Drug-Metabolizing Enzyme, Transporter and Nuclear Receptor Genetically Modified Mouse Models

Abstract

Determining the in vivo significance of a specific enzyme, transporter or xenobiotic receptor in drug metabolism and pharmacokinetics may be hampered by gene multiplicity and complexity, levels of expression and interaction between various components involved. The development of knockout (loss-of-function) and transgenic (gain-of-function) mouse models opens the door to the improved understanding of gene function in a whole body system. There is also growing interest in the development of humanized mice to overcome species difference in drug metabolism and disposition. This review, therefore, aims to summarize and discuss some successful examples of drug-metabolizing enzyme, transporter, and nuclear receptor genetically modified mouse models. These genetically modified mouse models have proven as invaluable models for understanding in vivo function of drug-metabolizing enzymes, transporters and xenobiotic receptors in drug metabolism and transport, as well as predicting potential drug-drug interaction and toxicity in humans. Nevertheless, concerns remain about interpretation of data obtained from such genetically modified mouse models in which the expression of related genes is altered significantly.

Introduction

Understanding the metabolism and disposition of drugs is critical for efficient drug discovery and development. Drugs may be structurally altered by Phase 1 oxidative enzymes [e.g., cytochrome P450 (P450 or CYP) and flavin-containing monooxygenase (FMO)] and Phase 2 conjugative enzymes [e.g., uridine 5'-diphospho-glucuronosyltransferase (UGT), sulfotransferase (SULT) and glutathione S-transferase (GST)] in liver, small intestine or other tissues. Drug accessibility, through intestinal absorption, tissue distribution, cell uptake and efflux, and renal and biliary excretion could be ultimately limited or facilitated by efflux [e.g., adenosine triphosphate (ATP)-binding cassette (ABC) family] or influx [e.g., solute carrier (SLC) family] drug transporters. Furthermore, levels of expression of enzyme and transporter proteins are subject to complex regulation by xenobiotic receptors [e.g., aryl hydrocarbon receptor (AhR), pregnane X receptor (PXR), and constitutive androstane receptors (CAR)], as well as epigenetic factors recognized recently (Gonzalez and Yu, 2006; Ingelman-Sundberg et al., 2007; Yu, 2007; Gomez and Ingelman-Sundberg, 2009; Yu, 2009). Differences in the functions or the degree of regulation of enzymes and transporters inevitably affect pharmacokinetic profiles, which may result in considerable inter-individual variability in drug response or toxicity.

Animals are common models used to predict drug metabolism and disposition in humans because they share many similarities in anatomy, physiology and genetics. Mouse models are of particular interest. The small size features relative low husbandry and maintenance costs. The fairly short life span, large litter size and fast breeding rate allow for growth of a large number of animals in a relatively short time, and therefore many studies can be carried out in a desired period. Mice are also of great value for genetic modification through transgenesis and targeted gene disruption; a target gene may be disrupted and a new gene inserted into the mouse genome (Manis, 2007). The resultant gene knockout and transgenic mouse models have been widely used for improved understanding of the role of specific enzyme, transporter or xenobiotic receptor in drug metabolism and disposition at the systemic level (Gonzalez, 2003; Gonzalez and Kimura, 2003; Henderson et al., 2003b; Henderson and Wolf, 2003; Dai and Wan, 2005; Xia et al., 2007; Muruganandan and Sinal, 2008; Lagas et al., 2009; Ou et al., 2010). In addition, species difference is a major confounding factor in metabolism and disposition of many xenobiotics such as chemical carcinogens and medications. The overall metabolic stability of the same drug could be very different across species. There are also human-specific metabolites that cannot be produced or produced at similar levels in wild-type mouse models, where pharmacological and toxicological properties of a particular metabolite need to be understood. Since species differences in drug metabolism and disposition are generally caused by the difference in composition, expression and/or function of enzymes, transporters or xenobiotic receptors, one approach to overcome this issue is to use humanized transgenic mouse models that can be developed by breeding a knockout mouse with a transgenic mouse created separately, or directly replacing a target mouse gene with the corresponding human gene or cDNA (“knock-in”). Studies with these humanized mouse models may not only provide mechanistic understanding of species difference but also improve prediction of risks of xenobiotics in humans (Henderson and Wolf, 2003; Liggett, 2004; Gonzalez and Yu, 2006). The following provides a brief review on enzyme, transporter and xenobiotic receptor gene knockout, transgenic and humanized mouse models that have been developed and used for studies on drug metabolism and disposition of xenobiotics or endobiotics.

Drug-Metabolizing Enzyme Genetically Modified Mouse Models

CYP1A1/CYP1A2

A number of Cyp1a1(−/−) and Cyp1a2(−/−) mouse lines have been generated and used to study the function of human CYP1A orthologues in xenobiotic metabolism (Pineau et al., 1995; Liang et al., 1996; Dalton et al., 2000), given a high degree of interspecies conservation in CYP1A1/1A2 expression, function and gene regulation. The Cyp1a2(−/−) mouse (Pineau et al., 1995) was among the first reported P450-knockout mouse models (Table 1), which was created by disruption of the exon 2 of Cyp1a2, and exhibited neonatal lethality associated with severe respiratory distress symptom, that ultimately was bred out yielding a mouse that was phenotypically indistinguishable from a wild-type mouse. Another Cyp1a2(−/−) mouse model was created by targeted disruption of the exons 2–5 of Cyp1a2 (Liang et al., 1996) (Table 1). The human and rodent CYP1A enzymes do exhibit some difference in metabolic capacity and substrate selectivity (Bogaards et al., 2000), and thus extrapolation of rodent data to humans is prone to error. Therefore, CYP1A1_1A2 humanized mice were generated by crossing the Cyp1a1(−/−), Cyp1a2(−/−) or Cyp1a1/1a2(−/−) mice with the CYP1A1_1A2(+/+) mice, which the latter were developed by incorporating a bacterial artificial chromosomes (BAC) containing human CYP1A1 and CYP1A2 into mouse genome (Jiang et al., 2005; Dragin et al., 2007) (Table 1). These knockout and humanized mouse models have been successfully used for studying the role of CYP1A1/1A2 in the toxicology of a number of environmental carcinogens including polycyclic aromatic hydrocarbons and heterocyclic aromatic amines/amides (Uno et al., 2004; Cheung et al., 2005a; Uno et al., 2006), and in the pharmacokinetics of many drugs including caffeine and theophylline (Buters et al., 1996; Derkenne et al., 2005).

Table 1.

Drug-metabolizing enzyme genetically modified mouse models.

Gene	Category	Example applications	Reference
Cyp1a1	Knockout	Gene expression and function	(Dalton et al., 2000)
Cyp1a2	Knockout	Potential physiological role of Cyp1a2 Metabolism of xenobiotics	(Pineau et al., 1995) (Liang et al., 1996)
Cyp1a1/1a2 CYP1A1/1A2	Knockout Humanized	Assess the risk of environmental toxicant	(Dragin et al., 2007)
Cyp1b1	Knockout	Evaluate the role of Cyp1b1 in DMBA carcinogenesis	(Buters et al., 1999)
Cyp2a5 CYP2A6	Knockout Transgenic	Role of Cyp2a5 in clearance of nicotine and cotinine Function of CYP2A6 in coumarin metabolism	(Zhou et al., 2009) (Zhang et al., 2005)
CYP2C18/ 2C19	Transgenic	Gene expression and function in drug metabolism	(Lofgren et al., 2008; Lofgren et al., 2009)
CYP2D6	Transgenic/ humanized	Debrisoquine metabolism and disposition Indolealkylamine metabolic pharmacogenetics	(Corchero et al., 2001) (Yu et al., 2003c)
Cyp2e1 CYP2E1	Knockout Humanized	Acetaminophen toxicity	(Lee et al., 1996) (Cheung et al., 2005b)
CYP3A4	Transgenic	Intestinal metabolism of midazolam, drug-drug metabolism, and physiological regulation	(Granvil et al., 2003) (van Herwaarden et al., 2005)
Cyp3a CYP3A4	Knockout Humanized	Docetaxel metabolism and pharmacokinetics	(van Herwaarden et al., 2007)
Cpr	Knockout	Drug metabolism and toxicity	(Shen et al., 2002; Otto et al., 2003) (Gu et al., 2003; Henderson et al., 2003a; Wu et al., 2003)
Gstp	Knockout	DMBA carcinogenesis Acetaminophen toxicity	(Henderson et al., 1998) (Henderson et al., 2000)
Sult1e1	Knockout	Physiological and pathological role of Sult1e1	(Qian et al., 2001; Tong et al., 2005)
Ugt1 UGT1	Knockout Transgenic Humanized	Tissue distribution, induction and hormonal regulation Physiological role in bilirubin homeostasis	(Nguyen et al., 2008) (Chen et al., 2005) (Cai et al., 2010; Fujiwara et al., 2010)

CYP2A6/Cyp2a5

Human CYP2A6 and the mouse orthologue Cyp2a5 share many common substrates (e.g., coumarin, nicotine and cotinine). To assess the role of CYP2A6 in drug metabolism in vivo, a CYP2A6(+/+) mouse model with liver-specific expression of human CYP2A6 was generated using the CYP2A6 cDNA, driven by a mouse liver-specific transthyretin promoter/enhancer (Zhang et al., 2005) (Table 1). More recently, a Cyp2a5(−/−) mouse model, in which the exon 9 of Cyp2a5 was disrupted with a tk-neo-bpA cassette, was developed on the C57BL/6J background (Zhou et al., 2009) (Table 1). CYP2A6 transgenic and Cyp2a5 knockout mice, showing no apparent developmental defects with normal viability and fertility, were used to establish that CYP2A6/Cyp2a5 is involved in the clearance of various drugs including coumarin, nicotine and cotinine (Zhang et al., 2005; Zhou et al., 2009).

CYP2C19

CYP2C19 metabolizes many drugs including antidepressants and proton pump inhibitors and has many genetic variations in different populations. A CYP2C18/2C19 transgenic mouse model was created using a CYP2C18/2C19-containing BAC clone (Lofgren et al., 2008; Lofgren et al., 2009) (Table 1). Gender difference in CYP2C18/C19 expression was observed at the mRNA level in transgenic mice, while only CYP2C19 protein was detectable in heterozygous CYP2C18/19(+/−) mice. In addition, CYP2C18/19(+/−) mouse liver microsomes showed higher catalytic capacity in metabolism of R-omeprazole and S-mephenytoin. These findings suggest this mouse line as an appropriate model for assessing hormonal regulation of CYP2C in vivo, as well as CYP2C19-mediated drug metabolism (Lofgren et al., 2008; Lofgren et al., 2009).

CYP2D6

There is remarkable species difference in xenobiotic metabolism between mice and humans mediated by the CYP2D family of P450s. Thus a humanized CYP2D6(+/+) (Tg-CYP2D6) mouse model was generated with a λ phage genomic clone carrying wild-type CYP2D6 gene (Corchero et al., 2001) (Table 1). Functional CYP2D6 protein was expressed in CYP2D6(+/+) mouse liver, intestine and kidney, resulting in an enhanced debrisoquine metabolism and clearance. Furthermore, urinary debrisoquine metabolic ratio (UMR) profiles (debrisoquine/4-hydroxydebrisoquine) in CYP2D6(+/+) versus wild-type mice closely resembled those in human CYP2D6 extensive metabolizers versus poor metabolizers, suggesting that CYP2D6(+/+) mice could be good animal models for studying CYP2D6 pharmacogenetics (Yu et al., 2004). Indeed, the CYP2D6(+/+) mice have been used to successfully define the role of CYP2D6 in pinoline metabolism (Jiang et al., 2009), as well as the effects of CYP2D6 status on harmaline pharmacokinetics and pharmacodynamics (Figure 1) (Wu et al., 2009). In contrast, when desipramine (DMI), another well-known CYP2D6 substrate, was administered to Tg-CYP2D6 and wild-type control mice, the UMRs of DMI were similar in the two genotyped mice (Shen and Yu, 2009). The results indicate that mouse enzymes have notable contribution to DMI 2-hydroxylation, which is mainly catalyzed by CYP2D6 in humans. In addition, this mouse model was used to identify potential endogenous substrates for CYP2D6 (Yu et al., 2003a; Yu et al., 2003b; Yu et al., 2003c; Yu et al., 2004), indicating its application to the assessment of CYP2D6-mediated reactions in vivo.

Figure 1.

Studies with wild-type and Tg-CYP2D6 mice revealed a significant (P < 0.05; two-way ANOVA) impact of CYP2D6 status on (a) harmaline pharmacokinetics (5 and 15 mg/kg, i.p.; N = 3 at each time point), and (b) drug effects on mouse marble-burying behavior (0–10 mg/kg, i.p.; N = 14 in each group) (Wu et al., 2009).

CYP2E1

To evaluate CYP2E1-mediated biotransformation in vivo, a Cyp2e1(−/−) mouse model was generated (Lee et al., 1996) (Table 1). The Cyp2e1(−/−) mice were less sensitive to acetaminophen (APAP) than wild-type mice, thus establishing an important role for Cyp2e1 in the activation of APAP to a hepatotoxic metabolite. Since species differences in CYP2E1 function and regulation were documented between mice and humans (Bogaards et al., 2000), a CYP2E1-humanized mouse model was developed by introducing a CYP2E1-containing BAC clone into the Cyp2e1(−/−) mice (Cheung et al., 2005b) (Table 1). Comparative studies revealed that CYP2E1-humanized and wild-type mice did differ in APAP toxicity. The CYP2E1-humanized mice should be useful in assessing the pharmacological and toxicological effects of CYP2E1 substrates.

CYP3A4

CYP3A4 is the most abundant P450 enzyme in human liver and gut, and contributes to the metabolism of many different drugs including benzodiazepines, HIV antivirals, immunomodulators and steroids. To develop a CYP3A4 transgenic mouse model, a BAC clone containing complete CYP3A4 gene was used (Granvil et al., 2003) (Table 1). In contrast to constitutive CYP3A4 expression in small intestines, CYP3A4 expression in transgenic mouse livers was subject to developmental and hormonal regulation (Figure 2) (Yu et al., 2005; Cheung et al., 2006). Pharmacokinetic studies in adult male mice with CYP3A4 expression only in the gut supported a role for CYP3A4 in first-pass metabolism of midazolam (Granvil et al., 2003). Furthermore, the impact of ketoconazole on midazolam pharmacokinetics in transgenic mice seems to be closer to that in humans (Table 2). This CYP3A4-transgenic mouse model may offer better assessment of intestinal CYP3A4-mediated drug metabolism and drug-drug interaction. Nevertheless, the expression of some mouse P450 enzymes (e.g., Cyp3a and Cyp2c) was altered in CYP3A4-transgenic mouse livers (Figure 2), which may be due to the change of epigenetic regulation (Li et al., 2009; Pan et al., 2009). The change in murine P450s might complicate interpretation of data obtained from these transgenic mouse models (Yu et al., 2005; Felmlee et al., 2008).

Figure 2.

Sexual dimorphism and developmental expression of human CYP3A4 transgene, as well as murine P450s. (a) Expression of the CYP3A4 transgene and murine Cyp3a and Cyp2b is dependent on gender and age [2–16 week (W)] in mouse livers (Yu et al., 2005). (b) Hepatic CYP3A4 and murine P450 expression profiles are changed in mice following growth hormone (GH) infusion, compared to the SHAM control (Cheung et al., 2006). HLM, human liver microsomes.

Table 2.

Effects of ketoconazole on midazolam pharmacokinetics (AUC_{(ketoconazole)}/AUC_(control)) in humans and animal models.

	Intravenous administration	Oral administration	Reference
Humans	5	16	(Tsunoda et al., 1999)
Rats	2	5	(Kotegawa et al., 2002)
Wild-type mice	n.a.	3	(Granvil et al., 2003)
Tg-CYP3A4 mice	n.a.	8	(Granvil et al., 2003)

Another CYP3A4-transgenic mouse model was created with an ApoE promoter-HCR1 driven expression cassette containing CYP3A4 cDNA (van Herwaarden et al., 2005) (Table 1). CYP3A4 was selectively expressed in transgenic mouse livers, leading to a reduced systemic exposure to intravenously-administered midazolam and cyclosporin A. A villin-CYP3A4-transgenic mouse model was also developed with selective CYP3A4 expression in small intestines (van Herwaarden et al., 2007). Furthermore, Cyp3a(−/−) mice lacking all functional murine Cyp3a genes were generated and bred with the ApoE- and villin-CYP3A4-transgenic mice to produce liver (Cyp3a−/−A) and intestine (Cyp3a−/−V) CYP3A4-humanized mice, respectively (van Herwaarden et al., 2007) (Table 1). Studies in the CYP3A4-genetically modified mice showed that docetaxel pharmacokinetics and toxicity were largely affected by Cyp3a. In addition, intestinal CYP3A4 was shown to determine the systemic exposure to orally administered triazolam, as well as the drug-drug interaction between ketoconazole and triazolam (van Waterschoot et al., 2009). These genetically modified mouse lines are useful models for understanding Cyp3a-mediated drug metabolism and drug-drug interaction. Nevertheless, change in mouse Cyp2c expression was observed in Cyp3a(−/−) mice (van Waterschoot et al., 2008). Since Cyp2c enzymes contribute to the metabolism of many Cyp3a substrates, it is necessary to be cautious when extrapolating data obtained from the related mouse models.

Cytochrome P450 reductase (CPR)

Disruption of CPR is anticipated to cause a global inactivation of metabolism catalyzed by all microsomal P450s because CPR is the sole electron donor for microsomal P450s. Indeed, ubiquitous disruption of Cpr led to embryonic lethality (Shen et al., 2002; Otto et al., 2003). The Cre-loxP approach was thus utilized to generate conditional knockout mouse lines lacking Cpr expression in livers (Gu et al., 2003; Henderson et al., 2003a; Wu et al., 2003) (Table 1). Other tissue-specific Cpr(−/−) mouse models were also created for intestine (Finn et al., 2007; Zhang et al., 2009), lung (Weng et al., 2007) and cardiomyocytes (Fang et al., 2008b), besides the global Cpr-low expression mice (Wu et al., 2005; Gu et al., 2007). These Cpr genetically modified mouse models were used to investigate the physiological role of Cpr (Gu et al., 2003; Wu et al., 2005; Mutch et al., 2006), and the importance of Cpr in the metabolic clearance or detoxification of many xenobiotic drugs (Pass et al., 2005; Henderson et al., 2006; Weng et al., 2007; Fang et al., 2008a; Fang et al., 2008b; Zhang et al., 2009).

Glutathione S-transferase pi (GSTP)

Human GSTP is one of the most important GST enzymes catalyzing the conjugation of glutathione (GSH) with electrophilic compounds and being responsible for cellular defense against endobiotic and xenobiotic toxins. A GstP1/P2(−/−) mouse model was created to study the function of GSTP in vivo (Henderson et al., 1998) (Table 1). Mice lacking GstP did not show any obvious phenotype, although some GstP null mice had higher body weights than controls. Studies using the GstP knockout mouse model have demonstrated that GstP could be an important metabolic determinant of 7,12-dimethylbenz anthracene (DMBA)-induced carcinogenesis (Henderson et al., 1998), APAP-induced hepatotoxicity (Henderson et al., 2000), and tobacco-induced endothelial dysfunction (Conklin et al., 2009).

Sulfotransferase 1E1 (SULT1E1)

Estrogen sulfotransferase (SULT1E1) catalyzes the sulfation of estrogens at the 3-hydroxyl position and contributes to estrogen homeostasis. A Sult1e1(−/−) mouse model was developed to examine the physiological role of Sult1e1 (Qian et al., 2001) (Table 1). Male Sult1e1(−/−) mice showed age-dependent Leydig cell hypertrophy/hyperplasia and seminiferous tubule damage, which could be recapitulated by estradiol supplementation. Further studies in Sult1e1(−/−) female mice demonstrate that Sult1e1 is a critical estrogen modulator in placenta and affects estrogen levels and placental thrombosis (Tong et al., 2005). This unique mouse model has not yet been used to study xenobiotic metabolism.

Uridine 5'-diphospho-glucuronosyltransferase 1 (UGT1)

A UGT1-transgenic mouse model was created using a BAC clone containing the entire human UGT1 locus (Chen et al., 2005) (Table 1). UGT1A proteins were differentially expressed in transgenic mouse liver and gastrointestinal tract. UGT1 transgene was inducible by AhR, PXR and PPARα ligands (Chen et al., 2005; Bonzo et al., 2007; Senekeo-Effenberger et al., 2007). The disruption of murine Ugt1 gene cluster resulted in severe hyperbilirubinemia in newborn Ugt1(−/−) mice, which resembles Crigler-Najjar Type I disease in humans (Nguyen et al., 2008) (Table 1). Furthermore, humanized mice expressing the Gilbert's UGT1A1*28 allele or the wild-type UGT1A1*1 allele were developed on the Ugt1-null background to evaluate UGT1A-dependent drug clearance (Cai et al., 2010), as well as the role of UGT1 in developmental hyperbilirubinemia and CNS toxicity (Fujiwara et al., 2010) (Table 1). These UGT1 genetically modified mice would be useful models to investigate regulation and function of UGT1 enzymes.

Drug Transporter Genetically Modified Mouse Models

P-glycoprotein (P-gp/MDR1/ABCB1)

Human P-gp, expressed at high levels on the apical membrane of enterocytes, biliary surface of hepatocytes, and the luminal (apical) side of kidney proximal tubule cells, largely affects drug absorption, distribution and excretion. To investigate the role of P-gp in vivo, mice deficient in Mdr1a [Mdr1a(−/−)] (Schinkel et al., 1994) or both Mdr1a and Mdr1b [Mdr1a/b(−/−)] (Schinkel et al., 1997) were generated (Table 3). Comparative studies in wild-type versus knockout mice demonstrated the importance of Mdr1a or Mdr1a/1b in brain uptake, pharmacokinetics and toxicity of many drugs including ivermectin, vinblastine, dexamethasone, digoxin, cyclosporine A, ondansetron, loperamide and rhodamine (Schinkel et al., 1994; Schinkel et al., 1995; Schinkel et al., 1996; Schinkel et al., 1997). For instance, Mdr1a-knockout mice showed 100-fold higher sensitivity to the neurotoxic pesticide ivermectin (Figure 3) and 3-fold higher sensitivity to the carcinostatic drug vinblastine (Schinkel et al., 1994). The Mdr1 knockout mice are useful animal models to evaluate the pharmacological and toxicological role of P-gp. Nevertheless, the obvious increase in Mdr1b expression in Mdr1a(−/−) mice (Schinkel et al., 1994) reminds one to be cautious when translating data from transgenic mice into humans.

Table 3.

Drug transporter genetically modified mouse models.

Gene	Category	Example applications	References
Mdr1a (Abcb1a)	Knockout	Drug absorption, distribution and elimination, including brain penetration	(Schinkel et al., 1994; Schinkel et al., 1995; Schinkel et al., 1996)
Mdr1a/1b (Abcb1a/1b)	Knockout	Drug distribution	(Schinkel et al., 1997)
Bsep (Abcb11)	Knockout	Bile acid distribution and excretion	(Wang et al., 2001; Wang et al., 2003)
Mrp1 (Abcc1)	Knockout	Disposition of xenobiotic drugs and endogenous leukotrienes	(Wijnholds et al., 1997) (Lorico et al., 1997)
Mrp2 (Abcc2)	Knockout	Glutathione excretion, bile flow and drug disposition	(Vlaming et al., 2006) (Chu et al., 2006)
Mrp3 (Abcc3)	Knockout	Acetaminophen disposition; biliary excretion	(Belinsky et al., 2005) (Zelcer et al., 2006)
Mrp4 (Abcc4)	Knockout	Drug resistance and brain penetration.	(Leggas et al., 2004)
Bcrp1 (Abcg2)	Knockout	Drug absorption, distribution and elimination, including the secretion of compounds into milk	(Zhou et al., 2002) (Jonker et al., 2002)
Ostα	Knockout	Bile acid disposition	(Rao et al., 2008)
Pept2 (Slc15a2)	Knockout	Role of Pept2 in drug distribution and excretion	(Rubio-Aliaga et al., 2003) (Shen et al., 2003; Shen et al., 2007)
Mate1 (Slc47a1)	Knockout	Renal secretion of metformin	(Tsuda et al., 2009)
Oct1 (Slc22a1)	Knockout	Hepatic uptake and intestinal disposition of organic cation drugs	(Jonker et al., 2001)
Oct2 (Slc22a2) Oct1/2 (Slc22a1/2)	Knockout	Renal excretion of drugs	(Jonker et al., 2003)
Oct3 (Slc22a3)	Knockout	Role of Oct3 in drug distribution	(Zwart et al., 2001)
Oat1 (Slc22a6)	Knockout	Excretion of organic anions and drugs	(Eraly et al., 2006)
Oat3 (Slc22a8)	Knockout	Uptake of organic anions and drugs	(Sweet et al., 2002) (Sykes et al., 2004)
OATP1B1 (SLCO1B1)	Transgenic	Hepatic uptake of mitoxantrone	(van de Steeg et al., 2009)
Oatp1b2 (Slco1b2)	Knockout	Drug disposition and hepatic uptake	(Zaher et al., 2008) (Lu et al., 2008) (Chen et al., 2008)

Figure 3.

Sharp difference in ivermectin (p.o.) toxicity between Mdr1a-null and wild-type mice (Schinkel et al., 1994).

Bile salt export pump (BSEP/ABCB11)

A Bsep knockout mouse model (Table 3) was developed after disruption of the coding region of Walker A of the N-terminal ATP-binding domain of Bsep (Wang et al., 2001). The Bsep(−/−) mice were viable and fertile, but showed growth retardation. In addition, the Bsep(−/−) mice exhibited various phenotypic characteristics including intrahepatic cholestasis, reduced secretion of cholic acid and output of bile salt, liver necrosis, high mortality, and change of the expression of many liver genes, supporting the critical role of Bsep in the transport of bile acid and lipid homeostasis (Wang et al., 2001; Wang et al., 2003). However, no study on drug disposition in this Bsep(−/−) mouse model has been reported.

Multidrug resistance-associated proteins (MRP/ABCC)

MRP1/ABCC1 is ubiquitously expressed in humans and has an important role in disposition of endobiotic and xenobiotic agents. Two different Mrp1(−/−) mouse lines (Table 3) were generated by targeted disruption of either the exons encoding the first ATP-binding domain of mouse Mrp1 gene (Wijnholds et al., 1997) or part of the second putative ATP-binding domain (Lorico et al., 1997). Both mouse lines exhibited increased sensitivity to anticancer drugs such as etoposide and vincristine (Lorico et al., 1997; Wijnholds et al., 1997), indicating the critical role of Mrp1 in drug disposition. Two Mrp2(−/−) mouse lines were also developed (Chu et al., 2006; Vlaming et al., 2006) (Table 3). Absence of Mrp2 resulted in a decreased biliary excretion of Mrp2 substrates and an increased systemic exposure to several drugs and carcinogens after oral administration, illustrating the critical role of Mrp2 in disposition of these substances.

To examine the function of MRP3 in vivo, two Mrp3 knockout mouse models (Table 3) were developed after targeted disruption of exons 6–8 (Belinsky et al., 2005) or exons 2–8 (Zelcer et al., 2006). The resulting Mrp3(−/−) mice have been employed to assess the function of Mrp3 in disposition of endogenous bile acids and bilirubin glucuronide (Belinsky et al., 2005; Zelcer et al., 2006), as well as morphine, morphine-6-glucuronide and APAP (Manautou et al., 2005; Zelcer et al., 2006). A Mrp4 knockout mouse model (Table 3) was also generated by disruption of exon 27, and utilized to evaluate the role of Mrp4 in renal elimination and brain penetration of topotecan (Leggas et al., 2004). The Mrp4(−/−) mouse model was further employed to demonstrate the contribution of Mrp4 to renal elimination of antiviral drugs (e.g., adefovir and tenofovir) (Imaoka et al., 2007), and cytotoxicity and tissue distribution of 9-(2-(phosphonomethoxy)ethyl)-adenine (Takenaka et al., 2007).

Breast cancer resistance protein (BCRP/ABCG2)

BCRP/ABCG2 plays an important role in disposition of many xenobiotic drugs (e.g., mitoxantrone, doxorubicin and topotecan) and metabolites (e.g., estrone-3-sulfate and estradiol-17-β-D-glucuronide). Two Bcrp1(−/−) mouse models (Table 3) were created to investigate the physiological and pharmacological function of BCRP/ABCG2 (Jonker et al., 2002; Zhou et al., 2002). Studies using Bcrp1(−/−) mouse models have revealed a critical role for Bcrp1 in protection of hematopoietic stem cells from toxins (Zhou et al., 2002), and in prevention of diet-dependent photoxocity (Jonker et al., 2002). In addition, the Bcrp1(−/−) mice were used to demonstrate the effects of Bcrp1 expression on absorption, distribution, and elimination of dietary carcinogens (e.g., PhIP) and medications (e.g., nitrofurantoin) (van Herwaarden et al., 2003; Breedveld et al., 2005; Jonker et al., 2005; Merino et al., 2005; Merino et al., 2006). The Bcrp1(−/−) mice may serve as unique animal models for understanding the disposition of drugs transported by BCRP.

Organic solute transporter α (OSTα)

OSTα and OSTβ are coexpressed on the intestinal basolateral membrane, forming a functional heteromer to transport many organic compounds such as taurocholate, digoxin and prostaglandin E1. An Ostα knockout mouse model (Table 3) was developed to evaluate the function of the Ostα-Ostβ heteromeric transporter in intestinal bile acid reabsorption after deletion of the proximal promoter and exons 1 and 2 of murine Ostα gene (Rao et al., 2008). The transileal transport study using the everted gut sac prepared from Ostα(−/−) mice demonstrated that Ostα-Ostβ heteromer is critical for intestinal transport of bile acids.

Peptide transporter 2 (PEPT2/SLC15A2)

A constitutive Pept2-deficient mouse model and a conditional Pept2 knockout mouse model (Table 3) were developed to examine the physiological and pharmacological role of PEPT2 (Rubio-Aliaga et al., 2003; Shen et al., 2003). Comparative studies in knockout versus wild-type mice demonstrated a critical role for PEPT2 in uptake of dipeptides (Rubio-Aliaga et al., 2003; Shen et al., 2003). Further studies revealed that renal PEPT2 was fully responsible for reabsorption of cefadroxil in kidney, and choroid plexus PEPT2 limited brain penetration of cefadroxil (Shen et al., 2007), indicating that the Pept2 knockout mouse model could be useful in the development of peptidomimetic drugs.

Organic cation transporters (OCT/SLC22A)

A number of Oct-knockout mouse models (Table 3) were developed to examine the physiological and pharmacological functions of OCT transporters (Jonker et al., 2001; Zwart et al., 2001; Jonker et al., 2003). Studies in the Oct1(−/−) mice demonstrated an important role for Oct1/Slc22a1 in the distribution and excretion of tetraethylammonium, as well as hepatic accumulation of neurotoxin 1-methyl-4-phenylpyridium and anticancer drug metaiodobenzylguanidine (Jonker et al., 2001). Studies in Oct2 single knockout and Oct1/2 double knockout mice indicated that Oct1 and Oct2/Slc22a2 together were critical for renal secretion of organic cations (Jonker et al., 2003). The mouse homolog of OCT3 was also disrupted to generate an Orct3(−/−) mouse line, which was used to reveal a critical role for Oct3/Slc22a3 in the accumulation of 1-methyl-4-phenylpyridium in the heart (Zwart et al., 2001).

Organic anion transporters (OAT/SLC22A)

OAT/SLC22 transporters are located on the barrier epithelia of diverse tissues such as the proximal tubule of kidney and choroid plexus of brain, where they mediate the absorption and excretion of many organic anions (e.g., β-lactam antibiotics). Disruption of murine Oat1/Slc22a6 or Oat3/Slc22a8 did not alter the fertility of the resulting gene knockout mice (Sweet et al., 2002; Eraly et al., 2006) (Table 3). Studies with Oat1(−/−) mice demonstrated a critical role for Oat1 in the renal excretion of organic anions and drugs (Eraly et al., 2006). The Oat3(−/−) mice were utilized to investigate the contribution of Oat3 to renal and brain uptake of several drugs including para-aminohippurate and estrone sulfate taurocholate (Sweet et al., 2002; Sykes et al., 2004).

Organic anion transporting polypeptides (OATP/SLCO)

To investigate the role of Oatp1b2 in hepatic uptake of xenobiotic drugs, Oatp1b2(−/−) mouse models (Table 3) were developed by the targeted disruption of exons 10–12 (Zaher et al., 2008) or exon 3 of Oatp1b2/Slco1b2 (Lu et al., 2008). The Oatp1b2(−/−) mice were used to demonstrate an important role for Oatp1b2 in the hepatic uptake, systemic disposition and toxicity of many xenobiotic drugs including phalloidin, microcystin-LR, pravastatin, rifampicin and lovastatin (Chen et al., 2008; Lu et al., 2008; Zaher et al., 2008). Recently, a human OATP1B1-transgenic mouse model with liver-specific OATP1B1 expression (Table 3) was generated and found to have significantly higher hepatic accumulation of methotrexate than wild-type control mouse (van de Steeg et al., 2009). These genetically modified mouse models should be of great value for defining the role of OATP in drug disposition.

Multidrug and toxin extrusion 1 (MATE1/SLC47A1)

MATE1/SLC47A1 is a proton-dependent transporter mainly expressed on the luminal side of renal proximal tubules and bile canaliculi, and mediates the excretion of many exogenous or endogenous organic cations. Targeted disruption of murine Mate1/Slc47a1 did not alter fertility of the resultant knockout mice (Tsuda et al., 2009) (Table 3). However, lack of Mate1 expression caused a sharp decrease in renal clearance of metformin and a considerable increase in systemic exposure to metformin, indicating an important role for Mate1 in metformin disposition.

Xenobiotic Receptor Genetically Modified Mouse Models

Aryl hydrocarbon receptor (AHR)

A number of Ahr knockout mouse models (Table 4) were developed independently through disruption of exon 1 or exon 2 (Fernandez-Salguero et al., 1995; Schmidt et al., 1996; Mimura et al., 1997). These Ahr-null mouse strains showed some physiological changes including lower growth rate, decreased fertility, and/or liver deficiency. Studies in Ahr(−/−) versus wild-type mice demonstrated a crucial role for Ahr in induction of Cyp1a enzymes, dioxin-induced toxicity and carcinogenesis. In addition, an AHR-humanized mouse model (Table 4), in which murine Ahr was replaced by human AHR after homologous recombination, was generated to investigate species difference in dioxin toxicity (Moriguchi et al., 2003). Compared with wild-type Ahr(+/+) mice, the AHR-humanized mice indeed exhibited a different induction profile following the treatment of AHR ligands. These genetically modified mice are useful models for predicting the risk of xenobiotic drugs in humans.

Table 4.

Xenobiotic receptor genetically modified mouse models.

Gene	Category	Example applications	References
Ahr	Knockout	Role of Ahr in gene regulation and liver function	(Fernandez-Salguero et al., 1995) (Schmidt et al., 1996) (Mimura et al., 1997)
AHR	Humanized	Species difference in AHR-mediated gene regulation	(Moriguchi et al., 2003)
Car (Nr1i3)	Knockout	Role of Car in gene regulation and liver toxicity	(Wei et al., 2000)
CAR (NR1I3)	Humanized Transgenic	Species differences in CAR-mediated gene regulation Gene regulation and bile acid detoxification	(Zhang et al., 2002; Huang et al., 2004) (Saini et al., 2004)
Pparα (Nr1c1)	Knockout	Gene regulation and peroxisome proliferation	(Lee et al., 1995)
PPARα (NR1C1)	Humanized	Species difference in PPARα-mediated gene regulation	(Cheung et al., 2004)
Pxr (Nr1i2)	Knockout	Role of Pxr in regulation of drug metabolism and disposition	(Xie et al., 2000) (Staudinger et al., 2001)
PXR (NR1I2)	Humanized	Species difference in PXR-mediated gene regulation	(Staudinger et al., 2001) (Ma et al., 2007)
Rxrα (Nr2b1)	Knockout	Physiological role of Rxrα (embryonic lethality)	(Kastner et al., 1994; Sucov et al., 1994)
Rxrα (hepatocyte)	Knockout	Regulation of endobiotics and xenobiotics metabolism	(Wan et al., 2000)

Pregnane X receptor (PXR/NR1I2)

PXR regulates the expression of many enzymes (e.g., CYP3A4 and UGT1A) and transporters (e.g., ABCB1) important in drug metabolism and disposition. Some Pxr(−/−) mouse models (Table 4) were developed to study its physiological and pharmacological role after targeted disruption of exons 2–3 or exon 1 (Xie et al., 2000; Staudinger et al., 2001). Induction of various drug-metabolizing enzymes and transporters by PXR ligands was abolished in mice lacking Pxr expression (Xie et al., 2000; Staudinger et al., 2001). In addition, human PXR-transgenic mice were generated using PXR cDNA or a BAC clone, and then crossbred with Pxr-null mice to produce PXR-humanized mouse models (Xie et al., 2000; Ma et al., 2007) (Table 4). Differential regulation of drug metabolism by human PXR and mouse Pxr was demonstrated by studies in wild-type versus PXR-humanized mice where they exhibited differences in ligand specificity (Figure 4). These findings indicate that PXR-humanized and Pxr-knockout mice are unique animal models to study PXR ligands and PXR-mediated drug-drug interaction.

Figure 4.

Differential induction of CYP3A4/Cyp3a by pregnenolone-16α-carbonitrile (PCN) and rifampicin (RIF) in wild-type and PXR-humanized mice (Ma et al., 2007). Mice treated with drug vehicle were used as control (CON).

Constitutive androstane receptor (CAR/NR1I3)

CAR is an important xenobiotic receptor regulating the expression of various enzymes (e.g., CYP2B6, UGT1A1 and GSTA1) and transporters (e.g., ABCC2). A Car(−/−) mouse model was developed to study the function of CAR in vivo (Wei et al., 2000) (Table 4). The absence of Cyp2b10 induction by Car activators (e.g., phenobarbital) in Car(−/−) mice supports the essential role for CAR in the regulation of Cyp2b10 expression (Wei et al., 2000). The Car(−/−) mice were resistant to APAP hepatotoxicity, indicating CAR as a modulator in APAP-induced toxicity (Zhang et al., 2002). A transgenic mouse model [VP-CAR(+/+)] with human CAR expression was also developed (Saini et al., 2004) (Table 4), which showed higher hepatic sulfation activity toward lithocholic acid (LCA) and increased resistance to LCA-induced hepatotoxicity, compared to wild-type mice (Saini et al., 2004). Another CAR-transgenic mouse model was created using human CAR cDNA, and crossbred with Car(−/−) mice to produce a CAR-humanized mouse model (Huang et al., 2004) (Table 4). Interestingly, meclizine was revealed as an agonist for mouse Car and an inverse agonist for human CAR by studies in wild-type, Car(−/−) and CAR-humanized mice, illustrating the potential application of these mouse models to preclinical assessment of CAR-mediated regulation of drug metabolism.

Peroxisome proliferator-activated receptor α (PPARα/NR1C1)

PPARα regulates CYP4A expression and provokes peroxisome proliferation and hepatocarcinogenesis in rodents. A Pparα(−/−) mouse model was developed after targeted disruption of exon 8 of murine Pparα (Lee et al., 1995) (Table 4). The absence of peroxisome proliferation and Cyp4a induction in Pparα(−/−) mice treated with Pparα ligands (e.g., clofibrate and Wy-14,643) demonstrated a critical role for PPARα in peroxisome proliferation and regulation of target genes (Lee et al., 1995; Gonzalez et al., 1998). In addition, a PPARα-humanized mouse model (Table 4) was generated to investigate the species difference in PPARα functionality (Cheung et al., 2004). Despite peroxisome proliferator induced hypolipidemia and many enzymes in fatty acid oxidation, synthesis, transport in both wild-type and PPARα-humanized mice, the increase in peroxisome and induction of cell cycle genes, hepatocellular proliferation and hepatomegaly were not observed in PPARα-humanized mice but only in wild-type mice (Cheung et al., 2004; Morimura et al., 2006; Foreman et al., 2009) (Table 5). Thus the PPARα-humanized mice may be a better animal model to predict risk of peroxisome proliferators (e.g., lipid lowering fibrate drugs) in humans.

Table 5.

Comparison of the responses to peroxisome proliferators (e.g., Wy-14,643) in wild-type, Pparα-knockout, PPARα-humanized mice and humans.

Response to peroxisome proliferators	Wild-type mice	Ppara-null mice	PPARa-humanized mice	Humans
PPARa expression	Murine Pparα	None	Human PPARα	Human PPARα
Increase in peroxisomes	Yes	No	No	No
Induction of enzymes in fatty acid oxidation, synthesis, transport	Yes	No	Yes	No
Induction of cell cycle genes	Yes	No	No	No
Cell proliferation	Yes	No	No	No
Hypolipidemia	Yes	No	Yes	Yes
Increased risk of cancer	Yes	No	No	No

Retinoid X receptor α (RXRα/NR2B1)

The RXRα is a heterodimer partner for PXR, CAR and PPARα regulating liver functions. The conventional Rxrα(−/−) mice (Table 4) developed by targeted disruption of exon 3 or exon 4 of RXRα/NR2B1 showed embryonic lethality, which was probably due to cardiac or placental developmental defects (Kastner et al., 1994; Sucov et al., 1994). Therefore, a Rxrα(−/−)_hep mouse model (Table 4), in which RXRα is conditionally and selectively disrupted in hepatocytes, was developed for investigation of the physiological functions of RXRα (Wan et al., 2000). While no gross liver abnormalities were found in the Rxrα(−/−)_hep mice, many hepatic metabolic pathways mediated by heteromerization between RXRα and xenobiotic receptors (e.g., PPARα, CAR and PXR) were changed (Wan et al., 2000; Wu et al., 2004; Dai et al., 2005). The Rxrα(−/−)_hep mouse line is an unparalleled animal model for studying the function of RXRα in the homeostasis of cholesterol, fatty acids, bile acids and steroids, as well as in drug metabolism and disposition.

Conclusions

Genetically modified mice with lost (knockout) or gained (transgenic) function are unique and valuable additions to in vitro and wild-type animal models for preclinical investigations. Studies with these mouse models help better understand the in vivo role of specific drug-metabolizing enzyme, transporter or xenobiotic receptor in drug metabolism and pharmacokinetics, and consequent effects on drug efficacy or toxicity. Development of humanized mouse models is of particular interest, which may ultimately overcome species difference that impairs accurate prediction of drug metabolism and disposition in humans. However, whether the level of transgene expression in mice could be translated into humans is undefined in many cases. In addition, the disruption of target genes or introduction of genes of interest can sometimes lead to changes in the expression of other related genes in the genetically modified mice. Mouse genetic backgrounds could also affect the results. One should not overlook the limitations when studying drug metabolism and pharmacokinetics with such genetically modified mouse models.