Advertisement

Transcriptional Regulation of CYP2C9 Gene

ROLE OF GLUCOCORTICOID RECEPTOR AND CONSTITUTIVE ANDROSTANE RECEPTOR*
      Although cytochrome P450 2C9 (CYP2C9) is a major CYP expressed in the adult human liver, its mechanism of regulation is poorly known. In previous work, we have shown that CYP2C9is inducible in primary human hepatocytes by xenobiotics including dexamethasone, rifampicin, and phenobarbital. The aim of this work was to investigate the molecular mechanism(s) controlling the inducible expression of CYP2C9. Deletional analysis ofCYP2C9 regulatory region (+21 to −2088) in the presence of various hormone nuclear receptors suggested the presence of two functional response elements, a glucocorticoid receptor-responsive element (−1648/−1684) and a constitutive androstane receptor-responsive element (CAR, −1783/−1856). Each of these were characterized by co-transfection experiments, directed mutagenesis, gel shift assays, and response to specific antagonists RU486 and androstanol. By these experiments we located a glucocorticoid-responsive element imperfect palindrome at −1662/−1676, and a DR4 motif at −1803/−1818 recognized and transactivated by human glucocorticoid receptor and by hCAR and pregnane X receptor, respectively. Identification of these functional elements provides rational mechanistic basis for CYP2C9induction by dexamethasone (submicromolar concentrations), and by phenobarbital and rifampicin, respectively. CYP2C9 appears therefore to be a primary glucocorticoid-responsive gene, which in addition, may be induced by xenobiotics through CAR/pregnane X receptor activation.
      CYP
      cytochrome P-450
      PXR
      pregnane X receptor
      GR
      glucocorticoid receptor
      GRE
      glucocorticoid-responsive element
      CAR
      constitutive androstane receptor
      RXR
      retinoid X receptor
      XREM
      xenobiotic responsive enhancer module
      DMEM
      Dulbecco's modified Eagle's medium
      PCN
      pregnenolone 16α-carbonitrile
      TCPOBOP
      1,4-bis[2-(3,5-dichloropyridyloxy)]-benzene
      CAR-RE
      constitutive androstane receptor-responsive element
      DTT
      dithiothreitol
      EMSA
      electrophoretic mobility shift assay
      TAT
      tyrosine aminotransferase
      Cytochrome P-450 (CYP)1is the generic name of a superfamily of heme-thiolate proteins that play a critical role in the oxidative metabolism of xenobiotics, including drugs, environmental pollutants and contaminants, and biological signaling molecules such as steroid hormones and biliary salts. CYP2C9 is a member of the CYP2C subfamily, which in man includes at least three other members, i.e. CYP2C8, CYP2C18, and CYP2C19 (
      • Goldstein J.A.
      • de Morais S.
      ). Accumulating evidence indicates that CYPC9 ranks second, after CYP3A4, among the most expressed drug-metabolizing enzymes in human liver (
      • Miners J.O.
      • Birkett D.J.
      ). CYP2C9 is involved in the metabolism of numerous substrates including phenytoin, tolbutamide, torsemide, S-warfarin, and numerous nonsteroidal anti-inflammatory drugs (reviewed in Refs.
      • Goldstein J.A.
      • de Morais S.
      and
      • Miners J.O.
      • Birkett D.J.
      ). In contrast to the large amount of data on the biochemistry, enzymology, pharmacology, and genetic polymorphism of CYP2C9, little is known on the inducibility of this gene in response to xenobiotics in humans. We recently demonstrated that CYP2C9 is inducible at the mRNA and protein levels in human hepatocytes in primary cultures in response to xenobiotics shown previously to beCYP3A4 and CYP2B6 inducers such as dexamethasone, rifampicin, and phenobarbital (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ). The concentration and time dependence of CYP2C9 mRNA expression in response to these three inducers, compared with those of CYP3A4 and CYP2B6 mRNAs, were consistent with the possible implication of at least three receptors in the inducible expression of CYP2C9: the glucocorticoid receptor (GR), the pregnane X receptor (PXR, also named steroid and xenobiotic receptor and pregnane-activated receptor), and the constitutive androstane receptor (CAR), respectively (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ,
      • Kliewer S.A.
      • Moore J.T.
      • Wade L.
      • Staudinger J.L.
      • Watson M.A.
      • Jones S.A.
      • McKee D.D.
      • Oliver B.B.
      • Willson T.M.
      • Zetterstrom R.H.
      • Perlmann T.
      • Lehmann J.M.
      ,
      • Sueyoshi T.
      • Kawamoto T.
      • Zelko I.
      • Honkakoski P.
      • Negishi M.
      ).
      Recent reports on PXR and CAR, two new members of the steroid receptor superfamily, have considerably clarified our understanding of the inducible regulation of CYP genes from families 2 and 3, in response to xenobiotics in rodents and in man (
      • Kliewer S.A.
      • Moore J.T.
      • Wade L.
      • Staudinger J.L.
      • Watson M.A.
      • Jones S.A.
      • McKee D.D.
      • Oliver B.B.
      • Willson T.M.
      • Zetterstrom R.H.
      • Perlmann T.
      • Lehmann J.M.
      ,
      • Sueyoshi T.
      • Kawamoto T.
      • Zelko I.
      • Honkakoski P.
      • Negishi M.
      ,
      • Bertilsson G.
      • Heidrich J.
      • Svensson K.
      • Asman M.
      • Jendeberg L.
      • Sydow B.M.
      • Ohlsson R.
      • Postlind H.
      • Blomquist P.
      • Berkenstam A.
      ,
      • Blumberg B.
      • Sabbagh W.J.
      • Juguilon H.
      • Bolado J.J.
      • van, M. C.
      • Ong E.S.
      • Evans R.M.
      ,
      • Lehmann J.M.
      • McKee D.D.
      • Watson M.A.
      • Willson T.M.
      • Moore J.T.
      • Kliewer S.A.
      ,
      • Honkakoski P.
      • Negishi M.
      ,
      • Kim J.
      • Kemper B.
      ). PXR forms a heterodimer with RXR, and this complex has been shown to activateCYP3A4 transcription through binding to an ER6 element present at position −160 in the proximal promoter (
      • Lehmann J.M.
      • McKee D.D.
      • Watson M.A.
      • Willson T.M.
      • Moore J.T.
      • Kliewer S.A.
      ). More recently, Goodwin et al. (
      • Goodwin B.
      • Hodgson E.
      • Liddle C.
      ) reported the presence of a distal element called xenobiotic responsive enhancer module (XREM) harboring both a DR3 and an ER6 motif, located at −7 kb, and they demonstrated that this element cooperates with the proximal ER6 to activateCYP3A4 transcription. PXR is activated by numerous compounds known to induce CYP3A expression, such as rifampicin, phenobarbital, clotrimazole, and dexamethasone (
      • Moore L.B.
      • Parks D.J.
      • Jones S.A.
      • Bledsoe R.K.
      • Consler T.G.
      • Stimmel J.B.
      • Goodwin B.
      • Liddle C.
      • Blanchard S.G.
      • Willson T.M.
      • Collins J.L.
      • Kliewer S.A.
      ). The apparentKd values of these compounds for the human PXR are in the micromolar (rifampicin, clotrimazole) or supramicromolar (phenobarbital, dexamethasone) range. In contrast to PXR, CAR is sequestered in the cytosol and translocates into the nucleus upon activation, notably in response to phenobarbital (
      • Honkakoski P.
      • Negishi M.
      ), presumably through several steps of phosphorylation (
      • Kawamoto T.
      • Sueyoshi T.
      • Zelko I.
      • Moore R.
      • Washburn K.
      • Negishi M.
      ). Like PXR, CAR forms a heterodimer with RXR. Several groups have identified a complex phenobarbital-responsive element module in CYP2B genes (
      • Park Y.
      • Li H.
      • Kemper B.
      ,
      • Trottier E.
      • Belzil A.
      • Stoltz C.
      • Anderson A.
      ), which was further characterized by Honkakoski et al.(
      • Honkakoski P.
      • Negishi M.
      ). The active element called NR1 is located at position −1663/−1683 in CYP2B6 and exhibits a DR4 motif (
      • Sueyoshi T.
      • Kawamoto T.
      • Zelko I.
      • Honkakoski P.
      • Negishi M.
      ). Only a few molecules among CYP inducers were shown to bind directly to human CAR. These include androstenol (and related compounds at supramicromolar concentrations), and clotrimazole or 5β-pregnane-3,20-dione (in the micromolar range). However, both androstenol and clotrimazole appear to be deactivators instead of activators of hCAR, whereas 5β-pregnane-3,20-dione appears as a true activator. Phenobarbital, a compound that has been shown to activate CAR through indirect mechanism (
      • Honkakoski P.
      • Negishi M.
      ), and other compounds known as CYP inducers (like rifampicin and dexamethasone) are not ligands of CAR (
      • Moore L.B.
      • Parks D.J.
      • Jones S.A.
      • Bledsoe R.K.
      • Consler T.G.
      • Stimmel J.B.
      • Goodwin B.
      • Liddle C.
      • Blanchard S.G.
      • Willson T.M.
      • Collins J.L.
      • Kliewer S.A.
      ).
      The aim of the present work was to investigate the molecular mechanism(s) of induction of CYP2C9 by dexamethasone, rifampicin, and phenobarbital. For this purpose, the 5′-flanking region of this gene was analyzed between +21 and −2088 by several tests, including transcriptional analysis of deletion fragments, co-transfection with nuclear receptor expression plasmids, directed mutagenesis, and gel shift assays. Our results suggest the existence of two functional responsive elements in the regulatory region ofCYP2C9: an imperfect palindromic glucocorticoid-responsive element (GRE) at −1662/−1676 and a CAR-responsive element (DR4) at −1803/−1818.

      EXPERIMENTAL PROCEDURES

       Materials and Reagents

      DMEM culture medium was purchased from Invitrogen (Cergy Pontoise, France). Dexamethasone, mifepristone (RU486), androstenol (5α-androst-16-en-3α-ol), rifampicin, pregnenolone 16α-carbonitrile (PCN), cycloheximide, and dimethyl sulfoxide (Me2SO) were purchased from Sigma. 1,4-Bis[2-(3,5-dichloropyridyloxy)]-benzene (TCPOBOP) was gift from P. Lesca (INRA, Toulouse, France). [γ-32P]ATP was purchased from Amersham Biosciences, Inc. (Amersham, England).

       CYP2C9 Promoter Constructs

      A 2.1-kb XbaI/BglII fragment ofCYP2C9 5′-flanking region was cloned in pBluescript vector (Stratagene, La Jolla, CA) after amplification by PCR using human DNA as a template and oligonucleotides sense (p2C9−2088/XbaI) 5′-ATCTACACATTATCTAGAATTCTTTCT-3′ and reverse (p2C9+21/BglII) 5′-GAGATCTTCTCTTCTTGTTAAGACAACCA-3′.CYP2C9−1545/+21 deletion was then obtained byBspEI digestion, blunt-ended with Klenow enzyme, and cloned into pBlueScript-cut SmaI. CYP2C9−340/+21 deletion was obtained by an EcoRI/EcoRI deletion. These fragments were then cloned into pGL3-basic vector usingKpnI/SacI enzymes. CYP2C9−1856/+21 deletion was obtained by a StuI/SmaI deletion in pGL3-basic vector. CYP2C9−1783, −1684, and −1648/+21 were amplified by PCR using pBS-CYP2C9−2088/+21 as a template and oligonucleotides sense p2C9−1783/KpnI 5′-CGGGGTACCCTGTAATTATTAATG-3′, p2C9−1684/KpnI 5′-CGGGGTACCCAACTGA-ACTGAATG-3′ and p2C9−1648/KpnI 5′-CGGGGTACCTTTGAGATGCAGGGCTTATG-3′ and reverse p2C9+21/BglII. They were then cloned into pGL3-basic digested with BglII/KpnI. Oligonucleotides corresponding to 2C9-GRE (5′-ACCCAACTGAACTGAATGTTTTGCTTGAA-3′) and 2C9-DR4 (5′-AAACCAAACTCTTCTGACCTCTCA-3′) were cloned into pGL3 promoter digested with SmaI.

       Other Plasmids

      The pTAT-GRE-TK-luc and pTAT-(GRE)2-TK-luc plasmids containing one or two copies of the consensus GRE upstream of a minimal herpes simplex virus thymidine kinase promoter and a luciferase reporter gene was kindly provided by Dr L. Poellinger (Karolinska Institute, Stockholm, Sweden). The wild hGR expression vector (pSG5-hGR) was kindly provided by Dr. J. C. Nicolas (INSERM, Montpellier, France). The pΔATG-hPXR expression plasmid was generated by PCR amplification of cDNA encoding amino acids 1–434 of hPXR (S. Kliewer, Glaxo-Wellcome, Research Triangle Park, NC) using oligonucleotides 5′-GGGTGTGGGGAATTCACCACCATGGAGGTGAGACCCAAAGAAAGC-3′ and 5′-GGGTGTGGGGGATCCTCAGCTACCTGTGATGCCG-3′ and insertion into pSG5 plasmid digested by EcoRI/BamHI (Stratagene, La Jolla, CA). The hCAR expression vectors were generated by PCR, using pCDM8-hCAR vector as a template (kindly provided by M. Negishi) and oligonucleotides sense hCAR/ATG 5′-CGGAATTCATGGCCAGTAGGGAAG-3′ and reverse hCAR/TGA 5′-AAAAAAGCGGCCGCCTCAGCTGCAGATCTCCTGG-3′ and cloned into plasmids pcDNA3 (Invitrogen, Groningen, The Netherlands) and pBSEN (
      • Pallisgaard N.
      • Pedersen F.S.
      • Birkelund S.
      • Jorgensen P.
      ) digested byEcoRI/NotI. The mCAR expression vector (pCR3-mCAR) was kindly provided by M. Negishi.

       Cell Culture, DNA Transient Transfections, and Reporter Gene Expression Assays

      Primary cultures of human hepatocytes were prepared and cultured as described elsewhere (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ). The HepG2 and HuH7 cell lines were obtained from the NIH ATCC repository (Bethesda, MD) and maintained without antibiotics. Human hepatocytes, HepG2, and HuH7 were transfected in suspension with Fugene-6 transfection reagent (Roche Molecular Biochemicals) according to the manufacturer's instructions; 80,000 cells were transfected with 250 ng of reporter plasmids, 25 ng of pSV-β-galactosidase control vector (Promega, Madison, WI), and either 25 ng of pSG5 (Stratagene, La Jolla, CA) or pSG5-hGR or pCR3-mCAR, or 250 ng of pBSEN or pBSEN hCAR expression vector in Opti-MEM I medium (Invitrogen). HepG2 and HuH7 cells were seeded in 24-well plates in DNA:liposome mix in DMEM supplemented with 10% fetal calf serum medium (Invitrogen) for 24 h. Then the medium was replaced by DMEM, and cells were treated for 16 h with dexamethasone, RU486, TCPOBOP, androstenol, rifampicin, PCN, or Me2SO. Luciferase and β-galactosidase assays were performed according to the specifications of the manufacturer (Promega).

       Site-directed Mutagenesis

      Site-directed mutagenesis was performed using a QuikChange™ site-directed mutagenesis kit (Stratagene) according to the recommendation of the manufacturer, and oligonucleotides including, a mutated 2C9-GRE-m1 (mutated bases in 5′-half underlined): 5′-GGTGGACCCAACTGCCCTGAATGTTTTGCTTGAAATGAAACC-3′, a mutated 2C9-GRE-m2 (mutated bases in 3′-half underlined): 5′-GGTGGACCCAACTGAACTGAATGCCTTGCTTGAAATGA-AACC-3′, a mutated 2C9-DR4-m1 (mutated bases in 5′-half underlined): 5′-CTAAATGTTATAAAACCCTTGTCTTCTGACCTCTCAATCTAGTC-3′, and a mutated 2C9-DR4-m2 (mutated bases in 3′-half underlined) 2C9-DR4-m2: 5′-GTTTATAAACCAAACTCTTCTCTGGTCTCAATCTAGTCAACTGGGG-3′.

       Electrophoretic Mobility Shift Assay (EMSA)

       EMSA with GR

      EMSA was performed using 5′-32P-labeled oligonucleotides TAT-GRE: 5′-GACCCTAGAGGATCTGTACAGGATGTTCTAGAT-3′ (Santa Cruz) or 2C9-GRE: 5′-ACCCAACTGAACTGAATGTTTTGCTTGAA-3′. Fifty thousand or 100,000 cpm of TAT-GRE or 2C9-GRE oligonucleotides, respectively, were incubated for 15 min at 4 °C in 12 μl of 10% glycerol, 10 mm Tris-HCl, pH 7.2, 100 mm KCl, 1 mm DTT in the presence of 4 μg of total proteins from Sf9 cells transfected with a recombinant GR-expressing baculovirus and Nonidet P-40 (0.1%) to minimize the binding of accessory proteins to GR. Competitions were performed with unlabeled oligonucleotides, including TAT-GRE, TAT-GRE mutant (5′-GACCCTAGAGGATCTCAACAGGATCATCTAGAT-3′), and a mutated 2C9-GRE-m3 mutant (mutated bases underlined: 5′-ACCCAACCAAACTGAATCATTTGCTTGAA-3′). For supershift assays, extracts were pre-incubated with 1 μg of GR antibody (Santa Cruz). Samples were loaded on a 4% polyacrylamide gel and submitted to electrophoresis at 20 mA in 0.25× TBE. The gel was analyzed using a PhosphorImager apparatus and ImageQuant software (Molecular Dynamics, Sunnyvale, CA).

       EMSA with CAR

      EMSA was performed using 5′-32P-labeled oligonucleotides 2B6-NR1 (5′-ACTGTACTTTCCTGACCCTGAAGA-3′) (
      • Sueyoshi T.
      • Kawamoto T.
      • Zelko I.
      • Honkakoski P.
      • Negishi M.
      ) or 2C9-DR4 (5′-AAACCAAACTCTTCTGACCTCTCA-3′). Fifty thousand cpm were incubated for 20 min at 4 °C in 20 μl of 10 mm Hepes, pH 7.9, 10% glycerol, 1 mmMgCl2, 0.1 mm EDTA, 100 mm KCl, and 0.3 mm DTT in the presence of 15 μg of total proteins from COS cells transfected or not with pcDNA3-hCAR. Competitions were performed with unlabeled oligonucleotides including 2B6-NR1, 2C9-DR4, 2C9–5′-CAR-RE (5′-GCCTTTGACTTACCTAAGTACTAAATGTTATAAAACC-3′, position −1856/−1818), 2C9–3′-CAR-RE (5′-AACCAAACTCTTCTGACCTCTCAATCTAGTCAACTGGGG-3′, position −1822/−1783), 2C9-DR4-m1 (5′-AAACCCTTGTCTTCTGACCTCTCA-3′), 2C9-DR4-m2 (5′-AAACCAAACTCTTCTCTTGTCTCA-3′), and 2C9-DR4-m3 (in which both half-sites were mutated). For supershift assays, extracts were pre-incubated with 1 μg of RXR antibody (Santa Cruz) or hCAR antibody (kindly provided by M. Negishi). Samples were loaded on a 4% polyacrylamide gel and submitted to electrophoresis at 20 mA in 0.25× TBE. The gel was analyzed using a PhosphorImager apparatus and ImageQuant software.

       EMSA with PXR

      EMSA was performed using 5′-32P-labeled oligonucleotides 3A4-ER6 (5′-TAGAATATGAACTCAAAGGAGGTCAGTGAGT-3′) (
      • Lehmann J.M.
      • McKee D.D.
      • Watson M.A.
      • Willson T.M.
      • Moore J.T.
      • Kliewer S.A.
      ) or 2C9-DR4 (described above). Fifty thousand cpm were incubated for 20 min at 20 °C in 20 μl of 5 mm Hepes, pH 7.8, 9% glycerol, 4 mm MgCl2, 0.05 mm EDTA, 1 mm DTT, 4 mm spermidine, 250 ng of dI-dC, and 1 μg of salmon sperm DNA in the presence of 2 μl of PXR and/or RXR proteins expressed using in vitro coupled transcription and translation (Promega). Competitions were performed using excess of unlabeled oligonucleotides, including 2C9-DR4, 3A4-ER6, and 3A4-DR3 (5′-GAATGAACTTGCTGACCCTCT-3′) (
      • Goodwin B.
      • Hodgson E.
      • Liddle C.
      ). Samples were loaded on a 4% polyacrylamide gel and submitted to electrophoresis at 20 mA in 0.25× TBE. The gel was analyzed using a PhosphorImager apparatus and ImageQuant software.

       Preparation and Analysis of Total RNA Extracted from Hepatocytes

      Total RNA was isolated using Trizol reagent (Invitrogen), from 107 cultured hepatocytes according to the manufacturer's instructions. For quality control, 30 μg of total RNA were analyzed by Northern blot using a rat glyceraldehyde-3-phosphate dehydrogenase cDNA probe (J. M. Blanchard, IGMM, Montpellier France). Tyrosine aminotransferase (TAT) mRNA was quantified by Northern blot using a specific probe (kindly provided by Dr. T. Grange, Institut J. Monod, Paris, France), and CYP2C9 mRNA was quantified by RNase protection assay as described (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ).

      RESULTS

       Identification of a Functional Glucocorticoid-responsive Element in the Regulatory Region of Gene CYP2C9

      In recent work, we observed that induction of CYP2C9 mRNA by dexamethasone paralleled that of TAT, a gene product known to be controlled by GR (
      • Schmid E.
      • Schmid W.
      • Jantzen M.
      • Mayer D.
      • Jastorff B.
      • Schutz G.
      ) in terms of time and concentration dependence in primary human hepatocytes (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ). This prompted us to look for a functional glucocorticoid-responsive element (GRE) in gene CYP2C9. For this purpose, the 5′-flanking region of CYP2C9 from −2088 to +21 was cloned upstream of a luciferase reporter gene driven by a SV40 promoter in pGL3 basic vector (p2088-luc), and several plasmids containing deletion constructs of this region were generated.
      The various plasmids harboring CYP2C9 deletions were transfected into HepG2 cells (in which GR expression is very low), with or without cotransfection of a plasmid expressing hGR (pSG5-hGR). Cells were then cultured for 16 h in the presence or absence of dexamethasone, and luciferase activity was measured. Plasmid pTAT-GRE-TK-luc harboring one GRE from TAT was used as control. As shown in Fig. 1, both p2088-luc and pTAT-GRE-TK-luc exhibited parallel response to increasing concentrations of dexamethasone (a plateau being reached at 10 nm) and similar extent of induction (15-fold). As these results suggested the presence of a GRE located on the −2088/+21 region of CYP2C9, a more detailed analysis of hGR-mediated transactivation of CYP2C9 deletion constructs was undertaken. The results are presented in Fig.2. No significant induction of the reporter gene was observed (2-fold induction maximum) without hGR transfection in any of the constructs examined, including the empty pGL3 basic vector. In contrast, when these plasmids were cotransfected with the hGR expression vector, a significant induction (10–20-fold) was observed in response to 100 nm dexamethasone for p2088-luc, p1856-luc, p1783-luc, and p1684-luc constructs. However, a major decrease in luciferase activity was observed with construct p1648-luc and shorter constructs including p1545-luc and p340-luc. These results suggest the existence of a GRE between −1684 and −1648.
      Figure thumbnail gr1
      Figure 1Transactivation of CYP2C9promoter by hGR. Plasmids p2088-luc (harboring the 2.1-kbCYP2C9 promoter) and pTAT-GRE-luc were cotransfected in HepG2 cells with pSG5 (control vector) or pSG5-hGR (hGR-expressing vector) and with pSV-β-galactosidase for transfection control. Cells were then treated with increasing concentration of dexamethasone for 16 h. Cell extracts were assayed for luciferase activity, which was normalized to β-galactosidase activity. Induction is expressed as the ratio of normalized luciferase activity in the presence of dexamethasone to this activity in the absence of dexamethasone.Error bars represent the standard deviations of five independent experiments.
      Figure thumbnail gr2
      Figure 2Identification of a functional GRE in theCYP2C9 promoter.Left panel, schematic representation of CYP2C9promoter constructs (from −2088 to +21). Right panel, transactivation of CYP2C9 promoter constructs by hGR. Plasmids harboring the CYP2C9 promoter constructs were cotransfected in HepG2 cells with pSG5 (control vector,white bars) or pSG5-hGR (hGR-expressing vector,black bars), and with pSV-β-galactosidase for transfection control. hGR was activated by treatment of cells with 100 nm dexamethasone for 16 h. Cell extracts were assayed for luciferase activity, which was normalized to β-galactosidase activity. Induction is expressed as the ratio of normalized luciferase activity in the presence of DEX to this activity in the absence of DEX.Error bars represent the standard deviations of 10 independent experiments.
      An oligonucleotide corresponding to the region −1684 to −1654 was then cloned in three copies in the pGL3-promoter to generate plasmid p2C9-(GRE)3-luc. As shown in Fig.3A, this construct was strongly transactivated by hGR (150-fold) in the presence of 100 nm dexamethasone. To verify that this transactivation was hGR-dependent, cells were treated with 1 μmRU486 (a prototypical hGR antagonist), in the absence or presence of dexamethasone. Although RU486 per se produced a moderate induction of luciferase activity, this compound drastically inhibited the dexamethasone-mediated transactivation of the construct. The full-length 2C9 promoter responded only modestly to dexamethasone when transfected in hepatocytes under the same conditions (data not shown). In parallel experiments carried out with plasmid pTAT-(GRE)2-TK-luc, similar observations were made (Fig.3B). Plasmid p2C9-(GRE)3-luc was also transfected in primary human hepatocytes and was transactivated by 100 nm dexamethasone as shown in Fig. 3C. Here again, RU486 completely suppressed dexamethasone-mediated transactivation.
      Figure thumbnail gr3
      Figure 3Effect of RU486 on transactivation of CYP2C9-GRE by hGR. HepG2 cells were cotransfected with pSG5 (control vector) or pSG5-hGR (hGR-expressing vector) and p2C9(GRE)3-luc (A) or pTAT(GRE)2-TK-luc (B) reporter vectors. Twenty-four hours later, cells were cotreated with 1 μmRU486 and/or 100 nm dexamethasone. C, human hepatocytes (culture FT187) were transfected with p2C9(GRE)3-luc and treated as indicated during 48 h. HepG2 cells and human hepatocytes were also transfected with pSV-β-galactosidase for transfection control. Cell extracts were assayed for luciferase activity, which was normalized to β-galactosidase activity. Induction is expressed as the ratio of normalized luciferase activity in the presence of dexamethasone and/or RU486 to this activity in the absence of dexamethasone or RU486.Error bars represent the standard deviations of three independent series of experiments.

       Characterization of the CYP2C9-GRE by Directed Mutagenesis

      Computer analysis of theCYP2C9−1684/−1654 region demonstrated the presence of two putative GRE half-sites separated by three nucleotides. To evaluate the role of these two sites, they were mutated sequentially by directed mutagenesis using p2088-luc as a template (Fig.4A). Plasmids p2088-GREwt-luc (wild type GRE sequence), p2088-GREm1-luc (mutations in the 5′-half-site of GRE), and p2088-GREm2-luc (mutations in the 3′-half-site of GRE) were then transfected in HepG2 cells in the presence of hGR expression vector and of 100 nmdexamethasone. As shown in Fig. 4B, the mutation of either half-site was sufficient to abolish transcriptional activation of the construct. The CYP2C9 region −1662/−1676 is hereafter referred to as 2C9-GRE.
      Figure thumbnail gr4
      Figure 4Effect of directed mutagenesis targeted to the GRE on transactivation of CYP2C9 promoter by hGR.A, schematic representation of theCYP2C9 promoter (−2088/+21) focusing on 2C9-GRE element (wt) and mutated 5′ half-site (m1) and 3′ half-site (m2). Mutants were obtained by directed-mutagenesis as described under “Experimental Procedures.”B, p2088-GREwt-luc, p2088-GREm1-luc, p2088-GREm2-luc, p1648-luc, or empty reporter plasmid (pGL3-basic) were cotransfected with pSG5-hGR expression vector in HepG2 cells. Twenty-four hours later, cells were treated with 100 nm dexamethasone for 16 h. HepG2 cells were also transfected with pSV-β-galactosidase for transfection control. Cells extracts were assayed for luciferase activity, which was normalized to β-galactosidase activity. Induction is expressed as the ratio of normalized luciferase activity in the presence of dexamethasone to this activity in the absence of dexamethasone. Error bars represent the standard deviations of five independent series of experiments.

       Characterization of the 2C9-GRE by Gel Shift Assays

      To determine whether hGR interacts directly with 2C9-GRE, a gel shift analysis was performed using baculovirus-expressed hGR. In a control experiment (Fig. 5A), we verified that the consensus TAT-GRE effectively binds hGR. A clear band revealing the complex was observed (lane 2), and this band was supershifted by anti-GR antibodies, as expected (lane 3). TAT-GRE oligonucleotide efficiently competed with itself (lanes 4 and 5), whereas the TAT-GRE mutant (see “Experimental Procedures”) did not (lane 6). 2C9-GRE oligonucleotide was a modest competitor for this binding (lanes 7 and 8), whereas the mutant 2C9-GREm3 (mutations in both half-sites of GRE, see “Experimental Procedures”) was not (lane 9). Similar experiments carried out with 2C9-GRE are shown in Fig. 5B. The efficient binding of hGR to 2C9-GRE is revealed by a clear band (lane 2); this band was supershifted by anti-hGR antibodies (lane 3). TAT-GRE in excess efficiently competed with this binding (lanes 4 and5), whereas mutated TAT-GRE did not (lane 6). These observations suggest that 2C9-GRE binds to hGR, although with a lower affinity than for TAT-GRE.
      Figure thumbnail gr5
      Figure 5Analysis of 2C9-GRE binding to hGR by electrophoretic mobility shift assay.A, analysis of TAT-GRE binding to hGR. Radiolabeled TAT-GRE oligonucleotide (50,000 cpm 32P) was incubated in the absence (lane 1) or presence (lanes 2–9) of DEAE-dextran-purified extracts of Sf9 cells transfected with recombinant hGR baculovirus (Bac hGR) before loading onto the gel. In parallel experiments, incubation was performed in the presence of anti-hGR antibodies (Ab-hGR, 1 μg, lane 3), 10–100-fold molar excess of unlabeled TAT-GRE (lanes 4 and5), 100-fold excess of mutant TAT-GRE (lane 6), 10–100-fold molar excess of unlabeled 2C9-GRE (lanes 7 and8), or 100-fold excess of double mutant 2C9-GREm3 (lane 9). B, analysis of 2C9-GRE binding to hGR. Radiolabeled 2C9-GRE oligonucleotide (100,000 cpm 32P) was incubated in the absence (lane 1) or presence (lanes 2–6) of DEAE-dextran-purified extracts of Sf9 cells transfected with recombinant hGR baculovirus (Bac hGR) before loading onto the gel. In parallel experiments, incubation was performed in the presence of anti-hGR antibodies (Ab-hGR, 1 μg, lane 3), 10–100-fold molar excess of unlabeled TAT-GRE (lanes 4 and 5), or 100-fold excess of mutant TAT-GRE (lane 6). S, shift; SS, supershift.

       Effect of Cycloheximide on CYP2C9 mRNA Induction in Primary Human Hepatocytes

      hGR is expressed constitutively in our hepatocyte cultures (
      • Pascussi J.M.
      • Drocourt L.
      • Fabre J.M.
      • Maurel P.
      • Vilarem M.J.
      ,
      • Pascussi J.M.
      • Gerbal-Chaloin S.
      • Fabre J.M.
      • Maurel P.
      • Vilarem M.J.
      ). We therefore anticipated that, if induction of CYP2C9 mRNA by dexamethasone is primarily mediated by hGR, protein synthesis should not be necessary for this process. To test this hypothesis, human hepatocytes were cultured in a dexamethasone-depleted medium for 48 h and then re-exposed to 100 nm dexamethasone for 24 h, either in the absence or presence of 10 μg/ml cycloheximide, a typical inhibitor of protein synthesis (control experiments indicated that protein synthesis is inhibited by more than 90% in these conditions). Analysis of CYP2C9 mRNA expression by RNase protection assay is reported in Fig.6. Clearly, cycloheximide affected neither CYP2C9 nor TAT mRNA induction by dexamethasone. In contrast, induction of CYP2C9 mRNA by rifampicin and phenobarbital was inhibited by 50% and by more than 75%, respectively, in cells cultured in the presence of cycloheximide. Similarly, induction of CYP3A4 mRNA by rifampicin was inhibited by more than 75% in cycloheximide-treated cells (not shown). In aggregate, our previous studies (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ), the data reported above on CYP2C9-GRE characterization, and the absence of any effect of cycloheximide on mRNA induction demonstrate that CYP2C9 is a primary hGR-responsive gene, in cultured human hepatocytes.
      Figure thumbnail gr6
      Figure 6Effect of cycloheximide on induction of CYP2C9 mRNA by dexamethasone, rifampicin, and phenobarbital in primary human hepatocytes.Upper panel, human hepatocytes (culture FT176) were cultured for 48 h in a dexamethasone-depleted medium, and then re-exposed to 100 nm dexamethasone for 24 h, either in the absence or presence of 10 μg/ml cycloheximide. Similar experiments were repeated in the absence or presence of 10 μm rifampicin or 500 μm phenobarbital. Cells were then harvested, and total RNA was extracted and analyzed for CYP2C9 mRNA expression by RNase protection assay. MW, size markers; NP, native undigested riboprobe. Middle and lower panels, Northern blot analysis of glyceraldehyde-3-phosphate dehydrogenase and TAT mRNAs, respectively, in the same total cellular RNA extracts. Densitometric analysis of CYP2C9 mRNA bands indicated no significant effect of cycloheximide on dexamethasone induction, and 50 and 75% reduction on rifampicin and phenobarbital induction, respectively.

       Identification of a CAR-responsive Element in the CYP2C9 Gene

      We have previously observed that induction of CYP2C9 mRNA in response to phenobarbital in primary human hepatocytes parallels that of CYP2B6 mRNA in terms of time and concentration dependence (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ). It was therefore suspected that CAR could account for the phenobarbital-mediated induction of CYP2C9.
      We looked for possible CAR-responsive element(s) in theCYP2C9 regulatory region by deletion analysis of transcriptional activity. The plasmids harboring the variousCYP2C9 deletion constructs described above were transfected into HuH7 cells, with or without cotransfection of a plasmid expressing either the mouse or human CAR (mCAR or hCAR). Cells were then cultured for 36 h and the luciferase activity measured. Plasmid p2B6-(NR1)3-luc, harboring three copies of the phenobarbital-responsive element (NR1) from CYP2B6 gene (
      • Sueyoshi T.
      • Kawamoto T.
      • Zelko I.
      • Honkakoski P.
      • Negishi M.
      ), was used as control. As shown in Fig. 7A, constructs p2088-luc and p1856-luc were transactivated to a similar extent by both mCAR and hCAR (6–8-fold induction). As expected, hCAR transactivated 2B6-NR1 (6-fold induction; data not shown, but see Fig. 9C). In contrast, construct p1783-luc and shorter constructs were not transactivated (data not shown). To confirm the implication of CAR in the transactivation of these fragments, p1856-luc was transfected into HuH7 cells, with or without cotransfection of mCAR or hCAR, and either in the absence or presence of androstenol (a deactivator of mCAR and hCAR) or TCPOBOP (an activator of mCAR). The results are shown in Fig.7B. Androstenol inhibited the mCAR-mediated transactivation of p1856-luc construct to a greater extent than the hCAR-mediated transactivation of this construct, as expected from previous observations (
      • Moore L.B.
      • Parks D.J.
      • Jones S.A.
      • Bledsoe R.K.
      • Consler T.G.
      • Stimmel J.B.
      • Goodwin B.
      • Liddle C.
      • Blanchard S.G.
      • Willson T.M.
      • Collins J.L.
      • Kliewer S.A.
      ). As expected, TCPOBOP was able to reverse the androstenol-mediated inhibition of p1856-luc transactivation by mCAR; this compound has no effect on hCAR (
      • Tzameli I.
      • Pissios P.
      • Schuetz E.G.
      • Moore D.D.
      ). In aggregate, these observations suggest that a CAR-responsive element (CAR-RE) is located in the region −1856/−1783.
      Figure thumbnail gr7
      Figure 7Basal transactivation of CYP2C9promoter by hCAR or mCAR.A, plasmids harboringCYP2C9 promoter constructs (p2088-luc, p1855-luc, and p1783-luc) were transfected in HuH7 cells with or without cotransfection of human (pBSEN-hCAR) or mouse CAR (pCR3-mCAR) expression vectors and with pSV-β-galactosidase for transfection control. Forty-eight hours later, cellular extracts were assayed for luciferase activity, which was normalized to β-galactosidase activity. Basal transactivation represents the ratio of normalized luciferase activity in the presence of CAR to this activity in the absence of CAR. Error bars represent the standard deviations of five independent experiments. B, plasmid p1856-luc was cotransfected with pCR3-mCAR or pBSEN-hCAR expression vectors in HuH7 cells, and with pSV-β-galactosidase for transfection control. Twenty-four hours later, cells were treated with 5 μm (mCAR) or 20 μm (hCAR) androstenol and/or 100 nm TCPOBOP (mCAR) for 16 h. Cell extracts were then assayed for luciferase activity which was normalized to β-galactosidase activity. Basal transactivation of the construct was arbitrarily fixed to 100% in the control (absence of androstenol or TCPOBOP). Error bars represent the standard deviations of five independent experiments. nd, not determined.
      Figure thumbnail gr9
      Figure 9Effect of directed mutagenesis targeted to 2C9-DR4 on transactivation of CYP2C9 promoter by hCAR or mCAR.A, schematic representation of theCYP2C9 promoter (−2088/+21) focusing on the 2C9-DR4 motif (p2088-DR4wt-luc) and mutated 5′ half-site (p2088-DR4 m1-luc) and 3′ half-site (p2088-DR4 m2-luc). Mutants were obtained by directed-mutagenesis as described under “Experimental Procedures.”B, p2088-DR4wt-luc, p2088-DR4 m1-luc, and p2088-DR4 m2-luc, were cotransfected in HuH-7 cells with either pCR3-mCAR or pBSEN-hCAR expression vectors and with pSV-β-galactosidase for transfection control. Forty-eight hours later, cell extracts were assayed for luciferase activity, which was normalized to β-galactosidase activity. C, plasmids p2B6(NR1)3-luc or p2C9(DR4)4-luc were cotransfected with pCR3-mCAR or pBSEN-hCAR expression vectors and activation was analyzed as inB. Relative basal transactivation of the constructs is represented by the ratio of normalized luciferase activity in the presence of CAR to this activity in the absence of CAR.Error bars represent the standard deviations of four independent series of experiments.

       Characterization of the 2C9-CAR-RE by Gel Shift Assays

      To further characterize this element, the −1856/−1783 fragment was split into two parts: a 5′ part from −1856 to −1818 (5′-CAR-RE) and a 3′ part from −1822 to −1783 (3′-CAR-RE) (Fig.8A). The ability of these oligonucleotides to interact with hCAR and compete with 2B6-NR1 was then tested by gel shift assays. We first used a radiolabeled 2B6-NR1 oligonucleotide and verified that this fragment binds hCAR (Fig.8B, lane 3), as described previously (
      • Sueyoshi T.
      • Kawamoto T.
      • Zelko I.
      • Honkakoski P.
      • Negishi M.
      ,
      • Honkakoski P.
      • Negishi M.
      ). This complex was supershifted by anti-RXR antibodies (lane 4) and did not form in the presence of anti-CAR antibodies (lane 5). As expected, an excess of unlabeled 2B6-NR1 competed with the radiolabeled probe for hCAR binding (lanes 6 and 7). The 5′ and 3′ parts of CAR-RE fragment were then used in competition experiments. Only the 3′ part competed efficiently with 2B6-NR1 (lanes 9 and 10), whereas the 5′ part did not (lane 8). Analysis of the nucleotide sequence of the 2C9–3′-CAR-RE revealed the presence of an imperfect DR4 motif between −1803 and −1818 (Fig. 8A). This motif is hereafter referred to as 2C9-DR4. An oligonucleotide harboring this DR4 strongly competed with 2B6-NR1 for hCAR binding (lanes 11and 12), whereas mutating either half-site of DR4 (DR4 m1 and DR4 m2) or both (DR4 m3) completely inhibited its properties as a competitor (lanes 13–15 for mutants DR4 m1–m3, respectively).
      Figure thumbnail gr8
      Figure 8Analysis of 2C9-CAR-RE binding to hCAR by electrophoretic mobility shift assay.A, schematic representation of the 2C9-CAR-RE (-1856/−1783) identified by deletion analysis of promoter transcriptional activity. The 3′ and 5′ parts of this region including nucleotides −1856 to −1818 and −1822 to −1783 are named 2C9–3′-CAR-RE and 2C9–5′-CAR-RE, respectively. The fragment from −1821 to −1799 is named 2C9-DR4. The putative DR4 element is written in bold. B, analysis of 2B6-NR1 binding to hCAR. Radiolabeled 2B6-NR1 oligonucleotide (50,000 cpm32P) was incubated in the absence (lane 1) or presence of 15 μg of nuclear extracts from untransfected COS cells (lane 2), or of hCAR-recombinant COS cells (COS hCAR, lanes 3–15) before loading onto the gel. In parallel experiments, incubation was performed in the presence of: anti-RXR antibodies (Ab-RXR, 1 μg, lane 4), anti-hCAR antibodies (Ab-hCAR, 1 μg, lane 5), 10–100-fold molar excess of unlabeled 2B6-NR1 (lanes 6 and 7), 100-fold excess of 2C9–5′-CAR (lane 8), 10–100-fold molar excess of unlabeled 2C9–3′-CAR (lanes 9 and 10), 10–100-fold molar excess of unlabeled 2C9-DR4wt (lanes 11 and 12), 100-fold excess of 2C9-DR4 m1 (mutant 5′ half-site, lane 13), 100-fold excess of 2C9-DR4 m2 (mutant 3′ half-site, lane 14), or 100-fold excess of 2C9-DR4 m3 (mutant both sites,lane 15). C, analysis of 2C9-DR4 binding to hCAR. Radiolabeled 2C9-DR4 oligonucleotide (50,000 cpm 32P) was incubated as indicated above (lanes 1–5). In parallel experiments, incubation was performed in the presence of 10-, 100-, or 200-fold molar excess of unlabeled 2B6-NR1 (lanes 6–8) or unlabeled ER6 (the CYP3A4 proximal ER6 element, lanes 9–11). S, shift; SS, supershift.
      The same experiments were repeated using a radiolabeled 2C9-DR4 wild-type oligonucleotide (Fig. 8C). This fragment efficiently complexed with hCAR (lane 3). The band revealing the complex was supershifted by anti-RXR antibodies (lane 4), and the binding was inhibited by anti-CAR antibodies (lane 5). Excess of 2B6-NR1 (lanes 6–8) or 3A4-ER6 (lanes 9–11) competed with 2C9-DR4. These experiments confirmed that a DR4 motif targeted by hCAR is located in the CYP2C9regulatory region between −1803 and −1818.

       Analysis of 2C9-DR4 Transcriptional Activity

      Next, the transcriptional activity of 2C9-DR4 was evaluated by transfection experiments. For this purpose, oligonucleotides DR4 m1 and DR4 m2 (Fig.9A) were used to mutate the p2088-luc template by directed mutagenesis. p2088-DR4wt-luc, p2088-DR4 m1-luc, and p2088-DR4 m2-luc were then cotransfected in HuH7 cells with or without mCAR or hCAR expression vector. As shown in Fig.9B, the mutation of a single half-site suppressed both hCAR- and mCAR-induced basal transactivation of p2088-luc.
      Oligonucleotide 2C9-DR4 was then cloned in four copies upstream of a luciferase reporter gene driven by a SV40 promoter in pGL3 promoter vector p2C9-(DR4)4. This plasmid was cotransfected with mCAR- and hCAR-expressing vectors in HuH7, and luciferase activity was measured 24 h later. Similar experiments were carried out using p2B6-(NR1)3 for comparison. Both reporter plasmids were strongly transactivated by mCAR (30–40-fold) and to a lesser extend by hCAR (5–7-fold), as shown in Fig. 9C. In sum, these experiments suggest that the DR4 motif located in the CYP2C9regulatory region between −1803 and −1818 is recognized and transactivated by hCAR and thus can account for induction of this gene by phenobarbital.

       Implication of PXR in CYP2C9 Regulation

      We have shown previously that CYP2C9 is inducible by rifampicin, a prototypical inducer of CYP3A4 and activator of hPXR (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ). Because CYP3A4 induction by rifampicin has been shown to be strictly dependent on PXR activation (
      • Lehmann J.M.
      • McKee D.D.
      • Watson M.A.
      • Willson T.M.
      • Moore J.T.
      • Kliewer S.A.
      ), we suspected that this receptor could be involved in the induction of CYP2C9 as well. To test this hypothesis, CYP2C9 p2088-luc and the deletion constructs were transfected in HepG2 or HuH7 with or without a hPXR expression vector. In parallel, control experiments were carried out using a reporter plasmid p3A4-(ER6)3-luc construct (harboring three copies of the CYP3A4 ER6 motif; Ref.
      • Lehmann J.M.
      • McKee D.D.
      • Watson M.A.
      • Willson T.M.
      • Moore J.T.
      • Kliewer S.A.
      ). Although p3A4-(ER6)3-luc was transactivated upon activation of PXR by rifampicin, as expected, neither of the CYP2C9constructs was transactivated by PXR in response to its classical activators, notably rifampicin or clotrimazole (data not shown).
      Sueyoshi et al. (
      • Sueyoshi T.
      • Kawamoto T.
      • Zelko I.
      • Honkakoski P.
      • Negishi M.
      ) demonstrated that PXR and CAR are able bind to and transactivate common DNA sequences, including CYP3A4-ER6. We therefore looked for the ability of PXR to bind to and transactivate the 2C9-DR4 motif described above. For this purpose, gel-shift assays were performed using 3A4-ER6 as a radiolabeled probe in the presence of RXRα and/or PXR produced by in vitro coupled transcription and translation (Fig. 10A). As reported previously, this fragment formed a complex with the heterodimer RXR/PXR (lane 4) and the band revealing this complex was supershifted by anti-RXR antibodies (lane 5). When used in excess, unlabeled 2C9-DR4 oligonucleotide competed with the 3A4-ER6 radiolabeled probe for the RXR/PXR heterodimer (lanes 6 and 7). Similar experiments carried out using 2C9-DR4 as a radiolabeled probe are shown in Fig. 10B. 2C9-DR4 formed a complex with RXR/PXR heterodimers (lane 4), and the band revealing this complex was supershifted by anti-RXR antibodies (lane 5). When used in excess, both 3A4-ER6 and 3A4-DR3 competed with the CYP2C9-DR4 radiolabeled probe for the RXR/PXR heterodimer (lanes 6–8 and 9–11, respectively).
      Figure thumbnail gr10
      Figure 10Analysis of 2C9-DR4 binding to PXR by electrophoretic mobility shift assay.A, analysis of 3A4-ER6 binding. Radiolabeled 3A4-ER6 oligonucleotide (50 000 cpm32P) was incubated in the absence (lane 1) or presence of 2 μl of RXR (lane 2), PXR (lane 3), or both proteins (lanes 4–7) produced by in vitro coupled transcription and translation before loading onto the gel. In parallel experiments, incubation was performed in the presence of anti-RXR antibodies (Ab-RXR, 1 μg, lane 5) or 10–100-fold molar excess of unlabeled 2C9-DR4 (lanes 6 and 7). B, analysis of 2C9-DR4 binding. Radiolabeled 2C9-DR4 oligonucleotide (50,000 cpm 32P) was incubated as indicated above (lanes 1–5). In parallel experiments, incubation was performed in the presence of 10-, 100-, or 200-fold molar excess of unlabeled 3A4-ER6 (lanes 6–8) or unlabeled 3A4-DR3. S, shift; SS, supershift.
      To determine whether PXR is able to transactivate the 2C9-DR4 motif, plasmids p2C9(DR4)4 and p3A4-(ER6)3 were transfected in HepG2 cells with or without mPXR or hPXR expression vectors. The cells were then treated for 16 h with either mPXR or hPXR activators, i.e. PCN or rifampicin, respectively, and luciferase activity was measured. In parallel experiments, the empty expression vector pSG5 and the empty reporter vector pGL3 promoter were used as controls. The data obtained from four independent series of experiments are collected in Table I. First, basal transactivation of 3A4-ER6 and 2C9-DR4 by mPXR was an order of magnitude greater than that observed in response to hPXR. Second, transactivation of 3A4-ER6 by mPXR was increased ∼3-fold upon PCN treatment, whereas, in comparison, transactivation of this construct by hPXR was less than doubled upon rifampicin treatment. This last point is in good agreement with observations reported by others, showing that transactivation of 3A4-ER6 is modest but greatly enhanced when this element is linked to the distal XREM harboring a DR3 motif (
      • Goodwin B.
      • Hodgson E.
      • Liddle C.
      ). Third, transactivation of 2C9-DR4 by mPXR was not affected by PCN treatment for unknown reason, whereas it was significantly (p = 0.0012) and reproducibly increased (1.6-fold) by hPXR in response to rifampicin treatment in the four series of experiments. In sum, these data suggest that PXR is able to bind to and modestly transactivate 2C9-DR4 element in response to rifampicin.
      Table IComparative analysis of 3A4-ER6 and 2C9-DR4 transcriptional activation by PXR
      OligonucleotidemPXRhPXR
      UTPCNUTRIF
      CYP3A4-ER615.9611.12.4
      13.353.60.91.5
      20.136.62.34.2
      4.214.91.31.6
      Average ± S.D.13.4 ± 6.741.5 ± 20.5
      Significance of difference between reporter activity in cells treated with inducerversus untreated cells: p value (pairedt test), 0.046.
      1.4 ± 0.62.4 ± 1.2
      Significance of difference between reporter activity in cells treated with inducer versusuntreated cells: p value (paired t test), 0.057.
      CYP2C9-DR47.88.00.91.5
      8.69.31.11.8
      5.72.81.11.6
      3.22.50.81.3
      Average ± S.D. 6.3 ± 2.4 5.7 ± 3.5
      Significance of difference between reporter activity in cells treated with inducer versusuntreated cells: p value (paired t test), 0.45 (no significant difference).
      1.0 ± 0.11.6 ± 0.2
      Significance of difference between reporter activity in cells treated with inducer versusuntreated cells: p value (paired t test), 0.0012.
      Plasmids p2C9(DR4)4 and p3A4-(ER6)3 were transfected in HepG2 cells with or without mPXR or hPXR expression vector. The cells were then treated for 16 h in the absence (UT) or presence of 10 μm PCN or rifampicin (RIF), respectively, and luciferase activity was measured and normalized with respect to β-galactosidase transfection control. Data represent the ratio of luciferase activity in PXR-transfected cells to luciferase activity in nontransfected cells, in four independent series of experiments. In parallel experiments, the empty expression vector pSG5 and the empty reporter vector pGL3 promoter were used as controls.
      1-a Significance of difference between reporter activity in cells treated with inducerversus untreated cells: p value (pairedt test), 0.046.
      1-b Significance of difference between reporter activity in cells treated with inducer versusuntreated cells: p value (paired t test), 0.057.
      1-c Significance of difference between reporter activity in cells treated with inducer versusuntreated cells: p value (paired t test), 0.45 (no significant difference).
      1-d Significance of difference between reporter activity in cells treated with inducer versusuntreated cells: p value (paired t test), 0.0012.

      DISCUSSION

      Here we report on the presence of two functional responsive elements in the regulatory region of gene CYP2C9, a GRE (imperfect palindrome at −1662/−1676) and a CAR-RE (DR4 motif at −1803/−1818) (Fig. 11). The presence of these two elements provides the mechanistic basis for the induction of CYP2C9 by dexamethasone (at submicromolar concentrations) and phenobarbital, respectively, in primary human hepatocytes. In addition, the finding that PXR may bind to and transactivate (although modestly) the 2C9-DR4 supports the fact that rifampicin is a known inducer of this gene both in vitro and in vivo(
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ,
      • Zilly W.
      • Breimer D.D.
      • Richter E.
      ).
      Figure thumbnail gr11
      Figure 11Comparison of 2C9-GRE and DR4 motifs with other motifs found on TAT or CYP2B6.
      The role of GR in the control of CYP gene expression has been the object of many investigations and controversies (
      • Schuetz E.G.
      • Guzelian P.S.
      ,
      • Schuetz E.G.
      • Schmid W.
      • Schutz G.
      • Brimer C.
      • Yasuda K.
      • Kamataki T.
      • Bornheim L.
      • Myles K.
      • Cole T.J.
      ). Although glucocorticoids are among the best inducers ofCYP3A, notably in rodents, available experimental data do not provide convincing evidence in favor of a direct an major implication of this receptor in CYP3A and CYP2Bgene regulation. However, other data clearly support the participation of GR in CYP gene induction. In particular, several groups have observed that in primary cultures of hepatocytes induction ofCYP2B and CYP3A genes in response to xenobiotics was enhanced by submicromolar concentrations of dexamethasone or physiological concentrations of glucocorticoids, which, by themselves, were not inducers (
      • Jarukamjorn K.
      • Sakuma T.
      • Miyaura J.
      • Nemoto N.
      ,
      • Parmentier J.H.
      • Schohn H.
      • Bronner M.
      • Ferrari L.
      • Batt A.M.
      • Dauca M.
      • Kremers P.
      ,
      • Sidhu J.S.
      • Omiecinski C.J.
      ,
      • Waxman D.J.
      • Morrissey J.J.
      • Naik S.
      • Jauregui H.O.
      ,
      • Wright M.C.
      • Wang X.J.
      • Pimenta M.
      • Ribeiro V.
      • Paine A.J.
      • Lechner M.C.
      ). We recently provided a likely scenario for these observations (
      • Pascussi J.M.
      • Drocourt L.
      • Fabre J.M.
      • Maurel P.
      • Vilarem M.J.
      ,
      • Pascussi J.M.
      • Gerbal-Chaloin S.
      • Fabre J.M.
      • Maurel P.
      • Vilarem M.J.
      ). Using primary cultures of human hepatocytes, we showed that expression of PXR, CAR, and their common heterodimeric partner RXR is positively regulated by dexamethasone and natural glucocorticoids. More detailed investigation led us to propose the hypothesis that GR controls the expression of PXR, CAR, and RXR, thus contributing indirectly to the inducible expression ofCYP genes, which are the targets of these nuclear receptors (see, for example, Fig. 6 (lanes 5, 6,9, and 10) for this effect).
      Computer analysis of the 2.1 kb of CYP2C9 5′-regulatory region revealed the presence of numerous GRE half-sites, most of which are nonfunctional (
      • de Morais S.
      • Schweikl H.
      • Blaisdell J.
      • Goldstein J.A.
      ). The 2C9-GRE element characterized here has a “classical” GRE structure with two half-sites separated by 3 nucleotides (Fig. 11). Both sites exhibit changes with respect to the consensus TAT-GRE: two in the 5′-half-site and one in the 3′-half. These changes might account for the lower affinity for GR of 2C9-GRE compared with TAT-GRE, as suggested by competition gel-shift experiments. Mutations of this GRE in either half-site or in both drastically decreased both its binding to and transactivation by hGR. In recent work, we observed that induction of CYP2C9 mRNA in response to dexamethasone paralleled that of TAT in terms of time and concentration dependence in primary human hepatocytes (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ). Notably, the maximum level of CYP2C9 mRNA was reached at 100 nmdexamethasone. In contrast, maximum induction of CYP3A4 mRNA in the same cells required at least 10–50 μm dexamethasone, whereas CYP2B6 mRNA was not inducible by dexamethasone alone. These observations are consistent with the fact that dexamethasone activates PXR only at concentrations greater than 10 μm (
      • Lehmann J.M.
      • McKee D.D.
      • Watson M.A.
      • Willson T.M.
      • Moore J.T.
      • Kliewer S.A.
      ) and is not a hCAR activator (
      • Moore L.B.
      • Parks D.J.
      • Jones S.A.
      • Bledsoe R.K.
      • Consler T.G.
      • Stimmel J.B.
      • Goodwin B.
      • Liddle C.
      • Blanchard S.G.
      • Willson T.M.
      • Collins J.L.
      • Kliewer S.A.
      ). In addition, we show here that induction of both CYP2C9 and TAT mRNAs by dexamethasone is insensitive to cycloheximide, whereas, in contrast, induction of CYP2C9 by rifampicin or phenobarbital is significantly reduced by this inhibitor. Collectively, the data obtained in the current and previous work (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ) demonstrate that CYP2C9 is a primary glucocorticoid-responsive gene. However, this is not the first report characterizing a functional GRE in CYP genes. Schuetzet al. (
      • Schuetz J.D.
      • Schuetz E.G.
      • Thottassery J.V.
      • Guzelian P.S.
      • Strom S.
      • Sun D.
      ) demonstrated the presence of a functional glucocorticoid-responsive element in the human CYP3A5 gene at position −891/−1109. However, instead of the “classical” palindromic structure (two half-sites separated by 3 nucleotides), they showed that this element was composed of two half-sites (TGTTCT) separated by 160 nucleotides. More recently, Pereira et al.(
      • Pereira T.M.
      • Carlstedt D.J.
      • Lechner M.C.
      • Gustafsson J.A.
      ) identified another functional GRE at position −1960 in the regulatory region of rat CYP3A1 (GGCACAnnnTGTTAT).
      Nuclear receptor binding elements are generally composed of two half-sites, more or less degenerated with respect to the consensus sequence AGGTCA (TGACCT in the opposite strand), organized either as direct, inverted or everted repeats, separated by a number of nucleotides varying from 1 to 7 (
      • Mangelsdorf D.J.
      • Evans R.M.
      ). The 2C9-DR4 identified here is located at −1810 bp upstream of the CYP2C9 transcription start site and organized in a direct repeat of two hexanucleotides separated by 4 bp (Fig. 11). Mutations of the element in either half-site or in both drastically decreased both its binding to and transactivation by hCAR. DR4 motifs were previously described in the NR1 sequence of CYP2B1/2 (
      • Kim J.
      • Kemper B.
      ), cyp2B10 (
      • Honkakoski P.
      • Negishi M.
      ), orCYP2B6 (
      • Sueyoshi T.
      • Kawamoto T.
      • Zelko I.
      • Honkakoski P.
      • Negishi M.
      ), as well as in human CYP3A4 (
      • Sueyoshi T.
      • Kawamoto T.
      • Zelko I.
      • Honkakoski P.
      • Negishi M.
      ,
      • Zelko I.
      • Negishi M.
      ). The 2C9-CAR element apparently exhibits a higher affinity for hCAR as compared with the 2B6-NR1 motif, as assessed by competition experiments in gel shift assay. The presence of such an element is in agreement with the fact that CYP2C9 mRNA is inducible by phenobarbital in parallel with CYP2B6 mRNA in terms of time and concentration dependence in primary cultures of human hepatocytes (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ). Thus, besides genes from the CYP2B and CYP3A subfamilies, the present work identifies CYP2C9 as a new target for CAR in man. It could be suspected that CYP2C9 induction by phenobarbital might be mediated by PXR as well, because phenobarbital is a PXR activator. However, the ED50 of phenobarbital for hPXR activation is >500 μm (
      • Moore L.B.
      • Parks D.J.
      • Jones S.A.
      • Bledsoe R.K.
      • Consler T.G.
      • Stimmel J.B.
      • Goodwin B.
      • Liddle C.
      • Blanchard S.G.
      • Willson T.M.
      • Collins J.L.
      • Kliewer S.A.
      ). We have shown (
      • Pascussi J.M.
      • Gerbal-Chaloin S.
      • Fabre J.M.
      • Maurel P.
      • Vilarem M.J.
      ) that phenobarbital induces the nuclear translocation of CAR in human hepatocytes at 100 μm. In addition, this compound is able to induce CYP2C9 mRNA in these cultures at concentrations in the range of 50 μm (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ). It is therefore suggested that, at concentration lower than 500 μm, phenobarbital inducesCYP2C9 via hCAR, whereas the implication of PXR at greater concentration (i.e. 1 mm) cannot be ruled out.
      CYP2C9 is inducible by rifampicin in vivo andin vitro (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ,
      • Zilly W.
      • Breimer D.D.
      • Richter E.
      ), but no fragment strongly responsive to PXR was found in the 2.1-kb region analyzed here. Our observations that PXR can bind to the 2C9-DR4 is interesting but might merely reflect the fact that nuclear receptors are able to bind with variable affinity to a variety of motifs including DR-3 and ER-6 (
      • Aumais J.P.
      • Lee H.S.
      • DeGannes C.
      • Horsford J.
      • White J.H.
      ). However, in response to rifampicin, hPXR produces a moderate (1.6-fold on average) but significant and reproducible induction of the transcriptional activity of 2C9-DR4. This supports the observation that CYP2C9 mRNA is induced by this compound in primary human hepatocytes and in vivo. It must be emphasized that, in the same series of experiments, we observed that induction of transcriptional activity of 3A4-ER6 by hPXR in response to rifampicin was of similar magnitude (1.7-fold, Table I). The modest activation of 2C9-DR4 by hPXR most likely explains the lack of activation of deletion constructs by this receptor and reflects a poor transactivation of this motif by hPXR in the in vitro context of these experiments. Another possibility is that another PXR-responsive element is present inCYP2C9 outside of the region analyzed here, i.e.upstream of −2.1 kb. A precedent for this possibility has been reported for CYP3A4 (
      • Goodwin B.
      • Hodgson E.
      • Liddle C.
      ). In this gene, the proximal PXR-responsive ER6 motif at position −160 has been shown to act cooperatively with a more distal element (XREM, harboring several nuclear receptor binding motifs including ER6 and DR3) located at approximately −7 kb, to mediate full induction of CYP3A4 by rifampicin. Our finding that PXR did not affect the transactivation activity of CAR on 2C9-DR4, either in the absence or presence of rifampicin, in co-transfection experiments (data not shown) is in favor of this possibility. This suggests that the affinity of hCAR for 2C9-DR4 is much greater than that of hPXR. Finally, the presence of hGR did not affect the modest transcriptional activity of hPXR in the context of the largest CYP2C9 promoter fragment −2088/+21 (data not shown).
      The presence of the two functional elements 2C9-GRE and 2C9-DR4 suggests a complex regulation of CYP2C9 in response to glucocorticoids and xenobiotics. First, the presence of the functional GRE implies that CYP2C9 expression should be constitutive under conditions where the GR is maintained in an activated state. Interestingly, CYP2C9 is one of the major forms ofCYP expressed in vivo (
      • Miners J.O.
      • Birkett D.J.
      ), and this gene is expressed constitutively in our hepatocyte culture system (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ), in which we have shown that GR is expressed constitutively and is functional (
      • Pascussi J.M.
      • Drocourt L.
      • Fabre J.M.
      • Maurel P.
      • Vilarem M.J.
      ,
      • Pascussi J.M.
      • Gerbal-Chaloin S.
      • Fabre J.M.
      • Maurel P.
      • Vilarem M.J.
      ). In contrast, for example, both CYP3A4 and CYP2B6 mRNAs rapidly decrease in these cultures to very low or undetectable levels, in the absence of a xenobiotic inducer. The finding that CYP2C9 mRNA is still detectable (although at very low level) in our hepatocyte cultures even in the absence of dexamethasone suggests that other transcription factors expressed in these cells, including, for instance, hepatocyte nuclear factors and/or CAAT/enhancer-binding proteins, are also involved in the transcriptional control of this gene. Interestingly, the implication of HNF4 has been clearly demonstrated in the regulation of CYP2C9 promoter activity (
      • Ibeanu G.C.
      • Goldstein J.A.
      ). Second, the presence of a CAR-RE (or CAR/PXR-RE) placesCYP2C9 among the group of genes regulated by xenobiotics such as phenobarbital (or rifampicin). In addition, activations by GR and by CAR/PXR should be additive. Again, this is what we observed in our primary hepatocyte cultures (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ,
      • Pascussi J.M.
      • Drocourt L.
      • Fabre J.M.
      • Maurel P.
      • Vilarem M.J.
      ,
      • Pascussi J.M.
      • Gerbal-Chaloin S.
      • Fabre J.M.
      • Maurel P.
      • Vilarem M.J.
      ) and here in co-transfection experiments including hGR and hCAR (data not shown). Thus, the maximum induction ratio of CYP2C9 mRNA in response to rifampicin or phenobarbital (3–4-fold) is lower than that of CYP3A4 or CYP2B6 mRNAs in response to the same compounds (>10-fold), merely because the basal level of CYP2C9 is greater (
      • Gerbal-Chaloin S.
      • Pascussi J.M.
      • Pichard-Garcia L.
      • Daujat M.
      • Waechter F.
      • Fabre J.M.
      • Carrere N.
      • Maurel P.
      ,
      • Pascussi J.M.
      • Gerbal-Chaloin S.
      • Pichard-Garcia L.
      • Daujat M.
      • Fabre J.M.
      • Maurel P.
      • Vilarem M.J.
      ).
      In conclusion, we have demonstrated in this work that CYP2C9is another member of the list of primary glucocorticoid-responsive and CAR/PXR-regulated genes.

      Acknowledgments

      We thank Drs. S. Kliewer, M. Negishi, L. Poellinger, and J. C. Nicolas for providing various plasmids and antibodies; Dr. F. Cadepond for providing the baculovirus-expressed hGR; and Dr. C. Young for careful reading of the manuscript.

      REFERENCES

        • Goldstein J.A.
        • de Morais S.
        Pharmacogenetics. 1994; 4: 285-299
        • Miners J.O.
        • Birkett D.J.
        Br. J. Clin. Pharmacol. 1998; 45: 525-538
        • Gerbal-Chaloin S.
        • Pascussi J.M.
        • Pichard-Garcia L.
        • Daujat M.
        • Waechter F.
        • Fabre J.M.
        • Carrere N.
        • Maurel P.
        Drug Metab. Dispos. 2001; 29: 242-251
        • Kliewer S.A.
        • Moore J.T.
        • Wade L.
        • Staudinger J.L.
        • Watson M.A.
        • Jones S.A.
        • McKee D.D.
        • Oliver B.B.
        • Willson T.M.
        • Zetterstrom R.H.
        • Perlmann T.
        • Lehmann J.M.
        Cell. 1998; 92: 73-82
        • Sueyoshi T.
        • Kawamoto T.
        • Zelko I.
        • Honkakoski P.
        • Negishi M.
        J. Biol. Chem. 1999; 274: 6043-6046
        • Bertilsson G.
        • Heidrich J.
        • Svensson K.
        • Asman M.
        • Jendeberg L.
        • Sydow B.M.
        • Ohlsson R.
        • Postlind H.
        • Blomquist P.
        • Berkenstam A.
        Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 12208-12213
        • Blumberg B.
        • Sabbagh W.J.
        • Juguilon H.
        • Bolado J.J.
        • van, M. C.
        • Ong E.S.
        • Evans R.M.
        Genes Dev. 1998; 12: 3195-3205
        • Lehmann J.M.
        • McKee D.D.
        • Watson M.A.
        • Willson T.M.
        • Moore J.T.
        • Kliewer S.A.
        J. Clin. Invest. 1998; 102: 1016-1023
        • Honkakoski P.
        • Negishi M.
        J. Biol. Chem. 1997; 272: 14943-14949
        • Kim J.
        • Kemper B.
        J. Biol. Chem. 1997; 272: 29423-29425
        • Goodwin B.
        • Hodgson E.
        • Liddle C.
        Mol. Pharmacol. 1999; 56: 1329-1339
        • Moore L.B.
        • Parks D.J.
        • Jones S.A.
        • Bledsoe R.K.
        • Consler T.G.
        • Stimmel J.B.
        • Goodwin B.
        • Liddle C.
        • Blanchard S.G.
        • Willson T.M.
        • Collins J.L.
        • Kliewer S.A.
        J. Biol. Chem. 2000; 275: 15122-15127
        • Honkakoski P.
        • Negishi M.
        Biochem. J. 1998; : 889-895
        • Kawamoto T.
        • Sueyoshi T.
        • Zelko I.
        • Moore R.
        • Washburn K.
        • Negishi M.
        Mol. Cell. Biol. 1999; 19: 6318-6322
        • Park Y.
        • Li H.
        • Kemper B.
        J. Biol. Chem. 1996; 271: 23725-23728
        • Trottier E.
        • Belzil A.
        • Stoltz C.
        • Anderson A.
        Gene (Amst.). 1995; 158: 263-268
        • Pallisgaard N.
        • Pedersen F.S.
        • Birkelund S.
        • Jorgensen P.
        Gene (Amst.). 1994; 138: 115-118
        • Schmid E.
        • Schmid W.
        • Jantzen M.
        • Mayer D.
        • Jastorff B.
        • Schutz G.
        Eur. J. Biochem. 1987; 165: 499-506
        • Pascussi J.M.
        • Drocourt L.
        • Fabre J.M.
        • Maurel P.
        • Vilarem M.J.
        Mol. Pharmacol. 2000; 58: 361-372
        • Pascussi J.M.
        • Gerbal-Chaloin S.
        • Fabre J.M.
        • Maurel P.
        • Vilarem M.J.
        Mol. Pharmacol. 2000; 58: 1441-1450
        • Tzameli I.
        • Pissios P.
        • Schuetz E.G.
        • Moore D.D.
        Mol. Cell. Biol. 2000; 20: 2951-2958
        • Zilly W.
        • Breimer D.D.
        • Richter E.
        Eur. J. Clin. Pharmacol. 1977; 11: 287-293
        • Schuetz E.G.
        • Guzelian P.S.
        J. Biol. Chem. 1984; 259: 2007-2012
        • Schuetz E.G.
        • Schmid W.
        • Schutz G.
        • Brimer C.
        • Yasuda K.
        • Kamataki T.
        • Bornheim L.
        • Myles K.
        • Cole T.J.
        Drug Metab. Dispos. 2000; 28: 268-278
        • Jarukamjorn K.
        • Sakuma T.
        • Miyaura J.
        • Nemoto N.
        Arch. Biochem. Biophys. 1999; 369: 89-99
        • Parmentier J.H.
        • Schohn H.
        • Bronner M.
        • Ferrari L.
        • Batt A.M.
        • Dauca M.
        • Kremers P.
        Biochem. Pharmacol. 1997; 54: 889-898
        • Sidhu J.S.
        • Omiecinski C.J.
        Pharmacogenetics. 1995; 5: 24-36
        • Waxman D.J.
        • Morrissey J.J.
        • Naik S.
        • Jauregui H.O.
        Biochem. J. 1990; 271: 113-119
        • Wright M.C.
        • Wang X.J.
        • Pimenta M.
        • Ribeiro V.
        • Paine A.J.
        • Lechner M.C.
        Mol. Pharmacol. 1996; 50: 856-863
        • de Morais S.
        • Schweikl H.
        • Blaisdell J.
        • Goldstein J.A.
        Biochem. Biophys. Res. Commun. 1993; 194: 194-201
        • Schuetz J.D.
        • Schuetz E.G.
        • Thottassery J.V.
        • Guzelian P.S.
        • Strom S.
        • Sun D.
        Mol. Pharmacol. 1996; 49: 63-72
        • Pereira T.M.
        • Carlstedt D.J.
        • Lechner M.C.
        • Gustafsson J.A.
        DNA Cell Biol. 1998; 17: 39-49
        • Mangelsdorf D.J.
        • Evans R.M.
        Cell. 1995; 83: 841-850
        • Zelko I.
        • Negishi M.
        Biochem. Biophys. Res. Commun. 2000; 277: 1-6
        • Aumais J.P.
        • Lee H.S.
        • DeGannes C.
        • Horsford J.
        • White J.H.
        J. Biol. Chem. 1996; 271: 12568-12577
        • Ibeanu G.C.
        • Goldstein J.A.
        Biochemistry. 1995; 34: 8028-8036
        • Pascussi J.M.
        • Gerbal-Chaloin S.
        • Pichard-Garcia L.
        • Daujat M.
        • Fabre J.M.
        • Maurel P.
        • Vilarem M.J.
        Biochem. Biophys. Res. Commun. 2000; 274: 707-713