Mammalian cytochromes P450 (P450s) play a major role in the metabolism and deposition of drugs and environmental chemicals (Rushmore and Kong, 2002). Understanding their functions in drug deposition is now a key factor in drug development and use. In the past, the only means of evaluating the P450 system in vivo in the deposition and toxicity of a drug was through the use of cytochrome P450 inhibitors (Fontana et al., 2005). This approach presents many difficulties for the interpretation of results, particularly as these inhibitors do not inhibit all P450 activity (Emoto et al., 2003), their effects are transient (Mugford et al., 1992), and they will inhibit other pathways, such as the function of drug transporters (Weiss et al., 2006). To circumvent these problems, we have developed a mouse where hepatic P450 activity is reduced by greater than 95% by the conditional deletion of cytochrome P450 oxidoreductase (POR). This was achieved by crossing mice carrying the Por gene flanked with loxP sites with mice expressing Cre recombinase under the control of the rat albumin promoter (Gu et al., 2003; Henderson et al., 2003a). Associated with the loss of hepatic POR, profound changes in the metabolism and deposition of a wide range of drugs and environmental chemicals have been documented (Arlt et al., 2005; Pass et al., 2005), which has validated the power of this model in establishing the role of the hepatic P450 system in pharmacodynamic and toxicological responses. Hepatic Reductase Null (HRN) mice developed normally and they were fertile; however, they exhibited a number of phenotypic changes associated with the loss of P450 function, including hepatic lipid accumulation, reduced bile acid production, increased constitutive P450 expression, and decreases in plasma cholesterol and triglyceride levels.

In the HRN model, deletion of the Por gene is liver-specific and occurs neonatally; thus, investigations of the role of the P450 system in other tissues/organs and at other stages of development are not possible. To extend the utility of this model through the deletion of POR, and P450 inactivation in other tissues, we have crossed mice containing floxed Por loci with mice where the expression of Cre recombinase is conditionally regulated by the rat CYP1A1 promoter (Campbell et al., 1996; Ireland et al., 2004). We have previously demonstrated that the use of this promoter provides a tightly regulated method for controlling the expression of transgenes in vivo by the administration of compounds that act through the Ah receptor (Campbell et al., 1996; Kantachuvesiri et al., 2001; Ireland et al., 2004).

In this study, we demonstrate that administration of Ah receptor ligands to these mice results in a time-dependent deletion of hepatic and gastrointestinal POR. Intriguingly, by changing the compound administered, it was possible to achieve either hepatic or hepatic plus gastrointestinal gene deletion. This model will provide a powerful tool for improving drug efficacy testing and to study the role of hepatic and gastrointestinal P450s in drug deposition and chemical toxicity.

Materials and Methods

Reagents. All chemicals were purchased from Sigma Chemical (Poole, Dorset, UK), except where indicated.

Production of Transgenic Mice.Por floxed mice (Por^lox/lox) were generated as described previously (Henderson et al., 2003a). A transgenic mouse line expressing Cre recombinase under the control of the rat cytochrome P450 CYP1A1 promoter (Cre^CYP1A1) (provided to us by Dr. Douglas Winton, Cambridge Institute for Medical Research, Cambridge, UK) (Ireland et al., 2004) was crossed onto the Por^lox/lox line to generate the mouse line Por^lox/lox/Cre^CYP1A1. This Por^lox/lox/Cre^CYP1A1 mouse line was subsequently backcrossed for six generations onto a C57BL/6 genetic background. Littermates with the Por^+/+ or Por^lox/lox genotype were taken as controls in the experiments described below. All mice were maintained under standard animal house conditions with a 12-h light/12-h dark cycle and free access to water and RM1 diet (SDS, Essex, UK). All studies were carried out on 12-week-old male mice in accordance with the Animal Scientific Procedures Act (1986) and after local ethical review.

Mouse Treatments. β-Naphthoflavone (BNF) (80 mg/kg body weight) was administered i.p. as a suspension in corn oil for 4 days at 24-h intervals. Tetrachlorodibenzo-p-dioxin (TCDD) was administered as a single i.p. dose in corn oil (37.5 μg/kg body weight). 3-Methylcholanthrene (3MC) was administered by i.p. in corn oil at a dose of either 40 mg/kg body weight for 4 days at 24-h intervals or as a single injection at various doses.

Immunoblotting and Biochemical Analysis. Microsomal fractions were prepared by differential centrifugation (Meehan et al., 1988), and protein concentrations were determined by the method of Bradford using bovine serum albumin as a standard (Bradford, 1976). Western blot analysis was carried out as described previously (Forrester et al., 1992) using 5 μg of microsomal protein per lane and polyclonal antisera raised against human POR (Smith et al., 1994), or rat P450s (Forrester et al., 1992). Polyclonal antiserum was raised against murine Cyp7a1 using a peptide corresponding to amino acids 114 to 134. Antibodies were detected using donkey anti-rabbit horseradish peroxidase IgG as a secondary antibody (Dako UK Ltd., Ely, Cambridgeshire, UK). Immunoreactive bands were visualized by chemiluminescence (ECL or ECL+; GE Healthcare, Little Chalfont, Buckinghamshire, UK). POR activity was determined as described previously (Smith et al., 1994). Hepatic P450 catalytic activity was determined using the fluorogenic substrate 7-benzyloxy-4-(trifluoromethyl)-coumarin (BFC) (BD Biosciences, Oxford, UK). Assays were performed at 37°C in white 96-well plates in a reaction volume of 150 μl containing 600 μM NADPH, 30 μg of microsomal protein, and a final substrate concentration of 40 μM. Excitation and emission wavelengths used were 409 and 530 nm, respectively.

Blood Chemistry. Animals were sacrificed by gradual elevation of CO₂ levels in accordance with the Animal Scientific Procedures Act (1986) (Schedule 1). Blood was collected by cardiac puncture into heparinized tubes and plasma prepared by centrifugation. Samples were stored at –70°C before analysis for alanine aminotransferase, nonfasting total cholesterol, and nonfasting triglycerides using commercially available assay kits (Thermo Trace, Alpha Laboratories, Hampshire, UK) on a Cobas Fara II centrifugal analyser (Roche, Hertfordshire, UK).

Histopathology and Immunohistochemistry. Tissue samples were fixed either in formalin/phosphate-buffered saline for 24 h before transfer to 80% ethanol for storage or snap-frozen in Cryo-M-Bed (Bright Instrument Co, Huntington, UK) and stored at –70°C. Formalin-fixed tissues were embedded in paraffin wax, sectioned (3 μm), and stained with hematoxylin and eosin or processed for immunostaining with the polyclonal anti-serum to human POR (Smith et al., 1994). Snap-frozen tissue was cryosectioned (10 μm) and processed for staining with Oil Red O to determine lipid content.

Results

Hepatic and Peripheral Deletion of POR Using the Ah Receptor Ligands. We have previously described a mouse model (HRN) where hepatic POR expression has been deleted, resulting in a profound reduction in hepatic P450 activity. These mice display a number of distinct phenotypic changes, including an enlarged fatty liver, reduced nonfasting plasma triglyceride and cholesterol levels, and induction of various P450 isozymes (Henderson et al., 2003a). In an attempt to extend the deletion of POR, and P450 inactivation, to other tissues, and to extend the utility of this model for investigating Por gene function at other developmental stages, we have generated an inducible POR knockout mouse model. Por^lox/lox mice were crossed with a nontargeted transgenic mouse line in which an 8.5-kilobase fragment of the rat CYP1A1 promoter was used to drive the expression of Cre recombinase (Cre^CYP1A1) (Ireland et al., 2004). Animals from the resulting Por^lox/lox/Cre^CYP1A1 line, in the absence of a chemical inducer, were phenotypically normal and indistinguishable from wild-type animals. There was also no excision of the Por gene in any tissue, demonstrating the absence of any background Cre recombinase activity.

The ability of Ah receptor ligands to regulate Cre recombinase expression, and as a consequence delete the Por gene, was investigated using a range of chemical agents. Treatment of Por^lox/lox/Cre^CYP1A1 mice with BNF led to almost complete loss of hepatic POR protein 8 days after administration and, additionally, in the upper section (duodenum) of the small intestine, the site of highest P450 expression (Zhang et al., 2003; Thörn et al., 2005), by day 6 (Fig. 1A). The lower POR-immunoreactive band that was detected in some small intestinal samples corresponds to a 68-kDa POR-proteolyzed cleavage peptide. In the small intestine, there was a very slight recovery in POR expression from day 10, possibly as a consequence of the higher cell regeneration rates found in epithelial cells lining the gut (Fig. 1A) (Ireland et al., 2004). POR protein levels were unchanged in the kidneys and lungs of treated Por^lox/lox/Cre^CYP1A1 mice (Fig. 1A), despite induction of Cyp1a1 in both tissues, as determined by Western blot analysis (data not shown). This demonstrates a clear separation between Cyp1a1 expression and that of Cre recombinase driven by the heterologous rat CYP1A1 promoter. The changes in hepatic and gastrointestinal (g.i.) POR levels were confirmed by immunohistochemical staining of formalin-fixed tissue sections with antisera to POR (Fig. 1B) and determination of POR activity by assaying the reduction of cytochrome c (Fig. 1C), which revealed that the recovery of small intestinal POR from day 8 to 10 was approximately 10%. The abolition of hepatic P450 catalytic activity as a result of POR deletion was confirmed using the fluorogenic CYP3A substrate BFC (Fig. 1D). However, abolition of gastrointestinal P450 catalytic activity, assessed using various P450 substrates, could not be confirmed due to the high P420 content of microsomal preparations, as determined by spectral analysis.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Deletion of murine cytochrome P450 oxidoreductase in Por^lox/lox/Cre^CYP1A1 mice following treatment with β-naphthoflavone. Por^lox/lox/Cre^CYP1A1 and Por^lox/lox (control) mice were treated with a multiple dose (80 mg/kg for 4 days) of BNF (see Materials and Methods), and tissues were harvested on the indicated day. POR protein levels were determined in various tissues from BNF-treated mice by Western blot analysis (A) and immunohistochemistry (B) with polyclonal antiserum to human POR (data shown is of tissues harvested on day 10). For Western blot analysis, samples were loaded at 5 μg of microsomal protein per lane, and all Por^lox/lox samples represent a pool of n = 2 (see Materials and Methods). C, POR activity was determined by cytochrome c reduction with all values representing the mean of n = 2; ▪, liver; , small intestine. D, hepatic P450 activity was determined using BFC as a substrate with all values representing the mean of n = 2. In all cases, n corresponds to the number of mice analyzed.

In view of the unexpected lack of POR deletion in the lung and kidney, where Cyp1a1 is known to be inducible, we carried out experiments with TCDD, one of the most potent inducers of the Ah receptor. Cyp1a1 levels were induced by a single dose of TCDD in all tissues examined as determined by Western blot analysis (Fig. 2A); however, POR protein levels were unchanged in the kidneys and lungs of treated Por^lox/lox/Cre^CYP1A1 mice (Fig. 2A). In contrast, a very marked loss of hepatic POR (53%) was observed 7 days after TCDD treatment (Fig. 2, A and B). In addition, a marked reduction, approximately 65%, in POR expression in the upper small intestine was also observed by this time; however, this was determined not be significant as in the case of BNF treatment, which produced approximately 84% reduction (Fig. 2, A and B). Similar to BNF-treated Por^lox/lox/Cre^CYP1A1 mice, a slight recovery of small intestinal POR activity, approximately 5%, was also observed in TCDD-treated Por^lox/lox/Cre^CYP1A1 mice. The abolition of hepatic P450 catalytic activity was confirmed as described for BNF treated mice (Fig. 2C), and similar to BNF-treated mice, the high P420 content of intestinal microsomal preparations, as determined by spectral analysis, prevented an accurate determination of P450 catalytic activity in this tissue. Treatment of Por^lox/lox mice with BNF or TCDD did not affect POR levels in any of the tissues examined, with the exception of liver where TCDD treatment induced POR levels. These results, therefore, show that the Ah receptor ligands BNF or TCDD can be used to conditionally delete the Por gene from the liver and small intestine of Por^lox/lox/Cre^CYP1A1 mice, thereby providing a model for investigation of POR function and the relative roles of the P450 system in these tissues in drug disposition.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Deletion of murine cytochrome P450 oxidoreductase in Por^lox/lox/Cre^CYP1A1 mice following treatment with tetrachlorodibenzo-p-dioxin. Por^lox/lox/Cre^CYP1A1 and Por^lox/lox (control) mice were treated with a single dose (37.5 μg/kg) of TCDD (see Materials and Methods), and tissues were harvested on the indicated day. POR protein and activity levels were determined in various tissues from TCDD-treated mice by Western blot analysis with polyclonal antiserum to human POR and rat CYP1A1 (A) and reduction of cytochrome c (B), respectively. For Western blot analysis, samples were loaded at 5 μg of microsomal protein per lane, and all Por^lox/lox samples represent a pool of n = 2 (see Materials and Methods; S.I., small intestine). For POR activity, all values represent the mean of n = 2; ▪, liver; , small intestine. C, hepatic P450 activity was determined using BFC as a substrate with all values representing the mean of n = 2. In all cases, n corresponds to the number of mice analyzed.

Hepatic Deletion of POR by 3-Methylcholanthrene. Treatment of Por^lox/lox/Cre^CYP1A1 mice with a multiple dose of 3MC led to complete and specific deletion of the hepatic Por gene, by day 10 (Fig. 3A). This was confirmed by immunohistochemical staining of formalin-fixed tissue sections with antisera to POR (Fig. 3B) and by measurement of enzyme activity. Hepatic microsomal POR activity was reduced from 123 ± 4 to 11 ± 1 μmol of cytochrome c reduced/min/mg microsomal membranes, i.e., >90%, in a similar manner to our previous studies with HRN mice. Ten days after 3MC treatment, the livers had become enlarged, increasing in weight by approximately 35% (Fig. 4B), and they looked pale, mottled, and friable; nonfasting levels of plasma cholesterol and triglycerides were reduced from 2.39 ± 0.20 to 1.32 ± 0.12 mM and from 0.73 ± 0.16 to 0.31 ± 0.06 mM, respectively (Fig. 4, C and D). In control mice, nonfasting levels of plasma cholesterol were measured as 2.90 ± 0.22 and 2.55 ± 0.29 mM and triglycerides as 1.35 ± 0.24 and 3.06 ± 1.39 mM, for corn oil versus 3MC treatment, respectively (Figs. 3D and 4C). Statistical analysis of these data showed a significant difference was only found in Por^lox/lox/Cre^CYP1A1 mice. 3MC treatment did not reduce POR levels in either the small or large intestine. The change in POR level in the large intestine shown in Fig. 3 is due to the apparent induction of POR by 3MC. These data show that conditional deletion of POR in Por^lox/lox/Cre^CYP1A1 mice resulted in a liver-specific POR null phenotype that was indistinguishable from that observed in the HRN model.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Hepatic deletion of murine cytochrome P450 oxidoreductase in Por^lox/lox/Cre^CYP1A1 mice following treatment with 3-methylcholanthrene. Por^lox/lox/Cre^CYP1A1 and Por^+/+/Cre^CYP1A1 (control) mice were treated with a multiple dose (40 mg/kg) of 3MC (see Materials and Methods), and tissues were harvested on day 10. POR protein levels were determined in various tissues by Western blot analysis (A) and immunohistochemistry with polyclonal antiserum to human POR (B). For Western blot analysis, samples are of individual mice, loaded at 5 μg of microsomal protein per lane.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Phenotypic changes in hepatic conditional P450 oxidoreductase null mice. Por^lox/lox/Cre^CYP1A1 and Por^+/+/Cre^CYP1A1 (control) mice were treated as outlined in Fig. 2. A, POR activity in microsomal fractions was measured using cytochrome c as a substrate as described under Materials and Methods (–, Por^lox/lox/Cre^CYP1A1 + corn oil; +, Por^lox/lox/Cre^CYP1A1 + 3MC). B, liver to body weight ratios. C, nonfasting plasma cholesterol levels. D, nonfasting plasma triglyceride levels. All values represent the mean ± S.E.M. of n = 4, where n corresponds to the number of mice analyzed. *, p ≤ 0.05; **, p ≤ 0.01; and ***, p ≤ 0.001. , corn oil; ▪, 3MC.

Dose-Dependent Deletion of Hepatic POR by 3MC. We have previously demonstrated that the expression of genes under the control of the CYP1A1 promoter can also be modulated by varying the dose of an Ah receptor ligand such as 3MC (Campbell et al., 1996). We have used this property to investigate the dose dependence of POR deletion by 3MC and to further understand whether the phenotypic changes in HRN mice, particularly the induction of cytochromes P450, are biochemically linked or involve multiple mechanisms.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Dose-dependent deletion of hepatic P450 oxidoreductase in Por^lox/lox/Cre^CYP1A1 mice using a single dose of 3-methylcholanthrene. Por^lox/lox/Cre^CYP1A1 and Por^lox/lox (control) mice were treated with a single dose of 3MC ranging between 5 and 40 mg/kg (see Materials and Methods). A, tissues were harvested on day 14, and POR and CYP protein expression levels determined on microsomal fractions by Western blot analysis (Forrester et al., 1992; Smith et al., 1994). Samples represent the pool of n = 3, where n corresponds to the number of mice analyzed, and samples were loaded at 5 μg of hepatic microsomal protein per lane. Standards used were human POR, human CYP2D6, rat CYP2B1, rat CYP1A1, rat CYP4A10, rat CYP2C6, His-tagged mouse Cyp3a11, and His-tagged mouse Cyp7a1. B, hepatic POR and lipid contents were visualized by either immunostaining with antiserum to POR of formalin-fixed sections or Oil red O staining of snap-frozen sections obtained at the indicated doses.

In these experiments, single doses of 3MC between 5 and 40 mg/kg were used, and the phenotypic changes were analyzed 14 days postadministration. Treatment of Por^lox/lox/Cre^CYP1A1 mice with a dose of either 20 or 40 mg/kg 3MC led to >90% loss of hepatic POR protein (Fig. 5A) and microsomal POR activity (Fig. 6A). A dose of either 5 or 10 mg/kg resulted in a 12 and 45% loss in POR, respectively (Figs. 5A and 6A); however, deletion of POR at these doses was not uniform within hepatocytes as determined by immunohistochemical staining of formalin-fixed tissue sections with antisera to POR (Fig. 5B). A similar staining pattern has been observed in the HRN model at early developmental stages and attributed to low levels of Cre expression (Henderson et al., 2003b; Gu et al., 2007). Consistent with previous reports, the loss of POR was accompanied by a marked increase in the expression of cytochromes P450 (Fig. 5A); however, in spite of this no hepatic P450 catalytic activity was detected (Fig. 6B). All cytochromes P450 examined showed a dose-dependent increase in expression with the exception of Cyp1a1 and members of the Cyp2d family (Fig. 5A). At lower doses of 3MC the induction of Cyp2b and Cyp3a proteins was distinguishable, with marked induction of Cyp3a11 occurring at all doses of 3MC used whereas for Cyp2b10 induction was only observed at doses above 10 mg/kg. Interestingly, the lower Cyp2b band, probably including Cyp2b9 (Li-Masters and Morgan, 2001), was induced at all doses of 3MC, similar to Cyp3a11, although at a lower -fold change. The data indicated that the hepatic level of the Cyp3a11 was more sensitive to changes in POR than the Cyp2b10, suggesting that different molecular mechanisms control the overexpression of these enzymes in the HRN mouse. In addition, significant changes in hepatic lipid accumulation (Fig. 5B), nonfasting plasma cholesterol (Fig. 6C) and triglyceride (Fig. 6D) levels, and liver size (Fig. 6E) were only observed in mice administered either 20 or 40 mg/kg 3MC, suggesting that these phenotypic changes are dependent on marked changes in hepatic POR activity. Therefore, liver enlargement, Cyp2b10 induction, liver lipid accumulation, and plasma lipid concentrations seemed linked (Henderson et al., 2003a; Weng et al., 2005), although there is no direct evidence suggesting Cyp2b10 is involved in the accumulation process, whereas Cyp3a11 induction seemed to involve at least in part a different mechanism.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Dose-dependent phenotypic changes in hepatic conditional P450 oxidoreductase null mice. Por^lox/lox/Cre^CYP1A1 and Por^lox/lox (control) mice were treated as outlined in Fig. 5. A, POR activity in hepatic microsomal fractions was measured using cytochrome c as a substrate. B, P450 activity was determined using BFC as a substrate. C, nonfasting plasma cholesterol levels. D, nonfasting plasma triglyceride levels. E, liver to body weight ratios. All values represent the mean ± S.E.M. of n = 3, where n corresponds to the number of mice analyzed. *, p ≤ 0.05; **, p ≤ 0.01; and ***, p ≤ 0.001. , +CYP1A1Cre; ▪, –CYP1A1Cre.

Discussion

The development of promoter systems for the conditional regulation of gene expression in both cell culture and animal models has proved extremely valuable in controlling the spatiotemporal expression of a protein. Inducible systems using the promoters of genes such as heat shock proteins (Schweinfest et al., 1988) and those regulated by steroids (Ko et al., 1989) have been developed, as have those using regulatory sequences of tetracycline resistance genes (Gossen and Bujard, 1992) and the estrogen receptor (Albanese et al., 2002; Hayashi and McMahon, 2002). We report here the use of the rat CYP1A1 promoter as a novel means of controlling the expression of Cre recombinase in the liver and small intestine and the generation a model for the conditional deletion of the Por gene in these tissues.

The activity of the CYP1A1 promoter is regulated by a wide range of compounds, such as polycyclic aromatic hydrocarbons, dioxins, flavones, and polychlorinated biphenyls. We have demonstrated that the Ah receptor ligands BNF and TCDD result in the deletion of POR from both the liver and small intestine. In contrast, 3MC administration resulted in the specific hepatic deletion of POR. Interestingly, the complete hepatic deletion of POR by 3MC could be achieved with a single dose of 20 mg/kg, with lower doses producing a dose-dependent decrease in the level of POR. The reason for the tissue selectivity observed is unclear, because these compounds are known to induce Cyp1a1 expression in a wide range of tissues in the mouse (Dey et al., 1999) and endogenous Cyp1a1 induction was observed in the liver, lung, and kidney, but it may be a consequence of random integration of the CYP1A1-Cre construct into a part of the genome that results in at least partial transcriptional silencing and failure of the construct to accurately reflect the endogenous pattern of CYP1A1 expression. Alternatively, the turnover rates of cells in the target tissues and/or the pharmacokinetics of the inducer may contribute to the unexpected expression pattern of Cre recombinase. The conditional deletion of POR in the liver using Por^lox/lox/Cre^CYP1A1 mice generated identical phenotypic changes to those observed when Cre recombinase was expressed using the rat albumin promoter (Gu et al., 2003; Henderson et al., 2003a), which included the increased expression of P450 isozymes. For certain of these isozymes, the induction is at least in part a consequence of increased transcription (Wang et al., 2005; Weng et al., 2005).

Spatiotemporal analysis of POR deletion is not possible in HRN mice as deletion of the gene occurs in the neonatal period when the albumin gene is switched on after birth. The inducible conditional model described in this study will allow us to analyze the temporal sequence of events that occur in the liver as a consequence of POR inactivation and that lead to the phenotypic changes observed (R. D. Finn et al., manuscript in preparation). Several interesting observations can be made from the dosage study shown here: 1) the extent of lipid accumulation in the liver is directly linked to level of POR deletion obtained and occurs only when this deletion is significant, and 2) the induction of Cyp2b and Cyp3a isozymes occurs via distinguishable pathways. Evidence to support the former observation comes from the generation of a hypomorphic POR deletion model (Weng et al., 2005; Wu et al., 2005), in which it was shown that mice with low levels of POR deletion did not accumulate lipid in the liver to the same extent as HRN mice. As for the latter observation, it has been known for many years that the regulation of Cyp2b10 and Cyp3a11 proteins overlap but can be distinguished, as highlighted by the findings that inducing agents such as dexamethasone, for example, can regulate CYP3A protein in rat liver without significantly affecting CYP2B levels (Meehan et al., 1988) and that the deletion of the transcription factors constitutive androstane receptor and pregnane X receptor, which control Cyp2b10 and Cyp3a11 regulation, respectively, only prevent the induction of one or the other of these proteins by certain agents, respectively (Wei et al., 2000; Xie et al., 2000; Staudinger et al., 2001).

Understanding the role of hepatic or g.i. metabolism in the absorption and disposition or toxicity of drugs is a key aspect of the drug development process and poor drug bioavailability in humans can be a consequence of metabolism in either tissue. It is also a key, but often neglected, aspect of the determination of clinical safety in general. Bioavailability can also be determined by drug transporters such as multidrug resistance protein 1 that are also expressed in the liver and g.i. tract. The availability of a model where both hepatic and g.i. P450s can be deleted will allow much more informative studies on the relative roles of transport and metabolism on drug deposition to be determined. The deletion of POR in the liver and g.i. tract, or only the liver depending on the inducing agent used, is fortuitous and allows a single animal model to be used to evaluate the role of hepatic or g.i metabolism in drug deposition and by inference the role of other extrahepatic enzymes. The effects of at least the hepatic deletion are long term, because we have recently found that POR deletion and the associated HRN phenotype is retained for at least 6 weeks (unpublished data). This is much longer than any inducing effect of 3MC on Ah receptor-mediated Cyp1a1 expression, which is lost 7 days after 3MC administration and a week before maximal POR deletion is observed on day 14 (Fig. 5A). Therefore, any potential compound effects on pathways of drug deposition can be avoided.

In conclusion, we have developed a transgenic mouse model in which the Por gene can be conditionally deleted, in a regulable manner, in the liver and gastrointestinal tract, providing a powerful tool for studying the role of hepatic and extrahepatic P450s in drug deposition and chemical toxicity. In addition, this model is allowing us to investigate the temporal sequence of events that result in P450 isozyme induction and to decipher the underlying molecular mechanism(s). As a means of inactivating P450 function, this model has several major strengths compared with the use of P450 inhibitors such as 1-aminobenzotriazole (Ortiz de Montellano and Mathews, 1981). Problems associated with using inhibitors include the following: 1) they are specific for individual isozymes, e.g., ketoconazole and quinidine inhibit CYP3A4 and CYP2D6, respectively, and those that are classed as nonspecific, such as 1-aminobenzotriazole or SKF-525A, show varied selectivity, have a wide K_i range (Emoto et al., 2003), and inhibit both microsomal and mitochondrial P450s (Colby et al., 1995; Soltis and Colby, 1998); 2) the duration and extent of inhibition depend on the pharmacokinetic properties of the inhibitor and the route of administration (Strelevitz et al., 2006); 3) they are potentially toxic and may be unsuitable for in vivo use, e.g., thiotepa (Turpeinen et al., 2006), or they may inhibit other pathways, e.g., chloramphenicol and its derivatives, such as cimetidine, affect mitochondrial protein synthesis, and antiretrovirals, such as tenofovir, affect P450s and drug transporters (Nekvindová et al., 2006; Weiss et al., 2006); 4) their use requires extensive preliminary studies to establish the correct dosing regimens and can generate uninterpretable results due to the complex interactions; and 5) sustained use at high doses is almost impossible, affecting the ability to perform efficacy screens in xenograft experiments. In contrast, genetic deletion of POR using the Cre/lox system 1) allows the specific inactivation of all microsomal P450s without effecting mitochondrial P450s; 2) allows tissue/organ-specific inactivation depending on the promoter driving Cre recombinase expression (the timing and duration of inactivation in a specific tissue/organ are dependent solely on the promoter system used, and, as demonstrated in the CYP1A1Cre model, the cell turnover rate of the target tissue); 3) does not require extensive preliminary dosing studies; and 4) does not require the sustained use of chemicals at high doses; thus, it is an ideal model system in which to perform efficacy screens.