Abstract
Gunn rats glucuronidate acetaminophen (APAP) at reduced rates and show increased susceptibility to APAP-induced hepatotoxicity. This defect is presumed to involve UDP-glucuronosyltransferase (UGT) 1A6, which is nonfunctional in Gunn rats, but it is currently unclear whether other 1A family members are also involved. In humans, two 1A isoforms are known to be active (1A6 and 1A9) but 1A6 form has a 25-fold lower apparent Km (2 mM). Rat liver microsomal APAP UGT activity is induced by in vivo treatment with β-naphthoflavone or oltipraz, an effect correlating with induction of 1A6 and 1A7. To address a possible role of 1A7 in APAP glucuronidation relative to other 1A forms, cDNAs encoding UGTs 1A1, 1A5, 1A6, 1A7, and 1A8 were expressed in human embryonic kidney cells and the contents of expressed enzyme in prepared membrane fractions determined by quantitative immunoblotting. At 2.5 mM APAP, 1A7 showed the highest specific activity (2.8 nmol/min/nmol 1A7 protein), followed by 1A6 (1.1 nmol/min/nmol), and 1A8 (0.27 nmol/min/nmol). 1A1 and 1A5 were essentially inactive. Kinetic comparisons indicated 1A7 had a similar apparent Km as 1A6 (4.7 versus 3.9 mM, respectively) but a 2.4-fold higher catalytic activity. These data suggest that in rats, 1A7 plays a major role in APAP glucuronidation and contributes to protection against APAP-induced hepatotoxicity. The involvement of other UGTs besides 1A6 is further underscored by the presence of significant residual APAP-glucuronidating activity by Gunn rat hepatocytes, indicating the activity of an unknown UGT2 family member.
APAP1(Tylenol) is a widely used over-the-counter analgesic/antipyretic drug and model hepatotoxin. Although it is considered safe at normal doses, at higher doses it is associated with a predictable, dose-dependent centrilobular hepatotoxicity (Black, 1984) by a mechanism involving its metabolism to a toxic quinone imine (Cohen and Khairallah, 1997;Bessems and Vermeulen, 2001). APAP undergoes detoxification by competing phase 2 conjugation reactions, glucuronidation and sulfation, which convert APAP to nontoxic conjugates for elimination in bile or urine. The main type of APAP conjugation in most species, including rat and human, is glucuronidation. It has been proposed that individuals who metabolize APAP at slower rates are at higher risk of developing hepatotoxicity, based partly on the known greater susceptibility of animals that are deficient in the glucuronidation of phenols, such as cats (Feloidae) (Court and Greenblatt, 2000; Bessems and Vermeulen, 2001) and the Gunn rat (de Morais and Wells, 1988, 1989).
A common assumption has been that the increased susceptibility of cats and Gunn rats to APAP toxicity is due to inactive UGT1A6, a major liver-expressed phenol UGT. The wild-type 1A62 enzyme from rat and human is effective in APAP catalysis (Bock et al., 1993), and in rats, 1A6 is induced by 3-methylcholanthrene (Munzel et al., 1994;Bock et al., 1999), which results in increased APAP glucuronidation (Gregus et al., 1990; Bock et al., 1993). In cats, the 1A6 gene is not expressed and possesses characteristics of a pseudogene (Court and Greenblatt, 2000). In Gunn rats, the UGT1A6 gene is expressed but is nonfunctional due to a frame shift mutation that inactivates the entire set of 1A enzymes (Iyanagi et al., 1989). However, evidence supports the involvement of other UGT isozymes in APAP glucuronidation. A second human UGT1A family member, the 1A9 form, was reported to have high activity toward APAP, although it exhibited significantly lower affinity (Km of ∼50 versus 2 mM for 1A6) (Bock et al., 1993). In addition, the observation that APAP glucuronidating capacity is not totally absent in Gunn rats (de Morais and Wells, 1989) suggests that a UGT2 isoform also contributes to APAP glucuronidation.
Although the rat is the best characterized animal model for drug metabolism studies, the UGT system from this species is not well understood. Rat liver expresses an array of UGT1A forms similar to that of human with some notable differences. The rat, like human, expresses 1A1 and 1A6 at relatively high levels in liver. However, 1A9 is a pseudogene in rats (Emi et al., 1995), and in this species, its function may be replaced by 1A7, which exhibits a similar preference for “bulky” phenol-type substrates. In humans, 1A7 is not significantly expressed in liver and is considered an extrahepatic UGT (Strassburg et al., 1997). 1A6 and 1A7 are both markedly induced in rat liver by exposure to inducing agents in the polycyclic aromatic hydrocarbon and dithiole thione families (Grove et al., 1997). Based on the reported characteristics of 1A9, we considered the possibility that rat 1A7 is also significantly active toward APAP. The main goal of the current study was to directly assess the APAP UGT activities of 1A7 and the other major UGT1A isoforms expressed in rat liver.
Materials and Methods
APAP hydrochloride and APAP glucuronide were obtained from Sigma Chemical (St. Louis, MO). BNF and phenobarbital were purchased from Aldrich Chemical (Milwaukee, WI) and Sigma Chemical, respectively. Oltipraz was from Aventis (Strasbourg, France). The lambda ZAP cDNA library synthesis kit and cloned Pfu DNA polymerase were from Stratagene (La Jolla, CA). [α-32P]dCTP (3000 Ci/mM) was from ICN Pharmaceuticals (Costa Mesa, CA). The human embryonic kidney-derived HEK 293 cell line and G418 sulfate were purchased from American Type Culture Collection (Reston, VA) and Mediatech (Herndon, VA), respectively. Human UGT1A common region anti-peptide antibody was purchased from GENTEST (Woburn, MA).
Animals and Drug Treatment.
All animal experiments were conducted according to the National Institutes of Health guidelines for the care and use of animals using protocols approved by the Virginia Commonwealth University Institutional Animal Care and Use Committee. Male rats (6–8 weeks old, 160–200 g) were obtained from Harlan Sprague-Dawley (Frederick, MD). After arrival, the rats were housed in steel-wire cages and acclimated for at least 1 week before the experiment during which they were given free access to food and water. In the inducer study, Sprague-Dawley rats were exposed for 3 days to oltipraz, BNF, or phenobarbital. Oltipraz was administered by oral gavage as a micronized suspension in 30% polyethylene glycol-8000 (75 mg/kg/day for 3 days). BNF was administered by intraperitoneal injection in corn oil (12 mg/ml, 80 mg/kg/day for 3 days). Phenobarbital was given as a single intraperitoneal injection (80 mg/kg) in water, followed by exposure for the next 3 days via drinking water (0.1%). In an experiment with Gunn rats, BNF (80 mg/kg/day) was delivered as a suspension in 30% polyethylene glycol by gavage for 3 days. Twenty-four hours after the final dose, livers were removed for the preparation of liver microsomes as previously described (Kessler and Ritter, 1997). The final microsome pellets were resuspended in 100 mM K2HPO4, pH 7.4, 1 mM EDTA, and 20% glycerol by homogenization (three passes in a Potter Elvehjem tube) and sonication (five 1-s pulses delivered at medium power on ice). The protein concentration was determined using the BCA Protein Assay kit (Pierce Chemical, Rockford, IL). Microsomes were stored frozen at −80°C until analysis.
Preparation of Recombinant Rat UGT-Expressing HEK Cell Membranes.
Complementary DNAs representing the major UGTs expressed in livers of oltipraz-treated rats were isolated by preparing and screening a lambda ZAP cDNA library with poly A+ RNA from an adult male Sprague-Dawley rat pretreated with oltipraz (300 mg/kg/day for 3 days). All probes used for screening corresponded to the 5′-end region of the respective cDNAs and were labeled by random priming in the presence of [α-32P]dCTP or dATP. Fragments representing UGT1A1 (bases 8–300 of GenBank accession number D38065) and UGT1A5 (bases 11–317 of GenBank accession number D38069) were obtained from rat liver cDNA by using polymerase chain reaction as described (Emi et al., 1995). Fragments representing UGT1A6 (bases 29–862) and UGT1A7 (bases 8–689) were derived from clones pR1A6 (Kessler and Ritter, 1997) and pLC14 (Grove et al., 1997). After plating and hybridization, filters were washed at low stringency (final wash in 2× standard saline citrate-0.05% SDS) and exposed to film. Under these conditions, it was established that the 1A5 and 1A7 probes would cross-react with other forms in the 1A2 to 1A5 and 1A7 to 1A9 subfamilies, respectively. For each probe, a total of 10 to 20 clones was plaque purified and analyzed by restriction endonuclease site mapping and partial nucleotide sequencing. This resulted in the successful isolation of full-coding clones for 1A1, 1A5, 1A6, and 1A7.3 The 1A8 cDNA was obtained by reverse transcriptase-polymerase chain reaction by using rat liver cDNA as template (D. Auyeung and J. Ritter, manuscript in preparation).
HEK cells stably expressing the five rat UGT cDNAs were obtained essentially as described previously for human 1A7 (Guillemette et al., 2000) with the following modifications. Four weeks after transfection, the medium was removed and individual cell colonies were selected by overlaying well separated colonies with sterile cloning disks (Scienceware, Pequannock, NJ) saturated with 1× Trypsin-EDTA (Invitrogen, Carlsbad, CA) for 1 min. The cloning disks were then transferred to individual T-25 flasks, and the cells were grown to confluence. Cells were maintained at 37°C under humidified air containing 5% CO2 in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum and 1 mg/ml G418.
HEK cell membranes were prepared by expanding the highest expressing cell clone for each isoform in a total of 24 T-75 flasks. At confluence, the cells were harvested and washed in ice-cold phosphate-buffered saline by using low-speed centrifugation. The final cell pellet was resuspended in 10 ml of 100 mM potassium phosphate, pH 7.4, containing 1 mM EDTA and 20% glycerol and then subjected to three consecutive freeze-thaw cycles and a brief sonication step (5 s on 30% power) to eliminate clumps. The resulting lysate was centrifuged at 5000 rpm for 10 min at 4°C in a DuPont SS34 rotor to pellet a heavy fraction containing nuclei and other large debris. The crude membrane fraction was then obtained by ultracentrifugation at 40,000 rpm for 45 min in a DuPont T1270 fixed angle rotor at 4°C and washed by repeating the resuspension and centrifugation steps. The final crude membrane fraction pellet was resuspended in 1 ml of 100 mM potassium phosphate, pH 7.4, containing 1 mM EDTA and 10% glycerol. Protein concentration was determined as described above.
APAP UGT Activity Determination.
APAP UGT activities were determined at 37°C in the presence of 50 mM Tris-Cl, pH 7.6, 10 mM MgCl2, 2.5 mM APAP, and microsomal/crude membrane protein (0.5 mg/ml). After preincubation for 5 min at 37°C, UDP-glucuronic acid was added (3 mM final concentration) to start the reaction. After 1 h, reactions were stopped by the addition of perchloric acid (5.7% final concentration). Precipitated protein was removed by centrifuging at 14,000 rpm for 2 min, and the supernatants were transferred to a fresh tube and stored at −80°C until the analysis.
APAP glucuronide in the postreaction supernatant was determined using a modification of a published HPLC APAP glucuronide assay method (Bock et al., 1993). The HPLC system consisted of a U6K injector, a model 510 pump, a Waters Resolve C18 Guard Pak precolumn cartridge, a Whatman Partisil 10 ODS-2 (4.6 × 250 mm) reverse phase column, a Lambda-Max model 481 spectrophotometer, and a Waters 740 data module. The analysis was run under isocratic conditions by using 96.25% 10 mM phosphoric acid, pH 2.3, and 3.75% acetonitrile as the mobile phase at a flow rate of 1.7 ml/min. Detection was by absorption at 254 nm. Under these conditions, APAP glucuronide, APAP sulfate, and parent APAP eluted at 6.7, 9.5, and 15.9 min, respectively. The position of the glucuronide was verified by carrying out control reactions (with and without UDP-glucuronic acid). Quantitation used a standard curve with authentic APAP glucuronide (50–1000 pmol). Under the conditions of the assay, the generation of APAP glucuronide was found to be linear with respect to time (0–60 min) and protein concentration (0.1–1 mg/ml).
Kinetic Studies.
In the experiments to determine the apparent Michaelis-Menten constants for APAP glucuronidation, the UDP-glucuronic acid concentration was 3 mM and the APAP concentration was varied between 0.33 and 40 mM. The apparent Km andVmax were calculated using Tallarida Pharm-PCS software (version 4.0), which estimatesKm andVmax based on standard linear regression analysis.
Western Immunoblot Analysis.
Relative levels of UGTs in rat liver microsomes and the UGT-expressing HEK cell membranes were determined using an enhanced chemiluminescence-based immunoblotting procedure essentially as described (Ritter et al., 1999; Guillemette et al., 2000). The phenol UGTs were analyzed using antisera raised in mice immunized with 6Xhis-tagged fusion proteins representing amino acids 14 to 132 of human 1A6 or 21 to 149 of rat 1A7. The human 1A6 antiserum displays moderate cross-reactivity toward rat 1A6 at a low dilution (1:500). Total UGT1A protein was analyzed using an anti-peptide antibody directed against the UGT1A common region (GENTEST). The absolute levels of expressed enzymes in crude cell membranes were established using a bacterially expressed 6Xhis-tagged fusion protein developed for this study. The fusion protein contains C-terminal residues 21 to 531 of r1A7 at its C terminus. The fusion protein was purified by standard Ni-NTA Sepharose affinity chromatography (QIAGEN, Valencia, CA). The concentration of the 58-kDa standard in the preparation was estimated by silver staining of SDS-polyacrylamide gel electrophoresis gels containing known amounts of human serum albumin standard (66 kDa).
Primary Rat Hepatocyte Cultures.
Hepatocytes from male Sprague-Dawley or Gunn j/j rats were obtained from the Hepatocyte Isolation Core Facility of the Liver Center at Virginia Commonwealth University. Cells (800,000/35-mm gelatin-coated dish) were plated in Williams' E medium containing 10% fetal bovine serum, 100 U/ml penicillin, 100 μg of streptomycin, 0.125 U/ml insulin, 1 μM T4, and 0.1 μM dexamethasone. After the cells had attached (∼4 h after plating), the medium was changed to serum-free medium. The following day, vehicle (0.1% v/v dimethyl sulfoxide) or vehicle containing oltipraz was added (50 μM final concentration) to induce 1A6 and 1A7 expression. Twenty-four hours later, medium with APAP (4 mM) was added, and aliquots were removed at the indicated times for analysis of APAP glucuronide.
Results
Effect of 1A6- and 1A7-Inducing Agents on Rat Liver Microsomal APAP UGT Activity.
An initial study was conducted to determine the effects of prototypical microsomal enzyme-inducing agents on relative liver microsomal 1A6 and 1A7 levels and APAP UGT activities. Antibodies raised against the N-terminal sequences of human 1A6 and rat 1A7 were tested for reactivity toward various UGT1A family isoforms expressed in HEK cells (to be described further below). The human 1A6 antisera was strongly reactive with r1A6 and had no apparent reactivity toward either r1A1, r1A5, r1A7, or r1A8 (Fig. 1A). In contrast, the rat 1A7 antisera also exhibited high selectivity, reacting preferentially with r1A7 and, to a lesser extent, r1A8.
Effect of UGT1A6/1A7 and UGT2B1-inducing agents on liver microsomal APAP UGT activities and phenol UGT expression in normal rats.
A, selectivity of the human 1A6 and rat 1A7 antisera toward different rat UGT1A isoforms. Membranes from different clonal UGT-expressing cell lines were prepared and analyzed by Western immunoblotting, as described under Materials and Methods. An immunoblot probed with human UGT1A common antibody is included for comparison. B, effects of inducers on UGT1A6, UGT1A7, or total UGT1A protein determined by immunoblotting with the indicated antibodies. Sprague-Dawley rats were treated subacutely with vehicle (VH), oltipraz (OTP), BNF, or phenobarbital (PB). Liver microsomes from two control or inducer-treated animals were prepared and analyzed as described underMaterials and Methods. The average magnitude of induction determined by densitometric analysis (n = 2) is indicated. C, APAP UGT activities determined by HPLC assay. The data represent the mean ± standard error of the mean (n = 3 rats/group).
The results of immunoblotting of rat liver microsomes from the vehicle and inducer-treated animals by using the h1A6 and r1A7 antisera are shown in Fig. 1B. The data suggest that both 1A6 and 1A7 were elevated to the greatest extent by BNF and oltipraz and to a lesser extent by phenobarbital. Although these data do not strictly rule out 1A8 induction in rat liver in response to BNF and oltipraz, results of reverse transcription-polymerase chain reaction analysis of control and inducer-treated rat liver mRNA (data not shown) and cDNA library cloning results (described in more detail below) suggest negligible expression of 1A8 compared with 1A7 in liver in either vehicle- or inducer-treated animals.
The effect of the inducer treatments on APAP UGT activities (Fig. 1C) was found to parallel the immunoblot data, with BNF and oltipraz elevating the activity to the greatest extent (2- and 1.8-fold, respectively) and phenobarbital having a weaker effect (1.2-fold). The lower effect of phenobarbital was not due to inefficient induction by this agent, because morphine UGT activity was found to be elevated 3.5-fold in the phenobarbital-treated rat liver microsomes (data not shown). Correlation analysis indicated significant positive associations between APAP UGT activity and the relative levels of both 1A6 and 1A7 determined by densitometric analysis (r2 values of 0.65 and 0.81, respectively).
The effect of having a nonfunctional UGT1A complex on rat liver microsomal APAP UGT catalysis is illustrated in Fig.2. Liver microsomes from BNF-exposed homozygous Gunn rats (j/j) showed 70% lower APAP UGT activity than corresponding samples from homozygous (+/+) normals (mean values of 0.46 versus 0.14 nmol/mg/min, respectively). Heterozygotes (j/+) exhibited an intermediate activity (0.36 nmol/mg/min). The data shown in Figs. 1 and 2 are consistent with a partial contribution of 1A6 to the total liver microsomal APAP UGT activity, but do not exclude a role for 1A7 or other UGT1A isoforms.
APAP glucuronidation activities of liver microsomes from UGT1A1-normal and UGT1A1-defective rats.
APAP UGT activities of liver microsomes from homozygous UGT1A1-normal (+/+), homozygous UGT1A-defective (j/j), or heterozygous UGT1A-defective (j/+) rats were treated with BNF and analyzed as described under Materials and Methods. Columns represent the activities determined for two different rats.
Cloning, Sequence Characterization, and Expression of Rat UGT1A Family cDNAs in HEK Cells.
To characterize the activities of the major UGT1A isoforms in rat liver, cDNA clones representing each of these forms were isolated from a lambda ZAP Sprague-Dawley rat liver cDNA library. Screening of the library with 5′-end directed probes indicated high abundance of 1A6 and 1A7-like positive clones followed in relative order by 1A1- and 1A5-like clones. Restriction analysis of purified clones suggested that 1A1, 1A5, 1A6, and 1A7 were the only UGT1A forms significantly expressed in adult male Sprague-Dawley rat liver.
Sequencing of full-coding clones selected for each form (plasmids p1-16, p5-11, p6-30, and p7-5, respectively) revealed several nucleotide sequence differences from sequences previously reported to GenBank. Most of these represented neutral substitutions having no effect on the coding sequence compared with published sequences, however, two of the clones exhibited differences (Table1). The 1A5 clone, p5-11, encodes a UGT with Phe rather than Leu at position 108 as reported in GenBank accession number D38069. The 1A7 clone, p7-5, differs from our previously reported 1A7 clone, pLC14 (Grove et al., 1997), at two positions. The newer clone is predicted to encode Asp and Glu at positions 394 and 403. Because the earlier clone was derived from an immortalized rat liver cell line, RALA LCS-3 cell line, it would appear that these differences reflect mutations acquired during propagation in culture. The 1A1 and 1A6 clones, p1-16 and p6-30 are each predicted to encode proteins identical to previously reported sequences (GenBank accession numbers RNU20551 and D83796, respectively).
Comparison of deduced amino acid sequences of rat UGT clones
During library screening and sequencing, we also obtained a complete-coding copy of rat 1A8 cDNA by using polymerase chain reaction with Pfu DNA polymerase (D. Auyeung and J. Ritter, manuscript in preparation). Sequence analysis of the 1A8 cDNA indicated several differences from the published 1A8 sequence (GenBank accession number D38063), which were verified by reamplifying and resequencing. These differences are also summarized in Table 1.
Clonal populations of HEK cells stably transfected with expression plasmids for each form were analyzed by Western blotting to compare the relative levels of expressed UGTs. Marked differences were apparent, with some of the UGTs exhibiting low expression in general (e.g., 1A5). The screening was repeated in an attempt to obtain clones with higher expression. However, only one clone set (r1A1) yielded a higher expressing isolate (r1A1-4). Clonal HEK cell populations r1A1-4, r1A5-2, r1A6-1, r1A7-3, and r1A8-1 with the highest relative expression were then selected and expanded in culture for the preparation of crude cell membrane fractions. The elimination of the nuclear and cytosolic fractions facilitated handling in enzymatic assays and enriched the UGT protein content ∼3-fold (data not shown). Relative differences in UGT content of the final membrane fractions were assessed by immunoblot analysis with UGT1A common region antibody in conjunction with densitometric image analysis (Fig. 1A). Quantitative immunoblotting of the 1A7-3 membranes by using affinity-purified His-tagged 1A721-531 standard (Fig.3C) enabled the absolute levels of UGT1A7 in the r1A7-3 lysates to be determined. These data together with the data for relative expression levels determined using the common region antibody (determined by densitometric analysis of the data in Fig. 1A) enabled the absolute UGT level in each preparation to be determined (range from 0.07–0.29 nmol/mg total protein).
Quantitative immunoblotting approach to determine the content of expressed 1A7 in clonal line r1A7-3.
Known amounts of a purified 6Xhis-tagged rat UGT1A7 fusion protein (r1A7FL) were electrophoresed in parallel with serially diluted samples of a clonal HEK cell line expressing r1A7 (r1A7-3) and analyzed by Western blotting with the r1A721-149 antisera. Densitometric analyses enabled estimation of the absolute level of expression of r1A7 protein in the r1A7-3 membranes.
Membranes from HEK cells expressing each of the five UGT clones were tested for the ability to catalyze APAP glucuronidation. Preliminary experiments with a [14C]UDP-glucuronic acid assay method suggested high activity of the 1A6 and 1A7 isoforms toward APAP, which was confirmed using the HPLC-based method (Table2). Activity was highest for the 1A7 crude membranes, followed by 1A6 and then 1A8, with little or no significant activity apparent for 1A1 or 1A5. Expressing the data on a per nanomole enzyme basis, 1A7 (2.8 nmol/min/nmol) was 2.6-fold more active than 1A6 (1.1 nmol/min/nmol) and >10-fold more active than 1A8 (0.27 nmol/min/nmol).
UGTs 1A6, 1A7, and 1A8 are catalytically active towards APAP
Kinetic Analysis of APAP UGT Activities in Expressing HEK Cell Protein and Liver Microsomes.
To compare the apparent Km andVmax of the two most active forms, the effect of varying the substrate concentration on initial reaction velocities was assessed. In these assays, UDP-glucuronic acid concentration was held constant at 3 0.0 mM. Recombinant 1A6 and 1A7 exhibited similar apparent affinities of 3.47 ± 0.4 mM versus 4.56 ± 1.5 mM, respectively (Fig.4A). TheVmax values were also comparable (0.60 ± 0.04 and 1.41± 0.10 nmol/min/mg protein, respectively). Expressed in units of specific activity, 1A7 had a 2.9-fold higher specific Vmax (7.83 versus 2.69 nmol/min/nmol).
Kinetic analysis of APAP glucuronidation catalyzed by rat 1A6 and 1A7 and comparison with rat liver microsomes with normal or defective UGT1A isozymes.
A, membranes from r1A6-1 and r1A7-3 were assayed for APAP UGT activity by using different concentrations of APAP and 3 mM UDP-glucuronic acid. The reciprocals of substrate concentration and reaction velocity were graphed to yield the Lineweaver-Burk plots shown. B, corresponding analyses were performed using liver microsomes from either a homozygous normal (+/+) or homozygous defective (j/j) Gunn rats.
The kinetic parameters were also determined for liver microsomes from homozygous normal (Gunn +/+) and UGT1A defective rats (Gunn j/j) (Fig.4B). For homozygous normal rats, theKm andVmax were 8.31 ± 1.7 and 1.85 ± 0.3, respectively. For homozygous UGT1A defective rats, the apparent Km increased (20.7± 5.3) and the Vmax was reduced (0.92 ± 0.20), consistent with the loss of the higher affinity 1A6 and 1A7 forms.
Induction of APAP Glucuronidation in Cultured Primary Rat Hepatocytes by Oltipraz.
An experiment was performed to study the glucuronidation of APAP by cultured primary hepatocytes from normal and UGT1A-deficient rats and the influence of pre-exposing the hepatocytes to oltipraz. Western analysis of lysate from control cells and cells exposed to oltipraz for 24 h before addition of APAP to the cell culture medium suggested that 1A6 and 1A7 UGT levels were elevated by oltipraz at the time of APAP addition (Fig. 5A). The glucuronidation of APAP was stimulated >2-fold in the oltipraz-exposed cells, whereas the sulfation pathway was not visibly altered (Fig. 5B). In contrast, hepatocytes from UGT1A defective Gunn rats (j/j) formed APAP-glucuronide at reduced rates compared with Sprague-Dawley hepatocytes, and there was no apparent stimulation of glucuronidation or sulfation by oltipraz (Fig. 5C).
Influence of oltipraz pretreatment on the formation of APAP glucuronide in cultured primary hepatocytes from normal and UGT1A-defective rats.
Freshly isolated rat hepatocytes (800,000 cells/35-mm well) were exposed to either dimethyl sulfoxide (0.1%) (open symbols) or 50 μM oltipraz added to the culture medium. After 24 h, the medium was changed, and APAP was added (4 mM final). At the times indicated, the culture medium was sampled for analysis of APAP glucuronide (squares) and sulfate (circles). A, immunoblot analysis for 1A6, 1A7, and UGT1A total protein in Sprague-Dawley hepatocytes. B, oltipraz stimulates APAP glucuronidation in Sprague-Dawley rat hepatocytes. C, oltipraz has no effect on APAP glucuronidation in homozygous defective (j/j) Gunn rat hepatocytes.
Discussion
APAP is commonly viewed as a specific “probe” substrate for the 1A6 isoform (Esteban and Perez-Mateo, 1999; Fisher et al., 2000). However, results in this study show that glucuronidation of APAP is catalyzed by multiple rat UGT isoforms, and that, at least in rats, APAP is a somewhat nonspecific substrate. Our data confirm a previous report suggesting high APAP activity of rat 1A6. Using rat 1A6 expressed in V79 Chinese hamster lung fibroblasts, Bock et al. (1993)reported an apparent Km of 2.7 mM, in good agreement with the value determined for HEK cell-expressed enzyme in the current study (3.9 mM). The activity of 1A7 toward APAP has not been previously described. We found that 1A7 has comparable APAP binding affinity and superior catalytic efficiency. These data suggest considerable potential for the 1A7 isoform to compete with 1A6 and contribute significantly to the capacity of rat liver to glucuronidate APAP. The percentage contributions of 1A6 and 1A7 to APAP glucuronidation will depend on their relative levels in the liver endoplasmic reticulum. Although these are not known precisely at present, both appear to be expressed at significant levels in rat liver, especially after treatment with inducing agents (Grove et al., 1997).
Our study also confirms earlier work indicating that 1A1, the major bilirubin-glucuronidating UGT, is not involved in APAP catalysis. Although it was observed that 1A1 did not significantly metabolize APAP, we found that the preparation was highly active in bilirubin glucuronidation (data not shown). This is of interest in view of reports linking Gilbert's syndrome with reduced APAP glucuronidation and a possible increase in susceptibility to APAP hepatotoxicity (Ullrich et al., 1987; de Morais et al., 1992; Esteban and Perez-Mateo, 1999). The primary cause of Gilbert's in the Caucasian population is reduced UGT1A1 expression related to the UGT1A1∗28 allele (TA repeat) in the 1A1 gene promoter. The suggestion by Esteban and Perez-Mateo (1999) that a significant percentage of Gilbert's patients (with the TA genetic change) may also possess a linked mutation in the phenol transferase gene appears to be possible based on a report by Lampe et al. (1999) that 8% of individuals homozygous for the ∗28 allele were also homozygous for the 1A6 variant allele (UGT1A6∗2) associated with reduced enzymatic activity (Ciotti et al., 1997).
The model for rat APAP glucuronidation catalyzed by UGT1A isoforms described in this study resembles that of humans with some apparent differences. A major similarity is that 1A6 in both species is expressed at high hepatic levels and exhibit similar apparent APAP binding affinity (2.0 mM for human versus 2.7–3.9 mM for rat) (Bock et al., 1993). Both species also express a member of the “bulky phenol” subgroup of UGT1A forms (UGT1A7-UGT1A10 subfamily) that is significantly active in APAP glucuronidation. In the case of human, this form is 1A9 rather than 1A7. There are apparent differences, however, in the kinetic properties of rat 1A7 and human 1A9, especially compared with the respective 1A6 forms from these two species. In contrast to rat, where 1A7 has similar apparent affinity and marginally higher catalytic activity than 1A6, human 1A9 was reported to possess much lower affinity (Km of ∼50 mM) than its 1A6 counterpart (2 mM). In the human study, comparisons of intrinsic clearance (Vmax/Km) for 1A6 and 1A9 were not possible, because the levels of expressed enzyme protein in transfected cells could not be determined. Thus, the relative Vmax of human 1A6 compared with 1A9 remains unclear at present.
Animal studies support the idea that the rate of APAP glucuronidation impacts the hepatotoxic potency of APAP (de Morais and Wells, 1988,1989, 1992; Court and Greenblatt, 2000). APAP is an intermediate clearance drug whose rate of elimination depends in part on the APAP conjugate-forming capacity of the liver. Factors that modulate the levels of APAP UGTs, for example, exposure to oltipraz, are predicted to affect APAP clearance and organ toxicity. In this report, we were able to show using the cultured hepatocyte model that induction of UGTs 1A6 and 1A7 by oltipraz was associated with a stimulation in the rate of APAP glucuronidation. In vehicle-exposed hepatocytes, we found that ∼26% of APAP was conjugated after 26 h, with the sulfoconjugate slightly predominating over the glucuronide (13.5 versus 12.3%, respectively). In hepatocytes exposed for 24 h to oltipraz, the total percentage conjugated increased to 37%, due almost entirely to an increase in the fraction glucuronidated. Oltipraz treatment shifted the ratio in favor of the glucuronide over the sulfate conjugate (22.8 to 14.0%, respectively). Despite our efforts, we were not able to demonstrate the association between UGT induction and protection against APAP toxicity in the cultured hepatocyte model. Others have reported that treatment of primary rat hepatocytes with APAP at concentrations in the range used in our study (4 mM) results in cytotoxic effects (Milam and Byard, 1985; Thibault et al., 1991). However, we were unable to observe any significant changes in cell morphology or biochemical indices (lactate dehydrogenase and aspartate transaminase) after a 24-h exposure to 4 mM APAP (data not shown). The lack of effect is likely related to the general insensitivity of rat as a model for APAP toxicity (Bessems and Vermeulen, 2001), due possibly to high glucuronidation or low bioactivation rates. It is noteworthy, however, that in hamsters, which are known to be highly sensitive to APAP, treatment with oltipraz has been shown to be protective against APAP-induced hepatotoxicity (Davies et al., 1991) by a mechanism involving induction of APAP glucuronidation (Davies and Schnell, 1991). These observations suggest that hamsters exhibit lower basal rates of APAP glucuronidation and that hamsters have corresponding oltipraz-inducible phenol UGTs.
Even under conditions when UGT1A isoforms are absent, it is clear that rat hepatocytes still possesses significant APAP glucuronide-forming activity. These data indicate a role for other UGTs in this process, presumably a UGT2B subfamily member. Although 2B1 and 2B12 are considered the major “xenobiotic-metabolizing” UGT2B family members, neither is active toward APAP (Pritchard et al., 1994; Green et al., 1995). Several UGTs in the rat 2B family, including 2B2, 2B3, and 2B6, each have been reported inactive toward simple phenolic substrates (Mackenzie, 1987). At least one remaining candidate to be investigated is the 2B5 form (GenBank accession number U27518). Data from the kinetic analysis of homozygous Gunn rat microsomes (Fig. 5) predict that this UGT has lower affinity for APAP (Km of ∼26.0) than either 1A6 or 1A7. Evidence has also been reported for a corresponding form in humans. Axelrod et al. (1957) described deficient, albeit not totally absent, APAP glucuronidation in a patient with severe hereditary unconjugated hyperbilirubinemia, who presumably had a UGT1A-inactivating mutation analogous to Gunn rats.
In summary, this study reveals evidence for the involvement of multiple UGT isoforms in the glucuronidation of APAP by rat liver. Two of these, 1A6 and 1A7, are members of the UGT1A family and are implicated in the increased sensitivity of homozygous Gunn j/j rats to APAP-induced hepatotoxicity. Both isozymes were found to possess similar kinetic properties in APAP catalysis. Their relative contributions to the total activity in vivo are predicted to depend on their levels of expression determined by dietary and/or environmental exposures. A third form that is active toward APAP, most likely from the UGT2 subfamily, remains to be identified.
Acknowledgments
We are indebted to the Bastones and the American Liver Foundation for the generous gift in memory of Louis John Bastone. We also thank William Gillespie for help with the rat 1A8 APAP UGT assays and Patricia Bohdan of the Hepatocyte Isolation Core Laboratory for preparing the primary rat hepatocytes.
Footnotes
-
This work was supported by the National Institute of Environmental Health Science Grant ES07762-06 (to J.K.R.). D.J.A. was supported, in part, by an Institutional Training Grant in Toxicological Sciences (ES078087) and a gift from the American Liver Foundation in memory of Louis John Bastone.
-
↵2 Bock et al. (1993) referred to the 3-methylcholanthrene-inducible phenol UGT as the UGT1A1 form. Under the current nomenclature guidelines (Mackenzie et al., 1997), this form is now designated UGT1A6.
-
↵3 The nucleotide sequences for the cDNA clones were submitted to the GenBank database and assigned accession numbersAF461734–38.
- Abbreviations used are::
- APAP
- acetaminophen
- UGT
- uridine 5′-diphosphoglucuronosyl transferase
- BNF
- β-naphthoflavone
- HEK
- human embryonic kidney
- HPLC
- high-performance liquid chromatography
- Received July 31, 2001.
- Accepted December 5, 2001.
- The American Society for Pharmacology and Experimental Therapeutics
