In vivo regulation of murine CYP7A1 by HNF-6: A novel mechanism for diminished CYP7A1 expression in biliary obstruction


  • Minhua Wang,

    1. Department of Surgery and Pediatric Surgery, University of Illinois at Chicago College of Medicine, Chicago, IL
    Search for more papers by this author
  • Yongjun Tan,

    1. Department of Biochemistry and Molecular Genetics, University of Illinois at Chicago College of Medicine, Chicago, IL
    Search for more papers by this author
  • Robert H. Costa,

    1. Department of Biochemistry and Molecular Genetics, University of Illinois at Chicago College of Medicine, Chicago, IL
    Search for more papers by this author
  • Ai-Xuan L. Holterman

    Corresponding author
    1. Department of Surgery and Pediatric Surgery, University of Illinois at Chicago College of Medicine, Chicago, IL
    2. Department of Biochemistry and Molecular Genetics, University of Illinois at Chicago College of Medicine, Chicago, IL
    • Department of Surgery/Pediatric Surgery, University of Illinois at Chicago, 840 S. Wood Street, M/C 958, Chicago, IL 60612
    Search for more papers by this author
    • fax: (312) 355-1497


Disruption of the enterohepatic bile acid circulation during biliary tract obstruction leads to profound perturbation of the cholesterol and bile acid metabolic pathways. Several families of nuclear receptor proteins have been shown to modulate this critical process by regulating hepatic cholesterol catabolism and bile acid synthesis through the transcriptional control of cholesterol 7-α hydroxylase (CYP7A1). Hepatocyte nuclear factor (HNF) 6 (also known as OC-1) is a member of the ONECUT family of transcription factors that activate numerous hepatic target genes essential to liver function. We have previously shown that hepatic expression of mouse HNF-6 messenger RNA (mRNA) and protein significantly decrease following bile duct ligation. Because CYP7A1 contains potential HNF-6 binding sites in its promoter region, we tested the hypothesis that HNF-6 transcriptionally regulates CYP7A1. Following bile duct ligation, we demonstrated that diminished HNF-6 mRNA levels correlate with a reduction in CYP7A1 mRNA expression. Increasing hepatic levels of HNF-6 either by infection with recombinant adenovirus vector expressing HNF-6 cDNA by growth hormone treatment leads to an induction of CYP7A1 mRNA. To directly evaluate if HNF-6 is a transcriptional activator for CYP7A1, we used deletional and mutational analyses of CYP7A1 promoter sequences and defined sequences −206/−194 to be critical for CYP7A1 transcriptional stimulation by HNF-6 in cotransfection assays. In conclusion, the HNF-6 protein is a component of the complex network of hepatic transcription factors that regulates the expression of hepatic genes essential for bile acid homeostasis and cholesterol/lipid metabolism in normal and pathological conditions. (HEPATOLOGY 2004;40:600–608.)

The molecular mechanism controlling bile acid formation and biliary circulation has been extensively studied because of the critical importance of hepatic cholesterol and bile acid metabolism in many important physiological processes such as digestion, lipid and vitamin uptake and distribution, steroid hormone and cell membrane function, and toxin elimination. The gene for cholesterol 7-α hydroxylase (CYP7A1) has been the focus of many studies because of the central role held by CYP7A1 in cholesterol and bile acid metabolic pathways. CYP7A1 is primarily regulated at the transcriptional level,1–3 and liver-specific transcription factors have been shown to be the dominant regulators of CYP7A1.4 The bile acid and lipid sensors belonging to the family of nuclear hormone receptors including hepatocyte nuclear factor (HNF) 4α have been implicated in this regulatory response through their interaction with the bile acid response element (BARE) I and BARE II of the CYP7A1 promoter.4 Several hepatocyte nuclear factors, including CCATT enhancer binding protein (C/EBP) β, albumin D-site binding protein (DBP),5 and HNF-3α (also known as FoxA1),6 also have been shown to have binding activities for CYP7A1.

HNF-6 (also known as OC-1) belongs to the family of ONECUT transcription proteins that transcriptionally activate target genes by using the bipartite ONECUT–homeodomain sequence to localize the HNF-6 protein to the nuclear compartment and bind to specific DNA sequences of target gene promoters.7–9 It is part of the network of hepatocyte-enriched transcription factors that specify the liver-differentiated phenotype by regulating the transcription of hepatic target genes.10In vivo, HNF-6 regulates hepatic glucose transporter 2 gene expression11 and the bile duct cell proliferative response to biliary obstruction.12 Consistent with these findings, mouse genetic studies demonstrated that HNF-6−/− mouse embryos fail to develop a gall bladder and exhibit severe abnormalities in both extrahepatic and intrahepatic bile ducts.13 Furthermore, HNF-6 transcription factor expression in hepatocytes is increased in response to growth hormone administration.14 Hepatocyte and biliary expression of HNF-6 are diminished during biliary obstruction,12 suggesting that HNF-6 may also participate in the liver adaptive response to biliary injury by down-regulating the expression of target genes essential to differentiated liver function. Among those potential HNF-6–regulated target genes, CYP7A1 is a good candidate because its regulation is likely to be affected when enterohepatic bile acid circulation is disrupted during biliary obstruction.

To define the role of HNF-6 in CYP7A1 transcriptional control, we first analyzed the steady-state messenger RNA (mRNA) expression of HNF-6 and CYP7A1 during bile duct ligation (BDL) and found that both genes share a parallel pattern of mRNA down-regulation. Using growth hormone and recombinant adenovirus-mediated overexpression of HNF-6, we showed that in vivo induction of HNF-6 mRNA activates CYP7A1 expression. To directly evaluate if HNF-6 is a transcriptional activator for CYP7A1, we used deletional and mutational analyses of CYP7A1 promoter sequences and defined sequences −206 through −194 to be critical for CYP7A1 transcriptional stimulation by HNF-6 in cotransfection assays. Electrophoretic mobility shift assay (EMSA) confirms that HNF-6 protein indeed binds to oligonucleotides bearing the corresponding HNF-6 binding sites for CYP7A1. These findings suggest that HNF-6 is an important in vivo transcriptional regulator of CYP7A1 and that it maintains and modulates bile acid biosynthesis in both the normal and injured cholestatic liver. This report demonstrates that HNF-6 hepatocyte nuclear factor is a component of the complex network of hepatic transcription factors that participate in the regulation of bile acid homeostasis and cholesterol/lipid metabolism genes.


CYP7A1, cholesterol 7-α hydroxylase; HNF, hepatocyte nuclear factor; BARE, bile acid response element; C/EBP, CCATT enhancer binding protein; DBP, albumin D-site binding protein; mRNA, messenger RNA; BDL, bile duct ligation; EMSA, electrophoretic mobility shift assay; GH, growth hormone; PBS, phosphate-buffered saline; RPA, ribonuclease protection assay; CMV, cytomegalovirus; nt, nucleotide; RNase, ribonuclease.

Materials and Methods

Animal Procedures.

Male CD1 mice (6–8 weeks old) were kept in a 12-hour light/12-hour dark cycle with free access to standard chow and water. Animals were sacrificed between 11 A.M. and 2 P.M. and whole livers were collected for total RNA and nuclear protein extraction. All animals received humane care according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals by the National Academy of Sciences (National Institutes of Health publication no. 86-25, revised 1985).


Following intraperitoneal administration of 0.24% Avidin (15 μL/g body weight), the common bile duct was divided near the liver hilum between 2 silk ligatures (n = 6). In sham-operated mice (n = 4), the duct was exposed without ligature.

Growth Hormone Administration.

Mice received an initial intraperitoneal injection of human recombinant growth hormone (GH) (somatropin [Norditropin], Novo Nordisk Pharmaceuticals Inc., Princeton, NJ) at 4 μg/g body weight for the priming dose followed 2 hours later with 3 μg/g body weight every 4–6 hours (n = 4) for 24 hours. Control mice received PBS (n = 3).

Adenovirus Infection.

Mice underwent tail vein injection of 200 μL of 2 × 1011 particles of recombinant adenovirus vectors expressing bacterial LacZ (AdLacZ) (n = 3) or mouse HNF-6 (AdHNF-6) cDNA (n = 4) for 24 hours. The construction and preparation of the replication-defective AdHNF-6 or AdLacZ have been described previously.11, 15, 16

Ribonuclease Protection Assay.

Total RNA was prepared from whole mouse liver using RNA-STAT-60 (Tel-Test “B” Inc., Friendswood, TX). Ten to twenty micrograms of total RNA were hybridized with [α-32P]UTP-labeled antisense RNA probes synthesized from the appropriate plasmid templates and digested with RNase ONE (Promega, Madison, WI) as described previously.17 Syntheses for antisense mouse HNF-6, HNF-4α, C/EBPα, FoxA2 (also known as HNF-3β), insulin-like growth factor 1, and cyclophilin riboprobes have been described.18 The mouse CYP7A1 ribonuclease protection assay (RPA) probe plasmid was obtained via reverse transcription polymerase chain reaction of mouse liver RNA (using primers 5′-GCA AGG ATC CTA CTT CTG CGA AGG CAT TTG G-3′ and 5′-GCC GGA ATT CAA ACA TCA CTC GGT AGC AGA A-3′) and cloned into the pBluescript SK(+) template, and an antisense RNA probe was synthesized from a BamHI digested template using T7 RNA polymerase. Expression levels were determined after exposure of the gel to phosphoimaging screens overnight. The signals were scanned with a Storm 860 phosphoimager, quantitated with the ImageQuant program (Amersham Biosciences, Piscataway, NJ), and normalized to cyclophilin mRNA levels.


In vitro–expressed HNF-6 protein was obtained by cloning the full-length recombinant mouse HNF-6 cDNA8 into pET28a(+) (Novagen, EMD Biosciences, San Diego, CA). The His-tagged HNF-6 fusion proteins expressed in Escherichia coli were purified with His Bind Quick Column (Novagen, EMD Biosciences) according to the manufacturer's instructions. Crude nuclear protein extracts were prepared from AdHNF-6–infected HepG2 cells; protein concentrations were determined using the Bradford method (Bio-Rad, Hercules, CA). The proteins were used for EMSA with double-stranded oligonucleotides containing HNF-6 binding sites that were radioactively end-labeled with [γ-32P]adenosine triphosphate and T4 polynucleotide kinase as described previously.15 Oligonucleotides with 5′BamHI or BglII overhang were synthesized from mouse CYP7A1 promoter containing the HNF-6 −206 to −194 binding sites (sequences 5′-GAT CCT TCG GCT TAT CGA CTA TTG CAG CTC TCA-3′ and 5′-GAT CTG AGA GCT GCA ATA GTC GAT AAG CCG AAG-3′ [binding sites are underlined]) (Integrated DNA Technologies, Coralville, IA). A previously well-characterized HNF-6 binding site from the −185/−144 glucose transporter 2 promoter region was used for positive control.11 The HNF-6 protein–DNA complexes were resolved from the unbound probe via electrophoresis on a nondenaturing gel and detected with autoradiography. Specificity of the binding was evaluated via coincubation with 100-fold molar excess of unlabeled oligocompetitor, mouse HNF-6 specific-antibody,18 or preimmune serum.

Expression Plasmid, Reporter Plasmid.

Expression vectors consisting of the cytomegalovirus (CMV) promoter driving the expression of mouse HNF-6 (pCMV-H6) have been described previously.8 The intact promoter construct pCYP-918; the deletion constructs pCYP-309, pCYP-193, and pCYP-147; and mutation constructs pCYP-309m1, m2 and m1, 2 (Table 3) were generated using polymerase chain reaction amplification of mouse genomic DNA (GenBank accession number GI 28495235, currently renamed GI 38078136). The polymerase chain reaction–amplified mouse CYP7A1 promoter regions were cloned into the pGL3-Basic firefly luciferase reporter plasmid (Promega). The following sense and antisense primers were used for polymerase chain reaction amplification of the designated mouse CYP7A1 promoter region: nucleotide (nt) −918 primer, 5′-CCG CTC GAG CTT CAA AGA GTT CCT GGA AC-3′; nt −309 primer, 5′-CGG GGT ACC TGA GCT CTT CTG TAG TGT G-3′; nt −193 primer, 5′-CGG GGT ACC GCA GCT CTC TGC TTG TTC TG-3′; nt −147 primer, 5′-CGG GGT ACC GTT CAA GGC CAG ATA ATG C-3′; and antisense nt +47 primer, 5′-GGA AGA TCT GGA GAA TCT GTG CTT AGC AA-3′. The primers for the mutation constructs were: nt −203 EcoRI-site antisense primer, 5′-CCG GAA TTC AAG CCG AAG GTC TGT CCC TC-3′; nt −199 EcoRI-site sense primer, 5′-CGG GAA TTC TAT TGC AGC TCT CTG CTT G-3′; nt −156 XbaI-site antisense primer, 5′-GCT CTA GAC TTC TCA GAA GAG GCT CCA G-3′; and nt −152 XbaI-site sense primer, 5′-GCT CTA GAT TAG TTC AAG GCC AGA TAA TGC-3′. All constructs were sequenced prior to use.

Cell Culture and Transient Transfection.

HepG2 cells were grown in six-well plates and cotransfected via the Fugene-6 reagent (Roche, Indianapolis, IN) with CMV-H6 or CMV control expression vectors (400 ng) and the CYP7A1 promoter firefly luciferase reporter constructs (1.6 μg). The CMV-Renilla luciferase plasmid was used to normalize for transfection efficiency (25 ng). Protein extracts were prepared following 24–36 hours of transfection. Luciferase activity was determined in triplicate from 4 separate experiments using the Dual-Luciferase Assay System (Promega). The mean fold induction of CYP7A1 promoter-driven luciferase activity by pCMV-H6 was calculated relative to luciferase values from CYP7A1 promoter transfected with pCMV control plasmid. Promoter-driven luciferase activities from the CYP7A1 deletion plasmids were normalized to luciferase activity of the intact promoter construct pCYP-918 cotransfected with pCMV.


Data are shown as mean ± SD. Intergroup differences were evaluated using Student's t test with the Analysis ToolPak in Microsoft Excel 2000 (Seattle, WA). A P value of less than .05 was considered significant.


HNF-6 and CYP7A1 mRNA Are Both Down-regulated During BDL.

Previous studies demonstrated that both hepatocyte and biliary expression of HNF-6 mRNA and protein are diminished during biliary obstruction,12 suggesting that HNF-6 may also participate in the liver transcriptional response to biliary injury. We therefore subjected CD1 mice to BDL and sham operation and sacrificed the mice 24 hours following the operation. Total liver RNA was prepared for ribonuclease (RNase) protection assays using antisense RNA probes specific to HNF-6 and CYP7A1 genes. Following 24 hours of BDL, RNase protection assays (RPA) showed that down-regulation of HNF-6 (Fig. 1A) is associated with similar reduction in the steady-state hepatic mRNA expression levels of CYP7A1 (Fig. 1B).

Figure 1.

Hepatic HNF-6 and CYP7A1 mRNA are down-regulated after 24 hours of BDL. Total RNA was isolated from the livers of sham-operated (n = 4) or BDL (n = 6) CD1 mice and analyzed along with control tRNA using RPA. Representative RPA gels for (A) HNF-6, (B) CYP7A1, and cyclophilin are shown. No hybridization signal was obtained with tRNA control (data not shown). The relative fold differences in HNF-6 and CYP7A1 mRNA levels between all sham-operated and BDL animals after normalization to cyclophilin mRNA levels show a statistically significant reduction in HNF-6 and CYP7A1 mRNA by fourfold and tenfold, respectively. BDL, bile duct ligation; HNF-6, hepatocyte nuclear factor 6; CYP7A1, cholesterol 7-α hydroxylase; Cyclo, cyclophilin; mRNA, messenger RNA.

AdHNF-6–Mediated Hepatic Overexpression of HNF-6 Leads to Up-regulation in CYP7A1 mRNA.

To directly examine the effect of HNF-6 overexpression on CYP7A1, we infected mouse livers with the recombinant adenovirus that uses the CMV promoter to drive expression of either control bacterial LacZ (AdLacZ) or the mouse HNF-6 cDNA (AdHNF-6). Total liver RNA were prepared for RNase protection assays after 24 hours of infection. Figure 2 shows that AdHNF-6–mediated increase in hepatic levels of HNF-6 leads to marked up-regulation of CYP7A1 mRNA relative to AdLacZ control (more than fourfold). Published data have shown that AdHNF-6 infection has no effect on hepatic expression of endogenous FoxA2, HNF-4α, nor C/EBPα or C/EBPβ.11 It is therefore likely that AdHNF-6–mediated induction of hepatic CYP7A1 expression results from the direct effect of increased HNF-6 levels. Taken together, these in vivo experiments suggest that increased levels of HNF-6 alone can transcriptionally activate the endogenous mouse CYP7A1 gene.

Figure 2.

Increased hepatic expression of HNF-6 mRNA by recombinant adenovirus expressing HNF-6 cDNA leads to hepatic CYP7A1 mRNA induction. Representative RPA gels of total liver RNA from control AdLacZ (n = 3) or AdHNF-6 (n = 4) infected mice show that HNF-6 mRNA induction in AdHNF-6 infected mice leads to more than fourfold up-regulation in CYP7A1 mRNA. AdLacZ, adenovirus expressing bacterial LacZ; AdHNF-6, adenovirus expressing hepatocyte nuclear factor 6.

GH-Mediated Hepatic Overexpression of HNF-6 Leads to Up-regulation in CYP7A1 mRNA.

Growth hormone administration has been shown to directly induce hepatic HNF-6 mRNA expression in hepatocytes through the signal transducer and activator of transcription 5 binding site on the HNF-6 promoter region.14 To evaluate the in vivo effect of GH-mediated increase in HNF-6 levels on CYP7A1 mRNA, we performed either GH or PBS injections for 24 hours. RNase protection assays showed that GH administration caused a twofold increase in hepatic levels of insulin-like growth factor 1 relative to PBS injection, confirming the efficacy of GH treatment (Fig. 3A). This increase in response to GH exposure is associated with similar twofold stimulation in hepatic levels of HNF-6 mRNA (Fig. 3B). GH-mediated HNF-6 increase, in turn, is associated with a fivefold up-regulation of hepatic expression of CYP7A1 (Fig. 3C) mRNA.

Figure 3.

GH treatment leads to induction of insulin-like growth factor 1, HNF-6, and CYP7A1. Representative RPA gels of mice treated for 24 hours with intraperitoneal injection of PBS control (n = 3) or GH (n = 4) show induction of (A) insulin-like growth factor 1 and (B) HNF-6 with GH treatment. (C) HNF-6 induction, in turn, was associated with a fivefold up-regulation in CYP7A1 mRNA. GH, growth hormone; IGF-1, insulin-like growth factor 1.

In hypophysectomized mice, a single dose of GH has been reported to lead to an up-regulation of FoxA2 and HNF-4α and inhibition of C/EBPα.14, 19 It has been proposed that GH induction of HNF-6 occurs via FoxA2 or HNF-4α transcriptional activation of HNF-6, or by GH effect on diminishing the repressive effect of C/EBPα on HNF-6 transcription.20 HNF-6 in turn can transcriptionally regulate HNF-4α,21 suggesting the possibility that CYP7A1 transcriptional induction during GH treatment may be directly mediated via HNF-4α through GH-mediated increase in hepatic expression of HNF-6. Using wild-type CD1 mice, however, GH has a minimal effect on the expression levels of FoxA2 (Fig. 4A), HNF-4α (Fig. 4B), and C/EBPα mRNA (Fig. 4C). Together with the AdHNF-6 experiments, our results suggest that HNF-6 induction of CYP7A1 mRNA is likely to be directly mediated by the increased hepatic expression of HNF-6, and not FoxA2, HNF-4α, or C/EBPα.

Figure 4.

HNF-6 mRNA up-regulation by GH treatment is not mediated by FoxA2, HNF-4α, or C/EBPα. Representative RPA gels show minimal differences in (A) FoxA2, (B) HNF-4α, and (C) C/EBPα mRNA levels between mice treated with PBS (n = 3) or GH (n = 4). GH, growth hormone; IGF-1, insulin-like growth factor 1; HNF-4α, hepatocyte nuclear factor 4α; C/EBPα, CCATT enhancer binding protein α; NS, not significant.

The CYP7A1 Promoter Region at nt −206/−194 Corresponds to the HNF-6 Responsive Element.

The 1-kb upstream sequences of the mouse CYP7A1 promoter region (GenBank accession number GI 28495235, currently renamed GI 38078136) were analyzed for the presence of HNF-6 consensus binding motif (DHWATTGAYTWWD where D is not C, H is not G, W = A or T, Y = C or T). Table 1 shows the three potential HNF-6 sites at nt −893/−881, nt −206/−194, and nt −159/−147. The DNA sequence alignment (using ClustalX 1.81) for the CYP7A1 mouse, rat, and human promoter regions (nt −1 to −249 for mouse) in Table 2 shows the binding sites for the two corresponding potential HNF-6 and other regulatory elements. Of note, the −206/−194 putative HNF-6 binding site partially overlaps with one of the DBP and C/EBPβ functional binding sites at the nt −208/−185 region of the CYP7A1 promoter,5 and the −159/−147 HNF-6 site also overlaps with sequences corresponding to the mouse HNF-4α binding site to the DR1 (direct repeat 1) site at nt −155/−143 sequence of the BARE II region.6

Table 1. DNA Sequences of HNF-6 Potential Binding Sites for CYP7A1 Promoter
  1. NOTE. Lower-case letters represent nucleotides that deviate from the HNF-6 binding sequence. Nucleotide abbreviations in the HNF-6 DNA binding consensus sequence are as follows: D is not C, H is not G, W = A or T, and Y = C or T.

Table 2. DNA Alignment of Mouse, Rat, and Human CYP7A1 Promoter Regions Corresponding to Mouse nt −1 to −249 Sequence
inline image
Table 3. Mouse CYP7A1 Promoter Construct pCYP-309 and Its Derivatives
  1. NOTE. Nucleotide position and sequences of the putative HNF-6 consensus binding sites of the pCYP-309 promoter constructs and the mutated nucleotides (underlined bold characters) of the pCYP-309m1, pCYP-309m2, and pCYP-309m1,2 derivatives are shown.

nt −309......  −206.......... −194....−159.......... −147........+47 
............  CTT ATC GAC TAT T    ....AAG ATG GAC TTA   G ....pCYP-309
............  CTT GAA TTC TAT T    ....AAG ATG GAC TTA    G ....pCYP-309m1
............  CTT ATC GAC TAT T    ....AAG TCT AGA TTA    G ....pCYP-309m2
............  CTT GAA TTC TAT T    ....AAG TCT AGA TTA    G ....pCYP-309m1,2

To further define the HNF-6 binding sites, we generated luciferase reporter gene constructs containing various deletions of the pCYP-918 promoter sequence as well as site-directed mutations in the HNF-6 binding sites. These deletion constructs included pCYP-918 (spanning the nt −918/+47 sequence, which contains all three of the HNF-6 potential binding sites), pCYP-309 (which contains the two proximal binding sites), pCYP-193 (which contains the HNF-6/HNF-4α site), and pCYP-147 (which has no HNF-6 binding site). Table 3 shows the sequences of the pCYP-309 derivatives bearing mutations at the nt −206/−194 (pCYP-309m1), nt −159/−147 (pCYP-309m2), or both nt −206/−194 and nt −159/−147 sites (pCYP-309m1, 2). Each of the CYP7A1 promoter luciferase constructs was transiently cotransfected with either pCMV control or HNF-6 expression vectors (pCMV-H6) into HepG2 cells. Protein extracts were prepared 24–36 hours after transfection, and luciferase activities were measured using the Dual-Luciferase Assay System (Promega).

As shown in Fig. 5, mutation of the −206/−194 and/or −159/−147 site lead to a sharp reduction in CYP7A1 basal promoter activity, confirming previously demonstrated importance of these sequences and their corresponding transcription factors to basal promoter function.5, 6, 22 Cotransfection of pCMV-H6 and promoter constructs with the intact −206/−194 site (pCYP-918 [lane b] and pCYP-309 [lane d]) led to a three- to fivefold induction in CYP7A1 promoter expression levels compared with pCMV control transfections. Mutation of the −206/−194 site with pCYP-309m1 (lane f) or pCYP-309m1, 2 (lane j) abolished the stimulatory effect of HNF-6 on CYP7A1 promoter activity, while mutation of the −159/−147 region (pCYP-309m2, lane h) alone maintained HNF-6 enhancing effect on reporter gene activity (tenfold stimulation over pCMV cotransfection). Together, these data define the −206/−194 sequence as an HNF-6–responsive element for CYP7A1.

Figure 5.

An HNF-6 responsive element resides at nt −206/−194 of the CYP7A1 promoter region. HepG2 cells were transfected with CMV or HNF-6 expression vector CMV-H6 and CYP7A1 wild-type reporter construct pCYP-918 bearing the three potential HNF-6 binding sites and its deletion derivatives pCYP-309, pCYP-193, pCYP-147, or the pCYP-309 mutated constructs: pCYP-309m1 and pCYP-309m2, which contain the mutated −206/−194 and −159/−147 sequences, respectively, and pCYP-309m1,2, which contains both mutations (mutated sites are represented by crossed boxes). Firefly luciferase levels are normalized to Renilla luciferase control as described in Materials and Methods. The reported luciferase activities are expressed as the fold change of pCYP7A1 promoter activity relative to the activity of the pCYP-908 promoter cotransfected with empty CMV expression vector. For each of the pCYP constructs, the fold stimulation of cotransfection with CMV-H6 relative to CMV is shown on the right side of the figure. Results of cotransfection assay of the pGL3-Basic firefly luciferase reporter plasmid with pCMV or pCMV-H6 vector were similar to those obtained with the pCYP-147 construct (data not shown). CMV, cytomegalovirus; CMV-H6, cytomegalovirus with hepatocyte nuclear factor 6 expression.

Having mapped the sequences essential for HNF-6 transcriptional activity, we performed EMSA using in vitro translated HNF-6 protein (Fig. 6A) or HNF-6–enriched nuclear protein extracts from AdHNF-6–infected HepG2 (Fig. 6B). EMSA experiments showed no HNF-6 protein/DNA complex to labeled double-strand oligonucleotides spanning the −159/−147 CYP7A1 promoter region (lane 10), but did show efficient binding of the HNF-6 protein to previously well-characterized HNF-6 binding sites from the −185/−144 glucose transporter 2 promoter region11 (lane 2) as well as the −206/−194 oligonucleotides corresponding to the HNF-6 response element (lane 6). In both cases, the protein/DNA complexes are competed off by excess unlabeled competitor (lanes 3 and 7) or HNF-6 specific antibody (lanes 4 and 8). EMSA using HNF-6 nuclear extracts confirms HNF-6 binding to the −206/−194 promoter region (lane 14). This is abrogated in the presence of excess cold probe or HNF-6 antibody (lanes 15, 16) but not preimmune sera (lane 17).

Figure 6.

HNF-6 recombinant protein binds to the nt −206/−194 CYP7A1 promoter site. EMSA using 300 ng of in vitro–translated (A) HNF-6 protein or (B) nuclear extract protein prepared from AdHNF-6 infected HepG2 cells showed HNF-6 protein binding to the 32P-labeled double-stranded −185/−144–bp glucose transporter 2 positive control probe (lane 2) and the −206/−194–bp CYP7A1 promoter sequences (lanes 6, 14; shown with arrows) but not the −159/−147–bp regions (lane 10). The HNF-6 protein–DNA complexes were competed off by homologous unlabeled oligonucleotides (lanes 3, 7, 15) and HNF-6 specific antibody (lanes 4, 8, 16) but not preimmune serum (lane 17). Glut-2, glucose transporter 2; NE, nuclear extract.

In summary, in support of our in vivo data showing that HNF-6 can induce CYP7A1 mRNA expression, these in vitro experiments demonstrate that HNF-6 protein can bind to the CYP7A1 nt −206/−194 promoter sequence and mediate transcriptional activation of the CYP7A1 promoter.


The importance of proper cholesterol and bile acid homeostasis in biological systems is reflected by the complex and redundant ways in which CYP7A1 is regulated in normal and pathological conditions. When the enterohepatic circulation is disrupted during cholestasis (e.g. biliary obstruction–, endotoxin-, or estrogen-induced cholestasis), toxic bile acids accumulate and proinflammatory cytokines (particularly tumor necrosis factor α) are released.23 The bile acid–responsive farsenoid X nuclear receptor provides negative feedback regulation of CYP7A1 by inducing short heterodimer protein–mediated inhibition of CYP7A1 promoter factor (also known as liver receptor homologue 1 or fetoprotein transcription factor) binding to the highly conserved bile acid responsive element BARE II region of CYP7A1 promoter.4, 24, 25 An alternative mechanism for the feedback inhibition of CYP7A transcription is through the effects of bile acids and inflammatory cytokines tumor necrosis factor α and interleukin 1β in recruiting the mitogen-activated protein kinase cascade (JNK pathways),26–29 resulting in the activation of the short heterodimer protein repressor protein26 or the inhibition of HNF-4α binding to the DR1 sequence of the CYP7A1 promoter BARE II region.28

HNF-6 hepatocyte transcription factor belongs to the network of hepatocyte-enriched nuclear factors that have potential binding sites for numerous target genes essential to liver differentiation and function. In vivo, HNF-6 also participates in the hepatic injury repair response to bile duct ligation by regulating the bile duct cell proliferative reponse to biliary obstruction.12 In this study, we demonstrate that HNF-6 can induce CYP7A1 gene expression both in mouse liver and in HepG2 cells in cotransfection assays. Using EMSA and transient cotransfection assay, we showed that HNF-6 binds to the −206/−194 site of the CYP7A1 promoter region and activates the luciferase reporter gene. Cotransfection of the reporter constructs bearing an intact −204/−196 HNF-6 site (which also partially overlaps with a DBP and C/EBPβ binding site) with the HNF-6 expression vector induced CYP7A1 promoter activity, while mutation of the overlapping HNF-4α/HNF-6 site (nt −159/−147) in the DR1 sequence of the BARE II region did not inhibit the transcriptional stimulation of the reporter gene by the HNF-6 expression vector. These results suggest that HNF-6 directly activates CYP7A1 transcription at the −204/−196 binding site independently of DBP, C/EBPβ, or HNF-4α. Although HNF-4α transcriptional activity at the direct repeat 1 site6, 22 and DBP and C/EBPβ activities at the nt −208/−185 site5 have been shown to affect CYP7A1 basal promoter function, the possibility of HNF-6 and HNF-4α interaction and/or HNF-6, DBP, and C/EBPβ interaction at their respective binding sites and the contribution of those interactions to CYP7A1 basal transcriptional activity has yet to be evaluated. Of note, transient transfection assays also showed that cotransfection of mouse FoxA1 and HNF-6 expression vectors has no effect on HNF-6–stimulated pCYP-309 reporter gene activity, while FoxA2 and HNF-6 cotransfection leads to a significant inhibition of HNF-6–induced CYP7A1 promoter activity at the −206/−194 site (data not shown). This is consistent with previous reports that in the context of HNF-6–dependent target genes, FoxA1 does not efficiently interact nor synergize with HNF-6, while FoxA2 can inhibit HNF-6 transcriptional activity by impairing HNF-6 DNA binding.9

Consistent with the in vitro data, downregulation of HNF-6 during liver injury after biliary obstruction is associated with a reduction in CYP7A1 mRNA levels, while increasing hepatic levels of HNF-6 by GH administration or by infection with the adenoviral HNF-6 expression vector leads to coinduction of HNF-6 and CYP7A1 mRNA. GH is an important in vivo transcriptional regulator of HNF-614 through GH–mediated signal transducer and activator of transcription 5 effect on HNF-6 promoter activity. Among many of the effects of GH on other downstream target genes, GH may also indirectly control hepatic gene expression by transcriptionally inducing HNF-4α and FoxA214 or suppressing C/EBPα19 transcription factors. Accumulating evidence suggests that maintenance of hepatocyte-specific expression of the HNF-6, HNF-4α, FoxA2, and C/EBPα transcription factors involves cross-regulation by one or more unrelated liver-enriched transcription factors.9, 14, 19, 21 They have thus been proposed to participate in a complex network of autoregulatory loops to control hepatic gene expression.20 The fact that we observed no change in hepatic expression of FoxA2, HNF-4α, or C/EBPα following GH administration to wild-type CD1 mice suggests that GH-mediated stimulation of CYP7A1 is likely mediated through increased HNF-6 levels and not through the HNF-4α protein. Furthermore, CYP7A1 mRNA induction by adenoviral-mediated hepatic overexpression of HNF-6 occurs without changes in the levels of endogenous FoxA2, HNF-4α, HNF-1α, C/EBPα, or C/EBPβ transcription factors,11 a finding consistent with a critical role of HNF-6 in mediating the in vivo induction of the CYP7A1 gene. Taken together with our HNF-6 cotransfection data on CYP7A1 basal promoter activity, published results on the role of the FoxA1, HNF-4α, C/EBPβ, and DPB hepatocyte nuclear factors on CYP7A1 gene regulation suggest that simultaneous binding of many of these transcription factors to distinct regulatory DNA sites might be important in providing synergistic transcriptional control of hepatic target genes. This underscores the physiological significance of combinatorial interaction among different hepatocyte transcription factors in modulating hepatic function.10

These in vitro and in vivo data demonstrate for the first time a new molecular mechanism of transcriptional regulation of the CYP7A1 gene that involves the hepatic HNF-6 transcription factor. This study also provides an alternative mechanism for CYP7A1 transcriptional inactivation during cholestasis through the loss of CYP7A1 transcriptional maintenance as a result of diminished expression of HNF-6. Further examination of the −206/−194 HNF-6 binding site by alignment of the mouse, rat, and human CYP7A1 promoter sequences shows that this region remains highly conserved, underscoring the evolutionary significance of this HNF-6–responsive element in CYP7A1 promoter activity.

In mice, GH administration increases Cyp7A1 enzymatic activity and lowers plasma cholesterol.30 A similar effect of GH in lowering serum cholesterol levels—reducing hyperlipidemia and thus improving cardiovascular risks—has been reported in patients on GH replacement therapy.31 Because the effect of growth hormone on the induction of CYP7A1 expression is in part mediated by HNF-6, these data provide a novel molecular basis for CYP7A1 regulation by GH through HNF-6. This implies that an HNF-6–based mechanism may be extended to a subset of GH-diverse effects on biological processes such as development, differentiation, growth, and cellular metabolism.

In summary, the biological role of HNF-6 in maintaining differentiated hepatic function includes cholesterol catabolism and bile acid biosynthesis through the transcriptional regulation of CYP7A1. During biliary obstruction, along with other liver transcription factors such as the nuclear hormone receptors and HNF-4α, HNF-6 participates in the complex liver repair response to cholestasis with the end result to down-regulate CYP7A1 transcription as part of a protective response to diminish the intracellular bile acid load and minimize hepatocellular injury. Further elucidation of HNF-6 biological function during cholestasis will further our understanding of the hepatic repair mechanism to injury and help design hepatoprotective therapeutic strategies for a broad spectrum of liver diseases including cholestases and cholangiopathies.


We thank Dr. V. Kalinichenko for critical evaluation of the manuscript.