934
Chem. Res. Toxicol. 2005, 18, 934-945
Articles Bioactivation of 2,3-Diaminopyridine-Containing Bradykinin B1 Receptor Antagonists: Irreversible Binding to Liver Microsomal Proteins and Formation of Glutathione Conjugates Cuyue Tang,*,† Raju Subramanian,† Yuhsin Kuo,† Sergey Krymgold,† Ping Lu,† Scott D. Kuduk,‡ Christina Ng,‡ Dong-Mei Feng,‡ Chad Elmore,§ Eric Soli,§ Jonathan Ho,§ Mark G. Bock,‡ Thomas A. Baillie,† and Thomayant Prueksaritanont† Departments of Drug Metabolism and Medicinal Chemistry, Merck Research Laboratories, West Point, Pennsylvania, and Department of Drug Metabolism, Merck Research Laboratories, Rahway, New Jersey Received February 16, 2005
The 2,3-diaminopyridine (DAP) moiety was found to represent a core structure essential for the potency of a new series of human bradykinin B1 receptor antagonists. However, incubation of 14C-labeled 2,3-DAP derivatives with rat and human liver microsomes resulted in substantial irreversible binding of radioactive material to macromolecules by a process that was NADPHdependent. Trapping the reactive species with GSH led to significant reduction of the irreversible binding of radioactivity, with concomitant formation of abundant GSH adducts. One type of thiol adducts (detected in both human and rat liver microsomes), resulting from addition of 305 Da to the parent compound, was observed with all 2,3-DAP compounds. These adducts also were detected in rat hepatocyte incubates, as well as in rat bile, following intravenous administration of 2,3-DAPs. Formation of the conjugates appeared to involve modification of the DAP ring, based upon mass spectral analysis of a number of representative GSH adducts; this was corroborated by detailed LC NMR analysis of one compound. Formation of this type of GSH conjugate was markedly reduced when the 2-amino methyl group linking the 2,3-DAP and the biphenyl moiety was replaced with an ether oxygen atom. It is postulated, therefore, that oxidation of the 2-amino group serves as a key step leading to the formation of reactive species associated with the DAP core. In addition, this step appears to be mediated primarily by P450 3A, as evidenced by the marked decrease in both the irreversible binding of radioactivity and the formation of the GSH adducts in human liver microsomes following treatment with ketoconazole and monoclonal antibodies against P450 3A. A mechanism for the bioactivation of 2,3-DAP is proposed wherein oxidation (dehydrogenation or N-hydroxylation followed by dehydration) of the 2-amino group, catalyzed by P450 3A, results in the formation of a highly electrophilic species, pyridine-2,3-diimine.
Introduction The low expression levels of the bradykinin B1 receptor in nondisease states and its inducibility upon injury have made it an attractive therapeutic target in the quest for more selective treatments of chronic nociceptive pain (1, 2). In contrast to the bradykinin B2 receptor, which is expressed constitutively and utilizes bradykinin and kallidin as natural agonists to evoke an acute pain response, the B1 receptor generally is present at low levels under normal conditions. However, the B1 receptor * To whom correspondence should be addressed. Tel: 215-652-9537. Fax: 215-993-3533. E-mail:
[email protected]. † Department of Drug Metabolism, West Point, PA. ‡ Department of Medicinal Chemistry. § Department of Drug Metabolism, Rahway, NJ.
is induced and expressed in peripheral tissues and cells within 2-5 h after tissue injury or inflammation and is activated by the metabolic fragments [des-Arg9]bradykinin and [des-Arg10]kallidin (1, 2). Animal models have shown that peptide-derived B1 antagonists can successfully block hyperalgesia produced by B1 agonists (3). In addition, B1 receptor knockout mice have displayed reduced sensitivity to painful stimuli, while appearing normal in all other respects (4). Of most interest is the finding that the B1 receptor is expressed constitutively in the central nervous system of mice (5), rats (6), and primates (7), suggesting a central role for the receptors along with the accepted peripheral mode of action. Therefore, a potent, selective, orally bioavailable B1 antagonist should have significant potential as a novel therapeutic agent for inflammatory pain.
10.1021/tx0500427 CCC: $30.25 © 2005 American Chemical Society Published on Web 05/25/2005
Bioactivation of 2,3-Diaminopyridines
Chem. Res. Toxicol., Vol. 18, No. 6, 2005 935
Figure 1. Structures of bradykinin B1 receptor antagonists 1-6.
Promising results have been reported for peptide antagonists (8, 9), and more recently nonpeptide antagonists (10, 11), which possessed high affinity for the human B1 receptor and exhibited activity in rodent pain models. While these antagonists have helped validate the B1 receptor as a therapeutic target, they lack the physical properties or pharmacokinetic characteristics required of an orally active pharmaceutical. Following investigations of a series of benzodiazepine and dihydroquinoxaline derivatives, compounds containing a substituted 2,3diaminopyridine (DAP)1 ring system and a biphenyl motif were identified and optimized for their potency and pharmacokinetic characteristics (12, 13). With an increasing awareness of potential toxicological liabilities that can result from metabolic activation of drug candidates, it is essential to identify and to minimize this undesirable property early in the drug discovery phase (14, 15). In this report, we disclose our efforts in evaluating bradykinin B1 receptor antagonists containing the 2,3-DAP ring system with the aid of both 14C-labeled tracers (to assess the propensity to undergo covalent binding to hepatic proteins), and LC-MS/MS analysis (to obtain information on pathways of metabolic activation through the identification of GSH adducts). The results of these studies strongly suggest that the 2,3-DAP moiety is subject to oxidative metabolism leading to the formation of electrophilic intermediates, which covalently modify proteins and react with cellular GSH.
Materials and Methods Chemicals and Enzyme Sources. Compounds 1-6 (Figure 1) and the 14C-labeled analogues of 1-3 were synthesized and purified at Merck Research Laboratories in West Point, PA, and Rahway, NJ, respectively (12, 13). NADPH, ketoconazole, sulfaphenazole, quinidine, and GSH were obtained from Sigma Chemical Co. (St. Louis, MO). HPLC grade solvents were purchased from Fisher Scientific (Pittsburgh, PA). A protein assay reagent kit was purchased from Pierce Chemical Co. (Rockford, IL). Pooled human (n ) 10) and rat (n ) 203) liver microsomal preparations were purchased from Xenotech LLC (Kansas City, 1 Abbreviations: DAP, diaminopyridine; GSH, glutathione; P450, cytochrome P450; CID, collision-induced dissociation; GS1, glutathione conjugates with a mass 305 Da greater than the corresponding parent compound and with the glutathione moiety attached to the pyridine ring; GS2, the cyclized form of GS1; GS3, glutathione conjugates where the site of attachment is at a position other than the pyridine ring.
KS). The mouse ascites containing monoclonal antibodies raised against cytochrome P450 (P450) 3A, 2C, and 2D6 were prepared in-house (16). Effect of Test Compounds on Human Liver Microsomal Testosterone 6β-Hydroxylase Activity. For nonpreincubation-dependent enzyme inhibition, pooled human liver microsomes (0.2 mg protein/mL) were incubated with testosterone (250 µM) and test compounds (0.1-10 µM) at 37 °C for 10 min in the presence of NADPH (1 mM). For preincubationdependent inhibition, microsomes were incubated with test compounds over the same concentration range in the presence of NADPH for 30 min, after which incubation with testosterone (250 µM) was carried out for 10 min. The reaction was quenched with an equal volume of acetonitrile, and the formation of 6βhydroxy-testosterone was measured as described previously (17). Incubation with Human and Rat Liver Microsomes. All incubations were performed in phosphate buffer (100 mM, pH 7.4) containing MgCl2 (10 mM). For determination of metabolic stability, human and rat liver microsomes (0.15 mg/mL) were incubated with test compounds (0.5 µM) for varying periods up to 60 min in the presence NADPH (1 mM). The remaining unchanged compounds were measured, and the linear range of parent compound consumption was used for the calculation of metabolic rates. For metabolite identification, test compounds (10 µM) were incubated with microsomes (1 mg/mL) for 60 or 120 min in the presence of both NADPH and GSH (5 mM). The reaction was terminated by adding two volumes of acetonitrile to precipitate proteins. The resultant supernatant was evaporated to dryness under a stream of nitrogen and reconstituted with 30% aqueous acetonitrile. The effect of chemical inhibitors selective towards different P450s on the formation of GSH adducts in human liver microsomes was investigated using the following inhibitors at the indicated concentrations: ketoconazole (1 µM; P450 3A), sulfaphenazole (2.5 µM; P450 2C9), and quinidine (2.5 µM; P450 2D6). The inhibitors were dissolved in 50% aqueous acetonitrile solution at concentrations 200-fold greater than the final concentrations. The effect of monoclonal antibodies was assessed by pretreating human liver microsomes with mouse ascites fluid containing the antibodies of interest (anti-P450 3A, anti-P450 2C, and anti-P450 2D6, 40 µL ascites fluid/mg protein) for 10 min at room temperature. Incubation with Rat Hepatocytes. Freshly isolated rat (250-300 g, male Sprague-Dawley) hepatocytes were prepared according to the method of Moldeu´s et al. (18). The viabilities of hepatocytes in the present studies were in the order of 92%, as assessed by trypan blue exclusion method. Incubations (total volume of 2 mL) of hepatocytes (2 million viable cells/mL) with 14C-labeled 1 (10 µM) were performed in Krebs-Henseleit buffer (pH 7.4) in 20 mL glass vials capped with white open-top polypropylene closures and 0.005′′ PTFE/0.12′′ silicone rubber
936
Chem. Res. Toxicol., Vol. 18, No. 6, 2005
septa. Each septum was equipped with two PrecisionGlide needles (Becton Dickson and Company, Franklin Lakes, NJ) as an inlet and an outlet for gas flushing. Incubations were performed with continuous shaking (80 times/min) and a gentle flush of gases (O2/CO2 95%/5%) at 37 °C. Acetonitrile (1 mL) was added to terminate the reaction 60 min later. Prior to radiochromatography, the samples were centrifuged to remove cell pellets. Animal Experiments. Experiments were performed according to procedures approved by the Merck Institutional Animal Care and Use Committee. One male Sprague-Dawley rat (288 g) purchased from Taconic Farms (Germantown, NY) was allowed free access to commercial rat chow and water. The animal was anesthetized with Nembutal, and its jugular vein and bile duct were cannulated with PE-10 tubing. An intravenous administration of 14C-1 (1.5 mg/kg) was given to the animal through the jugular vein catheter, and bile and urine were collected for a period of 6 h. Sample Analysis. Quantitative analysis of unchanged compounds in microsomal incubates from metabolic stability studies was accomplished by LC-MS/MS analysis. The separation of metabolites and parent compounds was achieved on a Betasil C18 column (2 mm × 50 mm, 5 µm, Keystone Scientific, Bellefonte, PA) using PE 200 binary pumps (Perkin-Elmer, Inc., Wellesley, MA). Solvent A consisted of 0.02% aqueous acetic acid adjusted to pH 4.5 with NH4OH and acetonitrile (90:10), and solvent B consisted of acetonitrile and water (90:10). The mobile phase was delivered at a flow rate of 0.5 mL/min with a linear increase of solvent B from 15 to 90% over 1 min and held at 90% for an additional minute. Equilibration was allowed for additional 1.5 min, giving a total chromatographic run time of 3.5 min. LC-MS/MS was performed on a Sciex model API 3000 triple quadrupole mass spectrometer (Concord, Ontario, Canada) interfaced to the column eluent via a Sciex turbospray probe operated at 450 °C. Operating conditions were optimized by infusion of a mixture of all analytes (2.5 µM each) at a flow rate of 5 µL/min, along with the LC flow (500 µL/min, solvent A/B ) 50/50). Selected reaction monitoring experiments in the positive ionization mode were performed using a dwell time of 200 ms per transition to detect ion pairs for each analyte. For metabolite identification, the chromatographic separation of test compounds and their metabolites was performed on a reverse phase C18 column (Synergi MAX-RP, 2.0 mm × 150 mm, 4 µm; Phenomenex, Torrance, CA) using a Rheos 4000 binary pump (LEAP Technologies, Carrboro, NC) with a flow rate of 200 µL/min. The mobile phases described above were used. The initial mobile phase was composed of 15% solvent B, which was increased linearly to 70% over a period of 20 min and held at that value for an additional 3 min. The gradient then was returned to 15% of solvent B in 1 min, and the column was equilibrated for 6 min prior to the next injection. Mass spectrometric analysis was performed on an LCQ ion trap mass spectrometer equipped with an electrospray ionization source (Finnigan MAT, San Jose, CA), as described previously (19). The14C-labeled analogue of I and its metabolites present in liver microsomal preparations, rat hepatotytes, and rat bile were separated on a Luna C18(2) column (4.6 mm × 250 mm, 5 µm, Phenomenex) using a Rheos 4000 binary pump (LEAP Technologies) with a flow rate of 1 mL/min. The aqueous mobile phase (solvent A) consisted of 10% acetonitrile in 0.02% aqueous acetic acid (pH 3.5), while the organic phase (solvent B) consisted of acetonitrile. The initial mobile phase was composed of 20% solvent B, which was maintained for 18 min and then increased linearly to 35% B over a period of 12 min, to 50% B over 10 min, and finally to 75% B over 1 min. This value was held for 5 min before returning to 20% B. The column was equilibrated for 4.5 min prior to the next injection. Postcolumn radiochemical detection and mass spectrometry were performed with a Packard C515 flow scintillation analyzer (PerkinElmer Life Sciences, Meriden, CT) and a Finnigan LCQ ion trap system (Thermo Finnigan MAT, San Jose, CA), respectively. The
Tang et al. postcolumn flow rate was split 1:9, with 100 µL/min being directed to the LCQ and 900 µL/min to the radiochemical director. Radiochemical detection was performed using a liquid flow cell (500 µL) with Packard scintillation cocktail (UltimaFloM) running at 3 mL/min. NMR and LC NMR Analysis. Proton NMR data were collected on a 500 MHz spectrometer (INOVA, Varian Inc., Palo Alto, CA) either on a 3 mm IFC flow probe or 3 mm inverse detection MIDG-3 probe (Varian Inc.). All reported 1H chemical shifts are relative to tetramethylsilane referenced with the solvent CD2HCN multiplet set to 1.99 ppm. The GS2 adduct of 2 was biosynthesized by incubating 2 (25 µM) in human liver microsomes (1 mg/mL) fortified with GSH (10 mM) and an NADPH regenerating system in a total reaction volume of 5 mL for 2 h. The reaction was quenched by adding 5 mL of cold acetonitrile. The supernatant was concentrated to ∼600 µL and filtered though a 0.22 µm amicon filter. A 500 µL aliquot of the filtered sample was then injected onto an HPLC system equipped with a PDA detector (Varian Inc.) with the eluent monitored at 297 nm. An analytical column (Betasil 2 mm × 250 mm, 5 µm, Keystone Scientific) and a flow rate of 0.3 mL/min were employed. Solvent A consisted of 100% D2O with 0.1% TFA-d, and solvent B consisted of 90% CD3CN with 10% D2O. The initial mobile phase was composed of 10% solvent B, which was maintained for 3 min and then increased linearly to 64% B over a period of 27 min, to 90% B over 5 min, and finally returned to 10% B over 0.1 min. Once the apex of the GS2 adduct peak (tR ∼ 24.5 min) was detected, the pump stopped after a precalibrated delay time of 58 s, at the end of which the metabolite band was considered to be parked in the probe flow cell and NMR data collection was initiated after 5 min. A stopped flow LC NMR spectrum of the parent molecule was obtained by injecting ∼50 µg of 2 under the same LC conditions. Covalent Binding of Radioactivity to Human and Rat Liver Microsomes. The 14C-labeled compounds 1-3 (10 µM, 0.2 µCi per incubate) were incubated at 37 °C with human and rat liver microsomes (1 mg/mL) in 100 mM potassium phosphate buffer (pH 7.4) containing 10 mM MgCl2, with or without NADPH (1.0 mM) in a total volume of 0.5 mL. Incubations (0.5 mL total volume) also were performed in the presence of 5 mM GSH to assess the effect of this nucleophile on the covalent binding of radioactivity to microsomal proteins. In addition, incubations were performed in the presence of ketoconazole (1 µM), sulfafenazole (2.5 µM), or quinidine (2.5 µM) to assess the effect of these selective chemical inhibitors on the degree of irreversible binding of drug-related material. Reactions were initiated by the addition of NADPH stock solution (10 mM), and the incubations were carried out for 60 min. At the end of incubation, the reactions were terminated with 1 mL of acetonitrile. Samples were vortex mixed and centrifuged at 10 °C. The resultant pellets were subjected to sequential washings with 2 mL portions of 70% aqueous acetonitrile, 0.6% trichloroacetic acid, ethanol, and a mixture of acetyl acetate, methanol, and trichloroacetic acid 0.6% (1:1:2 by vol). The samples were centrifuged again, and the supernatant was discarded. This step was repeated until the radioactivity in the washings from control samples (without incubation) was close to background (usually 4-5 times) before the next solvent was applied. After the final washing, the protein pellet was dissolved in 0.2 mL of 0.1 N NaOH by overnight incubation at 55 °C with gentle shaking (20 shakes/min). The radioactivity in a 0.125 mL aliquot of the resulting solution was measured by liquid scintillation counting following neutralization with 0.125 mL of 0.1 N HCl. Protein concentrations in the pellet solution were determined using the modified Lowry method with the Pierce protein assay kit. The final amount of radioactivity irreversibly bound to microsomal proteins was expressed as pmol equivalent bound per mg protein.
Bioactivation of 2,3-Diaminopyridines
Chem. Res. Toxicol., Vol. 18, No. 6, 2005 937
Table 1. Irreversible Binding of Radioactivity to Human and Rat Liver Microsomal Proteins upon Incubation with 14C-Labeled Compounds 1-3a covalent binding [pmol equivalent/(mg protein 60 min)]
a
human
rat
compd (10 µM)
no GSH
5 mM GSH
no GSH
5 mM GSH
1 2 3
4036 ( 208 3760 ( 408 2703 ( 48
601 ( 20 458 ( 49 316 ( 57
4780 ( 162 2512 ( 404 3698 ( 202
331 ( 6 357 ( 38 301 ( 15
metabolic rate [pmol/(min mg protein)] human rat 166 60 19
297 104 73
Data are expressed as means ( SD from three determinations. Table 2. Effect of Selective Chemical Inhibitors and Anti-P450 Antibodies on the Irreversible Binding of Radioactivity and Formation of Total GSH Adducts in Human Liver Microsomes upon Incubation with 14C-Labeled 1a antibody or inhibitor ketoconazole (1 µM) sulfafenazole (2.5 µM) quinidine (2.5 µM) anti-P450 3A anti-P450 2C anti-P450 2D6 b
Figure 2. Preincubation-dependent inhibition of P450 3A activity (testosterone 6β-hydroxylase) in human liver microsomal preparations incubated with compounds 2 and 3.
Results Effect on Testosterone 6β-Hydroxylase Activity in Human Liver Microsomes. Several 2,3-DAPcontaining compounds were evaluated for their effect on P450 3A-mediated testosterone 6β-hydroxylase activity in human liver microsomes both with and without preincubation. The results obtained with two representative compounds (2 and 3) revealed an inhibitory effect (Figure 2), whose magnitude was greater when the experiment was performed with preincubation. Although the degree of inhibition varied, similar results were observed with other 2,3-DAP-containing compounds (data not shown). Clearly, compounds from this structural class exhibited a propensity to act as preincubation-dependent inhibitors of P450 3A activity in human liver microsomes. Irreversible Binding to Human and Rat Liver Microsomal Proteins. Three compounds (1-3), whose rates of turnover in liver microsomes spanned a g5-fold
irreversible binding formation of GS1b (% of control) (% of control) 32 ( 5 93 ( 14 102 ( 8
