Abstract

Hepatic nuclear factor 4α (HNF4α) modulates the transcriptional activation of numerous metabolic genes in liver. In this study, gene-array analysis revealed that HNF4α overexpression increased peroxisome proliferator-activated receptorγ (PPARγ) greatly in cultured rat primary hepatocytes. PPAR-response-element-driven reporter gene expression could be elevated by HNF4α. Bioinformatics analysis revealed a high-affinity HNF4α binding site in the human PPARγ2 promoter and in vitro experiments showed that this promoter could be transactivated by HNF4α. The presence of HNF4α on the promoter was then confirmed by ChIP assay. In vivo, hepatic overexpression of HNF4α decreased cholesterol levels both in plasma and liver and several hepatic genes related to cholesterol metabolism, including scavenger receptor BI (SR-BI), were upregulated. The upregulation of SR-BI by HNF4α could be inhibited by a PPARγ antagonist in vitro. In conclusion, HNF4α regulates cholesterol metabolism in rat by modulating the expression of SR-BI in the liver, in which the upregulation of PPARγ was involved.

1. Introduction

Nuclear receptors are ligand-activated transcription factors that regulate such diverse physiological processes as reproduction, development, and metabolism. Hepatic nuclear factor 4α (HNF4α) is a member of the nuclear receptor superfamily and plays an essential role in development and function in different organs, including liver, kidney, intestine, and pancreatic β cells. HNF4α contributes to gene regulation of the liver and pancreatic islet by binding directly to many actively transcribed genes [1]. The role of HNF4α as a metabolic transcriptional regulator in both the liver and pancreas has been uncovered by transcriptome analysis. Like other transcription factors, HNF4α modulates transcriptional activation of genes involved in transportation and metabolism of glucose, amino acids, lipids, and vitamins and in bile acid biosynthesis. HNF4α was once considered an orphan nuclear receptor because its endogenous ligand was not clear. Structural biology evidence suggests that HNF4α can bind to the C14 to C18 long-chain fatty acids [2, 3]. Recently, Sladek group reported that linoleic acid could bind with HNF4α, but ligand occupancy appeared not having a significant effect on its transcriptional activity [4]. Thus, HNF4α may be responsible for a change in metabolic status (e.g., changes in fatty acid levels) beyond direct transcriptional modulation.

Clinical evidence revealed that loss-of-function mutations in HNF4α cause maturity onset diabetes of the young 1 (MODY1) [5], a type of early-onset diabetes with pancreatic β cell dysfunction. Patients with MODY1 have decreased blood triglycerides and cholesterol levels [6]. Gene variants of HNF4α are also associated with type 2 diabetes [7, 8]. The mechanism underlying hypolipidemia remains poorly understood. HNF4α-null mice die during embryogenesis [9], and mice lacking hepatic HNF4α expression showed increased lipid levels in the liver and greatly disturbed lipid metabolism. Liver-specific HNF4α-knockout mice showed profoundly decreased plasma levels of cholesterol, high-density lipoprotein-cholesterol (HDL-C), and triglycerides and elevated plasma level of bile acid [10]. Despite the investigation of numerous genes, the pattern of gene regulation of lipid homeostasis by overexpression of HNF4α is still unclear.

Previously, we reported that hyperinsulinemia downregulated HNF4α and its target genes through the upregulation of sterol responsive element binding proteins (SREBPs), other important transcriptional factors regulating cholesterol and fatty acid metabolism [11]. To further investigate the transcription profile of HNF4α in hepatocytes and its crosstalk with other transcription factors, we used a gain-of function model with adenovirus encoding HNF4α (Ad-HNF4α) and microarray analysis in hepatocytes and rat plasma and liver. Hepatic overexpression of HNF4α led to lower cholesterol levels both in plasma and liver, and another important metabolic nuclear receptor, peroxisome proliferator-activated receptor γ (PPARγ), and a scavenger receptor SR-BI were involved in the process. Our findings suggest that hepatic HNF4α has a more potent role in cholesterol than glucose and triglycerides metabolism.

2. Material and Methods

2.1. Cell Culture and Reagents

The human hepatoma cell line HepG2 was maintained in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% (vol/vol) fetal bovine serum (FBS, Hyclone, Logan, Utah, USA). HEK293 cells were cultured in DMEM with 5% FBS. Primary isolated and cultured hepatocytes were obtained from 8-week-old male Sprague-Dawley rats by the two-step collagenase perfusion technique [12]. Cells (3 × 106) were plated on 60-mm rat tail collagen- (Millipore-) coated dishes containing medium 1640 supplemented with 20 mM HEPES, 1 mM sodium pyruvate, and 10% FBS for 3 days before treatment. The cells were infected with the adenovirus described below for 24 hr, then cultured in medium containing the PPARγ antagonist BADGE (10 mM, Sigma, St. Louis, Mo, USA) or the same volume of relevant vehicle DMSO for an additional 24 hr. All cells were maintained at 37°C in 5% CO2, 95% air.

2.2. Adenovirus Construction and Infection

Ad-HNF4α was generated by use of the Adeno-X Expression System 2 (Qbiogene, Carlsbad, Calif, USA). The adenovirus vector was based on replication-deficient E1 and E3 adenovirus under the transcriptional control of the cytomegalovirus promoter. The cDNA encoding a full-length rat HNF4α was obtained by digestion from the pcDNR-CMV-rHNF4α construct, then inserted into the pLP-Adeno-CMV construct under the control of the Cre/loxP system. DNA linearized by PacI digestion was packaged into the virus in HEK293 cells. The recombinant adenovirus (Ad-CMV-rHNF4α) was grown in HEK293 cells, purified, and titrated according to the manufacturer’s instructions. HepG2 and rat primary hepatocytes were incubated with recombinant adenovirus at a multiplicity of infection (MOI) of 20 for 24 hr before experiments.

2.3. RNA Extracts and Quantitative Real-Time PCR Analysis

Total RNA was extracted by the Trizol Reagent method (Applygen, Beijing, China). The isolated RNA was converted into cDNA. Quantitative RT-PCR with the Brilliant SYBR Green QPCR system involved use of β-actin as an internal control with the MX3000P QPCR detection system (Stratagene, La Jolla, Calif, USA). The primer sequences are in the supplementary tables. In a separate experiment, the extracted total RNA underwent PCR array of liver-related genes (SuperArray, SABiosciences, Frederick, Md, USA) by Kangcheng Inc. (Beijing, China).

2.4. Western Blot Analysis

Cell lysates and rat liver extracts were resolved by 10% SDS-PAGE and transferred to a nitrocellulose membrane. Protein expression of sEH, GAPDH and β-actin was detected by use of polyclonal antibodies anti-HNF4α (Santa Cruz Biotechnology, Santa Cruz, Calif, USA), anti-SR-BI (NOVUS Biologicals, Littleton, Colo, USA) and anti-β-actin (Bioss, Beijing, China) followed by horseradish peroxidase-conjugated secondary antibody. The protein bands were visualized by the ECL detection system (Amersham, Stockholm, Sweden), and the densities of the bands were quantified by computer-assisted image analysis (NIH Image J).

2.5. Promoter Reporter Assay

A 1.6-kb fragment of PPARγ2 promoter was cloned by PCR from the genome of the HEK293 cell line with the oligonucleotide primers 5′-TCCAGAAGTGAGACCCTTTG- AG-3′ and 5′-CATGGAATAGGGGTTTGCTGTAAT-3′. The amplified fragment was sequenced and then inserted into the EZ-T construct (Genestar, Beijing, China), then cloned into the reporter vector pGL3 (Promega, Madison, Wis, USA) on the SmaI and KpnI sites and named PPARγ2-(−1505)-Luc. PPARγ2-(−893)-Luc and PPARγ2-(−502)-Luc vectors were constructed by the same strategy. Transfection experiments were carried out in 24-well plates with use of JetPEITM (PolyPlus-transfection, Illkirch, France). Luciferase and β-galactosidase assays followed the manufacturer’s instructions (Promega, Madison, Wis, USA).

2.6. Chromatin Immunoprecipitation (ChIP) Assay

ChIP assays were performed as described [13]. In brief, HepG2 cells were cross-linked and sonicated, then underwent immunoprecipitation (IP) with polyclonal anti-HNF4α. Normal IgG was used as an IP control, and the supernatant was an input control. After digestion with proteinase K, the resting DNA was extracted, and the PPARγ2 promoter containing the HNF4α consensus element was amplified by PCR with the primers 5′-TGACAAGACCTGCTCC-3′ and 5′-TACGCTGTTAGGTTGG-3′. The resulting DNA was resolved on 2% agarose gel and stained with ethidium bromide.

2.7. Animal Experiment

The investigation conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85–23, revised 1996). The animal experimental protocol was approved by the Peking University Institutional Animal Care and Use Committee. Sprague-Dawley rats were fed standard laboratory chow and tap water ad libitum and bred under a 12-h light/12-h dark cycle. The 8-week-old (~200 g) male rats were injected with Ad-rHNF4α at 1 × 109 plaque-forming units in 0.5-mL saline through the tail vein. The same amount of Ad-GFP was injected in the control group ( ). Levels of plasma glucose were measured by use of a portable glucometer (ACCU-CHEK II; Roche Diagnostics, Basel, Swiss). Seven days after the injection, the rats fasted for 6 hr and then were anesthetized and killed. Blood samples were collected from the saphenous vein, and plasma levels of lipoproteins and insulin were measured. Liver tissues were dissected after a PBS rinse and stored at −80°C.

2.8. Measurement of Plasma and Tissue Lipids

Rat plasma samples were collected, and lipoproteins were separated by fast performance liquid chromatography (FPLC) (Pharmacia Biotech, Sweden). The levels of cholesterol and triglycerides in plasma or the FPLC fractions were detected by use of an automated clinical chemistry analyzer kit (Biosino Biotech Inc., Beijing, China). For quantification of liver cholesterol and triglycerides, approximately 100 mg liver was homogenized and the extraction was dissolved with chloroform : methanol (2 : 1). The lipid fractions were dried under nitrogen gas and re-solubilized by phosphate buffered saline containing 1% Triton X-100 before measurement of cholesterol and triglyceride levels.

2.9. Statistical Analysis

Data were analyzed by the unpaired Student t test, one-way ANOVA, or Mann-Whitney test (GraphPad Prism4 software). All values are expressed as mean ± SEM. Differences were considered statistically significant at .

3. Results

3.1. Hepatic HNF4α Induces PPARγ Upregulation and Activation

Previously, we reported that hepatic HNF4α was downregulated in db/db diabetic mice [14]. To further investigate the role of HNF4α in liver metabolism, we overexpressed HNF4α in rat primary hepatocytes by Ad-rHNF4α infection and analyzed the expression of liver-related genes by PCR array. As shown in Table 1, known HNF4α target genes, such as glucose-6-phosphatase, catalytic subunit (G6pc) and liver type of pyruvate kinase (L-PK), were upregulated by HNF4α overexpression, as reported previously [14]. Surprisingly, we found a 19.29-fold increase of PPARγ expression with Ad-rHNF4α infection relative to the Ad-GFP control (Table 1). Meanwhile, the expression of Srebf1, a gene encoding a major transcriptional regulator SREBP-1c, which is important in triglyceride metabolism, showed a 1.53-fold decrease in expression (data not shown). The upregulation of PPARγ, including PPARγ1 and PPARγ2, by HNF4α overexpression was confirmed by quantitative real-time PCR (Figure 1(a)). To assess whether HNF-4α contributes to the transactivation of PPARγ, we performed PPRE-driven luciferase reporter assays in the human hepatic cell line HepG2. Rosiglitazone, the PPARγ agonist, was used as a positive control. Compared with control cells cotransfected with vehicle plasmid, the PPRE-luc activity was induced by HNF-4α co-transfection, and BADGE, the PPARγ antagonist, could diminish this effect (Figure 1(b)), which suggests that HNF-4α transactivated PPRE-luc activity through PPARγ.

3.2. HNF4α Is Involved in PPARγ2 Transactivation

Sequence comparison by BLAST analysis revealed more than 80% similarity in the PPARγ2 promoter but not PPARγ1 promoter between human and rat. We identified a putative HNF4α site on the human PPARγ2 promoter (in the region of −842 to −827) by the prediction program of the CREAD platform (PKUHSC) (Figure 2(a)). ChIP assay revealed that HNF4α could bind to the predicted HNF4α binding site on the PPARγ2 promoter (Figure 2(b)). Moreover, quantitative PCR with PPARγ2 promoter sequence-specific primers revealed that HNF4α overexpression enhanced the promoter occupancy of HNF4α on the PPARγ2 promoter (Figure 2(b)). Transient transfection assay with a series of 5′-deletion reporter constructs containing different lengths of the PPARγ2 promoter revealed that deletion of 893 to 502 bp in the promoter region led to a large decrease in the ability of HNF4α to transactivate the PPARγ2 promoter (Figure 2(c)).

3.3. Hepatic Overexpression of HNF4α in Rat Lowered Plasma and Liver Cholesterol Levels

To ascertain the role of HNF4α-upregulated PPARγ in liver in vivo, rats were injected with Ad-HNF4α or Ad-GFP intravenously and killed 7 days later. Overexpression of HNF4α in liver significantly increased hepatic HNF4α protein level (Figure 3(a)) as compared with Ad-GFP-infected controls. mRNA levels of HNF4α target genes (Pepck, G6p, Apob, Apoc3, and L-pk) in liver were elevated by Ad-HNF4α (Figure 3(b)). The rats showed normal body weight, liver weight, plasma insulin level (Table 2), and oral glucose tolerance (data not shown). Plasma level of triglycerides was increased but not significantly (Table 2). Oil-red O staining showed no lipid accumulation in the liver (data not shown). However, cholesterol levels were reduced in both plasma and liver (Figures 3(c) and 3(d)). Plasma lipoprotein analysis revealed decreased LDL-C but not HDL-C fraction (Figure 3(c)).

3.4. Selective Regulation of Hepatic Cholesterol-Regulatory Genes by HNF4α in Liver

To study the mechanism of HNF4α-regulated cholesterol metabolism in liver, we examined the expression of pivotal genes governing liver cholesterol metabolism in rat liver with HNF4α overexpression. qPCR results revealed elevated mRNA levels of hepatic PPARγ and its known target gene CYP7α1 [15], as expected (Figure 4(a)). Genes regulating lipoprotein uptake, such as LDL receptor and scavenger receptor class B type I (SR-BI), were also upregulated. However, genes regulating cholesterol de novo synthesis, efflux and metabolism, such as ABCA1, ABCG5, ABCG8, and HMG-CoA reductase showed no regulation by HNF4α. Intriguingly, the mRNA level of lecithin:cholesterol acyltransferase (LCAT), which converts cholesterol and phosphatidylcholines (lecithins) to cholesteryl esters and lysophosphatidylcholines on the surface of high-density lipoproteins [16], was elevated by HNF4α overexpression (Figure 4(a)). Therefore, cholesterol uptake genes seemed to be regulated by HNF4α in liver, possibly via PPARγ upregulation.

3.5. Upregulation of SR-BI by HNF4α Depended on PPARγ

SR-BI was found regulated by activators of PPARγ in human hepatocytes [17] and a PPRE was proved to exist in the human SR-BI promoter [18]. Western blot analysis confirmed the upregulation of SR-BI with HNF4α overexpression at the protein level in vivo in rat liver (Figure 4(b)) and in vitro in rat hepatocytes (Figure 5(b)). We further investigated the role of HNF4α in SR-BI upregulation in cultured hepatocytes. The PPARγ antagonist BADGE could attenuate HNF4α-induced SR-BI expression at both mRNA and protein levels (Figures 5(a) and 5(b)), which suggests that HNF4α upregulates SR-BI in a PPARγ-dependent manner.

4. Discussion

PPARγ is a ligand-activated transcription factor that regulates diverse biological activities and plays major roles in many diseases, including diabetes mellitus, metabolic syndrome, and atherosclerosis [19]. It is highly expressed in adipose tissue, where it plays an essential role in fat storage and the differentiation of adipocytes [20]. However, PPARγ is expressed at low levels in other tissues, including liver. An antidiabetic drug, pioglitazone, a known PPARγ activator, significantly improved lipid metabolism and insulin responsiveness and reduced the hepatic inflammatory response [21]. Among the hepatic expression PPARγ isoforms, though PPARγ2 has lower expression amount than PPARγ1, it is the only PPARγ isoform which could be regulated at the transcriptional level by nutrition [22]. Furthermore, PPARγ2 is the liver isoform that is ectopically induced in response to excess nutrition or genetic obesity [23]. Here, we show that PPARγ is a direct target gene of HNF4α in rat, and HNF4α overexpression may achieve its cholesterol-lowering effect via PPARγ-SR-BI upregulation. Martinez-Jimenez et al. showed that HNF4α can bind to and activate the PPARγ1 promoter, and the transcriptional cofactor Hes6 interacted with HNF4α to prevent the hepatic transactivation of PPARγ1 [24]. In addition to Martinez’s work, our data demonstrate a functional HNF4-responsive element on the PPARγ2 promoter. HNF4α could upregulate both PPARγ1 and PPARγ2 in our gain-of-function model with overexpression of HNF4α. We did not investigate the involvement of hepatic Hes6. Recently, HNF4α was proved to have the ability to bind with RXRα, a classical PPARγ transcriptional cofactor [25]. The fold activation of PPARγ by HNF4α overexpression differed in vivo and in vitro and that might result from the varied combinatiorial regulation of various factors including ligands and other transcription factors. The authors do not exclude the possibility that overexpression of HNF4α may change the expression pattern of PPARγ transcription variants.

Expression of PPARγ in the liver was augmented in murine steatosis, and adenovirus-mediated overexpression of PPARγ in the liver provokes steatosis [26, 27]. Surprisingly, we did not observe any significant change in hepatic steatosis in Ad-HNF4α-treated rats. In a previous study, coexpression of HNF4α increased the promoter activity of PPARα [28]. TZD18, a potent agonist with dual PPARα/γ agonist activities, affected lipid homeostasis, thus leading to an antiatherogenic plasma lipid profile [29]. Our results showed reduced LDL-C level and upregulation of PPARγ by HNF4α. HNF4α might be a dual PPARα/γ activator, and overexpression of HNF4α may be effective in treatment of type 2 diabetes and dyslipidemia and in preventing atherosclerotic cardiovascular disease.

Cholesterol homeostasis depends on proper control of cholesterol uptake, de novo synthesis, efflux, and metabolism. HNF4α is highly expressed in the liver and regulates numerous genes involved in energy metabolism. Decreased cholesterol, HDL cholesterol, and triglyceride levels found in the plasma of mice lacking hepatic HNF4α [10] suggests that HNF4α plays an important role in controlling hepatic lipid metabolism and transport. In our previous study, we found that HNF4α was sensitive to plasma insulin level and was downregulated by SREBPs in hyperinsulinemia diabetes [11] and hepatic SREBPs were downregulated by HNF4α overexpression in mice (unpublished data). Compared with knowledge from loss-of-function study [10], the one of the role HNF4α in gain-of-function animal models is limited. In the loss-of-function model of liver-specific HNF4α deficiency, authors attributed the decrease of the serum cholesterol to the reduced de novo cholesterol biosynthesis, VLDL secretion, and HDL biogenesis [10, 30]. However, our data revealed that hepatic HNF4α overexpression mainly affect the uptake of circulating cholesterol in rat liver. Overexpression of HNF4α in the liver lowered the plasma level of LDL-C but had little effect on HDL-C and triglyceride levels in plasma and liver. These data indicate that HNF4α overexpression might have a moderate cholesterol-lowering effect. Further dissection of genes pivotal in the control of hepatic cholesterol homeostasis in the rat liver revealed that HNF4α overexpression mainly enhanced the expression of genes involved in cholesterol uptake. This result is in accordance with those of a recent study of a similar gain-of-function mouse model [30].

The role of SR-BI in the hepatic hypocholesterolemic effect under hepatic HNF4α deficiency is still controversial: Hayhurst et al. attributed the cholesterol-lowering effect to increased hepatic SR-BI expression [10] while in an acute liver-specific loss of function model, Yin et al. demonstrated an elevation in liver SR-BI level [30]. In compensated for the mice data, we have provided in vivo and in vitro gain-of-function data on rats.

We found the mRNA level of LDLR and SR-BI but not ABCG5 were upregulated. Moreover, the fold upregulation of SR-BI is higher than LDLR in our in vivo data. It has been reported previously that HNF4α could interact with SREBP-2 on the sterol-response-element (SRE) of the LDLR promoter and hence modulate its transcriptional activity [28]. The hepatic upregulation of SR-BI was found to decrease LDL-C and HDL-C content in mice [31]. We further found overexpression of HNF4α upregulated the protein level of SR-BI in vivo. Our finding together with the data from the HNF4α-knockout mice research suggests that the regulation of SR-BI by HNF4α might be indirect. In this study, we showed that overexpression of HNF4α in isolated rat primary hepatocytes induced SR-BI, whereas inhibition of PPARγ diminished this effect, which suggests the involvement of HNF4α-induced PPARγ transactivation. In addition, rosiglitazone, a known PPARγ-agonist, significantly augmented the stimulation effect of HNF4α on SR-BI mRNA level. The synergistic effect of PPARγ-ligand and HNF4α overexpression gave us further evidence to support the idea that the elevation of SR-BI induced by enhanced HNF4α expression level is due to the HNF4α-induced transactivation of PPARγ. These findings indicate that HNF4α induced upregulation with the involvement of PPARγ transactivation in liver might, at least in part, account for the cholesterol-lowering effect of hepatic HNF4α overexpression.

We noticed that there was a significant difference of basal glucose levels with hepatic HNF4α overexpression than the control rats. This might attribute to the elevation of important glucose metabolic genes, including Pepck, G6pc, and L-pk. They are known HNF4α-target genes and can be upregulated to maintain an appropriate circulating glucose level while fasting. However, we did not observe any interruption caused by hepatic HNF4α overexpression in the oral glucose tolerance test on rats. The role of hepatic HNF4α in glucose metabolism is worth future investigation.

Rat livers exposed to Ad-PPARγ showed significantly less fibrosis than did controls [32], and overexpression of HNF4α alleviated hepatic fibrosis in rats with bile duct ligation [33]. HNF4α may prevent hepatic stellate cell activation and thus ameliorate liver steatosis through transactivation of PPARγ. Because of evidence of the interaction between PPARγ2 and HNF4α contributing to variation in insulin sensitivity in Mexican Americans [34], more work is needed to uncover the role of HNF4α and PPARγ in various human diseases.

In summary, we demonstrate that HNF4α transcriptionally upregulates PPARγ2 by directly binding to the PPARγ2 promoter. We also reveal the regulatory role of HNF4α in liver lipid metabolism by modulating the expression of cholesterol metabolism-associated genes. These effects were, at least in part, due to the upregulation of SR-BI through the activation of PPARγ by HNF4α.

Abbreviations

ABC:ATP-binding cassette transporter
Apob:Apolipoprotein B;
Apoc3:Apolipoprotein C-III;
G6p:Glucose-6-phosphatase;
Hes6:Hairy enhancer of split 6;
HNF4α:Hepatocyte nuclear factor 4α;
LCAT:Lecithin:cholesterol acyltransferase;
L-pk:L-pyruvate kinase;
Pepck:Phosphoenolpyruvate carboxykinase;
PLTP:Phospholipid transfer protein
PPAR:Peroxisome proliferator activated receptor;
SR-BI:Scanvege receptor B type I
Srebf1:Sterol regulatory element binding factor-1.

Acknowledgments

This work was supported by grants from the Major National Basic Research Grant of China (no. 2010CB912504), the National Natural Science Foundation of China (nos. 30821001, 30971063), and the “111” plan of China.