It has been known that diclofenac is biotransformed into chemically reactive metabolites, which bind covalently to liver microsomal proteins, including cytochrome P450 enzyme(s). We have investigated the ability and selectivity of diclofenac to inactivate P450 enzymes. Preincubation of microsomes of untreated rats with diclofenac in the presence of NADPH resulted in time-dependent loss of testosterone 2- and 16-hydroxylation activities. No effect of the preincubation was observed on ethoxyresorufin O-deethylase, pentoxyresorufin O-depentylase, or testosterone 6-hydroxylation activity. The time-dependent decreases in testosterone 2- and 16-hydroxylation activities followed the pseudo-first order kinetics and were saturable with increasing diclofenac concentrations. Reduced glutathione was not capable of protecting against the decrease in the enzyme activities. These data establish that a mechanism-based inactivation of CYP2C11 occurs during the oxidative metabolism of diclofenac. The diclofenac concentrations required to achieve the half-maximal rate of inactivation (KI) were 3 to 4 µM, which were close to Km for the low-Km components for diclofenac 4'- and 5-hydroxylation activities (7.29 and 4.43 µM, respectively). Anti-CYP2C11 IgG inhibited diclofenac 4'- and 5-hydroxylation activities, indicating that CYP2C11 is a major isozyme responsible for these aromatic oxidations. The preincubation of microsomes with 4'- or 5-hydroxydiclofenac did not cause a decrease in testosterone 2- or 16-hydroxylation activity, suggesting that neither of the primary metabolites is a precursor of the metabolite that inactivates CYP2C11. Therefore, a highly reactive intermediate(s) inactivating CYP2C11, probably arene-oxide, appears to be generated during the process of diclofenac 4'- and/or 5-hydroxylation. Diclofenac metabolism in human liver microsomes did not cause inactivation of CYP2C9, a major isozyme involved in diclofenac 4'-hydroxylation. Because the human microsomes have high diclofenac 4'-hydroxylation but not 5-hydroxylation activity, importance of the latter pathway in the inactivation is suggested.
Diclofenac is one of the nonsteroidal anti-inflammatory drugs widely used clinically. In relation to the diclofenac-induced hepatotoxicity, extensive studies have focused on biotransformation of diclofenac into chemically reactive metabolites capable of binding covalently to liver macromolecules (Boelsterli et al., 1995). Some protein targets of the reactive metabolites have been identified in the liver of animals administrated the drug (Pumford et al., 1993; Hargus et al., 1995; Wade et al., 1997; Seitz et al., 1998). In vitro studies with hepatocytes also showed the formation of the reactive metabolites (Kretz-Rommel and Boelsterli, 1994b; Gil et al., 1995). UDP-glucuronosyltransferase and cytochrome P450 (CYP1) enzymes were shown to mediate the metabolic activation (Kretz-Rommel and Boelsterli, 1993, 1994a; Hargus et al., 1994).
Acyl glucuronide is a common metabolite of carboxylic acid drugs such as acidic nonsteroidal anti-inflammatory drugs, which is often demonstrated as a reactive metabolite of these drugs (Spahn-Langguth and Benet, 1992; Boelsterli et al., 1995). The reactivity and potential toxicity of acyl glucuronide products are widely recognized (Kretz-Rommel and Boelsterli, 1993, 1994a; Hargus et al., 1994). It is known that diclofenac is oxidized mainly into two phenolic metabolites, 4'-hydroxydiclofenac and 5-hydroxydiclofenac (Stierlin et al., 1979) (Fig. 1). Although quinone imine metabolite of 5-hydroxydiclofenac has been proposed as a reactive metabolite of diclofenac (Brune and Lindner, 1992), definitive proof for the chemical structure of reactive metabolite by CYP enzymes had not been available. Recent studies provided information about the chemical nature of the reactive metabolite and CYP enzymes involved in its formation. Tang et al. (1999a) found benzoquinone imines as their reduced glutathione (GSH) conjugates, which were formed from 4'- and 5-hydroxydiclofenac in rats and human hepatocytes. They also isolated the GSH conjugates of benzoquinone metabolites in incubations of human liver microsomes with diclofenac in the presence of NADPH and GSH (Tang et al., 1999b). Shen et al. (1999) reported that covalent binding of diclofenac to human liver microsomes was CYP3A4-dependent, and benzoquinone imine, a decomposition product of 5-hydroxydiclofenac, bound covalently to human liver microsomes. Bort et al. (1999) reported that N,5-dihydroxydiclofenac was also found as a further metabolite of 5-hydroxydiclofenac, which was proposed to contribute to the hepatotoxicity of diclofenac
If a product formed by CYP-dependent metabolism is highly reactive, it should bind to the site of formation in the enzyme, resulting in mechanism-based inactivation of the CYP enzyme. The diclofenac metabolism to generate the benzoquinone metabolites mentioned above was shown to be catalyzed by CYP2B, CYP2C, and CYP3A enzymes in rats (Tang et al., 1999a), whereas it has been proposed that CYP2C11 is a target of a reactive metabolite of diclofenac in rat liver microsomes (Shen et al., 1997). In the present study, to investigate the relationship between the abilities of CYP enzymes to activate diclofenac and their tendency to be targets of the metabolites formed from diclofenac, we tested the ability and selectivity of diclofenac to inactivate CYP enzymes, including CYP2B, CYP2C, and CYP3A, in rat liver microsomes. The chemical nature of the reactive metabolite involved in the inactivation is also discussed.
Materials and Methods;
Chemicals. Diclofenac sodium, GSH, and tolbutamide were purchased from Wako Pure Chemical (Osaka, Japan); testosterone, 2-hydroxytestosterone, and 16-hydroxytestosterone, ethoxyresorufin, and pentoxyresorufin were from Sigma (St. Louis, MO); 6-hydroxytestosterone was from Steraloids Inc. (Wilton, NH); resorufin and sodium phenobarbital were from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan); -naphthoflavone was from Aldrich (Milwaukee, WI); and hydroxytolbutamide was from Ultrafine Chemicals (Manchester, UK). 4'-Hydroxydiclofenac and 5-hydroxydiclofenac were gifts from Novartis Pharma AG (Basel, Switzerland). Glucose 6-phosphate (G-6-P), glucose 6-phosphate dehydrogenase (G-6-PDH), and NADPH were purchased from Oriental Yeast Co., Ltd. (Tokyo, Japan). All other chemicals and solvents used were of analytical grade.
Liver Microsomes and P450 Enzyme. Male Wistar rats (2-months old) were obtained from Takasugi Experimental Animals (Saitama, Japan). The animals were housed in an air-conditioned room (25°C) under a 12-h light/dark cycle for 1 week before use. Food (commercially available pellet; Oriental Yeast Co., Ltd.) and water were given ad libitum. Sodium phenobarbital (80 mg/kg in physiological saline) or -naphthoflavone (80 mg/kg in corn oil) was given to the rats intraperitoneally for 4 days. The rats were killed along with untreated rats by decapitation 24 h after the final doses, and liver microsomal fractions were prepared according to the method of Omura and Sato (1964). Protein concentrations were assayed by the method of Lowry et al. (1951). Recombinant CYP2C11 expressed in microsomes of insect cells infected with baculovirus containing rat NADPH-P450 reductase and human cytochrome b5, and human liver microsomes (pooled fraction from 12 patients) were purchased from GENTEST (Woburn, MA).
Preincubation of Liver Microsomes with Diclofenac and Its Metabolites. Liver microsomes of untreated or inducer-treated male Wistar rats, microsomes from insect cells expressing CYP2C11, and pooled human liver microsomes were preincubated with diclofenac in the presence of NADPH to determine effects of the metabolic intermediates on microsomal monooxygenase activities. A 1-ml incubation mixture contained microsomes (0.5 mg of rat liver microsomes, 10 pmol of CYP2C11, or 0.25 mg of human liver microsomes), 10 mM G-6-P, 2 units of G-6-PDH, 10 mM MgCl2, 0.1 mM EDTA, and various concentrations of diclofenac in 0.15 M potassium phosphate buffer, pH 7.4). In some experiments, primary metabolites of diclofenac were used instead of diclofenac. After temperature equilibration (37°C, 5 min), preincubation of microsomes with diclofenac was started by adding NADPH (final 0.5 mM) and performed for various time periods up to 15 min (30 min for human). The subsequent incubation of the microsomes for the assay of enzymatic activities was started by the addition of a test substrate, testosterone, ethoxyresorufin, pentoxyresorufin, or tolbutamide.
Testosterone 2-, 16-, and 6-hydroxylation activities of the preincubated microsomes were determined according to the high-performance liquid chromatography (HPLC) method previously described (Masubuchi et al., 1995) at the testosterone concentration of 50 µM. Ethoxyresorufin O-deethylase (EROD) and pentoxyresorufin O-depentylase (PROD) activities of the microsomes were assayed by the fluorometric method to determine the resorufin formation (Burke et al., 1985) at the substrate concentrations of 1 and 5 µM, respectively. Tolbutamide methyl-hydroxylation activity was determined according to the HPLC method previously described (Miners et al., 1988) at the tolbutamide concentration of 2 mM. Liver microsomes from untreated rats were used for assays of testosterone oxidation activities; those from phenobarbital-treated rats were for PROD activity assays; those from -naphthoflavone-treated rats were for EROD activity assays; and human liver microsomes were for tolbutamide hydroxylation activity. All of the assays were performed under linear conditions of metabolite formation with regard to incubation time and protein concentration.
Assay of Diclofenac 4'- and 5-Hydroxylation Activities. Diclofenac 4'- and 5-hydroxylation activities were assayed according to the HPLC method of Leemann et al. (1993) with modifications. A 1-ml incubation mixture contained 0.5 mg of liver microsomes, 10 mM G-6-P, 2 units of G-6-PDH, 5 mM MgCl2, and various concentrations of diclofenac (1-320 µM) in 0.15 M potassium phosphate buffer, pH 7.4). After temperature equilibration (37°C, 5 min), the reaction was started by adding NADPH (final 0.5 mM), and the incubation was performed for 2.5 min. The reaction was terminated by 1 M sodium phosphate buffer, pH 5.0 and then flurbiprofen was added to the mixture as an internal standard. Diclofenac and its metabolites were extracted into diethyl ether, the organic layer was evaporated to dryness, and the residue was dissolved in 0.1 ml of a mobile phase for the HPLC, which consists of 100 mM sodium phosphate buffer, pH 7.4, including 0.02% triethanolamine and acetonitrile (7:3 by vol). The sample was applied to a reversed phase column (Inertsil ODS; GL Sciences Ltd., Tokyo, Japan). The UV absorbance intensity of diclofenac metabolites was monitored at 282 nm.
Immunoinhibition of Diclofenac Metabolism by an Antibody against CYP2C11. A polyclonal antibody against CYP2C11 raised in a goat was obtained from Daiichi Pure Chemicals (Tokyo, Japan). In immunoinhibition studies, microsomes were preincubated with various amounts of the antibody or preimmune serum at 25°C for 30 min, followed by adding other components of the incubation mixture and assay of diclofenac 4'- and 5-hydroxylation activities.
Data Analysis. Enzyme kinetic parameters (Km, Vmax) were analyzed according to a nonlinear least-squares regression analysis based on a simplex method (Yamaoka et al., 1981). Best fittings of the data were performed by weighting them with the reciprocal of the square of the activity. Pseudo-first order kinetic constants for the enzyme inactivation (k) were calculated from the initial slopes of the linear regression lines of the semilogarithmic plots of the remaining enzyme activity against the preincubation time. The reciprocal of k thus obtained was plotted against the reciprocal of the diclofenac concentration and then a concentration required for a half-maximum inactivation (KI) for the inactivation and a maximum inactivation rate constant (kinact) were determined from the intercepts on the abscissa and the ordinate, respectively. Results were represented as means ± S.E. Statistical significance was calculated by the Student's t test. For experiment involving more than two experimental groups, the groups were compared by analysis of variance, followed by Newman-Keuls multiple comparison test to determine significant differences between the group means
A previous immunochemical study has demonstrated that CYP2C11 is one of the microsomal target proteins of covalent binding of the diclofenac reactive metabolite (Shen et al., 1997). It was demonstrated that diclofenac inactivated CYP2C11 in a mechanism-based manner according to the following observations (Figs. 2-4): 1) NADPH dependence for the inhibition, 2) pseudo-first order kinetics for the time-dependent inactivation, 3) saturability of inactivation with increasing diclofenac concentrations, and 4) lack of protection against the inhibition by GSH. Selectivity toward one particular CYP enzyme, CYP2C11 in this case, is characteristic of suicide inactivation. More rapid inactivation of CYP2C11, which was observed by using expressed CYP2C11 instead of liver microsomes (Fig. 7), supported the conclusion.
The diclofenac concentrations required to achieve the half-maximal rate of inactivation (KI) were 3 to 4 µM. Kinetic analysis for diclofenac 4'- and 5-hydroxylation activities indicated that both of the reactions were catalyzed by more than one enzyme system (Table 2). The KI value for the inactivation were close to Km values for the low-Km components for diclofenac 4'- and 5-hydroxylation activities (7.29 and 4.43 µM, respectively), suggesting that the pathway(s) is relevant to the inactivation of CYP2C11.
There can be two ways to generate the reactive metabolite(s). One possibility is that the process of diclofenac 4'- and/or 5-hydroxylation is directly involved in generation of a reactive intermediate(s). Arene-oxide is one of the proposed metabolic intermediates generated during the aromatic hydroxylations, which are highly reactive and are involved in enzyme inactivation. A second possibility is that further metabolites of 4'- and 5-hydroxydiclofenac, which include benzoquinones and hydroxylamine and have been already proposed in relation to the diclofenac hepatotoxicity (Brune and Lindner, 1992; Bort et al., 1999; Shen et al., 1999; Tang et al., 1999a,b). However, the preincubation of microsomes with 4'- or 5-hydroxydiclofenac instead of diclofenac did not cause decrease in testosterone 2- or 16-hydroxylation activity (Fig. 6), indicating that the proposed further metabolite is not responsible for the inactivation of CYP2C11. We found that both aromatic hydroxylations of diclofenac were mediated by CYP2C11 (Fig. 7). Thus, it is concluded that diclofenac inactivates CYP2C11 during the 4'- and/or 5-hydroxylation processes, probably via the arene-oxide formation.
Human liver microsomes have higher diclofenac 4'-hydroxylation activity compared with rats, whereas 5-hydroxylation activity was very low. There was no evidence for mechanism-based inactivation of CYP2C9, a major CYP isozyme responsible for diclofenac 4'-hydroxylation, during the diclofenac metabolism in human liver microsomes (Fig. 9). It is reasonable to postulate that the reactive metabolite relevant to inactivation of CYP enzymes is not generated during diclofenac 4'-hydroxylation. Thus, diclofenac 5-hydroxylation rather than 4'-hydroxylation seems to be closely related to formation of arene-oxide, a possible candidate to inactivate CYP2C11. However, the possibility could not be excluded that diclofenac 4'-hydroxylation was mediated by CYP2C11 and CYP2C9 via different intermediates.
In summary, diclofenac is demonstrated to be a selective and mechanism-based inactivator of CYP2C11. Formation of a chemically reactive metabolite of diclofenac that inactivates CYP2C11 was a low-Km reaction. The major pathways leading to aromatic hydroxylations, especially 5-hydroxylation, appear to be directly involved in the formation of the intermediate that binds to CYP2C11, resulting in loss of catalytic activity