Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Cytochrome P450 (P450)¹ comprises a superfamily of enzymes involved in the oxidation of a great number of exogenous and endogenous compounds (Guengerich, 1995). In human livers, CYP3A4 is the major P450 enzyme, followed by CYP2C9 and CYP1A2 (Shimada et al., 1994). Other P450 enzymes in the CYP1, CYP2, and CYP3 families are relatively minor forms but can play important roles in metabolism of a variety of drugs. There are large interindividual variations in the contents and activities of several P450 forms and these variations lead to different roles of P450 in the oxidation of some substrates (Yamazaki et al., 1999b; Nakajima et al., 1999).

DPH (phenytoin, 5,5-diphenylhydantoin) is widely used as an anticonvulsant drug, showing nonlinearity in its blood concentrations in humans (Odani et al., 1997). It has been reported that DPH can show teratogenicity (Wells et al., 1989) and hypersensitivity reactions, including hepatitis (Haruda, 1997), and an arene oxide intermediate has been considered to be involved in its toxicity (Martz et al., 1977). There are many drug-drug interactions associated with DPH, including cases in which other drugs modify the pharmacokinetics of DPH as well as cases where DPH alters those of other drugs (Nation et al., 1990a,b). In cases of the drugs such as itraconazole, cyclosporine, and theophylline, these drug interactions would arise because DPH is a potent inducer of P450s (Adebayo, 1988; D'Souza et al., 1988; Ducharme et al., 1995). Therefore, DPH may increase the clearance of such drugs and lead to decrease their effects. With its narrow therapeutic range, drug interactions leading to alterations in plasma DPH concentration may be clinically important. Dexamethasone and rifampicin may increase the metabolism of DPH (Kay et al., 1985; Lackner, 1991), whereas other drugs (e.g., ticlopidine, taclolimus, and amiodarone) may decrease DPH metabolism (Nolan et al., 1990; Thompson, 1996; Klaasssen, 1998), leading to changes in the plasma concentration. However, some mechanisms of these interactions are still unknown (Klaasssen, 1998).

DPH metabolism has been well studied. Four oxidative metabolites, 4'-HPPH, 3'-HPPH, 3',4'-diHPPH, and 3',4'-dihydrodiol, are reported in humans (Fig. 1) (Maguire, 1988; Szabo et al., 1990). It has been suggested that DPH is mainly oxidized to 4'-HPPH by CYP2C9 and to a minor extent by CYP2C19 (Bajpai et al., 1996). Yasumori et al. (1999) have reported that the rate of microsomal formation of 4'-HPPH is approximately 3-fold higher than that of 3'-HPPH, but roles of human P450s other than CYP2C9 or CYP2C19 have not been reported in 4'-HPPH formation and the P450s responsible for formation of other metabolites are still unclear. Munns et al. (1997) demonstrated that 3',4'-diHPPH can be oxidized to semiquinone and quinone derivatives, which may covalently bind to microsomal proteins and lead to P450 inactivation or produce autoantibodies. The formation of products other than 4'-HPPH must be considered with regard to DPH metabolism and toxicity.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Proposed pathways of DPH metabolism in humans.

The present study was, therefore, undertaken to determine which P450 enzymes are most active in the primary and secondary hydroxylations of DPH in humans. We determined the formation activities of four oxidative metabolites of DPH. Different roles of P450s in individual human liver microsomes in 3',4'-diHPPH formation were investigated, along with the enhancement by liver cytosol.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. DPH was obtained from Wako Pure Chemicals (Osaka, Japan) and 4'-HPPH and 3'-HPPH were from Aldrich (Milwaukee, WI). Racemic mephenytoin, -naphthoflavone, and orphenadrine were purchased from Sigma (St. Louis, MO) and ketoconazole, furafylline, sulfaphenazole, and fluvoxamine were from Ultrafine Chemicals (Manchester, UK). TCPO was kindly provided by Dr. Tsutomu Shimada, Osaka Prefectural Institute of Public Health. S-Mephenytoin, testosterone, 7-ethoxyresorufin, tolbutamide, paclitaxel, and their metabolites and reagents used in this study were obtained from sources described previously or were of highest qualities commercially available (Nakajima et al., 1999; Yamazaki et al., 1999b).

Enzyme Preparations. Human liver microsomes were prepared in 10 mM Tris-HCl buffer (pH 7.4) containing 0.10 mM EDTA and 20% (v/v) glycerol as described previously (Yamazaki et al., 1999b). Liver samples HL-1, -3, -4, -5, -6, -9, and -10 correspond to those designated elsewhere as HL-104, -110, -111, -114, -127, -134, and -136 (Guengerich, 1995) and HL-C13, -C6, -C15, -C11, -C16, -C18, and -C19 (Shimada et al., 1999), respectively. Microsomal samples HL-1, -3, -4, and -10 contained total spectrally determined P450 levels (nmol/mg microsomal protein) of 0.35, 0.53, 0.32, and 0.45, respectively. Microsomal sample HL-1 had CYP2B6, CYP2C9, CYP2C19, and CYP3A4 levels of 9, 10, 1.2, and 26% total P450, respectively, as judged by immunoblot analysis. Sample HL-3 had CYP2B6, CYP2C9, CYP2C19, and CYP3A4 levels of 1.6, 12, 0.7, and 73% total P450, respectively; sample HL-4 contained CYP2B6, CYP2C9, CYP2C19, and CYP3A4 levels of 0.5, 21, 3.4, and 14%, respectively; and sample HL-10 had CYP2C9, CYP2C19, and CYP3A4 levels of 16, 1.3, and 46% of total P450, respectively. Information about CYP2B6 content in sample HL-10 was not available. Escherichia coli membranes expressing recombinant human P450/NPR were prepared as described previously (Yamazaki et al., 1999a) with cDNAs of CYP1A2, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1, 3A4, and 3A5 introduced (Parikh et al., 1997; Gillam et al., 1999). Recombinant human P450 enzymes were purified from membranes (Guengerich et al., 1996) and NPR and b₅ were from rabbit liver microsomes (Strittmatter et al., 1978; Guengerich et al., 1981). Rat liver microsomal epoxide hydrolase was prepared as described elsewhere (Guengerich et al., 1979). Anti-rat CYP2C13 and anti-rat CYP3A2 IgG fraction were obtained from Daiichi Pure Chemicals (Tokyo, Japan) and used for immunoinhibition experiments with human liver microsomes.

Enzyme Assays. DPH, 4'-HPPH, and 3'-HPPH hydroxylation activities were determined according to methods described elsewhere (Doecke et al., 1990; Munns et al., 1997; Yasumori et al., 1999) with slight modifications. The standard incubation mixture (final volume of 0.25 ml) contained human liver microsomes (1.0 mg protein/ml) and liver cytosol (5.0 mg protein/ml), 50 mM potassium phosphate buffer (pH 7.4), an NADPH-generating system consisting of 0.5 mM NADP⁺, 5 mM glucose 6-phosphate, and 0.5 unit glucose-6-phosphate dehydrogenase/ml, and DPH (100 µM). In some cases, membranes containing recombinant P450 (0.20 µM) and P450 reductase were used as an enzyme source with addition of NPR membranes and purified rabbit b₅ to the final molar ratio of P450:NPR:b₅ of 1:2:1. In reconstituted systems, recombinant human P450 (0.20 µM); NPR (0.80 µM); b₅ (0.20 µM); a phospholipid mixture (20 µg/ml) consisting of a 1:1:1 mixture (by mass) of L--dilauroyl-sn-glycero-3-phosphocholine, L--dioleoyl-sn-glycero-3-phosphocholine, and phosphatidyl serine; and sodium cholate (0.25 mM) were used. P450 inhibitors were dissolved in CH₃OH (with the exception of -naphthoflavone, which was dissolved in dimethyl sulfoxide) and the final concentration of solvent in the incubation mixture was <1.0%. Incubations were carried out at 37°C for 30 min and terminated by adding 2 ml of methyl-tert-butyl ether, 0.3 M NaCl, and 4 µM mephenytoin as an internal standard. The reaction mixture was extracted twice with methyl-tert-butyl ether. After centrifugation at 900g for 10 min, the organic phase was evaporated to dryness under a gentle N₂ stream. The residue was dissolved in a mixture of 20% CH₃CN:0.05% HCO₂H (v/v). Product formation was determined by HPLC with a C18 (5-µm) analytical column (150 × 4.6-mm i.d., Mightysil RP-18; Kanto Chemical, Tokyo, Japan) at 35°C. Elution was with a mixture of 20% CH₃CN:0.05% HCO₂H (v/v) at a flow rate of 1.5 ml/min with detection at 214 nm. Retention times of the products were as follows: 3',4'-dihydrodiol, 3.9 min; 3',4'-diHPPH, 4.7 min; 4'-HPPH, 6.9 min; and 3'-HPPH, 8.7 min. Assignment of 3',4'-dihydrodiol and 3',4'-diHPPH peaks were made by liquid chromatography/mass spectrometry analysis (H. Yamazaki, T. Komatsu, M. Saeki, Y. Minami, Y. Kawaguchi, M. Nakajima, and T. Yokoi, unpublished data). We present the activities for formation of 3',4'-diHPPH and 3',4'-dihydrodiol on the basis of chromatographic response using 4'-HPPH as a standard. In these assay conditions, the formation of 4'-HPPH at 100 µM DPH increased linearly for up to 60 min of incubation time and to 2.0 mg of microsomal protein per milliliter. In the presence of liver cytosol, 4'-HPPH formation from DPH was increased linearly up to 60 min. Results presented in this study were the means of duplicate determinations and the S.D. (ranges) in these values were less than 10% of the means.

Other Assays. Concentrations of P450 and b₅ (Omura and Sato, 1964) and protein (Lowry et al., 1951) were estimated as described. The contents of P450 enzymes in human liver microsomes were estimated by immunoblotting (Guengerich et al., 1982).

Statistical Analysis. Statistical analysis was performed by a computer program Instat (GraphPad Software, San Diego, CA) designed for Student's t test after the assumption of equal variance with an F test. The correlations between DPH oxidation activities and P450 levels or marker activities in different human liver microsomal preparations were analyzed using a linear regression analysis program (Instat; GraphPad Software). Kinetic analysis of substrate oxidations was estimated using a computer program (KaleidaGraph; Synergy Software, Reading, PA) designed for nonlinear regression analysis.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of Human Liver Cytosol on DPH Oxidation Activities Catalyzed by Human Liver Microsomes. DPH was incubated with human liver microsomes or cytosol and in combination in the presence of an NADPH-generating system. HPLC analysis revealed that unknown peaks from liver cytosol were observed, but no metabolites of DPH were detected (Fig. 2A). In Fig. 2B, 4'-HPPH formation was the predominant reaction catalyzed by human liver microsomes, and 3'-HPPH and 3',4'-diHPPH were detected to a minor extent. With addition of liver cytosol to liver microsomes (Fig. 2C), the formation of 3'-HPPH and 3',4'-diHPPH was increased. Under these assay conditions, no 3',4'-dihydrodiol could be detected.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Representative HPLC chromatograms of phenytoin metabolites catalyzed by human liver microsomes in the absence or presence of liver cytosol and by purified CYP2C19 in reconstituted systems. DPH (100 µM) was incubated at 37°C for 30 min in the presence of an NADPH-generating system with human liver cytosol (sample HL-10, 5.0 mg protein/ml) (A), with human liver microsomes (sample HL-10, 1.0 mg protein/ml) (B), or with human liver microsomes and liver cytosol (C). DPH (100 µM) was incubated as described above with purified CYP2C19 (0.20 µM) in reconstituted systems (D). Peak: 1, 3',4'-dihydrodiol; 2, 3',4'-diHPPH; 3, unknown peak associated with DPH, 4'-HPPH, or 3'-HPPH dissolved in this mobile phase; 4, 4'-HPPH; 5, 3'-HPPH; 6, mephenytoin (internal standard); 7, DPH.

DPH Oxidation Activities Catalyzed by Seven Human Liver Microsomal Samples in the Absence or Presence of Liver Cytosol. Formation of 4'-HPPH, 3'-HPPH, and 3',4'-diHPPH was determined in the absence or presence of individual liver cytosol with seven human liver samples (Fig. 3). When DPH was used as a substrate with liver microsomes, the major product was 4'-HPPH. 3'-HPPH formation was detected only with HL-10 liver microsomes and increased by addition of liver cytosol. In all samples, 3',4'-diHPPH formation activities in liver microsomes were enhanced by addition of cytosol. Ratios of 3',4'-diHPPH formation (to total products) were in the range of 7% (HL-10) to 43% (HL-5) in the presence of liver cytosol. Under these assay conditions, no 3',4'-dihydrodiol could be detected.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   4'-HPPH (A), 3'-HPPH (B), and 3',4'-diHPPH (C) formation activities from DPH catalyzed by human liver microsomes of seven samples in the absence () or presence () of liver cytosol. DPH (100 µM) was incubated at 37°C for 30 min with human liver microsomes (1.0 mg protein/ml) and liver cytosol (5.0 mg protein/ml) in the presence of an NADPH-generating system. Results are presented as means of duplicate determinations. ND, not detected (<0.1 pmol/min/mg protein).

View larger version (30K): [in this window] [in a new window]   Fig. 4.   3',4'-DiHPPH formation activities from 4'-HPPH or 3'-HPPH catalyzed by human liver microsomes of seven samples in the absence () or presence () of liver cytosol. 4'-HPPH (A, 1 µM; B, 10 µM; and C, 100 µM) or 3'-HPPH (D; 100 µM) was incubated at 37°C for 30 min with human liver microsomes (1.0 mg protein/ml) and liver cytosol (5.0 mg protein/ml) in the presence of an NADPH-generating system. Results are presented as means of duplicate determinations.

View this table: [in this window] [in a new window]   TABLE 2 Correlation coefficients (r) between oxidation activities of DPH, 4'-HPPH, and 3'-HPPH and P450 contents or typical drug oxidation activities in seven human liver microsomes Drug oxidation activities and P450 contents in these human liver microsomes were reported previously (Yamazaki et al., 1999b).

Metabolites Formation Activities Catalyzed by Recombinant Human P450 Enzymes Expressed in Escherichia coli Membranes. Ten forms of recombinant human P450 enzymes expressed in E. coli membranes with human NPR were used to elucidate which P450 forms are active in catalyzing DPH oxidation. 3',4'-DiHPPH formation from DPH could not be measured because peaks from membrane components overlapped with the peak of 3',4'-diHPPH and the level of production of this secondary metabolite was too low to detect above background. When DPH (100 µM) was incubated with P450/NPR membranes (0.20 µM P450) in the presence of an NADPH-generating system, CYP2C9 and CYP2C19 were highly active in formation of 4'-HPPH, and CYP2D6 and CYP2B6 showed weak activities (Fig. 5A). 3'-HPPH formation could not be observed with P450/NPR membranes. However, 3',4'-dihydrodiol formation, which was little observed with human liver microsomes, was efficiently catalyzed by CYP1A2, followed by CYP2C19 and CYP2E1 (Fig. 5B). The addition of purified rat epoxide hydrolase to P450/NPR membranes had no effect on 4'-HPPH or 3',4'-dihydrodiol formation (data not shown).

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Product formation from DPH, 4'-HPPH, or 3'-HPPH catalyzed by recombinant human P450 enzymes. DPH (100 µM) was incubated at 37°C for 30 min with P450/NPR membranes (0.20 µM P450) in the presence of an NADPH-generating system and the formation of 4'-HPPH (A) and 3',4'-dihydrodiol (B) are shown. 4'-HPPH (C) or 3'-HPPH (D) (100 µM) was incubated as indicated above. Results are presented means of duplicate determinations. ND, not detected (<1 pmol/min/nmol P450 for 4'-HPPH and 3',4'-dihydrodiol and <5 pmol/min/nmol P450 for 3',4'-diHPPH).

DPH and 4'-HPPH Oxidation Activities in Reconstituted Systems Containing Purified Human P450 Enzymes. Because 3',4'-diHPPH formation from DPH catalyzed by P450/NPR membranes could not be observed due to chromatographic interference, further studies were performed to determine the rates of 3',4'-diHPPH formation from DPH with purified recombinant human P450 enzymes (Table 3). The enzymes used were CYP1A2, CYP2C9, CYP2C19, and CYP3A4, which showed relatively high activities with P450/NPR membranes. As described above, CYP2C9 and CYP2C19 also had high catalytic activities in 4'-HPPH formation from DPH and CYP1A2 catalyzed 3',4'-dihydrodiol formation. 3'-HPPH formation activities by CYP2C19, CYP2C9, and CYP1A2 were low. 3',4'-DiHPPH formation from DPH was mainly catalyzed by CYP2C19, and to a minor extent, by purified CYP1A2. As indicated in Fig. 2D, CYP2C19 catalyzed the formation of all four DPH metabolites.

Effects of Chemical Inhibitors and Anti-P450 Antibodies on DPH Oxidation Catalyzed by Different Human Liver Microsomes in the Presence of Liver Cytosol. Effects of chemical inhibitors on microsomal 4'-HPPH formation were examined in the presence of liver cytosol (Fig. 6), using 100 µM DPH. Sulfaphenazole, an inhibitor of CYP2C9 (Mancy et al., 1996), inhibited 4'-HPPH formation in all human liver microsomes and fluvoxamine, an inhibitor of CYP2C19 and CYP1A2 (Jeppesen et al., 1996; Yamazaki et al., 1997), also inhibited the activities by approximately 50% (Fig. 6, A-D). Furafylline, an inhibitor of CYP1A2 (Tassaneeyakul et al., 1994), and ketoconazole, an inhibitor of CYP3A4 (Baldwin et al., 1995), showed no or weak inhibition. Orphenadrine, a partially selective inhibitor of CYP2B6 (Reidy et al., 1989), inhibited the activity in HL-1 microsomes by approximately 25% (Fig. 6A).

View larger version (64K): [in this window] [in a new window]   Fig. 6.   Effects of P450 inhibitors on DPH oxidation catalyzed by individual human liver microsomes in the presence of liver cytosol. DPH (100 µM) or 4'-HPPH (10 µM) was incubated at 37°C for 30 min with human liver microsomes (1.0 mg protein/ml) and liver cytosol (5.0 mg protein/ml) in the absence or presence of chemical inhibitors. Values in parentheses indicated concentrations of chemical inhibitors (µM). Results are presented as means of duplicate determinations and indicate percentage of control without inhibitor. A-D, control activities of 4'-HPPH formation from DPH (pmol/min/mg microsomal protein) were 9.7 (A, HL-1), 8.3 (B, HL-3), 7.3 (C, HL-4), and 12.7 (D, HL-10), respectively. E-H, control activities of 3',4'-diHPPH formation (from 4'-HPPH) were 6.1 (E, HL-1), 5.2 (F, HL-3), 3.9 (G, HL-4), and 5.8 (H, HL-10), respectively. N.D., not determined (because furafylline interfered with the 3',4'-diHPPH peak).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

It has been suggested that the anticonvulsant drug DPH is mainly metabolized to 4'-HPPH by CYP2C9 (Bajpai et al., 1996), but other metabolites have not been studied extensively. Therefore, the rates of formation of four oxidative metabolites, 4'-HPPH, 3'-HPPH, 3',4'-diHPPH, and 3',4'-dihydrodiol, from DPH were determined in human liver microsomes. 4'-HPPH formation was the major product, followed by 3',4'-diHPPH and 3'-HPPH. Although 3',4'-dihydrodiol (~1-10% of administered DPH) has been detected in human urine as a minor metabolite (Maguire, 1988; Szabo et al., 1990), it was not detected under these assay conditions with human liver microsomes and/or liver cytosol (Fig. 2).

Rates of DPH metabolites formation except 4'-HPPH were enhanced by addition of liver cytosol under conditions in which linearity for 4'-HPPH formation was observed. Although 4'-HPPH formation was not affected by addition of liver cytosol, it was enhanced by the addition of BSA or rabbit serum (Table 1). It has been reported that the addition of 4% BSA decreased K_m values of 4'-HPPH formation 20-fold (Ludden et al., 1997). This trend was also seen with 4'-HPPH formation upon the addition of liver cytosol. 3'-HPPH and 3',4'-diHPPH formation activities were enhanced to a large extent in the presence of liver cytosol. Heat stable enzymes might be involved in this enhancement because BSA or rabbit serum did not produce this effect and interindividual differences in the extent of the effects were observed (Table 1). Enhancing effects of cytosolic protein on microsomal P450 systems have been reported previously (Mori et al., 1984) and may proceed by a variety of mechanisms, including effects on microsomal enzymes, effects on stability of substrate and/or metabolites, and enhanced affinity of substrate to microsomal P450 enzymes. Our results strongly indicate that studies of DPH metabolism with human liver microsomes should be examined in the presence of individual liver cytosols.

In covalent binding studies of DPH reported previously (Roy and Snodgrass, 1988, 1990), glutathione [0.5-5 mM, an inhibitor that can trap epoxide intermediate(s)], and TCPO (1-2 mM, an inhibitor for epoxide hydrolase) have been used. Although microsomal 4'-HPPH formation was decreased by glutathione and increased by TCPO, 3',4'-dihydrodiol could not be detected. When epoxide hydrolase was added to the incubation mixture with P450/NPR membranes, 4'-HPPH and 3',4'-dihydrodiol formation were not affected (data not shown). We conclude that an arene oxide intermediate would have little contribution in DPH oxidation or that, if formed, it rearranges rapidly to the phenol.

It has been reported that 4'-HPPH formation is catalyzed by CYP2C9 and CYP2C19 (Yasumori et al., 1999). The substrate concentration of DPH used in this study was 100 µM, based on the clinical DPH plasma concentration of 40 to 80 µM (Bajpai et al., 1996). The major metabolite was 4'-HPPH and its formation correlated well with tolbutamide methyl hydroxylation activities and CYP2C9 contents, as reported previously (Doecke et al., 1991; Veronese et al., 1991). Sulfaphenazole showed the most effective inhibition on activities of all chemical inhibitors used, followed by fluvoxamine and -naphthoflavone (Fig. 6, A-D). Of 10 forms of recombinant human P450 enzymes (P450/NPR membranes and reconstituted systems) used (Fig. 5; Table 3), 4'-HPPH formation was catalyzed predominantly by CYP2C9 and CYP2C19. Although good correlation between DPH 4'-hydroxylation and ethoxyresorufin O-deethylation (or CYP1A2 contents) in human liver microsomes was observed (Table 2), the role of CYP1A2 in DPH 4'-hydroxylation is minor because of the small effect of furafylline and low activity of recombinant CYP1A2. In separate experiments, we confirmed that -naphthoflavone (100 µM) inhibited DPH 4'-hydroxylation catalyzed by recombinant CYP2C19 expressed in E. coli membranes by ~80% (data not shown).

Yasumori et al. (1999) reported that the rate of 3'-HPPH formation is approximately one-third of that of 4'-HPPH formation. In the present study, 3'-HPPH formation was very slow and detected only in sample HL-10. 3'-HPPH is rapidly hydroxylated to 3',4'-diHPPH by human liver microsomes or recombinant P450 enzymes (Figs. 4 and 5) causing little 3'-HPPH to accumulate in the reaction mixtures. 3'-HPPH formation was catalyzed by CYP2C19 and CYP2C9 (and CYP1A2 to a small extent), but the P450s responsible for 3'-HPPH formation could not be established with the combination of human liver microsomes and recombinant human P450 enzymes.

It has been shown that 3',4'-diHPPH may covalently bind to microsomal proteins after oxidation to semiquinone and quinone derivatives and lead to mechanism-based inactivation of the P450s that may be involved in initiation of a drug hypersensitivity reaction (Munns et al., 1997). Thus, we focused on 3',4'-diHPPH formation and investigated the roles of P450s in 3',4'-diHPPH formation from phenolic DPH products. When 4'-HPPH or 3'-HPPH was used as a substrate at 100 µM, 3',4'-diHPPH formation activities catalyzed by human liver microsomes were correlated with testosterone 6-hydroxylation activities and CYP3A4 contents. However, no correlations were observed between 3',4'-diHPPH formation (1 or 10 µM 4'-HPPH) and P450 levels or marker drug oxidation activities in liver microsomes (Table 2). These results indicate potential contributions of multiple P450 enzymes in 3',4'-diHPPH formation at low 4'-HPPH concentrations. Of the recombinant P450 enzymes examined, CYP2C19 was most highly active in 3',4'-diHPPH formation from phenols; CYP2C9 and CYP3A4 also catalyzed 3',4'-diHPPH formation. In human livers containing relatively high contents of CYP2C and low CYP3A4 (sample HL-4), fluvoxamine was the most effective inhibitor. However, 3',4'-diHPPH formation in sample HL-3 (high contents of CYP3A4 and low CYP2C) was inhibited by ketoconazole and enhanced by -naphthoflavone. In sample HL-10, fluvoxamine, ketoconazole, and sulfaphenazole all inhibited 3',4'-diHPPH formation to the same extent. These results support the contributions of CYP3A4, CYP2C9, and CYP2C19 in microsomal 3',4'-diHPPH formation. Different roles of human P450 enzymes in individual human liver microsomes were observed in drug oxidations such as azelastine N-demethylation (Nakajima et al., 1999) and troglitazone oxidation (Yamazaki et al., 1999b). In these cases, extrapolation of effects among individuals may be more complex.

In conclusion, our results suggest that microsomal DPH metabolism should be determined in the presence of individual liver cytosols. Rates of 3',4'-diHPPH formation were similar to formation of 4'-HPPH, a primary DPH metabolite. CYP2C19, CYP2C9, and CYP3A4 contribute significantly to 3',4'-diHPPH formation from primary hydroxylated metabolites. Roles of these P450s vary among different human liver samples having compositions of various P450 enzymes. These results suggest that 4'-HPPH formation and also other reactions (e.g., 3',4'-diHPPH formation) can be involved in nonlinearity of DPH plasma concentrations or adverse DPH reactions in humans.