Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Abnormally high dopamine activity has been associated with schizophrenic illness (Hollister, 1987). This is demonstrated by the fact that virtually all "typical" antipsychotic agents for the treatment of schizophrenia block postsynaptic dopamine receptors, particularly of the D₂ subtype (Hollister, 1987; Davis et al., 1991). However, most antipsychotic agents induce extrapyramidal side effects due to their inherent dopamine D₂ receptor antagonism, and are only partially effective in alleviating the "negative" symptoms (withdrawal, loss of drive, flattened affect) of schizophrenia (Reynolds, 1992).

The generalized hyperactive dopamine hypothesis has been challenged by the properties of clozapine, a nonclassical antipsychotic that shows only weak D₂ antagonist activity relative to its activity at other receptors (Fitton and Heel, 1990; Reynolds, 1992). Notably, clozapine elicits significantly fewer extrapyramidal side effects, and is more effective in relieving both the "positive" and "negative" symptoms of schizophrenia (Casey, 1989; Fitton and Heel, 1990). However, clozapine is indicated only for the management of severe and chronic schizophrenia refractory to classical antipsychotic therapy (Fitton and Heel, 1990), due to the high incidence of drug-induced agranulocytosis (Lieberman et al., 1988; Krupp and Barnes, 1989). Therefore, there is an unmet medical need for a safer and more effective treatment of schizophrenia.

The recent discovery of three additional subtypes of dopamine receptors, i.e., D₃, D₄, and D₅, has provided new potential targets for the treatment of schizophrenia (Sibley and Monsma, 1992; Taubes, 1994). Perhaps one of the most interesting findings associated with the occurrence of these receptors was that the D₄ receptors were reported to be concentrated 6-fold in the caudate nucleus of schizophrenic patients relative to controls (Seeman et al., 1993; Reynolds and Mason, 1995). Moreover, the nonclassical antipsychotic agent clozapine showed much greater affinity toward the D₄ than the D₂ receptor (Sibley and Monsma, 1992). Although recent evidence indicated the D₄ receptor as a promising target to treat schizophrenia, phase II clinical studies with the selective D₄ receptor antagonist L-745,870 (3-{[4-(4-chlorophenyl)piperazin-1-yl]-methyl}-1H-pyrrolo-2,3--pyridine)³ were ineffective in alleviating the symptoms of this psychiatric disease (Kulagowski et al., 1996; Bristow et al., 1997). This study reports the metabolism of L-745,870 in the rat, rhesus monkey, and human, and includes the identification of a novel N-acetylcysteine conjugate (M4), apparently formed through metabolic activation of the azaindole moiety of L-745,870.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. L-745,870 was synthesized at the Neuroscience Research Center, Merck Research Laboratories (Terlings Park, UK), whereas [¹⁴C]L-745,870 (9.7 µCi/mmol, 99.9% pure by HPLC) was prepared by the Labeled Compound Synthesis Group, Department of Drug Metabolism, Merck Research Laboratories (Rahway, NJ). The following metabolites also were synthesized at the Neuroscience Research Center: 1) 7-azaindole-3-methyl-mercapturic acid (M4), 2) 7-azaindole-3-carboxylic acid (M5), 3) N-hydroxy-p-chlorophenylpiperazine (M3), and 4) 3-formyl-7-azaindole (M1). An authentic standard of the GSH adduct of 3-methyl-7-azaindole was prepared by approach similar to that used for the preparation of the 3-methylindole N-acetylcysteine adduct (Skiles et al., 1991). p-Chlorophenylpiperazine (M2) was purchased from Aldrich (Milwaukee, WI). HPLC solvents, at the highest purity grade, were obtained from Fisher Scientific (Pittsburgh, PA).

Animal Studies. Three separate experiments in rats were carried out as part of an evaluation of the disposition of L-745,870. In one experiment, male Sprague-Dawley rats (267-368 g, n = 4) received [¹⁴C]L-745,870 at 1 mg/kg i.v., and urine was collected from 0 to 24 h. In the second experiment, animals were dosed i.v. (1 mg/kg, n = 4) with [¹⁴C]L-745,870, and blood samples were obtained via cardiac puncture at 1 h. Plasma was obtained by centrifugation at 15,000g for 5 min. In a third experiment, two bile duct-cannulated rats were dosed with [¹⁴C]L-745,870 (1 mg/kg i.v.), and bile samples were collected from 0 to 6 h (at 1-h intervals) and as a single pool from 6 to 24 h. All samples were collected over ice or dry ice and stored at 20°C until analyzed.

Human Studies. Two groups of 12 healthy male volunteers participated in a phase I clinical study, in which unlabeled L-745,870 (n = 9) or placebo (n = 3) was administered orally. One group received a single 10-mg dose per day for 10 days, whereas the second group received 25 mg/day for 14 days. Urine samples (0-24 h) were collected on the first and last day of the study.

Sample Preparation. Aliquots of urine (1-ml, rat and monkey; 5-ml, human) were either filtered through a 0.2-µm nylon filter or treated with four volumes of acetone or methanol to precipitate inorganic salts. The sample was concentrated under N₂ and dried under reduced pressure. Before liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis, the dried residue was reconstituted in aqueous formic acid (0.1%) containing 10% acetonitrile (200-400 µl). Plasma samples from rats (1-ml) and monkeys (0.5-ml from each of four monkeys, pooled from 2- and 4-h time points) were treated with two volumes of acetonitrile to precipitate proteins. The precipitate was removed by centrifugation, the supernatants were dried under N₂ or vacuum, and the residues were reconstituted in aqueous formic acid (0.1%) containing 10% acetonitrile for LC-MS/MS analysis. Aliquots of rat bile (10% of total volume collected) were pooled proportionally between 1 and 6 h to represent a 0 to 6 h sample, and analyzed directly by LC-MS/MS.

Isolation and Purification of M4. Isolation and purification of the major radioactive metabolite of L-745,870 from specimens of filtered rat urine (250 µl) were carried out by sequential HPLC analyses (Hewlett-Packard 1090; Hewlett-Packard Co., Palo Alto, CA), using a Zorbax Rx C-8 column (5 µm, 4.6 × 250 mm) and a UV detector ( = 290 nm). Elution was carried out at a flow rate of 1 ml/min with three different mobile phase systems. System 1 used a gradient of 0 to 10 min of 100% solvent B, followed by a linear increase in solvent A to 50% over 50 min, where A = acetonitrile and B = 0.1% aqueous trifluoroacetic acid (TFA) with triethylamine, pH 3.5. One-minute fractions were collected, and aliquots of each were analyzed by liquid scintillation counting. The fractions containing the greatest amount of radioactivity from this system were pooled and reduced to dryness under a stream of N₂. The residue was reconstituted in 0.1% aqueous TFA and injected onto system 2 with the following linear gradient: 0 to 40% A over 40 min, where A = acetonitrile and B = 0.1% aqueous TFA. Once again, 1-min fractions were collected, and aliquots of each fraction were analyzed by liquid scintillation counting. Finally, the fractions containing the majority of the radioactivity from system 2 were pooled and dried under N₂. The residue was reconstituted in 0.1% aqueous TFA and chromatographed in system 3, which was an isocratic mobile phase maintained at 7.5% A and 92.5% B, where A = acetonitrile and B = 0.1% aqueous TFA. Fractions were collected at 40-s intervals, and aliquots of each were analyzed by liquid scintillation counting. The fractions containing the majority of the radioactivity from system 3 were pooled and dried under N₂. These final residues then were dissolved in deuteromethanol for analysis by NMR.

NMR Spectroscopy. In addition to analysis by HPLC and mass spectrometry, metabolites of L-745,870 were characterized by NMR spectroscopy and comparison with authentic standards. ¹H NMR spectra were recorded on a Varian Unity spectrometer (Varian Inc., Palo Alto, CA) at 500 MHz using a Nalorac (Nalorac Corp., Martinez, CA) 3-mm microprobe. Samples (ca. 10-ug) were dissolved in deuteromethanol (CD₃OD). Chemical shifts are reported in ppm () and referenced to tetramethylsilane using residual solvent signal.

Identification of Metabolites by LC-MS/MS. The LC-MS/MS system consisted of a Sciex API III⁺ triple quadrupole mass spectrometer, interfaced to a HP1050 HPLC instrument equipped with a Zorbax Rx C-8 narrow bore column (3 µm, 2.1 × 150 mm; Mac-Mod Analytical Inc., Chadds Ford, PA). The mobile phase consisted of solvent A (acetonitrile) and solvent B (0.1% aqueous formic acid). Gradient elution was conducted from 5 to 55% A over 25 min, with a flow rate of 200 µl/min. The HPLC effluent was split such that 25% was directed to the mass spectrometer, whereas 75% was passed to a radiochemical detector (-RAM; IN/US, Tampa, FL). The radiochemical detector used a 100-µl liquid flow cell, and the HPLC effluent and scintillation cocktail (Ready Flow III; Beckman Instruments, Fullerton, CA) were mixed in a ratio of 1:2.

Quantitative Analysis of M4 in Human Urine. Urine samples (200-µl) were treated with an equal volume of 0.1% aqueous formic acid and analyzed directly by LC-MS/MS (100-µl injections). Conditions for LC-MS/MS were similar to those used for metabolite identification, except that the mobile phase was an isocratic mixture of 25% acetonitrile in 0.1% aqueous formic acid, and the mass spectrometer was operated in the selected reaction monitoring. M4 was detected using the parent  product ion pair of m/z 294  131. Quantitation was achieved by using an authentic sample of 3-[(N-acetylcysteine-S-yl)-methyl]-7-azaindole (M4) as the standard. A calibration curve in the range of 0.05 to 5 µg/ml of M4 was constructed from control (drug-free) human urine. The assay was accurate between 0.1 and 5 µg/ml (±20%). Concentrations of M4 in urine samples were found to fall within the range 0.5 to 4 µg/ml.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Metabolite Profiles. After a 1 mg/kg i.v. dose of [¹⁴C]745,870 to intact and bile duct-cannulated rats, an average of 48.3% of the dose was excreted in 0 to 24 h urine, whereas 25.3% of the dose was excreted in bile from 0 to 6 h. Radiochromatographic analysis of specimens of rat plasma (1-h), urine (0-24 h), and bile (0-6 h), after a 1 mg/kg i.v. dose of [¹⁴C]L-745,870, revealed the presence of several radioactive metabolites. Three of these metabolites, with retention times of 13.5 (M4), 15.5 (M5), and 18 min (M6), were characterized (Fig. 1). M4 was present in all three matrices, and accounted for 13, 34, and 10% of total radioactivity in rat plasma, urine, and bile, respectively. M5 was present only in urine (10% of total radioactivity), whereas M6 was present in plasma and urine, where it accounted for 62 and 17% of the total radioactivity, respectively. Parent drug was detected in plasma only, with a retention time of 21 min (Fig. 1A). The most abundant radioactive component (~50%) in bile is not parent compound and its identity has not yet been determined. In rhesus monkey, radiochromatograms of plasma (pooled from 2-4 h) and urine (0-24 h, containing 48.4% of the dose), after a 1 mg/kg i.v. dose of [¹⁴C]L-745,870 are shown in Fig. 2. In plasma, M5 was the most abundant peak detected, whereas it was absent in rat plasma. Similar to rat urine, M4 was the most abundant peak detected in monkey urine, but, unlike rats, monkeys also excreted numerous other minor, unresolved metabolites.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Radiochromatograms of plasma (A), urine (B), and bile (C) from rats dosed with [14C]L-745,870 (1 mg/kg i.v.).

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Radiochromatograms of plasma (A) and urine (B) from rhesus monkeys dosed with [14C]L-745,870 (1 mg/kg i.v.).

Identification of Metabolites. A product ion mass spectrum of the parent drug was obtained to facilitate structural characterization of the new metabolites by mass spectrometry (Fig. 3). The protonated molecular ion ([M+H]⁺) was present at m/z 327, which on CID gave rise to two abundant fragment ions at m/z 197 and 131. Both of these fragments resulted from a single cleavage at the methylene bridge between the azaindole and p-chlorophenylpiperazine moieties, with charge retention on either "half" of the molecule.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Product ion mass spectrum of L-745,870 obtained by CID of the [M+H]+ ion at m/z 327.  The proposed origin of the characteristic fragments at m/z 197 and 131 is as shown.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   1H NMR (500 MHz, CD3OD) spectrum of M4 purified from rat urine. *, Indicates signals from solvent and/or impurities.

View larger version (12K): [in this window] [in a new window]   Fig. 5.   Product ion mass spectra of M5 from rat urine (A) and the corresponding synthetic standard (B). The spectra were recorded in the positive ion mode and were obtained by CID of the [M+H]+ ion at m/z 163.

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Product ion mass spectra of M6 from rat plasma recorded in positive (A) and negative (B) ion modes, and obtained by CID of parent ions at m/z 423 and 421, respectively.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Total ion current chromatogram and extracted ion current chromatograms from the LC-MS/MS analysis of urine (2-6 h sample) from a human subject given an oral 10-mg dose of L-745,870. Ions with m/z values of 327, 197, 294, and 163 correspond to the [M+H]+ species of L-745,870, M2, M4, and M5, respectively.

Quantitative Determination of M4. After a single 10-mg oral dose of L-745,870 to nine human subjects, urinary excretion of M4 accounted for 17 ± 4% of the dose (Table 1). After ten successive daily 10-mg doses (to steady state), the corresponding figure was 34 ± 11% of the dose (Table 2).

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

In this study, it was shown that the dopamine D₄ antagonist L-745,870 undergoes extensive metabolism in the rat, rhesus monkey, and human, with only minor amounts of parent drug being excreted. As shown in Fig. 8, L-745,870 undergoes metabolism via three different pathways, namely, N-dealkylation, aromatic hydroxylation followed by sulfation, and glutathione conjugation, which leads to the formation of a mercapturic acid adduct. The two products of N-dealkylation, the acid M5 and the amine M2, were observed in all three species. However, formation of the sulfate conjugate of a hydroxylated derivative of L-745,870 was species-dependent, as it was observed in rats only.

View larger version (17K): [in this window] [in a new window]   Fig. 8.   Proposed scheme for the metabolism of L-745,870 in the rat, rhesus monkey, and human. 3-Formyl-7-azaindole (M1) was determined as the major metabolite from in vitro experiments in rat liver microsomes (P. Kari, K. Zhang, B. Arison and K. Vyas, unpublished results).

A notable finding of this in vivo study was the conversion of a substantial fraction of the dose of L-745,870 to a mercapturic acid adduct in all three species examined. These findings suggest that L-745,870 underwent biotransformation to a reactive electrophilic intermediate, which was trapped via conjugation with the physiological nucleophile GSH. Subsequent metabolism of this putative GSH conjugate would yield the corresponding mercapturic acid adduct M4. Interestingly, attempts to detect the GSH conjugate in specimens of bile and plasma from rats dosed with L-745,870 were unsuccessful, although it should be noted that liver tissue contains all of the enzymes necessary for the metabolism of GSH adducts to mercapturic acid derivatives (Hinchman et al., 1991). The intermediacy of a GSH conjugate of 3-methyl-7-azaindole en route to metabolite M4 remains speculative at this point.

Recently, Skordos and coworkers (1998) showed that 3-methylindole, a pneumotoxic compound (Nocerini et al., 1985; reviewed in Yost, 1989), is bioactivated to not only an imine methide intermediate (a product of lung-specific CYP2F catalysis; Thornton-Manning et al., 1996; Wang et al., 1998; Lanza et al., 1999), but also to a 2,3-epoxide intermediate. Although the precise mechanism by which L-745,870 undergoes metabolic activation is not understood at this time, an imine methide derivative represents a likely candidate for the reactive intermediate, based on the structure of the final metabolic product. Whereas the imine methide in question (Fig. 8) corresponds to the 7-aza derivative of the imine methide generated from 3-methylindole, this report appears to be the first example of metabolic activation of the 7-azaindole nucleus. It is tempting to speculate that the initial metabolic activation of L-745,870 occurs via N-oxidation of the piperazine nitrogen proximal to the azaindole moiety, and in vitro experiments are in progress to identify the underlying mechanism involved in the bioactivation of L-745,870.