Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

CIT is a new antidepressant of the selective serotonin reuptake inhibitor class (Hyttel, 1982). This drug is metabolized by N-demethylation to DCIT and didesmethylcitalopram, by deamination and further oxidation to a propionic acid derivative and by N-oxidation to CIT N-oxide (Oyehaug and Ostensen, 1984). N-Demethylation is the major metabolic pathway of CIT in humans (Oyehaug and Ostensen, 1984). The N-demethylated metabolites DCIT and didesmethylcitalopram are considered to be less potent than the parent compound as serotonin reuptake inhibitors (Hyttel, 1982).

(S)-Mephenytoin 4-hydroxylation shows a genetically determined polymorphism (Brøsen, 1990; Küpfer and Preisig, 1984; Wilkinson et al., 1989), and there are marked interethnic differences in the incidence of the PM phenotype; approximately 3 to 6% of Caucasian populations (Alván et al., 1990; Brøsen, 1990; Jacqz et al., 1988; Küpfer and Preisig, 1984; Wedlund et al., 1985; Wilkinson et al., 1989) and 13 to 23% of Oriental populations (Horai et al., 1989; Sohn et al., 1992) are PMs of (S)-mephenytoin 4-hydroxylation. The isoenzyme responsible for (S)-mephenytoin 4-hydroxylation has been shown to be CYP2C19 (Goldstein et al., 1994; Wrighton et al., 1993), and two mutations (m1 and m2) in the CYP2C19 gene have been described in Japanese PM subjects (de Morais et al., 1994a,b). The metabolism of several clinically important drugs used in psychoneuropharmacological treatment has been demonstrated to cosegregate with this genetically determined hydroxylation polymorphism (Dahl and Bertilsson, 1993; Spina and Capriti, 1994).

Sindrup et al. (1993) first reported in their human panel study that the metabolism of racemic CIT is at least partially under the pharmacogenetic control of (S)-mephenytoin 4-hydroxylase. In addition, we recently reported that racemic CIT competitively inhibits (S)-mephenytoin 4-hydroxylation in human liver microsomes, and we suggested that several selective serotonin reuptake inhibitors are substrates of CYP2C19 (Kobayashi et al., 1995). However, little information is available regarding the metabolism of CIT itself in human liver microsomes. In the present study, we investigated the N-demethylation of CIT in human liver microsomes and cDNA-expressed human P450s, to identify the principal isoform of P450s involved in this major metabolic pathway of CIT.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Drugs and chemicals. Racemic CIT hydrobromide and DCIT hydrochloride were kindly supplied by Lundbeck (Copenhagen-Valby, Denmark). Racemic mephenytoin and 4-hydroxymephenytoin were kind gifts from Dr. Küpfer (University of Berne, Berne, Switzerland). (S)- and (R)-Mephenytoin were separated from the racemic mixture of mephenytoin with a Chiralcel OJ column (10 µm, 4.6 × 250 mm; Daicel Chemical Co., Tokyo, Japan), as reported by Yasumori et al. (1990). Triazolam and its metabolites (- and 4-hydroxytriazolam) were supplied by Nihon Upjohn Co. (Tokyo, Japan). 17-Ethinylestradiol and ketoconazole were purchased from Sigma Chemical Co. (St. Louis, MO), and cyclobarbital was obtained from Tokyo Kasei Kogyo Co. (Tokyo, Japan). NADP⁺ and glucose-6-phosphate were purchased from Oriental Yeast Co. (Tokyo, Japan). Glucose-6-phosphate dehydrogenase was obtained from Boehringer Mannheim GmbH (Mannheim, Germany). Quinidine, acetonitrile and other reagents of analytical grade were purchased from Wako Pure Chemical Industries (Osaka, Japan).

Human liver microsomes. Human liver samples (n = 14) were obtained, as excess material removed during surgery on the liver, from Japanese patients who underwent partial hepatectomy at the Department of General Surgery, International Medical Center of Japan (Tokyo, Japan), as reported from our laboratory (Chiba et al., 1993; Kobayashi et al., 1995). All surgical procedures were performed for the removal of metastatic tumor(s) from the liver. The use of human liver tissue for this study was approved by the Institutional Ethics Committee of the International Medical Center of Japan. Less than 5 min passed between removal of the liver tissue and collection and freezing of samples in liquid nitrogen. The liver parenchymas of the non-tumor-bearing parts used in this study were shown later to be histopathologically normal in all cases. Liver samples obtained from patients with acute or chronic hepatitis, those with cirrhosis or those taking medications known to induce or inhibit the hepatic monooxygenase activity were not included in this study.

View this table: [in this window] [in a new window]   TABLE 1 Characteristics of the 14 human livers The R/S ratio was determined as the ratio of the rate of formation of 4-hydroxymephenytoin with 1 mM (R)-mephenytoin to that with 1 mM (S)-mephenytoin. Microsomal samples with a ratio of >0.7 were classified as those derived from the putative PM patients.

Assay with human liver microsomes. The basic incubation medium contained 0.1 or 0.2 mg/ml microsomes, 0.5 mM NADP⁺, 2.0 mM glucose-6-phosphate, 1 IU/ml glucose-6-phosphate dehydrogenase, 4 mM MgCl₂, 0.1 mM EDTA, 100 mM potassium phosphate buffer (pH 7.4) and 1 to 500 µM CIT, 100 µM (S)-mephenytoin or 25 µM triazolam, in a final volume of 250 µl. The mixture was incubated at 37°C for 30 min for CIT, 60 min for (S)-mephenytoin and 15 min for triazolam. All reactions were performed in the linear range with respect to protein concentration and incubation time. After the reaction was stopped by addition of 100 µl of cold acetonitrile, 50 µl of cyclobarbital (1.25 µg/ml in methanol) was added to the samples as an internal standard for assaying DCIT and 4-hydroxymephenytoin. To assay the two metabolites of triazolam (i.e., -hydroxytriazolam and 4-hydroxytriazolam), 50 µl of lorazepam (2.5 µg/ml in methanol) was added as an internal standard. The mixture was centrifuged at 10,000 × g for 5 min, and 50 or 100 µl of supernatant was injected into a HPLC system as described below.

HPLC conditions. The determination of DCIT was carried out by a modification of an HPLC method reported previously (Chiba et al., 1993). Briefly, the HPLC system consisted of a model L-6000 pump (Hitachi, Tokyo, Japan), a model L-4000 UV detector (Hitachi), a model AS-2000 autosampler (Hitachi), a model D-2500 integrator (Hitachi) and a 4.6 × 250 mm CAPCELL PAK C₁₈ UG120 column (Shiseido Co., Tokyo, Japan). The mobile phase consisted of 0.05 M potassium dihydrogen phosphate and acetonitrile at a ratio of 70:30 (v/v) and was delivered at a flow rate of 0.8 ml/min. The eluate was monitored at a wavelength of 205 nm. The column temperature was maintained at 30°C. Didesmethylcitalopram, the N-demethyl derivative of DCIT, and CIT N-oxide were not detected under these HPLC conditions. Determination of - and 4-hydroxytriazolam was carried out as described above, except that the mobile phase consisted of 10 mM potassium phosphate buffer (pH 7.4), acetonitrile and methanol at a ratio of 6:3:1 (v/v), and the eluate was monitored at 220 nm. Calibration curves were generated from 20 to 200 ng/ml by processing the authentic standard substances through the entire procedure. Analytes were quantified by comparison with the standard curves, using the peak-height ratio method. Determination of 4-hydroxymephenytoin was carried out as reported previously (Chiba et al., 1993). Intraassay (n = 6) coefficients of variation did not exceed 10% for all analytes.

Correlation study. The N-demethylation activities of CIT were compared with the 4-hydroxylation activities of (S)-mephenytoin and with the - and 4-hydroxylation activities of triazolam, using microsomes obtained from 10 human livers. The substrate concentrations used were 1 µM for CIT, 100 µM for (S)-mephenytoin and 25 µM for triazolam. Assays of (S)-mephenytoin 4-hydroxylation and triazolam - and 4-hydroxylation activities were performed in duplicate on the same day, with the same set of microsomal preparations.

Inhibition study. The effects of selective inhibitors or substrates (i.e., compounds acting as competitive inhibitors) of CYP2C19, CYP2D6 and CYP3A on the N-demethylation of CIT (10 and 100 µM) were studied. The isoform-selective inhibitors and alternative substrates used in this part of the study were 100 µM (S)-mephenytoin (CYP2C19) (Küpfer and Preisig, 1984), 10 µM quinidine (CYP2D6) (Guengerich et al., 1986b), 50 µM 17-ethinylestradiol (CYP3A) (Guengerich, 1988) and 10 µM ketoconazole (CYP3A) (Back and Tjia, 1991; Wrighton and Ring, 1994).

Immunoinhibition study. Human P450_MP (Shimada et al., 1986) and P450_NF (Guengerich et al., 1986a) were purified as reported previously. Polyclonal antibodies to CYP2C and CYP3A were raised in rabbits against human P450_MP and P450_NF, respectively, as described by Kaminsky et al. (1981). Anti-CYP2C antibodies used in the present study inhibited (S)-mephenytoin 4-hydroxylation (CYP2C19) and tolbutamide hydroxylation (CYP2C9) by >90%, whereas they did not inhibit testosterone 6-hydroxylation (CYP3A4) in human liver microsomes. Anti-CYP3A antibodies inhibited testosterone 6-hydroxylation (CYP3A4) by >80%, whereas they did not inhibit (S)-mephenytoin 4-hydroxylation (CYP2C19) in human liver microsomes.

Assay with cDNA-expressed human P450 isoforms. Microsomes from human B lymphoblastoid cells expressing human CYP1A2, CYP2A6, CYP2B6, CYP2C9, CYP2C19, CYP2D6, CYP2E1 and CYP3A4 (Gentest Corp., Woburn, MA) were used. The basic reactions were carried out as described for the human liver microsomal study. To examine the roles of individual P450 isoforms involved in CIT N-demethylation, each of the eight recombinant P450 isoforms (0.5 mg/ml protein concentration) described above was first incubated with 100 µM CIT for 120 min at 37°C, according to the manufacturer's recommended procedure. The rates of formation of DCIT from CIT were linear at least up to 60 min with both CYP2C19 and CYP3A4, which showed the catalytic ability of CIT N-demethylation under the same conditions. Accordingly, the following studies using recombinant CYP2C19 and CYP3A4 were carried out using an incubation period of 60 min and a protein concentration of 0.5 mg/ml.

Contribution of CYP3A4 and CYP2C19 to CIT N-demethylation in human liver microsomes. The percentage contributions of CYP2C19 and CYP3A4 to CIT N-demethylation in human liver microsomes were estimated by application of the RAF proposed by Crespi (1995). This approach makes the assumption that any effects on the rate of metabolism are independent of substrate, i.e., the rank order of rates of metabolism is the same for a particular P450 isoform and the same P450 isoform present in human liver microsomes, and any factor which affects the rate of metabolism for one substrate also does so equally for other substrates (Crespi, 1995). The validity of this assumption has not been rigorously tested, but for most enzymes an appropriate set of test compounds is available (Crespi, 1995). In the present study, we determined the RAF for CYP2C19 (i.e., RAF_CYP2C19) as the ratio of the activity of (S)-mephenytoin (100 µM) 4-hydroxylation, a specific metabolic reaction mediated via CYP2C19 (Goldstein et al., 1994; Wrighton et al., 1993), in human liver microsomes to that with recombinant CYP2C19. Triazolam (25 µM) -hydroxylation was used for the calculation of the RAF for CYP3A4 (RAF_CYP3A4), as a specific metabolic probe of CYP3A4 (Kronbach et al., 1989).

(1)

(2)

(3)

(4)

Estimation of CIT N-demethylation clearance with the first-order rate of metabolism. N-Demethylation clearances of CIT were estimated from the linear portion of the concentration-activity relationship. This is because our preliminary study showed that not only human liver microsomes but also recombinant CYP3A4 exhibited a curvilinear kinetic profile for the N-demethylation of CIT, although the possibility cannot be excluded that CIT enantiomers may show different kinetic profiles for the N-demethylation of CIT. Because we used a racemic mixture of CIT in the present study, we could not analyze these possibly complicated kinetics. Instead, we estimated the N-demethylation clearance of CIT with the first-order rate of metabolism from the linear portion of the concentration-activity relationship. Because the concentration-activity relationship was linear with up to 5 µM CIT with human liver microsomes and CYP3A4 and up to 25 µM CIT with CYP2C19, the parameters were estimated by linear regression analysis over these concentration ranges. All incubations for this estimation were carried out on the same days for each microsomal sample, to avoid errors due to changes in the activities over different storage periods for the microsomes.

Statistical analysis. Results are expressed as mean ± S.D. throughout the text. Differences in kinetic data and in estimated contributions of P450 isoform(s) between the putative EM and PM livers (table 1) were statistically evaluated using the Mann-Whitney U test. Correlation coefficients (r_s) were determined by the nonparametric technique (Spearman's rank correlation). A P value of <.05 was considered statistically significant.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Kinetic profile of CIT N-demethylation in human liver microsomes. Eadie-Hofstee plots for the N-demethylation of CIT (2.5-500 µM) in microsomes obtained from three putative EM (HL-3, -6 and -26) and two putative PM (HL-8 and -29) livers for (S)-mephenytoin are shown in figure 1. The plots showed biphasic curves, suggesting that each reaction showed multiple-enzyme kinetic behavior and/or K_m values of the enzyme(s) involved in the CIT N-demethylation were different for the CIT enantiomers. The curves for EM and PM microsomal samples were similar in biphasicity.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Eadie-Hofstee plots for the N-demethylation of CIT in human liver microsomes obtained from three putative EMs (, , ) and two putative PMs (, ) of (S)-mephenytoin. V, velocity of metabolite formation; S, substrate concentration. The individual R/S ratios were as follows: , 0.075; , 0.198; , 0.162; , 0.793; , 0.886.

Correlation study. The correlations in the individual activities for the N-demethylation of CIT vs the - and 4-hydroxylation of triazolam or the 4-hydroxylation of (S)-mephenytoin are shown in figure 2. The CIT N-demethylation activities showed a significant correlation with both - and 4-hydroxylation of triazolam (r_s = 0.818 and 0.851, respectively; P < .01) (fig. 2, A and B). Also, a significant correlation was observed between CIT N-demethylation and (S)-mephenytoin 4-hydroxylation (r_s = 0.773; P < .05) (fig. 2C). The (S)-mephenytoin 4-hydroxylation activities showed a significant correlation with both - and 4-hydroxylation of triazolam (r_s = 0.730 and 0.758, respectively; P < .05, respectively).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Correlations of triazolam -hydroxylation (A) and 4-hydroxylation (B) and (S)-mephenytoin 4-hydroxylation (C) with CIT N-demethylation in 10 human liver microsomal preparations, including microsomes obtained from four putative PMs of (S)-mephenytoin.

Inhibition study. The effects of four inhibitors or substrates on the formation of DCIT at 10 and 100 µM CIT in microsomes obtained from the three putative EM livers (HL-2, -6 and -27) for (S)-mephenytoin are shown in figure 3. Both ketoconazole and 17-ethinylestradiol inhibited the formation of DCIT by 68% at 10 µM CIT. At 100 µM CIT, these inhibitors showed 74% and 71% inhibition, respectively, of formation of this metabolite from CIT. Quinidine exhibited a minor inhibitory effect on the formation of DCIT at 10 and 100 µM CIT (11% and 9%, respectively). (S)-Mephenytoin also slightly inhibited the formation of DCIT at 10 and 100 µM CIT (17% and 18%, respectively). Interestingly, with microsomes (HL-8, -22 and -29) obtained from the putative PMs of (S)-mephenytoin, no inhibition by (S)-mephenytoin of the formation of DCIT at 100 µM CIT was observed, whereas ketoconazole, 17-ethinylestradiol and quinidine showed inhibition of 85%, 76% and 16%, respectively.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Mean ± S.D. of percentage inhibition of CIT N-demethylation activity at 10 and 100 µM CIT by 17-ethinylestradiol, ketoconazole, (S)-mephenytoin and quinidine in human liver microsomes obtained from three putative EMs of (S)-mephenytoin (R/S ratios ranging from 0.115 to 0.162).

Immunoinhibition study. Figure 4 shows the inhibition of N-demethylation of CIT by polyclonal antibodies raised against CYP3A or CYP2C. The addition of anti-CYP3A reduced the N-demethylation activity of CIT by approximately 70% at a concentration (2 mg IgG/mg microsomal protein) at which >80% of testosterone 6-hydroxylation was inhibited, whereas (S)-mephenytoin 4-hydroxylation was not. On the other hand, anti-CYP2C inhibited CIT N-demethylation by <10% at a concentration (2 mg IgG/mg microsomal protein) at which >90% of (S)-mephenytoin 4-hydroxylation was inhibited. The magnitude of inhibition by both antibodies was similar between the putative EM (HL-5) and PM (HL-29) microsomal samples.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Immunoinhibition of N-demethylation of CIT by anti-CYP3A (A) and anti-CYP2C (B) in human liver microsomes obtained from a putative EM () (R/S ratio of 0.204) and a putative PM () (R/S ratio of 0.793) of (S)-mephenytoin.

cDNA-expressed human P450 isoform study. Microsomes from human B lymphoblastoid cell lines expressing each of eight human P450 isoforms were examined in terms of the abilities of individual P450 proteins to catalyze CIT N-demethylation. CYP2C19 and CYP3A4 were found to catalyze the reaction (1.39 and 0.71 pmol/pmol P450/min, respectively), whereas other isoenzymes showed negligible activity for the N-demethylation of CIT (table 2).

Contributions of CYP2C19 and CYP3A4 to CIT N-demethylation in human liver microsomes. The activities of (S)-mephenytoin 4-hydroxylation and triazolam -hydroxylation in six different human liver microsomal preparations (HL-5, -6, -8, -22, -29 and -30) ranged from 2.9 to 48.3 pmol/mg/min and from 0.22 to 0.65 nmol/mg/min, respectively. The activities of (S)-mephenytoin 4-hydroxylation by CYP2C19 and of triazolam -hydroxylation by CYP3A4 were 4.59 and 1.73 nmol/nmol P450/min, respectively. RAF_CYP2C19 and RAF_CYP3A4 were thus estimated to range from 0.63 to 10.52 and from 0.13 to 0.38 pmol P450/mg protein, respectively (table 3). The CL_rec-CYP2C19 and CL_rec-CYP3A4 calculated from the linear ranges were 24.4 and 8.34 µl/nmol P450/min, respectively. Thus, the clearances by CYP2C19 and CYP3A4 contributing to CIT N-demethylation in six different human liver microsomes (CL_CYP2C19 and CL_CYP3A4) were estimated by using equations 1 and 2, respectively; the individual values are listed in table 3. The CL_HLM values estimated from the linear region (1-5 µM CIT) ranged from 1.72 to 3.67 µl/mg/min. The mean ± S.D. contributions of CYP2C19 to human microsomal CIT N-demethylation, calculated according to equation 3, were 7.35 ± 0.49 and 0.91 ± 0.16% for the microsomes of the putative EM (HL-5, -6 and -30) and PM (HL-8, -22 and -29) livers, respectively. The value for the putative EM microsomes was significantly (P < .05) greater than that for the PM microsomes. For CYP3A4, the mean ± S.D. contributions estimated by using equation 4 were 70.12 ± 15.07 and 58.99 ± 6.46% for the putative EM and PM microsomes, respectively; the values did not significantly differ between the two groups. The sum of the contributions of CYP2C19 and CYP3A4 was 94.26% in HL-5. However, the values in the other microsomes ranged from 52.63 to 71.75%.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The results suggest that CYP3A4 is primarily responsible for catalyzing the N-demethylation of CIT in human liver microsomes. This conclusion was inferred from the following observations. First, the CIT N-demethylation activities showed a significant correlation with both - and 4-hydroxylation activities for triazolam, a substrate of CYP3A4 (Kronbach et al., 1989), in microsomes prepared from 10 human livers (fig. 2). Second, ketoconazole and 17-ethinylestradiol, known as selective inhibitors or substrates of CYP3A (Back and Tjia, 1991; Guengerich, 1988; Wrighton and Ring, 1994), strongly inhibited the N-demethylation of CIT in microsomes obtained from both the putative EM and PM livers for (S)-mephenytoin 4-hydroxylation (fig. 3). Third, CIT N-demethylation was almost completely inhibited by the addition of anti-CYP3A antibodies (fig. 4). Fourth, the recombinant human CYP3A4 catalyzed CIT N-demethylation (table 2).

However, the N-demethylation of CIT was catalyzed not only by recombinant CYP3A4 but also by CYP2C19 (table 2), suggesting that CIT is a substrate of CYP2C19. This observation is consistent with our previous report that this drug competitively inhibited the 4-hydroxylation of (S)-mephenytoin in human liver microsomes (Kobayashi et al., 1995). This is also consistent with the observation that the N-demethylation activity of CIT was inhibited by (S)-mephenytoin in microsomes obtained from the putative EM livers (fig. 3), whereas no inhibition was observed in the PM microsomes. In addition, the N-demethylation activity of CIT correlated not only with the triazolam - and 4-hydroxylation activities but also with the (S)-mephenytoin 4-hydroxylation activity (fig. 2), although the latter correlation might depend on the result that the 4-hydroxylation activity for (S)-mephenytoin correlated with the - and 4-hydroxylation activities for triazolam. Moreover, the inhibition of CIT N-demethylation by (S)-mephenytoin in the EM microsomes was <20% (fig. 3), although anti-CYP2C antibodies scarcely inhibited the CIT N-demethylation (fig. 4). Taken together, these results suggest that CIT N-demethylation is catalyzed at least in part by CYP2C19 and thus that CYP2C19 contributes to this metabolic pathway of CIT in human liver microsomes to a much lesser extent than does CYP3A4.

Recently, Crespi (1995) proposed a corrective factor (i.e., RAF) to extrapolate the data obtained from cDNA-expressed P450s to those from human liver microsomes. The RAF was applied to interpret the data for the metabolic activation of tobacco smoke-derived nitrosoamine and aflatoxin b₁ obtained from cDNA-expressed P450s. However, little information has been available regarding the applicability of the RAF for assessing the magnitude of each of the P450 isoforms involved in the metabolism of therapeutic agents. Therefore, we applied the RAF to estimate the relative contributions of CYP2C19 and CYP3A4 to the overall CIT N-demethylation in each of the human liver microsomal samples used in the study. The estimated contributions of CYP3A4 and CYP2C19 to CIT N-demethylation were about 70% and 7%, respectively, in microsomes obtained from the putative EM livers (table 3). This further indicates that CIT N-demethylation is primarily catalyzed via CYP3A4. On the other hand, <10% of the overall N-demethylation was estimated to be catalyzed via CYP2C19 (table 3). These findings are consistent with the percentages of inhibition of CIT N-demethylation by CYP3A inhibitors or substrates (68-74%) (fig. 3) and by CYP2C19 substrate (17%) (fig. 3) in human liver microsomes, suggesting that the RAF may be a useful tool for estimating the contributions of human microsomal P450 isoenzymes to drug metabolism from the corresponding cDNA-expressed P450s. However, the concept of the RAF has been proposed based upon several assumptions (Crespi, 1995) (see "Materials and Methods"), and its validity has not been rigorously tested. Therefore, whether the proposed RAF concept would have wide applicability for other in vitro experiments like ours remains unknown and definitely requires further assessment.

With the limitations discussed above, the present approach using the RAF indicated that the CIT N-demethylation activities were almost completely explainable by metabolism via CYP3A4 and CYP2C19 in HL-5. However, some portions of CIT N-demethylation could not be explained by CYP3A4 and CYP2C19 alone in the remaining human liver microsomal samples. The findings suggest that enzymes other than CYP3A4 and CYP2C19 are partially involved in CIT N-demethylation in those microsomal samples, although it remains unclear from the present study which P450 isoform(s) may be involved. However, because CIT N-demethylation in human liver microsomes was completely inhibited by anti-CYP3A antibodies or selective inhibitors of CYP3A, CYP3A isoform(s) other than CYP3A4 (e.g., CYP3A5) may be responsible for CIT N-demethylation.

The present in vitro observation that the contribution of CYP2C19 to CIT N-demethylation was <10% in the putative EM liver microsomes is incompatible with findings from the in vivo human panel study by Sindrup et al. (1993). They reported that the N-demethylation clearance of CIT was about 2-fold greater in EM than in PM subjects for (S)-mephenytoin, indicating that about 50% of the total CIT N-demethylation was catalyzed by CYP2C19. We cannot offer any reasonable explanation for the discrepancy between the results of the present in vitro study and those of the in vivo study by Sindrup et al. (1993). However, we are tempted to assume that the hepatic microsomes designated as those from putative EM livers in the present in vitro study might have been obtained from Japanese patients heterozygous for CYP2C19. This is because the PM frequency for the 4-hydroxylation of (S)-mephenytoin is much greater in Japanese populations (22.5%) (Horai et al., 1989), compared with that in Caucasian populations (3-6%) (Alván et al., 1990; Jacqz et al., 1988; Wedlund et al., 1984). Therefore, the frequency of individuals heterozygous for the EM genotype (i.e., wt/m1 or wt/m2) (de Morais et al., 1994a,b) would be estimated to be much greater in Japanese populations, compared with Caucasian populations, implying that the contribution of CYP2C19 to the N-demethylation of CIT might be lower in Japanese human liver microsomes, compared with those from Caucasian subjects. Obviously, this assumption must be confirmed by interethnic in vitro and/or in vivo studies.

CIT is a chiral compound that is commercially available as a racemic mixture. Although (S)-CIT was reported to be metabolized more extensively than (R)-CIT in vivo (Rochat et al., 1995), little information is available concerning the stereoselective metabolism of CIT in in vitro studies. However, the results of our preliminary study indicated that the N-demethylation clearance of (S)-CIT via CYP2C19 is about 2 times greater than that of the R-enantiomer (K. Kobayashi, K. Chiba, T. Yamamoto, T. Ishizaki and Y. Kuroiwa, unpublished observations). Therefore, the discrepancy between the in vitro and in vivo CIT N-demethylation activities in humans might be attributable not only to the interethnic differences in CYP2C19 activity but also to stereoselective CIT N-demethylation by CYP2C19.

In conclusion, the present study using human liver microsomes and recombinant human P450 isoforms has strongly suggested that the N-demethylation of CIT is mediated mainly, but not exclusively, via CYP3A4 and partially via CYP2C19. In addition, the RAF proposed by Crespi (1995) appears to be a tool for estimating the contributions of the two P450 isoforms involved in CIT N-demethylation, thus bridging the gap between human liver microsomal and recombinant human P450 isoform studies.