The transport and metabolism of organic cationic drugs and endobiotics are essential tasks performed by hepatocytes. The SLC22A family of polyspecific organic cation transport (OCT) proteins includes important mediators of facilitated transport of organic cations (Koepsell et al., 2003; Jonker and Schinkel, 2004). The human organic cation transporter-1 (hOCT1; SLC22A1) is primarily expressed at the basolateral membranes of human hepatocytes and mediates the hepatic uptake of a variety of cationic drugs and endogenous substrates, such as dopamine, serotonin, metformin, cimetidine, verapamil, ganciclovir, and choline, from portal blood (Jonker and Schinkel, 2004). A net-positive charge is not absolutely required for OCT1 substrates, because hOCT1 has been shown to also transport certain anionic prostaglandins (Kimura et al., 2002). Studies with Oct1-null mice support the notion that Oct1 is the major physiological hepatic uptake system for small organic cations (Jonker et al., 2003).

The hOCT1 gene is located on chromosome 6 (Koehler et al., 1997), contains seven exons and six introns, and can produce several alternatively spliced mRNA isoforms (Hayer et al., 1999). Characteristic of the members of the SLC22 family, the hOCT1 protein contains 12 predicted α-helical transmembrane domains and a long hydrophilic loop between transmembrane domains 1 and 2 (Koepsell et al., 2003).

Changes in the expression level of hOCT1 are likely to influence the rate of hepatic clearance of hOCT1 substrates from sinusoidal blood. Although the signaling pathways that regulate the function of the hOCT1 protein have been recently studied (Ciarimboli et al., 2004), the mechanisms that control the expression level of the hOCT1 gene have not been previously investigated. In this study, we have analyzed the 5′-flanking region of the hOCT1 gene and identified two putative binding sites for the liver-enriched hepatocyte nuclear factor (HNF)-4α (NR2A1) between nucleotides -1479 and -1441 upstream of the transcription initiation site. We have shown that HNF-4α transactivates the hOCT1 promoter via direct binding to the identified HNF-4α-response elements.

In states of cholestasis the expression of many basolateral uptake transporters for drugs and bile acids is reduced to prevent intracellular accumulation of potentially toxic substances in hepatocytes (Lee and Boyer, 2000; Kullak-Ublick et al., 2004). Thus, we wanted to study whether the expression of the hOCT1 gene is also regulated by bile acids in hepatocyte-derived cultured cells. Furthermore, we investigated whether the small heterodimer partner (SHP), a transcriptional repressor induced by bile acids (Goodwin et al., 2000; Lu et al., 2000), plays a role in the regulation of the hOCT1 gene.

Materials and Methods

Chemicals. GW4064 was a gift from Dr. Daniel Berger (GlaxoSmithKline, Uxbridge, Middlesex, UK). Restriction enzymes were purchased from Roche Diagnostics (Rotkreuz, Switzerland), and the LigaFast rapid ligation kit was from Promega Catalys (Wallisellen, Switzerland). FuGENE 6 (Roche Diagnostics) was used for all transfections at a ratio of 3 μl/μg DNA. The oligonucleotides were synthesized by Microsynth (Balgach, Switzerland). Other chemicals were purchased from Sigma-Aldrich (Buchs, Switzerland), unless stated otherwise.

Plasmids. The promoter fragments derived from the hOCT1 gene were created by PCR using PuReTaq Ready-to-Go PCR beads from Amersham Biosciences (Otelfingen, Switzerland). Human genomic DNA (BD Biosciences Clontech, Basel, Switzerland) was used as a template, and oligonucleotide primers with engineered restriction sites are listed in Table 1. Point mutations within the DR-2-like elements of the hOCT1 promoter were created using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). The sequence of all constructs was confirmed by DNA sequencing. The pcDNA3.1-HNF-4α and pCMX-SHP constructs were kind gifts from Dr. David Moore (Baylor College of Medicine, Houston, TX) and Dr. David Mangelsdorf (University of Texas Southwestern Medical Center, Dallas, TX), respectively. The reporter vectors pGL3basic and pSV-β-galactosidase were from Promega Catalys, and pcDNA3.1 was from Invitrogen (Basel, Switzerland).

Cell Culture. The human hepatoma cell line Huh7 (LGC Promochem, Molsheim Cedex, France) was cultured in RPMI 1640 medium (Sigma-Aldrich) supplemented with 10% fetal bovine serum (Sigma-Aldrich), 100 U/ml penicillin, and 100 μg/ml streptomycin (Invitrogen). Cells were cultured at 37°C in a humidified atmosphere containing 5% CO₂.

Transient Transfections and Reporter Assays. Huh7 cells were seeded in 48-well plates at a density of 1 × 10⁵ cells/well and cotransfected with 400 ng of the luciferase constructs, together with the indicated amounts of the effector expression plasmids. To normalize the total amount of DNA transfected, pcDNA3.1 vector was added. To control for transfection efficiency, 100 ng of the pSV-β-galactosidase reporter plasmid was cotransfected. Thirty-six hours after transfection, cells were harvested in Passive lysis buffer (Promega Catalys), and luciferase activities were determined using the Luciferase assay system (Promega Catalys) in a Lumat LB 9507-2 luminometer (Berthold Technologies, Regensdorf, Switzerland). β-Galactosidase activities were quantified by a chlorophenol red-β-d-galactopyranoside-based colorimetric assay in a microplate reader (Molecular Devices, Sunnyvale, CA). Reporter activities obtained for the empty pGL3basic corresponding to each test condition as well as for the test construct containing the test promoter in the control conditions are set to 1, and -fold activities are shown relative to this. All transfection experiments were performed three times, each experiment containing triplicate wells for each set of conditions. Results are expressed as mean -fold activities ± S.D.

Electrophoretic Mobility Shift Assays. Oligonucleotides used in EMSAs were designed to have 5′-GATC overhangs when annealed, allowing radioactive labeling by fill-in reactions. Fifty nanograms of annealed oligonucleotides was labeled in a 20-μl reaction containing 200 U of SuperScript II RNase H^- reverse transcriptase (Invitrogen), 1× First-Strand Buffer (Invitrogen), 10 mM dithiothreitol, 250 nM dGTP/dCTP/dTTP, and 2 μCi of [α-³²P]dATP (Amersham Biosciences). Unincorporated nucleotides were removed using Microspin G-25 columns (Amersham Biosciences). Then, 3.3 μg (the HNF-4α consensus probe) or 10 μg (the hOCT1 probes) of nuclear extracts prepared from confluent Huh7 cells using the NE-PER kit (Perbio Science, Lausanne, Switzerland) or 1 μl (the HNF-4α consensus probe) or 3 μl (the hOCT1 probes) of the HNF-4α protein synthesized using the TnT rabbit reticulocyte lysate-coupled in vitro transcription/translation system (Promega Catalys) was used for DNA binding reactions. Approximately 50,000 cpm (0.5-1.0 ng) of the radioactive probes was used per reaction. Protein-DNA complexes were formed in binding buffer [20 mM Tris-HCl, pH 8.0, 60 mM KCl, 2 mM MgCl₂, 12% (v/v) glycerol, 0.3 mM dithiothreitol, and 87.5 ng/ml poly(dI-dC)·poly(dI-dC)] in a total volume of 20 μl for 10 min at 30°C, with gentle shaking. In supershift experiments, 1 μg of the HNF-4α (H-171; Santa Cruz Biotechnology Inc., Santa Cruz, CA) antibody was added to the extracts 1 h before the addition of the probe and incubated at 4°C. In competition experiments, 50-fold molar excess of the competing oligonucleotides was added simultaneously with the radioactive probe. Immediately after the binding reactions, the samples were loaded onto a pre-electrophoresed 5% (acrylamide/bis, 30:1) native acrylamide gel and run at 200 V in 0.5× Tris borate-EDTA for 1.5 h. After the runs, the gels were fixed in 10% (v/v) acetic acid for 10 min, rinsed with distilled water, dried down onto Whatman DE81 paper under vacuum, and exposed to Kodak BioMax MR-1 film (Sigma-Aldrich) at -70°C.

RNA Isolation, Reverse Transcription, and Real-Time PCR. Cells were seeded in six-well plates and, after reaching 80% confluence, transfected with 3 μg of the indicated expression plasmids or empty vectors. Where indicated, 12 h after transfection, the medium was supplemented with 50 μM chenodeoxycholic acid (CDCA), 200 nM farnesoid X receptor (FXR) agonist GW4064, or the vehicle DMSO. Total RNAs were extracted 36 h after transfection using the TRIzol reagent (Invitrogen). Two micrograms of each RNA preparation was reverse-transcribed by random priming (reverse transcription system; Promega Catalys). Five microliters of each resulting cDNA from a final reaction volume of 100 μl was used for real-time PCR, which was performed on an ABI Prism 7700 sequence detection system (Applied Biosystems, Rotkreuz, Switzerland) using the Taq-Man gene expression assays Hs00427550_m1 (Applied Biosystems) for human OCT1 and Hs00222677_m1 (Applied Biosystems) for human SHP. Constitutively expressed 18S rRNA assayed with eukaryotic 18S rRNA endogenous control (Applied Biosystems) was used as an internal standard for sample normalization. Relative levels of the hOCT1 and hSHP mRNAs were calculated using the ΔΔC_T (comparative threshold cycle) method. Each test was performed as a triplicate, and all experiments were repeated three times. The levels of the hOCT1 and hSHP mRNAs are expressed relative to the control sample, which was set to 1.

Statistical Analysis. All data presentation indicating variance represent mean ± S.D. Data from reporter assays and real-time PCR were analyzed by one-way analysis of variance followed by post hoc analysis using Tukey's test.

Results

Identification of Putative HNF-4α-Response Elements in the hOCT1 Promoter. HNF-4α plays an indispensable role in liver-specific gene expression (Hayhurst et al., 2001) and regulates several genes encoding hepatic membrane transporters (Jung and Kullak-Ublick, 2003; Popowski et al., 2005; Zollner et al., 2005). We hypothesized that the human OCT1 gene may also be a target for HNF-4α-mediated regulation. Thus, we initially performed an in silico analysis on the hOCT1 promoter region between nucleotides -3852 and +116 (GenBank AL353625), using a NUBIScan algorithm, designed to identify binding motifs for nuclear receptors (Podvinec et al., 2002). We were particularly interested in motifs in the DR-1 and DR-2 format (a direct repeat of AGGTCA-like hexamers, separated by one or two nucleotides, respectively), because these have been shown to interact with HNF-4α by random binding site selection (Fraser et al., 1998). We could identify two DR-2-like elements, separated by 11 nucleotides. DR-2 (A) is located between the nucleotides -1479 and -1466, and DR-2 (B) is located between the nucleotides -1454 and -1441 (Fig. 1, top line), relative to the transcriptional start site. The hOCT1 DR-2 (A) element differs from the consensus DR-2 element (Fig. 1, bottom line) by three nucleotides and DR-2 (B) by two nucleotides.

HNF-4α Transactivates the hOCT1 Promoter. To test whether the DR-2 elements identified above confer activation by HNF-4α, we cloned the region of the hOCT1 gene between the nucleotide positions -3852 and +116 containing these elements. This hOCT1 promoter fragment was then subcloned into a luciferase reporter vector, and Huh7 cells were cotransfected with this reporter construct, either without or with increasing amounts of the expression plasmid for HNF-4α. As shown in Fig. 2, the hOCT1 (-3852/+116) promoter construct was strongly transactivated by exogenously expressed HNF-4α in a dose-dependent manner.

We then created progressive 5′ deletions of the -3852/+116 fragment and transfected reporter constructs harboring these promoter variants into cells with or without cotransfection of the expression construct for HNF-4α. The hOCT1 promoter construct containing the nucleotides from -2620 to +116 was as potently activated by HNF-4α as the -3852/+116 construct (Fig. 3). A further deletion to the nucleotide position -1356 or -880 largely abolished the activation, indicating that the critical region that HNF-4α acts upon lies between nucleotides -2620 and -1356. This is consistent with our hypothesis that the two DR-2 elements identified above mediate the HNF-4α-activation of the hOCT1 promoter, because these are located between the nucleotide positions -2620 and -1356. We note that the hOCT1 promoter region -880/+116 still remains weakly inducible by HNF-4α, indicating that although the main response maps to the hOCT1 promoter region -2620/-1356, there may be more proximal promoter elements that mediate a weak response to HNF-4α overexpression.

Both DR-2 Elements in the hOCT1 Promoter Are HNF-4α-Response Elements. To functionally confirm that the two DR-2 elements in the hOCT1 promoter mediate the response to HNF-4α, we introduced point mutations (Fig. 1, middle three rows) predicted to abolish DNA binding by HNF-4α (Fraser et al., 1998) into the hOCT1 (-2620/+116)luc reporter construct. The construct containing either the wild-type DR-2 elements (WT), individually mutated DR-2 elements (mut A and mut B), or both DR-2 elements mutated together (mut A+B) were transfected into Huh7 cells together with the expression plasmid for HNF-4α. The three promoter variants harboring either single or double mutations in DR-2 elements were markedly less responsive to exogenously expressed HNF-4α than the wild-type promoter construct (Fig. 4), implying that both DR-2 elements are required for maximal activation by HNF-4α. The construct containing mutations in both of the two DR-2 elements similar to the construct lacking both DR-2 elements [hOCT1 (-880/+116)luc] remain weakly inducible by HNF-4α, suggesting that HNF-4α may exert additional effects on the hOCT1 gene expression that do not depend on the two DR-2 elements identified in this study.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Putative HNF-4α-response elements within the hOCT1 promoter. Sequence and location of the two adjacent DR-2-like elements of the hOCT1 promoter identified as potential HNF-4α-response elements are shown in the top row. Combined and individual point mutations created within the DR-2 elements and used in experiments presented in Figs. 4 and 5 are shown in the three middle rows (the mutated bases shown in boldface). In the bottom row, consensus sequences of HNF-4α-response elements (DR-1 or DR-2) are shown aligned to the DR-2 elements of the hOCT1 promoter.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

HNF-4α transactivates the hOCT1 promoter in a dose-dependent manner. Huh7 cells were transfected with the reporter construct hOCT1(-3852/+116)luc or the promoterless reporter vector pGL3basic. With each reporter, increasing amounts (0, 100, 200, and 400 ng; from left to right) of the CMV promoter-driven expression plasmid for HNF-4α were cotransfected. The pcDNA3.1 vector was used to normalize the total amount of transfected CMV promoter-containing expression plasmid to 400 ng. To calculate the -fold induction by HNF-4α, the reporter activities obtained for pGL3basic either in the presence or absence of the HNF-4α expression plasmid and that obtained for the hOCT1 promoter construct in the absence of exogenously expressed HNF-4α were set to 1. Results are expressed as mean -fold activities ± S.D. Asterisk (*) denotes statistically significant difference (p < 0.001) between the indicated data points.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Mapping the HNF-4α-response elements within the hOCT1 promoter. Huh7 cells were transfected with one of the indicated hOCT1 promoter constructs or the pGL3basic vector. With each reporter, 100 ng of either the expression plasmid for HNF-4α or the pcDNA3.1 vector was cotransfected. To calculate the -fold induction by HNF-4α, the reporter activities obtained for pGL3basic, either in the presence or absence of the HNF-4α expression plasmid, and those obtained for the hOCT1 promoter constructs in the absence of exogenously expressed HNF-4α were set to 1. Results are expressed as mean -fold activities ± S.D. Asterisk (*) denotes statistically significant difference (p < 0.001) between the indicated data points.

HNF-4α Can Bind to Both the DR-2 Elements from the hOCT1 Promoter in Vitro. To confirm that HNF-4α can bind to the two DR-2 elements identified within the hOCT1 promoter, we performed EMSAs using nuclear extracts from Huh7 cells. A protein-DNA complex of the same mobility was formed on the radiolabeled probe that corresponds to the nucleotides -1479 and -1441 (Fig. 5A, lane 4) and contains the two DR-2 elements, as on the probe containing the consensus HNF-4α binding site (Fig. 5A, lane 2). The in vitro binding of HNF-4α to the probe derived from the hOCT1 promoter is weaker than it is to the probe containing the consensus element, reflecting the deviations of the two DR-2 elements from the ideal binding site for HNF-4α (Fig. 1). When the same point mutants, as used in the functional assays (Fig. 4, hOCT1 mut A+B shown in Fig. 1), were introduced into both the DR-2 elements, no protein-DNA complex was formed on the -1479/-1441 probe (Fig. 5A, lane 10). Individual mutations in either DR-2 (A) or DR-2 (B) element reduced the ability of HNF-4α to interact with the probe (Fig. 5A, lanes 6 and 8, respectively), suggesting that both elements contribute to the interaction of HNF-4α with the wild-type probe. The mutations in the DR-2 (A) element resulted in a more pronounced loss of HNF-4α DNA binding than those in the DR-2 (B) element, indicating that the former element binds HNF-4α more avidly. For each probe, the identity of the protein-DNA complex was confirmed by a complete abolition of the complex upon incubation with an HNF-4α-specific antibody (Fig. 5A, lanes 1, 3, 5, 7, and 9).

To further confirm the specificity of HNF-4α binding to the DR-2 elements, we performed EMSA competition assays using in vitro translated HNF-4α protein and 50-fold molar excesses of the competitor oligonucleotides listed in Table 1. Antibodies against HNF-4α produced supershifted complexes (Fig. 5B, lanes 1 and 3), indicating that, similar to the endogenous HNF-4α present in nuclear extracts, in vitro translated HNF-4α efficiently binds to the hOCT1 (-1479/-1441) probe (Fig. 5B, lane 4), in addition to the HNF-4α consensus motif (Fig. 5B, lane 2). As expected, the binding of HNF-4α to the radiolabeled hOCT1 (-1479/-1441) probe was completely abolished in the presence of an excess of the oligonucleotide containing the consensus sequence for HNF-4α (Fig. 5B, lane 5). The competitors corresponding to the DR-2 elements derived from the hOCT1 promoter were also effective in reducing HNF-4α binding, the DR-2 (A) element more so than the DR-2 (B) element (Fig. 5B, lanes 6 and 8, respectively). This confirms that HNF-4α interacts with the DR-2 (A) element with higher affinity than with the DR-2 (B) element. Consistent with the results shown in Fig. 5A, excess of the hOCT1 DR-2 oligonucleotides carrying point mutations (shown in Fig. 1) had little effect on the interaction between the hOCT1 (-1479/-1441) probe and HNF-4α (Fig. 5B, lanes 7 and 9).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Both DR-2 elements in the hOCT1 promoter function as HNF-4α-response elements. Huh7 cells were transfected with the reporter construct hOCT1(-2620/+116)luc containing either wild-type or single- or double-mutated DR-2 elements, the hOCT1(-880/+116)luc lacking the two DR-2 elements, or the promoterless reporter vector pGL3basic. Each reporter plasmid was cotransfected with 100 ng of either the HNF-4α expression plasmid or the empty pcDNA3.1 vector. To calculate the -fold induction by HNF-4α, the reporter activities obtained for pGL3basic, either in the presence or absence of the HNF-4α expression plasmid, and that obtained for the hOCT1 promoter constructs in the absence of exogenously expressed HNF-4α were set to 1. Results are expressed as mean -fold activities ± S.D. Asterisk (*) denotes statistically significant difference (p < 0.001) between the indicated data points.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

HNF-4α specifically binds to both DR-2 elements within the hOCT1 promoter. Electrophoretic mobility shift assays were performed using nuclear extracts from Huh7 cells and either a consensus binding site for HNF-4α or sequences containing the wild-type or mutated DR-2 elements from the hOCT1 promoter as a radiolabeled probe. A, a protein-DNA complex is formed using the radiolabeled probe-spanning nucleotides -1479/-1441 of the hOCT1 promoter. The efficiency of complex formation is reduced upon mutating either of the two DR-2 elements and abolished when both the DR-2 elements are mutated. Preincubation of the extracts with an anti-HNF-4α-specific antibody prevented complex formation, confirming that HNF-4α is a component of this protein-DNA complex. B, competition experiment to verify the binding specificity of in vitro translated HNF-4α to the DR-2 elements of the hOCT1 promoter. A complex formed on the radiolabeled probe-spanning nucleotides -1479/-1441 of the hOCT1 promoter can be efficiently competed away by a 50-fold molar excess of the consensus HNF-4α binding site (CON). Likewise, the wild-type DR-2 elements from the hOCT1 promoter were effective as competitors, the DR-2 (A) element (WT A) being a stronger competitor than the DR-2 (B) element (WT B). The DR-2 (A) mut and DR-2 (B) mut competitor oligonucleotides containing point mutations listed in Fig. 1 were less efficient in reducing binding of HNF-4α on the hOCT1(-1479/-1441) probe than their wild-type counterparts.

Bile Acids and SHP Suppress Activation of the hOCT1 Gene by HNF-4α. The expression of several hepatic transporters is known to be modulated by bile acids (Eloranta and Kullak-Ublick, 2005). To investigate whether transactivation of the hOCT1 promoter by HNF-4α can be targeted by bile acids, we exogenously expressed HNF-4α in Huh7 cells, followed by treatment of cells with the bile acid CDCA. As shown in Fig. 6A, HNF-4α induces the endogenous hOCT1 gene similar to the reporter-linked hOCT1 promoter constructs studied above. The HNF-4α-mediated transactivation of the endogenous hOCT1 gene can be efficiently suppressed in Huh7 cells treated with the bile acid CDCA (Fig. 6A). Bile acids can activate various signaling cascades (Wang et al., 2002; Fang et al., 2004; Reinehr et al., 2004) and also serve as physiological ligands for the nuclear bile acid receptor FXR (Makishima et al., 1999; Parks et al., 1999; Wang et al., 1999). To investigate whether an FXR-mediated mechanism is involved in the suppression of hOCT1 mRNA expression, we used a synthetic agonist for FXR, GW4064 (Willson et al., 2001), which does not activate FXR-independent pathways of bile acid signaling. Similar to CDCA, incubation of Huh7 cells with GW4064 for 24 h counteracted the increase in the hOCT1 gene expression by HNF-4α (Fig. 6A), suggesting that the CDCA-mediated suppression indeed depends on FXR activation.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Activation of the endogenous hOCT1 gene by HNF-4α is suppressed by the bile acid CDCA, by the synthetic FXR agonist GW4064, and by exogenous expression of SHP. A, Huh7 cells were transfected with the HNF-4α expression plasmid and treated with 50 μM CDCA, 200 nM GW4064, or the vehicle DMSO for 24 h before isolating total cellular RNA. After reverse transcription, cDNAs were analyzed by quantitative real-time PCR using a hOCT1-specific gene expression assay. The signal obtained for the DMSO-treated cells was set to 1, and values for the samples from HNF-4α-transfected cells treated with DMSO, CDCA, or GW4064 are shown relative to this. B, the same RNA preparations that were analyzed for hOCT1 mRNA expression (A) were used to confirm induction of endogenous SHP mRNA levels upon treatment of cells with CDCA or GW4064. C, Huh7 cells were transfected with the HNF-4α expression plasmid with or without an expression plasmid for SHP. The effects of exogenous expression of these factors on expression level of the endogenous hOCT1 mRNA were determined by quantitative real-time PCR. Results are expressed as mean -fold activities ± S.D.

FXR-mediated transcriptional repression usually involves the FXR-induced transcriptional repressor SHP (Goodwin et al., 2000; Lu et al., 2000). Consistent with this finding, the endogenous level of SHP mRNA was strongly increased in Huh7 cells treated with CDCA and GW4064 (Fig. 6B). Because SHP has previously been shown to negatively interfere with HNF-4α-mediated transactivation in other promoter contexts (Lee et al., 2000; Zhang and Chiang, 2001; Chen and Chiang, 2003; Jung and Kullak-Ublick, 2003; Popowski et al., 2005), we investigated whether bile acids suppress activation of hOCT1 gene expression by HNF-4α via an SHP-mediated mechanism. As shown in Fig. 6C, upon exogenous expression of SHP, the HNF-4α-induced endogenous hOCT1 mRNA expression is effectively reversed to the same levels as obtained in the control conditions. In agreement with these results, the HNF-4α-mediated transactivation of the hOCT1 (-2620/+116)luc promoter construct was suppressed in cells cotransfected with both SHP and HNF-4α expression plasmids (Fig. 7). These results strongly imply that SHP is the executor of bile acid-dependent suppression of HNF-4α-mediated transactivation of the hOCT1 gene.

Discussion

HNF-4α is a liver-enriched nuclear receptor that plays a crucial role in hepatocyte differentiation and maintenance of the hepatic gene expression profile. Its target genes encode proteins involved in a range of physiological processes, notably cholesterol and glucose metabolism (Cereghini, 1996; Hayhurst et al., 2001). HNF-4α binds as a homodimer to its preferred DNA-response elements of the DR-1 or DR-2 configuration, as shown for a range of natural target promoters and by PCR-based binding site selection (Fraser et al., 1998). The crystal structure of the HNF-4α ligand binding domain suggests that it is constitutively bound by endogenous fatty acids (Dhe-Paganon et al., 2002; Wisely et al., 2002) and that its activity may not be readily modulated by other exogenous or endogenous ligands. However, HNF-4α may respond to extracellular and intracellular signals through pathways that modulate its phosphorylation status and, consequently, its activity (Viollet et al., 1997; Jahan and Chiang, 2005).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7.

SHP suppresses activation of the luciferase-linked hOCT1 promoter by HNF-4α. Cells were transfected with the hOCT1(-2620/+116)luc construct or the promoterless vector pGL3basic. Together with the reporter constructs, 100 ng of expression vector for either HNF-4α alone or for both HNF-4α and SHP was cotransfected. For all combinations of cotransfected vectors, the values obtained for pGL3basic were set to 1, and the values obtained for the hOCT1 reporter construct are shown relative to this. Results are expressed as mean -fold activities ± S.D. Asterisk (*) denotes statistically significant difference (p < 0.001) between the indicated data points.

HNF-4α has recently emerged as a key regulator of hepatic transport and metabolism of drugs and bile acids. Notably, one group of genes transactivated by HNF-4α is that encoding the cytochrome P450 enzymes CYP7A1 (Crestani et al., 1998), CYP8B1 (Zhang and Chiang, 2001), and CYP27A1 (Chen and Chiang, 2003) involved in the synthesis of bile acids from cholesterol. Recently, we have identified the gene encoding the human drug transporter organic anion transporter 2 (OAT2; SLC22A7) as a target for transactivation by HNF-4α via its direct binding to a DR-1 element within the promoter region (Popowski et al., 2005). Here, we have identified HNF-4α as a transactivator of the promoter of another member of the human SLC22A transporter gene family, namely the SLC22A1 gene coding for the hepatic drug transporter hOCT1. Both the luciferase reporter-linked hOCT1 promoter and endogenous hOCT1 mRNA expressions are activated by exogenously expressed HNF-4α in the human hepatoma cell line Huh7 (Figs. 2 and 6). By mutational analysis, we have further identified two adjacent HNF-4α binding sites within the hOCT1 promoter (Fig. 1 and 5), both arranged as a DR-2 element [nucleotides -1479/-1466 (A) and nucleotides -1454/-1441 (B), relative to the hOCT1 transcription start site]. The integrity of both of these DR-2 elements is required for maximal transactivation of the hOCT1 promoter by HNF-4α (Fig. 4). We are not aware of other promoter contexts in which such an arrangement of two closely adjacent functional HNF-4α elements would be used in their regulation. We note that both DR-2 elements of the hOCT1 promoter bind HNF-4α in vitro rather weakly, compared with the consensus DR-1 motif for HNF-4α (Fig. 5). It is conceivable that recruitment of HNF-4α to both of the two relatively weak DR-2 elements is required to cooperatively achieve potent activation of the hOCT1 promoter. Furthermore, low-affinity DNA binding could be beneficial in situations in which HNF-4α needs to dissociate from its DNA binding site to quickly and efficiently reduce the rate of transcription from the hOCT1 promoter. In other promoter contexts, variants of DR-2-type motifs are also known to bind and mediate transactivation by the nuclear receptor heterodimer retinoic acid receptor-retinoid X receptor. However, we did not see any effect on the hOCT1 promoter by retinoic acid receptor and retinoid X receptor (unpublished observations).

The sequences of the two DR-2 motifs are entirely conserved between the human and chimpanzee (GenBank NW_107986) OCT1 promoters, supporting the functional importance of these elements in regulation of the OCT1 genes in primates. However, no such conservation is readily apparent in the promoter regions of the rat and mouse Oct1 genes, and thus, it remains unclear whether HNF-4α plays a role in Oct1 gene regulation in these species. Two nuclear receptors of the peroxisome proliferator-activated receptor (PPAR) family, PPARα and PPARγ, have recently been shown to regulate the gene encoding the murine Oct1 homolog (Nie et al., 2005). In our studies, we did not find regulation of the human OCT1 gene by PPARα and PPARγ or their ligands (unpublished observations), further suggesting species specificity in the regulation of the OCT1/Oct1 promoters.

Changes in the concentration of bile acids within hepatocytes influence the expression levels of many genes encoding proteins that are involved in the control of drug metabolism and bile acid homeostasis (Kullak-Ublick et al., 2004; Eloranta and Kullak-Ublick, 2005). This enables the cell to adjust the uptake and subsequent metabolism of drugs and bile acids to tolerable levels in cholestatic situations when hepatocytes are burdened with elevated levels of bile acids. For example, expression levels of several human hepatic uptake transporters, such as organic anion-transporting polypeptide (hOATP1B1) (Jung and Kullak-Ublick, 2003), hOAT2 (Popowski et al., 2005), and Na⁺-taurocholate cotransporting polypeptide (Eloranta et al., 2006), have previously been shown to be suppressed by bile acids at the transcriptional level. The finding that HNF-4α-supported expression of hOCT1 was efficiently suppressed by the bile acid CDCA (Fig. 6A) is consistent with the previous report that the expression of the rat Oct1 is reduced in obstructive cholestasis (Denk et al., 2004). In further agreement, it has recently been shown that Oct1 mRNA expression is reduced in mice fed with a cholic acid-rich diet (Maeda et al., 2004). However, as mentioned above, because it is not known whether HNF-4α regulates the rodent Oct1 genes, the exact mechanisms by which bile acids exert their negative effects on OCT1/Oct1 expression are not necessarily conserved between the species.

Bile acids affect transcriptional rates of their target genes, mainly by acting as ligands for the nuclear receptor FXR (Makishima et al., 1999; Parks et al., 1999; Wang et al., 1999). Bile acids enhance the transactivation ability of FXR, thus leading to up-regulation of its direct target genes. One such direct FXR target gene encodes the transcriptional repressor SHP (Seol et al., 1996). SHP is a nuclear receptor that lacks a DNA binding domain and suppresses a range of other transcription factors through protein-protein interactions (for review, see Eloranta and Kullak-Ublick, 2005). In hepatocytes, SHP mediates, at least partly, the negative transcriptional bile acid response of genes, such as CYP7A1 (Goodwin et al., 2000; Lu et al., 2000), CYP8B1 (Zhang and Chiang, 2001), and hOAT2 (Popowski et al., 2005). In the current study, we show that the synthetic FXR ligand GW4064 and CDCA are equally effective in reversing the activation of endogenous hOCT1 mRNA expression in cultured cells (Fig. 6A), indicating that this effect probably depends on FXR. Furthermore, overexpression of the FXR-inducible repressor SHP counteracted HNF-4α-mediated activation of both the endogenous and reporter-linked hOCT1 promoters (Figs. 6A and 7).

In summary, HNF-4α activates hOCT1 gene expression, and this activation can be negatively targeted by bile acids via an SHP-mediated interference. Thus, two major genes of the human SLC22A family, hOCT1 and hOAT2, are both regulated in a similar manner. However, there are differences in the exact regulatory mechanisms of the two genes. Whereas there is a single DR-1 element with strong binding properties for HNF-4α in the hOAT2 promoter, the activation of the hOCT1 promoter by HNF-4α depends on two adjacent lower affinity DR-2 elements. Down-regulation of HNF-4α-supported expression of both the hOCT1 and hOAT2 genes by bile acids implies that, in cholestatic states where intracellular bile acid levels are elevated, the hepatic uptake of substrates via these two transporters may be coordinately reduced.