Roles of Human Hepatic Cytochrome P450s 2C9 ... - ACS Publications

Wong, B. K., Goldstein, J., Spielberg, S. P., and Rushmore, T. H. (1996) Characterization of an in vitro model for assessing polymorphic metabolism by...
0 downloads 0 Views 124KB Size
Download PDF
192

Chem. Res. Toxicol. 1999, 12, 192-199

Roles of Human Hepatic Cytochrome P450s 2C9 and 3A4 in the Metabolic Activation of Diclofenac† Wei Tang,* Ralph A. Stearns, Regina W. Wang, Shuet-Hing L. Chiu, and Thomas A. Baillie Department of Drug Metabolism, Merck Research Laboratories, Rahway, New Jersey 07065 Received September 29, 1998

Recently, it was shown that diclofenac was metabolized in rats to reactive benzoquinone imines via cytochrome P450-catalyzed oxidation. These metabolites also were detected in human hepatocyte cultures in the form of glutathione (GSH) adducts. This report describes the results of further studies aimed at characterizing the human hepatic P450-mediated bioactivation of diclofenac. The reactive metabolites formed in vitro were trapped by GSH and analyzed by LC/MS/MS. Thus, three GSH adducts, namely, 5-hydroxy-4-(glutathion-S-yl)diclofenac (M1), 4′-hydroxy-3′-(glutathion-S-yl)diclofenac (M2), and 5-hydroxy-6-(glutathion-S-yl)diclofenac (M3), were identified in incubations of diclofenac with human liver microsomes in the presence of NADPH and GSH. The formation of the adducts was taken to reflect the intermediacy of the corresponding putative benzoquinone imines. While M2 was the dominant metabolite over a substrate concentration range of 10-50 µM, M1 and M3 became equally important products at g100 µM diclofenac. The formation of M2 was inhibited by sulfaphenazole or an anti-P450 2C9 antibody (5-10% of control values). The formation of M1 and M3 was inhibited by troleandomycin, ketoconazole, or an anti-P450 3A4 antibody (30-50% of control values). In studies in which recombinant P450 isoforms were used, M2 was generated only by P450 2C9catalyzed reaction, while M1 and M3 were produced by P450 3A4-catalyzed reaction. Good correlations were established between the extent of formation of M2 and P450 2C9 activities (r ) 0.93, n ) 10) and between the extent of formation of M1 and M3 and P450 3A4 activities (r ) 0.98, n ) 10) in human liver microsomal incubations. Taken together, the data suggest that the biotransformation of diclofenac to M2 is P450 2C9-dependent, whereas metabolism of the drug to M1 and M3 involves mainly P450 3A4. Although P450s 2C9 and 3A4 both catalyze the bioactivation of diclofenac, P450 2C9 is capable of producing the benzoquinone imine intermediate at lower drug concentrations which may be more clinically relevant.

Introduction The nonsteroidal anti-inflammatory drug diclofenac causes a rare, but potentially severe, liver injury (1-3) which may be due to the formation of reactive metabolite(s) (4, 5). Using immunochemical detection, diclofenac was found to form protein adducts in the liver of treated mice and rats (6, 7) as well as in hepatocyte cultures (8, 9), apparently via activation of its carboxylic acid group. Several lines of evidence, such as the incorporation of radiolabel into proteins in incubations containing [14C]UDP-glucuronic acid, diclofenac, and rat hepatic microsomes (10), indicate that the acyl glucuronide metabolite of diclofenac is involved in the formation of those protein adducts. However, data also have been obtained that suggest that reactive metabolites of diclofenac are generated via a P4501-mediated pathway(s) (8, 11-13). For example, P450 2C11 was modified covalently in male rats to which diclofenac had been adminsitered, and the † A preliminary account of this study was presented at Experimental Biology, San Francisco, CA, April 1998. * Corresponding author: Department of Drug Metabolism, Merck & Co., P.O. Box 2000, RY80L-109, Rahway, NJ 07065. Telephone: (732) 594-4501. Fax: (732) 594-1416. 1 Abbreviations: P450, cytochrome P450; GSH, reduced glutathione; LC/MS/MS, liquid chromatography/tandem mass spectrometry; M1, 5-hydroxy-4-(glutathion-S-yl)diclofenac; M2, 4′-hydroxy-3′-(glutathionS-yl)diclofenac; M3, 5-hydroxy-6-(glutathion-S-yl)diclofenac.

covalent binding proved to be an NADPH-dependent process in rat liver microsomal incubations (11). Three GSH-conjugated metabolites, namely, 5-hydroxy-4-(glutathion-S-yl)diclofenac (M1), 4′-hydroxy-3′-(glutathion-Syl)diclofenac (M2), and 5-hydroxy-6-(glutathion-S-yl)diclofenac (M3), were identified in bile from rats receiving diclofenac (12). The formation of these non-protein thiol adducts most likely occurs via P450-catalyzed oxidation of diclofenac to benzoquinone imine intermediates followed by conjugation with GSH (12). In humans, the oxidative metabolism of diclofenac is catalyzed by P450s 2C9 and 3A4 (13-16). The P450 2C9mediated reaction gives rise to 4′-hydroxydiclofenac with a Km value of 4-6 µM, and the P450 3A4-catalyzed process results in 5-hydroxydiclofenac with a Km that is ∼20 times higher than that of P450 2C9 (13-16). In an earlier study in which human liver microsomes were used, only 5-hydroxydiclofenac was shown to undergo further metabolism to a reactive intermediate that bound to protein(s) (13). More recently, the GSH adducts M1M3, presumably derived from 4′-hydroxy- and 5-hydroxydiclofenac, were detected in human hepatocyte cultures treated with diclofenac (12). In this report, we describe the characterization of human hepatic P450-mediated bioactivation of diclofenac. Reactive metabolites formed in vitro were trapped by GSH and the resulting conju-

10.1021/tx9802217 CCC: $18.00 © 1999 American Chemical Society Published on Web 01/22/1999

Metabolic Activation of Diclofenac

gates analyzed by liquid chromatography/tandem mass spectrometry (LC/MS/MS). A correlation then was established between diclofenac concentrations and the P450 isoforms that were involved in bioactivation of the drug.

Materials and Methods Chemicals and Biochemicals. Diclofenac, GSH, 4-methylpyrazole, NADPH, quinidine, sodium diethyldithiocarbamate, and troleandomycin were purchased from Sigma Chemical Co. (St. Louis, MO). Trifluoroacetic acid was from Fisher Scientific (Fair Lawn, NJ), ketoconazole from Janssen Biotech (Olen, Belgium), furafylline from Research Biochemicals International (Natick, MA), and EDTA from Aldrich Chemical Co. (Milwaukee, WI). BondElut C18 solid phase extraction cartridge columns were obtained from Varian Chromatography Systems (Walnut Creek, CA). Sulfaphenazole was a gift from F. P. Guengerich (Vanderbilt University, Nashville, TN). The following recombinant P450 enzymes were purchased from Gentest Corp. (Woburn, MA): microsomes from a baculovirus-insect cell line expressing human P450 2C9 (Cys144 or Arg144) or 3A4 (+b5) and microsomes from a human lymphoblast cell line expressing human P450s 1A2, 2A6, 2C8, 2C19, 2D6, and 2E1. The P450 contents, activities, and specificities were assumed to be the same as those reported by the manufacturer. Chemical synthesis of M1 and M3 was carried out via oxidation of 5-hydroxydiclofenac by silver(I) oxide followed by conjugation with GSH (12). The synthesis of M2 was performed similarly using 4′-hydroxydiclofenac as the precursor (12). Anti-P450 3A4 antibody was prepared in rabbits by immunization with a synthetic peptide which was identical to residues 253-273 of P450 3A4 and was coupled to keyhole limpet hemocyanin (17). Anti-P450 2C9 antibody was a gift from J. Lasker (Alcohol Research and Treatment Center, Bronx Veterans Medical Center, Mt. Sinai School of Medicine, New York, NY). Instrumentation and Analytical Methods. LC/MS/MS experiments were performed on a Perkin-Elmer SCIEX (Toronto, ON) API III+ tandem mass spectrometer interfaced with an HPLC system consisting of two Shimadzu (Kyoto, Japan) LC-10A pumps and a static-bed mixer. Positive ion spray was used for ionization with a voltage of 5 kV. The orifice potential was 65 V. For collision-induced dissociation, argon was used as the collision gas at a thickness of 1.3 × 1014 atoms/cm2 and the collision energy was 30 eV. Chromatography was performed on a DuPont Zorbax (Wilmington, DE) Rx-C8 column (4.6 mm × 250 mm, 5 µm), and samples were delivered at a flow rate of 1 mL/min with a 1:25 split. The mobile phase consisted of acetonitrile/water containing 10% methanol and 0.05% trifluoroacetic acid and was programmed for a linear increase from 10 to 70% acetonitrile over the course of 30 min. Human Liver Microsomal Preparations. Human liver samples were obtained from Agouron Institute (La Jolla, CA), Memorial Sloan-Kettering Cancer Center (New York, NY), and Pennsylvania Regional Tissue Bank (Exton, PA). The medical history of the donors, in terms of disease state, alcohol consumption, and substance abuse, is available as Supporting Information. Liver microsomes were isolated from individual livers by differential centrifugation (18). Aliquots from each preparation were pooled on the basis of equivalent protein concentrations, resulting in two mixtures. Mixture 1 consisted of liver microsomes from five donors (samples 1-1 to 1-5, Agouron Institute and Memorial Sloan-Kettering Cancer Center) and mixture 2 of samples from 10 donors (samples 2-1 to 2-10, Pennsylvania Regional Tissue Bank) (Supporting Information). Specific P450 activities in these microsomal preparations were estimated using phenacetin O-deethylation for P450 1A2, tolbutamide methylhydroxylation for P450 2C9, bufuralol 1′-hydroxylation for P450 2D6, chlorzoxaone 6-hydroxylation for P450 2E1, and testosterone 6β-hydroxylation for P450 3A4 (17, 19).

Chem. Res. Toxicol., Vol. 12, No. 2, 1999 193 Incubations with Human Liver Microsomes and with Recombinant P450 Isoforms. Incubations were performed with microsomal mixtures 1 and 2, except where indicated. Briefly, diclofenac and GSH in phosphate buffer (pH 7.4) were added to human liver microsomes (1 mg of protein/mL) or recombinant P450 (0.5-1 mg of protein/mL) suspended in phosphate buffer (0.1 M, pH 7.4) containing EDTA (1 mM). The final volume was 0.5 mL, and substrate concentrations were 10, 25, 50, 100, 200, 500, and 1000 µM. The concentration of GSH was 5 mM. Microsomes containing an empty cDNA expression vector were used as controls. The mixture was incubated at 37 °C for 5 min before adding NADPH in phosphate buffer (final concentration of 1 mg/mL) to initiate the reaction. After incubation for an additional 30 min, the reaction was quenched by adding 10% aqueous trifluoroacetic acid (60 µL). Incubations of 4′-hydroxydiclofenac (10 µM) and 5-hydroxydiclofenac (50 µM) with human liver microsomes or with recombinant P450 isoforms were performed in a similar fashion. In the latter case, microsomes containing an empty cDNA expression vector were used as controls. In experiments involving P450 isoform-specific inhibitors, microsomes were preincubated with ketoconazole, sulfaphenazole, quinidine, and 4-methylpyrazole for 10 min and were preincubated with troleandomycin, diethyldithiocarbamate, and furafylline in the presence of NADPH for 15 min at 37 °C. The inhibitors were dissolved in methanol, and their final concentrations were 1 (ketoconazole), 20, and 40 µM (other inhibitors). The same amount of solvent was added into control incubations (final concentration of 0.2%). Reactions were quenched 30 min after addition of the substrate by adding 10% aqueous trifluoroacetic acid (60 µL). In immunoinhibition experiments, microsomes were preincubated with each antibody for 30 min at room temperature. Control incubations contained preimmune IgG. Thereafter, diclofenac, GSH, and NADPH were added and the incubations performed in a similar manner as described above. Detection and Quantitation of GSH-Conjugated Metabolites. Samples from incubations were applied to a C18 extraction cartridge column which had been prewashed with methanol and water. The column was washed consecutively with water and methanol. The methanol eluate was evaporated to dryness under a stream of nitrogen and the residue reconstituted in 300 µL of aqueous acetonitrile (60%) containing trifluoroacetic acid (0.05%). An aliquot of the resulting samples (80 µL) was injected onto the Zorbax Rx-C8 column and analyzed by LC/MS/MS. Identification of the metabolites was based on multiple reaction monitoring detection of four mass transitions, m/z 617 f 542, 617 f 488, 617 f 342, and 617 f 324 (12). The dwell times were 400 ms per channel.

Results Metabolite Formation in Incubations with Human Liver Microsomes. The detection of the GSHconjugated metabolites of diclofenac, M1-M3, was based on selective MS/MS monitoring coupled with HPLC separation. The characteristic mass transitions employed for metabolite identification corresponded to the loss of glycine (75 Da; m/z 617 f 542), loss of pyroglutamate (129 Da; m/z 617 f 488), cleavage between the cysteinyl C-S bond with charge retention on the aromatic moiety (m/z 617 f 342), and further loss of a water molecule from the fragment ion at m/z 342 (m/z 617 f 324), following collision-induced dissociation of the protonated molecular ions (MH+) (12). The metabolites of interest were identified by comparing their HPLC retention times and MS/MS product ions with those of synthetic reference compounds. When diclofenac was incubated with human liver microsomes in the presence of NADPH and GSH, the

194 Chem. Res. Toxicol., Vol. 12, No. 2, 1999

Tang et al.

Figure 1. LC/MS/MS detection of 5-hydroxy-4-(glutathion-S-yl)diclofenac (M1), 4′-hydroxy-3′-(glutathion-S-yl)diclofenac (M2), and 5-hydroxy-6-(glutathion-S-yl)diclofenac (M3) in incubations containing diclofenac, human liver microsomes, NADPH, and GSH. Four mass transitions, together with HPLC retention times, were used as criteria for metabolite identification (m/z 617 f 524, 617 f 488, 617 f 342, and 617 f 324). Substrate concentrations were (A) 10, (B) 50, and (C) 200 µM.

extent of formation of M1-M3 was found to be linear over a period of 45 min (data not shown), and therefore, incubations were performed for 30 min. Metabolite profiles, however, were dependent on substrate concentration. At 10 µM substrate, only M2 was detected (Figure 1A). The extent of formation of M2 increased with increasing substrate concentration, but the increase in the extent of M2 formation plateaued at ∼25 µM diclofenac. In cases where microsomes from individual livers (samples 2-1 to 2-10) were used as the sources of P450 enzyme(s), M2 production paralleled tolbutamide hydroxylation, a measure of P450 2C9 activity (Figure 2). While M2 was the predominant metabolite over a substrate concentration range of 10-50 µM, M1 and M3 gradually became equally important products as the concentration of diclofenac was raised (Figure 1C), such that the formation of M1 and M3 was maximal at ∼300 µM diclofenac. The level of production of M1 and M3 correlated well with the catalytic activity of testosterone6β-hydroxylase, namely, P450 3A4 activity, in individual human liver microsomal preparations (Figure 3). Metabolite M2 also was generated in incubations of 4′hydroxydiclofenac with human liver microsomes in the presence of NADPH and GSH, and M1 and M3 were formed from 5-hydroxydiclofenac. In the latter case, however, no difference was found between incubations containing human liver microsomes and those containing no protein. Metabolite Formation in Incubations with Recombinant P450 Enzymes. In incubations with recombinant P450 2C9, M2 was the only product detected at 10 µM diclofenac (Figure 4A). When the substrate concentration was increased to 1000 µM, the peaks

Figure 2. Correlation between the extent of formation of M2 and the specific activity of P450 2C9 (tolbutamide hydroxylation) in human liver microsomes (samples 2-1 to 2-10). The amount of M2 formed is expressed relative to the highest yield obtained among incubations with 10 microsomal preparations. The activity of P450 2C9 is expressed relative to the highest activity determined among 10 microsomal preparations. The substrate concentration was 10 µM. The correlation coefficient (r ) 0.93) was determined using linear regression analysis.

corresponding to M1 and M3 also were detected, although M2 remained the major product (Figure 5A). Incubations of diclofenac with two other members of the P450 2C family, 2C8 and 2C19, resulted in the detection of M1 and M3 (Figure 5B,C).

Metabolic Activation of Diclofenac

Figure 3. Correlation between the extent of formation of M1 and M3 and the specific activity of P450 3A4 (testosterone hydroxylation) in human liver microsomes (samples 2-1 to 2-10). The amount of M1 and M3 formed is expressed relative to the highest yield obtained among incubations with 10 microsomal preparations. The activity of P450 3A4 is expressed relative to the highest activity determined among 10 microsomal preparations. The substrate concentration was 100 µM. The correlation coefficient (r ) 0.98) was determined using linear regression analysis.

When 4′-hydroxydiclofenac at 10 µM was incubated with recombinant P450s 2C8, 2C9, 2C19, and 3A4, M2 was detected only in incubations containing P450 2C9. In these experiments, the substrate concentration was selected such that it was close to the Km value (4-6 µM) determined for the P450 2C9-catalyzed 4′-hydroxylation of diclofenac (14-16).

Chem. Res. Toxicol., Vol. 12, No. 2, 1999 195

Metabolites M1 and M3, but not M2, were detected in incubations with recombinant P450 3A4 over a substrate concentration range of 10-1000 µM (Figure 4B). However, when 10 µM diclofenac was incubated with a mixture of P450s 2C9 and 3A4 at a 2C9:3A4 ratio of 2:3, only M2 was detected (Figure 4C). This P450 concentration ratio was chosen to represent that found in human livers (20), and the results indeed reflected those observed at the same substrate concentration with native human liver microsomes (Figure 1A). Metabolites M1 and M3 also were detected in incubations containing recombinant P450 2E1 or 2D6 at g50 µM diclofenac and in incubations containing P450 1A2 or 2A6 at g200 µM diclofenac. These substrate concentrations were based on the reported Km value for the P450 3A4-catalyzed 5-hydroxylation of diclofenac (13), and therefore, P450 enzymes other than 2C9 and 3A4 may contribute to the biotransformation of the drug. Inhibition of Human Hepatic P450-Mediated Diclofenac Bioactivation. The formation of metabolite M2 seemed to be abolished at 10 µM diclofenac in incubations containing sulfaphenazole, a specific inhibitor of P450 2C9 (Table 1). Associated with this inhibition was the emergence of M1 and M3 (Supporting Information), neither of which was observed in control (containing 0.2% methanol) or in “regular” (containing no organic solvent) incubations at 10 µM substrate (Figure 1A). This is consistent with the results of studies in which recombinant P450 enzymes and diclofenac at 10 µM were used when M1 and M3 were detected in incubations with P450 3A4 but not in incubations with the mixture of P450 2C9 and P450 3A4 (vide supra). At a higher substrate concentration (100 µM), a 90% inhibition of M2 production by sulfaphenazole was accompanied by a 50% increase in the level of formation of M1 and M3 (Table 1). These results suggested a competition among P450 isoforms for

Figure 4. Formation of the GSH-conjugated metabolites of diclofenac in (A) incubations with recombinant P450 2C9, (B) incubations with recombinant P450 3A4, and (C) incubations with a mixture of P450s 2C9 and 3A4 at a 2C9:3A4 ratio of 2:3. The substrate concentration was 10 µM.

196 Chem. Res. Toxicol., Vol. 12, No. 2, 1999

Tang et al.

Figure 5. Formation of the GSH-conjugated metabolites of diclofenac in (A) incubations with recombinant P450 2C9, (B) incubations with recombinant P450 2C8, and (C) incubations with recombinant P450 2C19. The substrate concentration was 1000 µM. Table 1. Effect of P450 Isoform-Specific Inhibitors on the Formation of Reactive Metabolites from Diclofenac in Incubations with Human Liver Microsomesa

inhibitor sulfaphenazole (20 µM) sulfaphenazole (20 µM) sulfaphenazole (40 µM) troleandomycin (20 µM) troleandomycin (40 µM) ketoconazole (1 µM) furafylline (40 µM) quinidine (40 µM) diethyldithiocarbamate (40 µM) 4-methylpyrazole (40 µM)

M1 and [diclofenac] target M2 % M3 % (µM) P450 control control 10 100 100 50-400 50-400 50-400 100 100 100

2C9 2C9 2C9 3A4 3A4 3A4 1A2 2D6 2E1

0 10 10 100 100 100 100 200 50

NDb 141 160 45-50 40-50 40-50 100 400 45

100

2E1

100

100

a Diclofenac and GSH in phosphate buffer were added to human liver microsomes (1 mg of protein/mL) suspended in phosphate buffer (0.1 M, pH 7.4) containing EDTA. Ketoconazole, sulfaphenazole, quinidine, and 4-methylpyrazole were preincubated with microsomes for 10 min, and troleandomycin, diethyldithiocarbamate, and furafylline were preincubated with microsomes in the presence of NADPH for 15 min at 37 °C. The same amount of solvent was added to control incubations (0.2%). Reactions were initiated by adding the substrate and another portion of NADPH and proceeded for an additional 30 min. The products were analyzed by LC/MS/MS. b Metabolites M1 and M3 were not detected in control incubations at 10 µM diclofenac.

the substrate and a preference for metabolism of diclofenac by P450 2C9. The level of formation of M1 and M3 decreased to ∼50% of the control level in incubations containing the P450 3A4-specific inhibitors ketoconazole and troleandomycin (Table 1). Unlike the inhibition of M2 formation

by sulfaphenazole, which led to a significant increase in the extent of conversion of diclofenac to M1 and M3 (vide supra), the formation of M2 was not affected by the presence of ketoconazole or troleandomycin (Table 1). This was likely due to the fact that the P450 2C9catalyzed reaction was saturated at the substrate concentration used (100 µM). In support of this hypothesis was the observation that the amount of M2 formed in control incubations did not increase at substrate concentrations in excess of 25 µM (data not shown). Other chemical inhibitors examined were furafylline (P450 1A2), quinidine (P450 2D6), diethyldithiocarbamate (P450 2E1), and 4-methylpyrazole (P450 2E1). Nonspecific inhibition was seen with diethyldithiocarbamate; no effect was observed with furafylline and 4-methylpyrazole, while a stimulative effect was associated with quinidine (Table 1). The observed increases in the levels of formation of diclofenac metabolites in incubations containing quinidine were inconsistent with the reported inhibitory effect of quinidine on P450 3A4 (21, 22), and this phenomenon currently is under further investigation. Inhibition experiments also were performed with antiP450 2C9 and anti-P450 3A4 antibodies. Similar to results obtained with sulfaphenazole, M2 production was inhibited almost completely by the anti-P450 2C9 antibody, and it was noted that the inhibition occurred in concert with an increase in the level of formation of M1 and M3 (Table 2). On the other hand, the formation of M1 and M3 was inhibited by the anti-P450 3A4 antibody while M2 production was not affected (Table 2). When 10 µM 4′-hydroxydiclofenac was used as the substrate, the level of formation of M2 decreased to 30% of the control level in incubations containing the anti-

Metabolic Activation of Diclofenac

Chem. Res. Toxicol., Vol. 12, No. 2, 1999 197

Table 2. Effect of Anti-P450 Antibodies on the Formation of Reactive Metabolites of Diclofenac or Its Hydroxylated Metabolites in Incubations with Human Liver Microsomesa antibody

substrate

[substrate] (µM)

M2 % control

M1 and M3 % control

anti-2C9 antibody (2 mg of IgG/nmol of CYP) anti-2C9 antibody (4 mg of IgG/nmol of CYP) anti-2C9 antibody (4 mg of IgG/nmol of CYP) anti-3A4 antibody (2.8 mg of IgG/nmol of CYP) anti-3A4 antibody (5.6 mg of IgG/nmol of CYP) anti-3A4 antibody (5.6 mg of IgG/nmol of CYP)

diclofenac diclofenac 4′-hydroxydiclofenac diclofenac diclofenac 5-hydroxydiclofenac

25 25 10b 100-400 100 50b

10 5 30 100 100 NDc

200 230 NDc 50 30 100d

a Diclofenac and GSH in phosphate buffer were added to human liver microsomes (1 mg of protein/mL) suspended in phosphate buffer (0.1 M, pH 7.4) containing EDTA. Anti-P450 2C9 antibody or anti-P450 3A4 antibody was preincubated with microsomes for 30 min at room temperature. Control incubations contained preimmune IgG. Reactions were initiated by adding the substrate and another portion of NADPH and proceeded for an additional 30 min. The products were analyzed by LC/MS/MS. b The substrate concentrations were chosen such that they were similar to the reported Km values for P450-catalyzed metabolism of diclofenac. c ND, not detected in control incubations. d There was no difference observed between incubations containing human liver microsomes (control) and those containing no protein (blank) in terms of producing M1 and M3.

P450 2C9 antibody (Table 2). The formation of M1 and M3 from 50 µM 5-hydroxydiclofenac was not inhibited by the anti-P450 3A4 antibody (Table 2).

Scheme 1. Cytochrome P450-Mediated Bioactivation of Diclofenac Leading to the Formation of GSH-Conjugated Metabolites

Discussion It has been proposed that the P450-mediated bioactivation of diclofenac in rats entails two pathways, namely, (i) 4′-hydroxylation followed by further dehydrogenation to putative diclofenac-1′,4′-quinone imine and (ii) 5-hydroxylation followed by dehydrogenation to diclofenac2,5-quinone imine (Scheme 1) (12). In this study, LC/MS/ MS data were obtained which indicate that reactive benzoquinone imine intermediates, trapped in the form of their GSH conjugates, were formed in incubations involving human liver microsomes. In these experiments, formation of the GSH conjugates was found to be substrate concentration-dependent. Thus, when the diclofenac concentration was below 50 µM, M2 was the predominant product, generated primarily by P450 2C9 catalysis. This conclusion was based on the observations that the level of M2 production from diclofenac or from 4′-hydroxydiclofenac was reduced to less than 10-30% of control values in those incubations containing sulfaphenazole, a P450 2C9-specific inhibitor, or an anti-P450 2C9 antibody. Among several recombinant P450 enzymes tested, P450 2C9 was the only isoform that catalyzed the conversion of diclofenac or 4′hydroxydiclofenac to M2. These results are in accord with literature reports that conclude that P450 2C9 is involved in the 4′-hydroxylation of diclofenac (14-16) and further suggest that the conversion of 4′-hydroxydiclofenac to M2 also is catalyzed by P450 2C9. Formation of the GSH-conjugated metabolites of diclofenac requires three steps, i.e., (a) aromatic hydroxylation, (b) further oxidation of the resulting phenols to benzoquinone imine intermediates, and (c) conjugation with GSH. The first two steps are catalyzed by P450 enzyme(s). If a single P450 isoform was responsible for the two-step oxidation, and GSH conjugation with the reactive intermediate was not rate-limiting, the extent of production of the GSH conjugate(s) should correlate linearly with the activity of that P450 enzyme. Such a relationship was observed for the formation of M2. Upon examination of 10 individual human liver samples, a good correlation was established between the extent of formation of M2 and P450 2C9 activities (r ) 0.93) in those liver microsomal preparations. The formation of M1 and M3, on the other hand, was not a major pathway at diclofenac concentrations of 50 µM) required for the formation of M1 and M3, catalyzed by P450 3A4 and possibly by other P450 enzymes. In humans, peak plasma concentrations (Cmax) of diclofenac vary from 1.4 to 17 µM following a single oral dose of 100 mg (23-25). These values are in the range of the in vitro substrate concentrations where the biotransformation of diclofenac was dominated by the P450 2C9mediated pathway leading to M2. The amount of 4′hydroxydiclofenac, a precursor of M2, excreted in the urine of patients receiving 150 mg of diclofenac was ∼3fold higher than that of 5-hydroxydiclofenac (26-28). It may be argued, therefore, that although both P450s 2C9 and 3A4 are involved in metabolic activation of diclofenac, P450 2C9 is capable of producing reactive metabolite(s) at lower drug concentrations which may be more relevant clinically. However, the role of P450 3A4, and possibly other P450 isoforms, in the metabolism of diclofenac might become important in individuals with diminished P450 2C9 activity or who are exposed to higher hepatic drug concentrations. Among 10 human livers examined in this study, metabolite M2 was the predominant product in all cases at a substrate concentration of 10 µM, suggesting that the P450 2C9-mediated bioactivation of diclofenac probably is a common process in humans. It may be speculated that diclofenac-mediated hepatotoxicity might occur in those patients who have compromised defenses, due to either genetic or environmental factors, against reactive xenobiotic species. In summary, diclofenac has been shown to undergo P450-catalyzed metabolism in human liver preparations to reactive intermediates which can be trapped by GSH conjugation. At lower drug concentrations, this metabolic activation is catalyzed mainly by P450 2C9 via 4′hydroxylation. The metabolic activation at higher drug concentrations is mediated by P450s 2C9 and 3A4 which catalyze 4′- and 5-hydroxylation, respectively. It is interesting to note that GSH adduct formation from 5-hydroxydiclofenac does not require enzymatic catalysis, suggesting that once formed, 5-hydroxydiclofenac is susceptible to autoxidation to the benzoquinone imine derivative. Whether this relative high reactivity of 5-hydroxydiclofenac is related to the drug-induced toxicity remains to be investigated. However, on the basis of the results of this investigation, it would appear that further studies on the roles of P450 2C9 and 3A4 in diclofenac hepatotoxicity are warranted.

Tang et al.

Acknowledgment. We thank Ms. Deborah Newton (Merck Research Laboratories) for isolation of human liver microsomes and determination of P450 activities in those preparations and Dr. Anthony Y. H. Lu (Rutgers University) for valuable discussions. Supporting Information Available: Table S1 containing information on human liver donors and Figure S2 depicting inhibition of the biotransformation of diclofenac to M2 by sulfaphenazole in human liver microsomal incubations. This information is available free of charge via the Internet at http:// pubs.acs.org.

References (1) Breen, E. G., McNicholl, J., Cosgrove, E., McCabe, J., and Stevens, F. M. (1986) Fatal hepatitis associated with diclofenac. Gut 27, 1390-1393. (2) Purcell, P., Henry, D., and Melville, G. (1991) Diclofenac hepatitis. Gut 32, 1381-1385. (3) Helfgott, S. M., Sandberg-Cook, J., Zakim, D., and Nestler, J. (1990) Diclofenac-associated hepatotoxicity. JAMA, J. Am. Med. Assoc. 264, 2660-2662. (4) Banks, A. T., Zimmerman, H. J., Ishak, K. G., and Harter, J. G. (1995) Diclofenac-associated hepatotoxicity: Analysis of 180 cases reported to the food and drug administration as adverse reactions. Hepatology 22, 820-827. (5) Boelsterli, U. A., Zimmerman, H. J., and Kretz-Rommel, A. (1995) Idiosyncratic liver toxicity of nonsteroidal antiinflammatory drugs: Molecular mechanisms and pathology. Crit. Rev. Toxicol. 25, 207-235. (6) Pumford, N. R., Myers, T. G., Davila, J. C., Highet, R. J., and Pohl, L. R. (1993) Immunochemical detection of liver protein adducts of the nonsteroidal antiinflammatory drug diclofenac. Chem. Res. Toxicol. 6, 147-150. (7) Hargus, S. J., Amouzedeh, H. R., Pumford, N. R., Myers, T. G., McCoy, S. C., and Pohl, L. R. (1994) Metabolic activation and immunochemical localization of liver protein adducts of the nonsteroidal anti-inflammatory drug diclofenac. Chem. Res. Toxicol. 7, 575-582. (8) Kretz-Rommel, A., and Boelsterli, U. A. (1993) Diclofenac covalent protein binding is dependent on acyl glucuronide formation and is inversely related to P450-mediated acute cell injury in cultured rat hepatocytes. Toxicol. Appl. Pharmacol. 120, 155-161. (9) Gil, M. L., Ramirez, M. C., Terencio, M. C., and Castell, J. V. (1995) Immunochemical detection of protein adducts in cultured human hepatocytes exposed to diclofenac. Biochim. Biophys. Acta 1272, 140-146. (10) Kretz-Rommel, A., and Boelsterli, U. A. (1994) Mechanism of covalent adduct formation of diclofenac to rat hepatic microsomal proteins: Retention of the glucuronic acid moiety in the adduct. Drug Metab. Dispos. 22, 956-961. (11) Shen, S., Hatgus, S. J., Martin, B. M., and Pohl, L. R. (1997) Cytochrome P4502C11 is a target of diclofenac covalent binding in rats. Chem. Res. Toxicol. 10, 420-423. (12) Tang, W., Stearns, R. A., Bandiera, S. M., Zhang, Y., Raab, C., Braun, M. P., Dean, D. C., Pang, J., Leung, K. H., Doss, G. A., Strauss, J. R., Kwei, G. Y., Rushmore, T. H., Chiu, S.-H. L., and Baillie, T. A. (1999) Studies on hepatic cytochrome P450-mediated bioactivation of diclofenac in rats and in human hepatocytes: Identification of glutathione conjugated metabolites. Drug Metab. Dispos. (in press). (13) Shen, S., Marchick, M. R., Davis, M. R., and Pohl, L. R. (1997) 5-Hydroxylation of diclofenac by human cytochrome P4503A4: Possible role in idiosyncratic hepatotoxicity. Proceedings of 6th European ISSX Meeting, p 102, International Society for the Study of Xenobiotics, Cabin John, MD. (14) Leemann, T., Transon, C., and Dayer, P. (1993) Cytochrome P450TB (CYP2C): A major monooxygenase catalyzing diclofenac 4′-hydroxylation in human liver. Life Sci. 52, 29-34. (15) Wong, B. K., Goldstein, J., Spielberg, S. P., and Rushmore, T. H. (1996) Characterization of an in vitro model for assessing polymorphic metabolism by CYP2C9: Involvement in the mechanism of deficient Losartan metabolism. Proceedings of 7th North American ISSX Meeting, p 236, International Society for the Study of Xenobiotics, Cabin John, MD. (16) Transon, C., Lecoeur, S., Leemann, T., and Beaune, P. (1996) Interindividual variability in catalytic activity and immunoreactivity of three major human liver cytochrome P450 isozymes. Eur. J. Clin. Pharmacol. 51, 79-85.

Metabolic Activation of Diclofenac (17) Wang, R. W., and Lu, A. Y. H. (1997) Inhibitory anti-peptide antibody against human CYP3A4. Drug Metab. Dispos. 25, 762767. (18) Raucy, J. L., and Lasker, J. M. (1991) Isolation of P450 enzymes from human liver. In Methods in Enzymology (Waterman, M. R., and Johnson, E. F., Eds.) Vol. 206, pp 557-587, Academic Press, San Diego. (19) Newton, D. J., Wang, R. W., and Lu, A. Y. H. (1995) Cytochrome P450 inhibitors: Evaluation of specifities in the in vitro metabolism of therapeutic agents by human liver microsomes. Drug Metab. Dispos. 23, 154-158. (20) Shimada, T., Yamazaki, H., Mimura, M., Inui, Y., and Guengerich, F. P. (1994) Interindividual variations in human liver cytochrome P-450 enzymes involved in the oxidation of drugs, carcinogens and toxic chemicals: Studies with liver micromes of 30 Japanese and 30 Caucasians. J. Pharmacol. Exp. Ther. 270, 414-423. (21) Schellens, J. H., Ghabrial, H., van der Wart, H. H., Bakker, E. N., Wikinson, G. R., and Breimer, D. D. (1991) Differential effects of quindine on the disposition of nifedipine, sparteine and mephenytoin in humans. Clin. Pharmacol. Ther. 50, 520-528. (22) Bowles, S. K., Reeves, R. A., Cardozo, L., and Edwards, D. J. (1993) Evaluation of the pharmacokinetic and pharmacodynamic interaction between quinidine and nifedipine. J. Clin. Pharmacol. 33, 727-731.

Chem. Res. Toxicol., Vol. 12, No. 2, 1999 199 (23) Bliesath, H., Huber, R., Steinijans, V. W., Koch, H. J., Wurst, W., and Mascher, H. (1996) Lack of pharmacokinetic interaction between pantoprazole and diclofenac. Int. J. Clin. Pharmacol. Ther. 34, 152-156. (24) Miller, R. B. (1993) High-performance liquid chromatographic determination of diclofenac in human plasma using automated column switching. J. Chromatogr. 616, 283-290. (25) Davies, N. M., and Anderson, K. E. (1997) Clinical pharmacokinetics of diclofenac: Therapeutic insights and pitfalls. Clin. Pharmacokinet. 33, 184-312. (26) Stierlin, H., Faigle, J. W., Sallmann, A., and Kung, W. (1979) Biotransformation of diclofenac sodium (Voltaren) in animals and in man, I: Isolation and identification of pricipal metabolites. Xenobiotica 9, 601-610. (27) Stierlin, H., and Faigle, J. W. (1979) Biotransformation of diclofenac sodium (Voltaren) in animals and in man, II: Quantitative determination of the unchanged drug and principal phenolic metabolites, in urine and bile. Xenobiotica 9, 611-621. (28) Sawchuk, R. J., Maloney, J. A., Cartier, L. L., Rackley, R. J., Chan, K. K. H., and Lau, H. S. L. (1995) Analysis of diclofenac and four of its metabolites in human urine by HPLC. Pharm. Res. 12, 756-762.

TX9802217