Abstract
H 259/31 is a substituted benzimidazole developed as a structural analog of omeprazole. Metabolites of H 259/31 formed in human liver microsomes were identified by using the synthetic reference compounds and liquid chromatography/mass spectrometry. The predominant metabolic pathways found include oxidation of the sulfoxide to sulfone, oxidativeO-dealkylation of the cyclopropylmethoxy group to the corresponding pyridone and aromatic hydroxylation to give the phenolic derivative. Stereoselectivity in the metabolism of the enantiomers of H 259/31 was demonstrated in human liver microsomes. The sum of the formation intrinsic clearances of all three metabolites was higher for the S-enantiomer than that of the R-form, indicating that the S-enantiomer is eliminated more rapidly. It was also shown in the present study that the sulfone metabolite is subject to additional metabolism, which should be taken into account when determining the intrinsic clearance for formation of metabolites and when the relative importance of metabolic pathways is determined. Expressed enzymes indicate major involvement of cytochrome P-450 (CYP) 2C19 in the formation of the hydroxy derivative as well as in pyridone formation from the enantiomers of H 259/31. CYP3A4 and CYP2C9 seem to contribute as low-affinity enzymes in both reactions. The sulfone metabolite was formed mainly from CYP3A4. Stereoselectivity in CYP3A4-, CYP2C19-, and CYP2C9-mediated metabolic pathways was demonstrated.
The metabolism of substituted benzimidazoles such as omeprazole, lansoprazole, and pantoprazole is to a large extent dependent on cytochrome P-450 (CYP)13A4 and CYP2C19 (Simon, 1995; Andersson, 1996; Pearce et al., 1996). CYP3A4 metabolizes the sulfoxide functionality; the major site for CYP2C19 metabolism is the pyridine group for omeprazole and pantoprazole; and lansoprazole is mainly metabolized in the benzimidazole group (Pichard et al., 1995). Because CYP2C19 is polymorphically expressed in the human population, a small number of individuals who lack CYP2C19 will metabolize the benzimidazoles more slowly than the average. The dependence on CYP2C19 for metabolic clearance seems to be similar for the three benzimidazoles (Andersson, 1996; Andersson et al., 1998).
The sulfoxide functionality of these benzimidazole derivatives is a chiral center, and the enantiomers of, e.g., omeprazole has been shown to be stereoselectively metabolized in human liver microsomes and by expressed CYP2C19 with a higher CLint forR-omeprazole than for S-omeprazole (A.Ä., T.B.A., M. Antonsson, A.K. Naudot, I.S., and L.W., unpublished data). Furthermore, in vivo data clearly demonstrates enantioselective metabolism of omeprazole (Tybring et al., 1997), lansoprazole (Katsuki et al., 1996), and pantoprazole (Tanaka et al., 1997).
H 259/31 is a substituted benzimidazole and a structural analog of omeprazole (Fig. 1). The aim of the present work was to study the in vitro metabolism of the enantiomers of H 259/31 using human liver microsomes and expressed enzymes to identify metabolic pathways and to explore whether the different functional groups would result in other routes of metabolism as compared with other benzimidazoles. Special attention was paid to enantioselective metabolism.
Materials and Methods
Chemicals.
The following compounds were synthesized at Medicinal Chemistry, AstraZeneca R&D: racemic H 259/31 (R,S-5-fluoro-2-[[(4-cyclopropylmethoxy-2-pyridinyl)methyl]sulfinyl]-1H-benzimidazole) and its two enantiomers, H 326/42 ((S)-5-fluoro-2-[[(4-cyclopropylmethoxy-2-pyridinyl)methyl]sulfinyl]-1H-benzimidazole) and H 326/43 ((R)-5-fluoro-2-[[(4-cyclopropylmethoxy-2-pyridinyl)methyl]sulfinyl]-1H-benzimidazole) as the sodium salts, as well as the 3H-labeled racemate [R,S-5-fluoro-2-[[(6-[3H]4-cyclopropylmethoxy-2-pyridinyl) methyl]sulfinyl]-1H-benzimidazole; radiochemical purity by liquid chromatography (LC) 98%] and the main metabolites formed in human liver microsomes, H 259/84 (5-fluoro-2-[[(4-cyclopropylmethoxy-2-pyridinyl)methyl]sulfonyl]-1H- benzimidazole),H326/54(R,S)-5-fluoro-6-hydroxy-2[[(4-cyclopropylmethoxy-2-pyridinyl)methyl]sulfinyl]-1H-benzimidazole), and H 287/00 (5-fluoro-2 [[(4-oxo-1,4-dihydropyridin-2-yl)methyl]thio]-1H-benzimidazole) (see Fig. 1 for all structures and trivial names used throughout the text) and the internal standard, hydroxyomeprazole (R,S)-5-methoxy-2[[(4-methoxy-3-methyl-5-hydroxymethyl-2-pyridinyl)methyl]sulfinyl]-1H-benzimidazole). The optical purity of H 326/42 and H 326/43 was determined to be 99.0 and 99.4% enantiomeric excess, respectively. NADP and isocitrate dehydrogenase were obtained from Boehringer Mannheim (Bromma, Sweden) and isocitrate from Fluka (Stockholm, Sweden). All other chemicals were of analytical grade.
Microsomal Incubation.
Human liver samples (excess material removed during surgery on the liver) were obtained from Department of Surgery 1, Sahlgrenska Hospital, Göteborg, Sweden. The use of code numbers to identify the liver specimens ensured full protection of the privacy of the patients. Small cubes (1 cm3) were frozen in liquid nitrogen and stored at −70°C until preparation of liver microsomes. Microsomes were prepared according to the method of Ernster et al. (1962), and microsomal protein concentration was measured according to Lowry et al. (1951) using BSA as standard. The total CYP content was measured according to Omura and Sato (1964). Microsomes from a human lymphoblastoid cell line expressing human CYP1A2, CYP2C9, CYP3A4, CYP2C19, CYP2C8, CYP2D6, or CYP2E1 were purchased from Gentest Corp. (Woburn, MA). The microsomal preparations were stored at −70°C until use.
Tritium-labeled H 259/31 (5 μM) was incubated at a protein concentration of 1 mg/ml (liver HL100) for identification of formed metabolites. To generate amounts of metabolites sufficient for identification by LC/mass spectrometry, the sample was incubated for 60 min.
For the kinetic experiments, linearity in the formation rate of metabolites was established with respect to microsomal protein concentration and incubation time. Based on these results, kinetic studies were performed with a 30-min incubation time at a protein concentration of 1 mg/ml at 37°C. Reactions were started by the addition of substrate after a preincubation period of 1 min. Nine different concentrations of the two enantiomers, ranging from 5 to 1000 μM, were used in the kinetic experiments. The substrates were dissolved in 10% dimethyl sulfoxide (DMSO) and used fresh. The final concentration of DMSO in the incubation mixture was always 0.5%. Controls were prepared as above at each substrate concentration, but the extraction solvent was added before the substrate and the controls were incubated at room temperature (∼20°C). Duplicates were made of both the incubated samples and the controls. Three different human livers (HL100, HL105, and HL106) were used in the kinetic experiments.
The rate of disappearance of the sulfone metabolite was studied in microsomes from one human liver (HL100). The sulfone was dissolved in carbonate buffer (25 ml of 0.5 M Na2CO3 and 62.5 ml of 1.0 M NaHCO3 added to 912.5 ml of deionized water)/methanol (4:1). The final concentration of methanol in the incubation medium was 1%. Five different concentrations of the sulfone, ranging from 5 to 100 μM, were used in the experiment.
Analytical Procedure.
The incubated samples for identification of metabolites in human liver microsomes were centrifuged at 1000g, and 50 μl of the supernatant was injected onto the column. The metabolites formed were separated on a Kromasil C18 (150 mm × 4.6 mm i.d.; EKA Nobel, Surte, Sweden) analytical column protected by an RP-2 NewGuard (15 mm × 3 mm i.d.; Applied Biosystems, Foster City, CA) guard column. The pump and gradient controller were LKB 2150 and LKB 2152, respectively (Pharmacia/LKB, Uppsala, Sweden). The mobile phases were 5 and 55% acetonitrile in 50 mM aqueous phosphate buffer (pH 7.8) and 1 mM tetrapropyl ammonium hydrogen sulfate. The linear gradient profile was 5 to 30% acetonitrile in 30 min at a flow rate of 1.0 ml/min. The eluate was monitored by on-line radioactivity detection (model LB-501C; Berthold, Wildbad, Germany) in the homogeneous counting mode after a postcolumn addition of 4 ml/min of scintillation fluid (Rialuma; Lumac, Landgraaf, the Netherlands). The metabolites were identified using LC/mass spectrometry. In this case, the mobile phases were 5 and 55% acetonitrile in 2 mM aqueous ammonium acetate (pH 7). The mass spectrometer was a triple stage quadrupole operated in the atmospheric pressure ionization mode with an ion spray interface (API III; Perkin-Elmer Sciex, Thornhill, Canada). Retention times equal to those of the synthetic reference compounds, molecular ion mass, and fragmentation in the MS/MS mode were considered sufficient structural evidence for the metabolites depicted in Fig. 1.
The concentrations of the substrate and the three identified metabolites (in the following text referred to as sulfone, 6-hydroxy, and pyridone, Fig. 1) of H 259/31 were determined with the method described previously for omeprazole (Andersson et al., 1993a) with a few modifications. The quantitation of sulfone, 6-hydroxy, and pyridone was made by comparison with standard samples (containing known amounts of the synthesized racemic metabolites and microsomes but no NADPH) by using the peak area ratio method. Sulfone, 6-hydroxy, pyridone, and internal standard (hydroxyomeprazole) were dissolved in carbonate buffer (25 ml of 0.5 M Na2CO3 and 62.5 ml of 1.0 M NaHCO3 added to 912.5 ml of deionized water)/methanol (4:1); aliquots of these solutions were kept frozen (−20°C) until used. The samples were analyzed by an achiral method, and it was assumed that the metabolites formed retain the optical configuration of the parent enantiomer.
Calculations.
Estimates of enzyme kinetic variables were obtained by nonlinear regression analysis (PCNONLIN, version 4.2; SCI Software, Lexington, KY). Homoscedastic weighting was used in all regressions. The simple Michaelis-Menten equation, V = (Vmax · C)/(Km+ C), was fitted to the formation rates of 6-hydroxy and pyridone, respectively, versus substrate concentration, whereas a sigmoid model, equivalent to the Hill equation, V = (Vmax · CS)/(KmS+ CS), was fitted to the data on the formation rate of sulfone versus substrate concentration. The apparent maximum formation rate (Vmax) and the apparent substrate concentration resulting in one half of the maximum rate (Km) were estimated. In the case of the sulfone, a slope factor S, which is a parameter determined by the sigmoidicity of the curve, was also estimated. IntrinsicCLearance was calculated asCLint =Vmax/Km for the 6-hydroxy and pyridone and as CLint =Vmax/KmSfor the sulfone metabolite.
Microsome incubations with the metabolites indicated that the sulfone, but not the 6-hydroxy and the pyridone, was extensively more metabolized (data not shown). Because the sulfone itself has a rapid elimination when incubated, the rate of formation, i.e., theCLint for its formation, may be underestimated when data are treated in the traditional manner described above. An attempt was made to take the sequential metabolism of the sulfone into account when calculating the enzyme kinetic parameters for its formation. The principle of this alternative approach is to do a nonlinear regression analysis using a kinetic model describing the concentration of metabolite at the time of sampling (i.e., at 30 min) as a net effect of its formation and elimination. Because there is some loss of substrate during the incubation this is also accounted for. The rate of change of substrate and sulfone concentration can thus be expressed by two differential equations:
Incubations with cDNA-Expressed Human CYPs.
The formation rates of the metabolites related to CYP1A2, CYP2C9, and CYP3A4 activities were linear at 37°C for incubation times of up to 120 min when 20 or 200 μM H 259/31 was incubated with 0.5 to 1.5 mg/ml of microsomal protein and the NADPH-generating system. On the basis of these results, the kinetic studies were performed at a protein concentration of 1.0 mg/ml at 37°C over a time period of 120 min. Five different concentrations of the substrates, ranging from 20 to 500 μM, were used in these experiments. Furthermore, formation rate linearity of 6-hydroxy and pyridone by CYP2C19 was established for incubation times of up to 20 min with a microsomal concentration of 0.5 mg/ml and substrate concentrations of 2.5, 5, and 10 μM. The concentration range used in the kinetic study was 2.5 to 80 μM. The substrates were dissolved in 10% DMSO and used fresh. The final concentration of DMSO in the incubation medium was always 0.5%. Samples of the incubation mixture were taken in duplicate for quantitation of the 6-hydroxy and the pyridone metabolite. cDNA-expressed CYP2C8, 2D6, and 2E1 were shown to be unable to form any of the metabolites. Microsomes from lymphoblastoid cells transfected with the vector only were used as control and did not metabolize the compounds. The Michaelis-Menten equation was fitted to the unweighted data on the formation rates of metabolites and substrate concentrations using nonlinear regression (PCNONLIN, version 4.2, SCI Software). The maximum formation rate (Vmax) and the substrate concentration resulting in one half of the maximum rate (Km) were estimated. Intrinsic Clearance was calculated as
Results
Identification of Metabolites.
Metabolites formed from the tritium-labeled racemic parent compound in human liver microsomes were separated by LC with radiochemical detection. In this in vitro study, sulfoxidation to sulfone,O-dealkylation to the pyridone, and aromatic hydroxylation to the phenolic derivative were identified as the main metabolic pathways of H 259/31 (Fig. 1). The pyridone was quantified as the corresponding sulfide, as the dealkylation and the experimental conditions probably lead to the spontaneous nonenzymatic reduction of the sulfoxide functionality (S. von Unge, personal communication). A fourth minor metabolite was detected but not identified. About 50% of parent H 259/31 was consumed during the 60-min incubation. The three metabolites, sulfone, 6-hydroxy, and pyridone, were formed to 13, 35, and 38% in relation to the remaining parent compound, respectively. The identity of the metabolites was confirmed by comparison of LC retention times with the synthetic reference compounds and LC/mass spectrometry.
Formation of Metabolites in Human Liver Microsomes.
Representative Eadie-Hofstee plots for the formation of sulfone, 6-hydroxy, and pyridone are shown in Fig.2 a, b, and c, respectively. The plots for the O-dealkylation and the 6-hydroxylation were nonlinear for all three livers studied, indicating the involvement of multiple CYP isoforms in these reactions. Although the plots suggest that more than one enzyme is involved in the formation of these two metabolites, fitting of the data to a Michaelis-Menten expression for a two-enzyme system did not significantly improve the regression as compared with fitting of the data to a one-component expression. Parameters for a possible high-affinity component could thus not be estimated with the present data. The estimatedKm and Vmaxvalues when data was treated in the traditional way are shown in Table1, and the calculatedCLint values are shown in Table2. However, atypical Eadie-Hofstee plots were obtained for the formation of the sulfone (see Fig. 2a), and the common Michaelis-Menten equation did not result in a good fit to the data. Instead, a sigmoid Vmax model was used, resulting in the estimated parameters,Vmax and Km, and the sigmoid factor, S, shown in Table 1. The corresponding calculated CLint values are shown in Table 2. Incubation of the sulfone showed that its rate of disappearance was rapid during the 30-min experiment.Vmax was not obtained in the concentration interval studied and only CLint of the sulfone was estimated by linear regression. By taking this sequential metabolism into account when calculating the enzyme parameters for the formation of sulfone, substantially higher estimates ofCLint were obtained than when using the conventional method. The CLint estimates for the sulfone formation increased from 1.0 and 0.8 μl/min/mg, as obtained with the conventional method, to 9.7 and 1.6 μl/min/mg forS- and R-enantiomer, respectively, when sequential metabolism was taken into account. The estimated parameters generally had a reasonably high precision and the observed sulfone concentration was well predicted by the model (see Fig.3).
In all preparations, the formation, CLint, of 6-hydroxy, based on the conventional enzyme kinetic analysis, was much higher than that of sulfone and pyridone, suggesting this to be the most important route of metabolism. However, when the sequential metabolism of the sulfone was accounted for, the 6-hydroxy and the sulfone had similar rates of formation for the S-form, whereas formation of the 6-hydroxy metabolite was the dominant pathway for the R-form. A large interindividual variability was seen in the capacity of microsomes from different human livers to oxidize the parent compounds even though the relative importance of the metabolic pathways was similar in all preparations.
The formation, CLint, of the sulfone, 6-hydroxy, and pyridone of the S-form were 5.4, 3.1, and 1.4 times higher, respectively, than those of the R-form. The enantioselective rate of formation of the sulfone was evident also when the sequential metabolism of the sulfone was included in the calculations, the formation of the sulfone of the S-form being 5.9 times higher than that of the R-form using HL100 microsomes.
Formation of Metabolites by cDNA-Expressed Human CYPs.
The derived Km andVmax values and the calculatedCLint values are shown in Table3. According to these results, the formation, CLint, of sulfone in microsomes expressing CYP3A4 was significantly higher than that of 6-hydroxy and pyridone. A considerable stereoselectivity was observed in the formation of sulfone by CYP3A4. The formation,CLint, for the S- andR-enantiomers was 53.9 and 7.7 μl/min/nmol CYP, respectively. Incubations with microsomes expressing CYP1A2 also indicate major involvement of this isozyme in the formation of sulfone. Stereoselectivity was observed in the formation of 6-hydroxy by CYP1A2, the metabolite being formed nine times faster from theS-enantiomer than from the R-enantiomer. CYP2C9 was involved in the formation of all three metabolites and a great difference in 6-hydroxy formation between the two enantiomers was observed. The formation CLint of 6-hydroxy from the S-enantiomer was 42.4 μl/min/nmol CYP whereas that of the R-enantiomer was 4.6 μl/min/nmol CYP. The formation, CLint, of 6-hydroxy and pyridone metabolites in microsomes expressing CYP2C19 was shown to be very high, suggesting this to be a major (the dominant) enzyme in the production of these metabolites. Due to a very low Km, the enzyme kinetic parameter determinations show considerable variation. The sulfone metabolite, however, was not detected in CYP2C19 microsomes. In an attempt to predict CLintvalues in human liver microsomes by using the enzyme kinetic parameters from the cDNA-expressed enzymes in Table 3, we used the information from Shimada et al. (1994) and Yamazaki et al. (1997) on the average content of each CYP isoform in a microsomal preparation (Table4). The sum of the formation,CLint, of all three metabolites was 25.8 and 8.64 μl/min/mg microsomal protein, for the S- andR-forms, respectively, indicating that the S-form is cleared more rapidly than the R-form. The percentCLint was calculated for each isoenzyme, and CYP2C19 was found to contribute by 43 and 42%, respectively, to the total of the S- and R-form metabolism. CYP2C9 and CYP3A4 seem to be equally important for the metabolism of theS-form (25 and 29%, respectively), whereas the contributions of CYP2C9 and CYP3A4 are 37 and 16%, respectively, of the total metabolism of the R-form.
Discussion
The metabolic clearance of the S- andR-forms of H 259/31 in human liver microsomes is essentially dependent on the formation of sulfone, 6-hydroxy, and pyridone (Fig.1). The S-form is cleared more rapidly than theR-form, indicating a stereoselective metabolism of the enantiomers in human liver microsomes. It was also shown in this study that the sulfone itself was rapidly metabolized during the 30 min of incubation, which interfere with the kinetic calculations of sulfone formation. When the calculations were corrected for sequential metabolism, a substantially higher estimate for theCLint for sulfone formation was obtained. The difference in CLint was more pronounced for the S-form than for the R-form, which therefore potentiates the metabolic stereoselectivity. Furthermore, by taking sequential metabolism into account it was found that, instead of a minor metabolite, the sulfone formation was a major metabolite from the S-form and the relative importance of the metabolic pathways for CLint is: sulfone = 6-hydroxy > pyridone. Although theCLint of the sulfone formation from theR-form increased slightly when sequential metabolism was taken into account, the rank order for the metabolism of theR-form did not change (6-hydroxy > pyridone > sulfone). The results clearly show the importance of investigating the potential for sequential metabolism when kinetics of product formation is studied. If sequential metabolism is significant, the kinetics of the disappearance of the product has to be accounted for, preferably by kinetic modeling as suggested here.
It has been suggested that some CYP3A4-dependent reactions showing sigmoidal shapes when plotting initial metabolism rate versus substrate concentration (Michaelis-Menten plots) is due to activation of metabolism by the substrate itself (Schwab et al., 1988; Andersson et al., 1994; Shou et al., 1994; Ueng et al., 1997). As sulfoxidation of substituted benzimidazoles to sulfone is mainly mediated by CYP3A4, the cause of the atypical Eadie-Hofstee plots of sulfone formation from the enantiomers of H 259/31 may be explained by this concentration-dependent increase in activity of the enzyme. However, no metabolites were formed when the sulfone was incubated with cDNA-expressed CYP3A4. Consequently, the probable explanation for the atypical curves from the kinetic analysis of the sulfone formation is the sequential metabolism in human liver microsomes by isoenzymes other than CYP3A4.
Expressed enzymes were used to predict totalCLint for the metabolite formation and the relative importance of the individual enzymes. The predicted totalCLint in human liver microsomes based on the results from studies using expressed enzymes was in the same range as the value obtained in studies on human liver microsomes. The predicted clearance of the S-enantiomer was 3-fold that of the R-enantiomer. The totalCLint of the formation of individual metabolites from the expressed enzymes also predicted the formationCLint of the metabolites in human liver microsomes. These results suggest that expressed enzymes could be used to predict human liver microsomal metabolism of these benzimidazole derivatives.
Our studies using expressed enzymes suggest that CYP2C19 is involved in the metabolism of the enantiomers of H 259/31 and also that it is the quantitatively most important enzyme for the two enantiomers. The metabolic clearance in vivo should therefore be equally dependent on CYP2C19. A similar decrease in metabolic clearance of both enantiomers in CYP2C19 deficient subjects should therefore be expected.
Whereas CYP2C19 has been identified as the most important enzyme for the metabolism of H 259/31 and other benzimidazoles (Andersson et al., 1993b; Chiba et al., 1993; Simon, 1995; Pearce et al., 1996), CYP2C9 also seems to be an important enzyme for the metabolism of H 259/31. The pyridone and 6-hydroxy formation by CYP2C9 was responsible for 25 and 37% of total CLint of theS- and R-enantiomers, respectively.
Our studies show a clear stereoselectivity in the in vitro metabolic disposition of the enantiomers of H 259/31 with a 2-fold higher CLint for the S-form compared with the R-form in human liver microsomes. HigherCLint values were obtained for 6-hydroxy formation by CYP2C9 and CYP2C19, sulfone formation by CYP3A4, and pyridone formation by CYP2C19 when the S-enantiomer was used as substrate compared with those of the R-form. For omeprazole (Fig. 1), the overall in vivo metabolism in humans is also highly dependent on the activity of CYP2C19 (Andersson et al., 1993b;Andersson, 1996; Tybring et al., 1997). Thus, the formation of the main hydroxy derivatives of omeprazole and H 259/31 is catalyzed by the same enzyme. Interestingly, the hydroxylation takes place in opposite positions of the compounds, as H 259/31 is hydroxylated at carbon 6 of the benzimidazole moiety, and omeprazole is hydroxylated at the 5-methyl carbon of the pyridinyl group (Andersson et al., 1993a; Fig.4). The higher production of this metabolite from H 259/31 occurs when the S-form is docked into the active site of the enzyme. Sulfone formation has been described previously as a CYP3A4-dependent reaction for omeprazole (Andersson et al., 1993b) and other compounds of this benzimidazole class. This is true also for the enantiomers of H 259/31 although minor contributions of CYP2C9 and CYP1A2 were seen.
Stereoselective aspects of the human in vivo and in vitro metabolism in humans of omeprazole and other closely related compounds, e.g., lansoprazole and pantoprazole (Fig. 3), have recently appeared in the literature. The pharmacokinetics of the enantiomers of omeprazole after oral administration of the racemate to extensive and poor mephenytoin hydroxylators show that omeprazole is stereoselectively metabolized by CYP2C19 (Tybring et al., 1997). The plasma levels ofS-omeprazole were significantly higher than those of theR-form, suggesting that the hydroxy derivative was formed more readily from the R-enantiomer. In the case of lansoprazole, the plasma area under the curve (AUC) of theR-form is significantly higher than that of theS-form (Katsuki et al., 1996), i.e., the S-form is more rapidly cleared in humans. The most important enzyme in the in vitro metabolism of lansoprazole is CYP2C19, and the main route of metabolism is aromatic hydroxylation of the benzimidazole ring system (Karol et al., 1995; Pearce et al., 1996) (Fig. 3). As concerns pantoprazole, a preliminary report on the in vitro metabolism using human liver microsomes states that the main route of metabolism of racemic pantoprazole is demethylation of the 4-methoxypyridine group (Fig. 3) and that CYP3A4 and CYP2C19 are responsible for its formation (Simon, 1995). Thus far, data on the absolute configuration of the enantiomers of pantoprazole have not been published. However, mechanistic studies on titanium-tartrate-mediated enantioselective sulfoxidation of the corresponding sulfide suggest that (+)-pantoprazole is of the R-configuration (S. von Unge, personal communication). In the following discussion on stereoselectivity in the metabolism of the enantiomers of pantoprazole, we assume that (+)-pantoprazole is of the R-configuration by analogy to the enantiomers of omeprazole (von Unge et al., 1997), H 259/31, and lansoprazole (Katsuki et al., 1996). The pharmacokinetics of the enantiomers of pantoprazole was investigated in human subjects classified as extensive (EM) and apparent poor metabolizers (PM) (Tanaka et al., 1997), according toS-mephenytoin hydroxylation capacity (Sanz et al., 1989). In the EMs, the AUC values of S-pantoprazole were only slightly higher than those of R-pantoprazole, indicating that theR-form is the more favorable substrate of the two enantiomers. However, the small difference between the AUC values implies that the stereoselectivity in clearance is very low, which is in contrast to observations made for lansoprazole as outlined above. Interestingly, significant differences in the pharmacokinetics ofR- and S-pantoprazole were observed in the PMs in which the metabolism occurs without the influence of CYP2C19. The AUC values were 2.65 to 3.45 times greater for R-pantoprazole than for S-pantoprazole, indicating that the disposition in the PMs is the result of significant impairment in the metabolism ofR-pantoprazole, and that other enzymes participating in the elimination of pantoprazole e.g., CYP3A4, also exhibit considerable stereoselectivity. In the EMs, the stereoselectivity of the multiple enzymes involved in the metabolism e.g., CYP2C19 and CYP3A4, is probably counter directed so that there are no apparent differences in the overall disposition of the two enantiomers (Tanaka et al., 1997).
Taken together, the observations on the stereoselectivity in the human CYP2C19-mediated metabolism of compounds of the pyridinylsulphinylbenzimidazole class indicate that theR-enantiomers are preferentially metabolized in the pyridine group whereas the S-enantiomers are subject to metabolism in the benzimidazole by this enzyme (Fig. 3). Obviously, the conformation at the chiral center at the sulfur of the sulfoxide functionality plays an important role in directing the metabolic transformation to each respective side of the molecule. The sulfoxide oxygen, as a strong hydrogen acceptor, may thus be crucial to the favorable positioning of the substrate into the active site of the enzyme. As pointed out bySmith et al. (1997), there seems to be a prerequisite for a hydrogen bond-accepting functionality such as a sulfoxide oxygen for the successful docking of the substrate into the active site. Hydroxylation then takes place at a carbon five to seven bond lengths from the hydrogen bond acceptor. In the case of H 259/31, the lack of sulfone formation by CYP2C19 supports that the sulfoxide group is required for the stereoselective docking of the substrates into the active site of this enzyme. Sulfoxidation of the enantiomers of H 259/31, on the other hand, was mainly performed by CYP3A4, which also exhibited considerable stereoselectivity in favor of the S-form.
Ibeanu et al. (1996) identified the critical amino acids that determine the specificity of human CYP2C19. A model was proposed, based on the results of racemic omeprazole, showing omeprazole docked into the active site of the enzyme. S-mephenytoin (which is the model substrate of this enzyme) did not fit the model, however, and it was suggested that the specificity of CYP2C19 for S-mephenytoin requires a more complex enzyme configuration. On the basis of our results and published data on omeprazole and analogous compounds, however, we believe that the reason for the discrepancy is that racemic omeprazole was used and that it would be of great importance to test the single enantiomers in the proposed model. According to the model speculated on in our paper, one can expect that R-omeprazole and S-mephenytoin would fit into the same model.
In conclusion, the same CYP isoforms are involved in the metabolism of H 259/31 and omeprazole, despite the different functional groups at the pyridinylsulphinylbenzimidazole core structure. Sulfone formation is common to both compounds and is mediated via CYP3A4. Hydroxylation is performed by CYP2C19 as the main route of metabolism, albeit in opposite parts of the molecules. The absolute configuration of the sulfoxide appears to be important for directing hydroxylation to the most favorable position, leading to considerable stereo- and regioselectivity in the metabolic disposition of the enantiomers of H 259/31 by this enzyme.
Footnotes
-
Send reprint requests to: Angela Äbelö, AstraZeneca R&D Mölndal, S-431 83 Mölndal, Sweden. E-mail:angela.abelo{at}hassle.se.astra.com
- Abbreviations used are::
- CYP
- cytochrome P-450
- DMSO
- dimethyl sulfoxide
- EM
- extensive metabolizer
- PM
- poor metabolizer
- LC
- liquid chromatography
- AUC
- area under the curve
- Received December 31, 1998.
- Accepted September 15, 1999.
- The American Society for Pharmacology and Experimental Therapeutics