Abstract
Muraglitazar (Pargluva; Bristol-Myers Squibb), a dual α/γ peroxisome proliferator-activated receptor activator, is under development for treatment of type 2 diabetes. This study describes the biotransformation profile of carbon-14-labeled muraglitazar in plasma, urine, feces, and bile samples from male CD-1 mice [intact and bile duct cannulation (BDC)] after single oral doses of 1 and 40 mg/kg. The major drug-related component circulating in mouse plasma was the parent compound for up to 4 h postdose. Similar to excretion profiles of muraglitazar in humans, monkeys, and rats, urinary excretion was the minor and fecal excretion via the biliary route was the major elimination pathway for muraglitazar in mice. The parent compound was a minor component in urine, bile, and feces, indicating that muraglitazar was extensively metabolized in mice. Major biotransformation pathways of muraglitazar in mice included taurine conjugate formation, acyl glucuronidation, hydroxylation, and O-dealkylation. In addition to those metabolites previously identified in humans, monkeys, and rats (M1–M21), several unique metabolites identified in mice included taurine conjugates (M24, M25, M26a,b,c, and M31), oxazole-ring-opened metabolites (M27 and M28), glutathione conjugates (M29a,b and M30), a dihydroxylated metabolite (M32), hydroxylated metabolites (M33 and M35), and a dehydrogenated metabolite (M34). The taurine conjugate of muraglitazar, M24, was a major metabolite in mice, accounting for 12 to 15% of the total dose in BDC mice or 7 to 12% of the total dose in intact mice. None of these taurine and glutathione conjugates were found in the bile samples of humans, monkeys, or rats.
Muraglitazar (Pargluva, BMS-298585, N-[(4-methoxyphenoxy)-carbonyl]-N-[[4-[2-(5-methyl-2-phenyl-4-oxazolyl)ethoxy] phenyl]-methyl]glycine), a dual α/γ peroxisome proliferator-activated receptor activator, is being developed to treat type 2 diabetes (Devasthale et al., 2005). Muraglitazar was extensively metabolized in humans (Wang et al., 2006; Zhang et al., 2006). From a human absorption, distribution, metabolism, and excretion study with the urine and feces collection, after a single 10-mg oral dose of [14C]muraglitazar, fecal excretion represented >90% of muraglitazar elimination, and only oxidative metabolites were identified in feces of human subjects (Zhang et al., 2006). However, in a disposition study of [14C]muraglitazar in humans with bile collection after a single 20-mg oral dose, biliary excretion represented the major elimination pathway for muraglitazar, and glucuronide conjugates were the major metabolites in the human bile (Wang et al., 2006). Seventeen oxidative metabolites and several glucuronide conjugates of [14C]muraglitazar have been identified in humans and presented previously (Wang et al., 2006; Zhang et al., 2006). These oxidative metabolites included hydroxy muraglitazar (M8a, M10, M11, and M14), O-demethyl muraglitazar (M15), O-demethyl hydroxy muraglitazar (M2, M5, M6, and M7), oxazole-ring-opened metabolites (M9 and M16), O-dealkyl muraglitazar (M1), O-dealkyl O-demethyl muraglitazar (M21), dihydroxy muraglitazar (M3, M4, and M8), and muraglitazar 12-carboxylic acid (M12). For identification of oxidative metabolites, NMR data for M9, M10, M11, M14, and M16 were used, and M5, M10, M11, and M12 were synthesized (Zhang et al., 2006). Glucuronide metabolites included muraglitazar glucuronide (M13), the glucuronide of O-demethyl muraglitazar (M20), glucuronides of hydroxy muraglitazar (M18, three isomers), and glucuronides of hydroxy O-demethyl muraglitazar (M17, three isomers). Glucuronide metabolites were the major metabolites in human bile; however, the aglycones were found in feces, indicating hydrolysis in the intestines before excretion. Urinary elimination was very minor in humans (<5% of dose). Muraglitazar was the major circulating drug-related component in humans. After oral administration of [14C]muraglitazar to monkeys and rats, biliary elimination was also the major clearance pathway (data not shown).
To properly put into context any finding during toxicology or carcinogenicity studies, a thorough understanding of the biotransformation in the toxicology species relative to humans is necessary. This article describes the metabolism of muraglitazar and characterization of unique metabolites, including taurine and glutathione conjugates of [14C]muraglitazar in male CD-1 (BDC and intact) mice.
Materials and Methods
Materials. [14C]Muraglitazar (11.2 μCi/mg, 98.6% radiochemical purity) was synthesized at Bristol-Myers Squibb. The chemical structure and the position of the 14C label are shown in Fig. 1. Trifluoroacetic acid (TFA) and PEG-400 were purchased from Aldrich Chemical Co. (Milwaukee, WI). Ecolite liquid scintillation cocktail was purchased from ICN Biomedicals, Inc. (Costa Mesa, CA). Soluene-350 was purchased from PerkinElmer Life and Analytical Sciences (Boston, MA). All other organic solvents and the reagents were of HPLC grade or reagent grade. HPLC solvent A was 0.06% (v/v) TFA in water, and HPLC solvent B was 0.06% (v/v) TFA in acetonitrile.
The bile samples from rats, monkeys, and humans were used in this study as comparators to mouse bile samples, and they were collected for 0 to 24 h from rats (20 mg/kg dose), 0 to 48 h from monkeys (2 mg/kg dose), and 3 to 8 h from humans (10 mg/subject dose) after single oral administration of [14C]muraglitazar as described previously (Wang et al., 2006).
Animal Studies. The animal studies conformed to accepted criteria for care and use of laboratory animals and complied with all applicable sections of the Final Rules of the Animal Welfare Act regulation (United States Department of Agriculture, Animal Welfare Act, 9 CFR 3, 1991).
The BDC mice were prepared at least 24 h before dosing. Two 12-inch extension tubes (0.02“i.d.) were pulled through a spring tether with the use of a wire hook and filled with 2.5% taurocholic acid solution prepared in 0.9% sterile saline. One tube was marked at both ends and used for bile duct connection. The other tube was used for duodenal end connection, one end of which was clamped off with a hemostat. The unanesthetized mouse was gently restrained while the cannula was externalized. The loop was cut, and the tether end of the bile extension tube was expanded by dipping in xylenes for a few seconds. The bile end of the cannula was inserted into the expanded bile extension tube and the attachment sealed with a cyanoacrylate glue. The duodenum end of the cannula was inserted into the end of the other extension tube that had been expanded in xylenes and the attachment sealed with a cyanoacrylate glue. The hemostat attached to the bile extension tube was removed, and the flow of bile into the extension tube was observed. The extension tubes were pulled through the tether until it rested close to the animal's body. The jacket was put on over the tether and secured with light orthodontic elastics. The extension tubes were then attached to a dual-channel swivel suspended beside the cage. The bile collection tube (0.012” i.d.) attached to the swivel was filled with 2.5% taurocholic acid. A solution of 2.5% taurocholic acid was infused through the swivel and into the duodenum at a rate of 0.4 ml/h as a replacement fluid. The bile samples were collected by gravity flow.
As shown in Table 1, 42 CD-1 male mice were divided into four groups: group 1M (16 mice, intact) and group 3M (five mice, bile duct-cannulated) received a single dose of 1 mg/kg; group 2M (16 mice, intact) and group 4M (five mice, bile duct-cannulated) received a single dose of 40 mg/kg by oral gavage. The animals were fasted for approximately 4 h before dose administration. The dosing solutions of [14C]muraglitazar were prepared in PEG-400 1 day before the experiment and kept at room temperature. The actual amount of dosing solution administered to each animal was determined by weighing the loaded dosing syringe before dose administration and the emptied syringe after dose administration. The average doses for each groups were as following: 1M, 0.97 ± 0.03 mg/kg (10.9 ± 0.4 μCi/kg); 2M, 40.4 ± 2.1 mg/kg (453 ± 24 μCi/kg); 3M, 1.00 ± 0.03 mg/kg (11.2 ± 0.3 μCi/kg); and 4M, 40.6 ± 1.0 mg/kg (455 ± 11 μCi/kg). Four mice at each time point from groups 1M and 2M were euthanized by asphyxiation with air/carbon dioxide gas, followed by exsanguination via the abdominal vein at 1, 4, 12, or 24 h after dosing, and five mice in 3M and 4M groups were euthanized at 24 h after dosing. Blood was collected in tubes containing K3EDTA as the anticoagulant. After centrifugation, plasma was transferred into a labeled container, and glacial acetic acid was added immediately (20 μl/ml of plasma). Bile was collected on dry ice for 0 to 24 h after dosing from groups 3M and 4M. Glacial acetic acid was added to the bile samples (20 μl/ml bile) immediately upon thawing. Urine was collected on dry ice. A pooled 0 to 24-h urine sample was collected from four mice in the intact groups and five mice in the BDC groups. Glacial acetic acid was added to the urine samples (20 μl/ml urine) immediately upon thawing. Fecal samples were collected on dry ice. A pooled 0 to 24-h feces sample was collected from four mice in the intact groups and five mice from the BDC groups. Upon collection, fecal samples were weighed, ethanol/water (50:50, v/v) containing 2% glacial acetic acid (4 ml to 1 g of feces) was added, and then the samples were homogenized by a probe-type homogenizer. Unless otherwise noted, the pooled samples were obtained by combining half of each sample within the groups. All samples were stored at –20°C.
Preparation of Samples for Metabolite Profiling.Bile. A portion (0.5 ml) of the pooled bile sample (0–24 h) from groups 3M and 4M was mixed with 0.5 ml of methanol and then centrifuged for 10 min at 2000g. The supernatant was analyzed directly.
Plasma. Pooled plasma samples (1, 4, 12, and 24 h) were prepared by mixing 0.4 ml of plasma sample from each animal at each time point (groups 1M or 2M). One volume of pooled plasma sample was extracted twice with methanol/acetonitrile (50:50, v/v; 4 volumes each time). After the first addition of organic solvent, the sample was centrifuged at 2000g for 15 min. The supernatant was removed, and the second portion of organic solvent was added to the pellet. The sample was vortexed to resuspend the solid material and centrifuged at 2000g for 15 min. The combined supernatants were evaporated to dryness under a stream of nitrogen. The samples were reconstituted in 500 μl of acetonitrile/HPLC solvent A (20:80, v/v) and centrifuged at 2000g for 15 min before HPLC injection. The radioactivity recovery from the plasma extraction procedure was 95, 94, 72, and 61% for the 1-, 4-, 12-, and 24-h samples, respectively, from group 1M (1 mg/kg dose), and 95, 89, 73, and 59% for the 1-, 4-, 12-, and 24-h samples, respectively, from group 2M (40 mg/kg dose). The low overall level of radioactivity from the 24-h plasma might have contributed to the low extraction recovery at the late time points.
Urine. Before HPLC analysis, the pooled urine sample (0–24 h) from each group was centrifuged for 10 min at 2000g to remove any insoluble material.
Feces. The pooled feces homogenate sample (2 ml) was extracted twice with methanol/acetonitrile (50:50, v/v, 8 ml each time). After the first addition of organic solvent, the sample was centrifuged at 2000g for 15 min. The supernatant was removed, and a second portion of organic solvent was added to the pellet. The sample was vortexed to resuspend the solid material and centrifuged at 2000g for 15 min. The combined supernatants were evaporated to dryness at room temperature under a stream of nitrogen. The sample was reconstituted with 1 ml of HPLC solvent A/solvent B (80:20, v/v) and centrifuged before HPLC injection.
Determination of Radioactivity Recovery. Aliquots of the bile, plasma, or urine (25–100 μl) were mixed with Ecolite liquid scintillation cocktail (approximately 5 ml), and the amount of radioactivity was measured in a Packard Tri-Carb 2250 liquid scintillation counter (PerkinElmer Life and Analytical Sciences). Aliquots of feces homogenate (0.05–0.15 g) were mixed with Soluene-350 in a 25-ml glass scintillation vial and were incubated overnight with gentle shaking in a 60°C water bath. Upon cooling, 0.2 ml of 30% hydrogen peroxide and 15 ml of Ecolite were added. The vial was kept in the liquid scintillation counter for 1 or 2 days before counting.
HPLC/MS/MS Analysis. HPLC analysis was performed on an Agilent 1100 series HPLC system (binary pump, autosampler, and an adjustable wavelength UV detector; Agilent Technologies, Palo Alto, CA). Samples were injected onto a YMC ODS AQ reversed phase column (4.6 × 150 mm, 5 μ; Waters Corporation, Milford, MA) and separated after 75 min at a flow rate of 1.0 ml/min using mobile phases consisting of 0.06% TFA in water (solvent A) and 0.06% TFA in acetonitrile (solvent B). The solvent gradient was as follows: 5 to 25% B in 5 min, 25 to 40% B in 15 min, 40 to 53% B in 40 min, 53 to 60% in 3 min, 60 to 90% in 2 min, 90% B for 7 min, and 90 to 5% B in 3 min. The HPLC eluent went through a splitter so that 25% of the eluent was subjected to MS/MS analysis with a Finnigan LTQ ion trap mass spectrometer (ThermoFinnigan, San Jose, CA) using electrospray ionization (ESI) in the positive ion mode. The remaining 75% of the eluent was collected into LumaPlates-96 (PerkinElmer Life and Analytical Sciences) at 0.26-min intervals using a Gilson FC 204 96-well fraction collector (Gilson Medical Electronics, Middletown, WI) for radioactivity determination.
Accurate Mass Analysis. Pooled bile samples were analyzed by HPLC with Shimadzu LC-10AD binary pumps (Shimadzu, Kyoto, Japan) equipped with a LEAP autosampler (Carrboro, NC) on an YMC ODS AQ column (2.0 × 50 mm, 5 μ). The flow rate was 300 μl/min. Solvent B was maintained at 0% for 0.5 min, then increased to 70% in 15 min, then to 90% in 0.5 min, and then maintained at 90% for 2 min. For accurate LC/MS analysis, the HPLC eluent was directed to a Micromass Q-TOF API-US mass spectrometer that was equipped with a Lock-Spray and an ESI source (Waters, Milford, MA) and analyzed in the positive ion mode. An infused solution of a compound with a formula of C38H53N6O7 was used to provide the Lock Mass (m/z 705.3976). The quadruple time of flight was tuned to 18,000 resolution at half-peak height of the reference compound.
Radiochromatographic Analysis. HPLC fractions collected in 96-well plates (LumaPlate-96) were dried in an Automatic Environmental Speed Vac (Savant, Holbrook, NY). Each well was then counted for radioactivity for 10 min using a Microplate Scintillation and Luminescence TopCount Counter (model NXT; Packard Instrument Company, Meriden, CT). For preparation of biotransformation profiles (radiochromatograms), a background count of 2 to 3 cpm was subtracted from the cpm value of each fraction. The resulting net cpm values were plotted against time-after-injection using Microsoft Excel software. The relative distribution of each metabolite in a biotransformation profile was determined by summing of the radioactivity in the fractions corresponding to the metabolite divided by the total radioactivity collected during the entire HPLC run.
Synthesis of M24. 1-[3-(Dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride (100 mg, 0.5 mmol) was added slowly to an ice-cold solution of muraglitazar (200 mg, 0.387 mmol), taurine (2-amino ethanesulfonic acid, 48.4 mg, 0.387 mmol), 1-hydroxybenzotriazole (67.7 mg, 0.5 mmol), and N-methyl morpholine (0.1 ml, 1 mmol) in dimethyl formide (2 ml) at 0°C. The reaction mixture was stirred at room temperature for 3 h. The reaction mixture was neutralized with aqueous HCl (0.2 ml of a 12 N solution) and was then purified directly by preparative HPLC (continuous gradient from 70% A/30% B to 0% A/100% B [A = 90% H2O/10% MeOH + 0.2% TFA; B = 10% H2O/90% MeOH + 0.2% TFA] at a flow rate of 30 ml/min for 30 min on a Phenomenex LUNA 5 μ C18(2) 30 × 250 mm column (detection at 220 nm; Phenomenex, Torrance, CA) followed by lyophilization from water to give the compound (155 mg, yield 64%). NMR analysis was performed on a Joel 400 MHz spectrometer (Jeol USA Inc., Peabody, MA) in chloroform-d. Proton chemical shifts are reported in ppm relative to tetramethylsilane. MS (+ESI), [M + H]+: 624, MS fragments: m/z 292, 186, 416. 1H NMR (CDCl3; 400 MHz) δ: 2.53 (CH3, s), 2.93 (CH2, bs), 3.19 (CH2, bs), 3.61 (CH2, bs), 3.73 (CH3, s), 3.88–4.10 (CH2, d, J = 28 Hz), 4.29 (CH2, bs), 4.40–4.60 (CH2, d, J = 40 Hz), 6.70–6.79 (2 aromatic-H, d, J = 10 Hz), 6.85 (2 aromatic-H, bs), 6.98–7.05 (2 aromatic-H, d, J = 10 Hz), 7.15–7.25 (2 aromatic-H, d, J = 5 Hz), 7.30–7.50 (N-H, d, J = 28 Hz), 7.55–7.60 (2 aromatic-H, t, J = 10 Hz), 7.64–7.70 (aromatic-H, t, J = 10 Hz), 8.11–8.12 (2 aromatic-H, d, J = 5 Hz), 8.26 (1 acid-H, bs).
Results
Excretion and Recovery of Administered Dose. After a single oral dose of [14C]muraglitazar, the majority of radioactivity was recovered in feces of intact mice (approximately 80% of dose) and bile of BDC mice (>65% of dose) over a 0 to 24-h collection period (Table 1). Urine contained 7 to 16% of the administered dose. In group 1M (1 mg/kg dose, intact mice), 87% of the dose was recovered in excreta in the first 24 h after administration, whereas greater than 90% of the dose was excreted in the same period in the other three groups. Therefore, the fecal route via biliary excretion was the major elimination pathway for [14C]muraglitazar in mice.
Metabolite Profiles. Distribution of [14C]muraglitazar and its metabolites in plasma at 1, 4, 12, and 24 h is presented in Table 2. Distribution of muraglitazar and its metabolites in feces of intact mice and bile of BDC mice as the major elimination route is presented in Table 3. Representative radiochromatograms for plasma, urine, feces, and bile samples from groups 2M or 4M are shown in Fig. 2. The parent compound was the minor component in bile, feces, and urine samples, indicating that muraglitazar was extensively metabolized in mice.
Plasma. The concentrations of total radioactivity in plasma (microgram equivalents of muraglitazar per milliliter of plasma) at 1, 4, 12, and 24 h were 2.6, 1.7, 0.15, and 0.03 for group 1M and 78.6, 79.1, 3.4, and 3.0 for group 2M, respectively. Plasma samples at 1 h were similar to those at 4 h and contained 10- to 25-fold higher total radioactivity than those plasma samples at 12 and 24 h. The total radioactivity in 12- and 24-h samples were comparable. Therefore, similar to that in other species, the total radioactivity appeared to reach the maximum concentration at 1 h, and the exposure to the parent compound was proportional to doses. Metabolite profiles in plasma for both 1 and 40 mg/kg dose groups were similar. The parent compound represented the major radioactive peak in all plasma samples. M18b (glucuronide conjugate of hydroxy muraglitazar) was a trace metabolite at 1 h and a minor metabolite at 4 h. At 12 and 24 h, the level of M18b became comparable with that of the parent drug, and M34, a dehydrogenated metabolite, was evident as a minor metabolite. Three trace metabolites, M11, M14, and M24, were detected only by LC/MS in the 24-h plasma of the 40 mg/kg dose group.
Bile. The major portions of radioactive dose (>65%) were found in the bile at both 1 and 40 mg/kg doses (Table 1), and the parent compound only accounted for <4% of dose (Table 3). The metabolite profiles of bile were similar between the 1 and 40 mg/kg doses. Taurine conjugates M24 and M26(a,b,c), hydroxy metabolite M10, glucuronide conjugate M13, O-dealkylated metabolites M1 and M21, and the parent drug were prominent components in mouse bile. In addition, radioactive peaks corresponding to M4, M7, M9, M11, M12, M14, M15, M16, M18(a,b,c), M25, M27, M28, M29(a,b), M30, M31, M32, M33, and M35 were visible in the HPLC radiochromatogram (Fig. 2). Metabolite M2, M5, M6, M8, M8a, M17, M20, M26(d,e), and M34 were detected by LC/MS/MS in trace amounts in the 40 mg/kg dose group.
Metabolite M24, a taurine conjugate, was the most abundant drug-related species in the bile and accounted for 12.4 and 15.1% of the total dose for the 1 and 40 mg/kg dose groups, respectively. The total amount of taurine conjugates (M24, M25, M26a,b,c, and M31) accounted for 20.8 and 21.6% of the dose at 1 and 40 mg/kg. Three glutathione conjugates, M29a, M29b, and M30, accounted for 3.3, 0.8, and 1.8% of the dose at 1 mg/kg and 1.8, 0.9, and 1.1% of the dose at 40 mg/kg.
Feces. The major portions of the radioactive doses (>80%) were found in feces of intact mice (groups 1M and 2M) and small portions of radioactive doses (<12%) were found in feces of BDC mice (groups 3M and 4M). Radioactivity was distributed among M1, M8a, M9, M10, M11, M12, M14, M15, M16, M21, M24, M26(b,c), M27, M28, M32, M33, and muraglitazar in the fecal samples of intact mice. M1, M9, M10, M24, and muraglitazar were the major components in groups 1M and 2M. In addition, M3, M5, and M7 were detected by LC/MS as trace metabolites in the 40 mg/kg dose group.
M1, M9, M10, M12, M13, M14, M15, M18g, M20b, M21, M27, M28, and muraglitazar were the most prominent components in the fecal samples of BDC mice (groups 3M and 4M). The taurine conjugates M24 and M26(a,b) were not detected in the fecal samples from BDC mice of either dose group. M1 and M21 seemed to be the major metabolites in the feces of BDC mice but the minor metabolites in the feces of intact mice; however, these metabolites represented similar percentage of doses between the BDC and intact mice.
Urine. Urine represented a minor elimination pathway for [14C]muraglitazar. The predominant radioactive metabolites in the urine samples were M1 and M21 for all groups. The parent drug was a trace component in all mouse urine samples. M10, M12, M14, M17b, M18b, and M24 were detected only by LC/MS in the 40 mg/kg dose group.
Identification of Metabolites. More than 30 metabolites were found in mice after oral administration of [14C]muraglitazar. M1–M21 have been previously identified in humans after oral administration of [14C]muraglitazar (Table 4) (Wang et al., 2006; Zhang et al., 2006). These metabolites were assigned in this study by comparing retention time, MS/MS fragmentation, and relative elution order with previously identified metabolites. Multiple isomers (the same molecular weight but different HPLC retention times and/or MS/MS patterns) were detected for acyl glucuronides (M13, M17, M18, and M20), which may result from acyl migration and/or different hydroxylation positions (in the case of M17 and M18). Each isomer of these conjugates with the same molecular weight was assigned a unique metabolite number and letter designation. Multiple acyl migration isomers including M13a–d, M17a–j, M18a–g, and M20a–e (Table 4) were observed in LC/MS/MS analysis of mouse bile in this study. The structures of these acyl migration isomers were not further elucidated. The acyl migration isomers of glucuronide metabolites were not observed in human bile sample (Wang et al., 2006).
Metabolites M24–M35 were not detected in humans, monkeys, or rats after single oral doses of [14C]muraglitazar (data not shown). Their structural assignments in this study were based on the MS/MS fragmentation patterns (Table 4) and comparison to an authentic standard of M24. The MS/MS fragmentation pattern of muraglitazar ([M+H]+ = 517) showed cleavage adjacent to the carbamate nitrogen (C3-N) to give a fragment at m/z 292 (relative intensity 90%) and cleavage at the alkyl ether bond (C8-O) to give a fragment at m/z 186 (relative intensity 100%). The two cleavage points divided the molecule into three moieties: A, B, and C (Fig. 1).
M24. M24 had a molecular ion at m/z 624 (equivalent to parent + 107) and two major MS/MS fragment ions (m/z 292 and 186) similar to those of muraglitazar (Fig. 3), suggesting that modification occurred at the carboxylic acid side of the molecule. Accurate mass analysis indicated that the metabolite had an empirical formula of C31H33N3O9S (Table 5), which suggested that the modification resulted in the addition of a sulfur and a nitrogen atom. M24 was tentatively identified as the acyl taurine conjugate of muraglitazar. Thus, fragment ions at m/z 542 (624–82), m/z 499 (624–125), and m/z 471 (624–153) were proposed to result from the loss of H2SO4, taurine (H2NC2H4SO3H), and HC(O)-taurine [HC(O)NHC2H4SO3H], respectively (Fig. 3). The taurine conjugate standard M24 (Fig. 1) was subsequently prepared through chemical synthesis and subjected to full structural characterization. In addition to the same MS/MS patterns, coinjection of BMS-732106 and the bile sample showed that M24 and BMS-732106 had identical retention times, which confirmed the structural assignment of M24.
M25. M25 had a molecular ion at m/z 610 (equivalent to parent – 14 + 107) and two major MS/MS fragment ions (m/z 292 and 186) (Fig. 3). Accurate mass analysis suggested that this metabolite had an empirical formula of C30H31N3O9S (Table 5), which suggested that the modification resulted in the addition of a sulfur and a nitrogen atom. Thus, the structure of M25 was proposed to be the taurine conjugate of O-demethyl muraglitazar.
M26. Five metabolites with a molecular ion at m/z 640 (equivalent to parent + 16 + 107) were detected in bile and assigned as M26a, M26b, M26c, M26d, and M26e. The structures of the five isomers were proposed to be taurine conjugates of hydroxy muraglitazar. The key MS/MS fragments ions (Table 4 and Fig. 4) for these metabolites were: m/z 308 (292 + 16) and 187 for M26a; m/z 290 (292 + 16 – 18) and 187 for M26b; m/z 290 (29 – 2) and 202 (186 + 16) for M26c; m/z 432, 292, and 186 for M26d; and m/z 202 (186 + 16) and 308 (292 + 16) for M26e. These data suggested that M26a, b, c, and e had the hydroxyl groups on moiety A, whereas M26d had the hydroxyl group on moiety C.
M27 and M28. M28 had a molecular ion at 535 (equivalent to parent + 18) and a corresponding fragment ion of m/z 204 (186 + 18), suggesting the addition of H2O to moiety A. A proposed structure for M28 was shown in Fig. 5. M27 had a molecular ion at m/z 537 (equivalent to parent + 18 + 2) and a fragment ion at m/z 206 (186 + 18 + 2), suggesting the addition of H2O and two hydrogen atoms to moiety A. M27 was proposed to be formed via reduction of M28.
M30. M30 had a molecular ion at m/z 840 (equivalent to parent + 16 + 307), suggesting the addition of an oxygen atom, one hydrogen, and the elements of glutathione (GSH). Accurate mass analysis supported such a structure with a proper empirical formula. The glutathione conjugation structure was supported by two characteristic mass fragment ions at m/z 711 (84 – 129) and 765 (84 – 75), resulting from the loss of γ-glutamyl and glycine moieties of glutathione, respectively (Fig. 6) (Baillie and Davis, 1993; Jin et al., 1994). Fragment ions at m/z 615 (84 – 225), 597 (84 – 1 – 225), and 468 (840 – 1 – 22 – 129) suggested that the modification was not on moiety C (neutral loss of moiety C: 225). The fragmentation pattern (Fig. 6) of the predominant fragment ion m/z 308 suggested that it was a mixture of protonated glutathione ion (GSH+H+, m/z 308) and a daughter ion of m/z 840 after loss of GSH and moiety C (84 – 30 – 225). Thus, the fragment ions of m/z 179 (30 – 129), 162 (30 – 146), and 233 (30 – 75) were proposed to be derived from the protonated glutathione ion, whereas fragment ions of m/z 187 and 265 were proposed to be derived from the fragment ion of m/z 308 (840 – 307 – 225). These data indicated that glutathione was readily removed from the molecular ion of M30, consistent with an attachment of the glutathione to moiety A or B. Thus, M30 was proposed as hydroxy glutathionyl muraglitazar (Fig. 6).
M29a,b. M29a and M29b had molecular ions at m/z 856 (equivalent to parent + 16 + 307 + 16), indicating the addition of 16 (one oxygen atom) to the molecular weight found for M30. M29a and M29b were proposed to be isomers of dihydroxy glutathionyl muraglitazar. Fragment ions of M29a at m/z 709 (856 – 18 – 129), 691 (856 – 18 – 18 – 129), and 745 (856 – 18 – 18 – 75) gave support for the assignment of this glutathione adduct (Fig. 7). The fragment ion of m/z 595 (856 – 18 – 18 – 225) indicated that the glutathione was not on moiety C, whereas the fragment ion of m/z 290 (292 + 16 – 18) suggested that at least one hydroxyl group was on moiety A. The major fragment at m/z 515 (856 – 18 – 18 – 305) indicated that M29a lost the glutathione and the two water molecules simultaneously, which suggested that the glutathione was not on either phenyl rings of the moiety A or B. On the contrary, the fragment ion at m/z 409 may be attributed to an attachment of glutathione to phenylcyanide, implying that glutathione might attach to the phenyl ring of the moiety A. Thus, position of the glutathione group was not further defined.
The structural assignment of M29b as a glutathione adduct was supported by fragment ions at m/z 709, 691, and 763 (856 – 18 – 75) (Fig. 8). Fragment ions of m/z 613 (856 – 18 – 225) and 484 (856 – 18 – 225 – 129) indicated that the glutathione was not on moiety C, whereas the fragment ion of m/z 290 (292 + 16 – 18) suggested that at least one hydroxyl group was on moiety A. Accurate mass analysis of M29b provides additional support to the assignment with a proper empirical formula (Table 5). Therefore, M29a and M29b might be positional isomers of dihydroxy glutathionyl muraglitazar.
M31. M31 had a molecular ion at m/z 656 (equivalent to parent + 32 + 107), consistent with dioxygenation plus taurine formation. Fragment ions at m/z 306 (292 + 32 – 18) and 218 (186 + 32) suggested that both the added oxygen atoms were on moiety A, and at least one oxygen atom was attached to a carbon that yielded the corresponding water loss in the m/z 306 fragment ion. M31 was identified as a taurine conjugate of a dihydroxy muraglitazar derivative with both hydroxyl groups on moiety A (Fig. 5).
M32. M32 had a molecular ion at m/z 549 (517 + 32), consistent with dihydroxylation (Fig. 5). M32 showed fragment ions at m/z 306 (292 + 32 – 18), 218 (186 + 32), and 187. Based on these fragment ions, the addition of both hydroxyl groups was on moiety A.
M33 and M35. These two metabolites had the same molecular ion at m/z 533 (517 + 16), indicating mono-oxidation. Both metabolites showed fragment ions at m/z 308 (292 + 16) and 186, suggesting that the hydroxyl group was not located on moiety A and rather on moiety B (Fig. 5).
M34. M34 had a molecular ion at m/z 515 and a major MS/MS fragment at m/z 290, suggesting that it had lost two hydrogen atoms, and the modification occurred on either moiety A or B. The only plausible site for this transformation was between C-8 and C-9. M34 eluted after muraglitazar in the reversed phase HPLC system, which makes it unlikely that the m/z 515 ion resulted from dehydration of a hydroxy muraglitazar derivative (m/z 533) in the ionization source. Therefore, M34 was identified as an 8-ene muraglitazar derivative (Fig. 5).
LC/MS Profiles of Taurine and Glutathione Conjugates in Bile Samples From Other Species. The finding of taurine and glutathione conjugates in mouse samples prompted investigation of whether these previously unidentified metabolites were present in other species. Bile samples of rats (20 mg/kg dose), monkeys (2 mg/kg dose), and humans (10 mg/subject) after oral administration of [14C]muraglitazar from previous studies were reanalyzed with LC/MS methods used in this study. Figures 9 and 10 show LC/MS ion chromatographic profiles of glutathione conjugates M29(a,b) and M30 in bile samples of mice, rats, monkeys, and humans after oral administration of [14C]muraglitazar. None of the taurine or glutathione conjugates found in this study with mouse samples were detected in bile samples of rats, monkeys, and humans.
Discussion
The metabolism of muraglitazar was determined in intact and BDC male CD-1 mice. The work was done to compare the metabolite profiles in mice to those previously determined in humans (Wang et al., 2006; Zhang et al., 2006). The excretion profile for [14C]muraglitazar was similar between the low- and high-dose groups (Table 1). In addition, the metabolic profiles were similar for both groups in all matrices (bile, feces, plasma, and urine). These data suggested that there were no metabolism or disposition pathways saturated over this dose range.
After oral administration of [14C]muraglitazar, the major drug-related circulating component was the parent compound for up to 4 h postdose at both 1 and 40 mg/kg. Both the parent compound and the ether glucuronide of hydroxy muraglitazar M18b were the major components at 12 and 24 h postdose. O-Dealkylated metabolites M1 and M21 were the major metabolites in the mouse urine; however, urinary excretion was relatively minor. Biliary elimination of metabolites and the parent drug was the major clearance pathway at both low (1 mg/kg) and high (40 mg/kg) doses. This overall elimination profile was similar to what was observed in humans (Wang et al., 2006; Zhang et al., 2006) and monkeys and rats (data not shown). However, the metabolite profiles of bile and feces of mice were quite different from those of other species. M24, the taurine conjugate of muraglitazar, was found to be the most abundant metabolite, accounting for 12 to 15% of the dose in the bile of BDC mice or 7 to 12% of the dose in the feces of intact mice without bile collection. M13, the acyl glucuronide of muraglitazar, the most abundant drug-related species in human bile, was also a major metabolite in mouse bile but was not present in mouse feces. The acyl glucuronide was presumably hydrolyzed in the intestinal tract after biliary excretion. M10, a hydroxylated metabolite, was another major metabolite in mouse bile, although it was a minor metabolite in bile samples from other species.
In addition to those metabolite found in other species, several unique metabolites were found in mice (Fig. 5). Among them, a series of taurine conjugates (M24, M25, M26a,b,c,d,e, and M31) was found that accounted for more than 20% of the total administered dose for both the 1 and 40 mg/kg groups. These metabolites were present at high levels in bile and feces and were also detected in plasma and urine. The formation of taurine or glycine conjugates of bile acids is known to proceed via a two-step enzymatic reaction in liver (Lim and Jordan, 1981). The first step is catalyzed by acyl-CoA synthase, which converts bile acids to their respective acyl-CoA thioester. Bile acid-CoA:amino acid N-acyltransferase then transfers the acyl group to taurine or glycine to form the corresponding conjugates (Falany et al., 1994). The taurine conjugates of muraglitazar and its oxidative metabolites found in this study are presumably formed via acyl-CoA thioester intermediates as demonstrated for the formation of taurine conjugates of other xenobiotic carboxylic acids (Fears, 1985). The large amounts of taurine conjugates of muraglitazar in mice suggested that acyl-CoA synthases and acyltransferases in the mouse liver have the ability to use muraglitazar and its oxidative metabolites as substrates quite efficiently, whereas the taurine-conjugating enzymes in other species are not able to catalyze these reactions. Interestingly, the glycine conjugate of muraglitazar was not detected, which might be due to a change in relative rate of formation of glycine versus taurine conjugates with different acid structures as suggested by Czuba and Vessey (1982).
The glutathione adducts M30 and M29a,b are the second set of related metabolites that are unique to mice. These glutathione adducts were detected at very low levels only in mouse bile from the 40 mg/kg group, and they were not found in other species after oral administration of muraglitazar (Fig. 10). Thus, any possible biological/toxicological effect implicated by formation of these metabolites should also be unique to mice. These glutathione conjugate metabolites showed characteristic GSH conjugate fragmentation patterns upon LC/MS/MS analyses. The glutathione adduct detected with muraglitazar could be explained by an epoxidation of the oxazole ring of muraglitazar followed by a nucleophilic attack on the epoxide by glutathione (Fig. 11).
There are multiple mechanisms that could be postulated for the formation of M28; however, the most straightforward mechanism is through direct enzymatic hydrolysis. Electron-deficient carbon atoms in the oxazole ring can undergo nucleophilic/hydrolysis reactions. The enzymatic hydrolysis of an oxazole-containing compound to generate a ring-opened product has been described previously (Balani et al., 1994). It is plausible that the methyl ketone group in M28 would be reduced by an oxidoreductase to give M27. The carbonyl reduction by a cytosolic keto-aldo reductase provided an example for this conversion of M28 to M27 (Martin et al., 2006). The elution order of the two metabolites in the reversed phase column indicated that M27 was more polar than M28, which was consistent to proposed structures of M27 and M28.
The major site of hydroxylation of muraglitazar in humans occurred on moiety A (Zhang et al., 2006). The fragment ion of m/z 187 with an empirical formula of C12H11O2 was observed previously for M8a, M10, and M14. This fragment was postulated to be formed through a gas-phase replacement of a nitrogen atom with an oxygen atom after ether cleavage of the oxygenated fragment (187 = 186-NH+oxygen) in the mass spectrometer (Zhang et al., 2006). Alternatively, cleavage at the C3-N bond adjacent to the carbamate group and loss of phenyl cyanide from moiety A would also result in a fragment ion at m/z 187. A phenyl cyanide would be formed through a reverse Diels-Alder reaction in the gas phase of the mass spectrometer. The product ion of m/z 187 was characteristic to those metabolites with hydroxylation sites at C8, C9, or C12, such as M10, M14, and M8a, but was not seen in the parent or in other hydroxylated metabolites. Mouse cytochrome P450 enzymes seemed to exhibit different regioselectivity because hydroxylation was also observed on both moiety B (M33 and M35) and moiety C (M26d). There are three possible hydroxylation sites on moiety B: C3, C5, and C6. The low-intensity ratio of fragment ions at m/z 186 to m/z 308 (approximately 1:10) for M35 indicated that carbamate cleavage was much easier than ether cleavage in M35 for this metabolite possibly because of weakening of the C-N bond by resonance effects of a OH group at C3 or C5. On the other hand, M33 showed a “normal” intensity ratio of fragment ions at m/z 186 to m/z 308 (2:1) and suggested that the C-N bond was similar to that in muraglitazar, which would localize the OH group to C6.
The isomerization via acyl migration observed in this study was likely due to the conditions used for the sample collection. Mouse bile was acidified upon thawing by the addition of glacial acetic acid to a final concentration of 2% (v/v) in the present study. The acyl glucuronide of muraglitazar and its metabolites showed evidence of acyl migration to multiple isomers. This behavior was not observed in the previous studies in human bile (Wang et al., 2006). Because those studies were done with the addition of 5% (v/v) glacial acetic acid upon bile sample collection, it is concluded that the present protocol used for the mouse bile did not provide sufficient pH control and that the acyl migration found in this study was due to ex vivo effects. Acyl migration is known to be pH- and temperature-dependent (Sallustio et al., 2000). The results highlight the importance of proper sample collection and storage procedures for prevention of ex vivo acyl migration.
The full characterization of the biotransformation profile of drug candidates is not often conducted in mice. The fact that mouse is often a pharmacological model for many drug candidates and is also a species used for carcinogenicity study provides justification for the determination of the biotransformation profile of a new chemical entity in mice. As we learned from this study, the mouse generated all the metabolites found in the human and rat after administration of [14C]muraglitazar, as well as several additional metabolites not previously identified in the other species.
In summary, muraglitazar was the major drug-related circulating component in mice after oral administration. Fecal excretion via biliary elimination was the major elimination pathway for [14C]muraglitazar. However, the major metabolic pathway of muraglitazar in mice was taurine conjugation. Additional metabolic pathways were similar to those found in humans including glucuronidation, hydroxylation, O-dealkylation, and a combination of these modifications. The taurine conjugate was demonstrated to be a unique metabolic pathway of muraglitazar in mice.
Acknowledgments
We thank Drs. Gary Skiles and Mark Dominick for helpful discussions. We appreciate assistance provided by Drs. Lian Wei and Jiazhong Liu at ABC Laboratories (Columbia, MO) for the BDC mouse study.
Footnotes
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.105.007856.
-
ABBREVIATIONS: BDC, bile duct cannulation; TFA, trifluoroacetic acid; HPLC, high-performance liquid chromatography; MS/MS, tandem mass spectrometry; ESI, electrospray ionization; LC/MS, liquid chromatography/mass spectrometry; GSH, glutathione.
- Received October 13, 2005.
- Accepted February 2, 2006.
- The American Society for Pharmacology and Experimental Therapeutics