Human Cytochrome P450 1A2 Involvement in the Formation of

Jan 12, 2009 - Naruto, Tokushima 772-8601, Japan, Graduate School of Pharmaceutical Sciences, Hokkaido UniVersity,. Sapporo 060-0812, Japan, and ...
3 downloads 0 Views 888KB Size
Chem. Res. Toxicol. 2009, 22, 323–331

323

Human Cytochrome P450 1A2 Involvement in the Formation of Reactive Metabolites from a Species-Specific Hepatotoxic Pyrazolopyrimidine Derivative, 5-n-Butyl-7(3,4,5-trimethoxybenzoylamino)pyrazolo[1,5-a]pyrimidine Shunji Kuribayashi,*,† Kiyoto Goto,† Shinsaku Naito,† Tetsuya Kamataki,‡ and Hiroshi Yamazaki*,§ Nutrient/Drug Metabolism and Pharmacology, Otsuka Pharmaceutical Factory, Inc., 115 Tateiwa, Muya-cho, Naruto, Tokushima 772-8601, Japan, Graduate School of Pharmaceutical Sciences, Hokkaido UniVersity, Sapporo 060-0812, Japan, and Showa Pharmaceutical UniVersity, Machida, Tokyo 194-8543, Japan ReceiVed September 30, 2008

5-n-Butyl-7-(3,4,5-trimethoxybenzoylamino)pyrazolo[1,5-a]pyrimidine) (OT-7100) is a pyrazolopyrimidine derivative with potential analgesic effects. Exclusively limited elevations in serum levels of aspirateand alanine-aminotransferase were abnormally observed in a clinical study, in contrast to no toxicological potential to experimental animals. The aim of this study was to clarify the mechanism responsible for species-specific hepatotoxicity of this model compound. OT-7100 was primarily metabolized to a carboxylic acid derivative and an amino derivative (5-n-butyl-pyrazolo[1,5-a]pyrimidine, M-5) by hydrolysis in humans and rats. In human liver, pyrazolo[1,5-a]pyrimidine derivative M-5 was further metabolized to mainly M-23OH (a C-3-position hydroxyl derivative, 3-hydroxy-5-n-butyl-pyrazolo[1,5-a]pyrimidine). Studies with recombinant cytochrome P450s (P450s), correlation analysis using a panel of human liver microsomes as well as immunoinhibition with anti-P450 antibodies collectively suggested that human liver microsomal P450 1A2 preferentially metabolized M-5 to predominantly M-23OH. Human liver microsomes were capable of activating M-5 to a covalently bound metabolite faster than rat liver microsomes: reduced glutathione prevented the bindings. A cysteine adduct derivative of M-23OH at the C-6-position was structurally confirmed. On the contrary, rat liver microsomal P450 1A2 could metabolize M-5 to equally M-23OH, M-22OH (a C-6-position hydroxyl derivative, 6-hydroxy-5-n-butyl-pyrazolo[1,5a]pyrimidine), or an unknown metabolite. These results suggest that differences in the regiospecific metabolic function of human and rat P450 1A2 would be responsible for the human-specific metabolic activation of the primary metabolite of OT-7100 to a proximate form. It is presumed that hepatotoxicity associated with OT-7100 could be likely related to the formation of a human-specific reactive metabolite from M-23OH. OT-7100 activation by inducible P450 1A2 may therefore exhibit marked individual differences. Introduction Cytochrome P450 (P450) comprises a superfamily of enzymes involved in the oxidation of a large number of exogenous and endogenous compounds (1). In human liver, P450 3A4 is the major P450 enzyme, followed by P450 2C9 and P450 1A2 (2). Large interindividual variations in the contents and activities of several P450 forms lead to different roles of P450s in the oxidation of substrates associated with their pharmacological or toxicological actions (3, 4). It is believed that numerous factors contribute to the development of unpredictable adverse drug reactions such as an idiosyncratic drug reaction, including the metabolic balance between bioactivation and detoxification * To whom correspondence should be addressed. Shunji Kuribayashi, Nutrient/Drug Metabolism and Pharmacology, Preclinical Assessment Department, Research and Development Center, Otsuka Pharmaceutical Factory, Inc., 115 Tateiwa, Muya-cho, Naruto, Tokushima 772-8601, Japan. Phone: +8188-685-1151. Fax: +81-88-686-8185. E-mail: [email protected]. Professor Hiroshi Yamazaki, Showa Pharmaceutical University, 3-3165 Higashitamagawa Gakuen, Machida, Tokyo 194-8543, Japan. Phone/fax: +81-42721-1406. E-mail: [email protected]. † Otsuka Pharmaceutical Factory, Inc. ‡ Hokkaido University. § Showa Pharmaceutical University.

and formation of a reactive species capable of covalently binding to proteins (5, 6). It has been reported that the formation of reactive metabolites of the drug is as an initial step in adverse effects and that the step may be followed by covalent binding of the metabolites to the protein generating the reactive metabolites and/or other proteins, which then behave as neoantigens and trigger an abnormal immunological response, leading to adverse effects (7). In general, a quinone-type metabolite from drugs such as acetaminophen (8), diclofenac (9), or troglitazone (10) is considered to be one of the reactive intermediates, which are thought to be a rate-limiting step toward their toxicity. 5-n-Butyl-7-(3,4,5-trimethoxybenzoylamino)pyrazolo[1,5a]pyrimidine (OT-71001) is an amide moiety-bearing pyrazolopyrimidine derivative with a potential analgesic effect. We previously reported (11) that the bioavailabilities of OT-7100 after oral administration in rats, dogs, and monkeys were 36, 17, and 0.3%, respectively, despite the similar plasma concen1 Abbreviations: OT-7100, 5-n-butyl-7-(3,4,5-trimethoxybenzoylamino)pyrazolo[1,5-a]pyrimidine; M-5, 5-n-butyl-pyrazolo[1,5-a]pyrimidine; M-22OH, 6-hydroxy-5-n-butyl-pyrazolo[1,5-a]pyrimidine; M-23OH, 3-hydroxy-5-n-butyl-pyrazolo[1,5-a]pyrimidine.

10.1021/tx8003592 CCC: $40.75  2009 American Chemical Society Published on Web 01/12/2009

324

Chem. Res. Toxicol., Vol. 22, No. 2, 2009

Kuribayashi et al.

metabolite, M-23OH, derived from the primary metabolite, M-5, by oxidative metabolism. However, the mechanism responsible for the species-specific hepatotoxicity of this model compound is not known. The purpose of this study was to address the potential metabolic activation of OT-7100 through M-5 formation as well as to examine a possible mechanism associated with the development of hepatotoxicity. The roles of human P450 enzymes involved in the secondary metabolism of M-5 were investigated with recombinant human P450s and liver microsomes in the present study primarily at a substrate concentration of 5 µM, on the basis of the known human or monkey plasma concentration (11). The covalent bindings of reactive metabolites and structural analysis were also examined. In particular, we are interested in investigating the function of human P450 1A2 as one of the determinant factors for this metabolic activation in humans.

Experimental Procedures

Figure 1. Proposed model for quinone imine-dependent toxicity in hepatic cells after the administration of 5-n-butyl-7-(3,4,5-trimethoxybenzoylamino)pyrazolo[1,5-a] pyrimidine (OT-7100).

tration-time profiles of intravenously administrated OT-7100 in these animals. The primary factor influencing first-pass metabolism for OT-7100 was enzymatic hydrolysis in the small intestine, especially in monkeys (11). Subsequently, limited elevations in the serum levels of asparate aminotransferase and alanine aminotransferase were abnormally observed in a clinical study of OT-7100, in contrast to no toxicological potential of OT-7100 to experimental animals. No apparent differences in plasma OT-7100 levels of volunteers were found between the toxic cases and other groups in our limited observations (not shown). The reported metabolic pathway of OT-7100 is shown in Figure 1 (11). The OT-7100 metabolites included hydrolysates of the amide moiety of OT-7100 (M-19 and M-5) and oxidative products of the n-butyl group of OT-7100 (M-2 and M-3). The primary metabolite M-5 is assumed to be further metabolized to M-22OH (6-hydroxy-5-n-butyl-pyrazolo[1,5-a]pyrimidine) and M-23OH (3-hydroxy-5-n-butyl-pyrazolo[1,5-a]pyrimidine) by P450s, and additionally, the sulfate and glucuronate conjugates of M-22OH and M-23OH are excreted. A quinone imine metabolite formation from OT-7100 as one of active intermediate candidates resulting in induced hepatotoxicity may be postulated from the chemical structure of the secondary

Chemicals. OT-7100 (5-n-butyl-7-(3,4,5-trimethoxybenzoylamino)pyrazolo[1,5-a] pyrimidine; molecular weight of 384.44; Figure 1), its metabolites (M-2, M-3, M-5, M-19, M-22OH, and M-23OH), 14C-M-5 (the labeling position of this compound was the 5 position carbon atom in the pyrazolo-pyrimidine ring and is shown in Figure 1), and OT-7126 (5-n-pentyl-7-(3,4,5-trimethoxybenzoylamino)pyrazolo[1,5-a] pyrimidine; molecular weight of 398.47) as an internal standard for HPLC analysis were synthesized at Otsuka Pharmaceutical Factory (Tokushima, Japan). All of these chemicals were determined, by reversed-phase HPLC, to be >99.0% pure. Reduced glutathione, L-cysteine, and N-acetylcysteine were purchased from Sigma-Aldrich (St. Louis, MO). All other chemicals and reagents used were of analytical reagent grade. Enzymes. Pooled human liver microsomes (H0610) and individual human liver microsomes (Human Reaction Phenotyping Kit Ver. 5) were obtained from Xenotech, LLC (Lenexa, KS). Rat liver microsomes, pooled human liver microsomes (H161), and recombinant rat and human P450 enzymes coexpressed with NADPHP450 reductase in baculovirus-infected insect cells (Supersomes) were purchased from BD Gentest (Woburn, MA). Human P450 2A6, 2B6, 2C8, 2C19, 2E1, and 3A4, and rat P450 2A2, 2B1, and 3A2 were also coexpressed with cytochrome b5. Enzyme Assays. A typical incubation mixture (0.50 mL total volume) contained 0.50 mg/mL human liver microsomes (or 20 pmol equivalent recombinant P450/mL), 50 mM potassium phosphate buffer (pH 7.4), an NADPH-generating system (0.5 mM NADP+, 5 mM glucose 6-phosphate, 1 U/mL glucose-6-phosphate dehydrogenase), and OT-7100 or primary metabolite M-5 (5.0 and 50 µM). The primary substrate concentration of 5 µM was chosen because this was the peak blood concentration treated with the drug (11). In some experiments, monoclonal anti-P450 1A2 or P450 3A4 antibodies (BD Gentest, up to 50 µL/mg microsomal protein) were preincubated with pooled human liver microsomes for M-5 metabolism. After a 5 min preincubation, the reactions were initiated by the addition of the NADPH-generating system and were incubated at 37 °C for 30 min. The reactions were terminated by 0.50 mL of methanol, and then, OT-7126 was added as an internal standard. After centrifugation at 2,000g for 10 min at 4 °C, the supernatant was added to 5 mM ammonium acetate at a ratio of 1:1, and a 0.20 mL aliquot was injected onto the HPLC system. HPLC Analyses. OT-7100 and its metabolites were assayed using an HPLC method validated over a concentration range of 0.10-5.0 µM. OT-7100 and its metabolites were separated on a 250 mm × 4.6 mm i.d. Inertsil ODS-3V column (GL-Sciences, Tokyo, Japan) and detected at wavelengths of 215 nm (for M-19) and 230 nm (for OT-7100 and the other metabolites) using an HPLC system (LC-10A Series, Shimadzu, Kyoto, Japan). The column temperature was maintained at 40 °C. The mobile phase was 5 mM ammonium acetate (A) and acetonitrile (B). The conditions for

Human P450 1A2 in ReactiVe Metabolite Formation

Chem. Res. Toxicol., Vol. 22, No. 2, 2009 325

Figure 2. Metabolites of OT-7100 and M-5 formed by liver microsomes from humans and rats. OT-7100 (5 µM (A) or 50 µM (B)) and M-5 (5 µM (C) or 50 µM (D)) were incubated with liver microsomes (0.5 mg protein/mL) from humans and rats in the presence of an NADPH-generating system. Data are the average of duplicate incubations.

elution were as follows: 6 to 10% B (0-8 min), 10 to 20% B (8-10 min), 20 to 28% B (10-25 min), 28 to 80% B (25-36 min), and 80 to 80% B (36-43 min). Linear gradients were used for all solvent changes. The flow rate was 0.80 mL/min. M-5 and its metabolites were assayed using an HPLC method validated over a concentration range of 0.10-10.0 µM. M-5 and its metabolites were separated on a 250 mm × 4.6 mm i.d. Inertsil ODS-3V column and detected at wavelengths of 230 nm using an HPLC system (LC-10A Series). The column temperature was maintained at 40 °C. The mobile phase was 5 mM ammonium acetate (A) and acetonitrile (B). The conditions for elution were as follows: 13 to 25% B (0-8 min), 25 to 25% B (8-20 min), 25 to 35% B (20-25 min), 35 to 80% B (25-31 min), and 80 to 80% B (31-40 min). Linear gradients were used for all solvent changes. The flow rate was 0.80 mL/min. Covalent Binding of M-5 Derivatives to Liver Microsomal Proteins. Covalent binding was assessed according to a method described elsewhere (12). 14C-M-5 (5.0 µM) was incubated with human and rat liver microsomes (0.50 mg protein/mL) in the presence of the NADPH-generating system and reduced glutathione, L-cysteine, or N-acetylcysteine (0-50 mM) at 37 °C for 60 min. The reactions were started with the addition of the NADPHgenerating system and stopped with the addition of 3-fold cooled acetonitrile. After centrifugation at 2,000g for 10 min at 4 °C, the protein pellet was washed with 3.0 mL of methanol/diethyl ether (3:1) until no radioactivity was detected in the supernatant. The pellet was finally dissolved in 2.0 mL of 1 M NaOH at 60 °C for 90 min. An aliquot was neutralized with 0.50 mL of acetic acid before scintillation counting. Data are the mean ( SD of sextuplicate determinations and average of duplicate incubations, respectively, in the absence and presence of reduced glutathione, L-cysteine, or N-acetylcysteine. All statistical analyses were carried out with the InStat program (GraphPad Software, San Diego, CA). LC-MS Analysis and NMR Characterization. Pooled human liver microsomes were incubated with M-23OH (50 µM) and reduced glutathione or L-cysteine at 37 °C for 60 min in the presence of an NADPH-generating system. Incubations that lacked either the NADPH-generating system or glutathione and L-cysteine served

as negative controls. Reactions were terminated by the addition of ice-cold acetonitrile (1.0 mL). After centrifugation, the supernatant from each incubation was removed and evaporated to dryness. The residue was reconstituted in 90% acetonitrile in water (500 µL), vortex-mixed, and centrifuged. Aliquots (20 µL) of the final supernatant were analyzed by LC/ESI-MS and 1H NMR. Glutathione and cysteine conjugates of M-23OH formation were estimated using a JMS-700 MStation (JEOL, Tokyo, Japan) with a PU-980 HPLC system (JASCO, Tokyo, Japan). The mobile phase was 5 mM ammonium acetate (A) and acetonitrile (B). The conditions for elution were as follows: 2.0 to 50% B (0-30 min), 50 to 80% B (30-31 min), 80 to 80% B (31-35 min), and 80 to 2.0% B (35-36 min). Linear gradients were used for all solvent changes. The flow rate was 0.80 mL/min. The sample (20 µL) was injected onto the HPLC system. Ionization was conducted in positive ion mode at an orifice temperature of 80 °C and a desolvation plate temperature of 220 °C. The needle voltage was set to 2.0 kV, and the ring voltage was optimized at 45 V. 1 H NMR spectrum of M-23OH was recorded on NMR AC-300P spectrometers (Bruker BioSpin, MA, USA). This spectrum was referenced to the residual solvent resonances of DMSO-d6. The 1H NMR spectrum of M-23OH-cysteine conjugate was recorded on NMR JNM-AL400 spectrometers (JEOL, Tokyo, Japan). This spectrum was referenced to the residual solvent resonances of CD3OD-D2O. High-resolution LC/ESI-MS was recorded on a JMS700 mass spectrometer.

Results Primary Metabolism of OT-7100 in Liver Microsomes from Humans and Rats. OT-7100 or M-5 was incubated with liver microsomes from pooled humans and rats (Figure 2). M-5 and M-19 as well as the hydroxylation products on the n-butyl group (M-2 and M-3, Figure 1) were formed after the incubation of OT-7100 (5 and 50 µM) with liver microsomes from humans. M-5 was predominantly formed in rat liver microsomes. M19, M-2, and a small amount of M-3 were formed from OT-

326

Chem. Res. Toxicol., Vol. 22, No. 2, 2009

Kuribayashi et al.

Figure 3. Correlation between rates of M-23OH (O) and M-22OH ()) formation from M-5 (5 µM, A and B; 50 µM, C and D) and drug oxidation activities in 16 human liver microsomal samples.

Table 1. Correlation of the Rates of M-23OH and M-22OH Formation from M-5 with Various P450-Dependent Enzyme Activities in 16 Human Liver Microsomal Samplesa correlation coefficients (r) M-5:5 µM marker substrate reaction cytochrome P450 7-ethoxyresorufin O-dealkylation coumarin 7-hydroxylation S-mephenytoin N-demethylation paclitaxel 6R-hydroxylation diclofenac 4′-hydroxylation S-mephenytoin 4′-hydroxylation dextromethorphan O-demethylation chlorzoxazone 6-hydroxylation testosterone 6β-hydroxylation lauric acid 12-hydroxylation a

P450

M-23OH

1A2 2A6 2B6 2C8 2C9 2C19 2D6 2E1 3A4/5 4A9/11

0.21 0.91 0.51 0.02 0.10 0.37 0.09 0.02 0.11 0.03 0.59

M-5:50 µM M-22OH -

b

M-23OH 0.38 0.91 0.43 0.04 0.16 0.29 0.04 0.01 0.14 0.13 0.60

M-22OH 0.73 0.08 0.07 0.47 0.50 0.24 0.33 0.11 0.12 0.97 0.09

Marker activities were supplied by Xenotech, LLC. b -, M-22OH was not detected.

7100. M-5 formation activities in liver microsomes from rats were several fold higher than those from humans (Figure 2A,B). When M-5 was used as the substrate, M-23OH was the exclusive metabolite observed from M-5 in human liver microsomes at a low substrate concentration (5 µM, Figure 2C). At a high substrate concentration of 50 µM, a small amount of M-22OH was detected from M-5 with different pooled human liver microsomes (Figure 2D). In contrast, rat liver microsomes yielded M-23OH, M-22OH, and an unknown from M-5. Secondary Metabolism of M-5 by Liver Microsomes from Different Human Samples and by Recombinant P450 Enzymes. M-5 was incubated with different human liver microsomes to investigate the secondary metabolism (Figure 3). M-23OH formation rates from M-5 in liver microsomes were in the ranges of 0.004-0.055 and 0.057-0.35 nmol/min/mg protein at substrate concentrations of 5 and 50 µM, respectively.

Up to a 10-fold or more range in interindividual M-23OH formation from M-5 was observed. There were good correlations between P450 1A2-mediated 7-ethoxyresorufin O-dealkylation activities and formation rates of M-23OH from 5 and 50 µM of M-5 in a panel of human liver microsomes (r ) 0.91, Figure 3A,C). The minor M-22OH formation rates from M-5 at a high substrate concentration were correlated with P450 3A4-mediated testosterone 6β-hydroxylation activities (r ) 0.97, Figure 3D). As shown in Table 1, other marker substrate reactions for typical P450 isoforms did not show any significant correlations with M-23OH or M-22OH formation rates from M-5 in human liver microsomes. These results were also supported by immunoinhibition analyses that the major M-23OH formation from M-5 in human liver microsomes was suppressed by anti-P450 1A2 antibodies and that minor M-22OH formation was diminished by anti-P450 3A4 antibodies (Figure 4).

Human P450 1A2 in ReactiVe Metabolite Formation

Chem. Res. Toxicol., Vol. 22, No. 2, 2009 327

Figure 4. Inhibition by antihuman P450 1A2 (A) or P450 3A4 (B) on hydroxylation of M-5 by human liver microsomes. M-5 (50 µM) was incubated with pooled human liver microsomes in the absence or presence of antihuman P450 1A2 (A) or antihuman P450 3A4 (B). Control activities for M-23OH (b) and M-22OH ([) formations by liver microsomes from humans were 0.14 and 0.03 nmol/min/mg protein, respectively, and were defined as 100%. Data are the means of duplicate determinations.

Figure 5. Metabolism of M-5 formed by recombinant P450s. M-5 (5 µM (A) or 50 µM (B)) was incubated with the recombinant P450 system (0.02 µM P450) in the presence of an NADPH-generating system. Data are the average of duplicate incubations.

Human P450 1A2 preferentially metabolized M-5 to predominantly M-23OH at substrate concentrations of 5 and 50 µM. On the contrary, rat P450 1A2 metabolized M-5 to M-23OH (a C-3-position hydroxyl derivative), M-22OH (a C-6position hydroxyl derivative), or an unknown metabolite (Figure 5). Human P450 3A4 and rat P450 3A2 had limited activities compared with that of P450 1A2. Covalent Binding of M-5 after Metabolic Activation to Human Liver Microsomes. 14C-M-5 (5 µM) was incubated with liver microsomes from humans and rats in the presence of the NADPH-generating system (Figure 6). Human liver microsomes were capable of activating M-5 to a covalently bound metabolite (Figure 6). This covalent binding was also seen in the rat liver microsomes, but the amounts in the absence of reduced glutathione, L-cysteine, or N-acetylcysteine were sig-

nificantly lower than those in humans (p ) 0.005). Reduced glutathione, L-cysteine, or N-acetylcysteine dose-dependently inhibited the covalent binding of the metabolite(s) to liver proteins (Figure 6). Identification of M-23OH-Glutathione and M-23OHCysteine Conjugates. LC-MS analysis of extracts of human liver microsomal incubations containing M-23OH, the NADPHgeneration system, and reduced glutathione or L-cysteine revealed the presence of glutathione or cysteine conjugates, none of which were evident in control incubations that lacked one or both cofactors. The M-23OH-glutathione conjugate exhibited an [M + H]+ ion at m/z 512 and yielded the product ion spectrum shown in Figure 7A. High-resolution FAB/MS analysis of the M-23OH-glutathione conjugate yielded an [M + H]+ ion at m/z 512.1927, consistent with the M-23OH-glutathione conjugate molecular formula C20H30O7N7S (512.1920). The M-23OH-cysteine conjugate exhibited an [M + H]+ ion at m/z 324 and an [M - H]- ion at m/z 322 (Figure 7B,C). Highresolution FAB/MS analysis of the M-23OH-cysteine conjugate yielded an [M + H]+ ion at m/z 324.1125, consistent with the M-23OH-cysteine conjugate molecular formula C13H18O3N5S (324.1130). 1H NMR spectra of the M-23OH and M-23OHcysteine conjugate are shown in Figure 8A,B. The 1H NMR spectrum of M-23OH showed four proton singlets at δ 0.91, 1.33, 1.62, and 2.57. M-23OH showed resonances corresponding to -CH)C(NH2) (δ 5.81), -CH)N (δ 7.65), -NH2 (δ 7.31), and -OH (δ 8.41). The 1H NMR spectrum of the M-23OH-cysteine conjugate showed four proton singlets at δ 0.96, 1.39, 1.70, and 2.75. This substitution pattern of the M-23OH-cysteine conjugate was identical to that of M-23OH, which suggested no replacement of the n-butyl moiety. The M-23OH-cysteine conjugate showed resonances corresponding to -S-CH2 (δ 2.85, 3.63), -S-CH2-CH (δ 4.58), and -CH)N (δ 6.85).

Discussion It has been reported that drug-induced hepatotoxicity may be caused by active intermediates formed from a common toxicant acetaminophen (8) and idiosyncratic troglitazone (10) by P450s. Acetaminophen hepatotoxicity has been reported as being due to its biotransformation to a reactive and toxic intermediate, N-acetyl-p-benzoquinone imine, by P450 2E1, 1A2, 3A4, and 2A6 in humans (13-16). At therapeutic doses of acetaminophen, active N-acetyl-p-benzoquinone imine would be efficiently detoxified by conjugation with glutathione (17). However, following overdoses of the drug, depleted glutathione

328

Chem. Res. Toxicol., Vol. 22, No. 2, 2009

Kuribayashi et al.

Figure 6. Covalent protein binding of 14C-M-5 to liver microsomes from humans (A) and rats (B). 14C-M-5 (5 µM) was incubated with liver microsomes (0.5 mg protein/mL) from humans and rats in the presence of an NADPH-generating system and reduced glutathione (GSH), L-cysteine (L-Cys), and N-acetylcysteine (NAC). Data (and bars) are the mean ( SD of sextuplicate determinations and average of duplicate incubations, respectively, in the absence and presence of reduced glutathione, L-cysteine, or N-acetylcysteine.

Figure 7. LC-MS spectrum of the M-23OH-glutathione conjugate at m/z 512 ([M + H]+) (A) and the L-cysteine conjugate at m/z 324 ([M + H]+) (B), and m/z 322 ([M - H]-) (C) obtained in human liver microsomal incubations. The origins of the characteristics ions are as indicated.

has been reported to cause the metabolite to covalently bind to proteins (18) resulting in severe centrilobular hepatic necrosis in both humans and experimental animals (19, 20). In contrast to the cases of common active metabolites, human-specific idiosyncratic hepatotoxicity can be explored to occur at therapeutic doses from 1 in every 1,000 patients to 1 in every 100,000 patients, with a pattern that is consistent for each drug and for each class. Therefore, it would be almost impossible to detect idiosyncratic drug reaction in preclinical testing and during

clinical trials (7). One of the recent examples of drugs causing idiosyncratic hepatotoxicity and liver failure was an antidiabetic drug troglitazone (21). Nonclinical safety studies had not revealed any hepatotoxicity in experimental animals including monkeys, although the latter exhibit an oxidative metabolite profile similar to that of humans (22, 23). We similarly assumed that another hepatotoxic model drug OT-7100 would be metabolized to a reactive metabolite capable of covalently binding to cellular macromolecules (Figure 1).

Human P450 1A2 in ReactiVe Metabolite Formation

Chem. Res. Toxicol., Vol. 22, No. 2, 2009 329

Figure 8. 1H NMR spectra of the authentic M-23OH standard (A) and M-23OH-L-Cysteine conjugate (B) obtained in human liver microsomal incubations.

When OT-7100 was used as the substrate, primarily M-5 formation activities in liver microsomes from rats were several fold higher than those from humans (Figure 2). The hydrolase activities toward OT-7100 in liver microsomes from rats were higher than those of humans (11). Therefore, the differences in M-5 formations in liver microsomes between rats and humans were caused by these species differences in hydrolase activities. It should be mentioned that M-5 formations from OT-7100 were theoretically equivalent to another hydrolysate of the amide moiety (M-19). However, M-5 formations in liver microsomes from rats were 2 times higher than M-19 formations. In contrast, the main metabolite in rat plasma was M-19:M-5 was at lower levels than the other metabolites (unpublished observations). Taken together, hydrolysates of the amide moiety of OT-7100 (M-19 and M-5) and oxidative products of the n-butyl group (M-2 and M-3) were formed from OT-7100 (Figure 1) in the liver microsomes from rats and humans. There were no qualitative differences in primary metabolism toward OT-7100 in rats and humans. When the primary metabolite M-5 was used as the substrate, only M-23OH was detected in human liver microsomes, but M-23OH, M-22OH, and an unknown metabolite(s) were seen in rat liver microsomes at a low substrate concentration of 5 µM (Figure 2C). When M-23OH formation from M-5 by a panel of human liver microsomes screened for selective P450 marker activities was compared with marker P450 isoform activities in those same lines, the highest correlation was with the P450 1A2 marker activity 7-ethoxyresorufin O-dealkylation (r ) 0.91, Figure 3, Table 1). Antihuman P450 1A2 inhibited only M-23OH formation from M-5 (Figure 4). Recombinant human P450 1A2 preferentially metabolized M-5 to predominantly M-23OH, but recombinant rat P450 1A2 metabolized M-5 to M-23OH, M-22OH, and an unknown metabolite as well as those in liver microsomes (Figure 5). These results suggested that the

formation to M-23OH from M-5 in humans is catalyzed predominantly by P450 1A2. These lines of evidence also suggest that important differences exist between human and rat liver enzymes (P450 1A2) both in catalytic activity and regioselectivity of M-5 hydroxylation. Human P450 1A2 catalyzes the C-3 hydroxylation of M-5, the only hydroxylation product formed in human liver microsomes. However, rat P450 1A2 catalyzes both C-6 and C-3 hydroxylation of M-5, and unknown metabolite formation (Figure 5). In supporting the rapid clearance of M-5 in rat liver microsomes, in vivo M-5 clearance was confirmed by measuring OT-7100 metabolites in plasma after oral administration of OT-7100 to rats. The plasma concentration of another hydrolysate counterpart M-19 was higher than OT-7100, but M-5 concentration was too low to determine (results not shown). In supporting the selected hydroxylation of M-5 in humans in vivo, sulfate- and glucuronide-conjugates of the 3-position (but not 6-position) of M-5 (via M-23OH) were detected in urine after OT-7100 administration. In contrast, both 3- and 6-positions of sulfate- and glucuronide-conjugates (via both M-22OH and M-23OH) were observed in rat plasma after M-5 administration (results not shown). There are a number of reports on species differences for the metabolism by P450 1A2 in livers between human and experimental animals in terms of regioselectivity and enzymatic capacity. A typical example for different regioselectivity of P450 1A2 is caffeine (24, 25). Main human metabolites from caffeine have been reported to be 3-demethylated paraxanthine, followed by theobromine and theophylline in human liver microsomes (24). In contrast, rat liver microsomes have been shown to produce mainly 8-hydroxylated trimethyluric acid, followed by theophylline, paraxanthine, and theobromine (25). The typical catalytic efficiencies of human P450 1A2-mediated drug oxidations have been reported in 2-amino-3,8-dimethylimidazo[4,5-

330

Chem. Res. Toxicol., Vol. 22, No. 2, 2009

Kuribayashi et al.

Figure 9. Proposed mechanism for M-23OH-conjugate formation at the 6-position.

Figure 10. Proposed species differences in the metabolic pathways of M-5 between humans and rats.

f]quinoxaline or 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyrine (26, 27). Recombinant human P450 1A2 has been shown to have 10- to 19-fold higher N-oxidation activities than those of rat P450 1A2 (26). N-Oxidation of these procarcinogens were detected in human hepatocytes, in contrast to undetectable levels in rat hepatocytes pretreated with inducible agent 3-methylcholanthrene (27). The differences of the pathology of hepatotoxicity of OT-7100 in humans and rats are consistent with the metabolic specificity of human and rat P450 1A2 being responsible for the differences in M-5 metabolism. Interestingly, up to a 10-fold or more range in interindividual M-23OH formation from M-5 was observed, suggesting large interindividual differences based on inducible human P450 1A2 enzymes involved in metabolism to M-23OH from M-5 in humans. Development of hepatotoxicity in humans by M-5 metabolism may exhibit wide individual differences. We therefore focused on the differences of M-5 metabolism by P450 1A2. Human and rat liver microsomes were capable of activating M-5 to a covalently bound metabolite (Figure 6). The covalent binding to liver proteins was dose-dependently inhibited by reduced glutathione, L-cysteine, or N-acetylcysteine. In preliminary studies, any hepatotoxicity of OT-7100 or M-5 in cultured human HepG2 cells was not observed. A covalent binding of 14C-M-5 to 9,000 g of liver supernatant fraction was higher in humans than in rats after 6 h incubations. Taken together, liver microsomal bioactivation was a key determinant factor, supporting the present microsomal binding analysis (Figure 6). As shown in Figure 8B, the 1H NMR spectrum of the M-23OH-cysteine conjugate showed one aromatic proton signal (δ 6.85). By comparison to the 1H NMR spectrum of M-23OH (Figure 8A), if cysteine is substituted on the 3-position carbon, two aromatic proton signals should appear in the region from 5 ppm to 9 ppm in the 1H NMR spectrum (Figure 8B). Therefore, it was suggested that cysteine was substituted on the 6-position carbon of M-23OH (Figure 9). Thereafter, it must be substituted by the glutathione on the 6-position carbon of

the pyrazolo pyrimidine in the same way as in the case of the M-23OH-cysteine conjugate (Figure 7A). The M-23OH-glutathione conjugate was identified in incubation with human liver microsomes in the presence of M-5, the NADPH-generating system, and glutathione (results not shown). M-22OH might theoretically form another reactive quinone-imine to sulfhydryl adduct at the 3-position. However, we were not able to detect such metabolites as mentioned above in incubation with either human and rat liver microsomes or recombinant rat and human P450 1A2 in the presence of M-5, the NADPH-generating system, and glutathione (results not shown). We could consequently conclude that the ortho-quinone-imine derived from M-22OH would be more unstable than the para-quinone-imine derived from M-23OH and that the ortho-quinone-imine would degrade immediately before binding to another molecule. Proposed species differences in the metabolic pathways of primary metabolites of OT-7100 are shown in Figure 10. Because hydroxyl group is substituted on the 6-position carbon for M-22OH, quinone imine intermediates from M-22OH would not be bound to the proteins on the 6-position. Because the C-6positon of M-23OH is free, a quinone imine intermediate from M-23OH would be covalently bound to proteins in the absence of reduced glutathione. These results suggested that P450 1A2mediated C-3-hydroxylation of M-5 was the metabolically activating pathway and that those conjugation reactions represent an important detoxication pathway of OT-7100 in rats, which may not occur in some humans. As in the cases of acetaminophen (17-20), depletion of glutathione after high doses of OT-7100 to individuals, resulting in a saturation of detoxification, might be a causal factor in limited elevations in serum levels of aspartate- and alanine-aminotransferase abnormally observed in a clinical study, in which most cases (7 out of 11) occurred with >1400 mg/day of OT-7100 for >14 days after 28-day-treatments in a dose-escalation study (a total of 56 subjects for placebo and OT-7100). The increases were almost exclusively in aspartate- and alanine-aminotransferase; in some

Human P450 1A2 in ReactiVe Metabolite Formation

patients, small elevations in lactate dehydrogenase were seen but not in alkaline phosphatase or bilirubin levels. In conclusion, differences in the metabolic function of human and rat P450 1A2 would be responsible for the human-specific metabolic activation of primary metabolite M-5 of the parent compound OT-7100 to a proximate form. Our results collectively suggested that the hepatotoxicity associated with OT7100 would be most likely related to the formation of a reactive metabolite from M-23OH, which is free at the C-6-position and could yield a putative quinone imine metabolite. This reactive intermediate from M-23OH might be suggested to be the toxic molecule for OT-7100, but further conformational studies are necessary. However, the present species specificity of P450 1A2 alone may not be fully responsible for the species difference in hepatotoxicity of OT-7100 because any possibilities for species specificity in diminishing the active metabolites or repairing the damages could not be ruled out after metabolic activation. Consequently, glutathione in the liver was also found to be an important factor for detoxification of the reactive metabolites formed from this model compound. It is considered that elucidation of detoxication pathways and species differences in formation pathways of active intermediates between humans and experimental animals are necessary to assume human hepatotoxicity risk.

Chem. Res. Toxicol., Vol. 22, No. 2, 2009 331

(11)

(12)

(13) (14)

(15)

(16)

(17) (18)

Acknowledgment. This work was supported in part by the Ministry of Education, Science, Sports and Culture of Japan. (19)

References (1) Guengerich, F. P. (1995) Human Cytochrome P450 Enzymes. In Cytochrome P450 (Oritiz de Montellano, P. R., Ed.) pp 473-535, Kluwer Academic/Plenum Press, New York. (2) Shimada, T., Yamazaki, H., Mimura, M., Inui, Y., and Guengerich, F. P. (1994) Interindividual variations in human liver cytochrome P-450 enzymes involved in the oxidation of drugs, carcinogens and toxic chemicals: studies with liver microsomes of 30 Japanese and 30 Caucasians. J. Pharmacol. Exp. Ther. 270, 414–423. (3) Yamazaki, H., Shibata, A., Suzuki, M., Nakajima, M., Shimada, N., Guengerich, F. P., and Yokoi, T. (1999) Oxidation of troglitazone to a quinone-type metabolite catalyzed by cytochrome P450 2C8 and P450 3A4 in human liver microsomes. Drug Metab. Dispos. 27, 1260– 1266. (4) Watanabe, I., Tomita, A., Shimizu, M., Sugawara, M., Yasumo, H., Koishi, R., Takahashi, T., Miyoshi, K., Nakamura, K., Izumi, T., Matsushita, Y., Furukawa, H., Haruyama, H., and Koga, T. (2003) A study to survey susceptible genetic factors responsible for troglitazoneassociated hepatotoxicity in Japanese patients with type 2 diabetes mellitus. Clin. Pharmacol. Ther. 73 (5), 435–455. (5) Joshi, E. M., Heasley, B. H., Chordia, M. D., and Macdonald, T. L. (2004) In vitro metabolism of 2-acetylbenzothiophene: Relevance to zileuton hepatotoxicity. Chem. Res. Toxicol. 17, 137–143. (6) Park, B. K., Kitteringham, N. R., Maggs, J. L., Pirmohamed, M., and Williams, D. P. (2005) The role of metabolic activation in drug-induced hepatotoxicity. Annu. ReV. Pharmacol. Toxicol 45, 177–202. (7) Lee, W. M. (2003) Drug-induced hepatotoxicity. N. Engl. J. Med. 349, 474–485. (8) James, L. P., Mayeux, P. R., and Hinson, J. A. (2003) Acetaminopheninduced hepatotoxicity. Drug Metab. Dispos. 31, 1499–1506. (9) Bort, R., Ponsoda, X., Jover, R., Go´mez-Lecho´n, M. J., and Castell, J. V. (1999) Diclofenac toxicity to hepatocytes: a role for drug metabolism in cell toxicity. J. Pharmacol. Exp. Ther. 288, 65–72. (10) Kassahun, K., Pearson, P. G., Tang, W., McIntosh, I., Leung, K., Elmore, C., Dean, D., Wang, R., Doss, G., and Baillie, T. A. (2001)

(20)

(21) (22)

(23) (24)

(25)

(26) (27)

Studies on the metabolism of troglitazone to reactive intermediates in vitro and in vivo. Evidence for novel biotransformation pathways involving quinone methide formation and thiazolidinedione ring scission. Chem. Res. Toxicol. 14, 62–70. Kuribayashi, S., Ueda, N., Naito, S., Yamazaki, H., and Kamataki, T. (2006) Species differences in hydrolase activities toward OT-7100 responsible for different bioavailability in rats, dogs, monkeys and humans. Xenobiotica. 36, 301–314. Chan, W. K., Sui, Z., and Ortiz de Montellano, P. R. (1993) Determinants of protein modification versus heme alkylation: inactivation of cytochrome P450 1A1 by 1-ethynylpyrene and phenylacetylene. Chem. Res. Toxicol. 6, 38–45. Raucy, J. L., Lasker, J. M., Lieber, C. S., and Black, M. (1989) Acetaminophen activation by human liver cytochromes P450IIE1 and P450IA2. Arch. Biophys. 271, 270–283. Patten, C. J., Thomas, P. E., Guy, R. L., Lee, M., Gonzalez, F. J., Guengerich, F. P., and Yang, C. S. (1993) Cytochrome P-450 enzymes involved in acetaminophen activation by rat and human liver microsomes and their kinetics. Chem. Res. Toxicol. 6, 511–518. Thummel, K. E., Lee, C. A., Kunze, K. L., Nelson, S. D., and Slattery, J. T. (1993) Oxidation of acetaminophen to N-acetyl-p-aminobenzoquinone imine by human CYP3A4. Biochem. Pharmacol. 45, 1563– 1569. Chen, W., Koenigs, L. L., Thompson, S. J., Peter, R. M., Rettie, A. E., Trager, W. F., and Nelson, S. D. (1998) Oxidation of acetaminophen to its toxic quinone imine and nontoxic catechol metabolites by baculovirus-expressed and purified human cytochromes P450 2E1 and 2A6. Chem. Res. Toxicol. 11, 295–301. Miners, J. O., Drew, R., and Birkett, D. J. (1984) Mechanism of action of paracetamol protective agents in mice in vivo. Biochem. Pharmacol. 33, 2995–3000. Streeter, A. J., Dahlin, D. C., Nelson, S. D., and Baillie, T. A. (1984) The covalent binding of acetaminophen to protein. Evidence for cysteine residues as major sites of arylation in vitro. Chem.-Biol. Interact. 48, 349–366. Blake, K. V., Bailey, D., Zientek, G. M., and Hendeles, L. (1988) Death of a child associated with multiple overdoses of acetaminophen. Clin. Pharm. 7, 391–397. Roberts, D. W., Bucci, T., Benson, R. W., Warbritton, A. R., McRae, T. A., Pumford, N. R., and Hinson, J. A. (1991) Immunohistochemical localization and quantification of the 3-(cystein-S-yl)-acetaminophen protein adduct in acetaminophen hepatotoxicity. Am. J. Pathol. 138, 359–371. Smith, M. T. (2003) Mechanisms of troglitazone hepatotoxicity. Chem. Res. Toxicol. 16, 679–687. Herman, J. R., Dethloff, L. A., McGuire, E. J., Parker, R. F., Walsh, K. M., Gough, A. W., Masuda, H., and De la Iglesia, F. A. (2002) Rodent carcinogenicity with the thiazolidinedione antidiabetic agent troglitazone. Toxicol. Sci. 68, 226–236. Rothwell, C., McGuire, E. J., Altrogge, D., Masuda, H., and De la Iglesia, F. A. (2002) Chronic toxicity in monkeys with the thiazolidinedione antidiabetic agent troglitazone. J. Toxicol. Sci. 27, 35–47. Butler, M. A., Iwasaki, M., Guengerich, F. P., and Kadlubar, F. F. (1989) Human cytochrome P-450PA (P-450IA2), the phenacetin O-deethylase, is primarily responsible for the hepatic 3-demethylation of caffeine and N-oxidation of carcinogenic arylamines. Proc. Natl. Acad. Sci. U.S.A. 86, 7696–7700. Agu´ndez, J. A. G., Luengo, A., and Benitez, J. (1992) Caffeine demethylase activity in human and dark agouti rat liver microsomes. Comparison with aminopyrine N-demethylase activity. Drug Metab. Dispos. 20, 343–349. Turesky, R. J., Constablea, A., Faya, L. B., and Guengerich, F. P. (1999) Interspecies differences in metabolism of heterocyclic aromatic amines by rat and human P450 1A2. Cancer Lett. 143, 109–112. Langoue¨t, S., Welti, D. H., Kerriguy, N., Fay, L. B., Huynh-Ba, T., Markovic, J., Guengerich, F. P., Guillouzo, A., and Turesky, R. J. (2001) Metabolism of 2-amino-3,8-dimethylimidazo[4,5-f]-quinoxaline in human hepatocytes: 2-Amino-3-methylimidazo[4,5-f]quinoxaline8-carboxylic acid is a major detoxication pathway catalyzed by cytochrome P450 1A2. Chem. Res. Toxicol. 14, 211–221.

TX8003592