Abstract
The UDP-glucuronosyltransferase (UGT) 1A genes in humans have been shown to be differentially regulated in a tissue-specific fashion. Transgenic mice carrying the human UGT1 locus (Tg-UGT1) were recently created, demonstrating that expression of the nine UGT1A genes closely resembles the patterns of expression observed in human tissues. In the present study, UGT1A1, UGT1A3, UGT1A4, and UGT1A6 have been identified as targets of the peroxisome proliferator-activated receptor (PPAR) α in human hepatocytes and Tg-UGT1 mice. Oral administration of the PPARα agonist 4-chloro-6-(2,3-xylidino)-2-pyrimidinylthioacetic acid (pirinixic acid, WY-14643) to Tg-UGT1 mice led to induction of these proteins in either the liver, gastrointestinal tract, or kidney. The levels of induced UGT1A3 gene transcripts in liver and UGT1A4 protein in small intestine correlated with induced lamotrigine glucuronidation activity in these tissues. With UGT1A3 previously identified as the major human enzyme involved in human C24-glucuronidation of lithocholic acid (LCA), the dramatic induction of liver UGT1A3 RNA in Tg-UGT1 mice was consistent with the formation of LCA-24G in plasma. Furthermore, PPAR-responsive elements (PPREs) were identified flanking the UGT1A1, UGT1A3, and UGT1A6 genes by a combination of site-directed mutagenesis, specific binding to PPARα and retinoic acid X receptor α, and functional response of the concatenated PPREs in HepG2 cells overexpressing PPARα. In conclusion, these results suggest that oral fibrate treatment in humans will induce the UGT1A family of proteins in the gastrointestinal tract and liver, influencing bile acid glucuronidation and first-pass metabolism of other drugs that are taken concurrently with hypolipidemic therapy.
The catalytic reaction that utilizes UDP-glucuronic acid as a co-substrate for the formation of glucuronides from substrates such as steroids, bile acids, bilirubin, fatty acids, hormones, dietary constituents, and thousands of xenobiotics that include drugs, environmental toxicants, and carcinogens has evolved as a highly specialized function in higher organisms (Tukey and Strassburg, 2000). There exists in humans the superfamily of UDP-glucuronosyltransferases (UGTs), which is encoded by the UGT1 and UGT2 gene families (Mackenzie et al., 2005). The UGT1 locus is unique in that it encodes nine functional UGTs on a 220-kilobase stretch of DNA on chromosome 2 that has been shown to be regulated in humans in a strict tissue-specific pattern (Tukey and Strassburg, 2000; Gregory et al., 2004). The organization of the UGT1 locus allows for each of the UGT1A genes to be regulated uniquely on the basis of tissue as well as the ability to undergo induction, with no two tissues expressing the same complement of the nine UGT1A proteins.
We have cloned the entire UGT1 locus from a bacterial artificial chromosome, and, using this DNA, we created transgenic mice (Tg-UGT1) that express each of the nine UGT1A genes (Chen et al., 2005). The UGT1A genes in Tg-UGT1 mice have been shown to be regulated in a tissue-specific fashion that is similar to the regulatory patterns observed in humans (Tukey and Strassburg, 2000). Preliminary findings have shown that treatment of Tg-UGT1 mice with agents that activate the Ah receptor or the pregnane X receptor (PXR) leads to the induction of one or all of the UGT1A gene transcripts, depending on the tissue (Chen et al., 2005). These animals were also used to demonstrate that induction of UGT1A3 by the liver X receptor (LXR) correlated with increased glucuronidation levels of plasma primary and secondary bile acids (Verreault et al., 2006). Thus, these mice offer an opportunity to examine the regulatory or expression patterns of the human UGT1A genes in response to environmental or drug-induced exposure.
Fibrate hypolipidemic drugs are used worldwide in the treatment of dyslipidemia and are synthetic ligands for the peroxisome proliferator-activated receptor (PPAR) α. Along with other members of the PPAR family of nuclear receptors, γ and δ, PPARα activation is linked to regulation of lipid metabolism (Duval et al., 2004). After activation, PPARα heterodimerizes with the 9-cis-retinoic acid receptor (RXR) to bind to PPAR-responsive elements in DNA, leading to transcriptional activation of target genes. PPARα is widely expressed in liver, muscle, kidney, and intestine (Auboeuf et al., 1997) and mediates expression of genes involved in fatty acid β-oxidation (Desvergne and Wahli, 1999). In cell culture studies, activation of PPARα regulates UGT1A9 (Barbier et al., 2003b) as well as UGT2B4 expression (Barbier et al., 2003a). Thus, the UGT genes seem to be additional targets for PPARα. It is noteworthy that several of the human UGT1A proteins including UGT1A9 and UGT2B4 were shown to metabolize polyunsaturated fatty acids (Turgeon et al., 2003), providing evidence that the family of UGTs regulated by activated PPARα plays an important role in lipid metabolism. Because several of the UGT genes that are targeted for regulation by PPARα have been implicated in fatty acid metabolism, confirming the identity of the UGTs that are subject to regulation by PPARα in vivo will advance an understanding of the role of glucuronidation in dyslipidemia.
In the following study, we examined the influence of the fibrate and PPARα agonist WY-14643 on its control of the UGT1 locus in Tg-UGT1 mice. Significant regulation of UGT1A proteins was observed in the gastrointestinal tract and liver. The outcome of these results might predict that treatment with fibrates for clinical reasons will result in considerable up-regulation of the UGTs in the gastrointestinal tract, accelerating first-pass metabolism and affecting oral bioavailability of other drugs or nutrients that are consumed while patients are taking fibrate medication.
Materials and Methods
Reagents. Pirinixic acid (4-chloro-6-(2,3-xylidino)-2-pyrimidinylthioacetic acid), commonly known as WY-14643, and DMSO were purchased from Sigma-Aldrich (St. Louis, MO). An antibody to rabbit anti-human UGT1A (Albert et al., 1999) was kindly provided by Dr. Alain Belanger from Laval University, Quebec, Quebec, Canada. A rat CYP4A1 antibody originally generated as described previously (Okita and Okita, 1992) was donated to us by Dr. Eric Johnson from the Scripps Research Institute, La Jolla, CA. The rat CYP4A1 is most homologous to mouse Cyp4a10 (Okita and Okita, 2001), but in experiments outlined here the mouse protein(s) identified with the rat CYP4A1 antibody are referred to as Cyp4a. Human HepG2 cells that express a stably integrated PPARα cDNA were provided by Dr. Eric Johnson.
Animal Studies with the PPARα Ligand WY-14643. Animal studies were performed in compliance with the National Institutes of Health guidelines regarding the use of laboratory animals. Animals were provided with food and water ad libitum and maintained under controlled temperature (23°C) and lighting (12-h light/12-h dark cycles). Mice were pooled in groups of three for each treatment. WY-14643 was dissolved in DMSO-corn oil (50:50) at a concentration of 8 mg/ml. Wild-type (WT) and Tg-UGT1 mice were administered 100 μl of WY-14643 (40 mg/kg) by the oral route once a day for 3 days. Mice were sacrificed 24 h after the last dose, and the organs were collected and stored at –80°C until RNA and microsomal protein could be prepared. For plasma 24G-LCA analysis, mice were anesthetized by isoflurane inhalation 24 h after the last dose of WY-14643, and blood was collected by cardiac puncture into heparinized tubes. The formation of glucuronide conjugates was quantified by liquid chromatography/electrospray ionization-tandem mass spectrometry as described previously (Verreault et al., 2006)
Isolation and Treatment of Primary Hepatocytes. Human hepatocytes were purchased from InVitro Technologies (Baltimore, MD), seeded in 24-well collagen I-coated plates (VWR, West Chester, PA), and cultured in InVitro Gro CP medium for 48 h. The cells were then treated with either vehicle (DMSO) or Wy-14643 (75 μM) for 48 h. The isolation of primary hepatocytes from Tg-UGT1 mice has been described previously (Chen et al., 2005). After plating of hepatocytes on six-well collagen I-coated plates, the cells were allowed to attach for 3 h. The hepatocytes were then treated for 48 h with WY-14643. The cells were replenished with fresh media and WY-14643 after 24 h. For analysis of expressed UGT1A1 by Western blot analysis, hepatocytes were collected and lysed in buffer containing 0.05 M Tris-HCl, pH 7.4, 0.15 M NaCl, 0.25% deoxycholic acid, 1% Nonidet P-40, and a complement of protease and phosphatase inhibitors. The solubilized lysate was centrifuged for 20 min in a refrigerated Eppendorf centrifuge at 16,000g, and the supernatant was collected for Western blot analysis.
For isolation of RNA from human and Tg-UGT1 hepatocytes, the supernatant was removed, and a 1-ml solution of phenol-guanidinium isothiocyanate solution (TRIzol; Invitrogen, Carlsbad, CA) was added directly to the wells. After the solution was allowed to set for 30 min, it was removed and a Folsh separation was established after the addition of 200 μl of chloroform. The sample was briefly vortexed, and after a short pulse in an Eppendorf centrifuge the water phase was removed. The RNA was collected after the addition of 2 volumes of isopropyl alcohol and centrifugation at 16,000g in an Eppendorf centrifuge.
Isolation of Microsomes from Mouse Tissues. Combining three animals per group, the liver, small and large intestines, kidney, heart, and skeletal muscle were collected from treated and untreated WT and Tg-UGT1 mice. The intestinal tract tissues were dissected lengthwise and rinsed in 1.15% KCl before freezing on dry ice. The pooled tissues from each treatment group were pulverized under liquid nitrogen in a porcelain mortar. A sample of the pulverized tissue was homogenized in 5 volumes of chilled 1.15% KCl using a motorized glass-Teflon homogenizer. The homogenate was centrifuged at 2,000g for 10 min at 4°C, and the supernatant was further centrifuged at 9,000g for 10 min at 4°C. The resulting supernatant centrifuged at 100,000g for 60 min at 4°C in a floor-model Beckman SW-40Ti rotor. The pellet was resuspended in buffer (50 mM Tris-HCl, pH 7.4, 10 mM MgCl2, and 1 mM phenylmethylsulfonyl fluoride), and the protein concentration was determined by the Bradford method before storage at –80°C. For lamotrigine glucuronidation activity using microsomes, the method described previously was used (Chen et al., 2005).
Western Blot Procedures. All Western blots were performed using 4 to 12% NuPAGE bis-Tris-polyacrylamide gels as outlined by the supplier (Invitrogen). Twenty micrograms of microsomal protein was heated at 70°C for 10 min in loading buffer and resolved under denaturing conditions (50 mM MOPS, 50 mM Tris base, pH 7.7, 0.1% SDS, and 1 mM EDTA) before the proteins were transferred onto a nitrocellulose membrane using a semidry transfer system (Novex, San Diego, CA). The membrane was blocked with 5% nonfat dry milk in washing buffer (10 mM Tris-HCl pH 7.4, 0.15 M NaCl, and 0.05% Tween 20) for 1 h at room temperature, followed by overnight incubation at 4°C with primary antibodies (1:5000) prepared in washing buffer containing 0.02% sodium azide and 5% BSA. An anti-UGT1 common carboxyl terminus antibody made from an expressed fusion protein encoding amino acids 312 to 531 was produced in rabbits (Albert et al., 1999). The anti-human UGT1A1, UGT1A4, and UGT1A6 have been described previously (Ritter et al., 1999; Chen et al., 2005). Membranes were washed and exposed to horseradish peroxidase-conjugated secondary antibodies (Cell Signaling Technology Inc., Danvers, MA). After additional washes in washing buffer, the conjugated horseradish peroxidase was detected using the ECL plus Western blot detection system (PerkinElmer Life and Analytical Sciences, Boston, MA) and detected after exposure to X-ray film.
Measurements ofUGT1AGene Expression by Real-Time PCR. For RNA analysis, tissues from three animals were pooled, frozen in liquid nitrogen, and then pulverized to a fine powder. A sample of the mixture was added to 1 ml of acidic phenol-guanidinium isothiocyanate solution (TRIzol), and total RNA was extracted as described previously (Chen et al., 2005). For each reverse transcription (RT) reaction, 2 μg of RNA was first denatured by heating, and cDNA was synthesized in 20 μl using the Omniscript RT kit according to the manufacturer's instructions (QIAGEN, Valencia, CA). Each real-time PCR reaction was performed using a Mx4000 Multiplex quantitative PCR system (Stratagene, La Jolla, CA). For each reaction, the final volume was 20 μl and was composed of 10 μl of SYBR Green PCR mix, 2 μl of each primer (200 nM final), and 2 μl of each RT reaction. Each reaction was run in triplicate. Quantitative PCR analysis for UGT1A1, UGT1A3, UGT1A4, UGT1A6, and UGT1A9 mRNA levels were determined with the sense and antisense primers listed in Table 1. For UGT1A1 and UGT1A6, each PCR reaction was carried out at 95°C for 10 min, 95°C for 30 s, and 63°C for 60 s, followed by 72°C for 45 s, processed through 40 cycles. Amplification was followed by DNA melt at 95°C for 1 min and a 41-cycle dissociation curve starting at 55°C and ramping 1° every 30 s. For UGT1A3, each PCR reaction was carried out at 95°C for 10 min, 95°C for 30 s, and 58°C for 60 s, followed by 72°C for 45 s, processed through 40 cycles. To determine the quantitation of gene expression, the comparative threshold cycle method (ΔCT) was used (Livak and Schmittgen, 2001). To control for variations in mRNA quantity and quality, β-actin RNA was used as an internal control to calculate a relative Ct for the target molecules. Subtracting the Ct of the housekeeping gene β-actin from the Ct of the UGT gene of interest yields the ΔCt.
EMSAs. These experiments were carried out as described previously (Barbier et al., 2003b) by examining the ability of PPARα and RXRα to associate with putative DR1 binding sequences on the human UGT1A3 and UGT1A6 genes. The UGT1A3 DR1 oligonucleotides –5790 (5′-ggtaAGGTCACAGATCAacag-3′), the UGT1A6 promoter oligonucleotides –2680 (5′-ttatGGGTCGTAGGTCAtac-3′), and the UGT1A9 PPRE (5′-gacaTCACCTCTGACCTcaaggag-3′) were end-labeled with [γ-32P]ATP using T4-polynucleotide kinase, followed by incubation with PPARα and/or RXRα synthesized in vitro using the TnT Quick-Coupled Transcription/Translation System (Promega, Madison, WI). For competition experiments showing specificity, various concentrations of cold UGT1A9 DR1 sequence were included in the binding assays. The protein-DNA complexes were resolved by 4% nondenaturing polyacrylamide gel electrophoresis in Tris-borate-EDTA buffer.
UGT1A1Promoter Cloning and Expression. Enhancer DNA fragments have been described previously (Yueh et al., 2003). All experiments were conducted in triplicate using HepG2 cells and HepG2 cells that carry an integrated PPARα cDNA (PPARα-HepG2) (Hsu et al., 2001). For analysis of functional DR1 elements, oligonucleotides encoding three copies of the UGT1A1 (5′-gccaAGGGTAGAGTTCAgtgt-3′), UGT1A3 (5′-ggtaAGGTCACAGATCAacag-3′), and UGT1A6 (5′-ttatGGGTCGTAGGTCAtac-3′) DR1 sequences were synthesized and cloned into the SV40pGL3 expression vector. HepG2 and PPARα-HepG2 cells were cultured to ∼80% confluence in 12-well plates and transfected with 200 ng of reporter plasmid along with 5 ng of phRL-SV40 using LipofectAMINE 2000 reagent according to the manufacturer's protocol (Invitrogen). After 24 h, the cells were treated for 48 h with 100 μM WY-14643 dissolved in DMSO. Dual luciferase assays were conducted according to the manufacturer's instructions (Promega). While the cells were still attached, 200 μl of Passive Lysis Buffer (Promega) was added, and 20 μl of lysate was used for dual luciferase analysis using an LMax II384 luminometer. Firefly luciferase values were normalized to Renilla reniformis luciferase and protein concentration and reported as -fold increase over vehicle-treated cells.
Results
Induction of UGT1A Proteins by the PPARα Activator WY-14643 in Primary Hepatocytes. The treatment of human primary hepatocytes with WY-14643 (75 μM) led to the induction of UGT1A1, UGT1A3, UGT1A4, and UGT1A6 RNA, with the largest response being observed for UGT1A1 and UGT1A3 (Fig. 1). In comparison, when primary liver hepatocytes isolated from Tg-UGT1 mice were cultured in the presence of WY-14643 and the UGT1A gene transcripts were analyzed by real-time RT-PCR, the UGT1A1 gene transcript was found to be significantly induced (Fig. 1). A 2.5-fold induction of UGT1A6 was also observed. However, the UGT1A3 and UGT1A4 genes were not responsive to WY-14643 in hepatocytes from Tg-UGT1 mice. This result might reflect a species difference in control of these genes or possibly could be attributed to differences in the optimized tissue culture conditions used with the human and transgenic hepatocytes. Regardless, there seems to be consistency in the induction of UGT1A1 and UGT1A6.
Oral Administration of WY-14643 Induces Liver and Gastrointestinal Tract UGT1A Proteins in Tg-UGT1Mice. To examine the potential of activated PPARα to induce human UGT1A gene expression in vivo, Tg-UGT1 mice were treated with WY-14643 by oral gavage (40 mg/kg) every 24 h for 3 consecutive days. Mice were sacrificed 24 h after the last dose, and the tissues (three per treatment group) were collected and pooled. After pulverization of the tissue under liquid nitrogen, microsomal preparations were made from liver, small intestine, large intestine, and kidney. A sample of protein from each tissue was processed for Western blot analysis, and UGT1A protein was detected with an anti-UGT1A antibody that was produced to the common carboxyl terminus region of the human UGT1A proteins (Albert et al., 1999). Results from these experiments demonstrated that oral treatment with WY-14643 induced the human UGT1A family of proteins in the liver, small intestine, large intestine, and kidney (Fig. 2). Induction of Cyp4a protein, which is induced after activation of PPARα (Muerhoff et al., 1992), was demonstrated in the liver, small intestine, and kidney, but there was no detectable Cyp4a in the large intestine where we observed considerable UGT1A protein induction.
With the use of antibodies specific for human UGT1A1, UGT1A4, and UGT1A6 (Chen et al., 2005), oral WY-14643 treatment to Tg-UGT1 mice resulted in prominent induction of UGT1A1 and UGT1A6 in liver microsomes (Fig. 3). A very mild but observable induction of UGT1A4 was also evident. In human (Strassburg et al., 1999) and Tg-UGT1 liver (Chen et al., 2005), basal levels of UGT1A1, UGT1A3, UGT1A4, UGT1A6, and UGT1A9 RNA were detected. When we carried out real-time RT-PCR with specific primers to each of the human transcripts using liver RNA from male WY-14643-treated Tg-UGT1 mice, very prominent induction of UGT1A1 (500-fold) and UGT1A3 (120-fold) was observed (Fig. 4). The induction of UGT1A1 RNA corresponded to induction of protein (Fig. 3). Measurable inductions of UGT1A4 (3-fold) and UGT1A6 (4 fold) gene transcripts were also observed, and increases in these proteins were demonstrated. Although a specific UGT1A3 antibody is not available to examine protein expression, the 120-fold induction of UGT1A3 RNA by WY-14643 would lead us to speculate that UGT1A3 is highly expressed in liver after fibrate treatment.
When we examined extrahepatic microsomal preparations by Western blot analysis, UGT1A1 and UGT1A4 were induced in the small intestine, with no observable induction of UGT1A6 (Fig. 3). Small amounts of induced UGT1A1, UGT1A4, and UGT1A6 are seen in large intestine. In kidney, a tissue that contains substantial concentrations of PPARα, induction of only UGT1A6 was observed. There was no induction of these UGTs in other tissues such as heart or muscle. The differential response in induction of the UGT1A genes by WY-14643 most likely results from coordination of tissue-specific regulatory factors with PPARα.
It is interesting to note that the induction pattern of UGT1A6 in the different tissues closely resembles induction of the mouse Cyp4a protein (Fig. 3). This result indicates that PPARα is working in concert with additional tissue-specific transcriptional mechanisms to coordinately regulate these two genes. In addition, whereas induction of gene expression by activated PPARα is dependent upon tissue abundance of the receptor, the observations that UGT1A1, UGT1A4, and UGT1A6 are differentially induced clearly indicates that receptor abundance is not the sole factor dictating expression of target genes in the respective tissues.
Role of Induced UGT1A3 and UGT1A4 in Lamotrigine Metabolism. The formation of quaternary ammonium-linked glucuronides results from the attachment of glucuronic acid to aliphatic tertiary amino groups. In human liver microsomes, the kinetics of the reaction forming quaternary ammonium glucuronides was shown to be biphasic, displaying a high and low Km profile (Breyer-Pfaff et al., 1997). The enzyme responsible for the low Km formation was shown to be UGT1A4, whereas UGT1A3 exhibited low efficiency and contributed to the high Km (Green and Tephly, 1998; Green et al., 1998). It is noteworthy that the formation of N-linked quaternary ammonium glucuronides does not proceed efficiently in rodent models such as rats (Chiu and Huskey, 1998), so Tg-UGT1 mice may serve as a humanized animal model to examine the contribution of UGT1A expression toward detoxification of xenobiotics containing a tertiary amine moiety.
With the prominent induction of UGT1A4 protein in the gastrointestinal tract after WY-14643 treatment, we sought to examine the contribution of lamotrigine glucuronidation in liver and small intestinal microsomes. Lamotrigine is used therapeutically in humans as an antiepileptic drug, and forms a quaternary ammonium glucuronide, which is excreted in the urine (Remmel and Sinz, 1991). For these studies, Tg-UGT1 and WT mice (n = 3 per group) were treated orally with either vehicle or WY-14643 (40 mg/kg) for 3 days followed by the preparation of total RNA for real-time PCR quantification as well as microsomal preparation for analysis of lamotrigine glucuronidation.
As shown in Table 2, lamotrigine glucuronidation was detectable and induced in microsomes from WT mice, but the level of induced activity was 16-fold lower than that detected in untreated Tg-UGT1 mouse liver microsomes. Basal levels of either UGT1A3 or UGT1A4 are responsible for the elevated lamotrigine glucuronidation in Tg-UGT1 liver (Chen et al., 2005). In liver and small intestinal microsomes from male Tg-UGT1 mice, the basal levels of lamotrigine glucuronidation were comparable. After the oral administration of WY-14643, the rates of lamotrigine glucuronidation were dramatically induced in liver (10.4-fold). In light of Western blot findings that induction of UGT1A4 in liver was minimal (Fig. 3), the induction of lamotrigine glucuronidation must be dictated by induction of UGT1A3, as determined by real-time RT-PCR (Fig. 4). In small intestinal microsomes, WY-14643 elicited an 11.6-fold induction of lamotrigine glucuronidation activity. Because UGT1A4 is induced in this tissue as a result of WY-14643 treatment (Fig. 3), it can be speculated that the abundance of UGT1A4 underlies the majority of the induced catalytic activity in the gastrointestinal tract.
Induced UGT1A3 and Bile Acid Glucuronidation. During enterohepatic circulation, bile acids (BAs) enter the liver and are targets for glucuronidation (Pauli-Magnus et al., 2005). In humans, UGT1A3 has previously been demonstrated to be the primary enzyme underlying the hepatic C24-glucuronidation of LCA (Gall et al., 1999). In Tg-UGT1 mice, activation of the LXR leads to induction of UGT1A3, which is followed by the accumulation of LCA-24G in blood (Verreault et al., 2006). In agreement with our studies, recent findings have demonstrated that WY-14643 treatment of human hepatocytes induces UGT1A3 mRNA (Trottier et al., 2006). To evaluate whether the induction of UGT1A3 by PPARα activation in Tg-UGT1 mice can be correlated with increases in serum LCA-24G levels, the effects of WY-14643 treatment on LCA glucuronidation were evaluated in WT and Tg-UGT1 mice. Although there were detectable levels of LCA-24G in several WT and DMSO-treated Tg-UGT1 mice (Fig. 5), treatment of Tg-UGT1 mice with WY-14643 strongly induced levels of circulating LCA-24G over those in the other treatment groups. These experiments reinforce previous findings that up-regulation of UGT1A3 will have a significant impact on BA glucuronidation in vivo.
Identification of PPARα-Responsive Regions within theUGT1A1,UGT1A3, andUGT1A6Promoters. The significant induction of UGT1A1, UGT1A3, and UGT1A6 by WY-14643 in Tg-UGT1 liver suggests that these genes are targets of activated PPARα, and that the PPARα/RXRα transcriptional complex facilitates this induction by binding to PPREs on these genes. The PPREs that bind PPARα contain imperfect direct repeats separated by a single base (DR1). In scanning these genes for consensus DR1-like elements, EMSAs were performed using these response elements as probes to determine whether PPARα and RXRα were capable of binding to the PPREs. This approach resulted in the identification of functional DR1 elements located at –5790 on the UGT1A3 gene and –2692 on the UGT1A6 gene (Fig. 6). Both of these sequences were unable to bind expressed PPARα, but when incubated with PPARα and RXRα, their binding elicited an induced gel shift. In addition, binding of PPARα/RXRα to the UGT1A3 and UGT1A6 PPREs was inhibited when the binding reaction included wild-type UGT1A9 DR1 (Barbier et al., 2003b) but not when the incubations included oligonucleotide toward the mutated UGT1A9 DR1 sequence. Functional expression of the DR1 sequences was confirmed when they were concatenated and cloned into expression plasmids and shown to elicit WY-14643-induced promoter activity in HepG2 cells that overexpress PPARα.
To identify the potential DR1 sequence on the UGT1A1 gene, a UGT1A1 promoter construct (–3712 to –7) was cloned into the expression plasmid pGL3-basic and transfected into either HepG2 cells or PPARα-HepG2 cells. When these cells were treated with WY-14643, HepG2 cells showed no induction of luciferase activity (Fig. 7). In contrast, PPARα-HepG2 cells displayed a 2-fold induction of luciferase activity when treated with WY-14643 in comparison to vehicle (DMSO)-treated cells.
To localize the responsive element, a series of enhancer sequences were generated and transfected into the PPARα-HepG2 cells followed by treatment with WY-14643. WY-14643 treatment led to induced luciferase activity with those enhancer sequences (clones E1 and E3) that included the PXR, NR1, XRE, and putative DR1 binding regions (Fig. 8). This region has been identified as the phenobarbital-responsive enhancer module (PREM). When the PREM region was deleted (clone E2), no induction of luciferase activity was noted. DNA sequence comparisons to other DR1 elements demonstrated that the consensus recognition DR1 sequence identified on the UGT1A1 gene (Fig. 8) was identical to the CYP4A1 DR1 motif (Palmer et al., 1995). To further regionalize the binding region, a construct was made (clone E4, bases –3430 to –3272) (Figs. 6 and 7) that contained the PXR, NR1, and XRE sequences but with the identified DR1 element mutated by removing the 3′ region of the consensus binding region. After WY-14643 treatment of transfected HepG2-PPARα cells with E4 (Fig. 7), there was no increase in luciferase activity over DMSO-treated cells. Similar results were observed when the DR1 sequence was mutated in clone E3 (E3mt). Further evidence that the DR1 sequence provides functional activity in the presence of PPARα was confirmed by cloning the DR1 sequence into an expression plasmid and demonstrating that WY-14643 was capable of inducing reporter gene activity in PPARα-HepG2 cells. Combined, the putative UGT1A1 DR1 motif was shown to be essential for induction of UGT1A1 transcription by PPARα agonists.
Discussion
One of the more complicated model systems to study is presented with the human UGT1 locus, as nine functional UGT1A genes are individually regulated in a tissue-specific and inducible fashion through a process of exon sharing (Tukey and Strassburg, 2000). Through tightly controlled tissue-specific regulation, each of the nine exon 1 elements is individually spliced to common exons 2 to 5, producing UGTs that are encoded by an amino-terminal region that is highly variable and a carboxyl region that is identical. In the present study, oral administration of the PPARα activator and fibrate WY-14643 to Tg-UGT1 mice induced UGT1A1, UGT1A3, UGT1A4, and UGT1A6 in liver. It is unclear why UGT1A9 was not induced in liver, as WY-14643 had been shown previously to induce this protein in both human primary hepatocytes and HepG2 cells (Barbier et al., 2003b). UGT1A9 was not induced in human hepatocytes in this study, indicating that separate batches of human hepatocytes may respond differently to PPARα activators. However, the significant induction of UGT1A3 and UGT1A6 RNA in Tg-UGT1 liver was attributed to functional PPREs identified in the flanking region of these genes. Mapping of the responsive element on the UGT1A1 gene demonstrated that induction was regulated by the same 290-base pair distal PREM region that controls induction through activated Ah receptor (Yueh et al., 2003), PXR, and CAR (Xie et al., 2003). This region is specific for activated PPARα because previous studies had indicated that the DR1 region played little role in either CAR or PXR activation of the UGT1A1 gene (Sugatani et al., 2005).
WY-14643 treatment also facilitated the induction of UGT1A1, UGT1A4, and UGT1A6 in either the liver, gastrointestinal tract, or kidney. In kidney, only UGT1A6 was induced by oral administration of WY-14643. Although UGT1A1, UGT1A4, and UGT1A6 are induced in liver, there is virtually no induction of UGT1A6 in response to WY-14643 treatment in the small intestine where UGT1A1 and UGT1A4 are induced. WY-14643 activation of PPARα and induction of the UGT1A gene products is clearly occurring in concert with other tissue-specific regulatory factors. Although PPARα is abundant in tissues such as the liver, small intestine, kidney, heart, and muscle (Desvergne and Wahli, 1999), activation by oral treatment with WY-14643 leads to a unique complement of induced glucuronidation activity in each tissue.
In humans, the small intestine is rich in glucuronidation activity as is evident from the detection of multiple UGT1A gene transcripts (Strassburg et al., 1999, 2000). A similar pattern of expression is observed in Tg-UGT1 mice. The predominant induction of UGT1A1 and UGT1A4 in the small intestine by WY-14643 may be indicative of the expression patterns in humans after oral exposure to drugs that are PPARα agonists, such as those encountered with therapeutic doses of hypolipidemic drugs. Overexpression of the UGTs in tissues such as the small intestine and liver would be expected to accelerate metabolism of dietary constituents and therapeutic drugs that are ingested simultaneously with PPARα agonists. But this possibility can only be suggestive, because the role of PPARα in human liver is unclear. It has been suggested that the functional capabilities of PPARα in humans may not resemble those observed in mice because the hepatic concentrations of PPARα in mice are higher than those detected in human liver (Desvergne and Wahli, 1999). However, it can be noted from these studies that WY-14643 treatment does not facilitate a dramatic induction of mouse UGT1A protein expression in small intestine or liver, leading us to conclude that the human UGT1 locus is a more responsive gene target toward activated PPARα. This difference in sensitivity may result from PPREs being conserved on the human UGT1 locus and not conserved on the mouse Ugt1 locus.
The 3- to 5-fold increase in human hepatocyte UGT1A3 mRNA levels by WY-14643 noted previously (Trottier et al., 2006) as well as in these studies are reflected by a >100-fold induction in liver UGT1A3 mRNA after oral treatment to Tg-UGT1 mice. This is significant because UGT1A3 has been identified as the primary enzyme linked to the human glucuronidation of the BAs chenodeoxycholic acid and LCA (Trottier et al., 2006; Verreault et al., 2006). During cholestasis, plasma concentrations of conjugated BAs are elevated and excreted in the urine, because normal bile flow is interrupted. Cholestasis is sometimes treated clinically by the administration of drugs that activate the family of xenobiotic receptors, such as PXR and PPARα (Dohmen et al., 2004; Oo and Neuberger, 2004). It is noteworthy that we have shown previously that treatment of Tg-UGT1 mice with the PXR activator pregnenolone-16α-carbonitrile dramatically induces UGT1A3 in both the liver and gastrointestinal tract (Chen et al., 2005). Induction of liver UGT1A3 in Tg-UGT1 mice by the oral administration of LXR activators also resulted in significant accumulation of plasma chenodeoxycholic acid-24G and LCA-24G (Verreault et al., 2006). In the present studies, induction of UGT1A3 by WY-14643 and activation of PPARα is followed by increases in plasma LCA-24G levels. Because functional interruption in normal bile acid elimination can lead to cholestatic liver disease, increases in primary and secondary bile acid glucuronidation by induced UGT1A3 may be considered an additional detoxification step in reducing the accumulation of toxic levels of hepatic BAs.
The activation of PPARα in different tissues is thought to affect insulin sensitivity, lipoprotein synthesis and metabolism, inflammation, and cholesterol flux (Li and Palinski, 2006). Thus, the use of therapeutic agents such as fibrates that modulate these events has been keenly investigated. For example, type 2 diabetes predisposes individuals to coronary artery disease, a predictor that has called for the treatment of diabetes with fibrates. Recent clinical findings have demonstrated that fenofibrate treatment led to a significantly reduced rate of progression of coronary atherosclerosis in patients with type 2 diabetes (Rubins et al., 1996). In another study, gemfibrozil therapy resulted in the reduction of major cardiovascular events in patients with coronary artery disease (Rubins et al., 1996). In this same study, those individuals with the highest tertile of body mass and triglyceride level had the most dramatic risk reduction in coronary heart disease. This clinical study also indicated that fibrate therapy would be useful in those who display one or more of the major features associated with the metabolic syndrome: obesity, high triglyceride levels, insulin resistance, or diabetes. From our studies, it might be anticipated that the gastrointestinal tract and hepatic levels of the UGT1A proteins will be significantly induced when individuals are receiving therapeutic levels of fibrate therapy. Because glucuronidation in these tissues plays an important role in facilitating the metabolism and elimination of hundreds of known therapeutic agents (Tukey and Strassburg, 2000), individuals receiving fibrate therapy may be predisposed to accelerated metabolism of those drugs that are taken concurrently with fibrate therapy.
In conclusion, transgenic mice that express the UGT1 locus serve as a sensitive model to examine the contribution of PPARα activators on UGT1A expression. These data implicate the UGT1 locus as a target in the gastrointestinal tract and liver for activated PPARα, allowing us to speculate that the Tg-UGT1 mice may serve as a biological model to study the impact of human UGT induction on drug metabolism and other physiological events.
Footnotes
-
This work was funded by National Public Health Service Grants GM49135 and ES10337.
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.106.013243.
-
ABBREVIATIONS: UGT, UDP-glucuronosyltransferase; Tg-UGT1, transgenic UGT1; PXR, pregnane X-receptor; Ah receptor, aryl hydrocarbon receptor; LXR, liver X receptor; PPAR, peroxisome proliferator-activated receptor; RXR, retinoic acid X receptor; WY-14643, 4-chloro-6-(2,3-xylidino)-2-pyrimidinylthioacetic acid, pirinixic acid; DMSO, dimethyl sulfoxide; WT, wild-type; LCA, lithocholic acid; PCR, polymerase chain reaction; RT, reverse transcription; EMSA, electrophoretic mobility shift assay; DR1, direct repeat response element that is separated by 1 nucleotide; PPRE, PPAR-responsive element; SV40, Simian virus 40; BA, bile acid; PREM, phenobarbital-responsive enhancer module; CAR, constitutive androstane receptor; MOPS, 4-morpholinepropanesulfonic acid.
- Received October 9, 2006.
- Accepted December 5, 2006.
- The American Society for Pharmacology and Experimental Therapeutics