Abstract
The hypothesis was tested that sequence diversity in pregnane X receptor (PXR) cis-regulatory regions is a significant determinant of variation in inducible and constitutive CYP3A4 expression. A combination of comparative genomics and computational algorithms was used to select regions of the human PXR promoter and intron 1 that were resequenced in the polymorphism discovery resource 24 DNA subset. PXR single nucleotide polymorphisms (SNP) were then genotyped in donor human livers phenotyped for CYP3A4 and multidrug resistance protein 1 mRNA and primary human hepatocytes phenotyped for basal and rifampin-inducible CYP3A4 activity. The human PXR promoter [16.9 kilobase (kb)] was significantly larger than in rodents (2.9 kb). Eighty-nine SNPs were identified in the promoter and intron 1 of PXR. The SNPs most consistently associated with CYP3A4 phenotypic measures were a 44477T>C(-1359) promoter SNP (in linkage disequilibrium with SNP 463970, a 6-base pair deletion in intron 1a, and SNP 46551, a C nucleotide insertion in intron 1b); SNP 63396C>T in intron 1 (in linkage disequilibrium with SNP 63704A>G, a 63813(CAAA)(CA) variable repeat, and SNP 65104T>C); and SNP 56348C>A, SNP 69789A>G, and SNP 66034T>C. Donor livers with the variant PXR alleles had altered hepatic expression of PXR targets compared with livers with PXR wild-type alleles. These results identified PXR promoter and intron 1 SNPs associated with PXR target gene expression (CYP3A4) in donor livers and cultured hepatocytes and that a striking number of the linked intron 1 SNPs will affect putative binding sites for hepatic nuclear factor 3β (FOXA2), a transcription factor linked with PXR expression.
The pregnane X receptor (PXR), also known as the steroid and xenobiotic receptor and NR1I2, is a member of the orphan nuclear receptor family (Bertilsson et al., 1998; Blumberg et al., 1998; Lehmann et al., 1998). Several PXR isoforms exist with different amino terminal sequences. PXR and PAR2 are mutually and exclusively transcribed from exon 1a and exon 1b, respectively. PXR has a CTG translation start site in exon 2, whereas PAR2 has an ATG translation start site in exon 1b leading to an additional 39 amino acids at the amino terminal end. PXR and PAR2 are ligand-activated by many drugs and induce transcription of numerous drug detoxification genes, including CYP3A4 and ABCB1/multidrug resistance protein 1 (MDR1) encoding P-glycoprotein (Synold et al., 2001). CYP3A4 is involved in the oxidative metabolism of more than 50% of prescribed drugs along with maintenance of endogenous steroid homeostasis (Wrighton et al., 1996). Thus, administration of PXR ligands along with substrates of CYP3A4 and MDR1 can result in accelerated drug metabolism and clearance as a result of induction of CYP3A4 and MDR1 and is the basis of many drug-drug interactions. For example, combinatorial treatment with rifampin requires dosage adjustment to maintain therapeutic efficacy (Borcherding et al., 1992).
There is well documented variability in both basal and PXR-inducible CYP3A4 expression and activity within the population (Perry and Jenkins, 1986; Watkins et al., 1989; Kolars et al., 1992; Lown et al., 1997; Lamba et al., 2002, 2006). However, single nucleotide polymorphisms (SNPs) in the coding region of the CYP3A4 gene are not frequent enough to explain the observed variability in CYP3A4. The finding that expression of PXR and CYP3A4 is significantly related in human livers raised the possibility that PXR sequence variation might explain some differences in CYP3A4 expression (Chang et al., 2003; Lamba et al., 2004). However, sequence diversity in the PXR coding region is not sufficient to explain variation in either PXR or its transcriptional target CYP3A4 (Hustert et al., 2001; Zhang et al., 2001).
The present study used a combination of comparative genomics and computational algorithms to select regions of the human PXR promoter and intron 1 for resequencing. These SNPs were then related to CYP3A4 basal and inducible expression to identify SNPs contributing to the substantial variability observed in hepatic PXR and target gene expression.
Materials and Methods
In Silico Analysis of the PXR Gene to Identify Regions to Resequence. Several web-based bioinformatic tools were used to screen 37 kilobase (kb) of the PXR gene (16 kb promoter and 21 kb intron 1) for the presence of DNA response elements for liver-enriched transcription factors and for regions of high evolutionarily conservation between multiple species. Cister plot (http://zlab.bu.edu/~mfrith/cister.shtml), NUBISCAN (http://www.nubiscan.unibas.ch/), and Transfac (http://transfac.gbf.de/TRANSFAC/lists/matrix/matrixByName.html) were used to identify regions harboring clusters of DNA response elements for various transcription factors. Transcription factors (TF) for which matrices were available to do Cister include DR4, PXR, DR3, DR1, hepatic nuclear factor (HNF)4α, DR1, HNF1, HNF3α, HNF4, HNF3β, CCAAT enhancer binding protein (CEBP)α, CEBP, CEBPδ, HNF6, CEBPβ, GATA4, glucocorticoid receptor, HNF3γ, nuclear factor (NF)-1, NF-κB, SP1 and TATA, cAMP response element, ERE, NF-1, E2F, Mef-2, Myf, CCAAT, activator protein-1, Ets, Myc, GATA, LSF, SRF, CDX, and Tef. The UCSC (http://genome.ucsc.edu), evolutionary conserved region (ECR; http://ecrbrowser.dcode.org/), and rVISTA (http://genome.lbl.gov/vista/rvista/submit.shtml) genome browsers were used to identify regions of conservation between humans and other species on the PXR gene.
Subjects. Institutional Review Boards and Clinical Research Advisory Committees at St. Jude Children's Research Hospital and the University of Pittsburgh (Pittsburgh, PA) approved the use of tissue samples from organ donors.
Study Populations.Cohort I. A subset of 24 samples encompassing the same range of diversity as the complete set of Polymorphism Discovery Resource (PDR24) was purchased from the Coriell DNA repository (http://ccr.coriell.org/nigms/products/pdr.html) and was used for discovery of SNPs in the potential regulatory regions of PXR.
Cohort IIa. Primary human hepatocyte culture. Tissues were provided by the Liver Tissue Procurement and Distribution System (National Institutes of Health Contract N01-DK-9-2310) and the Cooperative Human Tissue Network. Forty-six human livers (Caucasians; males, n = 29, females, n = 17, all CYP3A5 nonexpressors) were procured from donor organs that were not suitable for whole organ transplantation or from remaining tissue after reduced allograft transplantation. Human hepatocytes were isolated (Strom et al., 1998), and cells were plated on collagen-coated six-well plates or 60-mm culture dishes and maintained in modified Williams' E medium (Hyclone, Logan, UT) for 48 h (washout phase) and then treated with vehicle or 10 μM rifampin for 48 h. The activity of CYP3A4 was determined by measuring the testosterone metabolism in cultured hepatocytes treated with vehicle (defined as basal controls) or rifampin (Kostrubsky et al., 1999).
Cohort IIb. Liver resource for hepatic mRNA quantification of PXR, PAR2, CYP3A4, and MDR1. Human liver tissue for cohort IIb was processed through Dr. Mary Relling's laboratory at St. Jude Children's Research Hospital and was provided by the Liver Tissue Procurement and Distribution System (National Institutes of Health Contract N01-DK-9-2310) and the Cooperative Human Tissue Network. Total RNA was isolated from the liver tissue from organ donors (46 Caucasian livers; males, n = 29, females, n = 17) using TRIzol (Invitrogen, Carlsbad, CA). First-strand cDNA was prepared using oligo(dT) primers [Invitrogen, Superscript reverse transcription-polymerase chain reaction (PCR) system]. Real-time PCR quantitation of PXR, PAR2, CYP3A4, MDR1, and the quality/housekeeping control glyceraldehyde-3-phosphate dehydrogenase was carried out using the QuantiTect SYBR Green PCR kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. cDNA was analyzed in duplicate by real-time PCR on an ABI PRISM 7900HT Sequence Detection System (PE Applied Biosystems, Foster City, CA). Primers used for real-time quantification are provided in Table 1. Specificity of amplification was confirmed in each case by performing melt curve analysis and agarose gel electrophoresis. The averaged Ct values for each gene were analyzed by the comparative Ct method to obtain relative mRNA expression levels (Lamba et al., 2004).
Resequencing PXR. Fifteen regions of interest, based on the in silico analysis (Fig. 1; Table 1), were sequenced for discovery of genetic variants in the PDR24 DNA samples. PCR amplifications were performed using primers described in Table 1. Amplification was carried out in a 1× PCR buffer using 50 ng of genomic DNA, 10 pmol each of forward and reverse primers, 0.2 mM deoxynucleoside-5′-triphosphate, and 1.5 units of Taq polymerase (Expand High Fidelity PCR System, Roche, Basel, Switzerland). PCR products were checked for the correct size by agarose gel electrophoresis. Before sequencing, unincorporated nucleotides and primers were removed by incubation with shrimp alkaline phosphatase (USB, Cleveland, OH) and exonuclease I (USB) for 30 min at 37°C, followed by enzyme inactivation at 80°C for 15 min.
Sequencing was carried out by an ABI Prism 3700 automated sequencer using the PCR primers or internal primers (see Table 1). Sequences were assembled using the Phred-Phrap-Consed package (University of Washington, Seattle, WA; http://droog.mbt.washington.edu/PolyPhred.html), which automatically detects the presence of heterozygous single nucleotide substitutions by fluorescence-based sequencing of PCR products (Nickerson et al., 1997; Rieder et al., 1998).
We should point out that because we resequenced PXR in DNA from the PDR24 (representing multiple racial populations), the SNP changes observed in the combined population (e.g., SNP44477T>C) were sometimes opposite to the frequency observed in the Caucasian subjects in this study. For example, the 44477T allele was more frequent in the PDR24, but the 44477C allele frequency was higher in the Caucasian subjects in this study.
Statistical Analysis. Because the phenotypic markers were not normally distributed, group differences were analyzed nonparametrically using the Wilcoxon rank sum test to compare binary groups (e.g., GG + GT versus TT). The Kruskal-Wallis test was used to compare three groups of genotype for each polymorphism (e.g., GG versus GT versus TT) with the phenotype.
All the statistical calculations were performed using statistical program R: A Language and Environment for Statistical Analysis (http://www.R-project.org). To assess the effects of combinatorial SNPs, samples were stratified by the most common informative genotypes (e.g., 63396 heterozygotes), and then the effect of additional genotypes on phenotypic measures within this group was determined as above.
Results
In Silico Analysis of the PXR Promoter and Intron 1 to Identify Regions to Resequence. The PXR promoter (16 kb) and intron 1 (24.9 kb; total 41 kb) were analyzed using various bioinformatic tools to judiciously choose regions for resequencing using the selection criteria of evolutionary conservation and/or regions that contained putative clusters of transcription factor binding sites because these regions are more likely to be functionally important. The UCSC, VISTA, and ECR browsers were used to perform comparative genomic analysis between different species. Figure 1A shows a snapshot from the UCSC browser with global alignment of the human PXR promoter and intron 1 compared with genomes of other species. This track shows a measure of evolutionary conservation in 17 vertebrates, including human, chimp, mouse, rat, chicken, fugu, and zebrafish based on a phylogenetic hidden Markov model, phastCons (Siepel et al., 2005). The taller the shaded peaks, the greater the extent of conservation between species in this region. The 400-base pair (bp) proximal promoter region showed the greatest evolutionary conservation between human and rodents. Analysis of the mouse and human PXR gene using Ensembl (http://www.ensembl.org/Homo_sapiens) (Fig. 2A) revealed that, although the genes flanking human PXR—telomeric 5′ [testis-specific AMY-1 binding protein (AAT1), also known as C3orf15] and centromeric 3′ [glycogen synthase kinase-3β (GSK3B)]—are conserved with mice, the distance between PXR and the upstream gene AAT1 was 16 kb in humans versus only 2.9 kb in mice. Although both the UCSC and ECR browsers (Fig. 1, A and B) indicate a region with a high degree of conservation in intron 1, this conserved region represents a retroposed pseudogene (prohibitin) present in human PXR intron 1 but not present in a syntenic region of mouse PXR (mouse PXR is on chromosome 6, whereas the pseudogene is on mouse chromosome 15). This result stems from the fact that both programs use global alignment to detect conserved regions, but as illustrated here, some caution must be used in the interpretation because the region of conservation may not always be syntenic.
Local gene-to-gene alignment of human and mouse PXR using rVISTA also indicated that the greatest evolutionary conservation was in the proximal promoter and intron 1 and that the intron 1 peak is absent (Fig. 1C). The Cister program was used to predict cis-element clusters by using matrices from Transfac and a hidden Markov model (Wingender et al., 2000; Frith et al., 2001) (Fig. 1D). The 15 regions of the PXR promoter and intron 1 meeting these selection criteria (Fig. 1D; Table 1) were targeted for resequencing.
SNP in the PXR Promoter and Intron 1. PXR promoter and intron 1 regions were resequenced in 24 samples from the Coriell polymorphism discovery resource (PDR24, cohort I). Eighty-nine SNPs/indels were identified (Table 2), five of them shared with the HAPMAP project. A comparison of LD between pairs of loci for PXR in Caucasian HAPMAP samples was obtained using the Marker program (http://www.gmap.net/marker) (Fig. 2B). This indicated that the PXR gene is broken down primarily into two large haplotype blocks. The tails of the two blocks intersect just upstream of the retrotransposed prohibitin pseudogene. Block one begins in the up-stream gene AAT1 and extends 3′, ending in intron 1 around SNP rs3030845. The second block picks up in intron 1 around SNP rs4566573 and extends 3′ into GSK3B (Fig. 2B; Table 2).
Effect of PXR SNPs on Basal and Rifampin-Inducible CYP3A4 Activity (Cohort IIa). SNPs with a minimum allele frequency of more than 5% and SNPs predicted to reside in a putative TF binding sequence based on in silico analysis (described later) were genotyped in DNA from primary human hepatocytes from 46 Caucasian subjects (all CYP3A5 nonexpressors) phenotyped in culture for basal and rifampicin-inducible CYP3A4 activity measured as testosterone 6β hydroxylation. PXR SNPs associated with CYP3A4 basal activity (vehicle-treated hepatocytes) (Table 3; Fig. 3). These SNPs were clustered in two regions (and different blocks): the proximal promoter region and the proximal 7 kb of intron 1 flanking exon 2. Because a number of informative SNP were in LD, a single SNP from each “LD” group was randomly selected as the “LD TAG SNP” to represent the linked SNP associated with CYP3A4. SNP 63396C>T was chosen as an intron 1 LD TAG SNP [it was in LD with intron 1 SNP: 63704G>A and a polymorphic repeat at 63813 (CAAA)5/6(CA)12/13; and in partial LD with intron 1 SNP 63877T>C and 66034C>T]. In the combined population, persons homozygous for the 63396C had 3-fold lower CYP3A4 activity (0.03 ± 0.02 pmol/min/mg protein) compared with persons heterozygous or homozygous for the 63396T allele (0.09 ± 0.05 pmol/min/mg protein, p = 0.006) (Table 3; Fig. 3A).
Because CYP3A4 is more highly expressed in female versus male livers, and because we have previously shown that for genes showing sexual dimorphism it is important to stratify by gender before looking for genotype/phenotype associations (Lamba et al., 2003), we next segregated the livers by sex and reran the analysis. Basal CYP3A4 activity was related to several linked SNPs, including the 44477T>C located at -1359 bp in the PXR promoter (Fig. 3C). Thus, SNP 44477 (-1359) was selected as the promoter LD tag SNP [in LD with a 6-bp deletion in intron 1b (SNP 46370) (Uno et al., 2004) and a 46551C nucleotide insertion in intron 1a]. Among females homozygous for the 44477T allele, testosterone 6β hydroxylase activity was ∼3-fold higher (0.18 ± 0.06 pmol/min/mg protein) compared with persons homozygous for 44477C (0.06 ± 0.04 pmol/min/mg protein; p = 0.037). Although the 63396 SNP was significantly associated with CYP3A4 activity in the combined population, it was not significantly associated with CYP3A4 in females. This may be because of the lower number of females homozygous for the 63396C allele (2/17) compared with males (4/29). Females homozygous (TT) for the intron 1b 58188T>C SNP had significantly lower (2.3-fold) basal CYP3A4 activity versus 58188 TC subjects (0.08 ± 0.04 versus 0.19 ± 0.05 pmol/min/mg protein; p = 0.026) (Fig. 3E; Table 3).
PXR SNPs associated with rifampin-induced CYP3A4 activity: on treatment with rifampin there was 6-fold variation in induction of CYP3A4 activity. The majority of samples with low CYP3A4 basal activity showed the highest induction, and those with the highest basal activity showed lower induction. Thus, the genotypes associated with lower basal activity were more likely to be associated with higher induction phenotypes and vice versa. Individuals homozygous for the LD tag SNP 63396C showed a 16.56 ± 3.1-fold induction in CYP3A4 activity versus 9.26 ± 6.1-fold induction in individuals homozygous for the major 63396T allele (Fig. 3B). Likewise, rifampicin-mediated -fold induction in CYP3A4 activity was 2 times higher in subjects with 63877CC/66034TT genotypes (17.3 ± 1.2) as compared with subjects homozygous with 63877TT/66034CC genotypes (8.8 ± 5.9). Similar effects were observed for the 68740T>A/69245C>T SNP (Table 3; Fig. 3F).
Combinatorial Effect of PXR SNPs (Cohort IIa). We attempted to infer PXR haplotypes using the PHASE 2.02 program because single variants might miss the combinatorial effect of SNPs in a haplotype (Stephens et al., 2001). However, the number of PXR haplotypes was too large to allow any analysis for association. Nevertheless, variation in PXR might be influenced by additional SNPs that were not captured by variation at a single marker (e.g., functionally important SNPs may be masked if they are present together with other SNPs that have a greater effect). Hence, we first segregated individuals by the two most common genotypes: the promoter 44477CT and intron 1 63396CT heterozygotes. Those SNPs whose association with CYP3A4 increased in statistical significance after stratification are shown in Table 4. For example, the 69789A>G SNP that independently showed marginal association with CYP3A induction (p = 0.058) (Table 3) increased significance among 44477CT carriers (p = 0.01) (Table 4); likewise, the association of SNP 68740 with CYP3A4 induction increased from p = 0.02 to p = 0.008 (Table 3 versus Table 4) on stratification. One novel SNP associated with increased basal CYP3A4, 56348C>A, was uncovered by this approach.
Effect of PXR SNPs on the Hepatic mRNA Levels of PXR, PAR2, CYP3A4, and MDR1 (Cohort IIb). PXR genotype analysis was next performed on donor human liver tissue phenotyped for mRNA expression of PXR and PAR2 and two of the targets, CYP3A4 and ABCB1/MDR1. The 69789A>G genotype was associated with lower MDR1 mRNA (p = 0.035), the effect being more pronounced in females (p = 0.008) (Fig. 4C). Among males, the 45005 C>T (referred to as 25385C>T by Zhang et al., 2001) located at -831 bp in the PXR promoter was associated with significantly lower hepatic PXR levels (p = 0.029) (Fig. 4A). Persons homozygous for the 63396 C intron 1 tag SNP had lower PAR2 mRNA levels compared with subjects with at least one 63396T allele (p = 0.038) (Fig. 4B).
We again determined whether there were any combinatorial interactions between SNPs after stratification by common genotypes. Among 44477CT heterozygous subjects: 1) the 69789A>G SNP was significantly associated with lower hepatic expression of CYP3A4 (Fig. 5B) and PXR (Fig. 5C); 2) subjects heterozygous for the 56348C>A genotype had lower CYP3A4 mRNA compared with subjects with the 56348 CC genotype (Fig. 5A); and 3) PXR showed a trend for being higher in persons carrying the 63396T allele.
Within the 63396 CT heterozygous stratified subgroup: the 46370 6-bp deletion (in LD with 44477T>C and 46551 C ins), which is present in an HNF1 binding site in the PAR2 5′UTR, was associated with lower expression of PAR2 in subjects having at least one allele with deletion (DD+ND) as compared with subjects with no deletion (NN) (Fig. 5F), and CYP3A4 mRNA tended to be lower in persons with the 44477(-1359) T allele (Fig. 5E).
In Silico Analysis for Functional Effect of SNPs on PXR Locus. We used Transfac to determine whether any PXR SNPs, particularly those associated with any of the hepatic traits measured, were located in TF binding sites or disrupting or creating any TF sites (Fig. 6). Interestingly, a number of SNPs in LD at the 3′ end of intron 1 were located in HNF3β binding sites: 1) the 63396 LD tag SNP is located in a putative HNF3β (also known as FOXA2, fork head transcription factor FoxA2) binding site; 2) the 63813 (CAAA)5/6 (CA)12/13 gain in polymorphic repeats increased the number of HNF3β binding sites; and 3) screening the HAPMAP data for any additional SNPs that were in LD with the linked 63396/63704/63813 SNP revealed a new SNP at 65104T>C (rs 2461823) that was also predicted to gain an HNF3β binding site. The 69245C>T change (which occurs in LD with the 68740T>A that was associated with lower CYP3A4 expression in hepatocytes) disrupts a binding site for cAMP response element-binding protein (Fig. 6). The 69789A>G SNP, which was associated with lower expression of multiple PXR targets in liver, was present in an HNF4 binding site, and the A>G change is predicted to result in the loss of the binding site. The 44477T>C SNP results in the loss of one binding site each for signal transducer and activator of transcription (STAT)1, STAT3, STAT6, HELIOSA, and nuclear factor of activated T cells (Fig. 6). The 44477 SNP is in LD with the 6-bp deletion that is located in an HNF1 binding site at position 46370 and a 46551 C ins that is present in a stress response element site. Although several other PXR SNP associated with hepatic traits were located in TF binding sites (Table 2), the functional significance to PXR expression/activity remains unknown.
Discussion
Potential regulatory regions in the PXR gene were screened for sequence variations that could explain the observed variability in the hepatic expression of PXR and its targets CYP3A4 and MDR1. Although the same genes in human and rodent flank the PXR gene, surprisingly, the human PXR promoter was significantly larger, containing an additional 13 kb compared with the rodent orthologs. Although there are regions of similarity between the conserved 2.9-kb promoter of rodent PXR and the proximal promoter of human PXR, the greatest degree of conservation is within the first 400 bp of the human proximal promoter. It is possible that rodents deleted the distal PXR promoter because blast analysis failed to identify nucleotide sequences corresponding to the distal 13-kb intergenic region 5′ of human PXR in the rodent genome. The functional implications of the additional 13 kb of promoter sequence to human PXR expression remain to be determined.
PXR is composed of two haplotype blocks. Resequencing of approximately 15 kb of the PXR promoter and intron 1 in DNA from the PDR24 subset identified 89 SNPs. SNPs that were most highly associated with PXR target gene expression were clustered in three regions, with each cluster containing multiple SNPs in LD. Although for clarity of presentation we randomly designated one of the SNPs in each cluster as the LD tag SNP, the functional effect could be caused by any of the single variants or synergistic effects of the SNPs in LD.
The results from the present study, in conjunction with other works, imply that HNF3β (also known as fork head transcription factor FoxA2) in particular and other hepatic transcription factors such as HNF4 and HNF1 may influence transcriptional activity of the PXR promoter and that SNPs in the TF binding sites alter PXR expression/activity. First, we found that multiple PXR SNPs associated with CYP3A4 expression were residing in and predicted to alter HNF binding sites. The 69789A>G SNP destroyed a potential HNF4 binding site and was associated with lower MDR1, CYP3A4, and PXR mRNA expression. This is consistent with our results (Tirona et al., 2003) and those of others (Kamiya et al., 2003; Kyrmizi et al., 2006) on the important role of HNF4 in regulating hepatic PXR expression. Most striking was the fact that the cluster of SNPs tagged with the 63396C>T SNP, in LD with the 63704G>A, 63813(CAAA)n/(CA)n, and 65104T>C, was each affecting HNF3β binding sites. Primary cultured human hepatocytes with the 63396T allele had higher basal CYP3A4 and conversely lower rifampin induction. Among livers stratified for the 44477CT genotype, those with the 63396T allele had higher PXR mRNA in donor liver (Fig. 5D). The combined genotype (63396T, 63813 increased repeats, and 65104C) is predicted by TRANSFAC to increase HNF3β binding sites compared with the 63396C/63813 fewer repeats/65104T genotype. The finding that SNPs affecting PXR target expression (and PXR levels) are predicted to alter intron 1 HNF3β binding sites is intriguing in light of studies showing that the human PXR proximal promoter is activated by HNF3β (Gibson et al., 2006), that HNF3β is recruited to the mouse PXR promoter (between -167 and -193 bp) during liver development (Kyrmizi et al., 2006), and that HNF3β cross-talks with PXR to modulate its drug-induced activation of lipid metabolism (and glucose levels) in fasting mice (Nakamura et al., 2007). Because HNF3β is an hepatic regulator of bile acid and glucose homeostasis (Rausa et al., 2000), this regulation of PXR by HNF3β and cross-talk with HNF3β not only suggests that HNF3β may be a master regulator of PXR but also that PXR, in addition to its known role in bile acid biology, may have some unexpected role in modulating blood glucose levels.
The 44477T>C(-1359 bp) tag1 promoter SNP (in LD with the 6-bp deletion at 46370) defined another cluster of informative SNPs. However, the relationship of this SNP to CYP3A4 expression differed between primary hepatocytes and donor livers. Hepatocytes with the 44477T allele had higher basal CYP3A4 but lower rifampin-mediated induction of CYP3A4, whereas among donor livers with the 63396CT genotype, those with the -1359 SNP had lower CYP3A4 and PAR2 mRNA. It is possible that the intrinsic differences between the environment of cultured hepatocytes and donor livers, and hence potential differences in the transcription factor levels, differentially affect the alleles.
The functional relevance of the 63813 SNP may also be related to the nature of this “SNP”—a polymorphic tetranucleotide (CAAA) that repeats five or six times followed by a dinucleotide repeat (CA) that occurs 12 or 13 times. There are numerous reports in the literature where polymorphic repeats have been associated with variability in mRNA expression. For example, a polymorphic CA repeat in intron 1 of epidermal growth factor receptor has been associated with altered epidermal growth factor receptor expression in breast cancer (Gebhardt et al., 1999). A repeat polymorphism in the ERβ gene has been implicated in menopausal and premenstrual symptoms (Takeo et al., 2005) and impaired healing in elderly persons, predisposing to venous ulceration (Ashworth et al., 2005).
The 45005T (-831 bp) promoter SNP, located in a putative NF-κB and ISFG-2 site, was associated with lower hepatic PXR mRNA expression in male livers in this study. Another group recently reported that the same 45005 (23585)C allele was associated with increased susceptibility to inflammatory bowel disease (IBD), Crohn's disease, and ulcerative colitis (Dring et al., 2006). This result was intriguing because it was recently reported (Zhou et al., 2006) that PXR knockout mice show significant inflammation within the intestine. Coupled with the fact that PXR suppresses the inflammatory response, these results suggested that persons with lower PXR expression (the 45005T allele in livers) might have a higher inflammatory response. Consistent with this notion, PXR expression was lower in patients with IBD (Dring et al., 2006). However, paradoxically the 45005 (25385) C allele was significantly more frequent in IBD patients than the control population (Dring et al., 2006). There are several possible explanations for the discrepant results. The SNP, which affects a putative NF-κB site, could behave quite differently in the noninflammatory versus inflammatory state, and there could be tissue-specific differences between liver and colon in the functional consequence of this SNP. A PXR 3′UTR SNP (rs1054190) and an intron 5 SNP were shown to modify disease course for patients with primary sclerosing cholangitis, a hepatic disease where bile duct pathology leads to inflammation, fibrosis, and ultimately cirrhotic disease (Karlsen et al., 2007). This suggests that in addition to PXR's well known role in modifying expression of drug detoxification genes, it can significantly affect the inflammatory response in mouse intestine and liver. The relationship of these newly identified PXR SNPs and risk of these inflammatory diseases should help to strengthen the identity of which PXR SNPs are important to PXR expression/function in various tissues and disease states. Moreover, the recent finding that PXR-/- mice develop hepatic steatosis (Nakamura et al., 2007), as well as the growing body of work implying a connection between PXR, HNF3β, and glucose homeostasis, suggests that PXR may be a modulator gene in these diseases and that PXR genotyping in other disease settings is warranted.
The results from the present study indicate that there exists substantial genetic variability in potential regulatory regions of PXR, that some SNPs are associated with PXR target gene expression, and that a number of these SNPs reside in or create or destroy putative transcription factor binding sites. The fact that the identical SNPs were associated with various PXR traits, despite limitations in measuring the phenotypic traits (which themselves show intrinsic genetic variation) in two nonidentical hepatic environments, strongly suggests that they are likely functionally important and need to be tested further for their relationship to PXR traits in well controlled studies.
Acknowledgments
We thank the Hartwell Center for DNA sequencing.
Footnotes
-
J.L. and V.L. contributed equally to this work.
-
This work is supported in part by the National Institutes of Health (NIH) Grant GM60346, NIH/National Institute of General Medical Sciences Pharmacogenetics Research Network and Database (U01GM61374, http://pharmgkb.org) under Grant U01 GM61393, and the NIH P30 CA21765 Cancer Center Support Grant, and by the American Lebanese Syrian Associated Charities (ALSAC).
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.107.016600.
-
ABBREVIATIONS: NR1I2/PXR, pregnane X receptor; PAR2, proteinase-activated receptor-2; MDR1, multidrug resistance protein 1; SNP, single nucleotide polymorphism; kb, kilobase; TF, transcription factor(s); HNF, hepatic nuclear factor; CEBP, CCAAT enhancer binding protein; NF, nuclear factor; ECR, evolutionary conserved region; PCR, polymerase chain reaction; bp, base pair; AAT1, testis-specific AMY-1 binding protein; GSK3B, glycogen synthase kinase-3β; STAT, signal transducer and activator of transcription; LD, linkage disequilibrium; IBD, inflammatory bowel disease.
- Received May 11, 2007.
- Accepted October 5, 2007.
- The American Society for Pharmacology and Experimental Therapeutics