Abstract
The oxidative metabolism of irbesartan, a new nonpeptide angiotensin II receptor antagonist, was investigated on 12 human fully characterized hepatic microsomes and purified cytochrome P-450 (CYP) isoforms. After incubation of microsomes with irbesartan and NADPH, four main hydroxy metabolites were formed, as confirmed by liquid chromatography-mass spectrometry analysis. Irbesartan oxidation follows Michaelis-Menten kinetics, consistent with the involvement of a single CYP isoform in these hydroxylation processes. Only a low interindividual variability (2-fold difference) was observed in drug oxidation, even in preparations lacking CYP2D6.Km and Vmax for irbesartan oxidation were 54 ± 6.5 μM and 0.62 ± 0.18 nmol/min/mg, respectively. Irbesartan oxidation correlated (r2 = 0.769) with tolbutamide (CYP2C9 substrate) 4-methyl-hydroxylation. Oxidation of irbesartan was markedly inhibited by sulfaphenazole (CYP2C9 inhibitor), but not by any of several other CYP inhibitors. In the same manner, both tolbutamide and warfarin (CYP2C9 substrates), were competitive-type inhibitors of irbesartan oxidation with Ki values of 500 and 30 μM, respectively. Moreover, irbesartan was a competitive-type inhibitor of tolbutamide 4-methylhydroxylation (Ki = 317 μM). Nifedipine also potentially decreased irbesartan oxidation, whereas neither ketoconazole and triacetyloleandomycin (CYP3A inhibitors), nor diltiazem and verapamil, (CYP3A4 substrates), exhibited an inhibitory effect. Additional studies demonstrated that nifedipine was an inhibitor of irbesartan (Ki = 20 μM) and tolbutamide oxidation processes, whereas irbesartan had no effect at all on nifedipine dehydrogenation. Enzyme kinetics suggest that nifedipine is a noncompetitive-type inhibitor of CYP2C9-mediated catalytic activities. Finally, only microsomes containing recombinant human liver CYP2C9 were capable of oxidizing irbesartan. These results provide evidence that CYP2C9 plays a major role in irbesartan oxidation.
Irbesartan is a new orally active nonpeptide angiotensin II receptor antagonist, marketed for the treatment of hypertension (Cazaubon et al., 1993). Previous studies performed on both human hepatic microsomal fractions (Perrier et al., 1994) and human hepatocytes in primary culture (G. Fabre, personal communication) demonstrated two primary routes of metabolism for irbesartan, including monohydroxylation at four different sites of the molecule and N-glucuronidation of the tetrazole moiety (Perrier et al., 1994).
Numerous approaches have been developed to identify the human cytochrome P-450 (CYP)1 isoform(s) (Guengerich, 1992;Nelson et al., 1993; Newton et al., 1995) involved in the oxidative metabolism of any given xenobiotic agent in vitro, including: 1) correlation of activities of isoform-selective substrates across a bank of human liver microsomes; 2) competitive inhibition of isoform-selective substrates with human liver microsomes; 3) effects of isoform-selective inhibitory probes, i.e.,Ki determinations, specificity-concentration relationship, and type of interaction; 4) determination of drug metabolism by heterologous human cells expressing drug-metabolizing enzymes (Remmel and Burchell, 1993); 5) characterization of drug metabolism in the presence of specific inhibitory antibodies; and 6) correlation of activity of considered pathway with specific CYP isoforms level of expression, as assessed by immunoblotting or enzyme-linked immunosorbent assays.
It is now well established that a combination of these approaches is required for the identification of the enzyme(s) responsible for the metabolism of a xenobiotic agent with a high degree of certainty.Rodrigues (1994) recently reported that selective chemical inhibitors should be fully characterized to determine a “window of selectivity” because most of these compounds, at high enough concentrations, will inhibit one or more additional CYP isoforms. By carrying out inhibitory studies, Bourrié et al. (1996) determined that at 10-fold their respective Kmand Ki values, substrates and inhibitors, respectively, demonstrated a selective inhibitory effect.
The main objective of this study was to identify the CYP isoform(s) responsible for the oxidation of irbesartan to its different hydroxy metabolites. Our approach involved 1) the screening of individual human liver microsomes previously characterized for their ability to metabolize specific CYP isoform-substrates; 2) the evaluation of isoform-specific CYP, either substrates or inhibitors, on irbesartan oxidation; 3) the metabolic characterization of the “substrate-irbesartan” and “inhibitor-irbesartan” interactions; and 4) the metabolic capacity of recombinant human liver CYP isoforms to oxidize irbesartan.
The characterization of the isoform(s) responsible for the in vitro metabolism of irbesartan has provided useful information on the design of clinical studies evaluating drug-drug interactions in vivo.
Materials and Methods
Chemicals.
Irbesartan (code SR47436-BMS 186295; Sanofi Recherche, Montpellier, France) was synthesized by the Chemical Development department of Sanofi Sisteron (Montpellier, France). Its chemical structure is illustrated in Fig. 1. Drug substrates and inhibitors used in this study were obtained from the following sources: dextromethorphan was a kind gift from Dr. B. Lacarelle (Faculté de Pharmacie, Marseilles, France); and furafyllin, sulfaphenazole, S-mephenytoin, tolbutamide and 4-methyl-hydroxytolbutamide were purchased from Ultrafine Chemicals (Manchester, United Kingdom). Aniline, ketoconazole, quinidine, diallyldisulfide, coumarin, phenacetin, α-naphthoflavone, nifedipine and its oxidized metabolite, and NADPH were obtained from Sigma Chemical Co. (St Louis, MO), and 7-ethoxyresorufin was purchased from Pierce Chemical Co. (Rockford, IL). All other chemicals and reagents used were of the highest commercially available quality.
Biological Preparations.
Individual human liver samples (n = 12) termed “HTL-x” were obtained from patients with secondary hepatic tumors. Immediately after removal, the samples were transferred in ice-cold Eurocollins buffer, transported to the laboratory, frozen in dry ice, and stored at −80°C until microsomes were prepared as described previously (Fabre et al., 1988a,b). Microsomes were stored at −80°C in 0.1 M KH2PO4buffer (pH 7.4) containing 0.1 mM EDTA and 20% glycerol (v/v).
Human liver microsomes were characterized for CYP isoform activity, and they exhibited considerable variability between microsomal samples in terms of catalytic activity (Table 1). Analytical methods used to evaluate isoform activities were previously reviewed (Bourrié et al., 1996).
Microsomes containing recombinant human liver CYP isoforms were obtained from Gentest Corp. (Woburn, MA). These included microsomes prepared from B lymphoblastoid cell line expressing CYP1A1 (20 pmol of P450/mg protein), CYP1A2 (40 pmol of P450/mg protein), CYP2A6 (60 pmol of P450/mg protein), CYP2B6 (55 pmol of P450/mg protein), CYP2C9 (10 pmol of P450/mg protein), CYP2D6 (160 pmol of P450/mg protein), CYP2E1 (40 pmol of P450/mg protein) and CYP3A4 (44 pmol P450/mg protein), as well as from control cells (0 pmol of P450/mg protein). The CYP content, activity, and linearity of product formation in these preparations were assumed to be the same as those asserted by the manufacturer. Incubations (80 pmol/ml) were conducted in 0.1 M KH2PO4 buffer (pH 7.4) after a 1-h incubation at 37°C, with 50 μM irbesartan and 1 mM NADPH.
Assays.
Total CYP content and proteins concentration were determined by the methods of Omura and Sato (1964) and Pollard et al. (1978), respectively. Incubations performed in 0.1 M KH2PO4 buffer (pH 7.4), were carried out at 37°C under continuous stirring.
Tolbutamide 4-methyl-hydroxylation was evaluated as described by Miners et al. (1988). Human hepatic microsomes (2 mg/ml; final protein concentration) were incubated for 30 min with increasing tolbutamide concentrations ranging from 83.3 to 500 μM and 1 mM NADPH. Enzymatic reaction was stopped by the addition of one volume of 0.15 M orthophosphoric acid. After removal of precipitated proteins by centrifugation, the incubation mixture was analyzed by high-pressure liquid chromatography (HPLC). The chromatograph was fitted with a C18 μBondapak column (300 × 3.9 mm; particle size = 10 μm; Waters Chromatography Div., Milford, MA). Unknown concentrations of 4-methyl-hydroxy-tolbutamide were determined by comparison of peak area with those of a calibration curve prepared with 4-methyl-hydroxy-tolbutamide concentrations ranging from 1 to 50 μM. The mobile phase was ammonium acetate buffer (10 mM, pH 4.3)-acetonitrile. A solvent programmer was set to deliver 20 to 40% acetonitrile over a 15-min linear gradient, followed by a second 5-min linear gradient from 40 to 75% acetonitrile.
Nifedipine oxidation was measured as previously described by Guengerich et al. (1986): human hepatic microsomes (final protein concentration = 0.25 mg/ml) were incubated for 10 min with increasing nifedipine concentrations ranging from 20 to 100 μM and 1 mM NADPH. Proteins precipitated with one-half volume of an acetonitrile/20% trichloroacetic acid mixture (50:50), were removed by centrifugation (2500g for 10 min) and the clear supernatant was analyzed by HPLC (ODS-2 column, 5-μm particle size; flow rate = 1.5 ml/min; UV absorbance = 254 nm; 0.1% diethylamine buffer adjusted at pH 6.0 with acetic acid/acetonitrile, 65/35; v/v).
Irbesartan Oxidation.
The standard incubation conditions were chosen based on the results of preliminary experiments varying both concentrations of substrate and microsomal proteins. Irbesartan (final dimethyl sulfoxide concentration in incubation medium never exceeded 0.1%) was used at concentrations ranging between 10 and 100 μM. At this 0.1% final concentration, dimethyl sulfoxide did not decrease the rate of irbesartan oxidation when compared with methyl alcohol as the organic solvent. Chauret et al. (1998) already reported that a 0.2% final DMSO concentration had no effect on CYP1A2, -2A6, and -2D6, but decreased by 10 to 20% CYP2C9 and -2C19, by 25% CYP3A4, and by more than 60% CYP2E1. Incubations performed in 0.1 M KH2PO4buffer were initiated by 1 mM NADPH in a water bath shaker at 37°C. Enzyme reaction was stopped by the addition of one volume of 20% trichloroacetic acid/acetonitrile (50:50) for one volume of incubation mixture.
Oxidation rate was determined by the amount of oxidized metabolite(s) formed per unit of time. Indeed, under optimal conditions, with saturable levels of substrate, determination of the quantity of metabolite(s) formed is a more accurate measurement than the disappearance of small amounts of substrates.
When studies were performed in the presence of specific CYP isoforms, substrates, and/or inhibitors, the latter were added in the incubation mixture just before NADPH addition, except when triacetyloleandomycin was used. For this CYP inhibitor, microsomes were first incubated with inhibitors and NADPH for 15 min to generate the suicide metabolite(s) (Chang et al. 1994). Then, preincubated microsomes were incubated with irbesartan and NADPH for an additional selected period of incubation at 37°C, and the formation of hydroxy-irbesartan metabolites was monitored.
HPLC Analysis.
The HPLC system consisted of a STAR Varian system equipped with a 9010 gradient pump, a 9050 UV-variable spectrophotomonitor and a 9095 automatic injector. Chromatography was conducted on a Superspher C8 (250 × 4.6 mm; particle size = 5 μm; Merck, Darmstadt, Germany) (Varian SA, Les Ulis, France) eluted at 1.0 ml/min with the following mobile phase; 20% acetonitrile/80% of 0.1% diethylamine (adjusted at pH 5.5 with acetic acid) to 40% acetonitrile over a 12-min linear gradient. UV detection was performed at 250 nm.
Statistical Analysis.
The apparent Michaelis-Menten parameters i.e.,Km,Kmapp, andVmax, for the formation of irbesartan, nifedipine, and tolbutamide metabolites, were estimated by nonlinear regression analysis with the EnzPack 3 software from BioSoft. TheKi for the different inhibitors were estimated from a plot ofKm/Vmax andKmapp/Vmax as a function of increasing inhibitor concentrations. In all cases, the parameters were estimated with linear regression analysis of data obtained from at least two different experiments.
Results
Chromatographic Analysis.
Typical HPLC chromatograms obtained after a 30-min incubation of human liver microsomes with 50 μM irbesartan, in the absence or the presence of 1 mM NADPH, are illustrated in Fig.2. In the presence of NADPH, the essential cofactor for CYP-dependent reactions, four additional chromatographic peaks, labeled A, B,C, and D according to their elution order under these chromatographic conditions, were observed. They were identified as monohydroxy derivatives of irbesartan by liquid chromatography-mass spectrometry analyses. Among these four derivatives, metabolite B was unambiguously characterized as the 3-hydroxy-butyl derivative, whereas for metabolites Cand D, hydroxyl groups were located on the diazo-spiro ring.
Metabolism of Irbesartan.
Irbesartan biotransformation was evaluated on 12 different human liver microsomal preparations (final protein concentration = 2 mg/ml) after a 30-min incubation of 10 μM irbesartan with 1 mM NADPH. Turnover rates ranged from 0.050 to 0.116 nmol/min/mg, indicating that the rates of product formation slightly varied from one individual to the other (Table 1). Regardless of the human liver microsomal preparation investigated and the rate of irbesartan oxidation, the four monohydroxy metabolites of irbesartan were detected. The quantitative formation of irbesartan monohydroxy metabolites was similar in the different human liver preparations, even in the preparation termed HTL-49, which expressed a very low turnover rate for dextromethorphanO-demethylation, i.e., a CYP2D6 probe reaction.
To evaluate kinetic parameters for irbesartan oxidation, human liver microsomes (2 mg/ml; n = 4) were incubated for 30 min with 1 mM NADPH and increasing irbesartan concentrations ranging from 10 to 100 μM, and individual irbesartan monohydroxy metabolites, as well as their sum, were determined. Kinetic parameters were fitted to the Michaelis-Menten equation by nonlinear regression analysis of the turnover rates (Fig. 3).Km andVmax values obtained for each monohydroxy metabolite, as well as their sum, and for four human liver preparations are summarized in Table 2. Because four metabolites were formed after incubation of irbesartan with microsomal preparations, these data should be taken as estimates only. Apparent Km andVmax for irbesartan oxidation were 54.0 ± 6.5 μM and 0.62 ± 0.18 nmol/min/mg, respectively. As expected, the in vitro kinetic studies with four different human liver microsomal preparations yield similar values forKm but different values forVmax. The differentVmax values undoubtedly reflect the fact that the four different livers are likely to have different CYP isoform contents.
Correlation between CYP Isoform Activities and Irbesartan Biotransformation.
Irbesartan oxidation rates were determined after a 30 min-incubation of human hepatic microsomal fractions (2 mg/ml) with 10 μM irbesartan in the presence of 1 mM NADPH. Over this incubation period, both the rate of irbesartan oxidation and the rates of formation of the four hydroxy metabolites were linear. The rates of irbesartan oxidation ranged from 0.050 to 0.116 nmol/min/mg in microsomes obtained from 12 different human liver samples, a 2-fold difference (Fig.4). The human liver microsomal fraction HTL-5, known to be high in CYP2C9-associated activity, had the highest irbesartan oxidation activity (Table 1). Irbesartan oxidation rates were compared with the catalytic activity of several selective human liver CYP isoforms. A good correlation (r2 = 0.769) between irbesartan oxidation and tolbutamide 4-methylhydroxylation was observed in the 12 human liver microsomal preparations (Table3, Fig. 4). No significant correlation with catalytic activity of other CYP isoforms (CYP1A1, CYP1A2, CYP2A6, CYP2D6, CYP2E1 and CYP3A4) was observed. Tolbutamide 4-methylhydroxylation was further compared with the formation of each single irbesartan metabolite. Nice correlation was also observed for metabolites A (r2 = 0.715),B (r2 = 0.667), andD (r2 = 0.700), also suggesting the involvement of CYP2C9 in their formation. For metaboliteC, a slightly lower but significant correlation was also obtained (r2 = 0.504). Therefore, no correlation at all was obtained with nifedipine dehydrogenation (r2 = 0.002).
Inhibitory Studies of Irbesartan Metabolism by Specific CYP Isoforms, Substrates, and Inhibitors.
To identify the CYP isoform(s) that catalyze irbesartan oxidation, 50 μM irbesartan (approximately the Kmvalue; Table 2) was incubated with human liver microsomes (2 mg/ml final protein concentration) and 1 mM NADPH in the absence or the presence of CYP isoform-selective substrates, i.e., 7-ethoxyresorufin [CYP1A1 (Tassaneeyakul et al., 1993);Km = 0.2 μM], [CYP1A2 (Sesardic et al., 1988; Butler et al., 1989); Km = 30 μM], coumarin [CYP2A6 (Pelkonen et al., 1985);Km = 0.3 μM], tolbutamide [CYP2C9 (Miners et al., 1988); Km = 200 μM], S-mephenytoin [CYP2C19;Km = 100 μM], dextromethorphan [CYP2D6 (Krombach et al., 1987); Km = 20 μM], aniline [CYP2E1 (Guengerich et al., 1991);Km = 15 μM], and nifedipine [CYP3A4 (Guengerich et al. 1986); Km= 15 μM]. The inhibitory effect of these different specific CYP isoform substrates was evaluated at 5- to 33-fold their respectiveKm value (Table4). TheseKm values have been examined recently (Bourrié et al., 1996) and are similar to those reported in the literature (Halpert et al, 1994).
The metabolism of irbesartan (50 μM; approximately theKm value) was also evaluated in the absence or the presence of CYP isoform-selective inhibitors, i.e., α-naphthoflavone [CYP1A1 (Chang et al., 1994);Ki = 0.01 μM; mixed-type inhibition], furafyllin [CYP1A2 (Sesardic et al., 1990);Ki = 3 μM; competitive-type inhibition], pilocarpine [CYP2A6; Ki= 4 μM; competitive-type inhibition], sulfaphenazole [CYP2C9;Ki = 0.3 μM; competitive-type inhibition], quinidine [CYP2D6; Ki = 0.4 μM; competitive-type inhibition], diallyldisulfide [CYP2E1 (Chen et al. 1994); Ki = 100 μM] and ketoconazole [CYP3A4 (Wrighton et al., 1994);Ki = 0.015 μM; mixed-type inhibition]. The selectivity of these different specific CYP isoform inhibitors was evaluated at 5- to 20-fold their respectiveKi value (Table 4). TheseKi values have been recently reported by Bourrié et al. (1996).
When combining data obtained in the presence of either CYP isoforms, substrates, or inhibitors, slight or no inhibitory effects on total irbesartan oxidation were observed after incubation of microsomes and irbesartan with “phenacetin-furafyllin,” “7-ethoxyresorufin-α-naphthoflavone.” Mephenytoin, even at 5-fold the Km value for CYP2C19, did not decrease irbesartan oxidation. For “coumarin-pilocarpine,” “aniline-diallyldisulfide,” “dextromethorphan-quinidine,” and “nifedipine-ketoconazole,” only either the substrate or the inhibitor, inhibited irbesartan oxidation, suggesting a nonspecific inhibition.
Because the metabolism of irbesartan was inhibited by both sulfaphenazole and tolbutamide, CYP2C9 was apparently involved. To better investigate the involvement of CYP2C9, the kinetics of tolbutamide and warfarin (Rettie et al., 1992) inhibitions on irbesartan oxidation was evaluated. Human hepatic microsomes (HTL-18; 2 mg/ml) were incubated with increasing irbesartan concentrations ranging from 10 to 100 μM in the absence or the presence of either fixed concentrations of tolbutamide (200 μM and 1 mM) or warfarin (20, 50, and 200 μM). Figure 5 illustrates the Lineweaver-Burk representations of the “tolbutamide-irbesartan” and the “warfarin-irbesartan” interactions.
Both tolbutamide and warfarin are competitive-type inhibitors of irbesartan oxidation as demonstrated by the intercept of the straight lines on the ordinate axis. Individual irbesartan monohydroxy metabolites were monitored, and each of them was similarly affected by tolbutamide and warfarin (data not shown). ComputedKi were 522 μM and 27.8 μM for tolbutamide and warfarin, respectively. The difference inKi values for tolbutamide andS-warfarin is consistent with theirKm values determined on human hepatic microsomes, i.e., 227.5 ± 72.4 μM (n = 6) (Bourrié et al., 1996) and 5.4 ± 0.4 μM, respectively.
The inhibitory effect of irbesartan on tolbutamide 4-methylhydroxylation, i.e., a reaction specifically catalyzed by CYP2C9, was also evaluated. Human hepatic microsomes (HTL-18; 2 mg/ml) were incubated for 30 min with 1 mM NADPH and increasing tolbutamide concentrations ranging from 83.3 to 500 μM in the absence or the presence of fixed irbesartan concentrations (50, 100, and 250 μM), and 4-methyl-hydroxytolbutamide derivative was monitored. As illustrated in Fig. 6, irbesartan is a competitive-type inhibitor of tolbutamide 4-methylhydroxylation with aKi computed around 317 μM.
Of particular interest was the observation that nifedipine completely abolished the formation of all irbesartan monohydroxy metabolites, whereas ketoconazole had no effect. This potential interaction between nifedipine and irbesartan was further investigated. Human hepatic microsomes (HTL-18; 2 mg/ml) were incubated with increasing concentrations of either nifedipine or irbesartan, and the effects of irbesartan (50, 100, 250, and 500 μM) or nifedipine (50 and 100 μM), respectively, were investigated. Figure7 illustrates the Lineweaver-Burk representation of the “nifedipine-irbesartan” interaction. These data demonstrate that nifedipine is a potent noncompetitive-type inhibitor (Ki = 46.2 μM) of irbesartan oxidation, whereas irbesartan had no effect at all on nifedipine oxidation.
To eliminate the involvement of CYP3A4 in irbesartan oxidation, the effect of either additional CYP3A4 substrates, i.e., diltiazem (Renton, 1985; Pichard et al., 1990) and verapamil (Renton, 1985; Pichard et al., 1990), or inhibitor, i.e., ketoconazole (Maurice et al., 1992;Baldwin et al., 1995; Wrighton and Ring, 1994) and triacetyloleandomycin (after a 15 min-preincubation period allowing the specific suicide inhibition of CYP3A4; Chang et al. 1994; Fabre et al., 1988a), was investigated. For ketoconazole, only concentrations up to 1.5 μM were investigated because theKi of ketoconazole for nifedipine and midazolam oxidation processes were computed to be around 0.03 μM (Wrighton et al., 1994; Bourrié et al., 1996). Results are illustrated in Fig. 8. Neither diltiazem and verapamil as CYP3A4 substrates, nor ketoconazole and triacetyloleandomycin as CYP3A4 inhibitors, exhibited an inhibitory effect on irbesartan monohydroxy metabolites formation. Under the same conditions, sulfaphenazole was a potent inhibitor of irbesartan oxidation.
Irbesartan Metabolism by Metabolically Competent Cell Lines.
To substantiate the results from inhibitory studies, the metabolism of irbesartan was investigated with the cloned human CYP microsomes (Gentest Corporation). Profiles of irbesartan oxidation obtained with control cells and those expressing CYP2C9 and CYP3A4 are illustrated in Fig. 9. Irbesartan was oxidized to its four monohydroxy metabolites only in incubations with CYP2C9. This metabolic profile was very similar to that obtained after incubation of human hepatic microsomal fractions with irbesartan. Only metaboliteC appeared to be quantitatively transformed at a slightly slower rate. After incubation with CYP3A4, metabolites A andB were not detected and only slight amounts of metabolitesC and D were observed. Irbesartan oxidation was demonstrated neither in control incubations nor in incubations with CYP1A1, CYP1A2, CYP2A6, CYP2B6, CYP2D6, and CYP2E1. The rate of irbesartan oxidation in incubations with recombinant human liver CYP2C9 and CYP3A4 was 3.88 and 0.08 pmol of total metabolites formed/pmol CYP/min, respectively. Although incubation with CYP3A4 was performed in the absence of cytochrome b5, these data confirmed the absence of the involvement (or the very slight involvement) of CYP3A4 in irbesartan oxidation.
Discussion
There is increasing awareness that combinations of in vitro studies provide a reliable means to identify the human enzymes involved in the metabolism of any given drug. The early knowledge of the potential human metabolism, as well as specific isoform(s) involved in this specific biotransformation process, is of great importance at several stages of drug development (Birkett et al., 1993; Miners et al., 1994). Once the enzyme(s) or isoform(s) responsible for the metabolism of a particular drug is (are) known, predictions may be made concerning drug-drug interactions likely to influence the pharmacokinetics, therapeutic response, and development of adverse reactions in specific patient groups (Cholerton et al., 1992; Wrighton and Stevens, 1992). Hence, drug interaction studies can be targeted to evaluate those likely to be relevant and clinically important with major savings in development time and resources (Tucker, 1992). Therefore, they obviate the need for the often needless pharmacokinetic studies performed in human subjects. This early knowledge of isoform involvement is of particular importance if the test compound or one of the coadministered drugs is characterized as having a relatively narrow therapeutic index, i.e., theophylline, warfarin, phenytoin.
Human liver microsomes and the techniques of correlation with marker activities, selective chemical inhibition and heterologous cDNA expression have been used to study the metabolism of irbesartan. The results collectively demonstrate that CYP2C9 plays a major role in the formation of the four hydroxyl derivatives of irbesartan in human liver microsomes.
Kinetic parameters were determined in human liver microsomes for the conversions of irbesartan to its monohydroxyl metabolites. The formation of each metabolite followed saturation kinetics with respect to substrate concentration. In addition, the Lineweaver-Burk plot of each data set was linear, suggesting the involvement of a single, or closely related, CYP isoform in irbesartan oxidation. TheKm determined for irbesartan oxidation was around 50 μM.
As demonstrated by the different approaches used, CYP1A1 and CYP1A2 were not involved in irbesartan oxidation.
Although a 50% decrease in irbesartan oxidation was observed in the presence of dextromethorphan, no inhibition at all was observed in the presence of quinidine, a specific and potent CYP2D6 inhibitor. Furthermore, as reported in Table 1, the rate of irbesartan oxidation was not decreased in preparations lacking CYP2D6, i.e., HTL-49. Finally, no correlation between irbesartan and dextromethorphan oxidation was reported and CYP2D6-expressing cells did not metabolize irbesartan.
Both pilocarpine and coumarin inhibited irbesartan oxidation. The inhibitory effect of pilocarpine on CYP2C9-catalyzed enzyme activities has already been reported (Bourrié et al., 1996). At 30-μM concentration, pilocarpine decreased coumarin 7-hydroxylation by 80% and tolbutamide 4-methylhydroxylation by 60%. Additional studies performed on the “coumarin-irbesartan” interaction demonstrated that irbesartan was only a poor inhibitor of CYP2A6 with a meanKi value of 176 μM (n = 6). Moreover, neither a correlation between coumarin and irbesartan metabolism, nor an oxidation of irbesartan by CYP2A6-expressing cells was demonstrated.
Although diallyldisulfide decreased irbesartan oxidation by approximately 30%, no inhibitory effect was observed after incubation with aniline. Bourrié et al. (1996) also reported the lack of specificity of diallyldisulfide. CYP1A-, 2A6-, -2D6, and -3A4 catalyzed activities were also inhibited at different extents. Hence, it appeared unlikely that CYP2E1 was involved in irbesartan oxidation because dimethyl sulfoxide had no inhibitory effect on this metabolic pathway. Finally, no correlation between aniline and irbesartan oxidation or the formation of irbesartan metabolites by CYP2E1-expressing cells were observed.
These different approaches lead to the conclusion that CYP1A1, -1A2, -2A6, -2C19, -2D6, and -2E1 are not involved in irbesartan oxidation.
Irbesartan oxidation was highly correlated (r2 = 0.769) to tolbutamide 4-methylhydroxylation, suggesting the involvement of CYP2C9 in irbesartan oxidation. No correlation at all was achieved with other isoform-selective substrate activities. As shown in Table 4 and Fig. 5, irbesartan oxidation was mainly inhibited by tolbutamide, sulfaphenazole and nifedipine. Tolbutamide and warfarin, two CYP2C9 substrates, were competitive-type inhibitors of irbesartan oxidation with Ki computed around 500 μM and 30 μM, respectively. These Ki values are consistent with their Km values, 227.5 ± 72.4 μM (Bourrié et al., 1996) and 5.4 ± 0.4 μM determined on human hepatic microsomes. Moreover, irbesartan was a competitive-type inhibitor of tolbutamide 4-methylhydroxylation, a reaction specifically catalyzed by CYP2C9. This suggests that CYP2C9 was mainly involved in irbesartan oxidation.
Of particular interest was the observation that irbesartan oxidation was potentially inhibited by nifedipine, a specific CYP3A4 substrate, whereas ketoconazole, a specific CYP3A4 inhibitor, did not affect irbesartan oxidation, and no correlation was demonstrated between irbesartan and nifedipine oxidation processes on a bank of human hepatic microsomes. Additional studies demonstrated that nifedipine was a potent noncompetitive-type inhibitor of irbesartan oxidation with aKi computed around 50 μM, whereas high irbesartan concentrations, i.e., 500 μM, had no inhibitory effects on nifedipine oxidation. These data suggest that nifedipine exhibited an inhibitory effect on CYP2C9 without being metabolized by this isoform. This hypothesis was demonstrated by different ways. First, irbesartan oxidation was investigated in the presence of either specific CYP3A4 substrates, i.e., verapamil (Km = 20 μM; Pichard et al., 1990) and diltiazem (Km = 25 μM; Pichard et al., 1990), or CYP3A4 inhibitors, i.e., triacetyloleandomycin (Ki = 10 μM; Pichard et al., 1990). Even at 5- to 20-fold their respectiveKm orKi values, these different probes have no inhibitory effect on irbesartan oxidation. Previous studies already reported that nifedipine was a relatively potent inhibitor of methyl-hydroxytolbutamide formation, with aKi value of 11 μM (Miners et al., 1988). Although a competitive-type inhibition was suggested, these authors suggested that reciprocal inhibition between tolbutamide and nifedipine did not necessarily indicate that these two compounds were metabolized by the same CYP isoform. The interaction between tolbutamide and nifedipine was further investigated on human hepatic microsomes. Nifedipine was a potent noncompetitive-type inhibitor of tolbutamide 4-methylhydroxylation with aKi computed around 20 μM (G.F. and M.B., personal communication), whereas tolbutamide had no effect at all on nifedipine dehydrogenation. These data demonstrated that the inhibitory effect of nifedipine on irbesartan metabolism was associated to a nonspecific inhibition of CYP2C9 and that CYP3A4 was not expected to play a major role in irbesartan oxidation.
Irbesartan was then incubated with microsomes containing different recombinant human liver CYP isoforms. The four monohydroxyl derivatives of irbesartan were observed only in the incubations of irbesartan with CYP2C9, confirming its involvement in these metabolic processes. Metabolites C and D were also detected after incubations with CYP3A4-engineered microsomes. Turnover rates for irbesartan oxidation were calculated at 0.08 pmol of metabolites formed/pmol CYP3A4/min and 3.88 pmol of metabolites formed/pmol CYP2C9/min, suggesting a 50-fold difference in the ability of these two isoforms to synthesize irbesartan metabolites. Recently, Shimada et al. (1994) reported that 70% of human liver CYPs were associated with CYP1A2 (13%), -2A6 (4%), -2C (18%), -2D6 (1.5%), -2E1 (7%), and -3A (30%) proteins. CYP3A and CYP2C proteins represent almost 40 to 50% of total hepatic CYP content. Based on the respective proportions of these two isoforms, as well as their specific turnover for irbesartan oxidation determined on CYP-engineered microsomes, CYP2C9 appears to be mainly involved in irbesartan oxidation.
In summary, a combination of in vitro procedures were used to demonstrate that CYP2C9 isoform was mainly responsible for human liver microsomal hydroxylation, regardless of the site of oxidation. Identification of the CYP isoform that catalyzes irbesartan oxidative metabolism allows prediction of those factors likely to affect irbesartan pharmacokinetic properties. In particular, drug interactions of potential therapeutic importance may be targeted for further studies.
Acknowledgments
We thank Dr. Patrick Maurel (Institut National de la Santé et de la Recherche Médicale U-128, Montpellier, France) and Dr. François Guillou (Sanofi Recherche, Montpellier, France) for stimulating and helpful discussions and their critical reading of the manuscript. Grateful thanks are also due to Josiane Villard for her invaluable help in the preparation of the manuscript.
Footnotes
-
Send reprint requests to: Dr. Gérard Fabre, Sanofi Recherche, Preclinical Metabolism and Pharmacokinetics Department 371 Rue du Professeur Blayac, 34184 Montpellier, Cédex 4, France. E-mail: Gerard.Fabre{at}TLS1.ELFSANOFI.FR
- Abbreviations used are::
- CYP
- cytochrome P450
- Received May 27, 1998.
- Accepted September 22, 1998.
- The American Society for Pharmacology and Experimental Therapeutics