Liver Biology and Pathobiology
Human UDP-glucuronosyltransferase (UGT)1A3 enzyme conjugates chenodeoxycholic acid in the liver†
Article first published online: 20 OCT 2006
Copyright © 2006 American Association for the Study of Liver Diseases
Volume 44, Issue 5, pages 1158–1170, November 2006
How to Cite
Trottier, J., Verreault, M., Grepper, S., Monté, D., Bélanger, J., Kaeding, J., Caron, P., Inaba, T. T. and Barbier, O. (2006), Human UDP-glucuronosyltransferase (UGT)1A3 enzyme conjugates chenodeoxycholic acid in the liver. Hepatology, 44: 1158–1170. doi: 10.1002/hep.21362
Potential conflict of interest: Nothing to report.
- Issue published online: 20 OCT 2006
- Article first published online: 20 OCT 2006
- Manuscript Accepted: 5 AUG 2006
- Manuscript Received: 20 JAN 2006
- Canadian Institutes of Health Research. Grant Number: MOP 118446
- Fonds de la Recherche en Santé Québec
- Faculty of pharmacy,Laval University (FER)
- Health Research Foundation of Rx&D-CIHR
Chenodeoxycholic acid (CDCA) is a liver-formed detergent and plays an important role in the control of cholesterol homeostasis. During cholestasis, toxic bile acids (BA) accumulate in hepatocytes causing damage and consequent impairment of their function. Glucuronidation, a conjugation reaction catalyzed by UDP-glucuronosyltransferase (UGT) enzymes, is considered an important metabolic pathway for hepatic BA. This study identifies the human UGT1A3 enzyme as the major enzyme responsible for the hepatic formation of the acyl CDCA-24glucuronide (CDCA-24G). Kinetic analyses revealed that human liver and UGT1A3 catalyze the formation of CDCA-24G with similar Km values of 10.6 to 18.6 μmol/L, respectively. In addition, electrophoretic mobility shift assays and transient transfection experiments revealed that glucuronidation reduces the ability of CDCA to act as an activator of the nuclear farnesoid X-receptor (FXR). Finally, we observed that treatment of human hepatocytes with fibrates increases the expression and activity of UGT1A3, whereas CDCA has no effect. In conclusion, UGT1A3 is the main UGT enzyme for the hepatic formation of CDCA-24G and glucuronidation inhibits the ability of CDCA to act as an FXR activator. In vitro data also suggest that fibrates may favor the formation of bile acid glucuronides in cholestatic patients. (HEPATOLOGY 2006;44:1158–1170.)
Bile acids (BA), the major end-products of cholesterol metabolism, are complex physiological molecules that are essential for the solubilization, absorption, and transport of dietary lipids in the intestine.1 The hepatic conversion of cholesterol into BAs produces cholic acid and chenodeoxycholic acid (CDCA), the major BAs found in human bile.1 These primary BAs are subsequently converted into the respective secondary deoxycholic acid and lithocholic acid (LCA) in the intestine.1 CDCA and LCA are also hydroxylated into hyocholic and hyodeoxycholic (HDCA) acids by hepatic cytochrome P450.1 BAs have recently been assigned a crucial regulatory function as ligands for specific members of the nuclear receptor family.1, 2 Indeed, the primary CDCA is a powerful endogenous activator of the farnesoid X-receptor (FXR), which upon activation, regulates a subset of genes involved in the control of BA homeostasis.3–5 However, due to their detergent properties, bile acids are inherently cytotoxic and their accumulation is observed in a variety of pathophysiological conditions, including intrahepatic cholestasis.6
Cholestasis, or impaired bile flow, is one of the most common and devastating manifestation of liver diseases.7 Any functional disturbance of the bile secretory process can lead to cholestatic liver disease, which is associated to intracellular accumulation of toxic bile components and consecutive cholestatic cell damages.7, 8 The causes of intrahepatic cholestasis are as varied as genetic alterations, inflammatory disorders such as primary biliary cirrhosis (PBC) or primary sclerosing cholangitis (PSC), pregnancy [intrahepatic cholestasis of pregnancy (ICP)], infiltrative or granulomatous diseases (malignant lymphoma or sarcoïdosis), alcohol and drugs.7, 8 Cholestasis is clinically characterized by elevated plasma concentrations of biliary constituents, such as BAs and their sulfate or glucuronide conjugates.9, 10 The therapeutic approaches used to delay the development and/or to reduce the complications of cholestatic diseases include synthetic bile acids (ursodeoxycholic acid),11 or drugs such as rifampicin or fibrates, which act as activators of the nuclear receptors, pregnane X-receptor (PXR)11 and proliferator-activated receptor alpha (PPARα),12, 13 respectively.
Hepatic glucuronidation is a major inactivating pathway for a huge variety of exogenous and endogenous molecules.14 The human UDP-glucuronosyltransferase (UGT) enzymes catalyze the glucuronidation reaction, which consists in the transfer of the glucuronosyl group from uridine 5′-diphosphoglucuronic acid (UDPGA) to active endogenous and exogenous molecules with oxygen, nitrogen, sulfur or carboxyl functional groups.14 The resulting glucuronide products are more polar, water soluble, generally less toxic and more easily excreted than the substrate molecules.14 In humans, 18 proteins were characterized and categorized into two major families, UGT1 and UGT2, according to their primary amino acid sequence homology.14 Because BAs are cytotoxic, their hepatic glucuronidation is a significant step in their sequential metabolism.8 The most abundant glucuronide conjugate reported in human plasma is CDCA-glucuronide, followed by LCA-glucuronide,10, 15 the concentrations of which are respectively increased by 50-fold and 10-fold, in cholestatic patients.10 Glucuronide conjugation involves either the 3α-hydroxyl group or the 24-carboxyl group of the steroid nucleus of primary bile acids, resulting in the formation of ether-type or acyl-type glucuronides, respectively.16–18 Human UGT1A3 was reported as the major enzyme for the hepatic production of LCA-24G, while UGT2B4 and UGT2B7 play important roles in the glucuronide conjugation of hyodeoxycholic acid, a 6α-hydroxylated metabolite of LCA.19, 20
This study was aimed at identifying which of the human UGT1A and UGT2B enzyme(s) catalyze(s) the formation of CDCA glucuronide derivatives and at investigating the consequences of glucuronidation on the FXR-mediated activity of CDCA. The effects of FXR and PPARα activators on the expression of the CDCA-conjugating UGT enzymes were analyzed in primary cultures of human hepatocytes and in hepatoma HepG2 cells.
Patients and Methods
Normal and deuterated BAs were purchased from Steraloids Inc. (Newport, RI) and C/D/N Isotopes Inc. (Pointe-Claire, Canada), respectively. Fulvestrant was synthesized in the chemistry division of our laboratory as reported.21 Wy14,643, fenofibrate, UDPGA, and all aglycones were obtained from Sigma-Aldrich Inc. (St. Louis, MO) and ICN Pharmaceuticals Inc. (Québec, Canada). We obtained [14C]UDPGA (285 mCi/mmol) and [γ-32P]ATP from NEN Life Science Products (Boston, MA). UGT baculosomes were from BD Biosciences (Mississauga, Ontario, Canada). Expression vectors and antibiotics were from Invitrogen (Burlington, Ontario, Canada). The anti-UGT1A and anti-UGT2B antibodies were kindly provided by Dr. A. Bélanger (Laval University, Québec, Canada)22 and the secondary antibody against rabbit IgG was purchased from Sigma-Aldrich (St. Louis, MO).
Human Liver Tissues.
|Liver Sample||Previous Numbering, From Sumida23||Gender||Age||Cause of Death||Drug Intake|
|4||K25||Female||39||Subarachnoid hemorraghe||Dopamine, Lidocaine|
|7||K28||Male||21||Bilateral cerebral infaction||Dopamine, vasopressin|
|8||K29||Male||19||Vehicle accident||Dopamine, vasopressin, mannitol, heparin, phenylephrine, furosemide, cefazolin|
Cell Culture and Treatment.
Cryopreserved human hepatocytes were obtained from In Vitro Technologies Inc. (Baltimore, MD). Human hepatoma HepG2 cells and human embryonic kidney 293 (HEK293) cells were obtained from the American Type Culture Collection (Manassas, VA) and were grown as reported.22, 24 For RNA analyses, 350,000 hepatocytes/well were seeded in 24-well plates and cultured in InVitroGro CP medium (In Vitro Technologies) for 48 hours. Cells were then treated with vehicles (ethanol or DMSO, 0.1% v/v) or activators (30 μmol/L CDCA, 250 μmol/L fenofibrate, or Wy14,643 at the indicated concentrations) for 24 or 48 hours, as indicated. For glucuronidation assays, 35,000 human hepatocytes/well were seeded in 96-well plates treated with vehicle [dimethyl sulfoxide (DMSO)] or Wy14,643 (75 μmol/L) for 48 hours and subsequently incubated in the presence of CDCA (60 μmol/L) for 1 hour. Media were then recovered for subsequent determination of CDCA-24G. HepG2 cells (20 × 106) were treated with Wy14,643 (75 μmol/L) for 36 hours and were then harvested for subsequent purification of microsomes and glucuronidation assays.
Production of UGT-Overexpressing HEK293 Cells.
UGT1A3-HEK293 and UGT2B7-HEK293 cell lines were obtained by transfecting 2 μg of the pcDNA6v5-His-UGT1A3 R11 and pcDNA6v5-His-UGT2B7 H268 plasmids, respectively, with the ExGEN reagent according to manufacturer instructions (Invitrogen). Stable transfectants were selected in medium containing 10 μg/mL blasticidin HCl. The inducible UGT1A3-HEK293 cell line was obtained following stable transfection with 2 μg of the pVgRXR plasmid and selection with zeocin (75 μg/mL). The pVgRXR-HEK293 cells were then transfected with a pIND-Sp1/hygro-UGT1A3 R11 plasmid (2 μg), and selected in the presence of hygromycin (150 μg/mL). The inducible expression of UGT1A3 was achieved with 5 μmol/L ponasterone A. Other UGT-HEK293 cell lines were as described.22.
Isolation of Microsomal Proteins.
Microsomes from HepG2 or UGT-HEK293 cells or from human liver samples (200 mg)23 were purified as reported.22 Briefly, cells and tissues were homogenized in K2HPO4 (0.1 mol/L), KH2PO4 0.1 mol/L (pH 7.4), glycerol 20%, EDTA 1 mmol/L, dithiothreitol 1 mmol/L, 2.5 μg/mL pepstatin and 0.5 μg/mL leupeptin, using a Potter-Glas-Col type homogenizer (Glas-Col LLC, Terre Haute, IN) with a Teflon pestle at 4°C. Homogenates were centrifuged at 12,000g, 4°C for 20 minutes. Supernatants were centrifuged at 105,000g for 1 hour at 4°C. Microsome pellets were resuspended in homogenization buffer at a concentration of 5 μg/μL and were stored at −80C until glucuronidation assays or western blot analyses.
Western Blot Analyses.
For western blot analyses, 10 μg of microsomes or baculosomes and 50 μg of total proteins were size-separated by 10% SDS-polyacrylamide gels, transferred onto nitrocellulose membranes which were stained with Ponceau S solution to ensure the equal protein loading on each lane. Membranes were hybridized with the EL-93 (anti-UGT2B) or RC-71 (anti-UGT1A) antibodies (1:2,000) as previously described.22 An anti-rabbit IgG horse antibody (1:10,000) conjugated with peroxidase (Sigma-Aldrich) was used as the second antibody. Immunocomplexes were visualized on hyperfilm (Kodak Corp., Rochester, NY) and quantified by BioImage Visage 110s (Genomic Solutions Inc., Ann Arbor, MI).
In Vitro Glucuronidation Assays.
All glucuronidation assays were performed in a final volume of 100 μL of the following assay buffer: 50 mmol/L Tris-HCl (pH 7.5), 10 mmol/L MgCl2, 8.5 mmol/L saccharolactone, 10 μg/mL phosphatidylcholine, 25 μg alamethicin, 2.5 μg/mL pepstatin, 0.5 μg/mL leupeptin at 37°C. Assays were ended by adding 100 μL of methanol with 0.02% butylated hydroxytoluene.
Screening for UGT enzymes reactive with CDCA (1, 100, or 200 μmol/L) was conducted with 10 μg of baculosomes or microsomes in the assay buffer supplemented with 7.5 μmol/L [14C]UDPGA and 92.5 μmol/L unlabeled UDPGA for 5 hours at 37C. The formation of CDCA-glucuronide was analyzed by thin layer chromatography as reported.22
Time-course analyses with human liver or UGT1A3-baculosomes (10 μg) were performed for 15 to 360 minutes in the same conditions as above in the absence of [14C]UDPGA but with 2 mmol/L unlabeled UDPGA. Furthermore, the terminating ethanol solutions were added 0.5 ng [2H4]LCA-24G or [2H4]CDCA-24G used as internal standards for the quantification by liquid chromatography coupled to mass spectrometry (LC-MS/MS).
Because the formation of CDCA-24G by human liver microsomes and UGT1A3-baculosomes is linear for 1 hour, all subsequent quantitative analyses (kinetics and treated cells) were performed for this duration in the presence of 2 mmol/L unlabeled UDPGA. The formation of CDCA glucuronide was determined by LC-MS/MS. Kinetic assays were performed in the presence of CDCA concentrations ranging from 1 to 250 μmol/L. The apparent Vmax values obtained in the presence of the 2 UGT1A3 clones were normalized according to UGT protein level in the microsomal extract as determined by western blot.
The formation of glucuronide conjugates of tauro- and glyco-CDCA (200 μmol/L) was also assayed with human liver microsomes and UGT1A3-baculosomes in the presence of radiolabeled UDPGA in the same conditions as indicated above for CDCA.
Quantification of Glucuronide Conjugates by Liquid Chromatography Coupled to Mass Spectrometry.
The formation of LCA- and CDCA-24G was analyzed by LC-MS/MS.25 The formation of fulvestrant glucuronide was quantified as previously reported.21 The chromatographic system consisted in an Alliance 2690 (Waters Corp., Milford, MA) equipped with a Synergie RP Hydro 100 × 4.6 mm, 4 μm column (Phenomenex, Torrance, CA) and coupled to a triple quadrupole mass spectrometer API 3000 (Applied Biosystems-Sciex, Concord, Canada).
Determination of CDCA-24 Glucuronide in Cell Culture Media.
Cell culture media (250 μL) were added to 2 mL of 1% formic acid in water containing [2H4]CDCA-24G, and glucuronide conjugates were subsequently extracted using a Strata X 60 mg column (Phenomenex, Torrance, CA) and eluted in a 90/10 methanol/water solution. Elutes were evaporated under nitrogen and reconstituted in 100 μL of 50/50 methanol/water to be analyzed by LC-MS/MS as above.25
RNA Isolation, Reverse Transcription and Real-Time PCR (Real-Time RT-PCR).
Total RNA was isolated from treated or control cells according to the Tri-Reagent acid:phenol protocol as specified by the supplier (Molecular Research Center Inc., Cincinnati, OH). The reverse transcription reaction was performed using 200 units of Superscript II (Invitrogen) with 1 μg of total RNA at 42°C for 50 minutes. The real-time PCR reactions were performed using an ABI Prism 7000 instrument from Applied Biosystems (Foster City, CA). For each reaction, the final volume of 20 μL was composed of 10 μL of SyBr Green PCR Mix, 2 μL of each primer (200 ng), and 6 μL of an RT product diluted 30 times. Oligonucleotides and PCR conditions for 28S were as described.24 Quantitative PCR analyses of UGT1A3 mRNA levels were determined using the primers UGT1A3 Sense (5′-CCA ATT CAG ACC ACA TGA CAT TCA-3′) and UGT1A3 Antisense: (5′-AGG AAG CCA CTA TCT CAG GAA TTT G-3′). Conditions for real-time PCR were 95°C for 10 minutes, 95°C for 15 seconds and 66°C for 60 seconds for 40 cycles. For each gene, the amplification efficiency and the accuracies of ΔΔCt UGTs versus 28S were tested using 2 to 5 log of concentrations of reverse-transcribed mRNA. The specific amplification of human UGT1A3 was ensured by sequencing real-time PCR products.
Transient Transfection Experiments.
All transient transfection analyses involved 50 × 103 cells per well of 24-well plates, which were transfected using the ExGen reagent (Invitrogen, Burlington, Canada) for 2 hours and with 100 ng of the firefly luciferase reporter plasmid pGL3 containing a promoter driven by 3 copies of a FXR response element (FXRE),24 in the presence or absence of 30 ng of expression plasmids for FXR and RXR24 and 30 ng of the pRL-NULL expression vector. Cells were activated with the indicated concentration of CDCA or CDCA-24G for 16 hours. Inducible UGT1A3-HEK293 cells were plated at the same density before being pretreated or not with ponasterone A (5 μmol/L) for 24 or 48 hours. Cells were further incubated for 6 hours in the absence of ponasterone, and then transfected and CDCA-activated as above. Luciferase and renilla activities were determined as reported.24
Electrophoretic Mobility Shift Assay (EMSA).
EMSAs using nuclear extracts (5 μg) from human hepatocytes (Active Motif, Carlsbad, CA) and a FXRE radiolabeled probe (5′-GAT CTC AAG AGG TCA TTG ACC TTT TTG-3′) were performed as described,24, 26 in the presence of ethanol (1% v/v), CDCA (30 μmol/L), CDCA-24G (30 μmol/L) and/or 0.4 μg of an anti-FXR antibody (sc1204, Santa-Cruz, La Jolla, CA). Protein-DNA complexes were visualized on a BioMax MS film (Kodak Corp., Rochester, NY) and quantified by BioImage Visage 110 optical densitometric image analyzer (Genomic Solution Inc. Ann Arbor, MI).
Correlation studies of mRNA levels and enzymatic assays, as well as nonparametric Student t test used to analyze for significant differences between the experimental groups were performed using the JMP v4.0.2 program (SAS Institute Inc., Cary, NC).
Human UGT1A3 Catalyzes the Glucuronide Conjugation of CDCA.
A screening assay realized with human liver microsomes, baculosomes (UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A7, UGT1A8, UGT1A9, UGT1A10, UGT2B4, UGT2B7, UGT2B15, and UGT2B17) and microsomes (UGT2B10, UGT2B11, and UGT2B28) in the presence of [14C]UDPGA, revealed that a radiolabeled glucuronide conjugate of CDCA is formed only in human liver and UGT1A3-baculosomes (Fig. 1A). A slower migrating conjugate was also detected with human liver microsomes. However the presence of the same band in a negative control performed without CDCA revealed that it does not correspond to CDCA glucuronide (Fig. 1A). CDCA is also found in human bile and plasma as a glycine or taurine conjugate;1, 27, 28 we therefore assayed glyco-CDCA and tauro-CDCA with human liver microsomes and UGT1A3-baculosomes. However, none of these enzymatic preparations led to the formation of any glucuronide conjugates (data not shown), indicating that tauro- and glyco-conjugation abolish the glucuronidation of CDCA.
Time-course analyses using a sensitive LC-MS/MS method revealed that human liver microsomes form the two glucuronide conjugates detected in vivo: CDCA-3G and CDCA-24G (Fig. 1B),16, 17 whereas UGT1A3 baculosomes only produce CDCA-24G (Fig. 1C). Furthermore, glucuronidation of both CDCA-3 and CDCA-24G by human liver microsomes and of CDCA-24G by UGT1A3 was linear for 1 hour (Fig. 1B-C). Identical results were obtained when replacing the UGT1A3 baculosomes by microsomes from UGT1A3-HEK293 cell lines (data not shown).
Glucuronidation assays performed over 1 hour (Fig. 1D) indicated that human liver microsomes form 20-fold higher levels of CDCA-24G than CDCA-3G, and further ascertained the major role that UGT1A3 plays in CDCA-24G formation. However, with this sensitive method, the production of CDCA-24G was also detected with UGT1A1, UGT1A4, UGT1A7, UGT1A8, UGT1A9, UGT1A10, and UGT2B7 enzymes but at a much lower rate than UGT1A3 (Fig. 1D), and low amounts of CDCA-3G were formed in the presence of UGT1A4 and UGT2B7 (Fig. 1D). Assays performed with 1 or 100 μmol/L CDCA or with all UGT enzymes overexpressed in HEK293 cells showed similar pattern of glucuronidation activities (data not shown). Western blot experiments established that the UGT protein content in UGT1A3-baculosomes was lower compared to UGT1A1, UGT1A8, and UGT1A10 (data not shown), demonstrating that the high glucuronidation activity was not a consequence of higher UGT1A3 protein levels in the preparation.
Human UGT1A3 Plays a Major Role in the Hepatic Glucuronidation of CDCA.
To further ascertain the role of UGT1A3 in the hepatic glucuronidation of CDCA, this BA and 2 other UGT1A3 substrates (the anti-estrogen fulvestrant and LCA) were incubated with microsomal preparations from 8 human livers (Fig. 2). All preparations produced both CDCA-3 and CDCA-24G with variable activity values; however the latter was the dominant species in all cases (Fig. 2A). Furthermore, despite the very limited number of microsome samples, a significant correlation was found both between the formation of fulvestrant-G and CDCA-24G (Rs = .80, P = .015) and between the formation of LCA-24G and CDCA-24G (Rs = .78, P = .019) (Fig. 2B-C).
Kinetic Parameters for the Formation of CDCA-24G by UGT1A3 and Human Tissues.
Microsome preparations from 2 human liver samples (liver 1 and liver 7) and from 2 different HEK293-cell lines were used to investigate the kinetic properties of CDCA-24G formation (Fig. 3). The 2 UGT1A3 clones presented similar Km values (Fig. 3A), whereas clone 1 displayed a more than 2-fold higher Vmax(app.) value compared to clone 2, for the formation of CDCA-24G. Western blot analyses revealed that the UGT1A3 protein level was 2.9-fold higher in clone 1 compared to clone 2 (Fig. 3B), and normalization of the Vmax(app.) by the protein content yielded to similar values (Fig. 3). Consequently, the 2 clones shared similar intrinsic clearance (CLINT) values of 2,698 and 3,367 μL/hour/mg proteins for clone 1 and 2, respectively.
The liver samples 1 and 7 yielded respective Km values of 10.6 and 18.6 μmol/L (Fig. 3C), but displayed different apparent Vmax(app.), which led to a 4.6-fold higher CLINT for liver 1 than for liver 7. The similar Km values obtained with UGT1A3 and liver microsomes further support the important role of UGT1A3 for the hepatic formation of CDCA-24G.
UGT1A3-Dependent Glucuronidation of CDCA Reduces its Biological Activity.
We then investigated whether glucuronide conjugation affects the ability of CDCA to bind and activate FXR (Fig. 4). In EMSA, incubation of a radiolabeled FXRE probe with nuclear extracts from human hepatocytes allowed the formation of a strong protein-DNA complex (Fig. 4A, lane 1), the formation of which was inhibited in the presence of an anti-FXR antibody (lane 4) confirming the presence of FXR in this complex. In presence of CDCA (30 μmol/L), the FXR-DNA complex was slightly increased (lane 2), whereas in presence of CDCA-24G, FXR binding remained more similar to the vehicle control (Fig. 4A, lane 3; Fig. 4B). However, the strong DNA binding of FXR in the absence of ligand rendered the effects of CDCA and CDCA-24G not statistically significant (Fig. 4B).
To further grasp the biological consequences of CDCA glucuronidation, HEK293 cells were transfected with a FXR responding luciferase construct in the absence or presence of FXR/RXR, and/or CDCA or CDCA-24G (Fig. 4C). The FXR/RXR heterodimer slightly increased the FXRE activity, an effect that was strongly enhanced with 15 μmol/L CDCA (Fig. 4C). By contrast, an identical concentration of CDCA-24G failed to activate the FXR-dependent induction of the reporter gene (Fig. 4C).
Tauro- and glyco-CDCA are able to bind and activate FXR, but fail to transactivate the receptor in transfection assays,3 due to defects in the cellular uptake of conjugated BAs. Therefore, to fully demonstrate that glucuronidation of CDCA inhibits its FXR activation property, the FXRE reporter construct was also transfected in the UGT1A3-HEK293 cell lines (Fig. 5A–C). The relative luciferase activity of the reporter transfected with FXR/RXR expression plasmids and activated with CDCA was reduced in UGT1A3-HEK293 cells when compared to UGT2B7-HEK293 cells (used as a negative control for CDCA-24G production) (Fig. 5A). Moreover, the CDCA-dependent activation of the reporter gene in the 2 UGT1A3 clones (Fig. 5A) was inversely correlated to both UGT1A3 protein levels (Fig. 5B) and the media concentration of CDCA-24G (Fig. 5C), suggesting that intracellular glucuronidation of CDCA by UGT1A3 reduce its ability to activate FXR (Fig. 5C).
These observations were further confirmed by using a cell line with inducible-expression of the UGT1A3 enzyme (Fig. 5D,F). As expected, CDCA produced a dose-dependent activation of the luciferase activity in noninduced UGT1A3-cells (Fig. 5D). Interestingly, induction of UGT1A3 expression with ponasterone A resulted in a time-dependent reduction of the CDCA-independent luciferase activity (Fig. 5D), in an induction of UGT1A3 protein levels (Fig. 5E) and in an increased formation of CDCA-24G (Fig. 5F). All these observations indicate that glucuronidation alters the biological activity of CDCA mediated by the ligand-activated transcription factor FXR.
PPARα, but not FXR, Regulates UGT1A3 Expression and Activity.
Previous studies illustrated the positive effects that CDCA and fenofibrate exert on the expression of the HDCA-conjugating UGT2B4 enzyme.24, 29 Therefore, we investigated whether these FXR and PPARα activators also modulate UGT1A3 expression and activity in human hepatic cells (Fig. 6). As expected, both treatments induced UGT2B4 mRNA levels (used as positive control), whereas only fenofibrate was a potent inducer of UGT1A3 expression (Fig. 6B) while CDCA failed to significantly modulate the UGT1A3 gene activity (Fig. 6A). The PPARα-dependent induction of UGT1A3 was further confirmed by using the potent PPARα agonist, Wy14,643, in human hepatocytes (Fig. 6C). In hepatoma HepG2 cells, Wy14,643 also dose-dependently induced UGT1A3 mRNA levels (Fig. 6D). Finally, in both human hepatocytes (Fig. 6E) and HepG2 cells (Fig. 6F), treatment with Wy14,643 resulted in an increased formation of CDCA-24G.
This study identified UGT1A3 as the major human hepatic UGT enzyme for the formation of CDCA-24glucuronide, and provides evidence to show that glucuronidation of this bile acid strongly affects its FXR-mediated biological activity. Previous studies revealed that UGT1A3 also plays an important role in the glucuronide conjugation of carboxyl groups of other bile acids such as LCA, its 6α-hydroxylated metabolite HDCA and its C20 short-chain homologs: etianic (C20αβ) and isoetianic acids (C20ββ).19 As for CDCA, UGT1A3 also conjugates LCA and HDCA at the 24-carboxyl position,19 which indicates that the stereospecificity of this UGT isoform is conserved among BA substrates.19 A significant amount of CDCA-3G was also formed in the presence of UGT2B7 and UGT1A4. Interestingly, the same isoforms were also identified as the human enzymes forming LCA-3G,19 thus suggesting that the stereospecificity of not only UGT1A3, but of all BA glucuronidating enzymes is conserved for the conjugation of these natural detergents.
Circulating levels of CDCA increase from less than 1 μmol/L in normal patients to more than 50 μmol/L in cholestatic patients with PBC.10, 30, 31 Therefore, the affinity of UGT1A3 for this substrate (i.e., Km value) as determined here, renders physiologically relevant the effect of this enzyme and suggests that its may contribute to decrease the hepatic content of CDCA in cholestatic patients. In hepatocytes, CDCA undergoes a variety of metabolic alterations, such as hydroxylation into hyocholic acid, amidation with taurine and glycine, sulfonation and/or glucuronidation (reviewed in Trottier et al.32). The exact contribution of UGT1A3 in the in vivo inactivation of CDCA during cholestasis remains to be determined exactly. Nevertheless, the previously reported 50-fold increase in circulating CDCA-glucuronide concentrations,10 supports a significant contribution of glucuronidation in the metabolism of this BA. Consistently with the previous identification of functional single polymorphisms in the coding region of the UGT1A3 gene,33, 34 we also observed a certain interindividual variation in the hepatic formation of CDCA-24G. Taken together, these observations underline the need for a large population study, in which not only CDCA-24G, but all the metabolites of this BA would be determined in cholestatic patients compared to normal subjects. Such a study may definitively establish the exact contribution of UGT1A3 to the metabolism of CDCA.
The absence of glucuronidation of glyco- and tauro-CDCA by UGT1A3 indicates that the presence of the glycine and taurine groups blocks the glucuronide conjugation. This observation also suggests that glucuronidation, instead of being an additive conjugation mechanism for CDCA, is a competitive one. In contrast to UGT1A3, human BA-sulfotransferase enzymes may also conjugate tauro- and glyco-CDCA, as demonstrated by the presence of significant concentrations of sulfo-tauro- and sulfo-glyco-bile acids in the gallbladder (reviewed in Bernstein et al.35). The presence of both sulfate and tauro-/glyco- groups on CDCA is allowed by their respective positions on the CDCA molecule, because BAs are sulfated at the 3-hydroxyposition or 7-hydroxyposition and amidated at their 24 carboxyl group.36, 37 By contrast, the absence of UGT1A3 conjugating activity with these substrates could have been anticipated since the 24 carboxyl group also corresponds to the site of glucuronidation for this enzyme. Contrary to glyco-CDCA and tauro-CDCA metabolites, which are secreted into the bile through the canalicular membrane of the polarized hepatocytes,8 CDCA-glucuronide is more efficiently excreted in the blood at the basolateral membrane of hepatocytes.7 These observations indicate that the UGT1A3-dependent glucuronidation of CDCA in replacement of glycine or taurine conjugation directs its elimination into the blood, for subsequent urinary excretion, when bile formation is compromised as observed during cholestasis.8, 38 An other major difference between amidated and glucuronidated CDCA concerns their respective ability to activate FXR. Indeed, previous reports demonstrate that tauro-CDCA and glyco-CDCA conjugates bind to and activate FXR with a similar affinity to unconjugated CDCA,4 thus indicating that these derivatives are still biologically active. By contrast, we observe here that glucuronidation reduces the ability of CDCA to enhance FXRE-mediated activation, thus supporting the idea that UGT1A3-dependent glucuronidation is an important inactivating pathway. While not being statistically significant, results from EMSA are indicative that CDCA-24G is less able to complex with FXR than CDCA. The lack of significance was mainly due to a strong ligand-independent binding of FXR to the radiolabeled probe, which resulted in a slight increase of the FXR-DNA complex. Despite that it cannot be definitively ascertained that glucuronidation reduces the ability of CDCA to bind FXR, it is tempting to speculate that glucuronidation strongly affects the amphipathic nature of the bile acid, a structural property at the basis of the specificity of FXR for BA.39 On the other hand, it is also possible that glucuronidation, without affecting the ability of CDCA to bind FXR, favors its excretion from cells thus reducing its availability to activate the receptor. Although the molecular mechanisms are still unclear, our data clearly demonstrate that CDCA glucuronidation affects its ability to modulate gene expression, which may have profound physiological consequences.
By contrast to the HDCA-conjugating UGT2B4 enzyme,24 we observe that CDCA fails to modulate UGT1A3 expression. This observation suggests that CDCA stimulates the conjugation of hydroxylated BAs, without affecting its own glucuronidation. The absence of CDCA effects on its UGT1A3-dependent glucuronidation constitutes a major difference between this primary bile acid and its secondary metabolite LCA. Indeed, in addition to being an FXR activator,4 LCA is a more specific endogenous ligand for the other BA sensor, PXR, a receptor that controls UGT1A3 expression.40, 41 When focusing on glucuronidation, all these data suggest that during cholestasis, CDCA and LCA will have complementary roles in the glucuronidation of these cholesterol metabolites. Indeed, accumulating CDCA may activate the glucuronide conjugation of 6α-hydroxylated BAs, while LCA would stimulate the glucuronidation of both primary (CDCA) and secondary (LCA) BAs.
An other key player in the control of BA homeostasis is the fibrate-activated PPARα receptor (reviewed in Eloranta et al.5). Upon ligand-activation, this transcription factor regulates the expression of genes involved in BA biosynthesis, transport and metabolism. The present characterization of UGT1A3 as a positively regulated target gene of PPARα further reinforces the importance of this receptor for the control of BA levels. Interestingly, Pineda Torra et al.42 suggested the existence of a cross-talk between the FXR and PPARα regulating pathways, since they demonstrated that PPARα expression is a target gene of FXR. Based on these observations, and in contrast to what we found, it would have been anticipated that CDCA treatment, by inducing PPARα, would also affect UGT1A3 mRNA levels. However, since in the present study, human hepatocytes and HepG2 cells experiments were performed in serum-free media, it remains possible that endogenous PPARα ligands were insufficiently abundant for activating UGT1A3 in CDCA-treated cells. Nevertheless, the observation that PPARα activators induce the expression and activity of UGT1A3 suggest that fibrates may constitute efficient molecules to stimulate BA elimination. Interestingly, fibrates are currently investigated for beneficial effects in the treatment of PBC, and recent studies demonstrated that these drugs are significantly effective in treating patients with asymptomatic43 and symptomatic PBC.12, 13, 44 In such patients, the use of fibrates improves liver function by resulting in a significant reduction of serum levels of alkaline phosphatase, γ-glutamyl transpeptidase, total cholesterol and immunoglobulin M.12, 13, 44
In conclusion, our findings thus suggest that fibrates may be effective in favoring the elimination of BA under the form of glucuronide conjugates in cholestatic livers. However, the in vitro data presented here will have to be validated in vivo by measuring the effects of fibrates treatment in humans. More generally, the results from this study suggest that any drug that stimulates UGT1A3 gene expression may be effective to stimulate BA glucuronidation in cholestatic patients.
Alain Bélanger (Québec, Canada) is acknowledged for providing UGT cDNAs and anti-UGT antibodies. We thank Chantal Guillemette (Québec, Canada) and Virginie Bocher for critical reading and helpful discussions on this work.
- 13Fenofibrate treatment in patients with primary biliary cirrhosis. Am J Gastroenterol 2002; 97: 2147–2149., , , .Direct Link: