Sumoylated PPARα mediates sex-specific gene repression and protects the liver from estrogen-induced toxicity in mice

As most metabolic studies are conducted in male animals, understanding the sex specificity of the underlying molecular pathways has been broadly neglected; for example, whether PPARs elicit sex-dependent responses has not been determined. Here we show that in mice, PPARα has broad female-dependent repressive actions on hepatic genes involved in steroid metabolism and immunity. In male mice, this effect was reproduced by the administration of a synthetic PPARα ligand. Using the steroid oxysterol 7α-hydroxylase cytochrome P450 7b1 (Cyp7b1) gene as a model, we elucidated the molecular mechanism of this sex-specific PPARα-dependent repression. Initial sumoylation of the ligand-binding domain of PPARα triggered the interaction of PPARα with GA-binding protein α (GABPα) bound to the target Cyp7b1 promoter. Histone deacetylase and DNA and histone methylases were then recruited, and the adjacent Sp1-binding site and histones were methylated. These events resulted in loss of Sp1-stimulated expression and thus downregulation of Cyp7b1. Physiologically, this repression conferred on female mice protection against estrogen-induced intrahepatic cholestasis, the most common hepatic disease during pregnancy, suggesting a therapeutic target for prevention of this disease.

Many metabolic pathways impacting whole-body physiology are expressed in a sexually dimorphic manner in the liver. In fact, the mammalian liver is the center for sex-specific controls that contribute to differences in energy homeostasis, lipid and steroid hormone metabolism, and the degradation of xenobiotics. Furthermore, in both humans and mice, sex affects inflammatory responses and hepatocellular carcinoma development (1–5). The regulation of the events that shape sex-specific physiological and pathological phenotypes is not well understood. At the molecular level, many sex-specific control mechanisms are unknown. Many of the 1,600 hepatic genes differentially expressed between sexes are regulated by differences in pituitary growth hormone (GH) secretion, which is pulsatile in males and continuous in females. The hormone affects gene expression by activating STATs (6, 7). Additional transcription factors known to activate genes controlling androgen metabolism, inflammation, and energy homeostasis, such as GA-binding proteins (GABPs), also contribute to hepatic sexual dimorphism (8–12). Furthermore, epigenetic modifications, including DNA methylation and methylation-sensitive transcription factors, participate in male-specific gene expression. Regulatory DNA methylation sites have been found in the male-specific Slp and Cyp2d9 promoters and other cytochrome P450 enzyme (Cyp) genes (8, 13). The CYPs metabolize diverse steroids, fatty acids, and many lipophilic drugs (14, 15). Though most pronounced in rat and mouse liver, sex differences in CYP expression also occur in humans (16). A prominent member of the CYP superfamily, oxysterol 7α-hydroxylase (CYP7B1), is enriched in male compared with female liver (17–21). The CYP7B1 protein represses androgen biosynthesis by modulating the availability of dehydroepiandrosterone (DHEA), the main precursor of the male hormone testosterone (18). In contrast to its negative effect on androgen production, CYP7B1 promotes estrogen receptor (ER) activity by catalyzing the clearance of 27-hydroxycholesterol, which functions as a competitive ER antagonist (22). The activated ER positively regulates the expression of Cyp7b1, which participates in estrogen-induced inflammation and hepatotoxicity, particularly in females (20, 21, 23). These pathologies affect susceptible women using estrogen-containing oral contraceptives or postmenopausal hormone replacement therapy (23). Moreover, estrogens cause intrahepatic cholestasis, the most common hepatic disease during pregnancy, which can result in intrauterine fetal death or spontaneous premature delivery (24). Therefore, understanding hepatic sexual dimorphism in steroid metabolism may help us to improve the efficacy of prevention measures and treatments based on steroid hormones. Similarly, knowledge of the molecular basis of sex-specific protective effects could be used for the development of treatments for inflammation- and hormone detoxification–related diseases.

It is likely that transcription factors other than those mentioned above play critical roles as determinants of sexual dimorphism, and their action is likely influenced by estrogen. In fact, signaling crosstalk between PPAR and ER have been documented (25), making PPAR an obvious candidate for involvement in the regulation of hepatic sexual dimorphism. Like the ER, the 3 different PPAR isotypes (PPARα, PPARβ [also called PPARδ], and PPARγ), belong to the nuclear receptor superfamily (26). The PPARα isotype regulates hepatic pathways associated with energy homeostasis, hepatic detoxification, the inflammatory response, and hepatocellular carcinogenesis (27–29). In addition to ligand binding, posttranslational modifications contribute to PPAR activity. Hence, sumoylation is implicated in the ligand-dependent transrepression of proinflammatory genes by PPARγ (30, 31). Sex-specific differences involving PPARα functions are poorly understood. Investigating sex-specific metabolic regulations may unveil novel functions of PPARα and help in the design of new prevention measures and treatments for dysfunctions related to sex-associated steroid detoxification.

With this objective in mind, we studied how PPARα impacts female hepatic metabolism. We compared the mRNA expression profiles in the liver of male and female mice in the presence or absence of PPARα. We discovered a marked sex-dependent inhibitory action of the receptor in the female liver. This transcriptional repression depended on the sex-specific sumoylation of PPARα; which triggers a cascade of events, described for the first time to our knowledge in this study, that result in the downregulation of many promoters.

PPARα-mediated inhibition of hepatic gene expression is more prevalent in females. We hypothesized that, among the genes differentially expressed in the male and female liver, a subset of genes is likely controlled by the hepatic regulator PPARα. To test this possibility, we compared the gene expression profiles of liver and heart, organs in which PPARα exerts regulatory functions, in female and male PPARα-null and WT mice. All animals were sacrificed at zeitgeber time 14 (ZT14; ZT00: lights on, ZT12: lights off), when the PPARα-stimulated genes were expected to reach maximum expression (32). Microarray data were analyzed using a linear model, with PPARα, sex, and their interaction as the factors examined (Supplemental Tables 1–5; supplemental material available online with this article; doi: 10.1172/JCI39019DS1). Genes with a statistically significant interaction factor were differentially affected by sex in PPARα-null mice compared with WT animals. Although the ablation of PPARα affected both the liver and heart as expected, there was a marked differential effect between the sexes only in the liver (Figure 1, A and B). The majority of genes with a significant interaction (false discovery rate [FDR], <5%) exhibited greater differences between PPARα-null and WT mice for females compared with males (Figure 1C). An analysis of these differences suggested a broader impact of PPARα-dependent repression in females. In contrast, fewer genes were specifically induced by PPARα in females, for example Rad51l1, which is involved in DNA repair. Together, these data demonstrated a sex-dependent function of PPARα in the liver, but not the heart, mainly directed toward the repression of gene expression in the female.

Figure 1

The hepatic sexual dimorphism of PPARα activity, with repressive functions in females. (A and B) Venn diagrams representing probe sets regulated in the heart (A) and liver (B). Microarray data were analyzed using a linear model with PPARα, sex, and their interaction as factors. For each factor, the total number of significant probe sets (FDR, <5%) is indicated in parentheses; the number of overlapping probe sets is indicated in the diagram. (C) Probe sets with a significant interaction factor. Log₂-fold differences between PPARα-null and WT for males and females are connected with a red line if the value is greater in females and a green line if the value is greater in males.

Repression of genes involved in steroid biosynthesis and the immune response. We concentrated on the liver and searched for specific pathways exhibiting PPARα-dependent sexual dimorphism using a gene set enrichment analysis strategy. We looked for gene sets that were significantly up- or downregulated based on the moderated t statistic calculated for the interaction factor. Sets of genes that differed significantly between females and males were involved in hepatic steroid metabolism (including cholesterol biosynthesis), the NADPH cytochrome P450 reductase–regulated (CPR-regulated) pathway, and immunity via the complement pathway (Figure 2A and Supplemental Table 6). All of these genes were upregulated in PPARα-null females compared with WT females but remained unchanged in PPARα-null males compared with WT males. Notably, in both humans and mice, dysfunctional CPR expression is linked to impaired steroidogenesis (33).

Figure 2

PPARα modulates hepatic steroid biosynthesis. (A) Heat maps of selected gene sets significantly enriched in genes with sexual dimorphism in PPARα-null mice. The log₂-fold differences between PPARα-null and WT female and male mice are presented; for each gene set, the color correspondence is indicated on the right. Genes are ordered in each set by significance of the change in expression in females, with the most significant on the left. Cholesterol biosynthesis, cholesterol biosynthesis canonical pathway; CR low liver up, upregulated in liver from mice with reduced liver expression of NADPH–cytochrome P450 reductase (CPR) versus normal controls; Comppathway, both the classic and alternative immune complement pathways that promote inflammation, foreign cell lysis, and phagocytosis. For data, see Supplemental Table 4. (B) Female PPARα-null (KO) and WT mice (n = 8 per group) were treated (black bars) or not treated (white bars) for 5 days with WY-14643. Plasma androstenedione and testosterone levels were determined. Values are mean ± SEM. *P ≤ 0.05, **P ≤ 0.01.

Since hepatic steroidogenesis is a crucial component of sexual dimorphism, we tested whether the ablation of PPARα impacts the levels of plasma steroids, which are physiological markers of steroid biosynthesis dimorphism. Female PPARα-deficient mice had lower plasma androstenedione and testosterone levels than WT mice (Figure 2B), suggesting a positive effect of PPARα on androgen biosynthesis. This effect was confirmed by treatment with the PPARα agonist WY-14643, which resulted in a significant increase in the levels of androstenedione and testosterone in WT but not PPARα-null female mice (Figure 2B). In contrast, the same treatment had no effect in males, which had much higher levels of testosterone, as it is mainly produced in the testis (data not shown). These results indicated that PPARα enhances testosterone levels in females by either directly stimulating its biosynthesis or inhibiting the catabolism of an androstenedione and testosterone precursor.

The activation function–2 domain is essential in the repression of Cyp7b1 by PPARα. In microarrays, CYP7B1 had one of the strongest positive interaction values, with a 4.1-fold increase in PPARα-null female livers compared with WT and no difference in males (FDR, 0.012; Supplemental Table 5). This result confirmed the sexual dimorphic expression pattern of a gene pivotal in steroid metabolism (17). In fact, CYP7B1 hydroxylates the testosterone precursor DHEA, diverting it from the sex hormone biosynthetic pathway in extragenital organs (20). Hepatic Cyp7b1 downregulation by PPARα would result in more DHEA being available for sex hormone biosynthesis and maintaining physiological levels of circulating androstenedione and testosterone (Figure 3A). Because this gene is strongly downregulated in females, and because of its role in estrogen-induced liver disease in women, we used it as a model to study the molecular mechanism by which PPARα exerts female-specific repression. As expected from the gene profiling data, quantitative PCR (QPCR) confirmed the much higher expression (11-fold) of Cyp7b1 in WT males than females (Figure 3B). Interestingly, in the absence of PPARα, PPARβ partially compensated for the repressive effect of the former on Cyp7b1 expression in females. In fact, only in double PPARα/β-null female mice did Cyp7b1 mRNA levels reach those of WT males (Supplemental Figure 1). After ligand treatment, expression was repressed more efficiently in males than females that presented lower basal expression (Figure 3, B and C). This inhibition was PPARα dependent, since it was not observed in PPARα-null mice.

Figure 3

The AF-2 domain is essential for PPARα repression of Cyp7b1. (A) Schematic representation of PPARα-modulated androgen production through regulation of Cyp7b1 in females (F). (B) Age-matched female and male WT and PPARα-null mice were treated with WY-14643 for 5 days, and hepatic Cyp7b1 expression was measured by quantitative real-time PCR (n = 6 per group) and (C) Western blotting. (D) Cotransfection of luc-reporter gene driven by a 1,086-bp CYP7B1 promoter fragment [pCYP7B1-luc(–1086/+189bp)] and mouse PPARα expression vector. 3X-PPRE-TK-Luc was used as a positive control. (E) The LBDα or N-terminal domain of PPARα was fused to the Gal4 DBD. The chimeras were cotransfected with the pCYP7B1-luc reporter gene. 5X-UAS-TK-Luc was used as a control for the activity of the PPARα N-terminal domain and LBD chimera (n = 4 per group). (F) PPARα lacking the 13 C-terminal amino acids (PPARαΔC13) failed to repress pCYP7B1-luc activity. (C–E) After transfection, cells were treated with WY-14643 (1 μM) or DMSO (vehicle; Control) for 20 hours. Values are presented as mean ± SEM (n = 3). *P ≤ 0.05, **P ≤ 0.01 versus control.

A 1,086-bp fragment of the human CYP7B1 promoter is known to mediate high basal activity in human kidney and liver cell lines (34). Using this promoter construct in NIH 3T3 cells, we reproduced the inhibitory effect of PPARα, which was enhanced by the PPARα ligand WY-14643, whereas a PPAR-positive reporter gene was stimulated (Figure 3D). Using chimeras containing either the N-terminal or ligand binding domain (LBD) of PPARα fused to the exogenous Gal4 DNA binding domain (DBD) in the presence or absence of PPARα ligand, we found that the LBD was responsible for the ligand-dependent repression of the Cyp7b1 promoter (Figure 3E). The last 13 amino acids (helix 12) of the activation function–2 (AF-2) domain of the LBD were necessary for mediating repression (Figure 3F). Collectively, these data show that the addition of the PPARα ligand in males mimics the repression observed in females and that helix 12 of the PPARα AF-2 domain is necessary for mediating this inhibition.

The repressor effect of PPARα is sumoylation dependent. Previous studies have shown that some ligand-activated nuclear receptors use sumoylation (small ubiquitin-related modification) to repress gene expression (30, 31). We explored the possibility that this posttranslational modification is implicated in sex-dependent repression by PPARα. First, knockdown of ubiquitin-like protein SUMO-1 conjugating enzyme (Ubc9) and protein inhibitor of activated STAT1 (PIAS1), two sumoylation rate-limiting ligases, abolished PPARα transrepression (Figure 4A). Second, an analysis of the PPARα amino acid sequence highlighted 3 lysine residues embedded in putative motifs for SUMO conjugation (Figure 4B). These observations suggested that sumoylation is involved in the repressive effects of PPARα. Next, we mutated the lysine residue to arginine in each of the putative sumoylation sites and tested the ability of these PPARα mutants to repress the Cyp7b1 promoter. The Lys358 mutation in helix 7 (Supplemental Figure 2) was sufficient to abolish the repressive activity of PPARα (Figure 4C), but the receptor conserved its transactivation potential (Figure 4D). Moreover, there was also a significant enhancement of PPARα sumoylation when small ubiquitin-related modifier 1 (SUMO1), SUMO2, and SUMO3 and PPARα were overexpressed in NIH 3T3 cells, demonstrating that PPARα has the potential to be modified by SUMO (Figure 4E).

Figure 4

A sumoylation-dependent mechanism mediates PPARα repression of Cyp7b1. (A) NIH 3T3 cells transfected with control and Ubc9- and PIAS1-specific siRNAs and treated with WY-14643 for 48 hours. White bars, control plasmid; black bars, PPARα plasmid. (B) Schematic representation and localization of 3 putative sumoylation sites in the PPARα LBD.Ψ indicates a hydrophobic amino acid. (C) Sumoylation sites were modified by site-directed mutagenesis (K358R, K185R, and K449R). The activity of these mutants was compared with WT mouse PPARα by cotransfections with pCYP7B1-Luc or (D) pHMGCS2-Luc (reporter gene driven by the HMGCS2 promoter), which is positively regulated by PPARα. For A, C, and D, values are presented as mean ± SEM (n = 3 per group). **P ≤ 0.01 versus control. (E) NIH 3T3 cells were cotransfected with expression vectors for WT or mutant (K358R) PPARα and c-myc-SUMO1, HA-SUMO2, or c-myc-SUMO3. Total proteins were immunoprecipitated using PPARα antibody and immunoblotted for c-Myc or HA-tag. After transfection, cells were treated with WY-14643 or vehicle (DMSO) for 20 hours. (F) Age-matched female (F) and male (M) PPARα-null and WT mice were treated with WY-14643 for 5 days. Hepatic nuclear proteins were immunoprecipitated with PPARα antibody and immunoblotted for SUMO1. The right panel presents an overexposure of the indicated area. Treated, WY-14643; Control, untreated.

To further explore the putative connection between PPARα sumoylation and hepatic sexual dimorphism, we measured the level of PPARα sumoylation in the liver of female and male mice. In untreated WT animals, sumoylated PPARα was detected in females but not males (Figure 4F). After treatment with the PPARα ligand, sumoylated PPARα increased to similar levels in both females and males (Figure 4F). These findings suggest that ligand-dependent PPARα sumoylation is determined by the agonist-induced conformation of the LBD, in which Lys358 is presented at the surface and therefore available for sumoylation, in contrast to the antagonist-induced conformation, in which case it is hidden (Supplemental Figure 2). This ligand-dependent conformational change is sufficient to explain why activated receptor is required for the repressive actions described in this study. Furthermore, the results indicate that sumoylation enzymes are similarly efficient in both sexes (Supplemental Table 3).

PPARα repression is mediated by an interaction with GABP. Transfection experiments identified the –144/–18 CYP7B1 promoter region as a mediator of the PPARα repressive effect (Figure 5A). Because no peroxisome proliferator responsive element (PPRE) was found in this region, it is likely that PPARα exerts its effect via interaction with other trans-acting factor(s) and/or cis-regulatory element(s). In this region, we found 7 potential binding sites for GABPs, heterodimeric transcription factors with widespread tissue distribution, among which the 3 most distal sites are well conserved between mammals (Figures 5B and Supplemental Figure 3A). Simultaneous cotransfection of human GABPα, which is the DNA binding subunit of the heterodimer, and GABPβ1, the effector subunit, increased the activity of the CYP7B1 promoter 8-fold, whereas separately transfecting GABPα and GABPβ1 had no effect (Figure 5C). Induction of the promoter by GABPα and GABPβ1 heterodimers was lost with the –18/+176 CYP7B1 promoter construct, which lacks the GABP-binding sites (Figure 5C). Mutation analyses showed that the centrally positioned GABP-binding site was necessary, in combination with either the proximal or distal site, to mediate basal GABP activation (Supplemental Figure 3B). Because GABP proteins are known to interact with other transcription factors (35), we investigated whether PPARα physically interacts with GABPα and thus interferes with GABP activity. An interaction between PPARα and GABPα was demonstrated by coimmunoprecipitation experiments. The interaction required the AF-2 domain of PPARα and an intact K358 sumoylation site (Figure 5D).

Figure 5

AF-2–dependent interaction between PPARα and GABP proteins on the CYP7B1 promoter. (A) Cotransfection experiments in NIH 3T3 cells using CYP7B1 promoter deletion constructs and a mouse PPARα expression vector or PSG5 as a control. Cells were then treated with WY-14643. (B) Schematic representation of the proximal CYP7B1 promoter region, which confers PPARα-dependent repression, with the positions of the GABP sites and the Sp1 site indicated. Sites conserved between mouse and human and sites only found in human are indicated. (C) NIH 3T3 cells were cotransfected with pCDNA3-flag-GABP expression vectors and a reporter gene driven by the –144/+189 CYP7B1 promoter region containing the GABP-binding sites or the –18/+176 promoter region, which lacks these sites. (A and C) Values are presented as mean ± SEM (n = 3 per group). **P ≤ 0.01. (D) NIH 3T3 cells were cotransfected with WT or mutant (Mut) PPARα (ΔC13 and K358R) expression vectors and the flag-GABP expression vector. Right: Reverse coimmunoprecipitation experiment. NT, nontransfected cells. (E) NIH 3T3 cells were cotransfected with a PPARα expression vector and Flag-GABPα expression vectors for WT or mutated LXXLL motif as indicated. (F) Same as in E, but IP was performed against Flag-tag, followed by Western blotting against PPARα. (G) Requirement of GABPα for mouse PPARα binding on the transfected human CYP7B1 promoter was determined by a ChiP assay and GABPα siRNA. (H) In vivo binding of PPARα and GABPα on the Cyp7b1 promoter was detected by a ChiP assay of hepatic nuclear protein using PPARα and GABPα antibodies. (I) Interaction of PPARα and GABPα on the Cyp7b1 promoter was detected by a re-ChIP assay (see Methods). Treated, WY-14643; Control, untreated.

Many nuclear receptor–interacting coactivators, including PGC-1, CBP, and RIP-140, contain one or more copies of the LXXLL motif necessary for their binding to nuclear receptors (29). We found that GABPα contains one such motif, LKKLL (amino acids 51–55), in its N-terminal domain. Strikingly, this motif is similar to the sequence found in the N-terminal region of PGC-1, a key energy metabolism coactivator known to interact with PPARα (36). Replacing leucines 54 and 55 with alanines impaired the ability of GABPα to interact with PPARα (Figure 5, E and F). Therefore, it is possible that PPARα uses GABPα to anchor itself to the Cyp7b1 promoter in vivo. This possibility was tested by ChiP. Knockdown of GABPα impaired the recruitment of PPARα on the human CYP7B1 promoter, indicating that GABPα is essential for PPARα-mediated repression (Figure 5G). Furthermore, we observed that PPARα and GABPα colocalized in vivo on this region of the mouse Cyp7b1 promoter (Figure 5, H and I). This experiment provided strong evidence for a GABPα LKKLL motif–mediated interaction between GABPα and the LBD of PPARα at the Cyp7b1 promoter.

PPARα-dependent repression and DNA methylation. Epigenetic modifications, particularly DNA methylation, have been implicated in hepatic sexual dimorphism (8, 13). The structure of the Cyp7b1 promoter, particularly the presence of an Sp1 site, which is known to stimulate the promoter, adjacent to the GABP sites suggests that its activity might be regulated by methylation (34). To test this possibility, we measured the expression of human CYP7B1 in hepatic cells treated with increasing concentrations of the DNA methyltransferase (Dnmt) inhibitor 5-aza-2-deoxycytidine (5-aza-dC). Enhanced CYP7B1 expression was observed (Figure 6A). To determine the role of Dnmt in PPARα-mediated repression, we performed transient transfection of a PPARα expression vector and a CYP7B1 promoter reporter gene in the presence of 5-aza-dC. The inhibitory effect of PPARα on CYP7B1 activity was abrogated by 5-aza-dC, suggesting that transcriptional repression is mediated by DNA methylation (Figure 6B). Moreover, knockdown of Dnmt3L, but not Dnmt1, increased CYP7B1 expression (Figure 6C) and abolished PPARα-dependent repression (Figure 6D). The methylation pattern of the native promoter in liver was analyzed by bisulfite sequencing, which relies on the conversion of unmethylated cytosine residues to uracil. A change in the methylation level of the above-mentioned Sp1 site, which contains a CpG dinucleotide, was observed. In untreated animals, methylation of the Sp1 site correlated with decreased expression levels of Cyp7b1 in mouse liver (compare Figure 6E and Figure 3A) and human liver cells (Figure 6F). In agreement with the repressive effect of ligand-bound PPARα, CpG methylation increased in males after WY-14643 treatment. A ChiP assay indicated that there is an inverse correlation between CpG methylation at the Sp1 binding site and the binding of Sp1 itself (Figure 6G). It is known that DNA methylation and histone trimethylation systems can synergize in complexes including methyl-CpG binding protein, nuclear receptor corepressor (NCoR), associated histone deacetylases, and histone methyl transferases (37). Similarly, we found that histone deacetylases, NCoR, and the trimethylation of histone 3 lysine 9 were implicated in the PPARα-mediated repression of Cyp7b1 (Supplemental Figure 4). Taken together, our observations suggest a functional interaction among sumoylated PPARα, epigenetic modifications, and chromatin remodeling in Cyp7b1 repression.

Figure 6

PPARα regulates epigenetic modification at the CYP7B1 promoter. (A) QPCR shows that CYP7B1 is activated in human hepatic cells (HepG2) treated with the Dnmt inhibitor 5-aza-dC (10 and 50 μM) for 4 days. (B) NIH 3T3 cells treated with WY-14643 and increasing concentrations of 5-aza-dC (10 and 50 μM) were transfected with a PSG5-mouse PPARα expression vector (PPARα) or the empty vector (control). (C) QPCR shows that knockdown of Dnmt3L, but not Dnmt1, stimulated CYP7B1 in HepG2 cells. (D) Knockdown of Dnmt3L, but not Dnmt1, abolishes PPARα repression of CYP7B1. (E) Hepatic genomic DNA was treated with bisulfite and the Sp1 site methylation level recorded. (F) After knockdown of human PPARα in HepG2 cells, Sp1 site methylation level was recorded and compared with CYP7B1 expression. Cells in D and F were treated with WY-14643 for 48 hours. In A–D and F, values are presented as mean ± SEM (n = 3). *P ≤ 0.05, **P ≤ 0.01 versus control. (G) In vivo binding of Sp1 to the Cyp7b1 promoter was detected by a ChiP assay of hepatic nuclear proteins using Sp1 antibody. Primers encompassing the GABPα and Sp1 binding sites were used for PCR. Treated, WY-14643; Control, untreated.

PPARα protects against estrogen-mediated hepatotoxicity. The repression of Cyp7b1 by PPARα suggests a possible protective effect of this receptor in estrogen-related liver diseases, including inflammation and cholestasis. Providing further support for this hypothesis, female PPARα-null mice displayed increased hepatic Cyp7b1 and complement protein 6 (C6) mRNA levels after exposure to the estrogen analog 17-α-ethynylestradiol (EE2) (Figure 7A). These changes correlated with increased plasma bilirubin levels and liver weight, two common markers of xenobiotic stress (Figure 7B and data not shown). To assess the potential protective effect of PPARα against hepatic estrogen toxicity, we treated female PPARα-null and WT mice with fenofibrate, a well-known PPARα activator used in clinics to treat hyperlipidemia. Fenofibrate treatment abolished the induction of Cyp7b1 and C6 expression by estrogen (Figure 7C). Moreover, fenofibrate lowered the high plasma bilirubin levels in estrogen-treated WT mice, but it had no effect on PPARα-null mice (Figure 7D). The same protective effect was observed with WY-14463 treatment (data not shown). These results underscore the potential of PPARα and its ligands for protecting against estrogen-induced intrahepatic cholestasis.

Figure 7

PPARα protects against estrogen-mediated hepatoxicity in females. (A and B) Female PPARα-null and WT mice treated (black bars) or not treated (white bars) for 5 days with EE2. (C and D) Female PPARα-null and WT mice were treated for 5 days with EE2, with (black bars) or without (white bars) fenofibrate. (A and C) Hepatic Cyp7b1 and C6 mRNA levels were measured by QPCR and normalized to the 36B4 expression level. (B and D) Plasma bilirubin concentration. In A–D, values are presented as mean ± SEM (n = 4). *P ≤ 0.05; **P ≤ 0.01. (E) Schematic representation of the PPARα-induced repression of Cyp7b1 in females. Sumoylation of PPARα promotes interaction with GABPs and HDACs, leading to DNA methylation by Dnmt3 and the displacement of Sp1 from its methylated core binding site. Removal of PPARα in female mice promotes the removal of the repression complex, which leads to Cyp7b1 stimulation. (F) Schematic representation of the protective effect of PPARα against estrogen-mediated inflammation and hepatotoxicity in females (F). αH3-K9 ac, αhistone 3 acetylated at Lys9; αH3-K9 tri-me, αhistone 3 trimethylated at Lys9.

Here we provide evidence for a marked sex-dependent difference in the downregulation of key hepatic functions by PPARα. This nuclear receptor has a broader repressive effect on gene expression in females, especially for genes involved in hepatic steroid synthesis and the complement system. Because the repression seen in females was mimicked in males with the addition of PPARα ligand, we postulated a mechanism in which the nuclear receptor interferes with coregulatory proteins and promoter-bound transcription factors without direct, sequence-specific interactions with DNA, referred to as transrepression (38).

Among the many downregulated genes, Cyp7b1 was chosen as a model because of its association with sex-related liver diseases. Our results demonstrate that the PPARα repression of this gene in females requires sumoylation of the receptor, which was mimicked in males after treatment with the hypolipidemic drug fenofibrate, suggesting a crucial role for the LBD of PPARα in female-specific repression. Although sumoylation-dependent transrepression was already observed in cell cultures for another PPAR isotype, PPARγ (30, 31), our study is the first to our knowledge to demonstrate a repressive role for sumoylated PPARα in an in vivo context: the female liver under normal physiological conditions.

Furthermore, the PPARα transrepression mechanism has not, to our knowledge, been previously observed for any other member of the receptor superfamily. Sumoylated PPARα interacts with GABP on the Cyp7b1 promoter via a functional nuclear receptor–interacting motif, LKKLL, in the N terminus of GABPα. The highly conserved LXXLL amphipathic α-helix motif, called the nuclear receptor box, is usually found in nuclear hormone receptor coactivators, suggesting the exclusive binding of GABPα or PPARα coactivators. In the unveiled transrepression mechanism, GABPα might neutralize coactivators when it recruits sumoylated PPARα together with NCoR, HDACs, and DNA and histone methylases to the Cyp7b1 promoter, in which no PPRE was identified. The GABP regulates the transcription of many viral and cellular genes; in particular, it plays a key role in the immune system by regulating T cell function and lymphocyte development (9, 10). The PPARα protein is a key player in inflammation (39) and modulates immune functions. Our study suggests that interactions between PPARα and GABP may be involved in these processes. In support of this idea, the proximal promoter region of the complement system and hepatic steroidogenesis genes, which were downregulated by PPARα, have GABP-binding sites (Supplemental Figure 3, B and C, and Supplemental Figure 5, B and C). Future work with specific hepatic GABPα–null mice should establish this putative, more general GABP-mediated PPARα repression mechanism.

The GABP and Sp1 proteins are known to cooperate in the activation of several GABP target genes, including lineage-specific genes (40). In the Cyp7b1 promoter, the GABP-binding sites are juxtaposed to an Sp1 site, which was methylated to a greater extent in female compared with male liver. Importantly, changes in DNA methylation have been associated with hepatic sexual dimorphism. Our finding establishes a link between PPARα sumoylation and the epigenetic modification of the Cyp7b1 promoter that correlates with reduced gene expression. Although other studies have shown regulation of gene activity through the methylation of the CpG core sequence in Sp1 sites (41, 42), our study has unveiled a mechanism by which PPARα uses DNA methylation to exert its transrepression activity (Figure 7E). This mechanism hinders Sp1 from binding to its methylated site and thus from stimulating Cyp7b1.

The ligand-dependent sumoylation of the PPARα LBD suggests that in females, this process is triggered by endogenous PPARα activators and, possibly, coactivators. In line with this hypothesis, we observed that the expression of fatty acid translocase/CD36 occurred predominantly in females, which suggests an increased availability of fatty acids for PPARα activation in the female liver (Supplemental Table 3). Interestingly, this difference in CD36 expression has also been observed in humans (43). The identity of PPARα activators enriched in females, such as specific ligands and coactivators, is currently being investigated.

The model in Figure 7F illustrates two complementary and not mutually exclusive mechanisms by which PPARα might repress estrogen-mediated hepatotoxicity: directly through the repression of the Cyp7b1 promoter or indirectly through its ability to increase testosterone levels, both leading to a reduction in ER activity. This model may help in the development of better treatments for estrogen-dependent liver diseases caused by high levels of estrogens, such as intrahepatic cholestasis of pregnancy (ICP), or for susceptible women undergoing postmenopausal hormone replacement therapy (44). In most areas of the world, ICP is observed in up to 1% of pregnancies, whereas it occurs in up to 15% of pregnancies in Chile and Bolivia. Today, the hydrophilic bile acid ursodeoxycholic acid is the most effective treatment for ICP (24), but it is not completely effective in preventing fetal death or premature delivery associated with fetal distress (45). Thus, due to its protective effect against estrogen-mediated hepatotoxicity, fenofibrate therapy might be an excellent alternative for treating ICP.

In addition to the fact that its corrective effects on estrogen decrease over time, postmenopausal hormone therapy is known to cause inflammation and liver damage in women (44, 46, 47). To prevent the deleterious effects of estrogen, androgen therapy is now a promising treatment for preventing menopausal problems in women, such as dyslipidemia, inflammation, and decreased libido (48–50). We demonstrated that the repressive effect of PPARα on Cyp7b1 leads to an increase in plasma androgen in females, suggesting a potential positive role in preventing aging-related problems in women.

In conclusion, we unveiled a molecular and physiological mechanism that explains, at least in part, sex-specific PPARα functions. Due to the crosstalk between this receptor and hepatic steroid signaling, new perspectives are emerging on the treatment of hepatic complications related to estrogen dysfunction in women.

View Supplemental data

We thank L. Michalik and N. Rotman for their comments on the manuscript. We are most grateful to D. Chollet and P. Descombes (Genomics Platform of the NCCR program Frontiers in Genetics) for their invaluable help with the Affymetrix transcription profiling studies; U. Schibler (University of Geneva) for help designing the transcription profiling experiment and stimulating discussions; M. Carrard (CIG, University of Lausanne) for measuring plasma bilirubin levels; C. Peter (CIG, University of Lausanne) for DNA sequencing; and V. Zoete (Molecular Modeling Platform, University of Lausanne) for providing Supplemental Figure 2. We are grateful to J.Y. Chiang, R.M. Evans, C.K. Glass, E.F. Johnson, C.R. Mueller, T. Osumi, and B. Staels for providing plasmids. This work was supported by the Swiss National Science Foundation, the Etat de Vaud, and the National Research Center “Frontiers in Genetics.”

Address correspondence to: Walter Wahli, Center for Integrative Genomics, National Research Center Frontiers in Genetics, University of Lausanne, Genopode Building, CH-1015 Lausanne, Switzerland. Phone: 41-0-21-692-41-10; Fax: 41-0-21-692-41-15; E-mail: walter.wahli@unil.ch.

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J. Clin. Invest.119:3138–3148 (2009). doi:10.1172/JCI39019.

1. Naugler, W.E., et al. 2007. Gender disparity in liver cancer due to sex differences in MyD88-dependent IL-6 production. Science. 317:121-124.
   View this article via: CrossRef PubMed Google Scholar
2. Bosch, F.X., Ribes, J., Diaz, M., Cleries, R. 2004. Primary liver cancer: worldwide incidence and trends. Gastroenterology. 127:S5-S16.
   View this article via: CrossRef PubMed Google Scholar
3. Crockett, E.T., Spielman, W., Dowlatshahi, S., He, J. 2006. Sex differences in inflammatory cytokine production in hepatic ischemia-reperfusion. J. Inflamm. (Lond.). 3:16.
   View this article via: PubMed Google Scholar
4. Ghebranious, N., Sell, S. 1998. Hepatitis B injury, male gender, aflatoxin, and p53 expression each contribute to hepatocarcinogenesis in transgenic mice. Hepatology. 27:383-391.
   View this article via: CrossRef PubMed Google Scholar
5. Mittendorfer, B. 2005. Sexual dimorphism in human lipid metabolism. J. Nutr. 135:681-686.
   View this article via: PubMed Google Scholar
6. Waxman, D.J., O’Connor, C. 2006. Growth hormone regulation of sex-dependent liver gene expression. Mol. Endocrinol. 20:2613-2629.
   View this article via: CrossRef PubMed Google Scholar
7. Jansson, J.O., Eden, S., Isaksson, O. 1985. Sexual dimorphism in the control of growth hormone secretion. Endocr. Rev. 6:128-150.
   View this article via: CrossRef PubMed Google Scholar
8. Yokomori, N., Kobayashi, R., Moore, R., Sueyoshi, T., Negishi, M. 1995. A DNA methylation site in the male-specific P450 (Cyp 2d-9) promoter and binding of the heteromeric transcription factor GABP. Mol. Cell. Biol. 15:5355-5362.
   View this article via: PubMed Google Scholar
9. Xue, H.H., et al. 2007. The transcription factor GABP is a critical regulator of B lymphocyte development. Immunity. 26:421-431.
   View this article via: CrossRef PubMed Google Scholar
10. Xue, H.H., et al. 2004. GA binding protein regulates interleukin 7 receptor alpha-chain gene expression in T cells. Nat. Immunol. 5:1036-1044.
    View this article via: CrossRef PubMed Google Scholar
11. Carter, R.S., Avadhani, N.G. 1994. Cooperative binding of GA-binding protein transcription factors to duplicated transcription initiation region repeats of the cytochrome c oxidase subunit IV gene. J. Biol. Chem. 269:4381-4387.
    View this article via: PubMed Google Scholar
12. Bannert, N., Avots, A., Baier, M., Serfling, E., Kurth, R. 1999. GA-binding protein factors, in concert with the coactivator CREB binding protein/p300, control the induction of the interleukin 16 promoter in T lymphocytes. Proc. Natl. Acad. Sci. U. S. A. 96:1541-1546.
    View this article via: CrossRef PubMed Google Scholar
13. Yokomori, N., Moore, R., Negishi, M. 1995. Sexually dimorphic DNA demethylation in the promoter of the Slp (sex-limited protein) gene in mouse liver. Proc. Natl. Acad. Sci. U. S. A. 92:1302-1306.
    View this article via: CrossRef PubMed Google Scholar
14. Anzenbacher, P., Anzenbacherova, E. 2001. Cytochromes P450 and metabolism of xenobiotics. Cell. Mol. Life Sci. 58:737-747.
    View this article via: CrossRef PubMed Google Scholar
15. Wang, H., et al. 2004. Cloning, expression, and characterization of three new mouse cytochrome p450 enzymes and partial characterization of their fatty acid oxidation activities. Mol. Pharmacol. 65:1148-1158.
    View this article via: CrossRef PubMed Google Scholar
16. Lamba, V., et al. 2003. Hepatic CYP2B6 expression: gender and ethnic differences and relationship to CYP2B6 genotype and CAR (constitutive androstane receptor) expression. J. Pharmacol. Exp. Ther. 307:906-922.
    View this article via: CrossRef PubMed Google Scholar
17. Stapleton, G., et al. 1995. A novel cytochrome P450 expressed primarily in brain. J. Biol. Chem. 270:29739-29745.
    View this article via: CrossRef PubMed Google Scholar
18. Martin, C., Bean, R., Rose, K., Habib, F., Seckl, J. 2001. cyp7b1 catalyses the 7alpha-hydroxylation of dehydroepiandrosterone and 25-hydroxycholesterol in rat prostate. Biochem. J. 355:509-515.
    View this article via: CrossRef PubMed Google Scholar
19. Pettersson, H., Holmberg, L., Axelson, M., Norlin, M. 2008. CYP7B1-mediated metabolism of dehydroepiandrosterone and 5alpha-androstane-3beta,17beta-diol--potential role(s) for estrogen signaling. FEBS J. 275:1778-1789.
    View this article via: CrossRef PubMed Google Scholar
20. Tang, W., Eggertsen, G., Chiang, J.Y., Norlin, M. 2006. Estrogen-mediated regulation of CYP7B1: a possible role for controlling DHEA levels in human tissues. J. Steroid Biochem. Mol. Biol. 100:42-51.
    View this article via: CrossRef PubMed Google Scholar
21. Tang, W., Norlin, M. 2006. Regulation of steroid hydroxylase CYP7B1 by androgens and estrogens in prostate cancer LNCaP cells. Biochem. Biophys. Res. Commun. 344:540-546.
    View this article via: CrossRef PubMed Google Scholar
22. Umetani, M., et al. 2007. 27-Hydroxycholesterol is an endogenous SERM that inhibits the cardiovascular effects of estrogen. Nat. Med. 13:1185-1192.
    View this article via: CrossRef PubMed Google Scholar
23. Yamamoto, Y., et al. 2006. Estrogen receptor alpha mediates 17alpha-ethynylestradiol causing hepatotoxicity. J. Biol. Chem. 281:16625-16631.
    View this article via: CrossRef PubMed Google Scholar
24. Pusl, T., Beuers, U. 2007. Intrahepatic cholestasis of pregnancy. Orphanet. J. Rare Dis. 2:26.
    View this article via: PubMed Google Scholar
25. Keller, H., Givel, F., Perroud, M., Wahli, W. 1995. Signaling cross-talk between peroxisome proliferator-activated receptor/retinoid X receptor and estrogen receptor through estrogen response elements. Mol. Endocrinol. 9:794-804.
    View this article via: CrossRef PubMed Google Scholar
26. Glass, C.K., Ogawa, S. 2006. Combinatorial roles of nuclear receptors in inflammation and immunity. Nat. Rev. Immunol. 6:44-55.
    View this article via: CrossRef PubMed Google Scholar
27. Desvergne, B., Michalik, L., Wahli, W. 2004. Be fit or be sick: peroxisome proliferator-activated receptors are down the road. Mol. Endocrinol. 18:1321-1332.
    View this article via: CrossRef PubMed Google Scholar
28. Mandard, S., Muller, M., Kersten, S. 2004. Peroxisome proliferator-activated receptor alpha target genes. Cell. Mol. Life Sci. 61:393-416.
    View this article via: CrossRef PubMed Google Scholar
29. Feige, J.N., Gelman, L., Michalik, L., Desvergne, B., Wahli, W. 2006. From molecular action to physiological outputs: peroxisome proliferator-activated receptors are nuclear receptors at the crossroads of key cellular functions. Prog. Lipid Res. 45:120-159.
    View this article via: CrossRef PubMed Google Scholar
30. Pascual, G., et al. 2005. A SUMOylation-dependent pathway mediates transrepression of inflammatory response genes by PPAR-gamma. Nature. 437:759-763.
    View this article via: CrossRef PubMed Google Scholar
31. Ghisletti, S., et al. 2007. Parallel SUMOylation-dependent pathways mediate gene- and signal-specific transrepression by LXRs and PPARgamma. Mol. Cell. 25:57-70.
    View this article via: CrossRef PubMed Google Scholar
32. Canaple, L., et al. 2006. Reciprocal regulation of brain and muscle Arnt-like protein 1 and peroxisome proliferator-activated receptor alpha defines a novel positive feedback loop in the rodent liver circadian clock. Mol. Endocrinol. 20:1715-1727.
    View this article via: CrossRef PubMed Google Scholar
33. Fluck, C.E., et al. 2004. Mutant P450 oxidoreductase causes disordered steroidogenesis with and without Antley-Bixler syndrome. Nat. Genet. 36:228-230.
    View this article via: CrossRef PubMed Google Scholar
34. Wu, Z., Chiang, J.Y. 2001. Transcriptional regulation of human oxysterol 7 alpha-hydroxylase gene (CYP7B1) by Sp1. Gene. 272:191-197.
    View this article via: CrossRef PubMed Google Scholar
35. Rosmarin, A.G., Resendes, K.K., Yang, Z., McMillan, J.N., Fleming, S.L. 2004. GA-binding protein transcription factor: a review of GABP as an integrator of intracellular signaling and protein-protein interactions. Blood Cells Mol. Dis. 32:143-154.
    View this article via: CrossRef PubMed Google Scholar
36. Vega, R.B., Huss, J.M., Kelly, D.P. 2000. The coactivator PGC-1 cooperates with peroxisome proliferator-activated receptor alpha in transcriptional control of nuclear genes encoding mitochondrial fatty acid oxidation enzymes. Mol. Cell. Biol. 20:1868-1876.
    View this article via: CrossRef PubMed Google Scholar
37. Yoon, H.G., Chan, D.W., Reynolds, A.B., Qin, J., Wong, J. 2003. N-CoR mediates DNA methylation-dependent repression through a methyl CpG binding protein Kaiso. Mol. Cell. 12:723-734.
    View this article via: CrossRef PubMed Google Scholar
38. Pascual, G., Glass, C.K. 2006. Nuclear receptors versus inflammation: mechanisms of transrepression. Trends Endocrinol. Metab. 17:321-327.
    View this article via: CrossRef PubMed Google Scholar
39. Devchand, P.R., et al. 1996. The PPARalpha-leukotriene B4 pathway to inflammation control. Nature. 384:39-43.
    View this article via: CrossRef PubMed Google Scholar
40. Xu, L., Komatsu, M. 2009. Promoter cloning and characterization of the anti-vascular proliferation gene, R-ras: role of ETS and Sp binding motifs. J. Biol. Chem. 284:2706-2718.
    View this article via: CrossRef PubMed Google Scholar
41. Brandeis, M., et al. 1994. Sp1 elements protect a CpG island from de novo methylation. Nature. 371:435-438.
    View this article via: CrossRef PubMed Google Scholar
42. Zhang, Q., Wang, H.Y., Liu, X., Wasik, M.A. 2007. STAT5A is epigenetically silenced by the tyrosine kinase NPM1-ALK and acts as a tumor suppressor by reciprocally inhibiting NPM1-ALK expression. Nat. Med. 13:1341-1348.
    View this article via: CrossRef PubMed Google Scholar
43. Stahlberg, N., et al. 2004. Female-predominant expression of fatty acid translocase/CD36 in rat and human liver. Endocrinology. 145:1972-1979.
    View this article via: CrossRef PubMed Google Scholar
44. Schreiber, A.J., Simon, F.R. 1983. Estrogen-induced cholestasis: clues to pathogenesis and treatment. Hepatology. 3:607-613.
    View this article via: PubMed Google Scholar
45. Lammert, F., Marschall, H.U., Matern, S. 2003. Intrahepatic cholestasis of pregnancy. Curr. Treat. Options Gastroenterol. 6:123-132.
    View this article via: CrossRef PubMed Google Scholar
46. Yang, S.X., et al. 2005. Estrogen replacement raises rat CRP without evidence of complement activation. Endocr. Res. 31:121-132.
    View this article via: CrossRef PubMed Google Scholar
47. Blumenfeld, Z., et al. 2007. High C-reactive protein levels are associated with oral hormonal menopausal therapy but not with intrauterine levonorgestrel and transdermal estradiol. Scand. J. Clin. Lab. Invest. 67:257-263.
    View this article via: CrossRef PubMed Google Scholar
48. Kingsberg, S., et al. 2007. Evaluation of the clinical relevance of benefits associated with transdermal testosterone treatment in postmenopausal women with hypoactive sexual desire disorder. J. Sex Med. 4:1001-1008.
    View this article via: CrossRef PubMed Google Scholar
49. Baulieu, E.E., et al. 2000. Dehydroepiandrosterone (DHEA), DHEA sulfate, and aging: contribution of the DHEAge Study to a sociobiomedical issue. Proc. Natl. Acad. Sci. U. S. A. 97:4279-4284.
    View this article via: CrossRef PubMed Google Scholar
50. Webb, S.J., Geoghegan, T.E., Prough, R.A., Michael Miller, K.K. 2006. The biological actions of dehydroepiandrosterone involves multiple receptors. Drug Metab. Rev. 38:89-116.
    View this article via: CrossRef PubMed Google Scholar
51. Bolstad, B.M., Irizarry, R.A., Astrand, M., Speed, T.P. 2003. A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics. 19:185-193.
    View this article via: CrossRef PubMed Google Scholar
52. Irizarry, R.A., et al. 2003. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 4:249-264.
    View this article via: CrossRef PubMed Google Scholar
53. Smyth, G.K. 2004. Linear models and empirical Bayes methods for assessing differential expression in microarray experiments. Stat. Appl. Genet. Mol. Biol. 3:Article3.
    View this article via: PubMed Google Scholar
54. Hochberg, Y., Benjamini, Y. 1990. More powerful procedures for multiple significance testing. Stat. Med. 9:811-818.
    View this article via: CrossRef PubMed Google Scholar
55. Mootha, V.K., et al. 2003. PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat. Genet. 34:267-273.
    View this article via: CrossRef PubMed Google Scholar
56. Efron, B., Tibshirani, R. 2007. On testing the significance of sets of genes. Ann. Appl. Stat. 1:107-129.
    View this article via: CrossRef Google Scholar
57. Subramanian, A., et al. 2005. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. U. S. A. 102:15545-15550.
    View this article via: CrossRef PubMed Google Scholar
58. Michalik, L., et al. 2005. Selective expression of a dominant-negative form of peroxisome proliferator-activated receptor in keratinocytes leads to impaired epidermal healing. Mol. Endocrinol. 19:2335-2348.
    View this article via: CrossRef PubMed Google Scholar
59. Lavery, D.J., Schibler, U. 1993. Circadian transcription of the cholesterol 7 alpha hydroxylase gene may involve the liver-enriched bZIP protein DBP. Genes Dev. 7:1871-1884.
    View this article via: CrossRef PubMed Google Scholar
60. Ripperger, J.A., Schibler, U. 2006. Rhythmic CLOCK-BMAL1 binding to multiple E-box motifs drives circadian Dbp transcription and chromatin transitions. Nat. Genet. 38:369-374.
    View this article via: CrossRef PubMed Google Scholar
61. Di-Poi, N., Desvergne, B., Michalik, L., Wahli, W. 2005. Transcriptional repression of peroxisome proliferator-activated receptor beta/delta in murine keratinocytes by CCAAT/enhancer-binding proteins. J. Biol. Chem. 280:38700-38710.
    View this article via: PubMed Google Scholar