Specific and Overlapping Functions of the Nuclear Hormone Receptors

Speciﬁc and overlapping functions of the nuclear hormone receptors CAR and PXR in xenobiotic response

PWei1,2 ABSTRACT J Zhang1 The products of the cytochrome P450 (CYP) genes play an important role 1 in the detoxification of xenobiotics and environmental contaminants, and DH Dowhan many foreign chemicals or xenobiotics can induce their expression. We have YHan1 previously shown that the nuclear hormone receptor CAR (Constitutive DD Moore1 Androstane Receptor, NR113) mediates the well studied induction of CYP2B10 gene expression by phenobarbital (PB) and 1, 4-bis-[2-(3, 5,- 1Department of Molecular and Cellular dichloropyridyloxy)] benzene (TCPOBOP). We have used the CAR knockout Biology, Baylor College of Medicine, Houston, mouse model to explore the broader functions of this xenobiotic receptor. TX, USA; 2Present address: X-Ceptor Therapeutics, San Diego, CA, USA In addition to the liver, CAR is expressed in the epithelial cells of the villi in the small intestine, and this expression is required for CYP2B10 induction in Correspondence: response to PB and TCPOBOP in those cells. In agreement with previous DD Moore, Department of Molecular and observations that CAR can bind to regulatory elements in CYP3A genes, CAR Cellular Biology, Baylor College of is also required for induction of expression of CYP3A11 in response to both Medicine, One Baylor Plaza, Houston, TX 77030, USA PB and TCPOBOP in liver. In males, CAR is also required for induction of liver Tel: 713 798 3313 CYP2A4 expression. In wild type animals, pretreatment with the CAR inverse Fax: 713 798 3017 agonist androstenol blocks the response of both the CYP2B10 and CYP3A11 E-mail: mooreȰbcm.tmc.edu genes to PB and TCPOBOP, and decreases basal CYP3A11 expression. CAR is also required for the response of CYP2B10 to several additional xenobiotic inducers, including chlorpromazine, clotrimazole and dieldrin, but not dexamethasone, an agonist for both the xenobiotic receptor PXR (Pregnane X Receptor NR112) and the glucocorticoid receptor. Chlorpromazine induction of CYP3A11 is also absent in CAR-deficient animals, but the responses to clotrimazole and dieldrin are retained, indicating that both of these inducers can also activate PXR (Pregnane X Receptor NR112). We conclude that CAR has broad functions in xenobiotic responses. Some are specific to CAR but others, including induction of the important drug metabolizing enzyme CYP3A, overlap with those of PXR. The Pharmacogenomics Journal (2002) 2, 117–126. DOI: 10.1038/ sj/tpj/6500087

Keywords: xenobiotic; drug metabolism; induction; cytochrome P450; CAR; PXR

INTRODUCTION Cytochrome P450 (CYP) enzymes play a signiﬁcant role in metabolism of both endogenous hormones and structurally diverse foreign chemicals. Exposures to foreign compounds or xenobiotics can lead to increased expression and activity of many CYP proteins and other drug metabolizing enzymes.1–3 This induction of CYP genes can have a major impact in CYP-dependent drug metabolism and drug–drug interactions of foreign chemicals.4,5 One major group of CYP inducers, the structurally diverse phenobarbital (PB)- Received: 28 September 2001 like compounds, induces a subset of CYP genes within the CYP1A, 2A, 2B, 2C, Revised: 6 December 2001 and 3A family.1,2 A signiﬁcant advance in understanding the widely studied Accepted: 12 December 2001 induction of CYP2B genes by PB and a number of other PB-like inducers was the Regulation of gene expression by CAR PWeiet al 118

identification of an upstream element, termed the PB expression profile of CAR in various embryonic stages and response element module (PBREM), required for this 8-week-old mice. In the developing animals, ␤-galactosidase response.6,7 A role for the nuclear receptor CAR (constitutive expression was not seen in the liver until the perinatal stage androstane receptor, NR113) in this induction was suggested at E19. This correlates well with the extremely limited by the demonstration that this element contains two bind- inducibility of the PB responsive CYP2B1 and CYP2B2 genes ing sites for CAR/RXR heterodimers.8 Further studies dem- in the rat at this stage.18 As expected, liver ␤-galactosidase onstrated that the highly potent PB-like inducer, TCPOBOP expression was relatively high in adult animals (Figure 1A, (1,4-bis[2-(3,5-dichloropyridyloxy)]benzene) can increase a; 1B, a). In adults, ␤-galactosidase activity was also observed CAR transactivation9 and functions as an agonist ligand.10 in small intestine (Figure 1A, b; Figure 1B, b), and stomach The significance of this activation was firmly established by (Figure 1B, c). Histologically, ␤-galactosidase activity was the complete absence of induction of the CYP2B10 gene in observed in hepatocytes in the liver and the epithelial cells liver in response to PB or TCPPOBOP in mice lacking the of the villi in the small intestine (Figure 1B, a and b). CAR gene.11 These mice also lost the liver hypertrophic and Expression was also observed in an unidentified subset of hyperplastic responses to these inducers and showed cells in the sub-mucosal layer of the stomach that, based on decreased drug metabolizing capability. Thus, these findings their reported expression of CYP3A,19 may be parietal cells indicated that CAR is a novel xenobiotic receptor that con- (Figure 1B, c). No detectable ␤-galactosidase activity was fers PB responses in vivo. observed in the heart, lung, spleen, mammary gland, uterus The closest relative to CAR within the nuclear receptor or ovary (data not shown). superfamily is a protein referred to as PXR (pregnane X receptor, NR112)12 or SXR (in humans).13 An extensive series of studies have demonstrated that PXR is also a xenobiotic receptor that plays a key role in activation of CYP3A genes in response to a wide range of compounds.14–16 While the functions of CAR in regulation of CYP2B gene expression and PXR in regulation of CYP3A expression are now clear, several lines of evidence suggest that this specificity is not absolute, and that the functions of these two nuclear receptors overlap. Thus, biochemical studies have demonstrated that PXR can bind to the CAR binding sites in the PBREM, and that CAR can bind to the PXR response elements identified in CYP3A genes.10,17 Partial overlap has also been demonstrated at the level of ligand binding.16 These results raise a number of important questions regarding the specificity of xenobiotic responses. We have pursued this issue by addressing several aspects of the function of CAR. The construct used to generate the CAR-deficient animals replaces the initial coding exons of the CAR gene with a ␤-galactosidase reporter, and we have used this reporter to further characterize CAR expression. In addition to liver, CAR is expressed in the epithelial cells of the villi of the small intestine. CAR-deficient animals are also defective in induction of CYP2B10 gene expression in these cells in response to PB and TCPOBOP. We have also used the CAR knockout animals to identify both CYP2A4 and CYP3A11 as direct targets of CAR, and to examine the relative roles of CAR and PXR in response to a series of additional xenobiotic inducers. These studies both identify specific functions of CAR and PXR and confirm their functional overlap, particularly in regulation of CYP3A gene expression.

Figure 1 Tissue-specific expression of CAR. Tissues or sections of RESULTS liver, small intestine and stomach, 6- to 8-week-old CAR hetero- Tissue-specific Expression of CAR zygote mice were prepared for X-Gal staining. (A) Whole mount of 11 In the previously described CAR knockout animals, the ␤- adult liver (a) and small intestine (b). (B) Sections of liver (a) stom- galactosidase gene replaces the first two exons of the CAR ach (b) and small intestine (c). Representative microphotography gene. Thus, ␤-galactosidase expression serves as a reporter illustrated ␤-galactosidase activity in the hepatocytes of liver (a), for CAR promoter activity in these animals. X-Gal staining epithelial cells of small intestine (c), and a subset of cells in stom- with CARϩ/Ϫ heterozygous mice was used to examine the ach (b).

The Pharmacogenomics Journal Regulation of gene expression by CAR PWeiet al 119

CAR is Essential for Induction of the PB-responsive Genes We have previously shown that CAR mediates specific induction of the mouse CYP2B10 transcript by PB or TCPO- BOP in the liver.11 Using a monoclonal antibody against rat CYP2B1, a homologue of mouse CYP2B10, we also detected significant induction of CYP2B10 protein in PB- or TCPO- BOP-treated wild-type mice. This induction is completely lost in the CAR knockout mice (Figure 2b). Similar results were observed in the small intestine, where both inducers increased CYP2B10 expression in the wild-type but not the knockout animals (Figure 3a). As expected, induction of CYP2B mainly occurred in the epithelial cells of the villi, in which CAR is expressed (Figure 3b). Phenobarbital is well known to induce additional cytochrome p450s and other drug metabolizing enzymes.1–3 To begin to define the role of CAR in such responses, we studied expression of two additional genes reported to be induced by PB and TCPOBOP, CYP3A11 and CYP2A4.20 CYP3A11 was of particular interest because CAR has been shown to bind to the ER-6 element from human CYP3A4 and the DR- 3 element from rat CYP3A1,10,17 and because of the well known importance of CYP3A enzymes in drug metabolism. In the wild-type animals, both CYP3A11 mRNA and protein were induced by both xenobiotics, with a particularly strong response to TCPOBOP. Loss of CAR expression did not have a large effect on the somewhat variable basal levels of CYP3A11 expression. This is consistent with the fact that in untreated animals CAR is sequestered in the cytoplasm of Figure 3 CAR mediates induction of CYP2B10 by PB or TCPOBOP in small intestine. (a) 8-week-old mice were treated with corn oil PB (24 h) and TCPOBOP (6 h). Twenty micrograms of total small intestine RNA were used for Northern blot analysis. Mstands for male and F for female. The blot probed with CYP2B10 was exposed for 48 h, and the control 18S probe was exposed for 4 h. (b) Mice were treated with either corn oil or TCPOBOP for 3 days. Sections of small intestine were subjected to in situ hybridization with an 35S-labelled antisense CYP2B10 probe. Representative microphotography of sections illustrated TCPOBOP-enhanced CYP2B10 expression in the epithelial cells of small intestine in wild-type animals (b).

wild-type hepatocytes, and thus should not contribute to basal CYP3A11 expression. In contrast, the induction of CYP3A11 in the xenobiotic-treated groups was clearly absent in the CAR-deficient animals (Figure 4). These results demonstrate that CAR, like PXR,15 is able to activate CYP3A gene expression in response to appropriate inducers. Expression of CYP2A4 is sexually dimorphic, with high levels in female liver and much lower levels in male liver.21 Figure 2 Loss of activation of CYP2B10 by PB or TCPOBOP in the This is thought to be due to differential effects of pulsatile CAR-deficient mice. (a) 8-week-old mice were treated with corn oil (CO), PB or TCPOBOP for 3 days. Twenty micrograms of total liver and more constant levels of growth hormone in males and RNA were used for Northern blot analysis. The blot probed with females. Interestingly, CAR-mediated induction of CYP2A4 CYP2B10 was exposed for 24 h, and the control 18S probe was by PB or TCPOBOP was also sexually dimorphic: significant exposed for 4 h. (b) Whole protein extract was used for Western induction of CYP2A4 was observed in wild-type, but not blot analysis. At least three mice were used for each treatment and knockout male animals (Figure 5). The higher basal levels of one representative blot is shown. Mstands for male and F for expression in female animals were not responsive to either female. inducer (data not shown). It is unclear whether the differ-

www.nature.com/tpj Regulation of gene expression by CAR PWeiet al 120

Figure 6 Androstenol pretreatment blocks TCPOBOP activation in vivo. 8-week-old wild-type male mice were treated with corn oil, TCPOBOP (T), androstenol (A) alone, or both T and A. Lane 1, corn oil for 4 days (CO →→→); lane 2, one injection of T for 4 days (T Figure 4 PB or TCPOBOP activation of CYP3A11 by CAR. RNA or →→→); lane 3, one injection of T for 24 h followed by three injec- protein samples were prepared and analyzed as in Figure 2. Twenty tions of A (one injection per day) (T → A → A → A); lane 4, one micrograms of total RNA were used, the blot probed with CYP3A11 injection of T for 1 day (→→T →); lane 5, three injections of A for was exposed for 16 h and the control 18S probe was exposed for 3 days (one injection per day), and one injection of T was given 4h. in between the last two doses of A (A → A → T → A); lane 6, three injections of A for 3 days (one injection per day) (A → A → A →). Twenty micrograms of total liver RNA were subjected to Northern blot analysis. The blot probed with CYP2B10 was exposed for 24 h, the blot probed with CYP3A11 was exposed for 16 h and the control 18S probe was exposed for 4 h.

Androstenol Pretreatment Blocks the Response to TCPOBOP Previous studies show that while TCPOBOP acts directly as a CAR agonist, the inverse agonist androstenol can block TCPOBOP activation by dissociating the coactivator SRC-1 (steroid receptor coactivator-1) from the CAR ligand binding domain.10,22 To determine whether androstenol can also abolish TCPOBOP-enhanced expression of endogenous CAR target genes, we performed two sets of 4-day treatments with wild-type male mice. In the ﬁrst set, one group of mice received corn oil only at day 1 (CO → - → - → -). The other group was injected with one dose of TCPOBOP (0.3 mg kgϪ1) and treated for 4 days (T → - → - → -). The last group received three injections of androstenol (150 mg kgϪ1, one Figure 5 PB or TCPOBOP activation of CYP2A4 by CAR is male spe- injection per day) 24 h after TCPOBOP treatment ciﬁc. 8-week-old mice were treated with corn oil (CO), PB or TCPO- (T → A → A → A). As expected, in the absence of any ligands BOP for 3 days. Twenty micrograms of total liver RNA pooled from basal expression of CYP2B10 was undetectable and that of three mice samples were subjected to Northern blot analysis. The blot probed with CYP2A4 was exposed for 24 h and the control 18S CYP3A11 was low (Figure 6, lane 1). TCPOBOP treatment probe was exposed for 4 h. strongly stimulated expression of both genes (lane 2). And- rostenol treatment following TCPOBOP injection only slightly inhibited TCPOBOP activation of both genes (lane ences in growth hormone responses or other factors such as 3). differences in endogenous CAR ligands contribute to this In the second set of treatments, one group of mice was lack of response. We conclude that CAR mediates induction treated with TCPOBOP at day 3 (- → - → T → -). The other of all three of the CYP genes tested in response to the two group of mice was treated with three injections of andros- xenobiotics, and is likely to mediate the response of other tenol (one injection per day). These mice also received one genes to the PB-like inducers. dose of TCPOBOP in between the second and the third dose

The Pharmacogenomics Journal Regulation of gene expression by CAR PWeiet al 121

of androstenol (A → A → T → A). Twenty-four hours after the last injection of androstenol, the mice were killed. The last group of mice was treated with three doses of androstenol only (A → A → A → -). In striking contrast to the limited effects of the inverse agonist after TCPOBOP pretreatment, pretreatment with androstenol completely abolished TCPOBOP induction of both CYP2B10 and 3A11 genes (compared to lanes 4 and 5). Interestingly, androstenol alone substantially decreased CYP3A11 expression (lane 6). As just noted, complete loss of CAR activity does not affect basal expression of CYP3A11, which is consistent with the cytoplasmic localization of this receptor under ordinary cir- cumstances. However, treatment with inverse agonists results in generation of a transcriptionally silent CAR/RXR complex22 that could account for the direct negative effect of the androstenol-bound CAR. For example, this inactive complex could displace basal levels of activated PXR or other transcription factors from the CYP3A11 promoter. Although it is also possible that the inverse agonist-bound CAR functions as an active transcriptional repressor in this context, we have not observed such activity in transient transfections and have not observed inverse agonist-dependent recruitment of corepressors. Overall, these results demonstrate that the inverse agonist can block CAR activation in vivo, and suggest that this inhibition of CAR activity may significantly impair the capacity of the liver to respond to xenobiotics. Figure 7 Responses of CYP2B10 to diverse PB-like chemicals by CAR Mediates Responses of CYP2B and CYP3A Genes CAR. 8-week-old mice were treated with indicated compounds for to Diverse Chemicals 24 h. (a) Thirty micrograms of total liver RNA pooled were sub- It has been previously reported that a wide variety of com- jected to Northern blot analysis. The blot probed with CYP2B10 pounds, including drugs, organic solvents, polychlorinated was exposed for 72 h and the control 18S probe was exposed for 4 h. (b) Twenty micrograms of total liver RNA were subjected to biphenyls, pesticides and plant products can stimulate Northern blot analysis. The blot probed with CYP2B10 was exposed 23 expression of CYP2B10 in primary hepatocytes. In order for 48 h (control and dieldrin treated), 72 h (clotrimazole treated) to determine whether CAR mediates the responses of and 24 h (dexamethasone treated). The blot probed with CYP3A11 CYP2B10 to such structurally diverse inducers, we treated was exposed for 16 h (control and dieldrin treated), 24 h the animals with the antipsychotic agent chlorpromazine, (clotrimazole treated), and 16 h (dexamethasone treated). The the pesticide dieldrin, the antifungal clotrimazole, and the control 18S probe was exposed for 4 h. synthetic steroid dexamethasone. CYP2B10 mRNA levels were induced by all four compounds in wild-type animals (Figure 7). In the cases of chlorpromazine, dieldrin and clo- responses in the CAR knockout animals are based on PXR trimazole, this induction was completely absent in the CAR activation, although it remains possible that the glucocort- knockout mice, providing compelling evidence that CAR icoid receptor could contribute to the dexamethasone mediates CYP2B10 response to all three of these inducers. responses.24 As expected from its function as an activator of both the While TCPOBOP is a direct agonist ligand for CAR, PB glucocorticoid receptor and PXR, but not CAR, the dexame- does not appear to bind CAR. Instead, induction of CAR thasone response was retained in the CAR knockout target genes is thought to be an indirect consequence of mice.11,15,24 The lack of dexamethasone response in gluco- the translocation of CAR from cytoplasm to nucleus in corticoid receptor-deficient animals indicates that this PB-treated hepatocytes.25,26 Dexamethasone has also been receptor is the predominant inducer of CYP2B10 reported to induce nuclear translocation of CAR in pri- expression.24 mary human hepatocytes, although this does not result The effects of these compounds on CYP3A11 expression in induction of CYP2B gene expression.27 Two approaches were also examined in the wild-type and CAR knockout ani- were taken to determine whether the various compounds mals. Chlorpromazine treatment did not change CYP3A11 act as CAR or PXR ligands. In the first, the ability of the expression (data not shown). However, dieldrin, clotrima- compounds to activate chimeras consisting of the DNA zole, and dexamethasone greatly increased CYP3A11 mRNA binding domain of Gal4 fused to either CAR or PXR was levels in both wild-type and CAR null mice (Figure 7b). tested. In the second, the effect of the compounds on Based on previous results,12,16 we assume that all of these coactivator recruitment was examined using a mam-

www.nature.com/tpj Regulation of gene expression by CAR PWeiet al 122

malian two-hybrid assay based on chimeras consisting of expected, the activation of CAR by TCPOBOP can be pre- the VP16 activation domain joined to CAR or the PXR vented by the inverse agonist ligand androstenol, though ligand binding domain, and the Gal4 DNA binding the basis for the striking and unexpected effects of the order domain fused to the receptor interaction domain from the of addition of the inhibitor and the activator remain to be coactivator SRC-3. As expected, from previous results, the explained. Perhaps more importantly, we have also found potent agonist TCPOBOP efficiently activated the Gal4- that CAR is essential for CYP2B10 induction in response to CAR chimera (Figure 8a) and promoted coactivator bind- a wider range of compounds, including chlorpromazine, ing (Figure 8b), while PB did neither. The negative control dieldrin and clotrimazole. Thus, CAR is clearly able to retinoic acid and the PXR ligands PCN and dexame- mediate the effects of a diverse group of xenobiotic inducers. thasone were also inactive. However, chlorpromazine, Previous results have suggested a significant overlap clotrimazole and dieldrin all significantly activated the between the activators of murine CAR and PXR, raising a Gal4-CAR chimera and at least modestly stimulated coac- number of intriguing questions regarding specificity. This tivator recruitment. Thus, all three of these compounds issue is complicated substantially by the issue of species dif- appear to function directly as CAR agonists. ferences, since the sequences of the ligand binding domains With PXR, the best activation of the Gal4-PXR chimera of both of these xenobiotic receptors differ significantly (Figure 8c) and coactivator recruitment (Figure 8d) was seen between rodents and humans and the corresponding recep- with the known ligands PCN and dexamethasone. Although tors can show quite different responses to various com- SXR/hPXR has been reported to be activated by TCPOBOP pounds. For example, the human CAR protein does not and PXR, the murine PXR is not.13,14 Thus, the lack of effects appear to bind or respond to TCPOBOP, a highly potent of these compounds and retinoic acid was not surprising. agonist for murine CAR.10,16 Similarly, the rodent PXR is The lack of effect of chlorpromazine was consistent with the unable to respond to rifampicin, a potent agonist of lack of induction of CYP3A mRNA in the CAR knockout ani- SXR/human PXR.13,14 Within the context of the murine pro- mals. Thus, this compound, like TCPOBOP, is a specific acti- teins, however, the in vivo functional results described here vator of CAR. In contrast, modest but significant activation are quite consistent with previous primarily biochemical and coactivator recruitment was observed with clotrimazole, results demonstrating that CAR and PXR share several and a more potent response was observed with dieldrin. ligands.16 Functionally, this overlap may be clearest in the These results, which are also consistent with the CYP3A case of clotrimazole, which has now been examined in both responses in the CAR knockout animals, indicate that these CAR and PXR knockout animals. Our results demonstrate two compounds function as agonists for both CAR and PXR. that clotrimazole treatment results in CAR activation and Overall, we conclude that among the series of inducers of induction of CYP2B10 expression. Studies by Xie et al,17 CYP gene expression tested in the knockout animals, TCPO- demonstrate that clotrimazole-activated CAR can also con- BOP, PB and chlorpromazine are specific activators of CAR tribute to CYP3A11 expression, since this response is (though PB does not function as a CAR agonist), dexame- retained in the absence of PXR function. Finally, clotrima- thasone is a specific activator of PXR, and dieldrin and clo- zole, which activates both PXR and SXR/hPXR in transfec- trimazole activate both xenobiotic receptors. tions and binds directly to at least the latter,16 can also activate PXR in vivo, since the results described here show that DISCUSSION loss of CAR does not prevent the induction of the PXR tar- Drug exposure can lead to an increased expression of spe- get CYP3A11. cific CYP enzymes as well as other drug-metabolizing pro- Although the effects of dieldrin have not been examined teins. The induction of these genes by xenobiotic chemicals in the PXR-deficient animals, the results described here is a common defense mechanism against the toxicity and demonstrate that it can activate both receptors in transfec- carcinogenicity of foreign compounds. This xenobiotic tions and also promote coactivator recruitment to both. In response can also greatly augment the metabolism and vivo, it can also activate both CAR, as demonstrated by loss clearance of therapeutic drugs. Major advances in our under- of CYP2B10 induction in the CAR knockouts, and PXR, as standing of the molecular basis for the induction of CYP suggested by the retention of the CYP3A11 response in genes by xenobiotics are being provided by the characteriz- these animals. ation of the functions of the newly identified xenobiotic The potent activation of CYP3A11 expression in response receptors CAR and PXR/SXR.2,3 to TCPOBOP and its loss in CAR knockout animals clearly One of the best characterized of the xenobiotic responses demonstrates that CAR can activate CYP3A expression. is the induction of CYP2B and other genes by PB and a num- Thus, the functional overlap between CAR and PXR extends ber of other structurally unrelated compounds.1–3 Results to target genes. This conclusion is consistent with previous with CAR knockout animals have clearly demonstrated that studies demonstrating the ability of CAR/RXR heterodimers CAR mediates the induction of the CYP2B10 gene in to bind to and transactivate previously characterized bind- response to the inducers PB and TCPOBOP in the liver.11 In ing sites for PXR/RXR heterodimers in CYP3A genes.10,17 this study, we have confirmed this conclusion and extended Vice versa, PXR/RXR heterodimers have also been shown to it to show that CAR is also required for CYP2B10 induction bind to and transactivate the CAR/RXR binding sites from in the small intestine. We have also examined the role of the CYP2B10 or CYP2B6 genes.10,17,28 This activation of CAR CAR in the responses to several additional xenobiotics. As targets by PXR is consistent with the effects of the additional

The Pharmacogenomics Journal Regulation of gene expression by CAR PWeiet al 123

Figure 8 Xenobiotic activation and coactivator recruitment to CAR and PXR. Xenobiotic stimulation of CAR and PXR activation (a and c). A vector expressing GAL4-CAR LBD (a) or GAL4-PXR LBD (c) fusion protein was cotransfected into HepG2 cells with G5E1b-Luciferase as a reporter in the presence of DMSO (solvent), 250 nM TCPOBOP (TCP), 10 ␮MPregnenolone-16alpha-carbonitrile (PCN), 1 mMpheno- barbital (PB), 10 ␮Mchlorpromazine (CHP), 10 ␮Mclotrimazole (CTZ), 20 ␮Mdieldrin (DIEL), 10 ␮Mdexamethasone (DEX) and 1 ␮M all-trans Retinoic Acid (AtRA). No activation was observed when transfecting GAL4-DBD alone in identical experiments (data not shown). Xenobiotic compounds enhance CAR and PXR interaction with the SRC-3 coactivator in vivo (b and d). A vector expressing the Gal4- SRC-3 RID fusion protein was cotransfected into HepG2 cells with a vector expressing the VP16-CAR (b) or VP16-PXR-LBD (d) fusion protein and G5E1b-Luciferase, in the presence of various compounds (as described above). No interaction/activation was observed in the absence of either GAL4 or VP16 fusion constructs in identical experiments (data not shown). Fold activation is calculated by setting the luciferase activity of GAL4-CAR-LBD (a), GAL4-PXR-LBD (c), GAL4-SRC-3 RID/VP16-CAR (b), and GAL4-SRC-3 RID/VP16-PXR LBD (d) treated with DMSO to a value of 1.

PXR speciﬁc agonists in wild-type mice or human hepato- and PXR activators that do not activate CAR cytes, and also the effects of a constitutively active VP16- (dexamethasone, PCN). Thus it is apparent that each recep- SXR protein in transgenic animals.17,28 tor mediates a speciﬁc subset of the xenobiotic response. It It should be emphasized that the overlap between CAR is not possible to predict whether a particular compound and PXR is not complete. There are murine CAR activators would activate CAR, PXR or both, considering the great that are ineffective on PXR (TCPOBOP, PB, chlorpromazine) structural diversity of the compounds that function directly

www.nature.com/tpj Regulation of gene expression by CAR PWeiet al 124

as ligands for the two proteins and also the compounding (90 mg kgϪ1), dieldrin (50 mg kgϪ1), chlorpromazine effects of both species differences and the indirect activation (30 mg kgϪ1), and dexamethasone (50 mg kgϪ1) for indi- of CAR by PB and presumably other compounds. Such deter- cated times. All compounds were purchased from Sigma (St minations will require careful characterization incorporat- Louis, MO, USA) unless indicated. For androstenol treat- ing not only biochemical studies of ligand binding, but also ment, the dose for TCPOBOP was 0.3 mg kgϪ1 and andros- in vivo functional studies. tenol 150 mg kgϪ1 (Steraloids Inc, Newport, RI, USA). For 3- The results described here suggest the existence of an day PB or androstenol treatment, mice were injected intra- intriguing additional level of complexity and specificity in peritoneally three times with the compound, one injection the functions of both CAR and PXR that is based on differ- per day. ential effects of different inducers. While the potent CAR agonist TCPOBOP is an effective inducer of both CYP2B10 X-Gal (5-Bromo-4-chloro-3-indolyl-␤-D-galactoside) and CYP3A11, CAR activation by chlorpromazine only Staining induces the former. Thus, induction of CYP2B10 expression Liver tissues were fixed at 4°C for 1 h with freshly made 4% appears to be a consistent consequence of CAR activation, paraformaldehyde in PBS then embedded in Tissue Freezing while the CYP3A11 response is not. Similarly, the activation Medium (Triangle Biomedical Sciences, Durham, NC, of PXR consistently results in CYP3A11 induction, but USA). Whole tissues or frozen sections of liver were stained among the compounds examined here only dexamethasone for 1.5 h at 37°C with a staining solution containing induced CYP2B10 expression in the CAR knockout animals, 1mgmlϪ1 X-gal, 5 mM potassium ferrocyanide, 5 mM pot- and this effect is thought to be due to the glucocorticoid assium ferricyanide, 2 mM MgCl2, 0.01% deoxycholate, 24 receptor. It is possible that the breadth of the response to 0.02% NP-40, 100 mM phosphate buffer, pH 7.3. Embryos a particular treatment is a relatively simple reflection of the or adult small intestine and stomach tissues were fixed at potency of receptor activation. Alternatively, individual room temperature for 2 h with fixative containing 0.2% glu- inducers could have more complex differential effects on the taraldehyde, 1.5% formaldehyde, 5 mM EGTA, 2 mM conformation of the activated receptors. This would be con- MgCl2, 100 mM sodium phosphate, pH 7.3. Whole-mount sistent with detailed studies of the estrogen receptor demon- staining of embryo and adult small intestine and stomach strating that binding of even closely related agonists or were performed at 37°C overnight with the same staining 29 antagonists can induce distinct receptor conformations. solution described above. They were then either preserved These differential responses complicate interpretation of in 3.7% formaldehyde or dehydrated and processed through the already complex effects of various inducers of these two Histo-Clear (National Diagnostics, Atlanta, GA, USA) genes and other targets in vivo. Particularly as the number embedding in paraffin. The sections were counter-stained of identified targets for the two receptors increases, it will with Nuclear Fast Red (Trevigen, Gaithersburg, MD, USA). be interesting to determine whether different activators of a specific receptor induce distinct patterns of target genes. Overall, we conclude that CAR plays a much broader role Northern Blot Analysis ␮ in xenobiotic response than simply mediating response of Total RNA (20–30 g) from individual mouse livers was sub- CYP2B10 to PB-like inducers. The central role of CAR in jected to Northern blot analysis. Preparation of probes for 11 CYP2B and 3A gene induction has wide implications, not CYP2B10 was previously described. Probes for CYP3A11 only in the mechanism of gene regulation, but also in the and CYP2A4 were prepared by RT-PCR with mouse liver understanding of the toxicological effects of environmental total RNA using the Superscript One-step RT-PCR System Ј contaminants and drug–drug interactions. CYP3A enzymes (Life Technologies). PCR primers for CYP3A11 were 5 - Ј Ј catalyze the metabolism of up to 60% of various commonly ccg cct cta gat gag atc gat gag gct ct-3 and 5 -ccg ccg gat Ј Ј used drugs,30,31 and the induction of this expression by one cct gcc ttt ctg aag aac tc-3 . The primers for CYP2A4 were 5 - Ј Ј drug can have significant undesirable consequences on the ccg cct cta gaa ggg tca cca agg aca cc-3 and 5 -ccg ccg gat ccc Ј effects of another. The results described here suggest that the cct tct ctg gct acc ttt-3 . use of cell culture- or transgenic animal-based test systems employing the human CAR protein may facilitate the identi- Western Blot Analysis fication of known drugs or drugs in development that show Mouse liver total protein was prepared by sonicating the such undesirable drug–drug interactions. This should allow tissues in Laemmli sample buffer without ␤-mecapto- the development of safer drugs that retain desirable thera- ethanol. Ten micrograms of total protein were electrophor- peutic activity but not undesirable effects on drug metab- esed on 10% polyacrylamide gels, and immunoblotted with olism. monoclonal anti-rat CYP2B1 (BE28.2) or anti-rat CYP3A1 (Ig8) antibodies (provided by Dr Erin G Schuetz, St Jude MATERIALS AND METHODS Children’s Research Hospital, Memphis, TN, USA). Both pri- Animal Treatment mary antibodies were followed by appropriate secondary At least three mice between 8–10 weeks old were used for antibodies coupled with horseradish peroxidase and each treatment. Mice were treated by intraperitoneal injec- developed with the enhanced chemiluminescence detection tion with corn oil, PB (100 mg kgϪ1) or TCBOPOP reagent ECL (Amersham Pharmacia Biotech, Piscataway, (3 mg kgϪ1, a gift from Dr Stephen Safe), clotrimazole NJ, USA).

The Pharmacogenomics Journal Regulation of gene expression by CAR PWeiet al 125

In Situ Hybridization 2 Waxman DJ. P450 gene induction by structurally diverse xenochem- Paraffin sections of adult small intestine were used in this icals: central role of nuclear receptors CAR, PXR, and PPAR. Arch Biochem Biophys 1999; 369:11–23. study. A fragment of CYP2B10 was amplified with RT-PCR 3 Honkakoski P, Negishi M. Regulation of cytochrome P450 (CYP) genes as described11 and subcloned into Xba I and BamH I sites of by nuclear receptors. Biochem J 2000; 347(Pt 2): 321–337. pBluescript (Stratagene, La Jolla, CA, USA). This vector was 4 Smith DA. Induction and drug development. Eur J Pharm Sci 2000; 11: linearized with Xba I. T7 polymerase was used to synthesize 185–189. 35 5 Thummel KE, Wilkinson GR. In vitro and in vivo drug interactions [ S]-labeled antisense CYP2B10 probes. Labeled probes were involving human CYP3A. Annu Rev Pharmacol Toxicol 1998; 38: 389– hybridized to tissue sections and mRNA expression was 430. detected by film emulsion with autoradiography. Sections 6 Honkakoski P, Moore R, Gynther J, Negishi M. Characterization of were counter-stained with Hematoxylin and then examined phenobarbital-inducible mouse Cyp2b10 gene transcription in primary hepatocytes. J Biol Chem 1996; 271: 9746–9753. using light- and dark-field microscopy. 7 Honkakoski P, Negishi M. Characterization of a phenobarbital-responsive enhancer module in mouse P450 Cyp2b10 gene. J Biol Chem 1997; Cell Culture and Transfections 272: 14943–14949. HepG2 cells were maintained in Dulbecco’s modified Eagle 8 Honkakoski P, Zelko I, Sueyoshi T, Negishi M. The nuclear orphan medium, supplemented with 10% fetal bovine serum receptor CAR-retinoid X receptor heterodimer activates the phenobarbital-responsive enhancer module of the CYP2B gene. Mol Cell Biol (Hyclone, Logan, UT, USA). For the transfections HepG2 1998; 18: 5652–5658. cells were plated into 24-well dishes with Dulbecco’s modi- 9 Sueyoshi T, Kawamoto T, Zelko I, Honkakoski P, Negishi M. The fied Eagle medium supplemented with 10% charcoal- repressed nuclear receptor CAR responds to phenobarbital in activat- stripped serum and transfected using the calcium phosphate ing the human CYP2B6 gene. J Biol Chem 1999; 274: 6043–6046. 10 10 Tzameli I, Pissios P, Schuetz EG, Moore DD. The xenobiotic compound precipitation method as described previously. The cells 1,4-Bis[2-(3,5-Dichloropyridyloxy)] Benzene is an agonist ligand for were washed with phosphate-buffered saline 12 h after the nuclear receptor CAR. Mol Cell Biol 2000; 20: 2951–2958. transfection and ligands were added. Transfections included 11 Wei P, Zhang J, Egan-Hafley M, Liang S, Moore DD. The nuclear recep- 200 ng of luciferase reporter plasmid, 50 ng ␤-galactosidase tor CAR mediates specific xenobiotic induction of drug metabolism. ␤ Nature 2000; 407: 920–923. internal control plasmid (pCMX- -galactosidase), 50–100 ng 12 Kliewer SA, Moore JT, Wade L, Staudinger JL, Watson MA, Jones SA et of pCMX expression vectors for GAL4 and/or VP16 fusion al. An orphan nuclear receptor activated by pregnanes defines a novel proteins and pGEM4 plasmid DNA to bring the total DNA to steroid signaling pathway. Cell 1998; 92:73–82. 1 ␮g per well. Cells were harvested and assayed for luciferase 13 Blumberg B, Sabbagh W Jr, Juguilon H, Bolado J Jr, van Meter CM, activity 24 h after the addition of the ligands, and reporter Ong ES et al. SXR, a novel steroid and xenobiotic-sensing nuclear ␤ receptor. Genes Dev 1998; 12: 3195–3205. expression was normalized to the -galactosidase activity 14 Lehmann JM, McKee DD, Watson MA, Willson TM, Moore JT, Kliewer (Tropix, Bedford, MA, USA), according to the manufacturer’s SA. The human orphan nuclear receptor PXR is activated by com- directions. Each data point represents the average of a quad- pounds that regulate CYP3A4 gene expression and cause drug interac- ruplicate experiment, and similar results were replicated in tions. J Clin Invest 1998; 102: 1016–1023. 15 Xie W, Barwick JL, Downes M, Blumberg B, Simon CM, Nelson MC et at least three independent experiments. al. Humanized xenobiotic response in mice expressing nuclear receptor SXR. Nature 2000; 406: 435–439. Plasmids 16 Moore LB, Parks DJ, Jones SA, Bledsoe RK, Consler TG, Stimmel JB et The pCMX-GAL4, pCMX-VP16, pCMX-GAL4-mCAR-LBD al. Orphan nuclear receptors constitutive androstane receptor and pre- fusion,10 pG5E1b-Luciferase reporter and pCMX-GAL4-SRC- gnane X receptor share xenobiotic and steroid ligands. J Biol Chem 32 2000; 275: 15122–15127. 3 RID have been described previously. pCMX-GAL4-hPXR- 17 Xie W, Barwick JL, Simon CM, Pierce AM, Safe S, Blumberg B et al. LBD was generated by PCR amplification of cDNA encoding Reciprocal activation of xenobiotic response genes by nuclear recep- amino acids 105–434 of hPXR using oligonucleotides tors SXR/PXR and CAR. Genes Dev 2000; 14: 3014–3023. 5Ј-GCGGGTACCAAGAAAGAGATGATCATGTCCGATGCCG 18 Asoh M, Tateishi T, Kumai T, Kobayashi S. Induction of hepatic CYP2B in foetal and neonatal rats after maternal administration of phenobar- CTG-3Ј and 5Ј-GGGTGTGGGGGATCCTCAGCTACCTG Ј bital. Pharmacol Toxicol 1999; 84:18–23. TGATGCCG-3 and insertion into pCMX-VP16 vector via 19 Kolars JC, Lown KS, Schmiedlin-Ren P, Ghosh M, Fang C, Wrighton Asp718/BamHI sites, with the sequence confirmed by in vitro SA et al. CYP3A gene expression in human gut epithelium. Pharmaco- translation and sequencing. genetics 1994; 4: 247–259. 20 Smith G, Henderson CJ, Parker MG, White R, Bars RG, Wolf CR. 1,4- Bis[2-(3,5-dichloropyridyloxy)] benzene, an extremely potent modu- ACKNOWLEDGEMENTS lator of mouse hepatic cytochrome P-450 gene expression. Biochem J We thank Drs Erin G Schuetz (St Jude Children’s Research Hospital) for 1993; 289: 807–813. providing antibodies to CYP2B1 and CYP3A1 and Steven Safe for TCPO- 21 Noshiro M, Negishi M. Pretranslational regulation of sex-dependent BOP. We also thank Dr Fred A Pereira for help with mouse histology. testosterone hydroxylases by growth hormone in mouse liver. J Biol This work was supported by a grant (R01 DK46546) from the National Chem 1986; 261: 15923–15927. Institutes of Health to DDM. 22 Forman BM, Tzameli I, Choi HS, Chen J, Simha D, Seol W et al. Andros- tane metabolites bind to and deactivate the nuclear receptor CAR- DUALITY OF INTEREST beta. Nature 1998; 395: 612–615. 23 Honkakoski P, Moore R, Washburn KA, Negishi M. Activation by None declared. diverse xenochemicals of the 51-base pair phenobarbital-responsive enhancer module in the CYP2B10 gene. Mol Pharmacol 1998; 53: REFERENCES 597–601. 1 Waxman DJ, Azaroff L. Phenobarbital induction of cytochrome P-450 24 Schuetz EG, Schmid W, Schutz G, Brimer C, Yasuda K, Kamataki T et gene expression. Biochem J 1992; 281 (Pt 3): 577–592. al. The glucocorticoid receptor is essential for induction of cytochrome

www.nature.com/tpj Regulation of gene expression by CAR PWeiet al 126

P-4502B by steroids but not for drug or steroid induction of CYP3A or 29 Paige LA, Christensen DJ, Gron H, Norris JD, Gottlin EB, Padilla KM et P-450 reductase in mouse liver. Drug Metab Dispos 2000; 28: 268–278. al. Estrogen receptor (ER) modulators each induce distinct confor- 25 Kawamoto T, Sueyoshi T, Zelko I, Moore R, Washburn K, Negishi M. mational changes in ER alpha and ER beta. Proc Natl Acad Sci U S A Phenobarbital-responsive nuclear translocation of the receptor CAR in 1999; 96: 3999–4004. induction of the CYP2B gene. Mol Cell Biol 1999; 19: 6318–6322. 30 Cholerton S, Daly AK, Idle JR. The role of individual human cyto- 26 Zelko I, Sueyoshi T, Kawamoto T, Moore R, Negishi M. The peptide chromes P450 in drug metabolism and clinical response. Trends Pharm- near the C terminus regulates receptor CAR nuclear translocation acol Sci 1992; 13: 434–439. induced by xenochemicals in mouse liver. Mol Cell Biol 2001; 21: 31 Guengerich FP. Cytochrome P-450 3A4: regulation and role in drug 2838–2846. metabolism. Annu Rev Pharmacol Toxicol 1999; 39:1–17. 27 Pascussi JM, Gerbal-Chaloin S, Fabre JM, Maurel P, Vilarem MJ. 32 Lee YK, Dell H, Dowhan DH, Hadzopoulou-Cladaras M, Moore DD. Dexamethasone enhances constitutive androstane receptor expression The orphan nuclear receptor SHP inhibits hepatocyte nuclear factor 4 in human hepatocytes: consequences on cytochrome P450 gene regu- and retinoid X receptor transactivation: two mechanisms for lation. Mol Pharmacol 2000; 58: 1441–1450. repression. Mol Cell Biol 2000; 20: 187–195. 28 Goodwin B, Moore LB, Stoltz CM, McKee DD, Kliewer SA. Regulation of the human cyp2b6 gene by the nuclear pregnane x receptor. Mol Pharmacol 2001; 60: 427–431.

The Pharmacogenomics Journal