Abstract
A new HPLC method was developed using a chiral column to efficiently separate four 1′′-hydroxybufuralol (1′′-OH-BF) diastereomers that are major metabolites of bufuralol (BF). Employing this method, we examined diastereomer selectivity in the formation of 1′′-OH-BF from BF racemate or enantiomers in four individual samples of human liver microsomes. Three different human liver microsomes showed a selectivity of 1′′R-OH < 1′′S-OH for BF enantiomers, which was similar to that of recombinant CYP2D6 expressed in insect cell microsomes, whereas one human liver microsomal fraction yielded a selectivity of 1′′R-OH > 1′′S-OH for BF enantiomers, which was similar to those of recombinant CYP2C19 expressed in insect cell microsomes. Recombinant CYP1A2 and CYP3A4 showed a selectivity similar to that of CYP2D6, but their BF 1′′-hydroxylase activities were much lower than those of CYP2D6. In inhibition studies, quinidine, a known CYP2D6 inhibitor, markedly inhibited BF 1′′-hydroxylation in the fractions of human liver microsomes that showed the CYP2D6-type selectivity. Furthermore, omeprazole, a known CYP2C19 inhibitor, efficiently suppressed the formation of 1′′-OH-BF diastereomers from BF in the microsomal fraction that showed the CYP2C19-type selectivity. From these results, we concluded that the diastereomer selectivity in the formation of 1′′-OH-BF from BF differs between CYP2D6 and CYP2C19, both of which can be determinant enzymes in the diastereoselective 1′′-hydroxylation of BF in human liver microsomes.
Cytochrome P450 2D6 (CYP2D6) is known to be one of the major drug-metabolizing CYP enzymes in human livers and is responsible for the major metabolic pathways of more than 60 clinically used medicines that are often prescribed (Gonzalez, 1996; Meyer and Zanger, 1997). Bufuralol (BF; Fig. 1), a β-adrenoceptor blocking agent, is a typical substrate of CYP2D6. As shown in Fig. 1, BF has an asymmetric carbon at the side chain, yielding enantiomers, 1′R-BF (R-BF) and 1′S-BF (S-BF). The major metabolic pathway of BF is 1′′-hydroxylation of the ethyl group at the 7-position of its aromatic ring by CYP2D6. The 1′′-hydroxylation produces a new chiral center in the BF molecule, thereby theoretically yielding four diastereomers [1′′R-OH-1′R-BF (1′′R-OH-R-BF), 1′′S-OH-1′R-BF (1′′S-OH-R-BF), 1′′R-OH-1′S-BF (1′′R-OH-S-BF) and 1′′S-OH-1′S-BF (1′′S-OH-S-BF)] of 1′′-hydroxybufurarol (1′′-OH-BF) from racemic BF (see Fig. 1).
For studies on the stereoselectivity in the oxidative metabolism of BF enantiomers by mammalian liver microsomes, Weerawarna et al. (1991)developed an HPLC method employing phenylethylurea derivatives of diastereomeric 1′′-OH-BFs and demonstrated that rat liver microsomes favored the formation of 1′′R-OH metabolite over 1′′S-OH metabolite from both R-BF andS-BF, whereas human liver microsomes favor the formation of 1′′S-OH metabolite over 1′′R-OH metabolite from the same substrates. Therefore, between rats and humans the stereoselectivity in the formation of 1′′-OH-BF diastereomers was shown to be reversed. However, there are no reports describing the profile of the diastereoselective metabolism of BF racemate in any animal species.
The pharmacological activity of BF as a β-antagonist is much higher in S-form than in R-form (Hamilton and Parkes, 1977). The metabolism of BF afforded several pharmacologically active metabolites such as diastereomers of 1′′-OH-BF and enantiomers of 1′′-oxo-BF, and the activity of 1′′-OH-S-BF is 2 to 8 times higher than that of S-BF (Machin et al., 1985). However, it is not known whether 1′′R-OH-S-BF or 1′′S-OH-S-BF is a eutomer or a distomer. In this context, CYP enzyme(s) that are mainly responsible for the 1′′-hydroxylation of BF may contribute to the pharmacological effects of BF as key enzymes responsible for the formation of pharmacologically active metabolites in vivo, at least in part.
We have been studying the stereochemistry in the oxidation of various chiral drugs by CYP2D enzymes in rodents (Narimatsu et al., 1994,1996b), monkeys (Narimatsu et al., 1996a, 2000) and humans (Narimatsu et al., 1999, 2000). In the present study, we developed a new HPLC method to simultaneously separate the four known 1′′-OH-BF diastereomers by employing a chiral column without time-consuming derivatization. This technique was used to study BF metabolism using human liver microsomes and recombinant CYP enzymes.
Materials and Methods
Materials.
Racemic BF hydrochloride was supplied by Daiichi-Kagaku Chemicals (Tokyo, Japan). Enantiomeric BF hydrochlorides and four 1′′-OH-BF diastereomers (1′′R-OH-R-BF, 1′′S-OH-R-BF, 1′′R-OH-S-BF, and 1′′S-OH-S-BF in Fig. 1) were gifts from Roche Diagnostics (Basel, Switzerland). The absolute configurations of the metabolites were confirmed by comparing their circular dichroism (CD) spectra with those reported previously (Weerawarna et al., 1991). The spectra were recorded on a JASCO J-720W spectropolarimeter equipped with a microcomputer (Dell Optiplex 466/Le). Ethoxyresorufin, tolbutamide and hydroxytolbutamide, S-mephenytoin and 4-hydroxymephenytoin, chlorzoxazone and 6-hydroxychlorzoxazone, and diazepam and 3-hydroxydiazepam were obtained from Daiichi Pure Chemicals (Tokyo, Japan). Debrisoquine (DB) and 4-hydroxydebrisoquine (4-OH-DB) as hemisulfates were supplied by F. Hoffmann La-Roche (Basel, Switzerland). Glucose 6-phosphate (G-6-P), G-6-P dehydrogenase, and NADPH were purchased from Oriental Yeast Co., Ltd. (Tokyo, Japan).
Enzyme Source.
Human liver microsomal fractions (four samples, three male and one female, Caucasians, age ranging from 38 to 65 years) were supplied by the Human and Animal Bridge Discussion Group (HAB) (Chiba, Japan). This study was approved by the Ethics Committee of the Faculty of Pharmaceutical Sciences, Okayama University. Recombinant CYP enzymes CYP1A2, CYP2C19, CYP2D6, and CYP3A4 expressed in insect cells (BTI-TN-5B1-4) were obtained as microsomal fractions from BD Gentest (Woburn, MA). CYP3A4 was the preparation coexpressed with cytochrome b5 in insect cells.
Enzyme Assays.
The following CYP enzyme-specific assays were performed by previously published methods as follows: CYP1A2, 7-ethoxyresorufin O-deethylation (Riley et al., 1995); CYP2C9, tolbutamide methylhydroxylation (Relling et al., 1990); CYP2C19, S-mephenytoin 4′-hydroxylation (Meier et al., 1985); CYP2D6, debrisoquine 4-hydroxylation (Kronbach et al., 1987); CYP2E1, chlorzoxazone 6-hydroxylation (Peter et al., 1990); CYP3A4, diazepam 3-hydroxylation (Mei et al., 1999).
BF 1′′-Hydroxylation Assay.
BF racemate or enantiomers were added to the reaction medium containing 5 mM G-6-P, 10 mM MgCl2, 1 mM NADPH, and 0.5 IU G-6-P dehydrogenase (final volume of 0.5 ml), which was preincubated at 37°C for 5 min. Reactions were started by adding microsomes (0.13–0.25 mg of protein for human liver microsomes; 0.021–0.063 mg of protein for recombinant CYP enzymes), and continued for 1 min. After the incubation, 1 ml of 1 N NaOH aqueous solution was added to stop the reaction, followed by addition of 1 ml of 1 M carbonate buffer (pH 9.6) and 25 μl of 1 mM (+)-bunitrolol (BTL) as an internal standard. Metabolites were extracted into 5 ml of ethyl acetate by vigorous shaking for 10 min and centrifugation at 2000g for 5 min at room temperature. The organic layer was taken and evaporated in vacuo, and the residue was dissolved in 200 μl of the mobile phase for HPLC as described below; an aliquot (10 μl) was subjected to HPLC.
HPLC Conditions.
HPLC conditions were: a Hitachi (Yokohama, Japan) 655-12A liquid chromatograph equipped with a Hitachil-7480 fluorescence detector, a Rheodyne (Rohnert Park CA) model 7125 injector, and a Shimadzu (Kyoto, Japan) C-R3A Chromatopac data processor; column, Chiralcel OD (4.6 mm × 250 mm; Daicel Co., Ltd., Tokyo, Japan); mobile phase,n-hexane/ethanol/diethylamine (95:5:0.1, by volume); flow rate, 0.5 ml/min; column temperature, 40°C; detection, fluorescence 252/302 nm (excitation/emission). Kinetic analysis was performed using concentration ranges from 2.0 to 500 μM for BF racemate and from 1.0 to 250 μM for BF enantiomers. Apparent Michaelis constants (Km) and maximal velocities (Vmax) were calculated using Eadie-Hofstee plots and least-squares analysis.
Results
As shown in a typical chromatogram (Fig.2), four authentic 1′′-OH-BF metabolites were sufficiently separated by employing a chiral column, Chiralcel OD (4.6 mm × 25 cm), and a mobile phase ofn-hexane/ethanol/diethylamine (95:5:0.1, by volume). The sequence of elution of the individual diastereomers in the microsomal samples was confirmed by spiking with each of the optically pure BF enantiomers. Blank microsomal extracts yielded no interfering peaks deriving from endogenous substances.
Calibration curves made by spiking known amounts of four 1′′-OH-BF diastereomers into ice-cold incubation medium and by extracting without incubation showed good linearity with correlation coefficients of at least 0.9995 for all of the metabolites within the concentration range evaluated (0.1–50 μM). The lowest detection limit, defined as 3 times the levels of baseline noise (S/N ratio = 3) was 0.015 μM for 1′′R-OH-R-BF, 1′′S-OH-R-BF, and 1′′R-OH-S-BF and 0.020 μM for 1′′S-OH-S-BF (injection volume was 10 μl of a 200-μl sample). The intra-assay variation employing three BF metabolite concentrations (0.1, 1.0, and 10.0 μM) gave coefficients of variation ranging from 0.45 to 5.99%.
Using the newly developed HPLC method, the metabolic profile of BF racemate or enantiomers produced by human liver microsomes was then examined. In this study, human liver microsomes from four subjects were used, and their drug-metabolizing enzyme activities are summarized in Table 1. Figure 3shows the formation of 1′′-OH-BF diastereomers from BF racemate (200 μM) or enantiomers (100 μM) by four lots of the microsomal fractions (HLM-003, HLM-014, HLM-059, and HLM-086). Among the four lots, HLM-014 showed the highest BF 1′′-hydroxylase activity followed by HLM-059, HLM-003, and HLM-086. The activities for BF racemate tended to be low compared with those for BF enantiomers, suggesting a metabolic interaction between BF enantiomers for the CYP enzyme(s) responsible for BF oxidation.
Interestingly, the three microsomal fractions (HLM-003, HLM-014, and HLM-086) showed similar profiles in the formation of 1′′-OH-BF diastereomers, but HLM-059 exhibited a different profile. That is, HLM-003, HLM-014, and HLM-086 showed a selectivity of 1′′R-OH < 1′′S-OH for both R-BF and S-BF, whereas HLM-059 exhibited the reverse selectivity of 1′′R-OH > 1′′S-OH for bothR-BF and S-BF. In Table 1, HLM-059 showed markedly higher activities than HLM-014 for S-mephenytoin 4′-hydroxylation and chlorzoxazone 6-hydroxylation as indices of CYP2C19 and CYP2E1, respectively. HLM-86 showed a chlorzoxazone 6-hydroxylase activity comparable to that of HLM-059. In addition,Mankowski (1999) reported that CYP2C19 also had a capacity to oxidize BF to 1′′-OH-BF. Therefore, CYP2C19 was thought to be mainly responsible for the apparently different selectivity exhibited by HLM-058 in the formation of 1′′-OH-BF.
To confirm this notion, the selectivity in the diastereomeric 1′′-OH-BF formation by recombinant CYP2D6 and CYP2C19 expressed in insect cells was examined in the second step. In this experiment, BF enantiomers (100 μM each) instead of BF racemate were employed as the substrate. In Fig. 3, the metabolic profiles of BF enantiomers by recombinant CYP2D6, CYP2C19, CYP1A2, and CYP3A4 are compared with the profile produced by human liver microsomes. The profile of CYP2D6 was similar to those of human liver microsomal fractions (HLM-003, HLM-014, and HLM-086), and the profile of CYP2C19 was similar to that of HLM-059. The selectivity (1′′R-OH < 1′′S-OH for bothR-BF and S-BF) in BF 1′′-hydroxylation by CYP1A2 and CYP3A4 was similar to that of CYP2D6, but the activities of CYP1A2 and CYP3A4 as BF 1′′-hydroxylases were much lower than those of CYP2D6 and CYP2C19.
In the third step, the kinetics for the formation of 1′′-OH-BF diastereomers from BF enantiomers by human liver microsomes and recombinant enzymes were studied. Typical Eadie-Hofstee plots for BF 1′′-hydroxylation by the human liver microsomal fraction (HLM-014) are shown in Fig. 4. 1R′′-Hydroxylation of BF enantiomers gave biphasic kinetics, but 1′′S-hydroxylation of BF enantiomers was monophasic. Similar results were obtained in kinetic experiments using HLM-003 and HLM-086 as enzyme sources. Kinetic parameters calculated are listed in Table2. Interestingly, all four 1′′-hydroxylation products of BF enantiomers by HLM-059 were found to be biphasic, whereas those produced by insect cell microsomes expressing recombinant CYP2D6, CYP2C19, CYP1A2, or CYP3A4 yielded monophasic kinetics. From the clearance values as well as theVmax values, the selectivity (1′′R-OH < 1′′S-OH for BF enantiomers) in BF 1′′-hydroxylation by human liver microsomal fractions HLM-003, HLM-014, and HLM-086 was similar to that by recombinant CYP2D6, whereas the selectivity (1′′R-OH >1′′S-OH for BF enantiomers) by human liver microsomal fraction HLM-059 was similar to that by recombinant CYP2C19.
As the last step, inhibition studies using quinidine as a CYP2D6 inhibitor (Newton et al., 1995) and omeprazole as a CYP2C19 inhibitor (Ko et al., 1997) were conducted to further confirm the involvement of the two CYP enzymes as major determinant enzymes for the diastereoselectivity in the metabolite formation in human liver microsomal fractions HLM-086 and HLM-059. As shown in the left panel of Fig. 5, quinidine inhibited BF 1′′-hydroxylase activity in HLM-086 in a concentration-dependent manner. As compared with the formation of 1′′R-OH metabolites from both BF enantiomers, the formation of 1′′S-OH metabolites was more markedly inhibited by the inhibitor. On the other hand, 1′′-hydroxylation of either BF enantiomer (100 μM each) by HLM-059 was not remarkably inhibited by quinidine (5 μM) (right panel of Fig. 5), at which concentration BF 1′′-hydroxylation by HLM-086 was markedly inhibited (left panel of Fig.5). It is noteworthy that approximately 40% of enantiomeric BF 1′′S-hydroxylase activities of HLM-059 (right panel of Fig.5, B and D) were decreased by the addition of quinidine.
Figure 6 shows the effects of omeprazole (10 μM) on BF 1′′-hydroxylation by human liver microsomes. The substrate concentration of 20 μM was chosen on the basis of theKm values (ca. 20 μM) for BF 1′′-hydroxylation by recombinant CYP2C19 (Table 2). Omeprazole did not remarkably affect 1′′-hydroxylation of BF enantiomers by HLM-086 at a concentration of 10 μM (left panel of Fig. 6), but enantiomeric BF 1′′R-hydroxylase activities of HLM-086 tended to be decreased by the addition of the inhibitor. In contrast, it efficiently inhibited the activity by HLM-059 (middle panel of Fig. 6). To further confirm its inhibitory effect on CYP2C19 under the conditions used, omeprazole (10 μM) was added to the reaction medium containing lymphoblastoid cell microsomes expressing recombinant CYP2C19. As shown in the right panel of Fig. 6, BF 1′′-hydroxylation by recombinant CYP2C19 was clearly decreased by the inhibitor.
Discussion
The results obtained in this study indicate that CYP2D6 in HLM-003, HLM-014, and HLM-086, and CYP2C19 in HLM-059 are the determinant enzymes for the diastereoselectivity in BF 1′′-hydroxylation by human liver microsomal fractions. The former three lots of human liver microsomes (HLM-003, HLM-014, and HLM-086) having the CYP2D6-type diastereoselectivity commonly showed biphasic kinetics for 1′′R-hydroxylation of BF enantiomers and monophasic kinetics for 1′′S-hydroxylation, whereas all four 1′′-hydroxylation products of BF enantiomers catalyzed by HLM-059 gave biphasic kinetics. These results suggest that some CYP enzyme(s) other than CYP2D6 or CYP2C19 are also involved in BF 1′′-hydroxylation by these lots of human liver microsomes.
As described before, Weerawarna et al. (1991) demonstrated that rat liver microsomes showed the diastereomer selectivity of 1′′R-OH > 1′′S-OH for BF enantiomers, which is coincident with that of CYP2C19, revealed in this study. In their data, human liver microsomes exhibited a reverse selectivity of 1′′R-OH < 1′′S-OH for BF enantiomers, which agrees with that of CYP2D6 in the present study. From its relatively high clearance values for BF 1′′R-hydroxylation, CYP2C19 may be a human liver microsomal component that causes biphasic kinetics in BF 1′′R-hydroxylation by human liver microsomal fractions.
In support of this notion, omeprazole and quinidine decreased BF 1′′S-hydroxylase activities of HLM-059 and BF 1′′R-hydroxylase activities of HLM-086, respectively, not remarkably but to some extent in the present inhibition study, indicating that biphasic kinetics observed in human liver microsomal fractions used are caused mainly by CYP2D6 and CYP2C19. As typically shown by the case of HLM-059, when BF 1′′-hydroxylation is used as an index of CYP2D6, there is the possibility that unless not only BF 1′′-hydroxylase activities but also the stereoselectivity is examined, a liver microsome sample that has high CYP2C19 but low CYP2D6 (or from a CYP2D6-poor metabolizer) could be mistakenly identified as a CYP2D6-extensive metabolizer sample.
Yamazaki et al. (1994) reported that CYP1A2 also contributes to BF oxidation although with a rather high Km value. In the present study, the capacities of recombinant CYP1A2 and CYP3A4 as BF 1′′-hydroxylases were examined, and their diastereoselectivity was found to be similar to that of CYP2D6. However, Km values of CYP1A2 and CYP3A4 were 10- to 20-fold those of CYP2D6, and clearance values for CYP1A2 and CYP3A4 were much lower than those of CYP2D6. These results suggest that CYP1A2 and/or CYP3A4 may become the determinant enzymes for the selectivity favoring BF 1′′S-hydroxylation in microsomal fractions from CYP2D6-deficient human livers or from human livers in which CYP1A2 or CYP3A4 are induced. From these results and speculations, it should be noted that BF 1′′-hydroxylation is not a simple, but a complicated index to which various CYP enzymes such as CYP2D6, CYP2C19, CYP1A2 and CYP3A4 having different diastereoselectivities may contribute.
It is noteworthy that CYP2C19 shows a reverse diastereoselectivity to that of CYP2D6. On the basis of the relationship between oxidative activities and chemical structures of substrates, various pharmacophore models have been proposed for the active site structures of CYP enzymes. The success of crystallization of partially modified CYP2C5 (Cosme and Johnson, 2000) has been of great help in making more accurate models. The results obtained in this study should provide important information for understanding similarities and differences between active site structures of CYP2D6 and CYP2C19.
In summary, a new chiral HPLC method to efficiently separate four 1′′-OH-BF diastereomers, major metabolites of BF, was developed. Using this method, diastereomer selectivity in the formation of 1′′-OH-BF from BF racemate or enantiomers was examined in four human liver microsomal fractions. Three microsomal fractions showed a selectivity of 1′′R-OH <1′′S-OH for BF enantiomers, which was similar to that of recombinant CYP2D6 expressed in insect cells, whereas one human liver microsomal fraction gave a selectivity of 1′′R-OH >1′′S-OH for BF enantiomers, which was similar to that of recombinant CYP2C19. Recombinant CYP1A2 and CYP3A4 showed a selectivity similar to that of CYP2D6, but their BF 1′′-hydroxylase activities were much lower than those of CYP2D6. quinidine, a known CYP2D6 inhibitor, markedly inhibited BF 1′′-hydroxylation in the fractions of human liver microsomes that showed the CYP2D6-type selectivity, and omeprazole, a known CYP2C19 inhibitor, efficiently suppressed the formation of 1′′-OH-BF diastereomers from BF in the microsomal fraction that exhibited the CYP2C19-type. From these results, it was concluded that the diastereomer selectivity in the formation of 1′′-OH-BF from BF are reversed between CYP2D6 and CYP2C19, both of which can be determinant enzymes in the diastereoselective 1′′-hydroxylation of BF in human liver microsomes.
Footnotes
-
DOI: 10.1124/jpet.102.036533
- Abbreviations:
- CYP
- cytochrome P450
- BF
- bufuralol
- R-BF
- 1′R-OH-bufuralol
- S-BF
- 1′S-OH-bufuralol
- 1′′R-OH-R-BF
- 1′′R-OH-1′R-bufuralol
- 1′′S-OH-R-BF
- 1′′S-OH-1′R-bufuralol
- 1′′R-OH-S-BF
- 1′′R-OH-1′S-bufuralol
- 1′′S-OH-S-BF
- 1′′S-OH-1′S-bufuralol
- BTL
- bunitrolol
- DB
- debrisoquine
- 4-OH-DB
- 4-hydroxydebrisoquine
- G-6-P
- glucose 6-phosphate
- Received March 20, 2002.
- Accepted May 28, 2002.
- The American Society for Pharmacology and Experimental Therapeutics