Inducibility of CYP3A Gene Expression

In humans, CYP3A4 is believed to play the central role in drug metabolism since it is responsible for the metabolism of the largest number of currently used drugs (Watkins, 1994). Enhanced transcription of CYP3A occurs primarily in the liver and intestines in response to xenobiotic exposure. Agents known to activate CYP3A transcription include steroid hormones having either glucocorticoid or antiglucocorticoid activities, macrolide antibiotics, imidazole antifungal agents, and phenobarbital and phenobarbital-like agents such as polychlorinated biphenyls and organochlorine pesticides (reviewed in Gonzalez et al., 1993). This observation, coupled with the remarkable versatility of CYP3A catalytic activities, creates a potential for drug-drug interactions. Extrapolating data from studies utilizing rodents has proven to be an unreasonable approach for risk assessment because there is a remarkable species-specific induction profile characteristic of this family of P450s (Wrighton et al., 1985). For example, PCN, a strong inducer of CYP3A in the rat, does not induce CYP3A6 in the rabbit, while RIF, a strong inducer of rabbit CYP3A6, is not an inducer in the rat. An understanding of the mechanism underlying this phenomenon required cloning and characterization of theCYP3A promoters and identification of inducer response elements and the factors that interact with these promoter sequences.

Characterization of the CYP3A Promoter

Studies of the rank order of agonist potency and efficacy, of agonist-antagonist relationships, and of the time course of induction of CYP3A in cultured hepatocytes by various steroid hormones led to the conclusion that glucocorticoids (GC) induced CYP3A by a nonclassical GR mechanism probably involving a separate receptor or recognition system (Schuetz and Guzelian, 1984). Verifying the presence of a nonclassical glucocorticoid pathway by detailed molecular studies required the isolation of the cloned genomic DNA encoding theCYP3A2promoter (Burger et al., 1992). The first functional characterization of the rat CYP3A promoter involved transient transfections of various chimeric reporter gene constructs into primary cultures of rat hepatocytes (Burger et al., 1992). These investigators reported that the Dex/PCN response element resided within 164 bp of the start of transcription, and demonstrated that this region of theCYP3A gene maintained the same Dex responsiveness and synergy with PCN that was previously shown in the intact rat liver (Schuetz and Guzelian, 1984). These deletion studies confirmed earlier investigations concluding that the induction of CYP3A apparently utilized a nonclassical glucocorticoid receptor pathway (Schuetz and Guzelian, 1984), in that reporter gene activation occurred only at high doses of GC, and that RU486, a potent GR antagonist, induced CYP3A reporter gene activity, while it blocked GC induction of murine mammary tumor virus promoter-driven reporter activity. The lack of binding sites for the GR in the Dex responsive gene fragment further supported the hypothesis that the CYP3A gene is regulated by steroids through a mechanism that differs from the classical GR-mediated pathway (Schuetz and Guzelian, 1984).

Subsequent studies in primary nonproliferating adult rat hepatocytes (Quattrochi et al., 1995) and in a rat hepatoma cell line (Huss et al., 1996) led to the identification of the rat CYP3A23 Dex/PCN response element. Although both groups of investigators mapped similar response elements, Quattrochi et al. (1995) identified a single 20-bp element (referred to as Fp1, Table 1) as conferring Dex or PCN inducibility, whereas Huss et al. (1996) reported inducible activity only when this same element (referred to as site B or DexRE-2) was linked to an additional upstream sequence (referred to as site C or DexRE-1, Table 1). The discrepancy likely lies in the use of primary hepatocyte cultures versus culture-adapted, continuously replicating hepatoma cell lines that might differ in the availability of transcription factors. Tests of patterns of binding of nuclear proteins to the CYP3A23 Dex/PCN DNA response element gave identical results in the absence or presence of inducers, suggesting that induction occurred via a mechanism in which the unliganded protein factor was constitutively bound to DNA, much as has been reported for the retinoid acid X receptor (RXR) family of nuclear receptors (Evans, 1988). Indeed, the CYP3A23 response element contained two copies of the half-site, AGG(T)TCA, which defines the consensus binding site for this family of nuclear receptors. Furthermore, the GR did not interact with the DNA at the CYP3A23 response element. Patterns of nuclear protein binding to the rat CYP3A23promoter were similar to those reported by Miyata et al. (1995), who tested the rat CYP3A2 gene. These authors identified three footprints of liver-specific binding proteins, one of which showed DNA sequence homology to the liver-enriched transcription factor HNF-4. Further characterization of the CYP3A23 response element and its binding proteins revealed that “site A” was bound and activated by the nuclear receptor HNF-4, and that the chicken ovalbumin upstream promoter transcription factor (COUP-TF) could interact in vitro with “sites B and C”, leading to the conclusion that the GC inducibility of CYP3A genes involves multiple binding sites for members of the nuclear receptor superfamily (Huss and Kasper, 1998; Quattrochi et al., 1998).

Species-Specific Induction of CYP3A Gene Expression

The early observation of interspecies differences in the induction of CYP3A led to studies that addressed whether these differences were due to gene structure or to an endogenous cellular factor. For example, the response element located in the rat CYP3A23 promoter [Fp1 (Quattrochi et al., 1995) or DexRE-2 (Huss et al., 1996)] is composed of two consensus half-sites organized as direct repeats separated by three nucleotides. Subsequent cloning and sequencing of the promoters of the rabbit CYP3A6 and humanCYP3A4, CYP3A5, and CYP3A7 genes identified distinct sequences related to the rat CYP3A23response element (Barwick et al., 1996; Pascussi et al., 1999).

These sequence comparisons could be interpreted in one of two ways: 1) the sequences of the response elements are closely related so that differences in species inducibility are conferred by the same transcriptional activator interacting at these DNA sites, or 2) the sequences are sufficiently different so that a transcriptional activator unique to each species interacts at these sites. To distinguish between these alternatives, Barwick et al. (1996) performed a suitably informative trans-species comparison study in which rat, rabbit, and human CYP3A response elements were transfected into primary cultures of rat or rabbit hepatocytes treated with one of three inducers, Dex, PCN, or RIF. They found that activation of theCYP3A response element-reporter gene acquired the induction characteristics of the cell type; for example, RIF was able to induce the rat CYP3A23 response element when expressed in the rabbit hepatocyte cultures, and PCN induced the rabbit and humanCYP3A response elements when expressed in rat cells. These studies clearly established the transcription factor, rather than theCYP3A promoter sequence, as the agent conferring species-specific induction on CYP3A gene transcription.

PXR and Its Role in Species-Specific Induction of CYP3A

PXR and the CYP3A Response Element.

Identification of the Dex/PCN response element of the CYP3Agenes provided the initial firm evidence for involvement of a nuclear receptor in CYP3A induction. Cloning of PXR (NR1I2), a PCN-activated nuclear receptor in mouse liver, by Kliewer's group at Glaxo Wellcome (Kliewer et al., 1998) confirmed this conclusion. Kliewer's group and others demonstrated that PXR, named the pregnane X receptor because of its strong activation by pregnane compounds, seemed to mediate CYP3A induction not only in mice, but through its homologous counterparts in rat, rabbit, and humans, as well (Bertilsson et al., 1998; Blumberg et al., 1998; Lehmann et al., 1998; Zhang et al., 1999; Savas et al., 2000). The human PXR receptor is also referred to as steroid and xenobiotic-sensing receptor (SXR) (Blumberg et al., 1998) or as the pregnane-activated receptor (Bertilsson et al., 1998). Because there is no generally accepted nomenclature for the human receptor, for the purposes of this review we have chosen to use the term “human PXR”. Mutational analysis of the previously identifiedCYP3A response elements established the PXR binding site of the rodent CYP3A promoter/enhancer as a direct repeat, DR-3 (TGAACTn3TGAACT) (Kliewer et al., 1998), and the rabbit and humanCYP3A response elements as everted repeats (TGAACTn6AGGTCA) (ER6; Lehmann et al., 1998) or inverted repeats (IR6; Blumberg et al., 1998). The PXR forms a heterodimer with RXR, a requirement for binding and activation. The heterodimer formed between PXR and RXR can interact with either the DR-3 or ER6/IR6 elements in CYP3A genes.

The human CYP3A4 gene appears to have not only the ER6 response element, located in the proximal promoter (Barwick et al., 1996), but also a second, distal xenobiotic responsive element module (referred to as XREM), located ∼7.7 kb upstream of the transcriptional start site (Goodwin et al., 1999). XREM contains a PXR binding site composed of an imperfect DR-3 motif. In transient cotransfections of human PXR expression vector and chimeric reporter constructs in HepG2 cells, XREM was required for induction of CYP3A4 by numerous drugs (Goodwin et al., 1999). However, these investigators were unable to demonstrate CYP3A4 reporter gene inducibility with constructions containing the proximal promoter alone. In contrast, chimeric CYP3A4 reporter gene constructs containing only a single copy of the CYP3A4 ER6, whether linked to a viral promoter (Barwick et al., 1996) or to the CYP3A4 cellular promoter (∼ −1 kb) (Xie et al., 2000a), are fully sufficient to confer inducible reporter gene expression in cultures of adult rat hepatocytes. Thus, the relevance to the living animal of the results obtained from such molecular transfer experiments may critically depend on the choice of cellular hosts.

A comparison of the PXR sequences from different mammalian species indicates that the PXR proteins share less than 80% amino acid identity in their ligand binding domains (LBD), while the DNA binding domain is ∼95% similar (Jones et al., 2000). This striking difference in LBD sequence across species is unusual for members of the nuclear receptor family which share greater that 90% amino acid identity in both LBD and DNA binding domains (Mangelsdorf et al., 1995). The recently identified constitutive androstane receptor (CAR), also seems to be an exception to the rule (Moore et al., 2000). Based upon their tissue distribution, DNA binding activity, and ability to mediate CYP3A induction, it has been suggested that PXRs are orthologs rather than closely related members of a subfamily of nuclear receptors (Bertilsson et al., 1998; Jones et al., 2000). The differences in LBD among species are assumed to be responsible for the selectivity in ligand binding, and thus the striking differences in the induction profiles among species.

PXR Ligands.

Following the cloning of the PXR from various species, numerous studies identified compounds capable of activating the receptor in a species-specific manner (see Table 2 for examples). The majority of these studies used a cell-based assay system in which CV-1 cells, or other non-liver-derived cells, were transiently transfected with a PXR expression vector and a synthetic oligomer defining the DR-3 or ER6 linked to a viral promoter-driven reporter construct (Bertilsson et al., 1998; Blumberg et al., 1998; Lehmann et al., 1998; Schuetz et al., 1998; Jones et al., 2000; Moore et al., 2000; Savas et al., 2000). For the most part, drugs known to induce CYP3A in vivo were also able to activate the PXR in a species-specific manner. Jones et al. (2000) performed comprehensive comparisons of the PXR from human, rabbit, rat, and mouse using reporter constructs containing two copies of the CYP3A23 DR-3 response element transfected into CV-1 cells. In general, activation of PXR from different species has been in good agreement with the reported induction of CYP3A expression in primary cultures of hepatocytes from these same species. For example, human and rabbit PXR were both efficiently activated by RIF, the antimycotic clotrimazole (CTZ), trans-nonachlor, and phenobarbital (PB); however, rabbit PXR was more sensitive than human PXR to activation by Dex, PCN, cyproterone acetate, and spironolactone. These studies demonstrated that there are clear differences in the activation profiles between human and rabbit receptors. To gain further support for the observations that structurally diverse compounds are ligands of the PXR, Jones et al. (2000) developed a competition radioligand binding assay using [³H]SR12813, a bisphosphonate ester shown to be a potent activator of both the human and rabbit PXR. These experiments showed good agreement between those compounds able to displace [³H]SR12813 from the PXR and those able to activate PXR in transfection assays.

PXR from each species were also activated by naturally occurring steroids. Although each displayed a distinct activation profile, in each case the most efficacious activator was a pregnane (Jones et al., 2000). 5β-Pregnane-3,20-dione was the most efficacious activator of human (Bertilsson et al., 1998; Jones et al., 2000), rat, and mouse PXR, while 17-OH-progesterone was the most potent activator of rabbit PXR (Jones et al., 2000). However, these pregnane compounds activated only at superphysiological concentrations, leaving unanswered the question of the identity of the natural ligand for PXR. Based on the findings that PXR is activated by pregnenolone and its metabolites at low micromolar concentrations, it has been suggested that the natural ligand of PXR is probably a pregnane, and that the PXR defines a novel endocrine signaling pathway (Kliewer et al., 1998). Alternatively, the PXR has been described as a steroid and xenobiotic “sensing” receptor, responding to a broad range of compounds that humans are exposed to in the environment, as well as to circulating steroids (Blumberg et al., 1998). These postulates are not mutually exclusive.

Recently, studies in primary cultures of rat hepatocytes have shed light on differences between rat and human receptors. Xie et al. (2000a) used primary cultures of rat hepatocytes to critically test whether the PXR is indeed the postulated cellular factor that confers species specificity to the induction of CYP3A. The human PXR was transiently expressed in primary cultures of rat hepatocytes, and the effects of a panel of steroid and nonsteroid inducers on expression of cotransfected cellular promoters of rat CYP3A23 or humanCYP3A4 linked to the firefly luciferase gene were examined. In control rat cells without expressed human PXR, CYP3A23 was strongly induced, as expected, by PCN, nifedipine, or RU486, whereas RIF, CTZ, PB, corticosterone, cortisol, 17β-estradiol, progesterone, pregnenolone, and cortisone produced little or no induction. In contrast, cotransfection of human PXR resulted in significant induction of CYP3A23 reporter by drugs known to be active in humans, including RIF, CTZ, PB, 17β-estradiol, and pregnenolone. Furthermore, the induction of CYP3A23 by nifedipine and RU486 increased significantly with added human PXR, indicating that these drugs activate both rat and human receptors, while the induction by PCN remained unchanged in the presence of human PXR, indicating that PCN specifically activated the rat receptor. These trans-species transfection assays clearly demonstrated that transfection of human PXR is sufficient to convert the induction response characteristics of hepatocyte from rat to human.

In Vivo Role of the PXR.

Transgenic mice have been developed to establish the role of the PXR in vivo (Xie et al., 2000a). Targeted disruption of the mouse PXR gene eliminated the induction of CYP3A by PCN. PXR-null transgenic mice harboring the human PXR gene, when challenged with drugs known to induce human CYP3A such as RIF and CTZ, displayed inducedCYP3A mRNA in the liver. Transgenic mice expressing a constitutively active human PXR were shown to develop sustained CYP3A expression, resulting in enhanced protection against challenges of xenobiotic toxicants (Xie et al., 2000a). This “humanized” rodent model system has potential applications for drug and toxicity screening.

The discovery of PXR represents an important advance both in identifying the transcriptional mediator of CYP3A induction and identifying the agent responsible for the striking species differences in the induction profiles. The ability of a single transcription factor to determine species differences in response to xenobiotics is in itself without precedent. The differences in the activation profiles among the PXRs clearly highlight the need to develop an appropriate model for drug screening and discovery.

Mechanisms: PXR, CAR, GR, and Others

The mechanism by which numerous structurally diverse compounds induce CYP3A expression can be for the most part explained by the PXR. However, several questions remain unanswered. The nuclear receptors COUP-TF and HNF-4 have been implicated as factors interacting with sequences, including those at the Dex/PCN response element, of the ratCYP3A23 promoter. What role they play in the responsiveness of CYP3A to inducers remains to be determined. For example, do these proteins interact directly with the PXR to produce a highly active transcriptional complex? One of the most pressing questions is whether the GR plays a role in CYP3A induction. Dex, an efficacious inducer of CYP3A in rodents and humans, seems to be a relatively weak activator of the PXR, suggesting that the GR may be involved in the induction of CYP3A in these species. It now seems that another nuclear receptor is involved in the regulation of drug-metabolizing enzymes. The CAR has been shown to mediate induction of both CYP3A and CYP2B. And finally, how does the wealth of information gathered over the past several years explain the well established observation of the synergistic effects of GCs and PCN on CYP3A expression?

Factors that Modulate the PXR-Mediated Induction.

Recent work has revealed the existence of other proteins that appear to modulate the induction of CYP3A expression. Several transcription factors have been identified that interact with the CYP3A23promoter and seem to play a role in modulating the expression of CYP3A23 (Huss and Kasper, 1998; Huss et al., 1999; Ogino et al., 1999). Among these are members of the nuclear receptor family, COUP-TF and HNF-4. A postulated mechanism is that COUP-TF competes with the PXR for binding to the DR-3 response element of the CYP3A23 gene or that binding of COUP-TF to the upstream imperfect ER6 (DexRE-1, Table1) may repress HNF-4-mediated basal expression (Huss and Kasper, 1998). It could be reasoned that if such interactions occur in vivo, then modulation of the induction response would depend on relative amounts of each transcription factor, as well as their affinities for the DNA binding sites. In addition to COUP-TF, an unidentified protein referred to as the “B” protein also may interact with the upstream imperfect ER6, and it was suggested that this protein functions as an accessory factor in cooperation with PXR/RXR and HNF-4 (Huss and Kasper, 1998). The identities of all of the proteins interacting at this upstream element remain unknown, although the possibility exists for the PXR itself to also interact weakly with the upstream imperfect ER6. It may be concluded that binding of PXR to the DR-3 of the CYP3A23response element is necessary but insufficient for optimal CYP3A23 expression (Quattrochi et al., 1995; Huss et al., 1996). Other factors may interact with other sites in the CYP3A gene or with the PXR at this DR-3 site through protein-protein interactions with the PXR. A summary of the currently postulated relationships of nuclear receptors to the regulation of the rat CYP3A23 promoter is illustrated in Fig. 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Model of CYP3A23 regulation.

The model is described in the text. Low dose Dex is 10⁻⁷ M or less, while high dose is 10⁻⁵ M. The underlined sequence, CACGTG, is the putative binding site for upstream stimulatory factor-1.

Role of the Glucocorticoid Receptor in CYP3A Induction.

An important unanswered question is whether the GR plays a role in the regulated expression of CYP3A. Some lines of evidence suggest that induction of CYP3A by GCs can also occur through a pathway distinct from the PXR. For example, there is a functional glucocorticoid response element (GRE) in a region of the rat CYP3A23 gene ∼2 kb upstream of the PXR/RXR binding site (Fig. 1) (Pereira et al., 1998). In addition, the human CYP3A5 gene promoter contains two GRE half-sites, separated by 160 bp, that confer GC responsiveness to reporter genes (Schuetz et al., 1996). Conversely, studies in a GR-null mouse model suggest the opposite. Schuetz et al. (2000), using mice containing a targeted disruption of the GR to test its role on induction of CYP3A, found that CYP3A induction by GC, RU486, or RIF can occur in the absence of the GR. However, since the concentrations of Dex used in these studies, 50 mg/kg, may have activated the PXR, it is not possible to discern whether the induction of CYP3A in the wild-type or null mice is mediated through the PXR. Moreover, inasmuch as dose response data were apparently not obtained, this study does not exclude the possibility of involvement of the GR to augment the induction process. In fact, recent studies demonstrating the ability of Dex to increase PXR levels in hepatoma cell lines and primary hepatocytes (Huss and Kasper, 2000; Pascussi et al., 2000a) lend support to the concept that the GR plays a role in the induction process (discussed below). Finally, it should be noted that the mouse CYP3Apromoters have not been characterized as extensively as the rat or human promoters. Thus, the possibility that the GR plays a role in the induction of rat CYP3A23 and/or human CYP3A4 cannot be excluded. Indeed, cotransfected GR was shown to enhance the GC induction of a CYP3A4 reporter gene construct in transfected HepG2 cells (El-Sankary et al., 2000).

Cross Talk: PXR and CAR.

Early studies of CYP3A induction profiles established that the induction of CYP3A and CYP2B by PB occurred through distinct mechanisms (Kocarek et al., 1990). The cloning of both PXR and CAR has provided insight into this divergent regulatory mechanism. CAR binds DNA as a heterodimer with RXR and activates gene transcription in a constitutive manner; however, CAR-mediated transcriptional activation can be inhibited by androstane metabolites (reviewed in Honkakoski and Negishi, 2000). The CAR/RXR heterodimer interacts with the PB response element module (PBREM) located in CYP2B promoters of rat, mouse, and human genes, and mediates PB-inducible gene transcription. The PBREM contains a nuclear factor 1 binding site flanked by two nuclear receptor binding sites composed of imperfect direct repeats of half-sites spaced by four nucleotides (DR-4 motif). The mechanism by which PB derepresses CAR is presently unknown, but may involve direct ligand binding. Phenobarbital is also a fairly strong activator of the human PXR, but it has little or no effect on the rodent PXR (Jones et al., 2000; Xie et al., 2000a). Each PB-activated receptor can then interact with its own recognition site (i.e., PXR for the DR-3/ER6 inCYP3A promoters, and CAR for the PBREM in CYP2Bpromoters). However, CAR was shown to bind to and activate a reporter gene through the CYP3A4 ER6 response element, thus establishing the possibility for cross talk between CAR and PXR (reviewed in Honkakoski and Negishi, 2000).

Recent studies have demonstrated that PXR and CAR have the potential to cross-regulate CYP3A gene expression by two independent mechanisms: 1) sharing of some ligands and 2) binding of PXR and CAR to each other's DNA response elements. Moore et al. (2000) compared PXR versus CAR for activation by their respective ligands. They found dual activation of receptors by a subset of compounds. For example, androstanol deactivated (i.e., repressed) mouse CAR, and human CAR to a lesser extent, while CTZ deactivated only human CAR. Both of these compounds activated PXR. In cotransfections of human XREM-CYP3A4 reporter constructions [containing the proximal ER6 and the upstream XREM, described by Goodwin et al. (1999)], the negative effects of androstanol and CTZ on human CAR were found to be insufficient to overcome their positive effects on human PXR. These studies led to the conclusion that the human PXR is the dominant regulator of CYP3A4 expression. Studies in primary cultures of rat hepatocytes, using natural CYP3A and CYP2B promoters, demonstrated that CAR could regulate CYP3A reporter gene activity, but only in response to its own ligands (i.e., the human PXR activator RIF had little effect on CAR activity). Likewise, cotransfection of human PXR resulted in induction of CYP2B reporter gene activity in response to RIF and RU486, two activators of the human PXR (Xie et al., 2000b). Studies comparing CAR/RXR with PXR/RXR binding to each other's response element established that PXR and CAR exhibited similar binding affinity toward the response element of the human CYP3A4gene. In contrast, while PXR could interact with the nuclear receptor-1 of the PBRE of CYP2B, this element showed preference for CAR (Xie et al., 2000b). The most convincing evidence was obtained when this same group of investigators demonstrated that even in PXR-null mice, the PXR ligands CTZ and PB were efficacious inducers of CYP3A. Thus, cross-regulation of P450 enzymes by distinct nuclear receptors can be viewed as a versatile adaptive mechanism by which numerous structurally diverse compounds induce enzymes that in turn catalyze their metabolism and elimination. Future study of xenoregulation in CAR/PXR double knockout mice would provide additional insight for this proposed cross talk and might disclose the presence of other response modifiers.

Synergy: Who Are the Players?

A final question concerns the identity of the factors mediating synergy. The GR has been implicated in playing a role in the previously reported synergistic interaction between PCN and Dex for induction of CYP3A (Schuetz and Guzelian, 1984). An open possibility is that the GR could interact with the PXR promoter transcription complex through protein-protein interactions. Another suggestion is that the GRE found upstream of the PXR/RXR binding site of the rat CYP3A1 gene might play a role in the synergistic effects of Dex and PCN on CYP3A expression (Pereira et al., 1998). Recently, it was shown that Dex treatment of rat hepatoma cells and primary cultures of human hepatocytes increased levels of both PXR and RXR (Huss and Kasper, 2000; Pascussi et al., 2000a). It was concluded from these studies that the effect of Dex on PXR mRNA accumulation was most likely through direct activation of the GR. Therefore, another possibility for the reported synergy is that physiological levels of GC may activate the GR and increase the production of both PXR and RXR proteins, which in turn can further function as transcriptional regulators of CYP3A. This idea would require the levels of both nuclear receptors to be rate limiting in the hepatocyte, and evidence suggests that this may be the case in vivo, at least for the PXR (Zhang et al., 1999). Cloning of the PXR and RXR promoters and the identification of functional GREs within the promoters would lend support for this theory.

Future

The cloning of the PXR and its involvement in CYP3A induction represents a critical advance in the field of drug metabolism. Its demonstrated interaction with structurally diverse chemicals provides an explanation for much of the pharmacology of this P450 gene subfamily. As with many of the nuclear receptors, the PXR likely will be found to play a role in human physiology, outside that of drug metabolism. If so, this will surely lead to studies directed at certain disease states and risk assessment. For example, Pascussi et al. (2000b) demonstrated that interleukin-6 negatively regulates both PXR and CAR, suggesting that this provides direct evidence for the molecular mechanism underlying one of the earliest observations: that xenobiotic and drug metabolism is markedly impaired during inflammation and infections. Furthermore, endocrine functions may be altered by the interaction of endocrine-disrupting chemicals, such as phthalic acid and nonylphenol, recently shown to activate the mouse PXR (Masuyama et al., 2000). It seems likely that CYP3A plays one or more essential roles in body homeostasis, since to date this enzyme system has been found in every human liver examined. However, there is no comparable database on the possibility that CYP3A induction itself could be polymorphic. Considering the growing evidence that CYP3A induction may be more complex than initially believed, involving not only the PXR, which itself may be under inducible regulatory control, but also other nuclear receptors and response-modifying proteins, a fruitful area of future research will be exploring human variation in CYP3A induction.