Transcription coactivator mediator subunit MED1 Is required for the development of fatty liver in the mouse


  • Liang Bai,

    1. Department of Pathology, Feinberg School of Medicine, Northwestern University, Chicago, IL
    2. Laboratory of Animal Fat Deposition and Muscle Development, College of Animal Science and Technology, Northwest A&F University, Shaanxi, China
    Search for more papers by this author
  • Yuzhi Jia,

    1. Department of Pathology, Feinberg School of Medicine, Northwestern University, Chicago, IL
    Search for more papers by this author
  • Navin Viswakarma,

    1. Department of Pathology, Feinberg School of Medicine, Northwestern University, Chicago, IL
    Search for more papers by this author
  • Jiansheng Huang,

    1. Department of Pathology, Feinberg School of Medicine, Northwestern University, Chicago, IL
    Search for more papers by this author
  • Aurore Vluggens,

    1. Department of Pathology, Feinberg School of Medicine, Northwestern University, Chicago, IL
    Search for more papers by this author
  • Nathan E. Wolins,

    1. Center for Human Nutrition, Department of Medicine, Washington University School of Medicine, St. Louis, MO
    Search for more papers by this author
  • Nadereh Jafari,

    1. Genomics Core Facility Center for Genetic Medicine, Feinberg School of Medicine, Northwestern University, Chicago, IL
    Search for more papers by this author
  • M. Sambasiva Rao,

    1. Department of Pathology, Feinberg School of Medicine, Northwestern University, Chicago, IL
    Search for more papers by this author
  • Jayme Borensztajn,

    1. Department of Pathology, Feinberg School of Medicine, Northwestern University, Chicago, IL
    Search for more papers by this author
  • Gongshe Yang,

    1. Laboratory of Animal Fat Deposition and Muscle Development, College of Animal Science and Technology, Northwest A&F University, Shaanxi, China
    Search for more papers by this author
  • Janardan K. Reddy

    Corresponding author
    1. Department of Pathology, Feinberg School of Medicine, Northwestern University, Chicago, IL
    • Department of Pathology, Feinberg School of Medicine, Northwestern University, 303 East Chicago Avenue, Chicago, IL 60611
    Search for more papers by this author
    • Supported in part by the National Institutes of Health (grants GM23750 and DK083163 to J.K.R; Clinical Nutrition Research Unit grant DK56351 to N.E.W); China Scholarship Council (to L.B.).

    • fax: 312-503-8241

  • Potential conflict of interest: Nothing to report.


Peroxisome proliferator-activated receptor-γ (PPARγ), a nuclear receptor, when overexpressed in liver stimulates the induction of adipocyte-specific and lipogenesis-related genes and causes hepatic steatosis. We report here that Mediator 1 (MED1; also known as PBP or TRAP220), a key subunit of the Mediator complex, is required for high-fat diet–induced hepatic steatosis as well as PPARγ-stimulated adipogenic hepatic steatosis. Mediator forms the bridge between transcriptional activators and RNA polymerase II. MED1 interacts with nuclear receptors such as PPARγ and other transcriptional activators. Liver-specific MED1 knockout (MED1ΔLiv) mice, when fed a high-fat (60% kcal fat) diet for up to 4 months failed to develop fatty liver. Similarly, MED1ΔLiv mice injected with adenovirus-PPARγ (Ad/PPARγ) by tail vein also did not develop fatty liver, whereas mice with MED1 (MED1fl/fl) fed a high-fat diet or injected with Ad/PPARγ developed severe hepatic steatosis. Gene expression profiling and northern blot analyses of Ad/PPARγ–injected mouse livers showed impaired induction in MED1ΔLiv mouse liver of adipogenic markers, such as aP2, adipsin, adiponectin, and lipid droplet-associated genes, including caveolin-1, CideA, S3-12, and others. These adipocyte-specific and lipogenesis-related genes are strongly induced in MED1fl/fl mouse liver in response to Ad/PPARγ. Re-expression of MED1 using adenovirally-driven MED1 (Ad/MED1) in MED1ΔLiv mouse liver restored PPARγ-stimulated hepatic adipogenic response. These studies also demonstrate that disruption of genes encoding other coactivators such as SRC-1, PRIC285, PRIP, and PIMT had no effect on hepatic adipogenesis induced by PPARγ overexpression. Conclusion: We conclude that transcription coactivator MED1 is required for high-fat diet–induced and PPARγ-stimulated fatty liver development, which suggests that MED1 may be considered a potential therapeutic target for hepatic steatosis. (HEPATOLOGY 2011;)

Nonalcoholic fatty liver disease is becoming a common chronic liver disorder with a morphological spectrum of liver pathology commencing with hepatic steatosis and steatohepatitis which may progress toward the development of cirrhosis and liver cancer.1, 2 Because the key aspects of lipid metabolism, including lipogenesis, fatty acid oxidation, lipoprotein uptake and secretion are regulated by the liver, an understanding of the regulatory mechanisms that influence hepatic lipid homeostasis and systemic energy balance is of paramount importance in gaining insights that might be useful in the management of fatty liver disease.1-4 In recent years, increasing attention is being focused on certain transcription factors/nuclear receptors that are known to serve as key regulatory molecules to influence hepatic lipid metabolism.3-5 In particular, the three members of the peroxisome proliferator-activated receptor (PPAR) subfamily of nuclear receptors, namely PPARα, PPARβ (also called PPARδ), and PPARγ, govern the regulation of liver lipid metabolism and thus influence the development of hepatic steatosis and fatty liver disease.2, 4, 5 Of the three members of the PPAR subfamily, PPARγ is critical for conserving energy as it contributes to adipogenesis,2, 4, 6-9 whereas both PPARα and PPARβ participate in energy expenditure.2, 5 PPARγ, which has two isoforms, PPARγ1, and an N-terminal 30–amino acid extended form PPARγ2 (henceforth referred to simply as PPARγ), is expressed at a relatively high level in adipose tissue, where it serves as a regulator of adipocyte differentiation and promotes energy storage in mature adipocytes.7, 8 Of particular interest is that overexpression of PPARγ in mouse liver leads to adipogenic hepatic steatosis (“hepatic adiposis”) and induces the expression of adipocyte-specific and lipogenesis-related genes.6 In contrast, liver-specific disruption of PPARγ exerts an opposite effect in that it dramatically reduces fatty liver.9, 10 Thus, PPARγ plays an important role in liver lipid metabolism and contributes to hepatic steatosis.

In the nucleus, PPARs heterodimerize with retinoid X receptor α and bind to peroxisome proliferator response elements in the promoter region of target genes.4, 11, 12 Transcriptional activity of nuclear receptors and other transcription factors requires certain coactivators and coactivator-associated proteins that include PBP/TRAP220 (Refs. 13-15) and /DRIP205/ARC/MED1 (henceforth referred to as MED1; reviewed in Refs. 15-17), SRC (steroid receptor coactivator)/p160 family of proteins)18 and others (reviewed in Refs. 17 and 19). Identification of an increasing array of coactivators in recent years raises new challenges about their specific functional role in PPAR action and lipid metabolism in liver.17, 18 Evidence indicates that coactivator MED1, the best-studied subunit of the 31 member mammalian Mediator complex,12-16 is required for PPARα-mediated transcriptional activity in vivo,20 for PPARα ligand-induced liver tumor development,21 and PPARγ-stimulated adipogenic differentiation in vitro,22 but the in vivo role of this and other coactivators in liver with regard to PPARγ function remains largely unknown. To delineate the in vivo function of coactivator molecules in PPARγ-stimulated adipogenic hepatic steatosis, we used genetically altered mouse lineages in this study and we demonstrate that deletion of MED1 in mouse liver (MED1ΔLiv) impairs high-fat diet–induced and PPARγ-stimulated hepatic steatosis, whereas deficiency of coactivators such as SRC-1, PRIC285, PRIP, and PIMT had no effect. Thus, liver MED1 contributes to hepatic steatosis as it is required for PPARγ function.


Ad/PPARγ, adenovirally-driven PPARγ; Ad/LacZ, adenovirally-driven βgalactosidase; Ad/MED1, adenovirally-driven MED1; ADRP, adipose differentiation-related protein; cDNA, complementary DNA; MED1, Mediator1; mRNA, messenger RNA; PCR, polymerase chain reaction; PIMT, PRIP-interacting protein; PPAR, peroxisome proliferator-activated receptor; PRIC285, PPAR-interacting cofactor 285; PRIP, PPAR-interacting protein; Q-PCR, quantitative polymerase chain reaction; SRC-1, steroid receptor-coactivator-1; TG, triglyceride.

Materials and Methods

This section is in the Supporting Materials and is available online.


MED1 Is Required for High-Fat Diet–Induced Hepatic Steatosis.

Wild-type (MED1fl/fl) and MED1ΔLiv mice were fed a high-fat diet (60% kcal fat) for 2, 4, 8, and 16 weeks. MED1fl/fl mice developed severe hepatic macrovesicular steatosis by 8 and 16 weeks on the high-fat diet but MED1ΔLiv mice exhibited only mild and spotty steatosis (Fig. 1A,B). Hepatic steatosis induced by the high-fat diet in MED1fl/fl mice was not associated with induction of PPARγ target gene aP2 but this protein was detected in PPARγ-induced hepatic adiposis (Fig. 1C).6 Glucose and insulin tolerance tests revealed that MED1ΔLiv mice fed a high-fat diet for 4 weeks (Supporting Fig. 1A) or 16 weeks (Supporting Fig. 1B) revealed lower glucose levels and exhibited greater insulin sensitivity (Supporting Fig. 1C) than MED1fl/fl mice. MED1ΔLiv mice also showed less weight gain on the high-fat diet compared with MED1fl/fl mice (Supporting Fig. 1D). These results suggest that MED1 deficiency increases glucose tolerance and insulin sensitivity.

Figure 1.

The MED1ΔLiv mouse is resistant to HFD-induced hepatic steatosis. (A) Gross and histological changes in liver of MED1fl/fl and MED1ΔLiv mice fed HFD for 2 and 4 months. (B) Liver weight/body weight ratio and (C) HFD-induced hepatic steatosis in MED1fl/fl mice is not associated with induction of aP2 protein as depicted using immunoblotting whereas hepatic adiposis resulting from PPARγ overexpression in mouse is associated with aP2 expression. Abbreviation: HFD, high-fat diet.

MED1 Is Required for PPARγ-Induced Adipogenic Hepatic Steatosis.

PPARγ, when overexpressed in liver, induces adipogenic hepatic steatosis along with increased expression of adipocyte-specific as well as lipogenesis-related genes.6 To investigate the role of MED1 in PPARγ-stimulated hepatic steatosis, we have used the conditional MED1 liver knockout mice.20 As expected, MED1fl/fl mice injected intravenously with 1 × 1011 adenovirus-PPARγ (Ad/PPARγ) particles revealed severe hepatic steatosis (Fig. 2A).6 In contrast, PPARγ overexpression failed to induce hepatic steatosis in MED1ΔLiv mouse (Fig. 2A). MED1ΔLiv mouse liver with PPARγ overexpression appeared essentially similar to the livers of uninjected MED1ΔLiv mice or those injected with Ad/β-galactosidase (Ad/LacZ) (Fig. 2A,B). Hematoxylin and eosin (H&E) and Oil Red O staining revealed no lipid accumulation in the MED1ΔLiv mouse liver except for a few large hepatocytes that escaped Cre-mediated gene deletion (Fig. 2C,D). In contrast, PPARγ overexpression in MED1fl/fl mouse liver resulted in a marked accumulation of lipid in hepatocytes (Fig. 2C,D). Immunohistochemical analysis confirmed MED1 nuclear staining in all hepatic parenchymal cells in MED1fl/fl mice, whereas only an occasional liver nucleus stained positive for MED1 in MED1ΔLiv mouse liver (Fig. 2C; MED1 IHC). In PPARγ overexpressing MED1fl/fl and MED1ΔLiv mouse livers nuclear localization of PPARγ was evident by immunohistochemistry (Supporting Fig. 2). In uninjected MED1fl/fl control livers, nuclear staining of PPARγ was not evident. Additionally, >70% of hepatocytes stained positively for β-galactosidase in MED1fl/fl mice after Ad/lacZ injection, revealing the efficacy of adenovirally-directed gene expression in liver (Fig. 2C; inset in Ad/LacZ).

Figure 2.

PPARγ overexpression fails to induce fatty liver in MED1ΔLiv mouse. (A) Liver photograph of MED1fl/fl and MED1ΔLiv mouse killed 6 days without or with Ad/LacZ or Ad/PPARγ injection. (B) Liver-body weight ratio (n = 5) (**P < 0.01). (C) Liver sections stained with H&E, and immunohistochemical (IHC) localization of MED1 in MED1fl/fl and MED1ΔLiv mouse livers after Ad/PPARγ injection. β-Galactosidase staining of MED1fl/fl mouse liver shows LacZ expression (inset). Arrows in C (MED1ΔLiv) point to an occasional large MED1-positive residual hepatocyte that escaped Cre-mediated excision of MED1 floxed alleles. (D) Oil Red O staining of liver sections confirms attenuation of steatosis in MED1ΔLiv mice following Ad/PPARγ injection.

The time-course of TG accumulation in MED1fl/fl and MED1ΔLiv mouse liver following Ad/PPARγ or Ad/LacZ tail vein injection is shown in Fig. 3A. Hepatic TG content remained nearly unchanged in MED1ΔLiv mice with PPARγ overexpression (Fig. 3A). In contrast, PPARγ overexpression resulted in significant elevation of liver TG content in MED1fl/fl mice at days 4 and 6 (Fig. 3A). Plasma TG and cholesterol levels did not change with PPARγ overexpression in MED1fl/fl and MED1ΔLiv mice (Fig. 3B-D), indicating that neither the hepatic secretion of very-low-density lipoproteins nor the plasma clearance of these lipoproteins was affected by the treatment with Ad/PPARγ.

Figure 3.

MED1 deficiency protects against PPARγ-stimulated hepatic TG accumulation. Hepatic (A) TG and (B) cholesterol in MED1fl/fl and MED1ΔLiv mice at day 0, 2, 4, and 6 following Ad/LacZ or Ad/PPARγ injection. Serum (C) TG and cholesterol (D) in MED1fl/fl and MED1ΔLiv mice 6 days after Ad/LacZ or Ad/PPARγ injection (n = 5).

Loss of MED1 Impairs the Expression of PPARγ-Stimulated Adipogenic Genes in Liver.

Because PPARγ overexpression failed to induce hepatic steatosis in the absence of MED1, we investigated the role of MED1 in the adipogenic action of PPARγ in liver. Dramatic increases in the messenger RNA (mRNA) levels of classic fat differentiation gene markers, such as aP2 were noted in mice expressing MED1 but not in MED1-null livers (Fig. 4A). Increases in the mRNA levels of stearoyl-CoA desaturase 1 (SCD-1), Foxo1, and glucose-6-phosphatase (G-6-P) were observed in MED1fl/fl mouse livers but not in MED1ΔLiv mouse liver following PPARγ expression (Fig. 4A). Expression levels of hepatic mRNA content of peroxisomal β-oxidation enzymes, namely fatty acyl-CoA oxidase (Acox1), enoyl-CoA hydratase/L-3-hydroxyacyl-CoA dehydrogenase (L-PBE), and 3-ketoacyl-CoA thiolase (PTL) in MED1ΔLiv mice increased to a lesser extent as compared to a modest level of induction observed in MED1fl/fl mice after PPARγ expression (Fig. 4A). These observations suggest that the peroxisomal β-oxidation pathway was activated as an attempt to burn the overload of fatty acid in steatotic liver.2, 6 PPARγ overexpression also increased fatty acid translocase (CD36) mRNA concentration in liver of both MED1fl/fl and MED1ΔLiv mice (Fig. 4A). Moreover, the mRNA expression of lipid droplet protein genes CideA6, 23 and S3-126, 24 was barely detectable in MED1ΔLiv mice, but strongly induced in MED1fl/fl mice following PPARγ treatment (Fig. 4B). Interestingly, the mRNA levels of fat-specific gene 27 (FSP27),6 adipose differentiation-related protein (ADRP),24 and tail-interacting protein of 47 kDa (TIP47)24 showed no differences in MED1ΔLiv and MED1fl/fl mouse livers (Fig. 4B). ADRP protein content was higher in the livers of PPARγ-injected MED1fl/fl mice but not in MED1ΔLiv mice (Fig. 4C). This is likely due to ADRP being stabilized by intracellular lipid.24 Immunofluorescence and confocal microscopy revealed reductions in S3-12, ADRP, and CideA content in MED1ΔLiv mouse livers expressing PPARγ when compared to MED1fl/fl mouse (Fig. 4D). Furthermore, primary hepatocytes isolated from MED1fl/fl and MED1ΔLiv mice, when infected with Ad/PPARγ for 12 hours, showed that the mRNA expression levels of aP2, S3-12, and CideA were significantly lower in MED1ΔLiv mouse–derived hepatocytes (Fig. 5A-C). On western blot analysis, aP2 induction was evident in MED1fl/fl hepatocytes but not in MED1-deficient hepatocytes after Ad/PPARγ infection (Fig. 5D). These data clearly demonstrate the importance of MED1 in PPARγ-inducible adipogenic gene expression in liver under both in vivo and in vitro conditions.

Figure 4.

Hepatic MED1 deficiency impairs PPARγ-induced adipogenic gene expression. Northern blot analysis to assess the expression levels of (A) adipogenic genes (aP2, adipsin, adiponectin, and others) and fatty acid β-oxidation genes Acox1, L-PBE, and PTL, (B) as well as genes encoding for some lipid droplet-associated proteins such as CideA, FSP27, ADRP, and S3-12 in the livers of MED1fl/fl and MED1ΔLiv mice 6 days after Ad/PPARγ injection or Ad/LacZ. RNAs 28S and 18S served as loading control. RNA from two mice is shown for each treatment. (C) Western blot analysis of PPARγ, aP2, C/EBPδ, and lipid droplet proteins in adenovirus-treated livers. Liver homogenates from MED1fl/fl and MED1ΔLiv mice injected with Ad/LacZ or Ad/PPARγ were subjected to SDS-PAGE. β-actin was used as loading control. (D) Immunofluorescence localization (a panels) displaying localization of lipid droplet proteins S3-12, ADRP, and CideA in MED1fl/fl and MED1ΔLiv mouse livers 6 days after Ad/PPARγ injection. Oil Red O staining (B) and merged pictures (C) are also shown. Confocal images (D, extreme right) for S3-12, ADRP, and CideA in MED1fl/fl livers are shown. Abbreviations: C/EBP, CCAAT/enhancer binding protein; SDS-PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis.

Figure 5.

MED1 null primary hepatocytes exhibit defects in PPARγ-stimulated expression of adipogenic genes. (A-C) Q-PCR analysis of mRNA of aP2, S3-12, and CideA. Total RNA from hepatocytes isolated from MED1fl/fl and MED1ΔLiv mice 12 hours after Ad/PPARγ infection was used for Q-PCR analysis. The specific amplification of target genes was normalized with 18 S rRNA signal and the fold-change values are shown. Data from three independent Q-PCR determinations are expressed as mean ± SD. (D) Immunoblot analysis of adipogenesis marker protein aP2 in hepatocytes from MED1fl/fl (lanes 1 and 2) and MED1ΔLiv (lanes 3 and 4) mice infected with Ad/LacZ (lanes 1 and 3) or Ad/PPARγ (lanes 2 and 4) for 12 hours. The nonresponsive gene β-actin is shown as loading control. Abbreviation: SD, standard deviation.

Gene Expression Profiling.

To ascertain the regulatory role of MED1 in PPARγ-stimulated adipogenic hepatic steatosis, we performed complementary DNA (cDNA) microarray analysis to check the global transcriptional profile in mouse liver 4 days after injection with Ad/LacZ or Ad/PPARγ. When four-fold change is used as the cutoff, over 260 genes were up-regulated in Ad/PPARγ-stimulated MED1fl/fl mouse liver (Fig. 6A; Supporting Table 1). Most of these genes are involved in adipogenesis and lipid and glucose metabolism, suggesting a transdifferentiation trend of hepatocytes toward adipocytes or the development of adipogenic steatosis. In the absence of MED1 in liver the levels of expression of these genes were markedly subdued. These observations clearly establish that MED1 plays a key role in facilitating the transcriptional regulation of PPARγ target genes (Fig. 6A). Data shown in the heat map reveal that the expression levels of 28 genes involved in PPARγ function are dramatically lower in MED1ΔLiv mouse liver when compared to MED1fl/fl mouse following Ad/PPARγ administration (Fig. 6B). These include adiponectin, Elovl4, caveolin-1, Fabp5, Psapl1, Cyp4a14, and Hkdc1, among others.6 Interestingly, several genes, including fibroblast growth factor 21 (FGF21),25, 26 Fads2, Fads6, Elovl2, Apoa4, and Acot1, showed increased expression in MED1ΔLiv mouse with PPARγ overexpression (Fig. 6B). We validated microarray results by quantitative polymerase chain reaction (Q-PCR), which showed remarkably lower expression of S3-12, promethin, Fabp5, Hkdc1, Insig2, Nfatc4, Apob48r, caveolin-1, and Hsd17b2 in MED1ΔLiv mouse liver compared to MED1fl/fl mouse after injection with Ad/PPARγ (Fig. 6C; see Supporting Table 2 for primers used for Q-PCR). These data clearly establish that the loss of hepatic MED1 results in an abrogation of induction of lipogenesis-related genes but MED1 is not required for the induction of CD36 and FSP27 (Fig. 4).

Figure 6.

Gene networks regulated by PPARγ that require MED1. (A) Global transcriptional profiling of MED1fl/fl and MED1ΔLiv mouse liver following PPARγ overexpression for 4 days. Red circles represent up-regulation of genes in PPARγ overexpressing MED1fl/fl mouse liver (circles) and red triangles represent increases in MED1ΔLiv mouse liver. Green circles represent down-regulated genes in MED1fl/fl mice injected with Ad/PPARγ. The intensity of red color indicates the degree of up-regulation in PPARγ overexpressing MED1fl/fl and MED1ΔLiv mouse liver. (B) Microarray analysis of data derived from MED1fl/fl and MED1ΔLiv mouse livers 4 days after injection with Ad/LacZ or Ad/PPARγ. The log ratios were presented by heat map. (C) Q-PCR analysis to confirm the expression patterns of selected genes in the livers of MED1fl/fl and MED1ΔLiv mice following PPARγ overexpression. Fold-change values are shown.

Forced Expression of MED1 Rescues Adipogenic Steatosis Stimulated by PPARγ.

To further confirm the role of MED1 in PPARγ-stimulated hepatic steatosis in vivo, MED1 was re-expressed in MED1ΔLiv mouse liver using adenovirally-driven MED1 (Ad/MED1) (Fig. 7A-F). As expected, re-expression of MED1 in MED1ΔLiv mouse liver restored the PPARγ-stimulated steatotic response (Fig. 7B,D,F). The relative liver weight of MED1ΔLiv mouse injected with both Ad/MED1 and Ad/PPARγ increased as compared to MED1ΔLiv mouse treated with Ad/MED1 and Ad/LacZ (Fig. 7G). Re-expression of MED1 in MED1ΔLiv mouse liver also restored the expression of adipogenesis-related genes in response to PPARγ (Fig. 7H,I). These data clearly establish the critical role for MED1 in PPARγ-stimulated hepatic steatosis.

Figure 7.

Ad/MED1 rescues PPARγ-inducible hepatic steatosis in MED1ΔLiv mice. MED1ΔLiv mouse liver 6 days after injection with (A) Ad/MED1 plus Ad/LacZ or (B) Ad/MED1 plus Ad/PPARγ. H&E (C, D) and Oil Red O (E, F) stains reveal rescue of steatosis in MED1ΔLiv mouse liver following (D,F) expression of MED1 (Ad/MED1) and PPARγ (Ad/PPARγ). (C,E) Livers from MED1ΔLiv mice treated with Ad/MED1 and Ad/LacZ served as controls. (G) Relative liver weight of MED1ΔLiv mice after injection with Ad/MED1 and Ad/PPARγ (n = 5). (H) Northern blots depict the expression levels of PPARγ and lipogenesis-related genes in MED1ΔLiv and MED1fl/fl mice injected with Ad/PPARγ alone or with Ad/MED1. RNAs 28 S and 18 S were shown as loading control. This blot includes two samples for each group from 6 days postinjection. (I) Protein expression of aP2 and C/EBPδ in livers from MED1ΔLiv and MED1fl/fl mice injected with Ad/MED1 and/or Ad/PPARγ. β-actin was used as loading control.

Coactivators SRC-1, PRIC285, PRIP, and PIMT are Redundant for PPARγ-Stimulated Hepatic Steatosis.

To assess the role of other PPAR-associated coactivators in PPARγ-stimulated hepatic steatosis, germline knockout SRC-1 (SRC-1−/−) and PRIC285 (PRIC285−/−) mice and liver conditional null PRIPΔLiv and PIMTΔLiv mice and their corresponding control mice were injected with Ad/PPARγ and killed 5 days later. Fatty liver developed in mice lacking SRC-1, PRIC285, PRIP, and PIMT and their corresponding intact floxed controls after Ad/PPARγ administration (Supporting Fig. 3A). Increases in liver/body weight ratio were essentially similar in knockout and intact mice following PPARγ overexpression (Supporting Fig. 3B-E). Northern blot analysis revealed similar levels of increases in hepatic mRNA levels of adipogenesis genes in knockout and control mice following PPARγ overexpression (Supporting Fig. 4). These data suggest that SRC-1, PRIP, PIMT, and PRIC285 are dispensable for PPARγ-stimulated fatty liver development whereas MED1 is necessary for PPARγ dependent transcription of downstream target genes and the development of hepatic steatosis.


The nuclear receptor PPARγ, which is expressed at the highest level in adipose tissue, is a key regulator of adipocyte differentiation, lipid storage in white and brown adipose tissues, and energy homeostasis.8-10 Overexpression of PPARγ in liver results in adipose tissue specific gene expression in hepatocytes and leads to the development of adipogenic hepatic steatosis (“hepatic adiposis”).6 These observations suggest a potential role for PPARγ in fatty liver conditions.6, 9, 10, 27 In this regard, it is worth noting that hepatic steatosis exhibited by leptin-deficient (ob/ob) mice9 and lipoatrophic A-ZIP/F-1 mice10 is associated with enhanced expression levels in liver of PPARγ and induction of lipogenic genes.9, 10 Therefore, as a corollary, one would expect that, under conditions of abnormal lipid accumulation in liver, reduction in hepatic PPARγ level would lead to attenuation of the magnitude of steatosis. Indeed, liver-specific disruption of PPARγ reduces the extent of hepatic steatosis in ob/ob mice9 and A-ZIP/F-1 mice.10

The present study provides evidence that transcription coactivator MED1 is required in the mouse for both the high-fat diet–induced (Fig. 1) and PPARγ-stimulated hepatic steatosis (Fig. 2). In the absence of MED1, steatotic response resulting from high-fat diet feeding as well as by PPARγ overexpression is markedly attenuated (Fig. 2). MED1 is a key component of Mediator complex, and is required for RNA polymerase II–dependent gene transcription.15, 16 Mediator functions as a bridging complex in transmitting signals from transcriptional activators, including a broad range of nuclear receptors, to the general transcription machinery.15, 16 The present study with MED1 liver conditional null mice establishes the in vivo function of this coactivator in high-fat diet–induced as well as PPARγ-stimulated gene expression and points to a new layer of regulatory complexity in the development of hepatic steatosis. It appears that PPARγ is not activated in steatotic livers resulting from high-fat diet feeding, as evidenced by the failure of induction of aP2 expression in such livers, and yet high-fat diet–induced steatosis appears to be dependent on MED1. If one were to establish that PPARγ is not activated in liver upon high-fat feeding, then MED1 has significant PPARγ-independent effects on hepatic steatosis. On the other hand, PPARγ-stimulated hepatic steatosis is dependent on MED1.

Hepatic adiposis induced by PPARγ overexpression in liver is characterized by excess accumulation of cytoplasmic lipid droplets and is associated with increased expression of a variety of genes involved in adipogenesis.6 Lipid droplets consist of a triacylglycerol core with a phospholipid monolayer on the surface in which several proteins including members of perilipin family are embedded.24, 28 Perilipins coat nascent lipid droplets during accelerated TG synthesis and are required for its storage.24 Recently, another family of proteins, known as the cell death-inducing DFFA-like effecter (Cide) family of proteins (CideA, CideB, and CideC/fat-specific gene 27 or Fsp27) has been found to be associated with lipid droplets and regulate lipid droplet metabolism.23, 29 Recent studies have shown that lipid droplet proteins are increased in steatotic livers of fatty liver dystrophic (fld) mice.30 Of particular interest is that several of these lipid droplet–associated proteins in liver are regulated by PPARγ and their induction is positively correlated with the development of hepatic steatosis,30 supporting existing evidence indicating a key role for PPARγ in the development of hepatic steatosis and ectopic induction of lipogenic genes.6, 9, 10, 27, 30 Fat droplet proteins S3-12 (perilipin-4) and CideA, although they were strongly induced in MED1fl/fl mice, are barely detected in PPARγ-overexpressing MED1ΔLiv mouse liver (Fig. 4B-D). ADRP (perilipin-2) protein expression was lower in PPARγ-overexpressing MED1ΔLiv mouse liver cells when compared to that of MED1fl/fl mouse (Fig. 4C,D). Accordingly, the failure of MED1ΔLiv mouse liver to develop hepatic adiposis implies that this coactivator is essential for PPARγ-stimulated gene expression and adipogenesis. When MED1 expression is restored by Ad/MED1 administration in MED1ΔLiv mouse liver, PPARγ-stimulated hepatic adiposis ensued, as expected, confirming the essential role of MED1 in PPARγ function vis-à-vis hepatic steatosis.

In addition to lipid droplet proteins, there is evidence to indicate that other metabolic pathways also influence hepatic lipid accumulation.4, 5, 31 Recently, FGF21, a member of the endocrine FGF subfamily of metabolic hormones, has emerged as a key regulator of glucose and lipid metabolism in liver.26 FGF21 reverses hepatic steatosis by lowering TG levels.26 This has been attributed to inhibition of nuclear sterol regulatory element binding protein-1 (SREBP-1) and of hepatic lipogenic, adipogenic, and glucose production pathways.26 FGF21 also increases energy expenditure by increasing fatty acid oxidation and improves insulin sensitivity in high-fat diet–induced obese mice.26 We now show that FGF21 expression is increased in MED1ΔLiv mouse liver after PPARγ overexpression (Fig. 6B). This might suggest that FGF21-regulated inhibition of SREBP-1 and of other adipogenesis-related genes, such as adipsin, adiponectin, caveolin-1, and SMAF1, contribute to the attenuation of hepatic steatosis in MED1ΔLiv mouse following PPARγ overexpression (Fig. 6B). However, it is unclear as to how FGF21 levels are increased in MED1ΔLiv mouse liver following PPARγ overexpression. FGF21 is regulated by both PPARγ and PPARα and because this regulation requires MED1 it is conceivable that other mechanism(s) also exist to maintain high FGF21 levels in MED1 null livers.

In this study, we also report that other coactivators, namely SRC-1, PRIC285, PRIP, and PIMT, are not essential for PPARγ-induced adipogenic steatosis (Supporting Fig. 3). PPARγ stimulated hepatic steatosis in these coactivator null mouse livers, indicating the redundancy of these coactivators in PPARγ function in vivo. Disruption of genes encoding for p160/SRC-1 family members (SRC-1, SRC-2, and SRC-3) singly, has been shown to be redundant for PPARα-regulated gene expression in mouse liver.17, 32 PRIC285, a component of PRIC complex, has been shown to interact with PPARα, PPARγ, TRβ1, ERα, and RXRα.33 PRIC285−/− mice showed no differences in the magnitude of pleiotropic responses when challenged with PPARα ligands, such as Wy-14,643 or ciprofibrate, implying that PRIC285 is not essential for PPARα function.34 We now demonstrate that PRIC285 is also unnecessary for PPARγ function in liver. Coactivator PRIP (NCoA6) and its associated protein PIMT (NCoA6IP), function as linkers between the two major multiprotein complexes anchored by CBP/p300 and MED1.35 PIMT interacts with coactivators CBP, p300, MED1, and PRIP in vivo and in vitro.35 PRIP-deficient mouse embryonic fibroblasts are refractory to PPARγ-stimulated adipogenic conversion and fail to express adipogenic marker aP2, a PPARγ-responsive gene.36 However, surprisingly, our in vivo observations indicated the development of severe fatty liver in PRIPΔLiv and PIMTΔLiv mice following PPARγ overexpression. These results suggest that under in vivo conditions, MED1 absence results in a dominant phenotype as compared to PRIP deletion. The nonessential role of PRIP and PIMT in the PPARγ pathway may be due to the compensation between PRIP and PIMT in vivo. In conclusion, this study provides evidence that MED1 is required for high-fat diet–induced and PPARγ-induced hepatic steatosis and that loss of MED1 protects against fatty liver under these conditions. It is possible that expression levels of MED1 in liver might also play a significant role in the progression of fatty liver disease by modulating lipotoxicity and influencing steatohepatitis and endoplasmic reticulum stress.