Transgenic Mice Overexpressing SREBP-1a in Male ob/ob Mice Exhibit Lipodystrophy and Exacerbate Insulin Resistance

Abstract

Sterol regulatory element–binding protein (SREBP)-1a is a key transcription factor that activates the expression of genes involved in the synthesis of fatty acids, triglycerides (TGs), and cholesterol. Transgenic mice that overexpress the nuclear form of SREBP-1a under the control of the phosphoenolpyruvate carboxykinase promoter (Tg-1a) were previously shown to display a lipodystrophic phenotype characterized by enlarged and fatty livers, diminished peripheral white adipose tissue (WAT), and insulin resistance. In the current study, we crossed these Tg-1a mice with genetically obese (ob/ob) mice (Tg-1a;ob/ob) and examined change in fat distribution between liver and adipose tissues in severe obesity and mechanism underlying the lipodystrophic phenotype in mice with Tg-1a. Tg-1a;ob/ob mice developed more severe steatohepatitis but had reduced WAT mass and body weight compared with ob/ob mice. The reduction of WAT mass in Tg-1a and Tg-1a;ob/ob mice was accompanied by enhanced lipogenesis and lipid uptake in the liver, reduced plasma lipid levels, impaired adipocyte differentiation, reduced food intake, enhanced energy expenditure, and extended macrophage infiltration and fibrosis in WAT. Despite the improved glucose tolerance, Tg-1a;ob/ob mice showed severe peripheral insulin resistance. Adenoviral hepatic expression of SREBP-1a mimicked these phenotypes. The “fat steal”-like lipodystrophy phenotype of the Tg-1a;ob/ob model demonstrates that hepatic SREBP-1a activation has a strong impact on the partition of TG accumulation, resulting in adipose-tissue remodeling by inflammation and fibrosis and insulin resistance.

Dysregulation of systemic lipid metabolism plays an important role in the development of metabolic disorders (1–3). Adipose tissue participates in the regulation of lipid homeostasis and energy balance, and its dysfunction results in obesity (excess adipose tissue) or lipodystrophy (insufficiency of adipose tissue) (4–7). Although these two conditions are pathologically opposite states, both are accompanied by similar metabolic consequences, including hepatic steatosis, insulin resistance, and type 2 diabetes mellitus (7).

Lipodystrophy is a pathological state of adipose tissue deficiency. Lipodystrophy may be generalized or partial, depending on the degree and location of fat loss (8–12). Both genetic and acquired factors are responsible for its occurrence (12–15). The understanding of the mechanisms by which lipodystrophy develops should facilitate the development of therapeutic strategies.

Sterol regulatory element–binding protein (SREBP) family members have been established as transcription factors, regulating the transcription of genes involved in the uptake, biosynthesis, and metabolism of cholesterol and fatty acids (16–18). SREBPs are synthesized as precursors bound to the endoplasmic reticulum and nuclear envelope. Under conditions of sterol deprivation, SREBPs are cleaved to liberate the amino-terminal portion containing a basic helix-loop-helix leucine zipper domain by site-1 and site-2 proteases. SREBPs enter the nucleus where they can bind to specific sterol response elements in the promoters of target genes (19). The mammalian genome encodes three SREBP isoforms, designated SREBP-1a, SREBP-1c, and SREBP-2, which have different characteristics. Although SREBP-2 activates the expression of genes governing the synthesis and uptake of cholesterol, SREBP-1c activates the expression of genes involved in the synthesis of fatty acids and triglycerides (TGs) (16, 19, 20). SREBP-1a is a potent activator of all sterol response element–regulated genes, including those mediating the synthesis of cholesterol, fatty acids, TGs, and phospholipids (19, 21).

To examine the role of SREBP-1 in the liver, transgenic mice that overexpress a truncated NH2-terminal segment of human SREBP-1a (TgSREBP-1a), which is the constitutive-active form of the protein under the control of the phosphoenolpyruvate carboxykinase (PEPCK) promoter, were created (21). These animals exhibited massive hepatic enlargement, owing to engorgement with TGs and cholesterol esters. An unexpected finding was a progressive atrophy of white adipose tissue (WAT) as the animals aged. Brown et al. (21) suggested that the decreased WAT mass in the transgenic mice may be a result of the “TG steal” syndrome in which very low–density lipoprotein (VLDL)–TG are preferentially cleared into liver and diverted from adipose tissue. However, the molecular explanation of this phenotype remains to be elucidated. Moreover, it is unknown how SREBP-1a activation affects adiposity in the setting of obesity, which is commonly associated with type 2 diabetes.

To explore the mechanism underlying the lipodystrophic phenotype in mice with TgSREBP-1a and to test if SREBP-1a activation affects the fat storage and metabolism under conditions of positive energy balance, we placed TgSREBP-1a mice in the genetically obese (ob/ob) mice and examined the role of SREBP-1a in the regulation of adipose tissue mass and function.

Materials and Methods

Animals

All animal husbandry and animal experiments complied with the guidelines of the University of Tsukuba’s regulations of animal experiments and were approved by the Animal Experiment Committee of the University of Tsukuba. Transgenic mice that overexpress a dominant-positive, truncated form of human SREBP-1a under the control of the rat PEPCK promoter (Tg-1a) on a C57BL/6 background were generated as described previously (21). We obtained Lep^ob/+ (ob/+) mice on a C57BL/6J background from The Jackson Laboratory (Bar Harbor, ME) and crossed them with Tg-1a mice to obtain Tg-1a;ob^+/− mice. Finally, male Tg-1a;ob^+/− mice and female ob^+/− mice were bred to generate Tg-1a;ob/ob mice. Measurement of body fat composition (total body fat, subcutaneous fat, and visceral fat) was performed by a computed tomography scan (La Theta LCT-100; Hitachi Aroka, Tokyo, Japan) after the mice were anesthetized with an intraperitoneal injection of pentobarbital sodium (50 mg/kg). Contiguous 0.5 mm slice images from the lumbar vertebra were used for analysis by La Theta software, version 1.62. All animals were housed in a pathogen-free barrier facility with a 12-hour light/12-hour dark cycle and were given free access to normal chow and water. Age-matched male littermates were used for all experiments, and animals in this study were harvested at 8 or 16 to 20 weeks old of age. Mice were euthanized during the early light phase after a 24-hour fast.

Metabolic measurements

Glucose, insulin, TG, total cholesterol (T-Cho), and free fatty acid (FFA) levels in plasma and TG and T-Cho levels in the liver were determined, as described previously (22). The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated as the fasting insulin levels (microunits per millileter) × fasting blood glucose levels (milligrams per deciliter)/405 (23). For the oral glucose tolerance test (OGTT), mice were orally administrated d-glucose (1 g/kg body weight) after an overnight fast (16 hours). For the insulin tolerance test (ITT), mice were injected intraperitoneally with regular human insulin (2 U/kg body weight; Eli Lilly, Indianapolis, IN). Blood samples were collected before injection and at specific times after injection (indicated in figures) to determine glucose and insulin levels.

Histological analysis

Livers and adipose tissues were removed; fixed in 10% buffered formalin; embedded in paraffin; cut into 4 or 6 µm-thick sections for liver and adipose tissue, respectively; and used for hematoxylin and eosin (H&E) staining and Masson trichrome (MT) staining, as described previously (24). Immunohistochemical staining for F4/80 (catalog no. ab6640; Abcam, Cambridge, United Kingdom) was also performed, as previously described (25). All images were acquired using a BZ-X710 microscope (Keyence, Osaka, Japan), and data were analyzed using a BZ-H3 analyzer (Keyence, Osaka, Japan). Adipocyte cell size was measured by using ImageJ software (National Institutes of Health, Bethesda, MD) and counted at least 300 cells in each sample.

Isolation of adipocytes and stromal vascular fractions

Epididymal adipose tissue was isolated from 8-week-old ob/ob and Tg-1a;ob/ob mice. The tissues were rinsed in saline, minced into small pieces, and digested for 1 hour at 37°C with collagenase (Sigma-Aldrich, St. Louis MO) in Krebs-Ringer buffer (12 mM HEPES, 121 mM NaCl, 4.9 mM KCl, 1.2 mM MgSO₄, and 0.33 mM CaCl₂, pH 7.4), supplemented with 0.1% glucose and 4% fatty acid-free BSA with gentle shaking. After filtering through a 250-μm nylon mesh, the tissue was centrifuged. The pelleted cells were collected as the stromal vascular fraction (SVF), and the floating cells were collected as the adipocyte fraction. Cells in each fraction were used for RNA extraction. Adipocytes isolated from wild-type (WT) and Tg-1a mice were used for FFA uptake assay using the Free Fatty Acid Uptake Assay Kit (Abcam, Cambridge, United Kingdom), according to the manufacturer’s protocol.

Primary adipocyte cell isolation and culture

The previously mentioned pellets containing the SVFs were filtered through a 40-µm cell strainer and centrifuged. Primary adipocytes were cultured in DMEM (Thermo Fisher Scientific, Waltham, MA), supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. The mouse primary adipocytes were then induced to differentiate into mature adipocytes (26), and after 8 days, the cells were stained with Oil Red O, followed by extraction of the absorbed stain using 100% isopropanol.

RNA extraction and quantitative real-time PCR analysis

Total RNA extraction from liver and WAT, cDNA synthesis, and quantitative real-time PCR were performed, as previously described (22, 25–30). Primer sequences for carbohydrate-responsive element–binding protein (ChREBP)α, ChREBPβ, arginase 1 (Arg1), CD36, C-type lectin domain family 10 member A, glucose transporter (GLUT)4, IL-10, integrin α X (Itgax), lipoprotein lipase (LPL), VLDL receptor (VLDLR), PR domain containing 16 (PRDM16), and cell death–inducing DNA fragmentation factor, α subunitlike effector A (CIDEA) are presented in Supplemental Table 1. mRNA expression levels were normalized to that of cyclophilin mRNA.

Indirect calorimetry

Whole-body oxygen consumption (VO₂) was measured using the Arco-2000 Mass Spectrometer System (Arcosystem, Kashiwa, Japan), as described previously (31). In brief, mice were housed individually in cages and acclimated for 3 days. VO₂ was recorded every 5 minutes for a period of 24 hours at 23°C under a 14-hour light/10-hour dark cycle. Food and water were available ad libitum. Data were calculated by normalizing with body weights.

Statistical analysis

Values were expressed as means ± SEM. Statistics were performed using Student t test (between two groups) or ANOVA (more than two groups), with P < 0.05 considered significant.

Results

Transgenic overexpression of nuclear SREBP-1a in the liver massively enlarges the liver but reduces WAT mass and body weight in ob/ob mice

Tg-1a mice show a massively enlarged liver and atrophic peripheral WAT, as described previously (32). Brown et al. (21) suggested that the decreased WAT mass in the transgenic mice may be a result of the TG steal syndrome in which VLDL-TGs are preferentially cleared into the liver and diverted from adipose tissue. We further extended this hypothesis to a therapeutic standpoint by placing this SREBP-1a overexpression in the ob/ob mice. There was no difference in the body weight between WT and Tg-1a (Fig. 1A) mice. In the ob/ob background, however, the body weight of double-mutant (Tg-1a;ob/ob) mice was dramatically reduced compared with ob/ob mice. Tg-1a;ob/ob mice became thin compared with ob/ob mice, but we observed upper-abdominal extension as a result of hepatomegaly (Fig. 1B). At the age of 12 weeks, Tg-1a;ob/ob mice showed significantly reduced epididymal WAT (eWAT) depots with grossly enlarged livers that weighed 2.4-fold more than the ob/ob livers (Fig. 1C–1E).

Figure 1.

Reduced adiposity and massive liver enlargement in Tg-1a;ob/ob mice. (A) Body weight changes of WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at indicated ages (n = 9 to 14 per group). (B) Photograph of 12-week-old WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice. (C) Photograph of livers and epididymal WATs (eWATs) of 12-week-old WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice. (D) Representative computed tomography images of transverse abdominal sections of 8-week-old WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice. (E) Liver and eWAT weights of WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 to 20 weeks old (n = 9 to 14 per group). Data are represented as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001. BW, body weight.

Transgenic overexpression of nuclear SREBP-1a in the liver decreases plasma levels of FFA and TG in ob/ob mice

Plasma lipid levels were compared among WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice (Fig. 2). All mice were fasted for 24 hours to activate transgene expression. Plasma levels of FFA, TG, and T-Cho were significantly decreased in Tg-1a mice compared with WT mice. Plasma levels of FFA and TG were significantly lower in Tg-1a;ob/ob mice than in ob/ob mice, whereas plasma T-Cho levels were similar between the two genotypes, suggesting that the decreased levels of plasma FFA and TG, rather than plasma T-Cho, are thought to mediate diminished adipose tissue mass by hepatic SREBP-1a activation.

Figure 2.

Plasma lipid levels of Tg-1a;ob/ob mice. Plasma levels of (A) FFA, (B) TG, and (C) T-Cho in WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 20 weeks old. Blood samples were collected from 24-hour fasted mice (n = 16 to 23 per groups). Data are represented as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001.

Severe hepatosteatosis in Tg-1a;ob/ob mice

To examine the effect of the expression of SREBP-1a transgene in livers of ob/ob mice, we compared the histology of liver from the four genotypes. Whereas external appearances indicated hepatosteatosis in the three experimental groups (Fig. 1C), analysis of liver histology confirmed a dramatic increase in lipid droplets in Tg-1a;ob/ob mice compared with Tg-1a and ob/ob mice (Fig. 3A). Hepatic contents of TG tended to be higher, and T-Cho was significantly higher in Tg-1a;ob/ob mice than in ob/ob mice (Fig. 3B and 3C). We evaluated the expression of SREBP-1a target genes, such as synthesis of fatty acid and cholesterol in the liver (Fig. 3D). Transgene expression was significantly lower in Tg-1a;ob/ob livers compared with Tg-1a livers (Fig. 3D, inset). Overexpression of SREBP-1a both in WT and ob/ob mice caused significant increases in the mRNA levels for endogenous SREBP-1a, SREBP-1c, and SREBP-2 and lipogenic enzymes of SREBP targets, such as fatty acid synthase (FAS), Elovl6, stearoyl-coenzyme A (CoA) desaturase-1 (SCD1), and glycerol-3-phosphate acyltransferase 1. Expression levels of ChREBPα and ChREBPβ, the transcription factors responsible for the coordinated induction of glycolytic and lipogenic genes, were significantly elevated in Tg-1a compared with WT mice. The expression levels of GLUT2 are similar in all groups. Expression of genes involved in cholesterol synthesis and uptake, such as low-density lipoprotein (LDL) receptor (LDLR), VLDLR, and hydroxy-3-methylglutaryl-CoA reductase, was significantly elevated in Tg-1a and Tg-1a;ob/ob mice compared with WT and ob/ob mice, respectively. Enhanced levels of these SREBP-1a target genes were generally proportional to the levels of expression of the SREBP-1a transgene. Expression levels of a hepatic steatotic transcription factor peroxisome proliferator–activated receptor γ (PPARγ)2 and a fatty acid transporter CD36 were similarly high in ob/ob mice and in Tg-1a;ob/ob mice, indicating robust TG synthesis by uptake of exogenous fatty acids in both groups. Acetyl-CoA:cholesterol acyltransferase 2 and LPL expression level were dramatically increased in Tg-1a compared with WT and tended to increase in Tg-1a;ob/ob compared with ob/ob mice. These results suggest that marked exacerbation of hepatosteatosis in Tg-1a;ob/ob mice is attributed to both exogenous energy influx by leptin deficiency, in addition to sustained robust hepatic lipogenesis, and cholesterogenesis by SREBP-1a overexpression.

Figure 3.

Severe fatty liver in Tg-1a;ob/ob mice. (A) Representative H&E staining from liver sections of WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 weeks old in the 24-four fasted condition. Scale bars represent 100 µm. (B) Liver TG and (C) T-Cho levels in WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 to 20 weeks old in the 24-hour fasted condition (n = 3 to 12 per group). (D) Real-time PCR analysis to measure various mRNA levels in livers of WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 to 20 weeks old in the 24-hour fasted condition (n = 3 to 12 per group). Results are represented as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001. ACAT2, acetyl-coenzyme A (CoA):cholesterol acyltransferase 2; FAS, fatty acid synthase; GPAT1, glycerol-3-phosphate acyltransferase 1; HMG-CoA, hydroxy-3-methylglutaryl-CoA; LDLR, low-density lipoprotein receptor; PPARγ, peroxisome proliferator–activated receptor γ; SCD1, stearoyl-CoA desaturase-1.

Aberrant morphology of adipose tissue in Tg-1a and Tg-1a;ob/ob mice

Next, we examined the morphologic characteristics of the eWAT in WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice. The eWAT from WT mice had mature adipocytes of uniform size containing unilocular fat droplets (Fig. 4A). The eWAT from Tg-1a mice exhibited adipocytes of various sizes, including scattered, larger adipocytes and also dense regions with smaller adipocyte dispersed among stromal cells, as well as areas with fewer adipocytes, infiltration of immune cells, and extensive fibrosis. Consistent with previously published findings, eWAT from ob/ob mice had large unilocular adipocytes and macrophage infiltration-forming characteristic crownlike structures (33, 34). Importantly, Tg-1a;ob/ob mice had a significantly increased, marked macrophage infiltration and extensive fibrosis in eWAT compared with either Tg-1a or ob/ob mice. Quantification of cell size revealed a wider size distribution, featuring both large and small adipocytes in Tg-1a mice compared with WT mice (Fig. 4B). However, Tg-1a;ob/ob eWAT showed a clear shift to smaller sizes in the distribution of adipocyte fractions compared with ob/ob mice that have distribution of marked, larger adipocytes than WT (Fig. 4C). We performed immunohistochemistry to detect F4/80, a macrophage marker, and MT staining for the detection of collagen fibers (Fig. 4D–4G). We found that F4/80-positive cells from eWAT of Tg-1a and ob/ob mice increased compared with those in WT eWAT, and predictably, there was markedly elevated Tg-1a;ob/ob eWAT compared with other genotypes (Fig. 4D and 4E). We also found that eWAT from Tg-1a, ob/ob, and Tg-1a;ob/ob mice exhibited substantially more interstitial fibrosis than WT tissue; adipose tissue fibrosis was especially prominent in Tg-1a/ob/ob mice (Fig. 4F and 4G). Collectively, these results suggest that transgenic overexpression of SREBP-1a in ob/ob mice causes an enhanced inflammatory phenotype in eWAT, characterized by infiltration of macrophages, tissue remodeling, and extensive fibrosis, representing lipodystrophy.

Figure 4.

Aberrant morphology of WAT in Tg-1a and Tg-1a;ob/ob mice. (A) Representative H&E staining from eWAT of WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 weeks old in the 24-hour fasted condition. Upper and lower panels show low and high magnification, respectively. (B and C) Histograms of adipocyte size from eWAT of (B) WT and Tg-1a mice and (C) ob/ob and Tg-1a;ob/ob mice at 16 to 20 weeks old. (D–G) Representative images and quantification of (D and E) F4/80 immunostaining and (F and G) MT staining of the eWAT sections from WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 weeks old. Scale bars represent 100 µm. Results are represented as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001.

Increased lipogenesis, impaired adipocyte differentiation, reduced fatty acid uptake, and increased inflammation in Tg-1a;ob/ob eWAT

To provide insight into how overexpression of SREBP-1a causes adipose tissue dystrophy, we evaluated the expression of genes for lipogenesis, adipocyte differentiation, inflammation, and fibrosis (Fig. 5). As a result of the control by the PEPCK promoter, the SREBP-1a transgene was expressed at only low levels in eWAT relative to the liver in Tg-1a and Tg-1a;ob/ob mice (Fig. 5A). Expression levels of endogenous SREBP-1a and SREBP-1c were significantly lower in the eWAT of Tg-1a, ob/ob, and Tg-1a;ob/ob mice than in the eWAT of WT mice, whereas the expressions of SREBP target genes, such as FAS and Elovl6 were significantly higher in the eWAT of Tg-1a and Tg-1a;ob/ob mice than in the eWAT of WT and ob/ob mice (Fig. 5B). Expression levels of ChREBPα and ChREBPβ were not higher in the eWAT of Tg-1a and Tg-1a;ob/ob mice than in the eWAT of WT and ob/ob mice. Expression of SCD1 tended to be higher in Tg-1a;ob/ob eWAT than in ob/ob eWAT (Fig. 5B). Expression levels of genes involved in adipocyte differentiation, such as CCAAT/enhancer-binding protein (C/EBP)β, C/EBPα, and PPARγ, were significantly lower in Tg-1a eWAT compared with WT eWAT (Fig. 5C). Expression levels of VLDLR and lipolysis-related genes, such as hormone-sensitive lipase and adipose TG lipase, were significantly lower in the eWAT of Tg-1a and Tg-1a;ob/ob mice compared with WT and ob/ob mice, respectively (Fig. 5C). Expression levels of a macrophage marker F4/80 and classically activated macrophage markers (M1), such as tumor necrosis factor α and Itgax, had a trend to increase in Tg-1a eWAT compared with WT eWAT (Fig. 5D). Consistent with previous studies (4, 35–37), expression levels of these genes were upregulated in ob/ob eWAT. Surprisingly, in Tg-1a/ob/ob eWAT, expression levels of these macrophage and proinflammatory M1 were not increased compared with ob/ob eWAT. Alternatively activated macrophage markers (M2), such as IL-10 and Arg1, tended to be high in Tg-1a eWAT and were similarly increased in both ob/ob eWAT and Tg-1a;ob/ob eWAT compared with WT eWAT (Fig. 5D). The expression levels of Arg1 tend to lower in Tg-1a;ob/ob eWAT than in ob/ob eWAT. As a macrophage polarity marker, the M1 to M2 macrophage ratio (calculated by dividing the relative expression of Itgax by the relative expression of Arg1) was significantly higher in Tg-1a;ob/ob eWAT than in Tg-1a eWAT and ob/ob eWAT (Fig. 5E), suggesting an overall inflammatory shift in Tg-1a;ob/ob adipose tissue macrophages.

Figure 5.

Gene expression in WAT. (A) Real-time PCR analysis to measure mRNA levels of transgene in liver and eWAT, (B) genes for SREBP-1 and ChREBP and lipogenic genes, (C) WAT differentiation and function, (D) inflammation, and (E) M1 to M2 macrophage ratio, calculated by dividing the relative expression of Itgax (M1 marker) by the relative expression of Arg1 (M2 marker) in eWAT from WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 weeks old. The eWAT of each group was collected after 24 hours of fasting (n = 3 to 7 per group). Results are represented as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001. ATGL, adipose triglyceride lipase; HSL, hormone-sensitive lipase; hSREBP-1a, human sterol regulatory element–binding protein-1a; TNFα, tumor necrosis factor α.

Evidence has accumulated indicating that unbalanced production of pro- and anti-inflammatory cytokines contributes to the development of adipose tissue dysfunction (36–38). Moreover, overexpression of SREBP-1a in adipose tissue caused adipocyte hypertrophy associated with inflammatory cytokine production (39). We then examined the expression of genes for adipocyte function and M1 and M2 macrophage markers in SVF and adipocytes from eWATs of ob/ob and Tg-1a;ob/ob mice (Fig. 6). Collagenase digestion of the eWAT from Tg-1a;ob/ob mice revealed that the human SREBP-1a transgene was expressed predominantly in mature adipocytes (Fig. 6A). Adiponectin, an adipocytokine, plays an important role in the regulation of glucose and lipid metabolism (40, 41), and PPARγ2, a key regulator of adipocyte differentiation (42, 43), was predominantly expressed in adipocyte, and expression levels of these genes were dramatically reduced in Tg-1a;ob/ob adipocytes compared with ob/ob adipocytes (Fig. 6B and 6C). The macrophage marker F4/80 and the proinflammatory M1 macrophage marker Itgax were expressed mainly in SVF (Fig. 6D and 6E). Expression levels of these genes were significantly higher in both SVF and adipocytes from Tg-1a;ob/ob mice compared with those from ob/ob mice. The M2 Arg1 was predominantly expressed in SVF, and the expression was significantly reduced in Tg-1a SVF compared with ob/ob SVF (Fig. 6E and 6F). In addition, the M1 to M2 macrophage ratio was significantly higher in Tg-1a;ob/ob SVF than in ob/ob SVF (Fig. 6G), suggesting an inflammatory shift in Tg-1a;ob/ob adipose tissue macrophages.

Figure 6.

Gene expression levels in mature adipocytes and SVF obtained from eWAT of ob/ob and Tg-1a;ob/ob mice. (A) Relative mRNA levels for the SREBP-1a transgene in SVF and adipocytes of eWATs from 8-week-old Tg-1a;ob/ob mice. (B–F) mRNA expression of mature adipocyte markers. (B) Adiponectin and (C) PPARγ and (D–F) SVF in the eWAT from ob/ob and Tg-1a;ob/ob mice at 8 weeks old (n = 3 to 4 per group). (G) M1 to M2 macrophage ratio, calculated by dividing the relative expression of Itgax (M1 marker) by the relative expression of Arg1 (M2 marker). (H) Representative images and (I) quantification of Oil Red O staining of differentiated adipocytes. Primary SVF cells isolated from the eWAT of 8-week-old WT and Tg-1a mice were differentiated under adipogenic conditions for 8 days. Scale bars represent 100 µm. (J) Relative mRNA levels for LPL in SVF and adipocytes of eWATs from 8-week-old Tg-1a;ob/ob mice (n = 3 to 5 per group). (K) FFA incorporation of isolated adipocytes from WT and Tg-1a mice (n = 5). Results are represented as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001. RFU, relative fluorescence unit.

Moreover, we tested the capacity of SVF cells, a source of preadipocytes isolated from eWAT of 8-week-old WT and Tg-1a mice, to differentiate to adipocytes in vitro (Fig. 6H and 6I). As assessed by Oil Red O staining, Tg-1a adipocytes had significantly smaller lipid droplets relative to WT adipocytes at day 8, when cells were fully differentiated, suggesting that overexpression of SREBP-1a under the control of the PEPCK promoter inhibits adipogenesis.

The lower ability of fatty acid uptake of adipocyte could cause WAT atrophy (44). A major portion of available fatty acids for adipocyte uptake is derived from LPL-mediated hydrolysis of plasma TGs (45, 46). Therefore, we examined the expression of LPL in SVF and adipocytes from eWATs of ob/ob and Tg-1a;ob/ob mice and found that LPL was predominantly expressed in adipocytes, and the expression level of LPL was significantly reduced in Tg-1a;ob/ob adipocytes compared with ob/ob adipocytes (Fig. 6J). Moreover, FFA incorporation activity was significantly lower in Tg-1a adipocytes than WT adipocytes (Fig. 6K). These results suggested that overexpression of SREBP-1a inhibits fatty acid uptake in WAT.

Collectively, these data suggest that the overexpression of SREBP-1a in adipocyte activates lipogenesis, inhibits adipocyte differentiation and fatty acid uptake, and causes unbalanced production of pro- and anti-inflammatory adipocytokines and chemokines in adipose tissue. Moreover, 5-bromo-2′-deoxyuridine staining strongly suggests that the SVF fraction from Tg-1a eWAT, especially Tg-1a;ob/ob eWAT, contains rapidly proliferating cells (Supplemental Fig. 1). The cell types of these proliferating cells, relationship of these cells, smaller cell-size population, screwing up of adipogenesis, and M1 and M2 composition are currently unknown and require future study.

To determine the contribution of the hepatic overexpression of SREBP-1a in the lipodystrophic phenotype in Tg-1a and Tg-1a;ob/ob mice, we administered an adenovirus carrying human nuclear SREBP-1a or green fluorescent protein to C57BL/6J mice. Human nuclear SREBP-1 mRNA was expressed only in the livers of Ad-BP1a mice (Supplemental Fig. 2A). Mice with adenovirus carrying human nuclear SREBP-1a exhibited significantly increased hepatic lipogenic gene expression (Supplemental Fig. 2B), liver weight (Supplemental Fig. 2C), and hepatic TG and T-Cho content (Supplemental Fig. 2D). In contrast, hepatic SREBP-1a overexpression significantly reduced plasma TG, FFA, and T-Cho levels (Supplemental Fig. 2E). Importantly, hepatic SREBP-1a overexpression significantly reduced eWAT weight (Supplemental Fig. 2F) without increasing the inflammatory response (Supplemental Fig. 2G); mRNA expression levels of genes involved in lipogenesis, adipocyte differentiation, and lipid uptake (Supplemental Fig. 2H); and food intake (Supplemental Fig. 2I). Thus, hepatic overexpression of SREBP-1a was sufficient to reduce WAT mass.

Reduced food intake and increased energy expenditure in Tg-1a;ob/ob mice

As appetite and energy expenditure are tightly linked to the control of adiposity (47, 48), we examined food intake and VO₂ (Fig. 7A–7C). Food intake was significantly lower in Tg-1a and Tg-1a;ob/ob mice compared with WT and ob/ob mice, respectively (Fig. 7A). There was no difference in VO₂ between WT and Tg-1a mice (Fig. 7B and 7C). In the ob/ob background, VO₂, which was decreased compared with ob^+/+ background, was significantly increased in Tg-1a mice compared with WT mice. These results indicated that the reduced food intake and increased energy expenditure could partially explain the observed leanness of Tg-1a and Tg-1a;ob/ob mice. A growing body of evidence from both animal and human studies suggests that activation of brown and beige adipocytes resulted in reduction of body weight and fat mass and improvement of glucose metabolism, including insulin sensitivity (49–51). The Tg-1a brown adipose tissue (BAT) had markedly enlarged and unilocular adipocytes in both WT and ob/ob backgrounds, probably caused by accumulation of lipids (Fig. 7D). We also observed that the expression of thermogenic genes, such as uncoupling protein 1, PRDM16, and CIDEA, was significantly reduced or unchanged in Tg-1a BATs from both WT and ob/ob backgrounds (Fig. 7E). Moreover, in contrast to what would be expected, Tg-1a inguinal WAT (iWAT) did not enhance beiging (Fig. 7F) and upregulate thermogenic genes in both WT and ob/ob backgrounds (Fig. 7G). Taken together, our data demonstrated that an increased energy expenditure in Tg-1a;ob/ob mice compared with ob/ob mice was not a result of increased thermogenesis in BAT and iWAT.

Figure 7.

Reduced food intake and enhanced energy expenditure in Tg-1a;ob/ob mice. (A) Food intake per day in WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 to 20 weeks old. (B) VO₂ in WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice during 24-hour light/24-hour dark cycles. (C) The area under the curve (AUC) during the (B) VO₂ measurements. (D) Representative H&E staining from brown adipose tissue (BAT) sections of WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 weeks old in the 24-hour fasted condition. Scale bars represent 50 µm. (E) Real-time PCR analysis to measure mRNA levels of genes for thermogenesis in BATs of WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 to 20 weeks old in the 24-hour fasted condition (n = 3 to 12 per group). Results are represented as means ± SEM. *P < 0.05; ***P < 0.001. (F) Representative H&E staining from inguinal WAT (iWAT) sections of WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 weeks old in the 24-hour fasted condition. Scale bars represent 100 µm. (G) Real-time PCR analysis to measure mRNA levels of genes for thermogenesis in iWATs of WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 to 20 weeks old in the 24-hour fasted condition (n = 3 to 12 per group). Results are represented as means ± SEM. *P < 0.05; ***P < 0.001. UCP1, uncoupling protein 1.

Severe insulin resistance in Tg-1a;ob/ob mice

Finally, we investigated the plasma glucose metabolism, as lipodystrophy in either human or mouse models is associated with severe metabolic disturbances, including hyperglycemia, hyperinsulinemia, glucose intolerance, and insulin resistance (12). As we reported previously (32), fasting plasma glucose levels were significantly higher in Tg-1a mice than in WT mice (Fig. 8A). Fasting plasma glucose levels of Tg-1a;ob/ob mice were similar to those of ob/ob mice. Fasting plasma insulin levels were significantly higher in Tg-1a;ob/ob mice than in ob/ob mice (Fig. 8B). The HOMA-IR was robustly increased in Tg-1a;ob/ob, ob/ob, and Tg-1a mice in this order of magnitude, indicating the presence of severe peripheral insulin resistance in Tg-1a;ob/ob mice (Fig. 8C). We next performed OGTT and found that plasma glucose concentrations after oral glucose loading were significantly lower in Tg-1a;ob/ob mice throughout the test (Fig. 8D). Insulin levels in Tg-1a;ob/ob mice were rapidly elevated following glucose loading to levels higher than ob/ob mice (Fig. 8E). ITT demonstrated impaired glucose removal and severe insulin resistance in Tg-1a;ob/ob mice compared with ob/ob mice (Fig. 8F). The size of pancreatic islets increased similarly in both ob/ob and Tg-1a;ob/ob mice (Fig. 8G). These results demonstrated that Tg-1a;ob/ob mice exhibit severe insulin resistance despite reduced adipose tissue mass compared with ob/ob mice.

Figure 8.

Overexpression of the nuclear form of SREBP-1a in the liver causes severe insulin resistance in ob/ob mice. Plasma levels of (A) glucose, (B) insulin, and (C) HOMA-IR in WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 to 20 weeks old of age. Blood samples were collected after 24 hours of fasting. (D) Plasma glucose and (E) insulin levels during OGTT in ob/ob and Tg-1a;ob/ob mice (n = 6 to 8 per group). (F) ITT in ob/ob and Tg-1a;ob/ob mice (n = 3 per group). (G) Representative H&E staining from pancreas sections of WT, Tg-1a, ob/ob, and Tg-1a;ob/ob mice at 16 weeks old. Scale bars represent 100 µm. Results are represented as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001.

Discussion

The current data provide an explanation for the previously unexpected observation that Tg-1a mice showed atrophy within peripheral WAT (21). These data suggest that hepatic SREBP-1a markedly affects the relationship between TG accumulation in the liver and lipids delivered from the adipose tissue. Overexpression of SREBP-1a under the control of the PEPCK promoter resulted in progressive and age-related lipodystrophy in mice that was markedly accelerated when Tg-1a mice were on the ob/ob background, resulting in increased inflammation and fibrosis in adipose tissue and severe insulin resistance. Furthermore, impaired adipocyte differentiation, reduced adipocyte fatty acid uptake capacity, reduced food intake, and increased energy expenditure observed in Tg-1a and Tg-1a;ob/ob mice could contribute to the leanness of these animals.

One of the reasons for the decrease in adipose tissue mass in the transgenic mice may be secondary to the accumulation of TGs in the liver. Because of SREBP-1a overexpression, the rate of synthesis of fatty acids and cholesterol is markedly increased in the liver, and plasma glucose is preferentially used for hepatic lipogenesis. Moreover, the cooperative overexpression of CD36, LDLR, VLDLR, and LPL in the transgenic liver may create a more progressive “fat steal” syndrome in which FFA and VLDL-TGs in the blood are preferentially incorporated into the liver and diverted from WAT. This explanation is supported by the finding that adenoviral overexpression of SREBP-1a in the liver resulted in enhanced lipogenesis and lipid uptake in the liver, reduced plasma lipid levels, and reduced adipose tissue mass without reducing food intake (Supplemental Fig. 2). It is intriguing to postulate, as another mechanism of fat steal syndrome, that overexpression of SREBP-1a in the liver might lead to the secretion of some unknown hepatic gene product (hepatokine), which would reduce adipose tissue mass or increase adipose tissue inflammation, leading to the incidence of lipodystrophy.

The precise molecular mechanism that counters reduction of body weight by overproduction of SREBP-1a in mice is currently unknown as a result of the limitation of the current study using PEPCK promoter Tg mice. Tg-1a did not change body weight, in combination with reduced adiposity and hepatomegaly, reflecting a fat steal syndrome or energy shift between organs. The energy derived from increased flux might be catabolized, as well as converted into hepatosteatosis and hepatomegaly.

The second possibility is that overexpression of SREBP-1a in adipocytes potentially inhibits their differentiation. We found that overexpression of SREBP-1a under the control of the PEPCK promoters inhibits adipogenesis, as assessed by the capacity of SVF cells, isolated from eWAT of WT and Tg-1a mice, to differentiate to adipocyte in vitro (Fig. 6H and 6I). It is possible that overproduction of SREBP-1a may interfere with activity of other transcription factors in adipocytes, although residual, mature adipocytes might be supported by enhanced lipogenesis. This finding is strikingly different from the previously generated adipocyte protein 2 (aP2)–SREBP-1a Tg mice that developed fully differentiated and markedly enlarged adipocytes (39). The dramatic differences in the phenotypes obtained in this and previous studies are likely a result of stage-specific expression and/or strength of the PEPCK and aP2 promoters. Further studies will be needed to elucidate the mechanisms of adipocyte or hepatic SREBP-1a activation that potentially affect adipocyte differentiation. Currently, we speculate that both liver and fat pathologies contributed to body weight reduction of Tg-1a;ob/ob mice.

Surprisingly, food intake was slightly but significantly reduced in Tg-1a and Tg-1a;ob/ob mice compared with WT and ob/ob mice, respectively (Fig. 7A). The mechanism of appetite reduction, caused by SREBP-1a overexpression, is unknown. It is well known that hormonal, neural, and metabolic signals regulate food intake (52). Because body adiposity can affect these signals, reduced food intake could be caused by the changes of adipokines and/or metabolites in mice overexpressing SREBP-1a. However, it was not the major cause of lipodystrophy, as adenoviral overexpression of SREBP-1a in the liver significantly reduced eWAT weight without reducing food intake in mice.

In addition to reduced food intake, Tg-1a;ob/ob mice exhibited the increased energy expenditure compared with ob/ob mice (Fig. 7B and 7C). The mechanism of enhanced energy expenditure, caused by SREBP-1a overexpression, is unknown. Two types of thermogenic fat cells, brown adipocytes and beige adipocytes, play a key role in the regulation of systemic energy homeostasis in mammals (50, 51). The increase of the thermogenic capacity of these adipocytes could contribute to enhance energy expenditure and protect mice against the development of obesity (50, 53, 54). However, the thermogenic function of BAT and iWAT did not increase, but rather decreased, in Tg-1a;ob/ob mice. Thus, the mechanisms that enhance energy expenditure in Tg-1a;ob/ob mice will require further study to fully elucidate.

Enhanced inflammation is a common feature of partial lipodystrophy (55–57). The decreased adipose tissue mass was most likely the result of adipocyte death in Tg-1a and Tg-1a;ob/ob mice, as evidenced by the decrease of adipocyte markers and by the presence of crownlike structures in their adipose tissue. PPARγ, a master regulator of adipogenesis, was decreased in Tg-1a and Tg-1a;ob/ob adipocytes. There are some reports that describe the relationship between PPARγ and lipodystrophy. Dominant-negative missense mutations may cause lipodystrophy by affecting adipogenesis (58). Moreover, PPARγ deficiency in adipose tissue results in severe lipodystrophy and hyperlipidemia (59). Therefore, it is possible that decreased PPARγ expression may cause lipodystrophy in Tg-1a and Tg-1a;ob/ob. The death and degeneration of adipocytes are believed to provoke macrophage infiltration, which helps clear lipids and adipocyte debris and allows for extracellular matrix remodeling (60, 61). In line with this, extensive fibrosis was observed in the eWAT of Tg-1a;ob/ob mice. Generally, the controlled balance between M1 and M2 macrophages is critical for the progression and resolution of inflammation (62–64). We observed that M1 macrophage marker Itgax expression was increased, but M2 macrophage marker Arg1 expression was reduced in the SVF fraction of Tg-1a;ob/ob mice, compared with the SVF fraction of ob/ob mice, suggesting that overexpression of SREBP-1a shifts the macrophage balance from M2 (Arg1⁺) to M1 (Itgax⁺) in ob/ob mice with increased macrophage numbers. As unrestrained inflammation invariably results in sustained tissue damage, this shift in eWAT macrophage phenotype may contribute to inflammation and fibrosis exacerbation in adipose tissue and systemic insulin resistance.

Our current study also suggests the potential involvement of SREBP-1a in fibrosis. Leptin and LDLR-mediated LDL cholesterol have also been suggested to be profibrotic factors (65, 66). However, this phenotype could be secondary to altered lipid metabolism and inflammation and requires further investigation. Collectively, mice overexpressing the nuclear form of SREBP-1a exacerbated inflammation and fibrosis, but there was no agreement between histology and mRNA expression levels.

Our data demonstrate that Tg-1a;ob/ob mice exhibit severe insulin resistance but have a larger capacity for glucose tolerance, presumably as a result of a large glucose disposal through the insulin-independent pathway into the larger liver mass and increased glucose incorporation into lipids via the activity of SREBP-1a. Liver weight in Tg-1a;ob/ob mice is twofold higher than in the ob/ob mice, suggesting that total glucose uptake by the liver should be increased in Tg-1a;ob/ob to a similar extent.

In conclusion, we propose a lipodystrophy mouse model via overexpression of SREBP-1a. Genes in adipogenesis, lipid metabolism, and caveolae formation have been identified as causing congenital-generalized and familial-partial lipodystrophy in humans (12). Our findings suggest that conditions that reduce the uptake of plasma lipids in adipocytes may also be potential causes of congenital lipodystrophy. Further study is required to understand the molecular mechanism by which SREBP-1a activation causes lipodystrophy. With the content of better understanding, the mechanism of SREBP-mediated metabolic pathway alteration will provide a therapeutic strategy in combating lipodystrophy.

Abbreviations

Abbreviations
 
• aP2

  adipocyte protein 2

 
• Arg1

  arginase 1

 
• BAT

  brown adipose tissue

 
• C/EBP

  CCAAT/enhancer-binding protein

 
• ChREBP

  carbohydrate-responsive element–binding protein

 
• CIDEA

  cell death–inducing DNA fragmentation factor, α subunitlike effector A

 
• CoA

  coenzyme A

 
• eWAT

  epididymal white adipose tissue

 
• FAS

  fatty acid synthase

 
• FFA

  free fatty acid

 
• GLUT

  glucose transporter

 
• H&E

  hematoxylin and eosin

 
• HOMA-IR

  homeostasis model assessment of insulin resistance

 
• Itgax

  integrin α X

 
• ITT

  insulin tolerance test

 
• iWAT

  inguinal white adipose tissue

 
• LDL

  low-density lipoprotein

 
• LDLR

  low-density lipoprotein receptor

 
• LPL

  lipoprotein lipase

 
• M1

  classically activated macrophage marker

 
• M2

  alternatively activated macrophage marker

 
• MT

  Masson trichrome

 
• ob/ob

  obese

 
• OGTT

  oral glucose tolerance test

 
• PEPCK

  phosphoenolpyruvate carboxykinase

 
• PPARγ

  peroxisome proliferator–activated receptor γ

 
• PRDM16

  positive regulatory domain containing 16

 
• SCD1

  stearoyl–coenzyme A desaturase-1

 
• SREBP

  sterol regulatory element–binding protein

 
• SVF

  stromal vascular fraction

 
• T-Cho

  total cholesterol

 
• TG

  triglyceride

 
• Tg-1a

  transgenic mice that overexpress the nuclear form of sterol regulatory element-binding protein-1a under the control of the phosphoenolpyruvate carboxykinase promoter

 
• Tg-1a
 
• ob/ob

  crossed transgenic-1a and obese mice

 
• TgSREBP-1a

  transgenic mice that overexpress a truncated NH2-terminal segment of human sterol regulatory element-binding protein-1a

 
• VLDL

  very low–density lipoprotein

 
• VLDLR

  very low–density lipoprotein receptor

 
• VO₂

  oxygen consumption

 
• WAT

  white adipose tissue

 
• WT

  wild-type

Acknowledgments

The authors are grateful to Professor Alyssa H. Hasty for critical reading of this manuscript. The authors also thank Katsuko Okubo, Yuko Tamai, and Chizuko Fukui for technical assistance and the members of our laboratories for discussion and helpful comments on the manuscript. The authors thank Enago for the English language review.

Financial Support: This work was supported by Grants-in-Aid for Scientific Research 15H02541 (to H. Shimano), 17H06395 (to H. Shimano), 15H03093 (to T.M.) and Program to Disseminate Tenure Tracking System (to T.M.) from the Ministry of Science, Education, Culture, and Technology of Japan.

Disclosure Summary: The authors have nothing to disclose.