Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

It has been well known that interindividual variations in the metabolic profile of many drugs are linked to genetic polymorphism in the cytochrome P450 (CYP) responsible for their metabolisms. The consequences of the variations may be of clinical significance with respect to either efficacy or toxicity of the drug as well as drug-drug interactions. CYP2C19 is one of the CYP isoforms concerning individual and ethnic variations of drug metabolisms that have been studied most extensively in recent years. CYP2C19-medicated S-mephenytoin 4'-hydroxylation shows a genetically determined polymorphism that has marked interracial differences, with the PM phenotype representing 2 to 5% of the Caucasian population but 13 to 23% of the Asian population (Wilkinson et al., 1989; Alván et al., 1990; Xiao et al., 1997). Metabolisms of a number of drugs, including diazepam (Qin et al., 1999), omeprazole (Shu et al., 2000), chloroguanide (Herrlin et al., 2000), and fluoxetine (Liu et al., 2001a, 2001b) in vivo or in vitro cosegregate with the CYP2C19 polymorphism.

Fluoxetine is a potent and selective serotonin reuptake inhibitor in the central nervous system and is widely used to treat depression and obsessive-compulsive behavior (Fuller et al., 1991; Gram, 1994). Despite the extensive use of fluoxetine clinically, it is thought that as much as 50% of its metabolism is still unclear. This study is very important because either fluoxetine metabolites are often therapeutically active or they may contribute to the side-effect profile of the parent drug. In addition, metabolites may compete with other exogenous substrates for catabolic enzymes and result in marked changes in human tissues and body fluid levels of these other drugs and their metabolites (Urichuk et al., 1997; Hiemke and Härtter, 2000). Up to now, the two major routes of metabolism for fluoxetine that have been identified are N-demethylation to norfluoxetine and O-dealkylation to TFMP (Altamura et al., 1994; Urichuk et al., 1997). Fluoxetine is metabolized extensively by the hepatic cytochrome P450 enzymes, and less than 2.5% of the drug is found unchanged in the human urine. Some studies have shown that polymorphic CYP2C19 and CYP2C9 appear to be the principal human CYP isoenzymes mediating N-demethylation of fluoxetine (von Moltke et al., 1997; Liu et al., 2001b) and CYP3A4 may make a minor contribution to it. To the best of our knowledge, there is no report as to which CYP isoforms are responsible for the O-dealkylation of fluoxetine, and TFMP has never been quantitated in human liver microsomes. Furthermore, fluoxetine O-dealkylation has not yet been characterized in vitro with respect to the individual CYP2C19 genotype status. Therefore, this study is of significance in that it investigates whether one or more isoenzymes of CYP are involved in the formation of TFMP from fluoxetine. In vivo and in vitro studies for fluoxetine metabolism have shown great interindividual variabilities consistent with CYP2D6 and CYP2C19 polymorphic characteristics (Hamelin et al., 1996; Liu et al., 2001a). Considering the involvement of multienzymes and the different contributions of various CYP isoforms in the O-dealkylation of fluoxetine, in this study we evaluated the correlation between fluoxetine O-dealkylase activity and S-mephenytoin 4'-hydroxylase and midazolam 1'-hydroxylase activity. To further determine the role of CYP2C19 in the O-dealkylation of fluoxetine, we assessed the relative contribution of CYP2C19 using different genotyped liver microsomes from three homozygous EMs, three heterozygous EMs, and three PMs of CYP2C19. Various selective chemical inhibitors and recombinant CYP1A2, 2C8, 2C9, 2C19, and 3A4 were also used to identify the isoforms of CYP involved in fluoxetine O-dealkylation.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Fluoxetine hydrochloride was supplied by Sigma/RBI (Natick, MA). TFMP and pentafluorobenzenesulfonyl chloride (PFBSC) were purchased from Aldrich Chemical Co. (Milwaukee, WI). 2,4-Dichlorophenol (internal standard) was supplied by Shanghai Chemical Center (Shanghai, China). Omeprazole was a generous gift from Astra Hässle AB (Mölndal, Sweden). Coumarin, quinidine, triacetyloleandomycin (TAO), diethyldithiocarbamate (DDC), NADP⁺, glucose 6-phosphate, and glucose-6-phosphate dehydrogenase were purchased from Sigma (St. Louis, MO). Sulfaphenazole was a gift from CibaGeigy Ltd. (Basel, Switzerland), and furafylline was kindly donated by Dr. W. Pfleiderer (Universität Konstanz, Wurzburg, Germany). Recombinant CYP1A2, 2C8, 2C9, 2C19, and 3A4A were purchased from Gentest (Woburn, MA). Acetonitrile of high-performance liquid chromatography grade was purchased from Tedia Company Inc. (Fairfield, OH). All other chemicals were of analytic reagent grade.

Human Liver Microsomes. Adult human liver tissues were collected from renal transplant donors and patients undergoing partial hepatectomy in our liver bank. The collection and use of human liver tissues in this study were approved by the Ethics Committee of Xiang-Ya School of Medicine, Central South University. Candidate patients for liver sample collection were those who did not suffer from acute or chronic hepatitis or cirrhosis and took no medications known to induce or inhibit CYP activity. Portions of surgical liver "waste tissues" distant from disease-affected regions and appearing visually normal was collected. The collection approaches of liver tissue and its morphologic and biochemical characterization were described elsewhere (von Bahr et al., 1980). Microsomes were prepared by differential centrifugation (von Bahr et al., 1980) and stored at 80°C ready for use. Microsomal protein concentration was determined by the method of Lowry et al. (1951).

Fluoxetine Metabolism in Vitro and GC-ECD Analysis. Fluoxetine metabolism in vitro was used in 0.1 M potassium phosphate buffer (pH 7.4) containing 1.0 mg/ml liver microsomal protein, reduced NADP (NADPH)-generating system, and various concentrations of fluoxetine with or without inhibitors in a final volume of 500 µl. The enzyme reaction was initiated by adding 20 µl of various concentrations of substrate and NADPH-generating system consisting of 1 mM NADP, 10 mM glucose 6-phosphate, 2 IU/ml glucose-6-phosphate dehydrogenase, and 10 mM MgCl₂. After the incubation at 37°C in a shaking water bath for 45 min, the reaction was stopped by cooling on ice and the addition of 100 µl of acetonitrile. Preliminary experiments showed that the formation of TFMP was linear to both incubation time over 60 min and microsomal protein concentration (0.5-2 mg/ml) at 37°C. Accordingly, the incubation time of 45 min and the microsomal protein concentration of 1 mg/ml were chosen in the present study.

Kinetics for TFMP Formation. Ten concentrations of fluoxetine (1-200 µM) were used to characterize the kinetics of fluoxetine O-dealkylation by liver microsomes from different CYP2C19 genotypes. Several enzyme kinetic equations proposed by Schmider et al. (1996) were used to fit the untransformed kinetic data (Figperfect, version 5.0; Soft Corporation, Durham, NC). The most appropriate model was obtained based on the dispersion of residuals and whether an F-test showed a significant reduction (P < 0.05) in the residual sum of squares. The next two equations best described the kinetics of fluoxetine O-dealkylation by EM (wt/wt and wt/m1) microsomes and PM (m1/m1) microsomes, respectively.

(1)

(2)

Inhibition Studies. The inhibitory effect of various selective inhibitors was examined in four EM microsomes at a substrate concentration of 100 µM fluoxetine. The selective inhibitors used were 200 µM coumarin (CYP2A6 substrate), 10 µM quinidine (CYP2D6 inhibitor), 20 µM DDC (CYP2E1 inhibitors), 50 µM TAO (CYP3A4 inhibitor), 25 µM furafylline (CYP1A2 inhibitor), 20 µM sulfaphenazole (CYP2C9 inhibitor), and 100 µM omeprazole (CYP2C19 substrate) (Andersson et al., 1994; Newton et al., 1995; Ko et al., 1997; Eagling et al., 1998). All inhibitors were dissolved in methanol except for DDC in distilled water. Because methanol has an inhibitory effect on CYP activity, solutions were dried before incubation. Furafylline, TAO, and DDC were preincubated with liver microsomal preparations and the NADPH generating at 37°C for 15 min before substrate was added.

Metabolism of Fluoxetine by cDNA-Expressed P450s. Recombinant CYP1A2, 2C8, 2C9, 2C19, and 3A4 were further used to assess the roles of CYP2C and CYP3A4 in the metabolism of fluoxetine. Fifty picomoles of human lymphoblast-expressed CYP1A2, 2C8, 2C9, 2C19, and 3A4 were coincubated with an NADPH-generating system, respectively, and the reaction was initiated by addition of 20 µl of substrate and incubated for 45 min. The reaction was stopped finally by cooling on ice and the addition of 100 µl of acetonitrile.

Correlation Studies. To further determine whether CYP2C19 and CYP3A4 are major CYP enzymes responsible for fluoxetine O-dealkylation, three substrate concentrations (5, 25, and 100 µM) of fluoxetine and 11 liver microsomes from EMs of CYP2C19 were used. The activities of fluoxetine O-dealkylation were correlated with the activities of 250 µM S-mephenytoin 4'-hydroxylation and 100 µM midazolam 1'-hydroxylation. The rate of formation of 4'-hydroxymephenytoin and 1'-hydroxymidazolam was determined using high-performance liquid chromatography as described by Xie et al. (1995) and by Carrillo et al. (1998), respectively.

Data Analysis. Duplicate incubations were used throughout the study. Data were analyzed by the paired and unpaired Student's t test and a one-way analysis of variance. The correlations between fluoxetine O-dealkylation in different liver microsomal preparations and S-mephenytoin 4'-hydroxylation and midazolam 1'-hydroxylation were determined by least-squares linear regression (SPSS for Windows 8.0; SPSS, Chicago, IL). A P value of < 0.05 was considered to be the minimum level of significance.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Kinetics for TFMP Formation. The kinetics of TFMP formation was studied in liver microsomes from nine subjects (three wt/wt, three wt/m1, and three m1/m1). After iteratively fitting the different enzyme kinetic models proposed by Schmider et al. (1996) to the untransformed data of each subject, the kinetics of TFMP formation in the six EMs (three wt/wt and three wt/m1) microsomes followed the two-enzyme Michaelis-Menten model (eq. 1), whereas the kinetics in the three PMs microsomes was best described by the single-enzyme Michaelis-Menten model (eq. 2). The kinetic parameters for fluoxetine O-dealkylation in EMs and PMs are shown in Table 1. The substrate versus velocity plots and the Eadie-Hofstee plots for TFMP formation showed a difference in kinetic behavior between the EM microsomes and the PM microsomes (Fig. 1). In PMs microsomes, the high-affinity component of fluoxetine O-dealkylation was absent. Compared with EMs, the formation of TFMP in PM liver microsomes was significantly slower, especially at a low substrate concentration (data not shown). Furthermore, the concavity of Eadie-Hofstee plots in the homozygous EMs (Fig. 1A) showed the involvement of at least two CYP enzymes in the reaction, which was more apparent for the homozygous EMs than for the heterozygous EMs. Mean apparent K_m values for the high- and low-affinity components were 4.6 and 60.2 µM in the homozygous EM liver microsomes and 9.1 and 70.9 µM in the heterozygous EM liver microsomes, and V_max values were 116 and 224 pmol/min/nmol of P450 protein in the homozygous EM microsomes and 113 and 191 pmol/min/nmol of P450 protein in the heterozygous EM microsomes. The mean intrinsic clearances (V_max1/K_m1) of the high-affinity component was 6.6 times that (V_max2/K_m2) of the low-affinity component in the homozygous EM microsomes and 4.4 times in the heterozygous EM microsomes.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Representative substrate versus velocity and Eadie-Hofstee plots for the formation of TFMP in human liver microsomes from different genotypes of CYP2C19, including homozygous EMs microsomes (A), heterozygous EMs microsomes (B), and PMs microsomes (C). The values are the means of duplicate incubations.

Inhibition with Selective Chemical Inhibitors. Omeprazole (100 µM) and TAO (50 µM) caused a mean 46.9% (P < 0.001) and 69.8% (P < 0.001) reduction in the formation of TFMP. At a substrate concentration of 100 µM, 10 µM quinidine inhibited this reaction to a minor extent. However, no inhibitory effect was observed for coumarin, furafylline, DDC, and sulfaphenazole. The addition of 100 µM omeprazole plus 50 µM TAO produced a maximum inhibition of 82.8% (P < 0.001) for TFMP formation (Fig. 2).

View larger version (25K): [in this window] [in a new window]   Fig. 2.   Effect of selective cytochrome P450 inhibitors, 200 µM coumarin, 25 µM furafylline, 10 µM quinidine, 20 µM DDC, 50 µM TAO, 20 µM sulfaphenazole, 100 µM omeprazole, and 100 µM omeprazole plus 50 µM TAO on the formation of TFMP in human liver microsomes from four EMs (two wt/wt, HL-4 and HL-11; two wt/m1, HL-10 and HL-22) at a substrate concentration of 100 µM fluoxetine. The values are the mean inhibition percentage (±S.D.). COU, coumarin; FUR, furafylline; QUI, quinidine; DDC, diethyldithiocarbamate; SUL, sulfaphenazole; OME, omeprazole.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Effect of omeprazole (left, A) and TAO (right, B) on the formation of TFMP in four EMs microsomes (two wt/wt, HL-4 and HL-11; two wt/m1, HL-10 and HL-22) at a low substrate concentration of 5 µM (solid circle) and at a high substrate concentration of 100 µM (open circle). The values are the mean inhibition percentage (±S.D.).

Inhibitory Effect of TAO on TFMP Formation by Different CYP2C19 Genotyped Microsomes. Three substrate concentrations (5, 25, and 100 µM) were used to assess the inhibitory effect of 50 µM TAO on TFMP formation in the nine liver microsomes (three wt/wt, three wt/m1, and three m1/m1). At a low substrate concentration (5 µM), TAO had a relatively minor inhibitory effect (<47%) on TFMP formation in both the homozygous EMs microsomes and heterozygous EMs microsomes. However, the mean percentage inhibition by TAO was lower in the homozygous EM microsomes than in the heterozygous EM microsomes (35.3 versus 47.0%, P < 0.05). With the increase of substrate concentration, the mean percentage inhibition of fluoxetine O-dealkylation increased to 47.4 and 59.3% at 25 µM fluoxetine and to 65.9 and 74.3% at 100 µM fluoxetine in the homozygous and the heterozygous microsomes, respectively. In the PM microsomes, 50 µM TAO almost abolished TFMP formation at all the three substrate concentrations (>90%) (Fig. 4).

View larger version (48K): [in this window] [in a new window]   Fig. 4.   Effect of 50 µM TAO on the formation of TFMP at different substrate concentrations (5, 25, and 100 µM) in liver microsomes from different CYP2C19 genotypes (three wt/wt, three wt/m1, and three m1/m1). The values are the mean inhibition percentage (±S.D.). FLU, fluoxetine.

Correlations Studies. The microsomal activities of fluoxetine O-dealkylation at low (5 µM), medial (25 µM), and high (100 µM) substrate concentrations were measured for 11 liver microsomes from EMs of CYP2C19. The rate of formation of 4'-hydroxymephenytoin and 1'-hydroxymidazolam in the these liver microsomes was also determined by previously described methods of Xie et al. (1995) and Carrillo et al. (1998), which were reflected with the CYP2C19 activity and CYP3A4 activity. At a low substrate concentration of 5 µM, a good correlation (r = 0.740, P < 0.01) was found between fluoxetine O-dealkylation and S-mephenytoin 4'-hydroxylation (Fig. 5A). However, with the increase of substrate concentration, the correlation coefficient between these two metabolic reactions decreased to r = 0.530 (P > 0.05) at 25 µM fluoxetine (Fig. 5B) and to r = 0.402 (P > 0.05) at a high substrate concentration of 100 µM (Fig. 5C), indicating no significant correlation. In contrast, The formation of TFMP at high (100 µM) and medial (25 µM) substrate concentrations showed close correlation with midazolam 1'-hydroxylation (r = 0.763, P < 0.01, Fig. 6C; r = 0.679, P < 0.05, Fig. 6B; respectively) whereas no significant correlation between these two metabolic reactions was found at a low substrate concentration of 5 µM (r = 0.424, P > 0.05) (Fig. 6A).

View larger version (12K): [in this window] [in a new window]   Fig. 5.   Correlation between S-mephenytoin 4'-hydroxylation and fluoxetine O-dealkylation at different substrate concentrations (A, 5 µM; B, 25 µM; C, 100 µM) in the liver microsomes of 11 Chinese individuals genotyped as EMs with respect to CYP2C19.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Correlation between midazolam 1'-hydroxylation and fluoxetine O-dealkylation at different substrate concentration (A, 5 µM; B, 25 µM; C, 100 µM) in the liver microsomes of 11 Chinese individuals genotyped as EMs of CYP2C19.

Effects of Recombinant CYP1A2, 2C8, 2C9, 2C19, and 3A4 on Fluoxetine Metabolism. Effects of recombinant CYP1A2, 2C8, 2C9, 2C19, and 3A4 on fluoxetine metabolism were observed only at low (5 µM) and high (100 µM) substrate concentrations (Table 2). The results showed that CYP2C19 and CYP3A4 were two major cytochrome P450 isoforms responsible for fluoxetine O-dealkylation, and CYP1A2, CYP2C8, and CYP2C9 catalyzed this reaction to a minor extent. At a substrate concentration of 5 µM, recombinant CYP2C19 produced a maximal catalyzing activity in the O-dealkylation of fluoxetine, however, CYP3A4 had higher catalyzing activity in this reaction than CYP2C19 at a substrate concentration of 100 µM (Fig. 7).

View larger version (24K): [in this window] [in a new window]   Fig. 7.   The rate of formation of TFMP from fluoxetine catalyzed by human lymphoblast-expressed CYP1A2, CYP2C8, CYP2C9, CYP2C19, and CYP3A4, respectively, at a low (5 µM, A) and high (100 µM, B) substrate concentrations.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

In general, the CYP2C19 oxidation polymorphism caused impaired drug metabolism and affected as many as 20% Asians but only 3% Caucasians. We observed a biphasic-enzyme kinetics for the formation of TFMP from fluoxetine in EM microsomes and a monophasic-enzyme kinetics in PM microsomes. These data indicated clearly that at least two CYP isoenzymes were involved in the O-dealkylation of fluoxetine with high- and low-affinity components. However, fluoxetine O-dealkylation lacked the high-affinity component and exhibited monophasic enzyme kinetics in PM microsomes. Furthermore, the intrinsic clearance of the high-affinity component for fluoxetine O-dealkylation was 6.6 and 4.4 times that of the low-affinity component in the homozygous EM microsomes and the heterozygous EM microsomes. These kinetic data suggested that CYP2C19 was the major CYP enzyme contributing to fluoxetine O-dealkylation in vitro. Moreover, the homozygous EMs showed a more typical substrate versus concentration plot and Eadie-Hofstee plot for the two-enzyme model compared with the heterozygous EMs (Fig. 1). This was probably due to the higher activity of fluoxetine O-dealkylation at a low substrate concentration in the homozygous EMs. In fact, some studies have shown that gene dose has an effect on CYP2C19 activity (homozygous EMs > heterozygous EMs > PMs) (de Morais et al., 1995; Shu et al., 2000). As a result, the genotype-related differences in CYP2C19 activity may result in different apparent enzyme kinetics for a metabolic reaction mediated by CYP2C19 and other enzymes between different genotyped liver microsomes. To our knowledge, this is the first study characterizing the in vitro enzyme kinetic behavior for the formation of TFMP from fluoxetine in human liver microsomes from different CYP2C19 genotypes.

Recently, omeprazole has been used as a potent selective inhibitor of CYP2C19 activity (Ko et al., 1997). At a concentration of 10 µM, omeprazole strongly inhibited the formation of both 4'-hydroxymephenytoin from S-mephenytoin (>90%) (Ko et al., 1997) and cycloguanil from proguanil (47%) in vitro (Funck-Brentano et al., 1997), two reactions catalyzed by CYP2C19. Therefore, to investigate the contribution of CYP2C19 to fluoxetine O-dealkylation, omeprazole was chosen as a specific selective inhibitor of CYP2C19 in this study. At a high substrate concentration of 100 µM, omeprazole caused a maximum of 46.9% reduction compared with the control value in the formation of TFMP, suggesting the involvement of CYP2C19. A selective and potent inhibitor of CYP3A4 (Pessayre et al., 1983; Xu et al., 1999), 50 µM TAO inhibited fluoxetine O-dealkylation by up to 69.8% at a high substrate concentration of 100 µM, indicating that CYP3A4 may also be the main enzyme involved in the O-dealkylation of fluoxetine.

To further assess the relative contribution of major CYP isoforms responsible for fluoxetine O-dealkylation in liver microsomes from Chinese individuals, inhibition experiments, heterologous expression experiments, and correlation studies were carried out. Within a range of concentrations of omeprazole (0-200 µM) and TAO (0-50 µM), we found that the inhibitory effect on TFMP formation by omeprazole was greater at a low substrate concentration (5 µM) than at a high substrate concentration (100 µM). However, TAO had a stronger inhibitory effect on this reaction at a high substrate concentration (100 µM) than at a low substrate concentration (5 µM). Furthermore, in the EM microsomes, TFMP formation was inhibited slightly by 50 µM TAO at a low substrate concentration. With the increase of substrate concentrations to 25 and 100 µM, the inhibition of TFMP formation by TAO substantially increased. In particular, under different substrate concentrations, the inhibitory potency of TAO on TFMP formation in the heterozygous liver microsomes was higher than that of homozygous liver microsomes, indicating that TAO had a markedly different inhibitory effect on fluoxetine O-dealkylation in the different CYP2C19 genotyped liver microsomes. Moreover, the correlation between fluoxetine O-dealkylation with S-mephenytoin 4'-hydroxylation declined with the increase of fluoxetine concentrations. However, the correlation between fluoxetine O-dealkylation with midazolam 1'-hydroxylation increased with the increase of substrate concentrations. These results showed that fluoxetine O-dealkylation had a substrate concentration-dependent contribution of CYP2C19 and CYP3A4. In addition, TFMP formation in the PM microsomes was inhibited almost completely by TAO (>90%), indicating that CYP3A4 is the principal enzyme responsible for this reaction in the PM microsomes. Finally, heterologous expression experiment further showed that recombinant CYP2C19 and CYP3A4 produced an maximum catalyzing effect on fluoxetine O-dealkylation at a low and a high substrate concentration, respectively. Accordingly, at therapeutically relevant substrate concentrations, e.g., at 5 µM, fluoxetine O-dealkylation was mediated predominantly via CYP2C19, with only a minor contribution of CYP3A4.

The fact that gene dose has an effect on drug metabolism has been reported before. Hamelin et al. (1996) reported that the metabolic ratio of fluoxetine N-demethylation in homozygous EMs to CYP2D6 is higher than that of heterozygous EMs. Recent studies from our laboratory have shown that the metabolism of S-mephenytoin (Shu et al., 2000) and diazepam (Qin et al., 1999) were gene dose-dependent. In this study, we found that at the therapeutically relevant substrate concentration of 5 µM, the mean TFMP formation of the three homozygous EM livers was significantly higher than that of the three heterozygous EM livers, indicating a gene dose effect on fluoxetine O-dealkylation. Thus, gene dose effect may result in differential inhibition of the affected CYP isoforms by substrate or inhibitors between different genotyped subjects, as demonstrated in our inhibition studies. In addition, recent studies conducted in our laboratory showed that the in vivo induction of CYP2C19 by rifampicin was gene dose-dependent (Feng et al., 1998). The higher proportion of heterozygous CYP2C19 EMs in Asian subjects is considered to be a cause of the differences between Caucasian and Asian subjects in the metabolism of chloroguanide (Herrlin et al., 2000) and diazepam (Qin et al., 1999). The genetic polymorphism of CYP2C19 is likely to be one of the major factors causing the interindividual and interracial differences of some drugs that are metabolized by CYP2C19.

In conclusion, polymorphic CYP2C19 is a high-affinity enzyme responsible for fluoxetine O-dealkylation in human liver microsomes. CYP3A4 is a low-affinity enzyme that contributes little to this metabolic reaction at the therapeutically relevant substrate concentration in EM microsomes but a lot to PM microsomes. The contribution of CYP2C19 to fluoxetine O-dealkylation is gene dose-dependent.