Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The cytochrome P450 (P450) proteins comprise a superfamily of heme-containing enzymes that are involved in the oxidative metabolism of many diverse hydrophilic compounds, including steroids, fatty acids, retinoids, and foreign chemicals, such as pharmaceutical drugs and other xenobiotics (Gonzalez, 1989; Dogra et al., 1998; Waxman, 1999). Transcriptional expression of P450 genes can be modulated both by endogenous compounds and structurally diverse exogenous compounds. Of particular interest is the tissue-selective transcriptional induction of P450 genes in response to xenobiotic inducers. We are studying the molecular mechanism by which phenobarbital, the sedative drug, induces expression of specific chicken P450 genes particularly in the liver (Hansen et al., 1989; Dogra et al., 1998; Dogra et al., 1999).

Xenobiotic-mediated mechanisms of induction of P450 genes have been identified that involve orphan receptors belonging to the nuclear receptor superfamily (Savas et al., 1999; Waxman, 1999). Of particular note is the constitutive androstane receptor (CAR), which mediates induction of CYP2B genes in response to phenobarbital and other chemicals (Honkakoski et al., 1998; Kawamoto et al., 1999) and pregnane X receptor (PXR) that activates CYP3A genes in response to many steroids and chemicals, including phenobarbital (Kliewer et al., 1999; Moore et al., 2000). The elegant work from Honkakoski et al. (1998), and more recently Handschin et al. (2000), has defined the central role of CAR in the mechanism underlying phenobarbital induction of rodent CYP2B and chicken CYP2H1 genes. A model of drug induction has been envisaged in which phenobarbital action results in the translocation of CAR into the nucleus, where it interacts with CAR binding sites in the 5' flanking regions of the responsive P450 genes (Honkakoski et al., 1998; Kawamoto et al., 1999; Tzameli et al., 2000). Similarly, CYP3A genes are activated in response to phenobarbital through the binding of PXR to the 5' flanking regions (Moore et al., 2000).

Genes from the CYP2B and CYP3A families can be induced not only by phenobarbital but also by the synthetic glucocorticoid agonist, dexamethasone (Honkakoski and Negishi, 1998; Schuetz et al., 2000). A long-standing question is whether the glucocorticoid receptor is involved in the dexamethasone-mediated induction of these genes. Recent studies using glucocorticoid receptor-null mice have established that this receptor is absolutely required for dexamethasone induction of CYP2B genes but not for CYP3A genes, where PXR mediates hormone induction (Schuetz et al., 2000). Because the glucocorticoid/progesterone antagonist RU486 inhibits phenobarbital induction of the rat CYP2B genes (Shaw et al., 1993), there is also the issue of whether the glucocorticoid receptor is involved in this induction mechanism in rodents. Interestingly, glucocorticoid receptor is not essential for phenobarbital induction of CYP2B and CYP3A genes in mice, which proceeds in glucocorticoid receptor null mice, although at diminished levels of induction (Schuetz et al., 2000). These findings are compatible with induction mechanisms involving CAR and PXR together with cooperative interactions from glucocorticoid receptor bound to known glucocorticoid responsive elements in the 5' flanking regions of these genes (Stoltz et al., 1998; Schuetz et al., 2000).

In the present study, we have investigated the chicken CYP2H1 gene, which is markedly induced by phenobarbital (Hansen et al., 1989; Dogra and May, 1996). Earlier, using transient transfection studies, we and others have characterized an upstream enhancer region that responds to phenobarbital and maximum induction dependent upon the binding of multiple proteins, including CAR (Dogra et al., 1999; Handschin and Meyer, 2000). We now report a novel situation, in which RU486 at high doses inhibits the phenobarbital-induced activation of the CYP2H1 gene; this inhibition does not seem to be mediated through either glucocorticoid/progesterone or PXR receptor. We discuss these findings in relation to the proposed mechanism of phenobarbital induction of this gene.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials and Plasmids. Phenobarbital was obtained from Faulding (South Australia), [³²P]dCTP and oligonucleotides were from GeneWorks Pty Ltd (South Australia). D-Threo-[dichloroacetyl-1-14C]chloroamphenicol (specific activity, 50-60 mCi/mmol) was from Amersham Pharmacia Biotech UK, Ltd. (Little Chalfont, Buckinghamshire, UK). Dexamethasone and cAMP were from Sigma (St. Louis, MO). Medroxyprogesterone acetate was a gift from Dr. Davidson (Flinders University, SA, Australia). RU486 was a kind gift from Roussel Uclaf (Romain Ville, France). Trichostatin A (TSA) was from Wako Pure Chemicals (Tokyo, Japan). Fertile white Leghorn chicken eggs were obtained from Hi-Chick (Gawler, SA, Australia). Construction of p4.8-SVCAT, p556-SVCAT, and pCYP-205CAT has been described previously (Dogra and May, 1997; Dogra et al., 1999). MMTV-luciferase was constructed by digesting pMSG (Pharmacia, Peapack, NJ) with HindIII/SmaI to release a 1.5 kb MMTV long terminal repeat fragment that was cloned in pBluescript. This construct was then digested with KpnI/SmaI to release MMTV, which was cloned into the multiple cloning site of pGL3-Basic vector (Promega, Madison, WI). The CAR-SV-luciferase and p556-SV-luciferase reporter plasmids were constructed by cloning four CAR binding sites from the phenobarbital-responsive enhancer (1637 to 1622) (Handschin and Meyer, 2000) and 556-bp enhancer region of the chicken CYP2H1 gene into a SalI site upstream of SV-luciferase in the pGL3 promoter vector (Promega), respectively. The sequence of the oligonucleotide containing two copies of the CAR binding site used in the construction of the CAR-SV-luciferase is: 5'-TCGAGTGAACTTCCTTGCCCTTCCTTGAACTTGAACTTCCTTGCCCTG-3'.

Preparation and Treatment of Primary Hepatocytes. Primary hepatocytes were prepared from 17-day-old chick embryos by a method described previously (Dogra and May, 1997). Briefly, chick embryos were decapitated and livers cannulated and perfused with 10 ml of EDTA/saline (2 mM EDTA in 0.9% NaCl) and 3 to 4 ml of 0.05% collagenase in Hanks' balanced salt solution. Pooled livers were incubated in fresh collagenase solution for 20 min, cut into small pieces, and further incubated at 37°C for 30 min with gentle pipetting every 5 min. Cells were washed with Hanks' balanced salt solution and treated with Sassa solution to lyse contaminating erythrocytes. Finally, hepatocytes were washed with Williams' E medium and plated at 6 to 8 × 10⁶ in 60-mm dishes in Williams' E medium supplemented with 10% Serum Supreme (Edward Keller Australia, Hallam, Victoria, Australia). Medium was changed after over night culture and cells were treated either with solvent only or with RU486 (40-100 µM) for 1 h before the following chemicals were added: phenobarbital (500 µM), TSA (1 µM), dexamethasone (1, 10, or 50 µM), or medroxyprogesterone acetate (100 nM) at the final concentrations shown. Hepatocytes were further incubated for 6 h and then were used to prepare total RNA (Chomczynski and Sacchi, 1987).

Northern Blot Analysis. Isolated total RNA was fractionated on formaldehyde-agarose gels, blotted onto NYTRAN membrane filters (Schleicher & Schuell), and UV cross-linked. The filters were prehybridized and then hybridized with cDNA probes labeled with [³²P]dATP by random priming using a DNA labeling kit (Pharmacia). The specific DNA probes were as follows: pCHPB15 for the CYP2H1 mRNA and a full-length cDNA clone for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA (Dogra and May, 1997). Each probe was added at an activity of 0.5 to 1.0 × 10⁶ cpm/ml. Filters were washed and quantified as described previously (Dogra and May, 1997).

Effect of RU486, Glutethimide, and AIA on Transiently Transfected Cells. Chick embryo primary hepatocytes (2 × 10⁷) were transfected by electroporation with 2 pmol of the following vectors: MMTV-Luciferase, p4.8-SVCAT, p556-SVCAT, or pCYP-205CAT. After transfection, each sample was split into two or three equal volumes and these were plated onto 60-mm dishes in Williams' E medium supplemented with 10% Serum Supreme (Edward Keller Australia) and 50 µg of gentamicin per ml of medium. Hepatocytes were incubated at 37°C overnight, after which medium was changed and hepatocytes were pretreated for 1 h with solvent only or RU486 (40 or 100 µM) before dexamethasone (0.1 µM), phenobarbital (500 µM), glutethimide (500 µM), or AIA (50 µg/ml) were added. The cells were further cultured for 48 h and then CAT or luciferase activities were determined.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

RU486 Inhibits Phenobarbital-Induced Increase in the Level of CYP2H1 mRNA. The effect of RU486 on the steady-state levels of mRNA for the CYP2H1 gene in phenobarbital-induced chick embryo hepatocyte cultures was examined. Hepatocytes were treated with RU486 (40-100 µM) for 1 h before the addition of phenobarbital at 500 µM. The cells were incubated for a further 6 h and total RNA isolated. The level of CYP2H1 mRNA was determined by Northern blot analysis using a probe specific for the CYP2H1 mRNA (Dogra and May, 1997). The filter was stripped and reprobed for GAPDH mRNA. The results for the autoradiographs of these filters are shown in Fig. 1A. Phenobarbital treatment produced an increase of approximately 9-fold in mRNA expression (lane 3 versus lane 1) as measured 6 h after phenobarbital addition. At concentrations up to 50 µM, RU486 had little effect on phenobarbital-induced expression of the CYP2H1 mRNA. However, at 100 µM, RU486 almost completely abolished mRNA induction (lane 7). This effect was seen in repeated experiments. RU486 at 100 µM had little effect on basal (noninduced) CYP2H1 mRNA levels (lane 2 versus lane 1). In addition, the level of GAPDH mRNA remained unchanged after treatment with phenobarbital or RU486 (up to 100 µM) either alone or in combination. As expected, the level of mRNA for the housekeeping form of 5-aminolevulinate synthase, the first enzyme of the heme biosynthetic pathway (May et al., 1995), was also induced by phenobarbital (Dogra and May, 1996), and this induced level was almost totally inhibited by RU486 (data not shown). It was possible that the inhibitory effect of RU486 on the phenobarbital induction could involve the recruitment of a corepressor with histone deacetylase activity (Heinzel et al., 1997; Nagy et al., 1997). However, as shown in Fig. 1B, Trichostatin A, a histone deacetylase inhibitor, did not reverse the inhibitory effect of RU486 on the CYP2H1 mRNA (lane 6 versus lane 2), suggesting that transcriptional silencing by RU486 does not involve a Trichostatin A-sensitive histone deacetylase. To test that the inhibition of phenobarbital induction does not result from RU486-mediated cell toxicity or irreversible cell damage, cells were pre-exposed to RU486 at 100 µM for 48 h and then inhibitor was removed by washing the cells. Subsequent induction of CYP2H1 mRNA by phenobarbital addition was identical to that of cells not pretreated with inhibitor (Fig. 1C, lane 3 versus lane 2). Hence, a high concentration of RU486 specifically inhibits phenobarbital induction of at least two genes, CYP2H1 and 5-aminolevulinate synthase (ALAS-1), but not basal expression of these genes, and the inhibitory response cannot be attributed to general cellular damage. In addition, we also examined the effect RU486 on the induction of the CYP2H1 mRNA by the phenobarbital-like inducers glutethimide and AIA. Glutethimide (500 µM) and AIA (50 µg/ml) induced the CYP2H1 mRNA level in a Northern blot to about the same level as phenobarbital. However, in marked contrast to the inhibitory action of RU486 on the phenobarbital response, it had no effect on the induction of CYP2H1 mRNA by these chemicals (data not shown).

View larger version (35K): [in this window] [in a new window]   Fig. 1.   Effect of RU486 on the induced levels of CYP2H1 mRNA. A, representative Northern blot analyses of steady-state levels of mRNAs are shown. Hepatocytes were pretreated with RU486 (40-100 µM), 1 h before the addition of PB at 500 µM. After a further 6-h incubation, total RNA was isolated and 15 µg was analyzed by Northern blotting. The filter was hybridized in turn with 32P-labeled probes specific for CYP2H1 (pCHB15) and GAPDH (control). Radiolabeled filters were quantified using a PhosphorImager (model 300A; Molecular Dynamics, Sunnyvale, CA), the level of CYP2H1 mRNA was standardized to that of GAPDH mRNA. B, chick embryo hepatocytes were pretreated with 100 µM RU486 for 1 h before the addition of PB at 500 µM with (lane 6) or without (lane 4) 1 µM TSA. Hepatocytes were also treated alone with 500 µM PB (lane 2), 100 µM RU486 (lane 3), 1 µM TSA (lane 5), or were untreated (lane 1). After a 6-h incubation, total RNA was isolated and analyzed for CYP2H1 and GAPDH mRNAs as described above. C, reversal of RU486 inhibition of CYP2H1 mRNA induction was analyzed in hepatocytes. Hepatocytes were untreated (lanes 1 and 2) or treated with RU486 at 100 µM for 48 h (lane 3), after which medium was replaced with medium lacking RU486. The cells were left for 18 h, at which time fresh medium without PB (lane 1) or with 500 µM PB (lanes 2 and 3) was added and the cells incubated for a further 6 h. Total RNA was isolated and analyzed for CYP2H1 and GAPDH mRNAs as described above.

Effect of Dexamethasone and Medroxyprogesterone Acetate on CYP2H1 mRNA Expression. The inhibitory action of RU486 raised the question of whether the glucocorticoid or the progesterone receptor is involved in phenobarbital induction of the CYP2H1 gene. To investigate this possibility, dexamethasone, a potent synthetic glucocorticoid agonist, was used. Chick embryo primary hepatocytes were treated with either 1 or 10 µM dexamethasone. The results of a Northern blot analysis of total RNA extracted from these cultures established that dexamethasone at a concentration of 10 µM failed to induce the endogenous gene (Fig. 2A, lane 4 versus lane 1). This finding is in contrast to the rat CYP2B1/2 genes, which are induced by dexamethasone (Kocarek et al., 1994; Sidhu and Omiecinski, 1995). To confirm that the induction of gene expression by dexamethasone can proceed in chick hepatocytes, a luciferase reporter construct driven by the dexamethasone-responsive MMTV long terminal repeat (MMTV-luciferase) was transfected into hepatocytes. Dexamethasone at a concentration of 0.1 µM induced this construct by about 2.7-fold above the control and importantly RU486 even at 25 µM significantly repressed induction (Fig. 2B) supporting the involvement of the glucocorticoid receptor. The effect of 1 or 10 µM dexamethasone on RU486-mediated inhibition of phenobarbital induction was examined next (Fig. 2C). It was found that dexamethasone did not reverse the inhibition by 100 µM RU486 (lanes 7 and 10), suggesting that RU486 does not suppress phenobarbital induction via a mechanism involving the glucocorticoid receptor. Dexamethasone at a concentration of 30 µM or more has been shown to activate orphan receptor PXR, which is known to bind CAR binding motif and thus may influence phenobarbital response. To test this possibility, chick primary hepatocytes were treated with 50 µM dexamethasone, alone or in combination with phenobarbital. RNA were prepared and analyzed by Northern Blot analysis. As shown in Fig. 2D, dexamethasone slightly induced CYP2H1 mRNA (which could be caused by activation by PXR), but it had little effect on the induction of CYP2H1 mRNA by phenobarbital (lane 4 versus 2; 18.8-fold versus 21-fold), indicating that activation of PXR does not affect phenobarbital response.

View larger version (42K): [in this window] [in a new window]   Fig. 2.   Effect of dexamethasone on the RU486-mediated inhibition of CYP2H1 mRNA induction. A, chick embryo primary hepatocytes were treated with either 1 or 10 µM Dex (lanes 3 and 4) or 500 µM PB (lane 2) for 6 h. Lane 1 is control RNA. After incubation, total RNA was isolated and analyzed for CYP2H1 and GAPDH mRNAs as described in Fig. 1. B, an MMTV-luciferase reporter construct was transfected into chick embryo hepatocytes by electroporation, each cuvette was halved, and then either Dex (0.1 µM) or Dex and RU486 (25 or 50 µM) were added to one dish and solvent was added to the control dish. The mean ± S.D. of three independent experiments repeated in duplicate is shown. Luciferase activity is represented as relative light units per 100 µg of protein. C, chick embryo primary hepatocytes were treated with either 1 or 10 µM Dex in combination with 500 µM PB or 100 µM RU486 or with PB and RU486 together. After an incubation of 6 h, total RNA was isolated and analyzed for CYP2H1 and GAPDH mRNAs as described above. Control cells were treated in the same way but were not initially exposed to Dex or RU486 (lanes 1 and 2) or PB (lanes 1 and 3). D, representative Northern blot analyses of steady-state levels of mRNAs are shown. Hepatocytes were treated with 500 µM of PB or 50 µM of Dex alone or together. After 6 h of incubation, total RNA was isolated and analyzed for CYP2H1 and GAPDH mRNA as described above in Fig. 1.

View larger version (48K): [in this window] [in a new window]   Fig. 3.   Effect of medroxyprogesterone acetate on CYP2H1 mRNA expression. Representative Northern blot analyses of the steady-state levels of mRNA is shown. Chick embryo hepatocytes were treated either with 500 µM of PB (lanes 1), 100 nM medroxyprogesterone acetate (MPA; lanes 3), or with PB plus medroxyprogesterone acetate plus 100 µM RU486 (lane 4). Lane 2 is control RNA. After 6 h of incubation, total RNA was isolated and analyzed for CYP2H1 and GAPDH (control) mRNAs as described in Fig. 1.

RU486 Inhibits Phenobarbital-Induction of CYP2H1/CAT Chimeric Constructs in Chick Hepatocytes. RU486 could inhibit phenobarbital-induced expression of the CYP2H1 gene through interference with either promoter activity (Dogra and May, 1997) or the activity of the upstream phenobarbital-responsive enhancer domain (Dogra et al., 1999). The effect of RU486 was investigated on the expression of the construct pCYP-205CAT, which contains the first 205 bp of the CYP2H1 promoter fused to the CAT gene as reporter. This construct has been shown to direct strong basal expression in transient transfection assays but does not respond to phenobarbital (Dogra and May, 1997). It has been postulated that this strong promoter is silenced in the native chromatin and that the drug releases this repression (Dogra and May, 1997). Transient transfection analysis showed that RU486 at 100 µM had no effect on the level of expression of pCYP-205CAT (Fig. 4, A and B), a finding that is in keeping with the Northern blot data, in which the basal level of CYP2H1 mRNA remained unaltered by RU486 treatment.

View larger version (46K): [in this window] [in a new window]   Fig. 4.   Effect of RU486 on the transient expression of p4.8-SVCAT, p556-SVCAT, and pCYP-205CAT. The p4.8-SVCAT, p556-SVCAT, and pCYP-205CAT constructs were transfected into chick embryo hepatocytes, each sample halved, and 500 µM PB or 40, 80, or 100 µM RU486 was added to one dish and solvent was added to the control dish. When analyzing the effect of PB and RU486 in combination, the control plate received PB (500 µM) and the other plate received PB (500 µM) and RU486 (40, 80, or 100 µM). A, a typical CAT assay of one such experiment is shown. B, the average ± S.D. of three independent experiments, repeated in duplicate and represented as a percentage of conversion of [14C]chloramphenicol to acetylated product is shown. C, the p556-SVCAT construct was transfected into chick embryo hepatocytes and each sample was divided into three equal parts and seeded in 60-mm dishes containing 5 ml of Williams' E medium plus 10% Serum Supreme. In each case, one dish was treated with solvent only for control and other dishes received either 500 µM PB or glutethimide (glut) or 50 µg/ml AIA with or without 100 µM RU486. A typical CAT assay is shown.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Schematic diagram of the 4.8-kb phenobarbital enhancer of CYP2H1 gene. Schematic diagram of the 4.8-kb phenobarbital enhancer region of CYP2H1 gene is shown. The 556-bp enhancer domain lies in the BglII/XhoI restriction fragment at the 3'-end of 4.8-kb enhancer region. Within the 556-bp enhancer domain, the 316-bp BglII/StuI region helps the phenobarbital-induced enhancer activity of the 240-bp region. Various transcription factor binding sites including CAR, CCAAT/enhancer-binding protein, HNF-1, and two unknown sites are shown.

RU486 Does Not Act through Any One of the Known Transcription Factors on the 556 bp Enhancer. In earlier work, we identified four transcription factor binding sites important for 556-bp enhancer activity, including HNF-1, CCAAT/enhancer-binding protein, and two unknown (Dogra et al., 1999). Mutagenesis of each site results in a lowering of phenobarbital induction by about 30%, whereas mutagenesis of multiple sites together almost completely abolishes induction (Dogra et al., 1999). To investigate whether repression of phenobarbital enhancer by RU486 is mediated through one of these binding sites, transient transfection assays, in which one of the four transcription factor binding-sites was mutated, were carried out in chick embryo primary hepatocytes using the 556-bp enhancer reporter constructs. In these assays, phenobarbital induction of the mutated enhancer constructs could still be inhibited by RU486 (data not shown). This suggested that RU486-mediated inhibition cannot be assigned to a single protein binding site in 556-bp enhancer.

View larger version (37K): [in this window] [in a new window]   Fig. 6.   Effect of RU-486 on the trans-activation by CAR. A luciferase reporter construct containing four copies of the CAR site upstream of the enhancerless SV40 promoter was cotransfected into LMH cells together with 50 ng of CAR expression vector and pRL-SV40 as an internal transfection control. Relative activity is expressed as a ratio of test to the control and as mean ± S.D. of three independent assays.

RU486 Could Affect Inhibition through Activation of PXR. A number of recent studies show that orphan nuclear receptor CAR plays a key role in mediating induction of genes by phenobarbital (Honkakoski et al., 1998; Sueyoshi et al., 1999). It has also been demonstrated that a CAR-related orphan receptor PXR can compete with CAR for binding to CAR binding sites on DNA (Xie et al., 2000). Therefore, RU486 could activate the orphan receptor PXR (Moore et al., 2000), which then competes with the binding of CAR (Xie et al., 2000) on the enhancer to abrogate the phenobarbital response. In an attempt to further evaluate this hypothesis, we investigated CAR-mediated expression of the 556-bp enhancer in transfected LMH cells in the presence of either RU486 (100 µM) or dexamethasone (50 µM), which are known to activate mPXR (Moore et al., 2000). As shown in Fig. 7, CAR activates the enhancer about 14-fold, and mPXR on its own has little effect (1.4-fold), whereas CAR and mPXR together activates this enhancer at a level about 65% of the activation with CAR alone. In the presence of dexamethasone or RU486, no significant effect on the CAR mediated activation of enhancer construct was observed. Also, induction of enhancer activity by CAR and mPXR together was of the same extent as with CAR alone. Induction of enhancer by mPXR alone in the presence of dexamethasone or RU486 was about 16-fold and 7.5-fold, respectively. From these data, we conclude that activation of mPXR by RU486 (or by dexamethasone) may not be responsible for the inhibitory action of RU486 on the CAR-mediated phenobarbital response.

View larger version (48K): [in this window] [in a new window]   Fig. 7.   Effect of PXR activation on the CAR induced induction of the 556-bp enhancer. LMH cells were transfected with p556-SV-luciferase reporter with or without 50 ng of CAR, 50 ng of mPXR expression clones alone or together, and pRL-SV40 as internal control. Transfected cells were treated with 50 µM Dex or 100 µM RU486. Cell extracts were assayed for luciferase activity. Relative activity is expressed as a ratio of test to the control and as a mean ± S.D. of three independent assays

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The most important finding in the present work is that RU486 strongly inhibits, in a dose-dependent fashion, the phenobarbital-induced expression of the endogenous hepatic CYP2H1 gene in chick embryo hepatocytes but does not affect basal expression of this gene (similar results were seen with the phenobarbital-induced ALAS1 mRNA). In this regard, our experiments with Trichostatin A indicated that chromatin remodeling is not a prerequisite for the inhibitory action of RU486. In marked contrast to phenobarbital effect, induction of the CYP2H1 mRNA by either of the two phenobarbital-type inducers, glutethimide or 2-allylisopropylacetamide, was unaffected by RU486, providing evidence that RU486 inhibitory action is drug selective.

Our Northern blot studies showed that dexamethasone was unable to antagonize the inhibitory effect of RU486 on drug induction of the CYP2H1 gene and, in addition, had no effect on either basal or phenobarbital-induced expression of the gene. The progesterone receptor specific activator (Smith et al., 1974) was also ineffective. These results demonstrate that the ability of RU486 to inhibit induction of the CYP2H1 gene is not caused by its antagonist activity toward the glucocorticoid or progesterone receptors. This is also further supported by the fact that no consensus glucocorticoid or progesterone receptor response elements are present in the 556-bp enhancer region. This situation is somewhat different from that of the rodent phenobarbital inducible CYP2B genes, in which functional glucocorticoid receptor elements have been located in the 5' flanking regions of the genes and contribute to the level of induced response (Stoltz et al., 1998; Schuetz et al., 2000).

Transient expression assays in transfected chick embryo hepatocytes demonstrated that the inhibitory target for RU486 lies in the upstream 556-bp drug-responsive enhancer region of the CYP2H1 gene and not in the proximal promoter, which is unresponsive to drug (Dogra and May, 1997). This finding agrees with the Northern blot data that basal expression driven by the promoter is not inhibited by RU486. Transcription factor binding-sites that participate in the phenobarbital-induced activation of 556-bp enhancer have been identified (Dogra et al., 1999). The differential effects of RU486 on the induction of 556-bp enhancer by phenobarbital and phenobarbital-type inducers (glutethimide and AIA) may be because one of these transcription factors is specifically activated by phenobarbital (not by glutethimide and AIA), and this activation is RU486-sensitive. Alternatively, phenobarbital, but not the phenobarbital-type inducers, could increase enhancer activity specifically through one of the enhancer proteins with RU486 affecting inhibition by promoting the binding of a repressor to this protein. However, this possibility seems unlikely, because RU486 does not alter the activity of any single protein, as demonstrated by analysis of phenobarbital-induced 556-bp enhancer constructs with mutations in each binding site.

The orphan receptor CAR binding plays a critical role in drug induction (Honkakoski et al., 1998; Handschin and Meyer, 2000) and a CAR binding site has been identified in the 556-bp enhancer region (Handschin and Meyer, 2000) that binds chicken xenobiotic-sensing orphan nuclear receptor, a homolog of CAR (Handschin et al., 2000). It was possible that RU486 inhibited phenobarbital induction through interference with CAR. An orphan receptor similar to CAR, PXR, is known to bind RU486 (Moore et al., 2000) and to compete with CAR for binding to CAR binding site (Xie et al., 2000). Hence, PXR activated by RU486 could prevent CAR binding to the enhancer. However, such a scenario did not explain why RU486 inhibits the action of phenobarbital and not that of the phenobarbital-type inducers (glutethimide and AIA), which presumably function through CAR. Also, as shown in Fig. 2D, treatment of chick primary hepatocytes with 50 µM dexamethasone, a concentration at which it is known to activate mPXR (Moore et al., 2000), did not affect induction of CYP2H1 mRNA by phenobarbital. These data suggested that activation of PXR does not seem to interfere with phenobarbital response.

Subsequent transient expression data in LMH cells on the induction of the 556-bp enhancer by exogenous CAR in the presence of mPXR, and neither of the two mPXR activators, RU486 and dexamethasone, supported the hypothesis of competition between these receptors. In fact, activated mPXR was a strong inducer of the enhancer. In other experiments with an artificial promoter containing multiple CAR binding sites, exogenous CAR-induced activation of this construct was not affected by RU486. We conclude from this work that RU486 inhibits phenobarbital induction at a step before enhancer activation by nuclear CAR.

Although CAR does not bind phenobarbital, it has been shown to bind TCPOBOP, a phenobarbital-type inducer of CYP2B genes (Tzameli et al., 2000). In the present system, TCPOBOP was not an inducer of the CYP2H1 gene. There is no information on the binding of glutethimide and AIA to CAR. One possibility is that RU486 inhibits a phenobarbital-dependent modification that is essential for nuclear import of CAR. For example, phenobarbital can activate a phosphatase that removes a phosphate group from CAR and allows nuclear import of CAR (Kawamoto et al., 1999) or it could activate a kinase (Dogra and May, 1996) that phosphorylates CAR to allow its nuclear import, with RU486 preventing phenobarbital action. Another possibility is that phenobarbital, but not glutethimide or AIA, must be metabolized to the true inducer (Moore et al., 2000), which then binds to CAR. RU486 could then inhibit the metabolism of phenobarbital to a true inducer. How RU486 could interfere with phosphorylation events or with the metabolism of phenobarbital is unclear.

A novel member of the steroid hormone receptor family could be involved in the inhibitory action of RU486. The requirement for a high concentration of RU486 may indicate weak binding to such a putative receptor; alternatively, RU486 may compete with a high-affinity endogenous ligand. What role such a receptor would play in the RU486 inhibitory action (e.g., by interfering with CAR phosphorylation events or the metabolism of phenobarbital) remains unclear. Nevertheless, our data provide evidence that although CAR may be central to the induction response, the mechanism by which the various phenobarbital-type inducers provoke CAR activation deserves further attention.