is a more potent transcriptional activator than SREBP-1c in cultured cells and liver. Here, we describe the physiologic consequences of overexpressing the nuclear form of SREBP-1a (nsrebp-1a) in adipocytes of mice using the adipocyte-specific ap2 promoter (ap2-nsrebp-1a). The transgenic ap2-nsrebp-1a mice develop markedly enlarged white and brown adipocytes that were fully differentiated. Adipocytes isolated from ap2-nsrebp-1a mice have significantly increased rates of fatty acid synthesis and enhanced fatty acid secretion. The increased production and release of fatty acids from adipocytes led, in turn, to a fatty liver. Overexpression of the alternative SREBP-1 isoform, nsrebp-1c, in adipose tissue inhibited adipocyte differentiation. As a result, the transgenic nsrebp-1c mice developed a syndrome that resembled human lipodystrophy, which included a loss of peripheral white adipose tissue, diabetes, and fatty livers (1). In striking contrast, nsrebp-1a overexpression in fat results in the hypertrophy of fully differentiated adipocytes, no diabetes, and mild hepatic steatosis. These results suggest that nsrebp-1a and nsrebp-1c have distinct roles in adipocyte fat metabolism in vivo.

4 4 INTRODUCTION Obesity is an ever-increasing public health concern that it is now estimated to afflict least one-third of the U.S. population (2). Understanding how progenitor cells differentiate into adipocytes may be critical for the development of anti-obesity therapies. Sterol regulatory element-binding proteins (SREBPs) 1 are a family of membranebound transcription factors that principally regulate lipid synthesis (3), but they have also been implicated in adipocyte differentiation (1,4,5). SREBPs belong to the larger basichelix-loop-helix-leucine zipper (bhlh-zip) family of transcription factors, but are unique because they are synthesized as ~1150 amino acid precursors bound to the endoplasmic reticulum and nuclear envelope (3). To be active, SREBP undergo a sequential two-step cleavage process to release the transcriptionally active NH 2 -terminal portion of the protein, designated the nuclear form. Once the nuclear form is released from the membrane, it can enter the nucleus and bind to the promoters of target genes to activate transcription. Three SREBP isoforms, designated SREBP-1a, SREBP-1c, and SREBP-2, have been identified in humans and animals (3). SREBP-1a and SREBP-1c are encoded by the same gene that undergoes alternative splicing in which the two transcripts are driven by different promoters (6,7). SREBP-1a and SREBP-1c differ only in the first exon, which encodes a portion of an acidic transcriptional activating domain (6,7). SREBP-1a has a longer activation domain, and is much more potent than SREBP-1c in transcriptionally activating known target genes in cultured cells (8) and liver (9). The relative amounts of the SREBP-1a and SREBP-1c transcripts differ among cells and organs. SREBP-1a is

5 5 the predominant transcript in most cultured cells. In most animal tissues, including liver and adipose tissue, the SREBP-1c transcript is predominant (10-12). The third SREBP isoform, designated SREBP-2, is produced by a different gene, and it contains a long acidic-activating domain resembling that of SREBP-1a (13). In liver, all SREBP isoforms are capable of activating the same families of genes, but they do so with varying relative efficiencies. SREBP-2 preferentially activates multiple genes in the cholesterol biosynthetic pathway, whereas SREBP-1a and SREBP- 1c preferentially activate genes involved in the synthesis of fatty acids and triglycerides (14). The function and transcriptional activation properties of SREBPs in tissues other than liver have been only partially characterized. Of particular interest is the in vivo function of SREBPs in adipocytes owing to the potential role SREBP-1 isoforms may have in adipogenesis (4). Overexpression of the transcriptionally active nuclear SREBP-1c (nsrebp-1c) in adipocytes of transgenic mice using the fat-specific ap2 promoter inhibited adipocyte differentiation and produced a syndrome in mice that shares the general features of congenital generalized lipodystrophy in humans (1). The amount of white adipose tissue (WAT) in the lipodystrophic mice is markedly reduced, and many of the residual cells have the histologic appearance of immature adipocytes. These changes are associated with a marked reduction in many mrnas encoding genes involved in adipogenesis such as C/EBP, PPAR, and adipsin. The majority of adipocytes in the brown adipose tissue (BAT) are histologically similar to immature white adipocytes. The mrnas for uncoupling protein-1 (UCP-1) and the additional adipogenic genes listed above were also

6 6 markedly reduced. The transgenic mice also exhibit profound insulin resistance, hyperglycemia, and enlarged fatty livers. To investigate the physiologic function of the SREBP-1a isoform in adipocytes, we have produced transgenic mice that overexpress the transcriptionally active nuclear SREBP-1a (nsrebp-1a) in WAT and BAT using the ap2 enhancer/promoter (ap2- nsrebp-1a mice). The phenotype of ap2-nsrebp-1a mice differs completely from that observed in ap2-nsrebp-1c mice (1). Overexpression of nsrebp-1a in adipose tissue does not inhibit adipocyte differentiation but rather activates the entire pathway of cholesterol and fatty acid biosynthetic genes, leading to increased lipid accumulation in the adipocytes. The white adipocytes from the transgenic mice are massively enlarged and contain excess triglyceride. The brown adipocytes contain large unilocular lipid droplets that produce a histologic appearance that resemble mature white adipocytes. Isolated adipocytes from ap2-nsrebp-1a mice manifest significantly increased rates of fatty acid synthesis, which are associated with increased fatty acid secretion into the medium. In vivo, this leads to a higher influx of FFA from adipose tissues into liver, resulting in a fatty liver.

7 7 EXPERIMENTAL PROCEDURES General materials and methods-dna manipulations were performed using standard molecular techniques (15). Oleic acid [9,10-3 H(N)] and [stearoyl 1-14 C]-stearoyl-CoA were purchased from American Radiolabeled Chemicals, Inc. (St. Louis, MO) and potassium[1-14 C]-palmitate was purchased from New England Nuclear Co. (Boston, MA). All other chemicals used were from Sigma (St. Louis, MO) unless otherwise stated. The plasma concentration of cholesterol, triglycerides, insulin, glucose, and free fatty acids, and the contents of liver cholesterol and triglyceride were measured as described previously (1,16). Transgene construction and generation of transgenic ap2-nsrebp-1a mice-an expression plasmid containing the mouse ap2 promoter fused to amino acids of the human SREBP-1a cdna was constructed as follows. A plasmid containing 5.4-kb of the ap2 gene promoter in pbluescript II SK(+) (designated pap2-pro) was a generous gift from Dr. B. M. Spiegelman. A 1.5-kb EcoRI-SalI cdna fragment of human SREBP-1a encoding amino acids was excised from the ppepck-srebp-1a460 plasmid (17) and ligated into EcoRI and SalI sites of the pcmv5 vector (D. W. Russell, unpublished construct). The resulting plasmid was designated pcmv5-srebp-1a460. A 1.7-kb EcoRI-SphI fragment containing SREBP-1a (amino acids 1-460), and a 3' polyadenylation signal was excised from pcmv5-srebp-1a460 and blunted using DNA polymerase I (klenow fragment). The resulting fragment was designated SREBP-1a460- polya. SREBP-1a460-polyA was subcloned into the SmaI site of pap2-pro to yield the final plasmid designated pap2-srebp-1a460. The integrity of this plasmid was confirmed by DNA sequencing.

8 8 Techniques used for generating transgenic mice were as previously described (18). A 7.1-kb NotI-ClaI fragment containing the ap2-promotor, SREBP-1a460, and poly (A +) tail from pap2-srebp-1a460 was gel-purified as described (17). A total of 613 fertilized eggs were microinjected with the ap2-srebp-1a construct and survived to the two-cell stage. Among the 74 offspring, 19 (26%) had integrated the transgene as determined by dot-blot hybridization of DNA from tail homogenates. Of 19 founder mice subjected to partial adipectomy, 10 (53%) expressed the human SREBP-1a transcript as determined by Northern blotting (17). Mice with high levels of expression were bred to C57BL/6J X SJL F1 mice (The Jackson Laboratories, Bar Harbor, ME), and four independent transgenic lines were established. The transgenic lines were designated ap2-nsrebp-1a lines A, B, C, and D, for lines 854-3, 951-4, 855-3, and in order of relative transgene mrna expression in WAT from highest to lowest. All mice were housed in colony cages and maintained on a 14 h light/10 h dark cycle, and fed Teklad Mouse/Rat Diet #7002 from Harlan Teklad Premier Laboratory Diets (Madison, WI). Immunoblot analysis-nuclear extracts and membrane fractions were prepared from mouse livers and adipose tissue as described previously (19). Immunoblot analysis of mouse SREBP-1 and SREBP-2 was carried out exactly as previously described (17,20). Blot hybridization of RNA-All cdna probes for Northern blot analysis have been described previously (12,17,20,21). Total RNA was prepared using an RNA STAT-60 TM (TEL-TEST"B". INC., Friendswood, TX). For Northern gel analysis, equal aliquots of total RNA made from adipose tissue from each mouse described in Table I (line A) were

9 9 pooled (total, 10µg), subjected to electrophoresis in a 1% formaldehyde agarose gel, and transferred to Hybond N + membrane (Amersham). Hybridization conditions and cdna probe preparations were carried out as described (1). The resulting bands were quantified by exposing the filter to a BAS 1000 Fuji PhosphorImager (Fuji Medical Systems, Stamford, CT). The fold-change was calculated after the signal was normalized to the signal generated by GAPDH. RNase protection assay-rnase protection assays for preadipocyte factor-1 (Pref-1) and mouse tumor necrosis factor-alpha (TNF- ) were done using the cdna template described previously (1). Anti-sense crna was transcribed with [ - 32 P]CTP (20 mci/ml) using bacteriophage T7 RNA polymerase (Ambion, Inc., Austin, TX). Specific activities of the transcribed RNAs were measured in each experiment and were in the range of X 10 9 cpm/ g for all RNAs except for -actin RNA, which was X 10 8 cpm/ g. Aliquots of total RNA (20 g ) from each sample were subjected to the RNase protection assay using a HybSpeed TM RPA kit (Ambion, Inc.). Each assay tube contained a crna probe for the mrna to be tested plus crna probe complementary to the mrna of actin. In preparing the probes, we adjusted the specific activity of the [ - 32 P]CTP to give a -actin signal comparable to the test mrnas. After digestion of RNase A/T1, protected fragments were separated on 8 M urea/4.8 % polyacrylamide gels. The gels were then dried and subjected to autoradiography using reflection film and intensifying screens (Dupont, Wilmington, DE). The dried gels were also analyzed quantitatively with a Bioimaging analyzer using BAS 1000 MacBAS software (Fuji Medical System). The level of -actin mrna in each RNA sample was used to normalize signals obtained for the test mrnas.

10 10 Acyl-CoA synthetase and stearoyl-coa activity measurements-liver acyl-coa synthetase (ACS) activity was measured in 8-wk-old non-fasted male mice (5 wild-type and 5 ap2- nsrebp-1a, line A) during the early light cycle. ACS activity was measured as described (22), with minor modifications. Briefly, the tissues were washed in 5 mm Tris- HCl buffer (ph 7.4) containing 0.25 M sucrose and the adventitia was carefully removed. The tissues were then minced, homogenized with 10 volumes of 5 mm Tris-HCl buffer (ph 7.4) containing 0.25 M sucrose, and centrifuged at 800 g for 5 min. Five g of protein from the supernatant was added to a reaction mixture that contained 150 mm Tris-HCl buffer (ph 7.4), 0.1% Triton X-100, 50 mm MgCl 2, 20 mm ATP, 200 mm potassium [1-14 C]palmitate (56 mci/mmol), 200 mm CoA-SH, 2.25 mm glutathione in a total volume of 250 l. This mixture was incubated for 5 min at 37 C. The reaction was terminated by adding 1 ml of isopropanol:heptane:1n H 2 SO 4 (40:10:1). Following thorough mixing, 0.35 ml of water and 0.6 ml of heptane was added and mixed by shaking for 5 min. The upper organic phase was removed and discarded. The lower water phase was washed twice with 0.6 ml of heptane and the radioactivity in 0.7 ml of the lower phase was counted. Hepatic stearoyl-coa desaturase (SCD) activity was measured in 8-wk-old nonfasted male mice (7 wild-type and 6 ap2-nsrebp-1a, line A) during the early light cycle, as described previously (21). Primary adipocytes-adipocytes were prepared and pooled from epididymal and brown fat pads from four 12-wk-old male wild-type and four littermate ap2-nsrebp1a (line A) mice. Fat depots were resected under aseptic conditions, and adipocytes were isolated by

11 11 collagenase digestion according to the Rodbell procedure (23) with minor modifications as described below. The fat pads were minced in Krebs-Ringer HEPES buffer (ph 7.4; containing 5 mm D-glucose, 2% BSA, 135 mm NaCl, 2.2 mm CaCl2 2H 2 O, 1.25 mm MgSO 4 7H 2 O, 0.45 mm KH 2 PO 4, 2.17 mm Na 2 HPO 4, and 10 mm HEPES). Adipose tissue fragments were digested in the Krebs-Ringer Hepes Buffer with 1.25 mg/ml rat type II collagenase at 37 C with gentle shaking at 60 cycles/min for 45 min. The resultant cell suspension was diluted in 13 ml cold Krebs-Ringer HEPES buffer. Isolated adipocytes were separated from undigested tissue by filtration through 4 layers of cheesecloth and washed three times in cold Krebs-Ringer HEPES buffer. For washing, cells were resuspended in 50 ml buffer and centrifuged at 400 g for 5 min. The final wash was in 13 ml of the culture medium (DMEM containing 10% fetal bovine serum). Floating cells were collected as primary adipocytes and plated on 60-mm collagen treated dishes (Cat. no ; Becton Dickinson Labware, Bedford, MA) and cultured at 37 C in 9% CO 2. Adipocytes were incubated with DMEM containing 10% FBS, Gentamycin (0.1 mg/ml) and 0.5 mm sodium [ 14 C]acetate (18 dpm/pmol) for 5 hr. Incorporation of [ 14 C]acetate into fatty acids and cholesterol was determined as previously described (8). The data are expressed as nanomoles of [ 14 C]acetate incorporated into fatty acids and picomoles of [ 14 C]acetate incorporated into cholesterol per g of cellular DNA. RESULTS Unregulated overexpression of nsrebp-1a exclusively in adipose tissue was achieved by the construction of a transgene encoding only the transcriptionally active fragment of human SREBP-1a driven by the adipose-specific enhancer/promoter of the ap2 gene (24).

12 12 The truncated SREBP-1a transgene product terminates prior to the membrane-attachment domain and, therefore, enters the nucleus directly without a requirement for proteolysis (3). We studied four lines of transgenic mice that were derived from independent founders (lines A-D). Fig. 1A shows a Northern blot analysis of mrna derived from the ap2-nsrebp-1a transgene in line A mice, confirming that the transgene mrna was expressed abundantly in both WAT and BAT. The transgene mrna level in WAT from lines B, C, and D was 70%, 35% and 30% the level of that measured in line A (data not shown). By Northern blotting, there was no evidence of transgene mrna expression in liver (Fig.1A), or in brain, spleen, and kidney (data not shown). As shown in Fig. 1B, transgenic ap2-nsrebp-1a (line A) mice had abundant nsrebp-1a protein expressed in adipose tissue. Phenotypic parameters in ap2-nsrebp-1a mice Table I lists the relevant phenotypic parameters measured in ap2-nsrebp-1a mice (lines A and B) and their respective wild-type littermates. The most profound abnormalities were in ap2-nsrebp-1a line A mice. Similar, but less profound changes were observed in the line B mice, which expressed lower levels of nsrebp-1a. The significant abnormalities noted in ap2-nsrebp-1a mice were as follows: 1) enlarged livers; 2) elevated hepatic triglyceride content; 3) enlarged interscapular brown fat pads; and 4) increased concentration of plasma FFAs. Plasma levels of cholesterol, triglyceride, insulin, and glucose were not significantly altered. Of note, the epididymal fat pad weights were not significantly different than those from littermate wild-type mice.

13 13 Histologic changes in fat of transgenic mice The histology of WAT and BAT from representative wild-type and ap2-nsrebp- 1a (line A) mice is shown in Fig. 2. The epididymal fat pad from a 40-day-old wild-type mouse has mature adipocytes of uniform size containing unilocular fat droplets (Fig. 2A). The epididymal fat pad from a 40-day-old transgenic littermate mouse also shows mature adipocytes with distinct unilocular vacuoles; however, a small percentage of adipocytes were significantly larger than wild-type adipocytes (Fig. 2B). This phenotype became more prominent as the mice aged. Fig. 2C shows WAT from a 70-day-old transgenic mouse. The diameter of most adipocytes from ap2-nsrebp-1a mice was markedly greater than adipocytes from wild-type mice. Many fat pads from mice of this age also showed an inflammatory response, characterized by a monocytic infiltration. Equally dramatic histologic changes were observed in BAT from ap2-nsrebp-1a mice. The BAT from a 40-day-old wild-type mouse consists of small multilocular adipocytes as shown in Fig. 2D. In contrast, most adipocytes in BAT from ap2-nsrebp- 1a mice were markedly enlarged and contained predominantly unilocular fat droplets. This resulted in a histologic appearance similar to that of normal WAT (Fig. 2E). Expression of mrnas for target genes in fat Fig. 3 shows multiple Northern blots measuring mrnas for genes encoding proteins involved in lipid metabolism and adipogenesis in WAT (Fig. 3A) and BAT (Fig. 3B) from wild-type and ap2-nsrebp-1a (line A) mice. Significant increases were measured in the mrnas for the LDL receptor as well as enzymes involved in cholesterol biosynthesis in WAT and BAT from transgenic mice. In particular, the mrna encoding

14 14 HMG-CoA reductase, a rate-controlling enzyme in the cholesterol biosynthetic pathway, was increased more than 20-fold in WAT and BAT. The mrnas for the fatty acid and triglyceride biosynthetic enzymes, acetyl-coa carboxylase-1 (ACC-1), fatty acid synthase (FAS), and glycerol-3-phosphate acyltransferase (GPAT) were also markedly increased in both adipose tissues of transgenic mice. WAT and BAT from transgenic mice showed normal mrna expression levels of genes encoding proteins that are associated with fully differentiated adipocytes: PPAR, C/EBP, and ap2. Of note, the mrna for UCP-1, which is highly expressed in BAT, was reduced by 90% in the interscapular brown fat pads from the transgenic mice. An RNase protection assay was used to determine whether overexpression of nsrebp-1a affected the expression of Pref-1 (Fig. 4A) and TNF- (Fig. 4B) in WAT and BAT. The level of Pref-1 mrna, a gene that is highly expressed in preadipocytes, was reduced by approximately 50% in both WAT and BAT of transgenic mice. The amounts of TNF- mrna did not change in either WAT or BAT of transgenic mice. Synthesis and secretion of lipid from primary adipocytes It is estimated that only 50-60% of the cells in adipose tissue are adipogenic lipidcontaining cells (25). The remaining 40% are mesenchymal cells, fibroblasts and endothelial cells. In the WAT of ap2-nsrebp-1a mice, infiltration of inflammatory cells was also observed. To assess the effect of nsrebp-1a overexpression on rates of fatty acid biosynthesis specifically in adipocytes, primary adipocytes were isolated from WAT and BAT, pulse-labeled with [ 14 C]acetate, and the amount of [ 14 C]-labeled fatty acids produced in the cells (Fig. 5A) and released into the medium (Fig. 5B) were measured.

15 15 The rates of [ 14 C]acetate incorporation into fatty acids in white and brown adipocytes from transgenic mice were increased 22- and 6-fold, respectively (Fig. 5A). Release of newly synthesized [ 14 C]-labeled fatty acids into the medium was significantly increased in primary adipocytes from WAT and BAT of transgenic mice (7- and 9-fold, respectively) (Fig. 5B). The increased release of fatty acids from white and brown adipocytes is likely responsible for the higher FFA levels measured in the plasma of ap2- nsrebp-1a mice. Histologic changes in livers of transgenic mice Fig. 6A compares the appearance of the liver from one representative ap2- nsrebp-1a mouse (line A) and the liver from a wild-type littermate. Livers from ap2- nsrebp-1a mice were slightly enlarged and exhibited a distinct pale color that was consistently different from that of wild-type mice. Fig. 6 (panels B and C), show the histologic appearance of the livers from a 2-month-old wild-type mouse and transgenic littermate mouse (line A), respectively. Many of the hepatocytes in the transgenic animal are distended by fat droplets. There is neither generalized inflammation nor overt signs of necrosis. Similar macroscopic and histologic appearances were also observed in the livers of ap2-nsrebp-1a line B mice (data not shown). Measurements of SREBPs and enzymes related to lipid accumulation in liver To determine whether the hepatic triglyceride accumulation in ap2-srebp-1a mice was due to enhanced expression of genes involved in lipid synthesis, we first measured the mrna and protein levels of all SREBPs. Previously, we showed that in the fatty livers of the lipodystrophic ap2-nsrebp-1c mice, hepatic nsrebp-1c was

16 16 significantly elevated, which led to the activated lipogenic target genes and increased rates of fatty acid synthesis (26). Fig. 7A shows immunoblots of the precursor and nuclear forms of SREBP-1 and SREBP-2 in livers from wild-type and ap2-nsrebp-1a (line A) littermate mice. No significant changes were detected in the amounts of nuclear or precursor forms of SREBP-1 and SREBP-2. Fig. 7B contains mrna levels of genes encoding proteins involved in synthesis of cholesterol and fatty acids, and fatty acidmodifying enzymes. No substantial changes were measured in the mrnas for genes involved in the fatty acid biosynthesis such as ACC-1 and FAS, nor in cholesterol biosynthetic genes such as HMG-CoA synthase and HMG-CoA reductase. The lack of change in the mrnas for lipogenic enzymes was reflected in a lack of change in synthetic rates for fatty acid and cholesterol in intact livers as measured in vivo by the incorporation of tritiated water (data not shown). The mrnas for enzymes that modify fatty acids, namely, ACS, SCD, and GPAT, were all slightly increased. To determine whether the change measured in the mrnas led to increased enzymatic activity of ACS and SCD, we measured the enzymatic activities of each in livers from wild-type and ap2- nsrebp-1a mice (Table II). The livers from transgenic mice had more than 2-fold higher activities for both ACS and SCD, compared with littermate controls. Taken together, these results suggest that the lipid accumulation in livers of ap2- nsrebp-1a mice is not due to increased de novo synthesis of fatty acids, but due to the increased influx of FFA from plasma. The influx of FFA is likely the stimulus that leads to increased mrnas for the fatty acid modifying enzymes, possibly through PPAR mediated transcriptional activation (27,28). Correlation between transgene expression and fatty liver

17 17 To determine whether the degree of fatty liver correlated with the level of SREBP-1a transgene expression in adipose tissue, we measured the triglyceride content in the livers from the four independent lines of ap2-nsrebp-1a mice. These mice produced varying amounts of human nsrebp-1a transgene mrna in white fat pads as determined by Northern blot analysis. Fig. 8 shows that the amount of hepatic triglyceride accumulation in the four lines of transgenic mice correlated with the level of mrna expression of the nsrebp-1a transgene in white fat from each founder mouse. This further supports the hypothesis that the lipid accumulation in liver is a direct result of increased FFA delivery from the peripheral adipose tissue overexpressing transcriptionally active SREBP-1a.

18 18 DISCUSSION In this report, we show that nsrebp-1a overexpression in adipose tissue of mice activates both the cholesterol and fatty acid biosynthetic pathways by increasing the levels of mrnas encoding the key genes in each pathway. This activation resulted in enhanced rates of cholesterol and fatty acid synthesis in both BAT and WAT. These changes are similar to those found in the livers of transgenic mice overexpressing nsrebp-1a (17) and cultured cells that overexpressed nsrebp-1a (8). As a consequence of enhanced synthesis of lipids, the adipocytes in the white fat pads of transgenic mice accumulated excess amounts of triglyceride, which in turn caused severe adipocyte hypertrophy that was associated with an inflammatory reaction. Brown fat pads were markedly enlarged due to hypertrophy of the brown adipocytes, which showed the histologic appearance of mature white adipocytes. The ap2-srebp-1a mice developed a fatty liver presumably from an increased of fatty acids derived from adipose tissue. Altered white adipocyte structure/function The levels of nsrebp-1a transgene expression in the WAT and BAT of ap2- nsrebp-1a (line A) transgenic mice were similar to the transgene levels of nsrebp-1c measured in the adipose tissues of ap2-nsrebp-1c mice (data not shown) (1). Despite similar transgene expression levels in the ap2-srebp-1a and ap2-srebp-1c mice, the effects on adipocyte differentiation and function were distinctly different. The white fat pads of 5-wk-old lipodystrophic ap2-nsrebp-1c mice were composed predominantly of immature adipocytes; characterized by an abundant eosinophilic cytoplasm, centrally located nucleus, and a small but distinct fat droplet (1).

19 19 The mrna levels of genes normally expressed in differentiated adipocytes (PPAR, C/EBP, and adipsin) were markedly reduced in WAT from ap2-nsrebp-1c mice suggesting that these cells were metabolically undifferentiated. In addition, the abundance of Pref-1 transcripts, a prominent marker of immature adipocytes was increased in WAT pads of the lipodystrophic ap2-nsrebp-1c mice (1). In contrast, the white fat pads of 5-wk-old ap2-nsrebp-1a mice consisted predominantly of fully differentiated adipocytes, many of which were hypertrophic. The mrnas levels of genes associated with fully differentiated adipocytes were expressed at normal levels and mrnas encoding lipogenic enzymes were markedly increased. Our laboratory previously reported that during the differentiation of 3T3-L1 cells, the abundance of SREBP-1a increases, which is followed by the induction of mrnas for genes encoding fatty acid, cholesterol and triglyceride biosynthetic enzymes (10). Together, these observations suggest that SREBP-1a contributes to adipogenesis through the activation of lipid synthesis. The exact mechanisms by which SREBP-1a and SREBP-1c differentially influence adipocyte differentiation in vivo remains to be elucidated. In cultured cells and in liver, SREBP-1a is a much more potent activator of virtually all target genes controlling cholesterol and fatty acid biosynthesis than is SREBP-1c (8,9). In HepG2 cells, overexpression of nsrebp-1a increased the transcription of PPAR presumably by activating the promoters of PPAR and (29). However, in fat of ap2-nsrebp-1a mice, there was no change in the abundance of PPAR transcripts suggesting that overexpression of SREBP-1a does not induce the transcription of this key regulator of

20 20 adipogenesis in vivo. However, this does not exclude the possibility that SREBP-1a activates genes that produce a ligand that activates PPAR (30). In contrast to the markedly elevated levels of plasma insulin and glucose observed in the lipodystrophic ap2-nsrebp-1c mice, ap2-nsrebp-1a transgenic mice had normal plasma insulin and glucose levels, indicating that glucose homeostasis was normal. The insulin-resistance and ensuing diabetes in the lipodystrophic ap2-nsrebp-1c mice were the consequences of the dramatic reduction in the levels of circulating leptin that resulted from the failure of adipocyte differentiation (31). There was no evidence, however, of inadequate adipocyte differentiation in the ap2-nsrebp-1a mice and circulating levels of leptin were normal. The normal leptin levels likely accounts for the lack of an effect on glucose metabolism. In preliminary studies, we have found mild hyperinsulinemia does develop in geriatric age ap2-srebp-1a mice. Although we have not investigated the mechanism responsible for the hyperinsulinemia, it can not be a result of altered adipogenesis as was observed in the lipodystrophic ap2-srebp-1c mice. Altered brown adipocyte structure/function The interscapular brown fat pad in ap2-nsrebp-1a mice was enlarged and resembled white fat in color and texture. Histologically, the brown adipocytes were almost indistinguishable from normal mature white adipocytes. The change in brown adipocyte morphology was accompanied by a significant reduction in UCP-1, the major differentiation marker of the brown adipocytes. A similar change of morphology and abundance of UCP-1 mrna was reported in the BAT of ob/ob mice (32). This change in the brown adipocytes of ap2-nsrebp-1a and ob/ob mice may be simply a result of

21 21 excess lipid synthesis within the brown adipocyte which causes a morphologic and metabolic conversion to a white adipocyte. Adipocyte fatty acid secretion and hepatic steatosis The livers of ap2-nsrebp-1a mice exhibited a moderate accumulation of triglyceride as compared with massive accumulation of triglyceride in livers of ap2- nsrebp-1c mice. We believe that the source of the hepatic lipid was entirely different in the transgenic ap2nsrebp-1a and ap2-nsrebp-1c mice. Neither the amount of hepatic nuclear SREBPs nor the rate of hepatic fatty acid synthesis was altered when nsrebp-1a was overexpressed in adipose tissue. These results are in marked contrast to the changes that occur in the liver of ap2-nsrebp-1c. In these mice, hyperinsulinemia stimulates SREBP-1c transcription, which leads to an increase in nsrebp-1c protein. Increased nsrebp-1c protein results in the transcriptional activation of genes encoding the fatty acid biosynthetic enzymes, which significantly increased rates of hepatic fatty acid synthesis (12,26). This resulted in liver triglyceride levels that were more than 10-fold higher than wild-type. In contrast, ap2-nsrebp-1a mice had normal levels of insulin and glucose, and normal hepatic levels of SREBP-1c and mrnas encoding lipogenic enzymes. The liver has two sources of fatty acids-endogenously synthesized fatty acids, and fatty acids derived from the peripheral fat stores or diet. Increased fatty acid influx from adipose tissue may contribute to the hepatic steatosis associated with obesity and type 2 diabetes (33). Before fatty acids can be incorporated into triglycerides, phospholipids or cholesteryl esters, they must be activated to their acyl-coa form by an acyl-coa synthetase. Fatty acids can stimulate the expression of the ACS gene and its activity in a

22 22 number of different tissues and cell types (34,35). The combination of increased plasma FFAs, increased activities of the hepatic fatty acid-modulating enzymes (ACS and SCD), and unchanged endogenous hepatic fatty acid synthesis, suggests that the fatty livers observed in ap2-nsrebp-1a mice are purely a result of the very high fatty acid biosynthetic and secretion rates present in adipocytes. In addition, there was a significant correlation between the levels of the transgene expressed in white adipose tissue of independently derived lines of ap2-nsrebp-1a mice and the degree of hepatic triglyceride accumulation. Comparison of ap2-nsrebp-1a, ap2-nsrebp-1c and SREBP-1 knockout mice phenotype Table 3 summarizes the phenotypes observed in transgenic ap2-nsrebp-1a, transgenic ap2-nsrebp-1c (10) and SREBP-1 knockout mice (20). The SREBP-1 lack both SREBP-1a and SREBP-1c proteins. Interestingly, the SREBP-1 knockout mice had normal amounts of adipose tissue that were populated with fully differentiated adipocytes that expressed normal levels of adipocyte-specific transcripts (20). The lack of any recognized defect in adipose tissue suggests that SREBP-1a and SREBP-1c are dispensable for adipocyte differentiation. It is also possible, however, that there is sufficient SREBP-2 in adipose tissue to compensate for the loss of SREBP-1. Overexpression of SREBP-1a in adipocytes promotes lipid synthesis and accumulation, both in cultured 3T3-L1 adipocytes (10), and in vivo, which suggests that SREBP-1a promotes lipid accumulation during adipose development. Overexpression of SREBP-1c in developing adipocytes inhibits their differentiation through an, as yet, unknown mechanism. It is possible that overproduction of nsrebp-1c may be interfering with the

23 23 activities of other transcription factors required for adipogenesis, which could include nsrebp-1a, nsrebp-2 and/or other basic-helix-loop-helix transcriptional factors. A common phenotypic consequence of the overexpression of both SREBP-1a and SREBP- 1c in adipose tissue was fatty liver. In the case of overexpression of nsrebp-1a in adipose tissue, fatty livers develop as a consequence of an increased influx of adiposederived FFA into the liver (Table III). Overexpression of SREBP-1c in adipose tissue inhibits adipogenesis, causing a deficiency in leptin, which results in hyperinsulinemia, and increased hepatic fatty acid biosynthesis (12,26,31). The dramatically different phenotypes that resulted from overexpressing these two closely related SREBP-1 isoforms in fat was unexpected. The SREBP-1a and SREBP-1c proteins have identical DNA binding domains and only differ in the length of their transactivation domains. The current results raise the possibility that SREBP-1a and SREBP-1c have tissue-specific transcriptional activation properties that are modulated by as yet unidentified co-activators or co-repressors. A more thorough understanding of the pathways regulated by each SREBP-1 isoform may be elucidated in the future by comparing genome-wide tissue-specific gene expression patterns in SREBP-1 transgenic and knockout mice.

24 24 Acknowledgments We thank Drs. Michael S. Brown and Joseph L. Goldstein for their critical reading of the manuscript. We also thank our former colleague Hitoshi Shimano for reagents and helpful suggestion; Beth Hinnant and Scott Clark for excellent technical assistance; Mario Villarreal, Elizabeth Lummus, Shanna Maika for excellent help with the animal studies. This work was supported by grants from the National Institutes of Health HL , the Perot Family Foundation, the Moss Heart Foundation, and the W.M. Keck Foundation. J.D.H. is a Pew Scholar in the Biomedical Sciences and the recipient of the Established Investigator Grant from the American Heart Association. 1 The abbreviations used are: ACS, acyl-coa synthetase; ACC-1, acetyl-coa carboxylase-1; bp, base pair; BAT, brown adipose tissue; CoA, coenzyme A; FAS, fatty acid synthase; FFA, free fatty acids; GAPDH, glyceraldehyde-3-phophosphate dehydrogenase; GPAT, glycerol-3-phosphate acyltransferase; HMG-CoA, 3-hydroxy-3- methylglutaryl coenzyme A; kb, kilobase pair; Pref-1, preadipocyte factor-1; SCD, stearoyl-coa desaturase; SCAP, SREBP cleavage-activating protein; SREBP, sterol regulatory element-binding protein; TNF-, tumor necrosis factor-alpha; UCP-1, uncoupling protein-1; WAT, white adipose tissue.

25 25 References 1. Shimomura, I., Hammer, R. E., Richardson, J. A., Ikemoto, S., Bashmakov, Y., Goldstein, J. L., and Brown, M. S. (1998) Genes Dev. 12, Mokdad, A. H., Bowman, B. A., Ford, E. S., Vinicor, F., Marks, J. S., and Koplan, J. P. (2001) JAMA 286, Brown, M. S., and Goldstein, J. L. (1997) Cell 89, Tontonoz, P., Kim, J. B., Graves, R. A., and Spiegelman, B. M. (1993) Mol. Cell. Biol. 13, Kim, J. B., and Spiegelman, B. M. (1996) Genes Dev. 10, Yokoyama, C., Wang, X., Briggs, M. R., Admon, A., Wu, J., Hua, X., Goldstein, J. L., and Brown, M. S. (1993) Cell 75, Hua, X., Wu, J., Goldstein, J. L., Brown, M. S., and Hobbs, H. H. (1995) Genomics 25, Pai, J.-t., Guryev, O., Brown, M. S., and Goldstein, J. L. (1998) J. Biol. Chem. 273, Shimano, H., Horton, J. D., Shimomura, I., Hammer, R. E., Brown, M. S., and Goldstein, J. L. (1997) J. Clin. Invest. 99, Shimomura, I., Shimano, H., Horton, J. D., Goldstein, J. L., and Brown, M. S. (1997) J. Clin. Invest. 99, Shimomura, I., Bashmakov, Y., Shimano, H., Horton, J. D., Goldstein, J. L., and Brown, M. S. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, Shimomura, I., Bashmakov, Y., Ikemoto, S., Horton, J. D., Brown, M. S., and Goldstein, J. L. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, Hua, X., Yokoyama, C., Wu, J., Briggs, M. R., Brown, M. S., Goldstein, J. L., and Wang, X. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, Horton, J. D., Goldstein, J. L., and Brown, M. S. (2002) J. Clin. Invest. 109, Sambrook, J., and Russell, D. W. (2001) Molecular cloning: a laboratory manual, 3rd Ed. 3 vols., Cold Spring Harbor Laboratory Press, New York 16. Ishibashi, S., Brown, M. S., Goldstein, J. L., Gerard, R. D., Hammer, R. E., and Herz, J. (1993) J. Clin. Invest. 92, Shimano, H., Horton, J. D., Hammer, R. E., Shimomura, I., Brown, M. S., and Goldstein, J. L. (1996) J. Clin. Invest. 98, Hofmann, S. L., Russell, D. W., Brown, M. S., Goldstein, J. L., and Hammer, R. E. (1988) Science 239, Sheng, Z., Otani, H., Brown, M. S., and Goldstein, J. L. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, Shimano, H., Shimomura, I., Hammer, R. E., Goldstein, J. L., Brown, M. S., and Horton, J. D. (1997) J. Clin. Invest. 100, Shimomura, I., Shimano, H., Korn, B. S., Bashmakov, Y., and Horton, J. D. (1998) J. Biol. Chem. 273, Bar-Tana, J., Rose, G., and Shapiro, B. (1971) Biochem. J. 122, Rodbell, M. (1964) J. Biol. Chem. 239, Ross, S. R., Graves, R. A., Greenstein, A., Platt, K. A., Shyu, H.-L., Mellovitz, B., and Spiegelman, B. M. (1990) Proc. Natl. Acad. Sci. U. S. A. 87,

26 Ailhaud, G., Grimaldi, P., and Negrel, R. (1992) Annu. Rev. Nutr. 12, Shimomura, I., Bashmakov, Y., and Horton, J. D. (1999) J. Biol. Chem. 274, Miller, C. W., and Ntambi, J. M. (1996) Proc. Nat. Acad. Sci. U. S. A. 93, Martin, G., Schoonjans, K., Lefebvre, A.-M., Staels, B., and Auwerx, J. (1997) J. Biol. Chem. 272, Fajas, L., Schoonjans, K., Gelman, L., Kim, J. B., Najib, J., Martin, G., Fruchart, J. C., Briggs, M., Spiegelman, B. M., and Auwerx, J. (1999) Mol. Cell Biol. 19, Kim, J. B., Wright, H. M., Wright, M., and Spiegelman, B. M. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, Shimomura, I., Hammer, R. E., Ikemoto, S., Brown, M. S., and Goldstein, J. L. (1999) Nature 401, Jacobsson, A., Stadler, U., Glotzer, M. A., and Kozak, L. P. (1985) J. Biol. Chem. 260, Unger, R. H. (1995) Diabetes 44, Suzuki, H., Kawarabayasi, Y., Kondo, J., Abe, T., Nishikawa, K., Kimura, S., Hashimoto, T., and Yamamoto, T. (1990) J. Biol. Chem. 265, Schoonjans, K., Watanabe, M., Suzuki, H., Mahfoudi, A., Krey, G., Wahli, W., Grimaldi, P., Staels, B., Yamamoto, T., and Auwerx, J. (1995) J. Biol. Chem. 270, McGarry, J. D. (1992) Science 258,

27 27 FIGURE LEGENDS FIG. 1. Amount of human truncated nsrebp-1a transgene mrna and nuclear protein in WAT (epididymal fat) and BAT from transgenic ap2-nsrebp-1a (line A) mice. (A) Total RNA was isolated from WAT, BAT, and liver from 5-wk-old male mice (5 wild-type and 5 ap2-nsrebp-1a littermate mice), and equal aliquots were pooled. Aliquots of pooled RNA (10 g) were subjected to electrophoresis and blot hybridization with 32 P-labeled human SREBP-1a and GAPDH probes. Tg; human SREBP-1a transgene mrna. (B) Immunoblot analysis of the human nuclear SREBP-1a transgene protein in nuclear extracts from WAT and BAT of wild-type (WT) and transgenic ap2- nsrebp-1a mice (Tg). Adipose tissue from mice described in (A) was pooled, and 30 g of nuclear protein extract was loaded into each lane. Tg; human ap2-nsrebp-1a transgene-derived protein. FIG. 2. Representative histologic sections of WAT and BAT. Epididymal fat pads from 40-day-old wild-type (panel A) and transgenic ap2-nsrebp-1a (line A) littermate mouse (panel B). Epididymal fat pad from 70-day-old transgenic ap2-nsrebp-1a (line blot hybridization. Total RNA was isolated from WAT and BAT of the 5-wk-old wild- A) mouse (panel C). Interscapular brown fat pads from 40-day-old wild-type (panel D) and transgenic ap2-nsrebp-1a (line A) littermate (panel E) mouse. Note that the adipocytes of the white and brown fat pads from the transgenic mice are markedly enlarged (hemotoxilin and eosin stain; magnification, 112 x). FIG. 3. Amounts of various mrnas in WAT (panel A) and BAT (panel B) from wild-type (WT) and transgenic ap2-nsrebp-1a (line A) (Tg) mice as measured by

28 28 type and transgenic ap2-nsrebp-1a (line A) mice described in Table I. Equal aliquots of total RNA from each mouse were pooled (10 µg, total) and subjected to electrophoresis and blot hybridization with the indicated 32 P-labeled cdna probe. The amount of radioactivity in each band was quantified as described in Experimental Procedures. The fold-change for each mrna, relative to that of wild-type, was calculated after correction for loading differences with GAPDH. These values are shown below each blot. FIG. 4. Levels of mrnas for Pref-1 (A) and TNF- (B) in wild-type (WT) and transgenic ap2-nsrebp-1a (line A) (Tg) mice as measured by RNase protection assay. Aliquots of total RNA (20 g) isolated from white fat (WF) and brown fat (BF) described in Fig. 2 were hybridized in solution for 10 min at 68 o C to the 32 P-labeled crna probes for Pref-1 and TNF-, all in the presence of a crna probe for -actin as described in Experimental Procedures. After RNase digestion, the protected fragments were separated by gel electrophoresis and exposed to film for 16 h at -80 o C. The radioactivity in the gels was quantified, normalized to the -actin signal, and expressed as the fold change-relative to the mrna level of wild-type mice. FIG. 5. Lipid synthesis and secretion rates from primary adipocytes of wild-type (hatched bar) and transgenic ap2-nsrebp-1a (line A) littermates (closed bar) mice. Primary adipocytes were prepared as described in Experimental Procedures. The adipocytes were plated on 60-mm collagen-treated dishes and incubated with DMEM containing 10% fetal bovine serum for 3h. The cells were then incubated with 0.5 mm sodium [ 14 C]acetate (18 dpm/pmol) in DMEM supplemented with 10% fetal bovine

29 29 serum for 5 hr, after which the media and cells were harvested for measurement of [ 14 C]- labeled fatty acids and cholesterol. FIG. 6. Representative gross and histologic sections of liver from wild-type and transgenic ap2-nsrebp-1a mice. (A) Liver from a 2-month-old wild-type male mouse and a transgenic ap2-nsrebp-1a littermate mouse (line A). (B and C) Histologic sections of liver from a 2-month-old wild-type male mouse (panel B) and a transgenic ap2-nsrebp-1a littermate mouse (line A) (panel C). Note the marked vacuolization due to lipid accumulation in the liver of the transgenic mouse (hemotoxilin and eosin stain; magnification, 112 x). FIG. 7. SREBP protein and various mrna levels in livers of wild-type and ap2- SREBP-1a mice. (A) Immunoblot analysis of SREBP-1 and SREBP-2 protein in membranes and nuclear extracts from livers of five 5-wk-old male wild-type and 5 transgenic ap2-nsrebp-1a (line A) littermate mice. Livers from the two groups of mice described in Table I were separately pooled, and aliquots of membrane protein (50 g) and nuclear extract protein (30 g) were subjected to 8% SDS-PAGE. Immunoblot analysis was performed as described in Experimental Procedures. P and N denote the precursor and cleaved nuclear forms of SREBP, respectively. (B) Amounts of various mrnas in livers from wild-type (WT) and transgenic ap2-nsrebp-1a (line A) (Tg) mice as measured by blot hybridization. Total RNA was isolated from livers of 5-wk-old wild-type and ap2-nsrebp-1a (line A) mice described in Table I. Total RNA from each mouse was pooled, and 10 µg aliquots were subjected to electrophoresis and blot hybridization with the indicated 32 P-labeled cdna probe. The amount of radioactivity in

30 30 each band was quantified as described in Experimental Procedures. The fold-change for each mrna, relative to that of the respective wild-type, was calculated after correction for loading differences with GAPDH. These values are shown below each blot. FIG. 8. Correlation between the amounts of SREBP-1a transgene mrna expression in WAT and the hepatic triglyceride content from 4 independent lines of transgenic ap2-nsrebp-1a mice. Transgene expression was measured by Northern blotting in the WAT from the four independent transgenic founder mice (A-D), and liver triglyceride contents were measured in 3-month-old F1 mice derived from each founder. The triglyceride content data is plotted as the mean ± S.E.M. for each line. The number in each group is as follows: 5 wild-type; 3 line D; 4 line C; 4 line B; and 5 line A.

31 TABLE I Phenotypic comparison of wild-type and ap2-nsrebp-1a mice Line A Line B 5 week old 10 week old 4 week old Parameter Wild-type Transgenic Wild-type Transgenic Wild-type Transgenic Sex 5 males 5 males 6 males 8 males 4 males 4 males Body weight (g) 19.4 ± ± ± ± ± ± 0.9 Liver weight (g) 1.3± ± 0.1 ** 1.4 ± ± 0.18 ** 0.88 ± ± 0.03 Epididymal fat weight (g) 0.13 ± ± ± ± 0.03 * ± ± Brown fat weight (g) ± ± ± ± ± ± ** Liver cholesterol content (mg/g) 2.3 ± ± ± ± ± ± 0.10 Liver triglyceride content (mg/g) 8.4 ± ± 1.2 ** 4.2 ± ± 0.4 ** 7.3 ± ± 1.1 ** Plasma cholesterol (mg/dl) 80 ± 5 66 ± 5 83 ± ± ± ± 7 Plasma triglycerides (mg/dl) 199 ± ± ± ± ± ± 13 Plasma insulin (ng/ml) 1.1 ± ± ± ± ± ± 0.3 Plasma glucose (mg/dl) 148 ± ± ± ± ± ± 19 Plasma free fatty acid (mm) 752 ± ± 159 ** 602 ± ± 161 * 686 ± ± 141 * Each value represents the mean ± SEM. Wild type mice were littermates of transgenic animals. *Values in parentheses denotes level of statistical significance (Student s t test); *P < 0.05 and **P < Downloaded from by guest on June 28, 2018

32 TABLE II Hepatic fatty acyl-coa synthase and stearoyl-coa desaturase activities Mouse Genotype Enzyme Activity Wild-type ap2-nsrebp-1a Fatty Acyl-CoA synthase (nmol/min/mg protein) 51 ± 6 (n=5) 146 ± 18 (n=5) (*p<0.001) Stearoyl-CoA desaturase (pmol/min/mg protein) 890 ± 48 (n=7) 2028 ± 270 (n=6) (*p<0.001) Each value represents the mean ± SEM. Wild-type mice were littermates of transgenic mice (line A). 8-wk-old mice were used for activity measurements as described in Experimental Procedures. *Values in parentheses denote the level of statistical significance (Student s t test).

33 TABLE III Phenotypic comparison of transgenic ap2-nsrebp-1a, transgenic ap2-nsrebp-1c and SREBP-1 knockout mice Mouse Genotype Parameter ap2-nsrebp-1a ap2-nsrebp-1c Srebp-1 -/- Adipocytes Hypertrophic Undifferentiated Normal Liver triglycerides Normal Plasma FFAs Normal Normal Plasma insulin Normal Normal Leptin levels Normal Normal Insulin resistance Absent Present Absent

34 Fig.1 A. Tissue White Fat Brown Fat Liver Genotype WT Tg WT Tg WT Tg B. 28S 18S GAPDH Tissue Genotype WT White Fat Tg Brown Fat WT Tg * 5 6 Tg kd Tg

35 Fig. 2

36 Fig. 3 A LDL Receptor HMG-CoA Synthase HMG-CoA Reductase ACC FAS GPAT GAPDH Fold Change 6.8 PPARγ 9.7 C/EBPα 22 ap2 3.2 Insulin Receptor 4.6 Glut4 6.2 Leptin GAPDH Fold Change B 1.0 LDL Receptor 1.0 HMG-CoA Synthase 0.9 HMG-CoA Reductase 1.0 ACC 0.9 FAS 0.8 GPAT GAPDH Fold Change PPARγ C/EBPα ap2 Insulin Receptor GLUT4 Leptin UCP1 GAPDH Fold Change

37 Fig. 4 Protected Fragments Pref-1 Actin A WF BF Fold Change B WF BF Fold Change TNF-α Actin E-104.ai

38 Fig. 5 [ 14 C]Acetate Incorporation into Fatty Acid (nmol / hr per µg DNA) [ 14 C]Acetate Incorporation into Cholesterol (pmol / hr per µg DNA) A. Synthesis in Cell B. Release to Medium White Fat Brown Fat Wild-type ap2-srebp-1a White Fat Brown Fat C. Synthesis in Cell D. Release to Medium White Fat Brown Fat White Fat Wild-type ap2-srebp-1a Brown Fat E-106.ai

39 Fig. 6

40 Fig. 7 A SREBP-1 SREBP-2 Genotype WT Tg Membranes KD 111 P Nuclear Extract B GADPH LDL Receptor WT 1 2 HMG-CoA Synthase 3 4 HMG-CoA Reductase N Squalene Synthase Tg Fold Change WT ACC FAS Acyl-CoA Synthase Stearoyl-CoA Desaturase Tg GPAT WT Tg GADPH Fold Change E-105.ai