Expression of Constitutive Androstane Receptor, Hepatic Nuclear Factor 4 , and P450 Oxidoreductase Genes Determines Interindividual Variability in Basal Expression and Activity of a Broad Scope of Xenobiotic Metabolism Genes in the Human Liver

Identification of genetic variation predictive of clearance rate of a wide variety of prescription drugs could lead to cost-effective personalized medicine. Here we identify regulatory genes whose variable expression level among individuals may have widespread effects upon clearance rate of a variety of drugs. Twenty liver samples with variable CYP3A activity were profiled for expression level and activity of xenobiotic metabolism genes as well as genes involved in the regulation thereof. Regulatory genes whose expression level accounted for the highest degree of collinearity among expression levels of xenobiotic metabolism genes were identified as possible master regulators of drug clearance rate. Significant linear correlations (p < 0.05) were identified among mRNA levels of CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, MRP2, OATP2, P450 oxidoreductase (POR), and UDP-glucuronosyltranferase 1A1, suggesting that these xenobiotic metabolism genes are coregulated at the transcriptional level. Using partial regression analysis, constitutive androstane receptor (CAR) and hepatic nuclear factor 4 (HNF4 ) were identified as the nuclear receptors whose expression levels are most strongly associated with expression of coregulated xenobiotic metabolism genes. POR expression level, which is also associated with CAR and HNF4 expression level, was found to be strongly associated with the activity of many cytochromes P450. Thus, interindividual variation in the expression level of CAR, HNF4 , and POR probably determines variation in expression and activity of a broad scope of xenobiotic metabolism genes and, accordingly, clearance rate of a variety of xenobiotics. Identification of polymorphisms in these candidate master regulator genes that account for their variable expression among individuals may yield readily detectable biomarkers that could serve as predictors of xenobiotic clearance rate. Interindividual variation in drug clearance rate is often responsible for toxicity or inefficacy of prescription drugs. Systemic drug clearance rate is determined by hepatic expression and activity of phase I oxidative cytochromes P450 (P450s), phase II conjugative enzymes, and transporter proteins. Expression of these metabolic enzymes is coordinately regulated by a network of transcription factors (Pascussi et al., 2004; Xu et al., 2005) exemplified in Fig. 1. The network is composed of ligand-activated nuclear receptors that recognize a variety of endogenous and xenobiotic compounds to activate transcription of metabolic enzymes involved in biotransformation and transport. Multiple nuclear receptors can recognize response elements of the same target gene (Fig. 1) and may control their own expression as well as the expression of other nuclear receptors in the pathway (Maglich et al., 2002; Pascussi et al., 2004). In addition, these regulatory proteins share a common pool of coregulators and the common heterodimeric partner, the retinoid X receptor (RXR). The role of xenobiotic metabolism gene polymorphisms in determining clearance rate of specific prescription drugs has been extensively studied. The polymorphisms of several P450 genes can predict corresponding P450 activity level (Ingelman-Sundberg et al., 1999). However, it has become increasingly clear that trans-acting factors are also involved in determining expression and activity of xenobiotic metabolism genes. For example, the activity of CYP3A4, the most active P450 in the human liver, exhibits a high degree (at least 40-fold) of interindividual variation (Paine et al., 1997); however, CYP3A4 polymorphisms are not accurate predictors of CYP3A4 activity. Indeed, CYP3A4 activity exhibits a unimodal distribution among the population, suggesting that a multitude of factors are involved in dictating CYP3A4-mediated pathways (Wilkinson, 1996). Estimates suggest that 90% of the variation in CYP3A4 activity may be attributed to genetic factors (Ozdemir et al., 2000). Accordingly, polymorphisms of genes involved in the regulatory network affecting This study is supported by National Institutes of Health Grant R01 CA53596 and Centers of Biomedical Research Excellence (COBRE) P20 RR021940 as well as the Molecular Biology Core supported by the COBRE grant. Article, publication date, and citation information can be found at http://dmd.aspetjournals.org. doi:10.1124/dmd.107.016436. ABBREVIATIONS: P450, cytochrome P450; MRP2, multidrug resistance-associated protein 2; OATP2, organic anion-transporting polypeptide 2; POR, P450 oxidoreductase; UGT1A1, UDP-glucoronosyltransferase 1A1; CAR, constitutive androstane receptor; HNF4 , hepatic nuclear factor 4 ; RXR, retinoid X receptor; SXR, steroid and xenobiotics receptor; PCR, polymerase chain reaction; GR, glucocorticoid receptor. 0090-9556/07/3509-1700–1710$20.00 DRUG METABOLISM AND DISPOSITION Vol. 35, No. 9 Copyright © 2007 by The American Society for Pharmacology and Experimental Therapeutics 16436/3247157 DMD 35:1700–1710, 2007 Printed in U.S.A. 1700 at A PE T Jornals on Sptem er 0, 2017 dm d.aspurnals.org D ow nladed from CYP3A4 expression or activity likely have effects upon the rate of CYP3A4-mediated metabolism. Recent studies suggest that variation in trans-acting master regulator genes may account for differential expression of a large proportion of genes and determine complex traits (Schadt et al., 2003; Yvert et al., 2003). Indeed, complex phenotypes are often dictated by variation across entire pathways (Evans and Relling, 2004). Thus, polymorphisms in master regulator genes could serve as biomarkers to predict drug clearance rate (Carlberg and Dunlop, 2006). Genotyping for single polymorphisms predicting a complex phenotype would be more practical and cost-effective than genotyping for many individual xenobiotic metabolism genes. Variable nuclear receptor expression level may determine target gene expression level by affecting constitutive activity in the case of some nuclear receptors or by modulating the degree of response to ligands in others. For example, CAR mRNA expression level is highly variable (at least 240-fold) among individuals (Chang et al., 2003), and livers expressing high levels of CAR have enhanced expression of CAR target genes (Finkelstein et al., 2006). Additionally, the abundance of common coregulator proteins and RXR may serve as limiting factors to nuclear receptor-mediated transcriptional control. Finally, the rate of regeneration of P450s by P450 oxidoreductase (POR), the obligate P450 regenerative enzyme in mammals (Wu et al., 2005), may dictate phase I metabolism rate. A number of isolated studies have previously identified interindividual variation in expression levels of nuclear receptors such as CAR, SXR, and hepatic nuclear factor 4 (HNF4 ) and their correlations with target genes (Pascussi et al., 2001; Chang et al., 2003; Lamba et al., 2003). However, a systematic screening of the effects of variable expression of nuclear receptors and their coregulators has not been conducted. Here we hypothesize that variation in expression level of previously unidentified master regulator genes is responsible for determining basal expression level and activity of a broad scope of xenobiotic metabolism genes in the human liver. In this study we use linear regression analysis to identify associations between nuclear receptors and xenobiotic metabolism genes as well as to identify patterns of collinearity. We identify a high degree of transcriptional coregulation among the xenobiotic metabolism genes CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, multidrug resistance-associated protein 2 (MRP2/ABCC2), organic anion-transporting polypeptide 2 (OATP2/SLCO1B1), POR, and UDP-glucuronosyltranferase 1A1 (UGT1A1). We also identify the nuclear receptors CAR and HNF4 as extensive regulators of xenobiotic metabolism whose expression level modulates basal transcription of a broad scope of xenobiotic metabolism genes. Finally, we identify POR expression level, which is closely associated with expression level of CAR and HNF4 , to be a putative limiting factor for phase I enzyme activity. Materials and Methods Basal Enzyme Activities. Twenty human liver samples with variable CYP3A activity were obtained from the National Disease Research Interchange (Philadelphia, PA) and the Midwest Transplant Network (Westwood, KS). Demographics for human liver samples are listed in Table 1. Considering that CYP3A activity is probably indicative of overall liver function, using samples with variable CYP3A activity increases the statistical power of this study. Variable samples allow for the identification of significant correlations while minimizing the number of rare liver samples used. All liver samples have tested negative for human immunodeficiency virus (HIV) and hepatitis and were not suspected of harboring other infectious diseases. Liver microsomes were prepared as described previously (Obach et al., 2001) to determine P450 enzyme activity rates as described by Madan et al. (1999) using the probe drugs listed in Table 2. Incubation conditions were controlled to ensure linear metabolite formation with respect to reaction time and protein concentration. Identification of Relevant Nuclear Receptor, Coregulator, and Xenobiotic Metabolism Genes. Nuclear receptors and coregulators were selected for the study based upon their ability to trans-activate common xenobiotic metabolism genes. Selection was performed using Ingenuity Pathways Analysis (Ingenuity Systems, Inc., Redwood City, CA) and a literature search (Pascussi et al., 2004; Xu et al., 2005). mRNA level was subsequently quantified for 10 nuclear receptors, 5 coregulators, 9 cytochrome P450 enzymes, 3 transport proteins, and the obligate regenerative enzyme for P450 enzymes, POR, as listed in Table 3. Quantitative Real-Time PCR. Liver samples were lysed in the TRIzol reagent (Invitrogen, Carlsbad, CA) and total RNA was isolated according to the manufacturer’s protocol. RNA purity was confirmed by a 260:280 nm absorbance ratio greater than 1.5, and agarose gel electrophoresis with ethidium bromide staining was used to confirm integrity of 18S and 28S ribosomal RNA bands. Reverse transcription was carried out using MMLV Reverse Transcriptase (Invitrogen) according to the manufacturer’s protocol. Primers and probes for real-time PCR amplification were designed using Primer Express 3.0 according to the manufacturer’s instructions (Real-Time PCR Chemistry Guide, Applied Biosystems, Foster City, CA) or as described in the literature (Table TABLE 1 Demographics and CYP3A4/5 activity of individual human liver samples Smoking and alcohol use data were obtained from families of the donors. Recent smoking activity is defined as regular use in the 5 years before death. Liver microsomes were prepared and P450 enzyme activity rates were determined as described under Materials and Methods. Liver Sample No. Age Gender Ethnicity Recent Smoker? Alcohol Use? CYP3A4/5 Activity yr pmol/mg protein/min 1 1 M C N N 448 2 1 M H N N 207 3 4 F C N N 9890 4 9 M C N N 2900 5 17 M C Y Y 3620 6 24 M C N Y 5580 7 36 F C Y N 3940 8 42 M C Y Y 268 9 42 M C N Y 772 10 43 M C N Y 1960 11 49 M C N Y 4410 12 49 F C Y N 16,900 13 52 M C N N 1640 14 54 F C N.D. N.D. 1425 15 56 F C Y N 20,300 16 59 M H Y Y 16,400 17 61 F C Y N.D. 765 18 65 M AA N Y 9100 19 66 M C N Y 888 20 66 M C Y Y 201 C, Caucasian; H, Hispanic; AA, African American; N.D., not determined. FIG. 1. An example of coordinated regulation of xenobiotic metabolism pathways and cross-talk among transcription factors. Arrows indicate transcriptional regulation. Refer to Table 3 for gene abbreviations. 1701 REGULATION OF DRUG METABOLISM GENE EXPRESSION at A PE T Jornals on Sptem er 0, 2017 dm d.aspurnals.org D ow nladed from