Experimental Procedures

Materials.

Steroids, chemicals, and molecular biology reagents were obtained from Sigma-Aldrich (St. Louis, MO). Vitrogen was obtained from the Collagen Corporation (Palo Alto, CA). Matrigel substratum was purchased from Collaborative Biomedical Products (Bedford, MA). Human recombinant insulin (Novolin R) was purchased from Novo Nordisk Pharmaceuticals, Inc. (Princeton, NJ). Trizol reagent, other cell culture reagents, and a random primer DNA labeling kit were obtained from Invitrogen (Carlsbad, CA). The ECL Western blotting kit was purchased from Amersham (Arlington Heights, IL). SDS-polyacrylamide gel electrophoresis reagents and precast polyacrylamide gels were obtained from Bio-Rad Laboratories (Hercules, CA). Nylon hybridization membranes were obtained from PerkinElmer Life Sciences (Boston, MA).

Primary Cultured Human Hepatocytes.

High quality human livers that were judged to be unsuitable for transplantation were obtained from the Transplant Society of Michigan (Ann Arbor, MI). Donor livers were harvested, cold-perfused and preserved with the intent to transplant but were made available for research following a secondary evaluation by the surgical pretransplant team. The cold ischemia time for all livers used in this study was less than 24 h. Hepatocytes were prepared from the left hepatic lobes, using some modifications of the method described by Strom et al. (1996). Briefly, the portal vein was cannulated with silicon tubing (5/32“ o.d.), the tubing was advanced into the branch leading to the left lobe, and perfusion was begun with Hanks' balanced salt solution (HBSS) lacking calcium and magnesium and containing 10 mM HEPES, 5 mM EGTA, 100 U/ml penicillin, and 100 μg/ml streptomycin (Ca/Mg-free HBSS/EGTA). The portal branch leading to the right lobe and the hepatic artery were ligated. After verifying outflow through the hepatic vein, most of the right lobe was dissected free from the liver, and the remnant, containing the left lobe, was placed into a sterile Stomacher bag, which was submerged in a water bath maintained at 37°C. The lobe was perfused at a flow rate of ∼100 ml/min with 1 liter of Ca/Mg-free HBSS/EGTA, followed by 2 liter of Ca/Mg-free HBSS (containing 10 mM HEPES, penicillin, and streptomycin). Following these initial perfusions (total time ∼30 min), the warmed liver was perfused at a flow rate of ∼60 ml/min with 1.7 liters of HBSS (containing calcium and magnesium) supplemented with 0.5% bovine serum albumin, 0.05% collagenase (type IV, Worthington), penicillin, and streptomycin. Following the collagenase perfusion, softened sections of the left lobe were dissected, placed into a sterile beaker and chopped with scissors, and then 500 ml of HBSS containing 0.5% bovine serum albumin, 0.02% collagenase (type IV, Worthington), penicillin, and streptomycin was added to the mixture. The tissue was incubated at 37°C for 10 min with gentle shaking, and released cells were filtered first through sterile gauze and then through 250-μm nylon mesh and were collected into 250-ml centrifuge bottles. Hepatocytes were pelleted by centrifugation at 50g for 3 min and were washed twice with HBSS containing 0.5% bovine serum albumin, penicillin, and streptomycin, and once with Williams' medium E containing 0.25 U/ml insulin, penicillin, and streptomycin (defined as standard Williams' medium E) that was also supplemented with 10⁻⁷ M triamcinolone acetonide. Following these washes, the final hepatocyte pellet was resuspended in standard Williams' medium E, and cell yield and viability were estimated by counting trypan blue-stained samples, using a hemocytometer. The average yield was 1.1 × 10⁹ viable cells with an average cell viability of 79% (with all but one hepatocyte preparation >82% cell viability). The hepatocytes were diluted into standard Williams' medium E containing triamcinolone acetonide and 10% fetal bovine serum and were plated at 3 million cells/dish onto 60-mm dishes that were precoated with Vitrogen, unless otherwise indicated (Runge-Morris et al., 1999). After 3 to 10 h, the medium was replaced with standard Williams' medium E containing triamcinolone acetonide but lacking serum. At approximately 24 h after plating, culture medium was replaced with standard Williams' medium E containing 600-μg Matrigel but lacking triamcinolone acetonide. From 24 h onward, the hepatocytes were maintained in medium that was not supplemented with triamcinolone acetonide. Forty-eight or 72 h after plating, hepatocytes were treated for 24 h as described in the individual figure legends (3 to 6 dishes per treatment group).

Northern and Western Blot Analyses.

The cDNA probes for human SULT1A1 (Wilborn et al., 1993) and SULT2A1 (Falany et al., 1989; Comer et al., 1993) were prepared as previously described. The human CYP3A cDNA probe that was used in Fig.1B was a generous gift from Dr. Erin G. Schuetz (St. Jude Children's Research Hospital, Memphis, TN). This CYP3A7 cDNA probe hybridizes with all known human CYP3A forms. Total RNA was prepared from primary cultured human hepatocytes using the same techniques that were previously applied to primary cultured rat hepatocytes (Runge-Morris et al., 1996). Samples of total RNA were fractionated on denaturing agarose gels, transferred onto nylon filters and hybridized with ³²P-labeled human SULT1A1 or SULT2A1 cDNA probes, as detailed previously for rat sulfotransferase cDNA probes (Runge-Morris et al., 1996). To normalize Northern blots for subtle differences in RNA loading and transfer, filters were stripped of radio-labeled probe following autoradiography and rehybridized with a ³²P-labeled 7S cDNA probe as described previously (Runge-Morris et al., 1996). For ECL Western blot analysis of sulfotransferase protein levels, human hepatocyte cytosol samples were prepared using methods that had previously been applied to the isolation of cytosol protein from primary cultured rat hepatocytes (Runge-Morris et al., 1996). Immunoreactive human SULT2A1 protein expression was determined using a commercially available polyclonal antibody to human SULT2A1 (PanVera Corp., Madison WI), and expressed SULT2A1 protein (Comer et al., 1993) was used as a protein standard. Western blots were normalized for variations in protein loading and transfer by reincubation with a commercial antibody to human β-actin (Santa Cruz Biotechnology Inc., Santa Cruz, CA). The autoradiographs for both Northern and Western blots were quantified using a scanning laser densitometer and the ImageQuant software package (Molecular Dynamics, Sunnyvale, CA).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Northern blot analysis of the effects of DEX treatment on human SULT1A1 and SULT2A1 mRNA expression in primary cultured human hepatocytes.

Primary cultures of human hepatocytes were prepared as described underExperimental Procedures from hepatocyte donor “A”. Forty-eight hour old cultures were treated with 0.1% dimethyl sulfoxide vehicle (DMSO controls) or with DEX at the indicated log molar concentrations (3 culture dishes per treatment group). Cultures were harvested for total RNA at 96 h after plating. Pooled total RNA samples (30 μg/lane) were analyzed on Northern blots that were sequentially hybridized with ³²P-labeled cDNA probes to human (A) SULT1A1, SULT2A1, or (B) CYP3A. Following autoradiography, Northern blots were stripped of radiolabeled probe and rehybridized with a ³²P-labeled cDNA probe to 7S RNA, an abundant cytoplasmic RNA that is routinely employed to normalize Northern blots for variations in RNA loading and transfer. Northern blots revealed three discernible SULT2A1 mRNA bands (an upper, middle, and lower band) and one distinct band for SULT1A1. Graphical representations of densitometrically quantified and normalized Northern blot data are presented in the right-hand panel.

Transient Transfection of SULT2A1–5′ Reporter Constructs in Primary Cultured Rat Hepatocytes, and Cotransfections with a Dominant Negative PXR Expression Construct.

Transfection studies were performed using luciferase reporter constructs containing either 1938 bases of the rat SULT2A-40/41 5′-flanking region, 1463 bases of the human SULT2A1 5′-flanking region or a concatamerized PXR-responsive element (CYP3A23-DR3 nuclear receptor motif). The rat SULT2A-40/41-containing reporter was described previously (Runge-Morris et al., 1999). The 5′-flanking region of the human SULT2A1 (GI 908761, GI 806711, and GI 17482918) spanning from −1463 to + 48 bp, relative to the transcription start site that was previously defined by Otterness et al. (1995) (GI 806711), was obtained by polymerase chain reaction (PCR) amplification using TaqPlus Precision polymerase (Stratagene, La Jolla,CA) and human genomic DNA as template. The forward (5′-GCGACGCGTTTCCCAACTTGCCTTTGAAG-3′) and reverse (5′-GCGCTCGAGGCGTGGTGTGAGGGTTTC-3′) PCR primers were selected using Oligo Primer Analysis Software (Molecular Biology Insights, Cascade, CO). The underscored bases of the primers indicateMluI and XhoI sites that were added to the 5′-ends of the primers and were used for ligation of the amplified product into the corresponding sites of the pGL3-Basic firefly luciferase reporter plasmid (Promega Biotec, Madison, WI). The CYP3A23-DR3-containing reporter construct was prepared by ligating three copies of a double-stranded oligonucleotide (top strand sequence: 5′-GTAGATGAACTTCATGAACTGTCTA-3′; complements to the AGTTCA nuclear receptor motif are underscored) upstream of a minimal herpes simplex virus thymidine kinase promoter, which had been preligated into pGL3-Basic.

To prepare a cDNA encoding a dominant negative PXR receptor, we applied the strategy that was successfully used for preparing dominant negative LXR receptors (Venkateswaran et al., 2000), whereby bases encoding the C-terminal portion of the receptor, encompassing the AF-2 subdomain of the ligand binding domain, were deleted. Alignment of the mouse PXR amino acid sequence (GI 2852329) to that for mouse LXRα (GI 7305321) indicated that deletion of the bases encoding the 11 C-terminal amino acids of PXR should produce a receptor lacking its AF-2 subdomain. A cDNA encoding this receptor was prepared by PCR, using Pfupolymerase, 100 ng of a mouse PXR cDNA clone (gift from Dr. Steven Kliewer, GlaxoSmithKline, Research Triangle Park, NC) as template, and primers corresponding to bases 1–22 (counting from the translation initiation codon) and 1260–1244. The sequence of the forward primer was 5′-GCGGGTACCGCCACCATGAGACCTGAGGAGAGCTGGA-3′, and the sequence of the reverse primer was 5′-GCGTCTAGAGGTCATCATGGGGTGGCAAAGGGT -3′. The underscored bases of the primers indicate KpnI andXbaI restriction sites, whereas the bolded bases of the forward and reverse primers represent a Kozak consensus sequence (Kozak, 1996) and tandem translation stop codons, respectively. Following amplification for 20 cycles, the product was digested withKpnI and XbaI and ligated into the corresponding sites of the pcDNA3.1 expression plasmid (Invitrogen). The identities of the cloned fragments with human SULT2A1 and mouse PXR were verified by the Center for Molecular Medicine and Genetics DNA Sequencing Facility at Wayne State University.

Due to the limited availability of primary cultured human hepatocytes, luciferase reporter constructs were transiently transfected into 48 h-old primary cultured rat hepatocytes, essentially using methods that were previously described for the transcriptional analysis of the rat SULT2A-40/41 gene (Runge-Morris et al., 1999). For each transfection, a plasmid mixture consisting of 800 ng of one of the reporters, 20 ng of either the dominant negative PXR expression plasmid or pcDNA3.1 (as empty vector control) 0.25 ng pRL-CMV (expressing Renilla eniformis luciferase to control for transfection efficiency) and 179.75 ng of pBluescript II KS⁺ (to balance total amounts of DNA to 1 μg) was combined with 6.25 μg of Lipofectin (Invitrogen). In preliminary experiments, we determined that 20 ng of the dominant negative PXR plasmid markedly inhibited PCN-mediated activation of reporter gene expression from the DR3-containing plasmid, without affecting phenobarbital-mediated activation of expression from a CYP2B1 5′-luciferase reporter plasmid (manuscript in preparation). Transfectants were treated for 24 h with test reagents as described in the figure legends and then harvested for the measurement of luciferase activity using a dual luciferase reporter assay system (Promega Biotec) according to the manufacturer's instructions and a Dynex model MLX luminometer. Transient transfection data were analyzed by one-way analysis of variance followed by the Newman-Keuls multiple comparison test (GraphPad Software Inc., San Diego, CA).

Results

We previously demonstrated that SULT1A1 and SULT2A mRNA and immunoreactive protein expression are induced in response to treatment with the potent synthetic glucocorticoid DEX, in primary cultured rat hepatocytes (Runge-Morris et al., 1996). We also showed that glucocorticoid-inducible SULT1A1 (unpublished data) and SULT2A mRNA expression (Wu et al., 2001) occurs in primary cultured mouse hepatocytes (Wu et al., 2001). To determine whether DEX-inducible SULT1A1 and SULT2A1 expression is conserved in humans, the effects of treatment on SULT1A1 and SULT2A1 expression were investigated in primary cultured human hepatocytes. As in rat (Runge-Morris et al., 1996) and mouse (Wu et al., 2001) hepatocytes, the mRNA expression of SULT2A1 in primary cultured human hepatocytes, prepared from four different donors (here designated A through D), increased in response to DEX treatment in a concentration-dependent manner (Figs. 1 and3-5).

In hepatocyte cultures prepared from liver donor “A” (Fig. 1A), three mRNA sizes estimated at ∼1100, 1300, and 1800 bp could be discerned on Northern blots that were hybridized with the human SULT2A1 cDNA probe (Fig. 1A), as has been described in the literature (Falany et al., 1995). To determine whether the different mRNA species were coregulated, the relative amounts of the different bands were quantified separately. As is apparent from the histogram representations, expression of the three different transcripts appeared to be modified in parallel. Relative to vehicle-treated controls, the greatest stimulation in SULT2A1 mRNA expression occurred following the treatment of human hepatocyte cultures with pharmacological concentrations of DEX. Relative to DMSO-treated controls, band intensities of the upper, middle, and lower SULT2A1 mRNA bands increased by 11.7-, 13.1-, and 10.7-fold, respectively, following incubation of cultured hepatocytes with DEX (10⁻⁵ M) (Fig. 1A). For comparison, the DEX-inducible expression of CYP3A mRNA, which is known to be regulated by the PXR transcription factor (Lehmann et al., 1998), was demonstrated in the same human hepatocyte preparation (Fig. 1B).

To establish that DEX-inducible SULT2A1 mRNA expression results in corresponding increases in SULT2A1 protein expression, the effects of DEX treatment on SULT2A1 immunoreactive protein levels were examined in hepatocyte cultures that were prepared from the same donor. Western blot analysis showed that, relative to vehicle-treated controls, the treatment of primary cultured human hepatocytes with concentrations of DEX ranging from 10⁻⁸ M to 10⁻⁵ M produced from 6- to 17-fold increases in SULT2A1 protein expression (Fig. 2), thus recapitulating the levels of DEX-inducible SULT2A1 mRNA expression that were observed. In contrast to the glucocorticoid-inducible expression of SULT1A1 that has been previously described in primary cultured rat (Runge-Morris et al., 1996) and mouse (unpublished) hepatocytes, and in bovine tracheobronchial epithelial cells (Schauss et al., 1995), SULT1A1 mRNA expression did not increase in response to DEX treatment in primary cultured human hepatocytes (Fig. 1A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Western blot analysis of the effects of DEX treatment on human SULT2A1 protein levels in cytosol prepared from primary cultured human hepatocytes.

Primary cultured human hepatocytes from hepatocyte donor A were prepared as described under Experimental Procedures. Forty-eight hour old cultures were treated with 0.1% DMSO (control) or with DEX at the indicated log molar concentrations (six culture dishes per treatment group). Cultures were harvested for cytosol protein at 96 h after plating. Human SULT2A1 protein standard (100 ng) was applied to the first lane, and pooled cytosol hepatocyte samples (125 μg/lane) were analyzed by ECL western blotting, using a commercial polyclonal antibody to human SULT2A1. A SULT2A1 protein signal of the expected subunit size (∼35 kDa) was identified in each of the hepatocyte cytosol samples and confirmed in the SULT2A1 protein standard (first lane). The Western blot was normalized for variations in protein loading and transfer by reincubation with a commercial antibody to human β-actin. To delineate semiquantitative changes in the amounts of SULT2A1 protein expressed, the cytosol protein concentrations that were used in the analysis were within the linear range of ECL detection for the SULT2A1 antibody.

In deference to the potential for variations among donor livers in hepatocyte culture quality, as well as the possibility for interindividual differences in glucocorticoid-responsiveness that are attributable to the polymorphic expression of one or more components of the cellular response machinery, the reproducibility of DEX-inducible SULT2A1 expression and DEX-refractory SULT1A1 expression was examined in hepatocyte cultures prepared from three additional human livers. In repeated studies, Northern blots revealed two rather than three predominant SULT2A1 mRNA bands. Nevertheless, as depicted in Fig. 1, the incubation of cultured hepatocytes with higher concentrations of DEX produced measurable increases in SULT2A1 mRNA levels, although the magnitude of DEX-inducible expression varied among hepatocyte preparations. Relative to DMSO-treated controls, the data in Fig.3 (donor B) demonstrated that DEX (10⁻⁶ M) treatment produced 4.5- and 4.2-fold increases in the upper and lower SULT2A1 mRNA band intensities, respectively. In Figs. 4 and5 (donors C and D), DEX treatment also produced increases in the amounts of SULT2A1 mRNA. For example, relative to DMSO-treated controls, DEX (10⁻⁵ M) treatment increased the expression of the upper and lower SULT2A1 mRNA bands by 2.2- and 3.8-fold, respectively (Fig. 4), whereas in Fig. 5, the intensities of the upper and lower SULT2A1 mRNA bands increased by 3.0- and 1.7-fold, respectively, in response to DEX (10⁻⁵ M) treatment (Fig. 5). By contrast, none of the four hepatocyte preparations showed appreciable (greater than 2-fold) changes indicative of DEX-inducible SULT1A1 mRNA expression (Figs. 1, and 3-5).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

The effects of DEX treatment on human SULT1A1 and SULT2A1 mRNA expression in primary cultured human hepatocytes.

Primary cultured human hepatocytes were prepared from an additional liver donor (donor “B”) as described under Experimental Procedures. Northern blots to assess the expression of human SULT1A1 and SULT2A1 were performed and analyzed as described in the legend to Fig. 1. For hepatocyte donor B, Northern blots revealed two discernible SULT2A1 mRNA bands (an upper and a lower band) and one distinct band for SULT1A1. Graphical representations of normalized Northern blot data are presented in the right-hand panel.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

The effects of DEX treatment on human SULT1A1 and SULT2A1 mRNA expression in primary cultured human hepatocytes.

Primary cultured human hepatocytes were prepared from an additional liver donor (donor “C”) as described under Experimental Procedures. Northern blots to assess the expression of human SULT1A1 and SULT2A1 were performed and analyzed as described in the legend to Fig. 1. For hepatocyte donor C, Northern blots revealed two discernible SULT2A1 mRNA bands (an upper and a lower band) and one distinct band for SULT1A1. Graphical representations of normalized Northern blot data are presented in the right-hand panel.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

The effects of glucocorticoid and antiglucocorticoid treatments on human SULT1A1 and SULT2A1 mRNA expression in primary cultured human hepatocytes.

Primary cultured human hepatocytes were prepared from liver donor “D”, as described under Experimental Procedures, except that the hepatocytes had been plated onto dishes that had been precoated with Matrigel. Northern blots to assess the expression of human SULT1A1 and SULT2A1 were performed and analyzed as described in the legend to Fig. 1. Seventy-two hour old cultures were treated for 24 h with DMSO (0.1%, vehicle control), DEX (10⁻⁷ or 10⁻⁵ M), antiglucocorticoid RU486 (10⁻⁶ M), DEX (10⁻⁷ M) + RU486 (10⁻⁶ M), or TA (10⁻⁷ or 10⁻⁵ M). For hepatocyte donor D, Northern blots revealed two discernable SULT2A1 mRNA bands (an upper and a lower band) and one distinct band for SULT1A1. Graphical representations of normalized Northern blot data are presented in the right-hand panel.

Currently, two receptors are known to be involved in glucocorticoid-mediated regulation of gene expression, namely the classical glucocorticoid receptor, which controls the effects of physiological concentrations of glucocorticoid in a variety of metabolic processes, and the PXR (also known as steroid and xenobiotic receptor or SXR in human), which mediates the effects of higher doses of selected glucocorticoids (e.g., DEX), as well as pregnanes, secondary bile acids (e.g., lithocholic acid), and a host of xenobiotics, on expression of certain genes that encode drug-metabolizing enzymes (e.g., CYP3A family members). In support of a role for the PXR transcription factor as the prime mediator of DEX-inducible SULT2A1 gene transcription, we were able to demonstrate that both SULT2A1 and CYP3A are DEX-inducible within the same human hepatocyte preparation (Fig. 1). To dissect the salient mechanism(s) that underwrite DEX-inducible SULT2A1 expression, primary cultured human hepatocytes were treated with DEX in the presence of the antiglucocorticoid RU486 or were treated with triamcinolone acetonide, a potent ligand for the glucocorticoid receptor, but a poor inducer of rat CYP3A (Schuetz and Guzelian, 1984) (Fig. 5). While SULT1A1 mRNA levels were not markedly altered in response to any of the steroid treatments, SULT2A1 mRNA expression was induced by both DEX and triamcinolone acetonide (TA) treatments, at the lowest concentration tested (10⁻⁷ M), supporting some involvement of the classical glucocorticoid receptor (Fig. 5). However, cotreatment with DEX (10⁻⁷ M) and RU486, at a concentration (10⁻⁶ M) that is insufficient to activate PXR (Kliewer et al., 1998), did not effectively block DEX-inducible SULT2A1 expression (Fig. 5), suggesting that as in the rat, DEX-inducible human SULT2A1 expression likely includes a nonglucocorticoid receptor-mediated component, possibly involving the PXR.

To target the role of the PXR transcription factor as a central mediator of DEX-inducible SULT2A1 expression, transient transfection studies were conducted in primary cultured rat hepatocytes using reporter constructs containing 5′-flanking regions of the human SULT2A1 gene (shown in Fig. 6) and the rat SULT2A-40/41 gene, with the DR3 nuclear receptor motif of the CYP3A23 gene used as a PXR-responsive control (Fig.7). Treatment with DEX or PCN, a prototypical ligand for the rat PXR, robustly activated the DR3-containing reporter plasmid and also significantly and proportionately activated transcription of the human SULT2A1 and rat SULT2A-40/41 reporter constructs (Fig. 7). In addition, cotransfection of hepatocyte cultures with a plasmid expressing a dominant negative PXR essentially ablated DEX- and PCN-inducible expression from the DR3 reporter plasmid (relative to the responses that were produced in cultures cotransfected with empty vector), while also eliminating PCN-inducible expression from the rat and human SULT2A reporter plasmids. In addition, cotransfection with the dominant negative PXR partially, but significantly, decreased DEX-inducible expression from both the rat and human constructs (Fig. 7), again supporting roles for both PXR- and non-PXR-mediated components in DEX-inducible SULT2A expression.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Sequence of the 5′-flanking region of the human SULT2A1 gene.

The nucleotide sequence of the human SULT2A1 5′-flanking region that was prepared by PCR and used in transient transfection studies is shown. The transcription start site that was identified by Weinshilboum and coworkers (Otterness et al., 1995) is indicated by the broken arrow. The locations of several putative (A/G)G(G/T)TCA nuclear receptor half-site motifs are also depicted (shaded areas).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7

Transient transfections of primary cultured rat hepatocytes with SULT2A-luciferase reporter plasmids and effects of cotransfected plasmid expressing a dominant negative PXR.

Primary cultured rat hepatocytes were transiently transfected with luciferase reporter plasmid containing either human SULT2A1 5′-flanking region, rat SULT2A-40/41 5′-flanking region or a PXR-responsive CYP3A23-DR3 nuclear receptor motif together with either 20 ng pcDNA3.1 plasmid (pcDNA, empty vector control) or 20 ng of PXR-DN-pcDNA (PXR-DN, dominant negative PXR plasmid), as described under Experimental Procedures. Transfected cells were treated for 24 h with either DMSO (0.1%, vehicle), PCN (10⁻⁵ M), or DEX (10⁻⁵ M) and harvested for measurement of luciferase activities. The results depict firefly/Renilla reniformis luciferase activities for each of the transfected reporter constructs. Each bar represents the mean-normalized luciferase activity ± S.D., with n = 3 wells per treatment group. Statistical analysis consisted of one-way analysis of variance, followed by the Neuman-Keuls multiple comparison test. Groups that do not share a capital letter are significantly different from each other (p < 0.05).

Discussion

The application of in vivo and in vitro rodent models has proved to be an invaluable tool to investigators who are committed to understanding the complex role of steroid signals in the transcriptional regulation of gene expression. In vivo, the treatment of rats with pharmacological doses of DEX was previously shown to produce differential effects on SULT1A1 and SULT2A isoform-specific gene expression in male and female rat liver (Liu and Klaassen, 1996b). In vitro, we reported that glucocorticoid-inducible SULT1A and SULT2A gene expression occurs as a consequence of mechanism(s) that act directly on the hepatocyte (Runge-Morris et al., 1996).

To probe the molecular mechanism(s) that underwrite glucocorticoid-inducible SULT gene expression in the rat, transient transfection studies were conducted in primary cultured rat hepatocytes. These studies supported a central role for the glucocorticoid receptor transcription factor in the regulation of SULT1A1 by physiological concentrations of glucocorticoids (Duanmu et al., 2001). Although the rat SULT1A1 gene did not contain a consensus GRE, an integrated analyses of SULT1A1 reporter plasmids identified acis-acting glucocorticoid responsive region in the rat SULT1A1–5′-flanking sequence which contained two candidate GRE-like sequences (Duanmu et al., 2001).

In contrast to the straightforward scenario for glucocorticoid-inducible rat SULT1A1 gene expression, the role of nuclear receptors in the transcriptional regulation of rat SULT2A-40/41 has materialized as a more complex and multifaceted process. Of the three rat hepatic SULT2A isoforms, the expression and regulation of SULT2A-40/41, a form that is robustly expressed both in rat liver and in primary cultured rat hepatocytes, has been best characterized. Growth hormone has previously been shown to be an important regulator of age- and gender-dependent expression of SULT2A-40/41 in rat liver (Liu and Klaassen, 1996a; Ueda et al., 1997). Moreover, as male rats age beyond puberty, the expression of hepatic SULT2A-40/41 declines with rising androgen levels (Chatterjee et al., 1990). Previous studies on the 5′-flanking region of the rat SULT2A-40/41 gene indicated that the androgen receptor is involved in regulation of androgen-repressible SULT2A-40/41 gene transcription, but also suggested that the androgen receptor transcription factor does not bind directly tocis-acting sequences in the SULT2A-40/41 5′-flanking region (Song et al., 1998).

An emerging body of evidence has implicated members of the “orphan nuclear receptor” family of transcription factors in the regulation of genes that encode proteins involved in cholesterol and bile acid biosynthesis, metabolism, and transport (Kliewer et al., 1999). Recent data have identified the bile acid chenodeoxycholic acid as a physiological ligand for the farnesoid X receptor (FXR) (Makishima et al., 1999). The ligand activated FXR · RXR transcription factor typically binds to IR1 nuclear receptor motifs in target gene sequences (Forman et al., 1995) and thereby mediates the transcriptional repression of the rate-limiting enzyme in bile acid synthesis, cholesterol 7α-hydroxylase, or activates the transcription of the human bile salt export pump (Makishima et al., 1999). Similarly, the PXR was initially discovered to be activated by steroidal ligands such as pregnenolone, PCN and DEX (Kliewer et al., 1998), and more recently by the hydrophobic secondary bile acid, lithocholic acid (Staudinger et al., 2001). The ligand bound PXR · RXR transcription factor induces gene transcription by binding to conserved DR3 or ER6 motifs in target genes such as CYP3A23 and CYP3A4 (Kliewer et al., 1998).

Alternatively coined by the name “bile acid sulfotransferases”, the SULT2A enzymes in both rat (Barnes et al., 1989) and human (Radominska et al., 1990) liver have long been considered to be an integral part of the bile acid metabolism defense machinery that protects the liver against the cholestatic effects of toxic bile acids. Recently, an IR0 motif that is located within the proximal promoter region of the rat SULT2A-40/41 gene has been implicated as a recognition sequence for FXR and PXR. In transfected human HepG2 cells and in Caco-2 enterocytes, the FXR · RXR transcription factor was shown to bind the IR0 and trans-activate SULT2A-40/41 (also called “STD”) reporter constructs (Song et al., 2001). Similarly, we found evidence to suggest that the rat SULT2A-40/41 gene, which does not contain a cognate PXR recognition sequence (i.e., DR3 or ER6), is nevertheless transcriptionally activated by PXR ligands. For example, in previous transfection studies conducted in primary cultured rat hepatocytes, the treatment of cultures with the PXR ligand PCN resulted in the trans-activation of SULT2A-40/41–5′ reporter gene constructs that contained an intact IR0 (Runge-Morris et al., 1999).

In contrast to the relatively more comprehensive work that has been done on SULT gene regulation in the rat, the molecular mechanisms that control human SULT1A1 and SULT2A1 gene transcription have not been well characterized. Therefore, the present study was undertaken to define the extent to which transcriptional regulatory mechanisms may be conserved across species. Unlike the rat SULT1A1 gene, which is most likely regulated by a glucocorticoid receptor-mediated mechanism, the expression of SULT1A1 was not glucocorticoid-inducible in primary cultured human hepatocytes. In accord with this observation, a sequence comparison of the rat and human SULT1A1 gene structures indicates possible reasons why the regulation of the two genes may be quite dissimilar. Whereas the rat SULT1A1 gene contains a candidate GRE motif within the glucocorticoid-responsive region of its 5′-flanking sequence (Duanmu et al., 2001), no such element is apparent in the human gene. A striking feature of human SULT1A1 gene regulation is the utilization of two alternative promoters, giving rise to two SULT1A1 isoforms. Although we have demonstrated that the rat SULT1A1 gene contains multiple transcription start sites within close proximity of one another, there is currently no evidence that transcriptional regulation of rat SULT1A1 involves alternative promoter usage. In any case, the lack of glucocorticoid inducibility of the human SULT1A1 suggests that this is not a priority mechanism for the transcriptional control of this gene.

By contrast to SULT1A1, our results demonstrate that glucocorticoid regulation is conserved for SULT2A in primary cultured human hepatocytes. We showed that the mRNA and protein expression of human SULT2A1 gene, like its rodent counterparts, is induced in response to treatment with concentrations of DEX that would be expected to activate the PXR transcription factor. As further evidence in support of an incisive role for the PXR transcription factor in DEX-inducible rat and human SULT2A expression, cotransfection of either a human SULT2A1 or rat SULT2A-40/41 reporter construct into primary cultured rat hepatocytes showed that the induction of reporter gene expression by PXR ligands could be significantly reduced by the presence of a dominant negative PXR.

Computer-based analysis of the human SULT2A1 5′-flanking sequence failed to reveal a consensus nuclear receptor motif [i.e., direct, inverted or everted repeat of (A/G)G(G/T)TCA]. However, visual inspection indicated several nuclear receptor half-sites located within 200 bp of the SULT2A1 transcription start site (Fig. 6), and it should be noted that sequences exhibiting substantial divergence from the consensus have been demonstrated to be functional nuclear receptor response elements. For example, Yoshikawa et al. (2001) recently demonstrated that an oxysterol-responsive region of the sterol regulatory element binding protein-1 promoter contained two LXR response elements that did not conform to the consensus DR4. Therefore, the functionality of potential nuclear receptor motifs in the SULT2A1 5′-flanking region awaits verification through systematic characterization in transfection studies.

The PXR transcription factor is involved in the regulation of DEX-inducible CYP3A expression (Kliewer et al., 1998), and as a nuclear receptor with unusually broad ligand specificity (Watkins et al., 2001), regulation by this relatively malleable transcription factor may enlarge the capacity of lynchpin drug-metabolizing enzymes such as CYP3A4 and SULT2A1, to detoxify a wider range of xenobiotic substrates. By the same token, the predominant localization SULT2A1 to the liver, an organ which is not only the seat of drug metabolism but also of cholesterol and bile acid biosynthesis, places SULT2A1 in a prime position for transcriptional regulation by the sterol and bile acid intermediates that function endogenously as PXR, FXR, or LXR ligands. To establish the precise molecular mechanisms that are responsible for the regulation of SULT1A1 and SULT2A1 expression in human hepatocytes, future studies will emphasize and refine the dynamic interactions ofcis- and trans-acting factors that combine to orchestrate SULT2A gene transcription.