Peroxisome proliferator-activated receptor-γ (PPARγ) is considered to be one of the master regulators of adipocyte differentiation. PPARγ2 is abundantly expressed in mature adipocytes and is elevated in the livers of animals that develop fatty livers. The aim of this study was to determine the ability of PPARγ2 to induce lipid accumulation in hepatocytes and to delineate molecular mechanisms driving this process. The hepatic cell line AML-12 was used to generate a cell line stably expressing PPARγ2. Oil Red O staining revealed that PPARγ2 induces lipid accumulation in hepatocytes. This phenotype is accompanied by a selective upregulation of several adipogenic and lipogenic genes including adipose differentiation-related protein (ADRP), adipocyte fatty acid-binding protein 4, sterol regulatory element-binding protein-1 (SREBP-1), fatty acid synthase (FAS), and acetyl-CoA carboxylase, genes whose expression levels are known to increase in steatotic livers of ob/ob mice. Furthermore, the PPARγ2-regulated induction of both SREBP-1 and FAS parallels an increase in de novo triacylglycerol synthesis in hepatocytes. Triacylglycerol synthesis and lipid accumulation are further enhanced by culturing hepatocytes with troglitazone in the absence of exogenous lipids. These results correspond with an increase in the lipid droplet protein, ADRP, and the data demonstrate that ADRP functions to coat lipid droplets in hepatocytes as observed by confocal microscopy. Taken together, these observations propose a role for PPARγ2 as an inducer of steatosis in hepatocytes and suggest that this phenomenon occurs through an induction of pathways regulating de novo lipid synthesis.
- adipose differentiation-related protein
- lipid droplets
- nonalcoholic fatty liver disease
- peroxisome proliferator-activated receptor-γ2
obesity is quickly becoming a global epidemic in both industrialized and developing nations. Obesity, insulin resistance, and the onset of type 2 diabetes mellitus are now hypothesized to impact lipid accumulation and cellular dysfunction in a variety of organs (46, 47). The liver is an example of this, and there is a direct correlation between the degree of obesity and the prevalence and severity of nonalcoholic fatty liver disease (NAFLD) (2, 7). The pathology of NAFLD is similar to alcohol-induced liver injury and results in an excessive accumulation of lipids in the liver but occurs in patients who deny alcohol abuse (3). Therefore, it is of interest to determine whether transcription factors directing adipogenesis are also regulating lipid accumulation in hepatocytes.
Several murine models of obesity and diabetes, including ob/ob, A-ZIP, aP2/DTA, and KKAy, develop fatty livers that express enhanced levels of peroxisome proliferator-activated receptor-γ (PPARγ) (6, 14, 15, 20, 30). PPARγ has two major isoforms, γ1 and γ2, generated from the same gene by alternative splicing (18, 48). PPARγ2 is highly expressed in adipose tissue and is upregulated in steatotic livers of ob/ob mice, whereas PPARγ1 is found at low levels in many tissues (17, 32, 44, 48). Often referred to as the “master regulator” of adipogenesis, PPARγ participates in the transcriptional activation of numerous adipogenic and lipogenic genes important for adipocyte maturation, lipid accumulation, and insulin-sensitive glucose transport, including aP2, CCAAT/enhancer-binding protein-α (C/EBPα), perilipin, and GLUT4 (4, 31, 33, 34, 45, 51, 52).
A number of studies have demonstrated enhanced expression of lipogenic genes and increased expression of PPARγ in animal models of steatotic liver (20, 30, 55). Moreover, a role for PPARγ has been established in the maintenance of a steatotic phenotype in the liver (20, 30). The molecular mechanism by which PPARγ induces steatosis has not been thoroughly examined. To determine how PPARγ induces a fatty phenotype in hepatocytes, a hepatic stable cell line expressing PPARγ2 was created. The results of this study reveal that PPARγ2 induces lipid accumulation in hepatocytes. The data also reveal that PPARγ2 selectively upregulates adipogenic and lipogenic gene expression. Culturing of PPARγ2-expressing hepatocytes in the absence of serum (exogenous lipids) results in lipid accumulation, suggesting that de novo lipid synthesis may be an important mechanism contributing to steatosis in hepatocytes. A PPARγ2-regulated increase in de novo lipid biosynthesis was confirmed by studies examining incorporation of [14C]acetate in triacylglycerol molecules. This phenomenon was enhanced in cells cultured in the absence of exogenous lipids alone or in combination with troglitazone. It is of interest that the lipid droplet-associated protein [adipose differentiation-related protein (ADRP)] was upregulated by PPARγ2 and functions to uniformly coat droplets in hepatocytes while perilipin expression was not detected at the level of mRNA or protein. These data also demonstrate that ADRP mRNA and protein expression were enhanced in an in vivo model of steatotic liver, the ob/ob mouse, while perilipin expression was detected primarily in adipose tissue. Taken together, this body of work demonstrates a role for PPARγ2 as a regulator of steatosis in hepatocytes and suggests that PPARγ2 is inducing this process by enhancing hepatic lipogenesis. Moreover, the data show that, while PPARγ2 is capable of selectively upregulating several adipogenic genes, it does not appear to induce a transdifferentation of hepatocytes into adipogenic cells.
MATERIALS AND METHODS
Dexamethasone (DEX), 3-isobutyl-1-methylxanthine, insulin, and insulin, transferrin, and selenium (ITS) liquid supplement were purchased from Sigma (St. Louis, MO). Leupeptin, aprotinin, and puromycin were purchased from American Bioanalytical (Natick, MA), while DMEM and DMEM-F-12 as well as calf serum and TRIzol were purchased from Invitrogen (Carlsbad, CA). FBS was obtained from Hyclone (Logan, UT), and [14C]acetate was obtained from PerkinElmer Life Sciences (Boston, MA). Troglitazone was a gift from Johnson and Johnson (Parke-Davis/Warner Lambert, Ann Arbor, MI). PPARγ2pBABE-Puro and pBABE-Puro retroviral expression vectors were kind gifts of Dr. B. Spiegelman (Dana Farber Cancer Institute, Harvard Medical School, Boston, MA).
Monoclonal anti-PPARγ antibody and polyclonal C/EBPα were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), polyclonal aP2 serum (kindly provided by Dr. D. Bernlohr, University of Minnesota), polyclonal ADRP (anti-rabbit) and perilipin (anti-goat) antibodies (kindly provided by Dr. C. Londos, National Institutes of Health, Bethesda, MD), and perilipin (anti-rabbit) antibody was kindly provided by Dr. A. Greenberg (Tufts University, Boston, MA). Polyclonal anti-fatty acid synthase (FAS) antibody was obtained from Novus Biologicals (Littleton, CO), polyclonal anti-acetyl-CoA carboxylase (ACC) was purchased from Upstate (Lake Placid, NY), polyclonal anti-ACRP30 (adiponectin) was obtained from Affinity BioReagents (Golden, CO), and FITC-conjugated donkey anti-rabbit and Cy3-conjugated donkey anti-goat were obtained from Jackson ImmunoResearch (West Grove, PA).
AML-12 hepatocytes were obtained from Dr. N. Fausto (University of Washington, Seattle, WA) and cultured as described previously (49). Briefly, cells were plated and grown for 2 days postconfluence in DMEM-F-12, 1× ITS, 100 nM DEX, 2.5 mg/ml puromycin, and 10% FBS, at which point (day 0) cells were maintained in the presence or absence of 10% FBS and 5 μM troglitazone. Murine 3T3-L1 preadipocytes were cultured, maintained, and differentiated as described previously (41, 50). Briefly, cells were plated and grown for 2 days postconfluence in DMEM supplemented with 10% calf serum. Differentiation was then induced (day 0) by changing the medium to DMEM containing 10% FBS, 0.5 mM 3-isobutyl-1-methylxanthine, 1 μM DEX, and 1.67 μM insulin. After 48 h, cells were maintained in DMEM containing 10% FBS. Human embryonic kidney (HEK)-293 cells were cultured in DMEM with 10% FBS.
Studies were approved by Boston University School of Medicine Institutional Animal Care and Use Committee. Eight-week-old ob/ob male mice and age- and sex-matched controls (C57BL/6J mice) were purchased from Jackson Labs (Bar Harbor, ME). Mice were fed a standard chow diet ad libitum and kept on a 12:12-h light-dark cycle for 4 wk. Animals were euthanized at 12 wk of age. Liver and fat tissues were harvested from the animals for total RNA and protein analysis.
Plasmids and retroviral transfection/infection materials.
Retroviral transfections and infections were carried out using VSV-G and GAG-POL vectors (Stratagene) and FUGENE 6 (Roche, Indianapolis, IN), according to the manufacturer's instructions with minor modifications. Briefly, HEK-293T cells were plated in DMEM with 10% FBS. Cells were grown to ∼75% confluence in a 60-mm-diameter dish. Individual cultures of cells were transfected with Fugene 6 and 2 μg of either pBABE-puro or PPARγ2-pBABE, along with 2 μg each of vesicular stomatitis virus G and GAG-POL expression plasmids (pVpack) with a ratio of FUGENE to DNA of 6 μl:1 μg DNA in accordance with standard methods from the FUGENE 6 protocol. Twenty-four hours after transfection, media were removed and fresh DMEM with 10% FBS was added to the cells. HEK-293 medium containing high-titer retrovirus was collected after 24 h and filtered through a 0.45-μm low-protein binding filter. The viral filtrate supernatant was used to infect target cells that were seeded in a 60-mm dish at 50% confluence. Cells were infected with retrovirus in the presence of 10 μg/ml polybrene, and the infection was allowed to proceed for an additional 24 h. At this stage, cells were changed into fresh growth medium for a period of 72 h. After 3 days, 2.5 μg/ml of puromycin were added to the medium to facilitate the selection of stable cell lines, which requires 2–3 wk in the presence of an antibiotic.
Total RNA was extracted from animal tissue, hepatocyte, and fibroblast cell lines with TRIzol, according to the manufacturer's instructions. After quantification, mRNA was assayed in equivalent amounts of total RNA by RT-PCR as described previously (25). All primers were synthesized from Integrated DNA Technologies (Coralville, IA) based on published sequences in the gene banks. Primer sequences, annealing temperatures, cycle numbers, and product sizes are listed in Table 1. All PCRs were performed in the linear range of cycle number for each set of primers, and the corresponding products were analyzed by 1.0% (wt/vol) agarose gel electrophoresis and revealed with ethidium bromide.
Protein analysis, gel electrophoresis, and immunoblotting.
Cells were washed three times with cold PBS and scraped in Western lysis buffer (300 μl/60-mm plate or 500 μl/100-mm plate) consisting of the following: 20 mM Tris·HCl, pH 7.4, 150 mM NaCl, 10% glycerol, 2% Nonidet P-40, 1 mM EDTA, pH 8.0, 20 mM sodium fluoride, 30 mM sodium pyrophosphate, 0.2% sodium dodecyl sulfate, 0.5% sodium deoxycholate, 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, 1 mM sodium vanadate, 50 μM leupeptin, and 5 μM aprotinin. Samples were incubated on ice with frequent vortexing for 15 min and centrifuged for 20 min at 18,000 g. Protein content of each supernatant was quantified using a bicinchoninic acid (BCA) kit (Pierce, Rockford, IL). Sixty micrograms of each supernatant sample of proteins were separated by electrophoresis through a 12% polyacrylamide gel (ACC and FAS 6% gel) and transferred to 0.45 μM Immobilin-P polyvinylidene difluoride membrane (PerkinElmer Life Sciences). After transfer, membranes were blocked in 5% (wt/vol) nonfat dry milk in PBS-0.1% Tween 20 and probed with the antibodies specified. Horseradish peroxidase-conjugated secondary antibodies (Sigma) and enhanced chemiluminescence (ECL) substrate kit (PerkinElmer Life Science) were used in the detection of specific proteins.
Oil Red O staining.
Oil Red O staining was performed following the procedure described previously, with minor modifications (52). In brief, cells were washed twice with PBS and fixed with 7.5% formaldehyde in PBS for 15 min. After two washes in PBS, cells were stained for at least 1 h in freshly diluted Oil Red O solution (6 parts Oil Red O stock solution and 4 parts H2O; Oil Red O stock solution is 0.5% Oil Red O in isopropanol). The stain was then removed, and the cells were washed twice with water. Images of cells stained with Oil Red O were obtained with a Canon Powershot SD300 digital camera. Cells were also photographed using a phase-contrast Olympus IX70 microscope in combination with a MicroFire digital camera (Image Processing Solutions, North Reading, MA) at ×100 and ×200 magnifications.
Immunofluorescent staining and confocal microscopy.
For ADRP and perilipin staining, cells were fixed in 4% paraformaldehyde for 30 min at room temperature. After a brief rinse with PBS (pH 7.4), cells were permeablilized with 0.2% Triton X-100 for 5 min and blocked with donkey IgG (Sigma) for 60 min. Cells were incubated with primary antibodies, including rabbit anti-murine ADRP (1:100) and goat anti-murine perilipin (1:100), for 2 h at room temperature. After three washes in PBS, cells were incubated for 1 h with appropriate secondary antibodies, including FITC-conjugated donkey anti-rabbit (1:100) and Cy3-conjugated donkey anti-goat (1:100). Cellular fluorescent staining profiles were examined with a Zeiss 510 confocal laser-scanning microscope (Carl Ziess, Thornwood, NJ). Coverslips stained with only the respective secondary antibodies were used to establish the fluorescence exposure settings, and all experimental images were analyzed at the same voltage and aperture settings as used for controls.
Incorporation of [14C]acetate into lipid and measurement of total triglyceride accumulation.
Incorporation of [14C]acetate into lipid was performed as described previously, with minor modifications (35, 36). Briefly, cells were plated and grown to confluence on 24-well tissue culture-treated dishes. Two days postconfluence, media were changed and cells were maintained in the presence and absence of 10% FBS and 5 μM troglitazone. Twenty-four hours before harvest, sodium [1,2-14C]acetate (55.3 mCi/mmol) was added to each well at a concentration of 0.5 μCi/ml of medium. At the time of harvest, the media were removed, and the cells were washed three times with cold 1× PBS. Lipids were extracted with hexane-propan-2-ol (3:2, vol/vol), and de novo triacylglycerol synthesis was evaluated by drying the extracts under a stream of nitrogen, redissolving them in toluene, and subjecting them to thin-layer chromatography. The plates were developed in hexane-diethyl ether-acetic acid (70:30:1, by vol). Staining with iodine was used to reveal the migration of standards, which allowed counting of triacylglycerols using an Instant Imager Electronic Audioradiographer (PerkinElmer) and Instant Imager Electronic Autoradiography software (Packard, Meriden, CT). In addition, cell layer protein was solubilized with 0.2 N NaOH and quantified using a BCA kit. Results are expressed as total counts per minute of [14C]acetate in triacylglycerol per microgram of protein.
Constitutive expression of PPARγ2 in hepatocytes induces lipid accumulation.
To gain further insight into the role of PPARγ2 as a regulator of lipid accumulation in the liver, PPARγ2 was stably introduced into an immortalized hepatocyte cell line, the AML-12 hepatocytes. This cell line was formed from a transgenic mouse overexpressing human transforming growth factor-α, a potent hepatocyte mitogen. The AML-12 hepatocytes grow indefinitely in culture, are non-tumorigenic, and maintain expression of liver-specific genes (49). To determine the ability of PPARγ2 to induce lipid accumulation in hepatocytes, cells expressing PPARγ2 or vector alone were cultured in full media containing 10% FBS for a period of 7 days in the absence or presence of the PPARγ ligand and troglitazone, at which point the cells were fixed, stained, and assessed for neutral lipid accumulation. As depicted in Fig. 1A, Oil Red O staining of hepatocytes demonstrates that increased lipid stores are present in hepatocytes expressing PPARγ2. This phenomenon appears to be independent of treatment with troglitazone, as comparable levels of staining were observed in cells treated with or without ligand, suggesting that an endogenous PPARγ ligand is either present in the media or produced by the cells. These results concur with a report demonstrating that overexpression of PPARγ1 in the mouse liver induces steatosis and expression of adipocyte-specific genes regardless of whether animals were treated with or without troglitazone (55). Closer inspection reveals the presence of lipids in droplets within the PPARγ2 hepatocytes, whereas minimal staining was observed in hepatocytes containing the control pBABE vector (Fig. 1B).
PPARγ2 selectively induces adipogenic and lipogenic protein expression in hepatocytes.
Next, the capacity of PPARγ2 to induce adipogenic and lipogenic protein expression was determined. Hepatocyte cell lines expressing PPARγ2 or pBABE were cultured in the absence and presence of troglitazone for 5 days and subjected to Western blot analysis using antisera directed against several adipogenic and lipogenic proteins. As a positive control for adipogenic gene expression, 3T3-L1 fibroblasts were induced to differentiate into adipocytes and harvested for total cellular protein 5 days postinduction. As expected, perilipin, PPARγ, adiponectin, and aP2 were abundantly expressed in the 3T3-L1 adipocytes. Both aP2 and perilipin are direct downstream targets of PPARγ and are abundantly expressed in mature adipocytes (31, 45). It is of interest that expression of PPARγ2 in both the absence and presence of troglitazone correlated with an induction of aP2 expression, whereas perilipin expression was not detected in the hepatocytes (Fig. 2). In comparison, ADRP was expressed in hepatocytes and was increased by the presence of PPARγ2 in a troglitazone-independent manner. These results suggest that ADRP may be functioning to coat lipid droplets in hepatocytes. Adiponectin is the most abundant adipose tissue protein and has been shown to enhance insulin sensitization and glucose tolerance in animal models of diabetes and obesity (9, 54). Expression of PPARγ2 in an adipogenic fibroblast cell line induces adiponectin gene expression (23). In contrast, the data here demonstrate that adiponectin expression was not induced by PPARγ2 in hepatocytes. Sterol regulatory element-binding protein-1 (SREBP-1) positively regulates many genes functioning in lipid uptake and de novo fatty acid synthesis including FAS. SREBP-1 directly activates transcription of FAS via binding to two sterol regulatory elements present in the FAS promoter (8, 29). Both FAS and SREBP-1 were increased in cells constitutively expressing PPARγ2 (Fig. 2). These results suggest that PPARγ2 may be enhancing lipid accumulation via activation of SREBP-1.
Effect of PPARγ2 on the expression of adipogenic and hepatic genes.
The results shown in Fig. 2 demonstrate that PPARγ2 selectively induces expression of several adipogenic proteins. Although neither perilipin nor adiponectin was detected by Western blot analysis, it is plausible that PPARγ2 positively regulates transcriptional activation of these genes. Peroxisome proliferator response elements exist within both the aP2 and perilipin promoters, and PPARγ2 has been shown to activate transcription of these genes in adipogenic cell lines (4, 31, 44). Therefore, to determine the ability of PPARγ2 to induce adipogenic and lipogenic gene expression in hepatocytes, cells expressing PPARγ2 or control vector were cultured in the absence and presence of troglitazone, and total cellular RNA was harvested after 5 days in culture. Similar to results obtained from Western blot analysis, RT-PCR revealed that PPARγ2 selectively induces aP2 and SREBP-1 mRNA in a troglitazone-independent manner. Neither perilipin nor adiponectin mRNA expression was detected in the PPARγ2-expressing hepatocytes (Fig. 3A). Additionally, although PPARγ2 expression was associated with an increase in ADRP protein as observed by Western blot analysis (Fig. 2), there appeared to be no effect of PPARγ2 on ADRP at the level of mRNA, suggesting that the increase in ADRP occurs at a posttranscriptional level. The effect of PPARγ2 on expression of hepatocyte-specific genes was also examined. When PPARγ2-pBABE and pBABE control cells were compared, no detectable differences were observed in albumin or α1-antitrypsin expression, two proteins that are abundantly expressed in a normal differentiated liver (Fig. 3B). It is of interest that expression of PPARγ2 results in a decrease in hepatocyte nuclear factor 4 (HNF-4) expression in hepatocytes, suggesting that PPARγ2 may either directly or indirectly regulate its expression. A liver-specific knockout of HNF-4 in the mouse results in decreased expression of genes encoding proteins involved in the export of very low-density lipoproteins from the liver. These animals also accumulate excess lipids in their livers (26). Although expression of PPARγ2 correlates with a selective induction of several adipogenic and lipogenic genes, these hepatocytes still express levels of liver-specific genes similar to those observed in the control hepatocytes as well as the liver (Fig. 3, A and B). This suggests that PPARγ2 is not transdifferentiating the cultured hepatocytes into adipocytes but may be upregulating expression of genes that normally function in hepatocyte lipid metabolism.
Steatotic livers from ob/ob mice express enhanced levels of adipogenic and lipogenic genes.
To confirm the in vivo significance of the results obtained in this in vitro model of steatosis, adipogenic and lipogenic protein and mRNA expression were examined using an in vivo model of steatotic liver, the ob/ob mouse. Twelve-week-old ob/ob mice and age- and sex-matched controls were euthanized, and total protein and RNA extracts were prepared from liver and adipose tissue. Livers from the wild-type mice were lean with a normal red-brown color, whereas livers from the ob/ob mice were enlarged with a fatty, yellow phenotype. Wild-type animals weighed 30.5, 28.9, and 32.3 g, while the weights of the ob/ob animals were 57.6, 57.2, and 60.2 g. Paralleling the in vitro data, ADRP was highly induced in the livers of ob/ob animals, and both perilipin and adiponectin were detected primarily in fat tissue (Fig. 4A). An abundance of ADRP protein and apparent lack of perilipin in ob/ob livers revealed a unique difference in the pattern of gene expression in steatotic livers compared with adipose tissue. Steatotic livers of ob/ob mice express aP2 (although at lower levels than observed in adipose tissue), suggesting that PPARγ is activated in the livers of these animals. In addition, the lipogenic enzymes FAS and ACC are also elevated in the ob/ob livers compared with the livers of their wild-type counterparts.
To determine mRNA expression, total RNA extracted from the animals was subjected to semiquantitative RT-PCR. It is of interest that expression of PPARγ mRNA and its downstream target aP2 is elevated in ob/ob livers compared with wild-type controls (Fig. 4B). Additionally, ADRP mRNA expression was elevated in ob/ob livers compared with wild-type livers. Mice deficient in leptin are hyperlipidemic (5), and studies suggest that long-chain fatty acids stimulate ADRP gene expression at the level of transcription (19). This could be a potential mechanism for the increase in ADRP observed in the ob/ob livers. Paralleling the Western blot data, and coinciding with the in vitro model, mRNA expression of perilipin and adiponectin was detected exclusively in adipose tissue. Previous studies demonstrated that expression and ligand activation of PPARγ2 are capable of inducing C/EBPα gene expression in a preadipocyte, fibroblast cell line (24). In this experiment, although PPARγ mRNA is elevated in ob/ob livers, no visible increase in C/EBPα mRNA expression was observed. These observations parallel the results obtained in the PPARγ2-expressing hepatocytes. The results suggest that de novo lipid synthesis induced by PPARγ2 may be an important factor contributing to lipid accumulation both in vitro and in vivo.
Lipid accumulation induced by PPARγ2 in hepatocytes is enhanced by serum deprivation in combination with troglitazone treatment.
The preceding experiments indicate that PPARγ2 induces lipid accumulation in hepatocytes and show that it also selectively enhances expression of both SREBP-1 and FAS, suggesting that lipid synthesis may be contributing to steatosis in hepatocytes. Therefore, it was of interest to investigate the contribution of de novo synthesis to lipid accumulation. To address this question, hepatocytes were cultured in the presence (lipid positive) or in the absence (lipid negative) of serum that supplies an exogenous source of lipid. In addition, cells were treated with and without troglitazone for 7 days. On day 7, the cells were fixed and stained for neutral lipid accumulation with Oil Red O. Despite the lack of an exogenous source of lipid, the greatest amount of intracellular lipid accumulation is evident in PPARγ2-expressing hepatocytes cultured in the absence of serum (Fig. 5). Unlike the data reported in Figs. 1–3, in which the cells were maintained with serum, troglitazone treatment enhanced lipid accumulation in the serum-deprived cells. In comparison, cells containing the control vector alone accumulated substantially less lipid under all conditions. These data draw a parallel with the previous experiments, which show a PPARγ2-regulated increase in both SREBP-1 and FAS, and further support the hypothesis that de novo lipid synthesis contributes to lipid accumulation in hepatocytes.
PPARγ2 enhances de novo triacylglycerol synthesis in hepatocytes.
The PPARγ2-regulated increase in expression of both SREBP-1 and FAS, combined with the finding that PPARγ2-expressing cells accumulate lipid in the absence of an exogenous source of lipid (serum), suggests that PPARγ2 may increase hepatocyte lipid stores by enhancing de novo lipid synthesis. On the basis of these observations, the effect of PPARγ2 on de novo lipid synthesis was assessed. Multiple reports have shown that liver triacylglycerol stores are increased in steatotic livers of both ob/ob and A-ZIP/F-1 mice (20, 30). Therefore, the effect of PPARγ2 on de novo triacylglycerol synthesis was examined. Hepatocytes were cultured in the presence and absence of serum and troglitazone for 1 or 7 days. Twenty-four hours before harvest, [14C]acetate was added and its incorporation into triacylglycerols assessed as described in materials and methods. The data in Fig. 6 demonstrate that PPARγ2 increases de novo triacylglycerol synthesis at both time points. Figure 6, inset, depicts de novo triacylglycerol synthesis on day 1 on a smaller scale. It is evident from this graph that, on day 1 of the experiment, PPARγ2 is enhancing de novo lipid synthesis under all conditions. The levels of triacylglycerol synthesis in cells harvested on day 7 were higher than those of day 1 hepatocytes. Although it appears that ligand enhances de novo triacylglycerol synthesis, it is not required for the PPARγ effect. These results in combination with the data in Fig. 5 further support the hypothesis that de novo lipid biosynthesis contributes to lipid accumulation in PPARγ2-expressing hepatocytes.
ADRP protein expression increases with troglitazone treatment and serum deprivation in PPARγ2 hepatocytes.
The upregulation of ADRP in PPARγ2-expressing hepatocytes and the lack of expression of perilipin suggest that ADRP may be important in coating lipid droplets in steatotic hepatocytes. Therefore, it was of interest to determine whether ADRP expression also increases under conditions that enhance lipid accumulation and de novo triacylglycerol synthesis. Concurring with the previous experiments, Western blot analysis shows a substantial increase in the expression of ADRP in hepatocytes expressing PPARγ2 cultured with troglitazone in the absence of serum (Fig. 7). Furthermore, these results coincide with an upregulation of the lipogenic enzymes FAS and ACC. In contrast, expression of perilipin, adiponectin, and aP2 was not increased by culturing cells in the absence of exogenous lipids. These experiments demonstrate a link among lipid accumulation, de novo triacylglycerol synthesis, and an upregulation of ADRP, FAS, and ACC expression in hepatocytes and provide further evidence that similar to perilipin, ADRP is stabilized by lipid accumulation in cells.
ADRP coats lipid droplets in hepatocytes expressing PPARγ2.
To determine whether ADRP is functioning to coat lipid droplets in hepatocytes, cells expressing PPARγ2 or control vector were cultured for a period of 7 days, fixed, and immunostained with antibodies specific for ADRP and perilipin. Results in Fig. 8 demonstrate that ADRP uniformly coats lipid droplets in PPARγ2-expressing hepatocytes. Although staining for ADRP was detected in the control hepatocytes, it appeared to be diffused throughout the cytoplasm and was not localized to the surface of lipid droplets. This result is consistent with prior data from this body of work, which demonstrates that ADRP is expressed in control hepatocytes that accumulate substantially less lipids in droplets compared with PPARγ2-expressing hepatocytes (Figs. 2 and 5). Perilipin was present on a majority of droplets in differentiated 3T3-L1 adipocytes (Fig. 8) but was not detected on the surface of droplets in hepatocytes (data not shown). These results coincide with the observations that neither mRNA nor protein expression for perilipin was detected in the in vitro and in vivo models of steatosis. These data confirm that, although PPARγ2 regulates perilipin in adipocytes, it does not induce perilipin in hepatocytes. On the basis of these experiments, it is apparent that ADRP plays an important role in coating lipid droplets in steatotic hepatocytes.
Over the past few years, several studies have established a role for hepatic PPARγ in the development and maintenance of steatosis in the liver (20, 30, 55). A liver-specific knockout of PPARγ in both A-ZIP/F-1 and ob/ob mice results in decreased lipid stores in the livers of these animals and reduced expression of several genes important to adipocyte differentiation and lipid metabolism (20, 30). The mechanisms leading to increased lipid stores in hepatocytes remain unclear. The goal of this present study was to determine whether PPARγ2 induces steatosis in hepatocytes and to identify mechanisms regulating this process.
In a normal liver, most synthesized lipids are packaged into very low-density lipoprotein particles, which are exported and transported in the serum to peripheral tissues. The liver does not typically function to store excessive amounts of lipids as energy reserves for the body. One mechanism that could lead to increased lipid stores in hepatocytes is an increase in de novo lipid biosynthesis. This present study demonstrates that PPARγ2 is capable of inducing lipid accumulation in hepatocytes as observed in Fig. 1. Furthermore, SREBP-1 as well as ACC and FAS are increased in hepatocytes expressing PPARγ2 (Figs. 2 and 7). This finding suggests that a mechanism by which PPARγ2 drives lipid accumulation in hepatocytes is by augmenting pathways of hepatic de novo fatty acid biosynthesis. To support this hypothesis, hepatocytes were cultured in the absence and presence of serum (exogenous lipids) and treated with and without the PPARγ ligand troglitazone. It is of interest that in the absence of serum, PPARγ2-expressing hepatocytes accumulated lipids and that this phenomenon was enhanced by troglitazone (Fig. 5).
On the basis of these observations, the ability of PPARγ2 to regulate de novo lipid synthesis was examined. Coinciding with the studies performed on lipid accumulation mentioned above, de novo triacylglycerol synthesis was consistently higher in PPARγ2-expressing hepatocytes compared with control cells on both days 1 and 7 (Fig. 6). Additionally, culturing PPARγ2-expressing hepatocytes in the absence of serum alone or in combination with troglitazone increased triacylglycerol synthesis. These conditions also led to a severalfold increase in the expression of the lipogenic enzymes FAS and ACC, as well as the lipid droplet protein ADRP (Fig. 7). These experiments suggest that the absence of exogenous lipids in combination with a PPARγ ligand synergistically interact to enhance both lipid synthesis and accumulation. It is likely that several events working in concert are responsible for this result. One hypothesis is that pathways regulating cell growth and division are downregulated due to the absence of FBS from the medium, which contains many important growth factors that positively impact these processes. Expression of PPARγ is activated following the switch between growth and differentiation during adipocyte differentiation (50). Therefore, these conditions may create an environment in hepatocytes that favors PPARγ activity. Second, an initial PPARγ-independent mechanism may also be involved. It is evident from Fig. 6 that serum deprivation alone enhances hepatic lipogenesis. It is well known that cholesterol homeostasis is regulated by SREBP-1, and the absence of cholesterol in the medium may favor SREBP-1 activation, leading to an upregulation of lipid synthesis in hepatocytes (12, 13, 39). Additionally, activation of SREBP-1 is associated with the production of lipid molecules that have been shown to bind to PPARγ with high affinity and displace radiolabeled thiazolidinediones. Therefore, activation of SREBP-1 may also sustain PPARγ2 activity via increasing production of endogenous PPARγ ligands (27). Finally, polyunsaturated fatty acids (PUFAs) suppress SREBP-1 activity (53) and when ob/ob mice are fed a diet supplemented with PUFAs, SREBP-1 activity decreases, leading to a reduction in liver triacylglycerol content (37). Therefore, culturing hepatocytes in the absence of FBS and PUFAs may result in a relief of repression on SREBP-1 activity. It is conceivable that a combination of these factors results in the severalfold increase observed for lipid accumulation and synthesis. Further studies are necessary to determine the exact mechanisms showing how PPARγ2 induces SREBP-1 in hepatocytes.
As shown in Fig. 1, lipids are sequestered within droplets in hepatocytes. Two droplet proteins, ADRP and perilipin, have been studied extensively in adipocytes. ADRP is a ubiquitously expressed protein and is found at the highest levels in adipose tissue, but it is also expressed in other tissues, including the lung, liver, testis, and spleen (11). In contrast, perilipin expression is limited to adipocytes and steroidogenic cells (21, 22, 38). During preadipocyte differentiation, ADRP initially functions to coat lipid droplets but is replaced by perilipin in mature adipocytes (11). A peroxisome proliferator response element exists within the perilipin promoter and in vitro studies have demonstrated that PPARγ binds to this element and directs expression of the perilipin gene (4, 31). Therefore, it was of interest to determine whether PPARγ2 induces perilipin expression in hepatocytes. As shown in Figs. 2 and 3A, ADRP is abundantly expressed in hepatocytes, and its expression is increased in cells expressing PPARγ2; however, perilipin was not detected. Analysis of livers from ob/ob mice compared with wild-type mice also demonstrates an increase in ADRP abundance and absence of perilipin, suggesting that the major lipid droplet protein in steatotic livers is ADRP as opposed to perilipin. It is likely that this property of ADRP will have significant consequences when the metabolism of stored triacylglycerols in steatotic hepatocytes is considered compared with adipocytes.
Several possibilities exist that could explain how ADRP expression is increased and why perilipin is not activated. Perilipin is stabilized posttranslationally by lipid loading of cells (10). The NH2 termini of perilipin and ADRP contain two regions of sequence homology (28). Therefore, similar to perilipin, ADRP may be stabilized posttranslationally by lipid synthesis and accumulation in hepatocytes. Additionally, in vivo, ADRP mRNA expression is enhanced in steatotic livers of ob/ob mice compared with normal liver tissue from wild-type mice (Fig. 4B). Long-chain fatty acids have been shown to stimulate ADRP mRNA expression (19). Because ob/ob mice are known to develop hyperlipidemia, it is conceivable that this condition enhances ADRP gene expression. These studies demonstrate that conditions that increase triacylglycerol synthesis (Fig. 6) and lipid accumulation (Fig. 5) also lead to an increase in ADRP expression (Fig. 7). The abundance of ADRP and lack of perilipin expression may have important ramifications for NAFLD. Perilipin stringently regulates lipolysis in adipocytes. Droplets coated with perilipin exhibit low rates of basal lipolysis and respond robustly to stimulated lipolysis (40, 42). In contrast, ADRP-coated droplets are associated with higher rates of basal lipolysis and have an attenuated response to stimulated lipolysis (42, 43).
Current theories on NAFLD propose that excessive levels of fatty acids may lead to the production of oxygen free radicals that damage hepatocyte membranes, triggering an inflammatory response and hepatocyte apoptosis (1). Therefore, the presence of ADRP and absence of perilipin on lipid droplets in hepatocytes may result in higher levels of intracellular fatty acids potentiating hepatocyte apoptosis and the progression of NAFLD. Further studies are necessary to delineate the mechanisms controlling expression of these lipid droplet proteins in hepatocytes and to determine the clinical relevance of these findings.
These experiments suggest that PPARγ2 is capable of selectively upregulating a subset of adipogenic genes in hepatocytes but does not appear to induce a complete adipogenic program, suggesting tissue-specific regulation of lipid homeostasis as well as pathology. The fact that neither adiponectin nor perilipin was detected in either the PPARγ2 hepatocytes or in the livers of ob/ob mice supports this theory, as both of these genes are positively regulated by PPARγ (4, 23, 31). Hepatocytes and adipocytes are derived from distinct embryonic germ layers, the endoderm and mesoderm, respectively (16, 56, 57). These developmental differences could result in conditions within hepatocytes that limit PPARγ2's ability to positively regulate adipogenic gene transcription. Moreover, expression of PPARγ2 does not lead to a dedifferentiation of hepatocytes, as these cells maintain expression of liver-specific genes. Taken together, these observations support the hypothesis that PPARγ2 is regulating steatosis in hepatocytes by enhancing hepatic lipogenesis and is not inducing a transdifferentiation of hepatocytes into adipogenic cells.
In conclusion, the aforementioned data suggest a role for PPARγ2 as a regulator of steatosis in hepatocytes and propose that this process occurs through an induction of SREBP-1 and its downstream targets. It is still unclear whether PPARγ2 is directly or indirectly activating SREBP-1. Further studies are necessary to clarify this mechanism. This study and others support PPARγ's role in directing lipid accumulation and the maintenance of steatosis in hepatocytes (20, 30, 55). Additionally, the presence of ADRP on lipid droplets in steatotic hepatocytes may also be of clinical significance to NAFLD, and future studies are warranted to determine whether PPARγ is a therapeutic target for this disease. Information gained from these investigations may provide insight into the mechanisms regulating steatosis and may be clinically relevant for future treatment of NAFLD.
This work was supported by National Institutes of Health Grants DK-51586, DK-58825, and PO1-HL-13262 and American Heart Association Grant 0455846T.
The technical assistance of J. Lavanture is deeply appreciated, and we thank Dr. C. Chu for advice on confocal microscopy.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 by American Physiological Society