Different metabolites of human hepatotoxic pyrazolopyrimidine derivative 5-n-butyl-pyrazolo[1,5-a]pyrimidine produced by human, rat and monkey cytochrome P450 1A2 and liver microsomes.

Drug-induced liver injury is one of the most frequent single causes of safety-related withdrawals of drugs from the market [1]. In a clinical study of 5-n-butyl-7-(3,4,5-trimethoxybenzoylamino)pyrazolo[1,5-a]pyrimidine (OT-7100, an amide moiety-bearing pyrazolopyrimidine derivative with potential analgesic effects) [2], limited elevations in the serum levels of aspartate or alanine aminotransferase were occasionally observed in human beings; these elevations were not predicted from regulatory animal or in vitro studies [3]. There were only limited pre-clinical studies with monkeys, that is non-human primate species, before the clinical trial in human beings. Moreover, the underlying molecular mechanisms of the elevation in levels of liver damage markers are not yet fully elucidated, although no apparent differences in plasma OT-7100 levels were found in volunteers between the toxic cases and other groups in our limited observations. The purpose of this study was to characterize and understand the metabolites of 5-n-butyl-pyrazolo[1,5-a]pyrimidine (M-5, the primary metabolite of OT-7100) [2] produced by monkey liver microsomal P450 enzymes in comparison with those produced by human and rat enzymes. We report herein that M-5 was metabolically activated by monkey liver microsomal P450 1A2 in a different manner from rat P450 1A2, but in a similar way to that of human P450 1A2. 5-n-Butyl-pyrazolo[1,5-a]pyrimidine (a primary metabolite of OT-700, M-5), 3-hydroxy-5-n-butyl-pyrazolo[1,5-a]pyrimidine (M-23OH), 6-hydroxy-5-n-butyl-pyrazolo[1,5-a]pyrimidine (M-22OH) and 5-n-pentyl-7-(3,4,5-trimethoxybenzoylamino)pyrazolo[1,5-a]pyrimidine (OT-7126) as an internal standard for HPLC analysis were synthesized at Otsuka Pharmaceutical Factory (Tokushima, Japan). All these chemicals were determined by reversed-phase HPLC to be >99.0% pure. Reduced glutathione was purchased from Sigma-Aldrich (St. Louis, MO, USA). Human liver microsomes were obtained from Xenotech (Lenexa, KS, USA). Rat and monkey liver microsomes and rat P450 1A2 in baculovirus-infected insect cells (Supersomes) were purchased from BD Bioscience (Woburn, MA, USA). Recombinant human and monkey P450 1A enzymes co-expressed with NADPH-P450 reductase were prepared [4,5]. All other chemicals and reagents used were of analytical reagent grade. A typical in vitro incubation mixture (0.50 mL total volume) contained liver microsomes from human beings, rats or monkeys (0.48 mg/mL) or recombinant human, rat, or monkey P450 1A1 or 1A2 (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 M-5 (5.0 μM) in the presence of reduced glutathione (50 mM). After a 5-min. pre-incubation, the reactions were initiated by addition of the NADPH-generating system and incubation was continued 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 2000 × g 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 LC system. M-5 and its metabolites were assayed using UV-LC methods validated over a concentration range of 0.10–10.0 μM. M-5 and its metabolites in incubation mixtures were separated on a 250 × 4.6 mm i.d. Inertsil ODS-3V column (GL Sciences, Tokyo, Japan) and detected at wavelengths of 230 nm 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 elution were as follows: 13–25% B (0–8 min.), 25–25% B (8–20 min.), 25–35% B (20–25 min.), 35–80% B (25–31 min.) and 80–80% B (31–40 min.). Linear gradients were used for all solvent changes. The flow rate was 0.80 mL/min. in LC assays. Monkey and rat P450 1A primary sequences were aligned with a crystal structure of human P450 1A2 (Protein Data Bank code 2HI4) [6] using MOE software (ver. 2009.10; Computing Group, Montreal, Canada) for modelling of the three-dimensional structure. Prior to docking simulation, the energy of the P450 structures was minimized using the CHARM22 force field. Docking simulation was carried out for M-5 binding to P450 enzymes using the MMFF94x force field distributed in the MOE Dock software [7]. Twenty solutions were generated for each docking experiment and ranked according to the total interaction energy (U value). Summarized in table 1 is the metabolite formation from M-5 (5.0 μM) catalysed by liver microsomes and recombinant P450 1A enzymes in the presence of an NADPH-generating system and reduced glutathione (50 mM). Human liver microsomes and human P450 1A2 preferentially metabolized M-5 to M-23OH (peak 5 in 1-3, a C-3-position hydroxyl derivative). Rat liver microsomes and rat P450 1A2 rapidly and differently metabolized M-5 to M-22OH (peak 4, a C-6-position hydroxyl derivative) and an unknown metabolite(s) (peak 2, fig. 1B,E). Monkey liver microsomes and monkey P450 1A2, but not monkey P450 1A1, metabolized M-5 to M-22OH and M-23OH in similar amounts (peaks 4 and 5) (fig. 1A,D). A M-23OH–glutathione conjugate (peak 1 [3] was also detected in the presence of glutathione and NADPH after incubation of M-5 with all enzyme sources except for the case of monkey P450 1A1. Representative LC chromatograms of in vitro metabolites from 5-n-butyl-pyrazolo[1,5-a]pyrimidine (M-5) produced by incubation with liver microsomes from human beings (A), rats (B) and monkeys (C) and by incubation with human P450 1A2 (D), rat P450 1A2 (E), and monkey P450 1A1 and 1A2 (F,G) in the presence of reduced glutathione. Liver microsomes (0.48 mg/mL) or recombinant P450 enzymes (20 pmol/mL) were incubated with 5.0 μM of 5-n-butyl-pyrazolo[1,5-a]pyrimidine (M-5) for 30 min. with an NADPH-generating system and reduced glutathione (50 mM). Peak 1, a glutathione conjugate of M-23OH; Peak 2, rat-specific unknown product-1; Peak 3, unknown product-2; Peak 4, M-22OH, Peak 5, M-23OH; Peak 6, M-5 (substrate); and Peak 7, internal standard. Docking simulation of interaction of 5-n-butyl-pyrazolo[1,5-a]pyrimidine (M-5) with human P450 1A2 (A), rat P450 1A2 (B) and monkey P450 1A2 (C). The haem group of the P450 is shown in the lower part of each of the figures. Metabolic pathways of 5-n-butyl-7-(3,4,5-trimethoxybenzoylamino)pyrazolo[1,5-a]pyrimidine (OT-7100), M-5 and M-23OH in monkeys. Molecular docking of M-5 to the active sites of reported human P450 1A2 (Protein Data Bank code 2HI4) [6] and modelled rat and monkey P450 1A2 enzymes was investigated (fig. 2). In human beings and monkeys, the C-3-carbon of M-5 was positioned towards the active sites of P450 1A enzymes and the ligand–P450 interaction energies (U values) were found to be −15.2 and −28.7, respectively (fig. 2A,C). In contrast, the C-6-carbon of M-5 docked into the active site of rat P450 1A2 with a U value of −18.7 (fig. 2B). These findings suggest that monkey P450 1A2 may have some similarities as human P450 1A2 but belong to a third group of catalytic function in terms of activation and inactivation of this species-dependent target compound M-5. M-5 is the primary metabolite of former drug candidate OT-7100 (fig. 3), which had a potential analgesic effect and was without apparent toxicity in rats and dogs [3]. Proposed species differences in the metabolic pathways of the primary metabolites of OT-7100 could be shown because the hydroxyl group is substituted at the C-6-position for M-22OH, but expected quinone imine intermediates from M-22OH would not be bound to proteins at the C-6-position (fig. 3). Because the C-6-position of M-23OH is free (fig. 3), the quinone imine intermediate formed from M-23OH would be covalently bound to proteins in the absence of reduced glutathione [4]. In our preliminary study, treatment of rats with 1000 mg/kg of OT-7100 or M-5 produced no apparent hepatotoxicity. These results suggest that human and monkey P450 1A2-mediated C-3-hydroxylation of M-5 was the metabolically activating pathway of OT-7100. Therefore, it could be summarized that M-5 is metabolized by the human P450 1A2 enzyme to form M-23OH (table 1), which conjugates with a peptide to form an adduct that has liver toxicity. In rats, the same enzyme forms some M-23OH but leads predominantly to M-22OH (table 1), which shows no toxicity. These metabolic differences between rats and human beings highlight some of the perils of testing drug candidates for metabolite toxicity. Therefore, it is of interest to know whether the monkey could be a good predictor in pre-clinical studies for human-specific liver toxicity of drug candidates like OT-7100. The depletion of glutathione that occurred after high doses of OT-7100 were administered to patients in a clinical study, resulting in a saturation of detoxification, might be a causal factor in the limited elevations in serum levels of aspartate and alanine aminotransferase observed. Most of these cases (seven of 11) occurred with >1400 mg/day of OT-7100 for >14 days after 28-day treatments in a dose-escalation study [3]. After this unexpected clinical event, monkeys were orally treated with 1000 mg/kg of OT-7100; this resulted in undetectable levels of M-5 in plasma and no hepatotoxicity. On single treatment of monkeys with 1000 mg/kg of M-5, limited elevations (2–3 times) in serum levels of aspartate and alanine aminotransferase were observed, but there were no increases in alkaline phosphatase or γ-glutamyl transpeptidase levels. In this study, bioactivation of this primary metabolite was effectively catalysed by recombinant monkey and human P450 1A2 in vitro through C-3-hydroxylation to form the glutathione adduct. The hepatotoxic proximate compound M-5 was activated by monkey and human liver microsomal P450 1A2 to reactive intermediate(s) and could relatively non-specifically bind to biomolecules, such as those in the livers of humanized mice [4]. All three species in this study may form reactive metabolite to some degree or the other; it would be the glutathione detoxication mechanism (possibly mediated by glutathione transferase) that may ultimately account for the species differences in toxicity. Further investigation of the glutathione levels and function of glutathione S-transferases in monkey livers [8] would be of interest with regard to the in vivo deactivation of drugs in successful pre-clinical studies. The monkey is a non-human primate species of great importance to drug metabolism studies and toxicity tests in pre-clinical studies because of its evolutionary closeness to human beings [9]. In this species, numerous P450s have been identified but have not been fully characterized [9]. Cynomolgus P450 1A1 and P450 1A2 are functional drug-metabolizing enzymes involved in ethoxyresorufin and caffeine metabolism [5]. Cynomolgus P450 1A1 and P450 1A2 mRNAs are predominantly expressed in liver, where the expression level of the former is much greater than that of the latter [5,10] Despite the relatively lower expression of P450 1A2 in liver, this study showed that P450 1A2 was responsible for metabolite formation of the drug, providing evidence for the potential importance of P450 1A2 in drug metabolism studies. In conclusion, monkey liver P450 1A2 exhibits different functionality from rat P450 1A2 but similar functionality to human P450 1A2, in the bioactivation of primary metabolite M-5 of OT-7100; this finding was supported by in silico docking simulations and indicates the usefulness of using monkey liver enzymes to give accurate pre-clinical predictions. The authors thank Drs. Norie Murayama and Makiko Shimizu for their assistance. This work was supported in part by the Ministry of Education, Culture, Sports, Science and Technology of Japan (H.Y.)