Abstract
(3R,4R)-4-Amino-1-((4-((3-methoxyphenyl)amino)pyrrolo[2,1-f] [1,2,4]triazin-5-yl)methyl)-3-piperidinol (BMS-690514) is a potent inhibitor of human epidermal growth factor receptors 1, 2, and 4 and vascular endothelial growth factor receptors 1 through 3. BMS-690514 is an oral oncologic agent currently being developed for the treatment of patients with advanced non–small cell lung cancer and breast cancer. In this investigation, a series of studies was conducted to determine the biotransformation of [14C]BMS-690514 after oral administration to rats, rabbits, and dogs. After administration of a single oral dose of [14C]BMS-690514 to rats and dogs, the majority of the radioactive dose (61–71%) was recovered in the feces, whereas 18 to 20% was eliminated in urine. In bile duct-cannulated rats, 83 and 17% of the administered radioactivity was recovered in the bile and urine, respectively, suggesting that biliary secretion was a major route for the elimination of BMS-690514-derived radioactivity in rats. The parent compound underwent extensive metabolism in both species, with <12% of the administered radioactivity recovered as BMS-690514 in the excreta samples. Metabolite profiles in plasma were qualitatively similar in rats, rabbits, and dogs. Unchanged BMS-690514 was a prominent drug-related component in the plasma profiles from all the species. However, multiple metabolites contributed significantly to the circulating radioactivity, particularly for rabbit and dog, in which metabolites comprised 73 to 93% of the area under the time curve (0–8 h). Circulating metabolites included M6, a direct O-glucuronide conjugate; M1, a hydroxylated metabolite; and glucuronide conjugates of hydroxylated and O-demethylated metabolites. Overall, the results from these studies suggested that BMS-690514 was well absorbed and highly metabolized through multiple pathways in these preclinical species.
The human epidermal growth factor receptor [HER, including epidermal growth factor receptor (EGFR)] and vascular endothelial growth factor receptor (VEGFR) signaling pathways are key components in cancer pathogenesis and progression (Herbst et al., 2005; Tortora et al., 2008). Whereas multiple agents targeting these individual pathways have shown utility in treating advanced non–small cell lung cancer (NSCLC) and other solid tumors (Fukuoka et al., 2003; Kris et al., 2003; Shepherd et al., 2005; Sandler et al., 2006; Reck et al., 2009), recent literature has suggested the potential for enhanced antitumor efficacy through the simultaneous inhibition of both EGFR and VEGFR pathways (Tabernero, 2007; Tortora et al., 2008).
(3R,4R)-4-Amino-1-((4-((3-methoxyphenyl)amino)pyrrolo[2,1-f] [1,2,4]triazin-5-yl)methyl)-3-piperidinol (BMS-690514) is a highly selective and potent inhibitor of multiple receptors in the HER and VEGFR families, including EGFR, HER2, and HER4, and VEGFRs 1 through 3. It is currently under clinical development as an oral treatment for NSCLC, metastatic breast cancer, and other solid malignancies (Bahleda et al., 2009). Preclinical studies have shown that BMS-690514 exerts its antiproliferative and proapoptotic effects on NSCLC cell lines expressing activating mutations such as exon 19 del and L858R, as well as deactivating mutations such as T790M, known to impart resistance to other EGFR tyrosine kinase inhibitors (de La Motte Rouge et al., 2007), and in human xenograft models dependent on EGFR and HER2 signaling (Wong et al., 2007).
The pharmacokinetic properties of BMS-690514 in preclinical studies supported an oral clinical development program in patients with solid tumors (Marathe et al., 2010). BMS-690514 was well absorbed in mouse, rat, and dog, with mean bioavailabilities ranging from 29.4% in dogs to nearly 100% in rats; the mean bioavailability in monkey was only 8.1%, possibly because of high first-pass metabolism. The volume of distribution ranged from 1.9 to 8.7 l/kg in these animal species, indicating distribution outside the vascular compartment. In incubations with hepatocytes from animals and humans, BMS-690514 was metabolized to several oxidative metabolites, including hydroxylated metabolites and an O-demethylated metabolite. In addition, a direct glucuronide conjugate was formed in most animal species and was a prominent metabolite in human hepatocytes (Marathe et al., 2010).
In vivo disposition and metabolism information is critical in all phases of a fully integrated drug development program, and these studies are typically performed with either C-14 or tritium-labeled material to provide detailed quantitative information on the parent drug and its metabolites. In the current studies, the comparative metabolism of BMS-690514 was assessed in rats, rabbits, and dogs after oral administration of [14C]BMS-690514. Comprehensive mass-balance and metabolite profiling of plasma and excreta samples was performed for rat and dog, the primary species used for toxicological evaluation of the compound. In addition, plasma metabolite profiles were obtained from rabbit, a supplemental species used in reproductive toxicology studies. Metabolite profiles from these studies were retained for comparison with clinical samples. Metabolites of BMS-690514 were identified by liquid chromatography/tandem mass spectrometry (LC/MS/MS). Selected metabolites were isolated and further characterized by NMR.
A previous study with bile duct-cannulated (BDC) rats using nonradiolabeled BMS-690514 showed that the parent drug and metabolites were excreted via biliary and renal routes (Marathe et al., 2010), with glucuronide conjugates mainly excreted in the bile. It has been known that glucuronides typically undergo hydrolysis in the gastrointestinal tract by gut flora before excretion in the feces (Parker et al., 1980; Christopher et al., 2008). Therefore, the current studies included administration of [14C]BMS-690514 to BDC rats to understand its overall disposition, particularly for conjugated metabolites excreted in the bile. Based on these in vivo animal metabolism and disposition data, as well as the results from earlier in vitro studies (Marathe et al., 2010), potential human in vivo biotransformation pathways of BMS-690514 were proposed.
Materials and Methods
Chemicals.
[14C]BMS-690514 with a specific activity of 36.8 μCi/mg and radiochemical purity of 98.96% was supplied by the Radiochemistry Group of the Department of Chemical Synthesis, Bristol-Myers Squibb (Princeton, NJ). The C-14 label was evenly distributed among the six carbons of the methoxyaniline ring as indicated in Fig. 1. Nonradiolabeled BMS-690514 and reference standards for M2 (the 3-methoxy, 4-hydroxy aniline metabolite) and M6 (an ether glucuronide of BMS-690514) were also supplied by Bristol-Myers Squibb Research and Development.
All of the chemical reagents were purchased from Sigma-Aldrich (St. Louis, MO) except where indicated. Solvents used for high-performance liquid chromatography (HPLC) analysis were of HPLC grade. Ecolite liquid scintillation mixture was purchased from MP Biomedicals (Solon, OH). Permafluor E+ and Emulsifier-Safe scintillation mixtures were obtained from PerkinElmer Life and Analytical Sciences (Waltham, MA).
Dosing of Animals and Collection of Samples.
All the animal studies were conducted after approval of protocols by the Institutional Animal Care and Use Committee in approved facilities. For each study, a dosing solution was prepared by dissolving appropriate amounts of radiolabeled and nonradiolabeled BMS-690514 in dilute hydrochloric acid (0.01–0.02 N) to achieve the desired specific activity.
Intact Rat Study.
Twenty-one fasted male Sprague-Dawley rats (Charles River Laboratories, Inc., Wilmington, MA), weighing approximately 290 to 320 g, were each administered an oral solution dose of [14C]BMS-690514 at a target dose of 15 mg/kg (100 μCi/kg). The animals were fasted before dosing. The dose solutions were administered by oral gavage in a volume of 5 ml/kg. Terminal blood samples were collected from one group of animals (n = 3 animals/time point) at 1, 2, 4, 8, and 24 h after dosing, and blood samples were centrifuged to obtain plasma for biotransformation analyses. Another group of three animals were individually housed in metabolism cages for the separate collection of urine and feces. Urine was collected at 0 to 12 h, 12 to 24 h, and in 24-h intervals up to 168 h. On collection, urine samples were acidified with concentrated formic acid, such that the final acid concentration was 1% (v/v). Feces were collected in 24-h intervals from 0 to 168 h. The fecal samples were homogenized with 4 volumes of 20% ethanol in water. Cage rinses were performed after each collection interval, and a cage wash was performed at the end of the study period. Excreta and cage residue samples were analyzed for total radioactivity. A third group of animals were received from the supplier instrumented with jugular vein catheters for serial blood collection. Blood samples (0.25 ml/time point) were collected before dosing and at 0.5, 1, 2, 4, 8, 12, and 24 h after dosing for determining the pharmacokinetics of total radioactivity in plasma.
BDC Rat Study.
Three BDC Sprague-Dawley rats, weighing 262 to 283 g, were each administered an oral dose of [14C]BMS-690514 at a target level of 15 mg/kg (100 μCi/kg). The animals were fasted before the dosing, and the dose solutions were administered by oral gavage in a volume of 5 ml/kg. Blood samples, 1 ml each, were collected from each rat at 1 and 4 h postdose, and terminal blood samples were collected from each animal at 24 h for confirmation of exposure. Urine and bile samples for biotransformation profiling were collected over the intervals of 0 to 8 h and 8 to 24 h postdose. On collection, urine samples were acidified with formic acid to a final concentration of 1% (v/v). The fecal samples were collected over the 0- to 24-h time interval and homogenized with 3 volumes of 20% methanol in water.
Intact Rabbit Study.
Three female New Zealand White rabbits, weighing approximately 3.5 to 4.0 kg, were each administered an oral dose of [14C]BMS-690514 at a target dose level of 5 mg/kg (30 μCi/kg). The animals were fasted overnight before dosing, and doses were administered by oral gavage in a dose volume of 2 ml/kg. Blood samples (5 ml/time point) were collected from each animal predose and at 1, 2, 4, 8, and 24 h after dosing for identification of metabolites in plasma.
Intact Dog Study.
Three male beagle dogs, weighing approximately 8.5 to 10.7 kg, were each administered a single oral dose of [14C]BMS-690514 at a target dose of 10 mg/kg (20 μCi/kg). The animals were fasted before dosing. The doses were administered by oral gavage in a volume of 2 ml/kg. Blood samples (2 ml/time point) were collected from each animal predose and at 0.5, 1, 2, 4, 8, 12, and 24 h after dosing for determining the pharmacokinetics in plasma. Additional blood samples for biotransformation analysis were collected from each animal predose and at 1, 2, 4, 8, and 24 h after dosing. All the blood samples were centrifuged to obtain plasma. Urine was collected from 0 to 12 h, 12 to 24 h, and in 24-h intervals thereafter, up to 168 h; the samples were acidified with formic acid to a final concentration of 1% (v/v). Feces were collected from 0 to 168 h in 24-h intervals. The fecal samples were homogenized with 4 volumes of 20% ethanol in water. Cage debris was collected after each daily postdose excreta collection; a cage wash was performed at the end of the study period. Excreta and cage residue samples were analyzed for total radioactivity.
Analysis of Radioactivity.
Portions of plasma, bile, and urine (50–100 μl) and cage wash/rinse samples (300-μl volumes) were added to 5 to 15 ml of Ecolite (MP Biomedicals) or Emulsifier-Safe scintillation mixture (PerkinElmer Life and Analytical Sciences) and analyzed directly by liquid scintillation counting. Portions of fecal homogenate from the BDC rat study (approximately 0.2 g each) were incubated overnight at 60°C with Soluene-350 (PerkinElmer Life and Analytical Sciences) to solubilize the samples. After cooling, 0.2 ml of 30% hydrogen peroxide was added to each sample. The samples were mixed with scintillation mixture (Ecolite) and counted on a Packard Tricarb 3200CA liquid scintillation counter (PerkinElmer Life and Analytical Sciences). In other studies, aliquots of fecal homogenates, 0.2 to 0.5 g, were air-dried and combusted with a model 307 or model 387 Sample Oxidizer (PerkinElmer Life and Analytical Sciences). The liberated 14CO2 was trapped into Carbo-Sorb E (PerkinElmer Life and Analytical Sciences), mixed with Permafluor E scintillation fluid, and analyzed by an LS6500 liquid scintillation counter (Beckman Coulter, Inc., Brea, CA). The efficiency of the sample oxidizer was determined before combusting experimental samples. Three levels of radioactivity corresponding to low, medium, and high [14C]standards were placed into cones and pads and combusted. The combusted standards were analyzed by liquid scintillation counting along with equivalent levels of the [14C]standards spiked into scintillation fluid. The average disintegrations per minute in combusted samples were compared with the spiked samples to determine the combustion efficiency. Combustion efficiency for fecal samples was considered acceptable if the recovery of radioactivity was >90%. In all cases, scintillation data were corrected for counting efficiency using appropriately prepared quench curves. The radioactivity in plasma, urine, bile, and feces was reported as nanogram-equivalents of BMS-690514 per milliliter or per gram or as a percentage of the administered dose.
Preparation of Biological Samples for Analysis.
Representative pooled urine samples from intact rats (5% pool, 0–168 h), BDC rats (2% pool, 0–24 h), and dogs (5% pool, 0–168 h) were prepared by combining a constant percentage of the volume excreted from each animal during each collection interval. Representative pooled fecal homogenate samples from intact rats (2% pool, 0–168 h) and dogs (2% pool, 0–168 h) were prepared by combining an approximately constant percentage of weight excreted from each animal during each collection interval. Representative pooled bile samples from BDC rats (2% pool, 0–24 h) were prepared by combining a constant percentage of the bile volume excreted from each animal during each collection interval. Plasma samples were segregated by collection time, and equal volumes from each animal were combined.
Pooled plasma and fecal samples were extracted with methanol/acetonitrile (50:50, v/v) at a ratio of 3 volumes of solvent per 1 volume of sample. The mixtures were vortex-mixed, sonicated for 5 min, and centrifuged (centrifuge 5810R; Eppendorf AG, Hamburg, Germany) at 2500g for 20 min. The resulting supernatants were transferred into new tubes, and the precipitated pellets were extracted two additional times with 2 volumes of methanol/acetonitrile/water (25:25:50, v/v/v). The total extraction recovery for all the samples ranged from 85 to 99%. The replicate extracts for a given sample were combined and evaporated to dryness under nitrogen in a TurboVap Evaporator (Zymark, Inc., Hopkinton, MA). The resulting residues were reconstituted in methanol/acetonitrile/water (25:25:50, v/v/v). Pooled urine and bile samples were centrifuged at 9000g for 10 min before biotransformation profiling and mass spectrometric analysis.
Radiochromatographic Analysis of Metabolites.
Biotransformation profiling was performed on a Surveyor HPLC system equipped with an autosampler and photodiode-array detector (Thermo Fisher Scientific, Waltham, MA). Sample components were separated on a Waters (Milford, MA) YMC octadecylsilane-AQ HPLC column (4.6 × 150 mm, 120 Å, 5-μm particle size) maintained at ambient temperature. The mobile phase consisted of two solvents: mobile phase A [95% water and 5% acetonitrile (v/v) containing 10 mM ammonium acetate, pH 5.0] and mobile phase B (100% acetonitrile). The linear gradient program used for HPLC separation was as follows: hold isocratic at 0% B (0–5 min), linear gradient from 0 to 28% B (5–50 min), linear gradient from 28 to 90% B (50–55 min), and hold isocratic at 90% B (55–60 min). The mobile phase flow rate was 1.0 ml/min, and the column eluate was monitored at a wavelength of 307 nm. An accurate postcolumn splitter (LC Packings, Inc., San Francisco, CA) was installed after the photodiode array detector, diverting 25% of the eluate to a Thermo Fisher Scientific LCQ Deca XP-Plus ion trap mass spectrometer and the remaining 75% of the eluate to a fraction collector (Gilson, Inc., Middleton, WI), where the flow was collected in 96-well Luma plates (PerkinElmer Life and Analytical Sciences) with a collection rate of 0.25 min/well. The plates were evaporated to dryness on a Savant Speed-Vac (Thermo Fisher Scientific) and counted for radioactivity (10 min/well) with a TopCount Microplate Scintillation Counter (PerkinElmer Life and Analytical Sciences).
The LCQ Deca XP-Plus mass spectrometer (Thermo Fisher Scientific) was equipped with an electrospray ionization source operated in positive-ion mode with a capillary temperature of 350°C and spray voltage of 5 kV. The nitrogen gas flow rate, capillary voltage, and the tube lens voltages were periodically adjusted to give maximal sensitivity or fragmentation of drug-related components. Product ions were generated via collision-induced dissociation with helium using normalized collision energy of 30% and a precursor ion isolation width of m/z 2.0.
For exact mass measurement, 25% of the column eluate was diverted to an LTQ Orbitrap Discovery (Thermo Fisher Scientific), which was equipped with an electrospray ionization source operated in positive-ion mode. The capillary temperature was 350°C, and the spray voltage was 4 kV. The resolving power for full-scan event and MS/MS spectrometric events was set at 30,000 and 15,000, respectively. Before analysis of study samples, the instrument was calibrated with a calibration solution; the instrument was deemed acceptable for use if the mass accuracy for all the components in the calibration solution was within 5 ppm of their expected values.
Pharmacokinetic Analyses.
For each species, the concentration of total drug-related radioactivity in plasma samples (ng-Eq/ml) was determined from the mean radioactivity of n = 3 animals/time point and the administered specific activity. The concentration of parent compound was determined by multiplying the percentage contribution of the BMS-690514 in the pooled radioprofile by the total radioactivity concentration for each time point. The area under the time curve (AUC) (0–8 h) values for total radioactivity and BMS-690514 were determined using Microsoft (Redmond, WA) Excel with the trapezoidal method. The concentrations at 1, 2, 4, and 8 h were used in the calculation of AUC (0–8 h) for total radioactivity and BMS-690514 in rats, rabbits, and dogs.
Isolation of Metabolites M6, M7, M8, and M24.
Metabolites M6, M7, M8, and M24 were isolated from rat bile collected after oral administration of [14C]BMS-690514. Before HPLC separation, the pooled bile sample was concentrated on a Waters Sep-Pak Vac 35 cc (10 g) C18 solid-phase extraction cartridge preconditioned with methanol, acetonitrile, water, and then 50 μM ammonium acetate, pH 5. The bile sample (diluted approximately 5-fold in 50 μM ammonium acetate) was applied to the cartridge. After loading the sample, the cartridge was washed with water, and metabolites were then eluted with acetonitrile/water (50:50, v/v). The acetonitrile/water eluate was evaporated to dryness and reconstituted in 10 mM ammonium acetate, pH 5.
M6, M7, M8, and M24 were isolated from the concentrated sample using a Waters 2695 HPLC equipped with a photodiode-array detector and a YMC octadecylsilane-AQ 4.6 × 150-mm column (Waters). The mobile phase consisted of two solvents: mobile phase A [95% water and 5% acetonitrile (v/v) containing 10 mM ammonium acetate, pH 5.0] and mobile phase B (100% acetonitrile). The linear gradient program used for HPLC separation was as follows: hold isocratic at 0% B (0–5 min), linear gradient from 0 to 25% B (5–45 min), hold isocratic at 25% B (45–50 min), linear gradient from 25 to 90% B (50–55 min), and hold isocratic at 90% B (55–60 min).
Treatment of M24 with β-Glucuronidase.
Isolated M24 was treated with β-glucuronidase at an enzyme concentration of 800 U/ml in 40 mM sodium acetate buffer, pH 5.1. The reaction was conducted at 37°C in a shaking water bath for 90 min. At the end of reaction, an equal volume of ice-cold acetonitrile was added to quench the reaction. The samples were vortex-mixed and centrifuged at 13,000 rpm for 10 min, and the supernatant was analyzed by LC/MS/MS on the LTQ-Orbitrap system (Thermo Fisher Scientific) described above.
NMR Analysis.
BMS-690514 and isolated metabolites M6, M7, and M8 (approximately 200 μg each) were each dissolved in D2O/acetonitrile-d3 (50:50, v/v) and transferred to 3-mm NMR tubes. The samples were analyzed using Bruker Advance 600- and 700-MHz NMR spectrometers equipped with a 5-mm cryo triple resonance probe (Bruker Daltonics, Billerica, MA) at 30°C. All the 1H chemical shifts were reported on the δ scale (ppm) downfield from tetramethylsilane. 13C chemical shift data were deduced from heteronuclear multiple-bond correlation (HMBC) spectroscopy. The peak assignment was based on 1H,1H-1H-gradient correlation spectroscopy and 1H-13C-HMBC spectroscopy analysis. NMR prediction software was obtained from Advanced Chemistry Development, Inc. (Toronto, ON, Canada).
Results
Excretion of the Radioactive Dose.
The recovery of radioactivity in the urine and feces of intact rats and dogs and in urine, bile, and feces from BDC rats after administration of a single oral dose of [14C]BMS-690514 is summarized in Table 1. In intact rats and dogs, the majority of the radioactive dose was recovered in the feces. The fecal excretion of radioactivity in intact rats and dogs in the 0- to 168-h interval postdose was 71.2 and 60.8% of the dose, respectively. For both species, a significant portion of radioactivity was also eliminated in the urine. Urinary excretion of the dose was 18.4 and 20.1% in intact rats and dogs, respectively. The total recovery of the radioactive dose in these two studies was >89%. After oral administration to BDC rats, 83.4% of the administered dose was recovered in the bile, 17.1% in the urine, and 5.8% in the feces over the 24-h interval postdose.
Plasma Concentration Profiles of Radioactivity and BMS-690514.
The plasma concentration versus time pharmacokinetic profiles for total radioactivity and BMS-690514 after administration of single oral dose of [14C]BMS-690514 to intact rats, rabbits, and dogs are shown in Fig. 2, A through C. For all three species, the plasma concentration of total radioactivity was significantly higher than the plasma concentration of BMS-690514 at all the time points up to 8 h. The AUC (0–8 h) values for BMS-690514 and total radioactivity were 5420 and 9456 ng-h/ml in rats, 709 and 10,167 ng-h/ml in rabbits, and 3340 and 12,389 ng-h/ml in dogs, respectively. Based on these exposure data, BMS-690514 comprised approximately 57, 7, and 27% of the AUC (0–8 h) of total radioactivity in rats, rabbits, and dogs, respectively (Fig. 2).
Metabolite Profiles in Plasma.
Representative biotransformation profiles of BMS-690514 in plasma from rats, rabbits, and dogs at 2 h postdose are shown in Fig. 3, A through C, and the distribution of drug-related components in rat, rabbit, and dog plasma is summarized in Table 2. Plasma biotransformation profiles were qualitatively similar between the species. Parent drug BMS-690514 (P), a direct glucuronide conjugate (M6), a hydroxylated rearrangement product (M1), and glucuronide conjugates of oxidative metabolites (M7, M8, and M9) were prominent drug-related components in the plasma from rats, rabbits, and dogs.
Metabolite Profiles in Urine.
Representative radioprofiles of pooled urine samples collected in the 0 to 168 h after oral administration of [14C]BMS-690514 to rats and dogs are shown in Fig. 4, A and D, respectively. The urinary biotransformation profiles were qualitatively similar in rats and dogs. In total, BMS-690514 and 10 drug-related products were identified in the urine profiles in these two species (Table 3). Unchanged parent drug (P), oxidative metabolites M2, M3, and M37, and glucuronide conjugates M6, M7, M8, and M9 were the prominent drug-related components in the urine from both species. Approximately 5.0 and 1.0% of the administered radioactivity was recovered as parent compound in the urine of rats and dogs, respectively. Several early eluting peaks (eluting before 10 min) present in the profiles from the two species could not be identified because of high mass spectrometric interference. These early eluting peaks accounted for 5.9 and 10.7% of the radioactivity in pooled rat and dog urine (0–168 h), respectively.
Metabolite Profiles in Feces.
Representative radioprofiles of pooled fecal samples collected in the 0 to 168 h after oral administration of [14C]BMS-690514 to rats and dogs are shown in Fig. 4, B and E, and the distribution of metabolites is shown in Table 3. BMS-690514 (P) and metabolites M2, M3, and M37 were prominent peaks in each of the fecal profiles. The parent compound accounted for 6.8 and 2.3% of the administered dose in the feces of rats and dogs, respectively. Metabolites M2 and M37 together accounted for 6.9 and 15.6% of the dose in rat and dog feces, respectively, whereas M3 accounted for <8.2 and 16.3% of the dose in these matrices. M6 accounted for 4.8% of the dose in fecal samples from dogs but was not identified in the feces from rats. Other prominent biliary glucuronide-conjugated metabolites (M7, M8, and M9) were not detected in the fecal samples from either species.
An additional oxygenated metabolite, M26, was detected in the fecal profile from rats. M26 coeluted with M3 in the rat fecal profile, and together these two metabolites accounted for 8.2% of the administered dose in rat feces. Several early eluting peaks, before 8 min and between 15 and 20 min, in the rat fecal metabolite profile could not be identified because of high mass spectrometric background interference. These unknown peaks comprised 35.6% of the radioactivity in the pooled rat feces (0–168 h). Similar peaks were not observed in the fecal profiles from dog.
Metabolite Profiles in Bile.
The parent compound and 13 metabolites were found in the bile from BDC rats (Table 3; Fig. 4C). Unchanged parent drug accounted for 5.2% of the dose secreted in the bile of BDC rats. M6, a direct glucuronide conjugate of BMS-690514, was a prominent metabolite in rat bile, accounting for 7.2% of the administered dose. Other prominent biliary metabolites found in rats were M7, M8, and M9, glucuronide conjugates of oxidative metabolites. It is interesting to note that the early eluting unknown metabolite peaks observed in rat feces were not observed in rat bile.
Identification of Metabolites.
Metabolites of BMS-690514 were characterized by LC/MS/MS. Similarities in HPLC retention times and mass spectrometric fragmentation patterns to available reference standards facilitated the identification of metabolites M2, M3, and M6. Structures of other metabolites were proposed based on their mass spectrometric fragmentation patterns relative to BMS-690514 or these known metabolites. In addition, the structural assignments of metabolites were further confirmed by high-resolution mass spectrometry (Table 4). A list of the metabolites of BMS-690514 found in the in vivo samples from rats, rabbits, and dogs is compiled in Table 4, along with the mass spectrometric fragmentation data supporting the structural characterization of each of the metabolites. Proposed structures and metabolic pathways for the formation of the detected metabolites are shown in Fig. 5. Hydroxylated metabolite M1 and the glucuronides M6, M7, and M8 were isolated from in vitro incubations or from rat bile for further characterization by NMR.
The BMS-690514 reference standard had a protonated molecule [M+H]+ of m/z 369 and a major MS2 fragment at m/z 253 resulting from cleavage of the carbon-nitrogen bond between the nitrogen of the aminopiperidine ring and the methyl carbon of the methyl pyrrolotriazine group (Table 4). The metabolites of BMS-690514 generally underwent similar fragmentations, enabling structure identification of metabolites by comparison of changes in the masses of fragment ions relative to the parent.
Metabolite M1 had a protonated molecule [M+H]+ of m/z 385 (16 Da higher than BMS-690514), and the exact mass measurement suggested that it was a monooxygenated metabolite. The MS2 spectrum showed major fragment ions of m/z 367 (loss of water) and m/z 253 (Table 4). M1 was isolated from in vitro incubations with human liver microsomes, and its identity was confirmed by NMR, a hydrogen/deuterium exchange experiment, and high-resolution mass spectrometric analysis. The results from these experiments suggested that M1 was a metabolite that had undergone monohydroxylation on the pyrrolotriazine ring and then underwent an intramolecular rearrangement in which the pyrrolotriazine moiety of the parent had rearranged into a pyridotriazine for M1. Detailed experiments conducted to elucidate the structure of M1 will be described in a separate manuscript.
Metabolite M2 is a monohydroxylated metabolite of BMS-690514, with the site of hydroxylation on the 3-methoxyaniline group. The reference standard for metabolite M2 had a protonated molecule [M+H]+ of m/z 385 (16 Da higher than BMS-690514) and a major MS2 fragment at m/z 269 resulting from cleavage of the nitrogen-carbon bond between the nitrogen of the aminopiperidine ring and the methyl carbon of the methyl pyrrolotriazine group (Table 4). Comparison of the mass spectrometric fragmentation pattern of M2 in biological samples with the authentic M2 reference standard provided additional verification for the identity of M2 in plasma, urine, fecal, and bile samples.
Metabolite M3 had a protonated molecule [M+H]+ of m/z 355 (14 Da lower than BMS-690514), and the exact mass measurement indicated that it was a demethylated metabolite. The MS2 spectrum had a major fragment at m/z 239, which was 14 Da lower than the corresponding fragment ion of the parent BMS-690514 molecule (m/z 253) (Table 4). These data were consistent with a metabolite that had undergone O-demethylation of the methyl group on the methoxyaniline moiety.
Metabolite M6 had a protonated molecule [M+H]+ of m/z 545 (176 Da higher than BMS-690514), suggesting that it was a direct glucuronide conjugate of the parent molecule. The MS2 spectra of M6 showed a minor fragment ion at m/z 369 (neutral loss of 176 Da, indicative of a glucuronide) and a major fragment ion at m/z 253 (Table 4). M6 was isolated from rat bile, and the structure was identified by NMR. One- and two-dimensional NMR results showed that chemical shift differences were observed for some of the 1H and 13C atoms of the aminopiperidinal moiety in M6 compared with the parent compound (Fig. 6). In addition, a key long-range 1H-13C correlation between H3 and C29 of the glucuronide was observed and confirmed the attachment site of the glucuronide at position 17 of the parent (Fig. 6). The results from these experiments localized the site of glucuronidation on the hydroxyl group of the aminopiperidine moiety.
Metabolites M7, M8, and M27 each had a protonated molecule [M+H]+ of m/z 561 (192 Da higher than BMS-690514) and major product ions at m/z 269 and m/z 385 (neutral loss of 176 Da) (Table 4). The mass spectra indicated that these three metabolites are glucuronide conjugates of monohydroxylated BMS-690514. The m/z 269 fragment was similar to that observed for M2 and suggested that the site of hydroxylation for these metabolites was on the pyrrolotriazine or aniline groups.
M7 and M8 were isolated from rat bile and were further characterized by NMR. Comparison of the chemical shifts of M7 with BMS-690514 revealed several 1H and 13C chemical shift differences, e.g., proton 15 exhibited an upfield shift of 0.32 ppm compared with parent, and carbon 15 shifted upfield by 15.7 ppm. More importantly, proton 16 was absent in M7, and carbon 16 has shifted downfield by 21.8 ppm, consistent with the site of hydroxylation on carbon 16 (Fig. 6). Key long-range 1H-13C correlations were observed between H29 and C16, as well as between H15 and C16 (Fig. 6). These correlations and chemical shift changes are consistent with attachment of an ether glucuronide on the C16 of the pyrrolotriazine moiety.
NMR analysis suggested that the isolated M8 peak from rat bile contained two metabolites, M8a and M8b. The two metabolites in the mixture were characterized using one- and two-dimensional NMR methods. Based on proton integration in the 1H spectrum, M8a was a major component, and M8b was a minor component. A proton from the methoxyaniline moiety was absent in the spectra of both metabolites; for M8a, H23 was absent, and in metabolite M8b, H25 was absent. Chemical shift differences were also observed for the 1H and 13C atoms of the methoxyaniline moiety. Note that carbon 23 shifted downfield by 35.9 ppm in M8a, and carbon 25 shifted downfield by 28.2 ppm in M8b, suggesting that these carbons are the sites of hydroxylation for M8a and M8b, respectively (Fig. 6). Key long-range 1H-13C correlations between H29 and C23 in M8a and between H29′ and C25′ in M8b confirm the site of glucuronidation in both metabolites (Fig. 6). For M8a, glucuronidation occurred at the hydroxyl in position 23; for M8b, glucuronidation occurred at the hydroxyl on position 25 (Fig. 6).
Metabolites M9 and M28 each had a protonated molecule [M+H]+ of m/z 531 (176 Da higher than M3). The MS2 spectrum of M9 showed major fragment ions at m/z 415, 355 (neutral loss of 176), and 239 (Table 4). The fragment at m/z 415 resulted from cleavage of the carbon-nitrogen bond between the nitrogen of the aminopiperidine ring and the methyl carbon of the methyl pyrrolotriazine group. The fragment at m/z 239 resulted from cleavage of this same bond after loss of the glucuronide moiety. These data suggested M9 is a glucuronide conjugate of M3, with the site of glucuronidation on the pyrrolotriazine or aniline groups. The MS2 spectrum of M28 showed fragment ions at m/z 355 (neutral loss of 176), 293, and 239 (Table 4). The fragments at m/z 293 and 239 resulted from cleavage of the carbon-nitrogen bond between the nitrogen of the aminopiperidine ring and the methyl carbon of the methyl pyrrolotriazine group. In particular, the fragment at m/z 293 corresponded with the glucuronidated aminopiperidine moiety. These data suggested that M28 is a glucuronide conjugate of M3, with the site of glucuronidation on the hydroxyaminopiperidine moiety (Fig. 5). The exact mass measurement further confirmed that M9 and M28 were glucuronides of demethylated metabolites.
M14 and M36 both had a protonated molecule [M+H]+ of m/z 547, which was 192 Da higher than M3 and MS2 fragment ions at m/z 431, 371 (loss of 176 Da), and 255 (Table 4). The fragment at m/z 431 resulted from cleavage of the carbon-nitrogen bond between the nitrogen of the aminopiperidine ring and the methyl carbon of the methyl pyrrolotriazine group. The fragment at m/z 255 resulted from cleavage of this same bond after loss of the glucuronide moiety, and it was 16 Da higher than m/z 239, which was the corresponding fragment of M3. These data suggest that M14 and M36 are glucuronide conjugates of monooxygenated M3, and these proposed structures were supported by the results of exact mass measurement.
M24 had a protonated molecule [M+H]+ of m/z 589, which was 220 Da higher than BMS-690514. The MS2 spectrum of M24 showed major fragment ions at m/z 413 (neutral loss of 176) and 253 (Table 4). The fragment at m/z 413 (44 Da higher than the [M+H]+ of the parent molecule) and the fragment at m/z 253 (same fragment as the parent molecule) suggested that carboxylation had occurred on the aminopiperidine moiety. After treatment with β-glucuronidase, M24 was completely hydrolyzed and converted to BMS-690514 (Fig. 7, A and B), which confirmed that M24 was a glucuronide metabolite. Although the carbamic acid intermediate was not observed in the incubation mixture by LC/MS/MS because of the instability of this species (Shaffer et al., 2005; Gunduz et al., 2010), the fragmentation of M24 has the signature fragmentation pattern of N-carbamoyl glucuronides, which produced the product ion of m/z 413, 44 Da higher than the parent drug (Dow et al., 1994; Liu et al., 2001; Shaffer et al., 2005; Gunduz et al., 2010). The exact mass measurement confirmed that M24 is a carbamoyl glucuronide of BMS-690514 (Fig. 7, C and D).
M26 had a protonated molecule [M+H]+ of m/z 385 (16 Da higher than BMS-690514), and the exact mass measurement suggested that it was a monooxygenated metabolite. The major MS2 fragment at m/z 269 further suggested that the addition of 16 Da had occurred on the pyrrolotriazine or aniline groups (Table 4).
M30 had a protonated molecule [M+H]+ of m/z 367 (2 Da lower than BMS-690514). The exact mass measurement suggested a formula with the loss of two hydrogen atoms from the parent. A major MS2 fragment ion at m/z 269, which was 16 Da higher than the corresponding fragment ion from BMS-690514 (m/z 253), suggested that an addition of 16 Da had occurred on the pyrrolotriazine or aniline groups (Table 4). A structure consistent with these data is one in which a monooxygenation (M + 16) occurred on the pyrrolotriazine or aniline groups and a dehydration (M-18) occurred on the aminopiperidine moiety.
M32 had a protonated molecule [M+H]+ of m/z 371 (16 Da higher than M3) and a major MS2 fragment ion at m/z 255 (Table 4). The fragment at m/z 255 was 16 Da higher than the corresponding fragment ion of M3, suggesting that the site of modification was on the pyrrolotriazine or aniline groups. These data are consistent with a metabolite that had undergone O-demethylation (M − 14) and monooxygenation (M + 16) on the pyrrolotriazine or aniline groups. This purposed structure has a consistent molecular formula with the one suggested by the exact mass measurement.
M35 had a protonated molecule [M+H]+ of m/z 383 (2 Da less than M2) and a major MS2 fragment ion at m/z 269 (Table 4). The result of exact mass measurement suggested an addition of oxygen atom and loss of two hydrogen atoms from the parent molecule. The fragment at m/z 269 was the same as the corresponding fragment ion of M2, suggesting an addition of oxygen occurred on the pyrrolotriazine or aniline groups. A structure consistent with these data is one in which an M + 16 oxidation occurred on the pyrrolotriazine or aniline groups and a dehydrogenation (M − 2) occurred on the aminopiperidine moiety.
M37 had a protonated molecule [M+H]+ of m/z 385 (16 Da higher than BMS-690514) and a major MS2 fragment at m/z 269, suggesting that the addition of 16 Da had occurred on the pyrrolotriazine or aniline groups (Table 4). M37 and M2 coeluted on the HPLC and had similar MS and MS2 spectra. Incubation of human liver microsomes with deuterated BMS-690514 (labeled on the 16 position) generated unlabeled M37, suggesting that M37 was hydroxylated at the 16 position and the deuterium label at this position was lost on the formation of M37. The experimental details conducted to elucidate the structure of M37 will be described in a separate manuscript.
Discussion
The objectives of the current studies were to evaluate the metabolism and disposition of [14C]BMS-690514 in rat and dog, two species used in long-term toxicology studies. In addition, circulating levels of parent compound and metabolites were quantified in these species and in rabbit, an additional investigative species.
Earlier studies showed that formation of M6, a glucuronide conjugate, was an important pathway for biotransformation of BMS-690514 in hepatocyte preparations from various species (Marathe et al., 2010). Therefore, besides intact rat and dog studies, this study included an assessment of bile from BDC rats after administration of [14C]BMS-690514 to understand the role of conjugative metabolism in the disposition of BMS-690514 and the possible contribution of biliary secretion of parent drug.
After administration of a single oral dose of [14C]BMS-690514, the recovery of radioactivity was 95.4 and 89.6% in intact rats and dogs, respectively, indicating good mass balance (Table 1). In both species, the majority of the administered radioactivity (>60%) was recovered in the feces. After oral administration of [14C]BMS-690514 to BDC rats, approximately 83 and 17% of the dose was recovered in bile and urine, respectively, suggesting that BMS-690514 was well absorbed in rats. The parent compound comprised <12% of the radioactive dose recovered in the excreta from intact and BDC rats and dogs, suggesting that BMS-690514 was primarily cleared by metabolism in these species.
Pharmacokinetic profiles of total radioactivity and BMS-690514 from rat, rabbit, and dog plasma (Fig. 2, A–C) suggested that metabolites contributed substantially to the AUC of total radioactivity over the 0- to 8-h interval. Consistently, the metabolite profiles from the 2-h plasma samples (Fig. 2, A–C; Table 2) confirmed that for all the species, particularly for rabbit and dog, multiple metabolites contributed a large portion of the plasma radioactivity at that time point. The metabolite profiles in plasma from rats, rabbits, and dogs were qualitatively similar (Fig. 3, A–C). In all the species, BMS-690514 and multiple oxidative and conjugative metabolites were detected in the plasma; M1, a hydroxylated metabolite; M6, a direct glucuronide conjugate; and M7, M8, and M9, glucuronide conjugates of oxidative metabolites, were prominent circulating metabolites in all the species. The parent compound was the predominant peak in rat plasma, and rat tended to have a greater abundance of oxidative metabolites (M1, M3, and M26) in the circulation than other species. In rabbits and dogs, glucuronides M6 through M9 were the most prominent plasma metabolites. In the plasma metabolite profiles from intact rat, there was a prominent but unidentified component that eluted at approximately 26 min (unknown, Fig. 3A). This peak was present in the 1- to 8-h plasma samples from intact rats, but it was not present in the plasma profiles from BDC rats (data not shown), rabbits, or dogs, nor was it present in the bile, urine, or fecal profiles from rats, rabbits, or dogs. The mass spectrum of the unknown peak did not show a distinct protonated molecule. In addition, no drug-related component was detected under the unknown peak when the rat plasma sample was analyzed on an LTQ-Orbitrap (Thermo Fisher Scientific) under either positive or negative ionization modes, even after processing the data with a mass defect filter (Zhang et al., 2003, 2009). It is possible that the radioactive component in the unknown peak was not readily ionized under the mass spectrometric conditions used in these experiments.
Approximately 5% of the administered dose was excreted as unchanged parent compound in rat bile after oral administration of [14C]BMS-690514. Glucuronide conjugates of the parent (M6) and oxidative metabolites (M7–M9, M27, and M28) accounted for another 31.8% of the dose in the rat bile. Glucuronide-conjugated metabolites also comprised the majority of the biliary radioactivity after oral administration of [3H]BMS-690514 to BDC dogs (data not shown). In contrast, BMS-690514 and oxidative metabolites comprised most of the excreted dose in the feces of intact rats and dogs, and few to no glucuronide metabolites were detected (Fig. 4; Table 3). The difference in the bile and fecal profiles was attributed to hydrolysis of the conjugative metabolites in the gastrointestinal tract before excretion in the feces (Parker et al., 1980; Christopher et al., 2008). M6, which was hydrolyzed in the gastrointestinal tract, could be reabsorbed as the parent compound, leading to further oxidative metabolism. These data suggested that oxidative metabolism may ultimately be the primary clearance pathway for BMS-690514, even though glucuronide conjugates were responsible for the majority of the biliary radioactivity in BDC rats and dogs.
These results show the importance of collecting and profiling bile to understand the complete disposition of a compound, such as BMS-690514, which has a significant amount of conjugated metabolites excreted via the bile. Bile collection may also provide insight for other biotransformations. For example, several unidentified peaks observed in rat feces were not present in rat bile. These data suggest that the unknown compounds were likely generated by microflora residing in the gastrointestinal tract of the rat. Without bile data, these peaks would have been attributed to systemic metabolism.
Together, the fecal and bile metabolite profiles suggested that the primary in vivo metabolic pathways of BMS-690514 in all the species were hydroxylation to form M1, M2, and M37; O-demethylation to form M3; and direct glucuronidation to form M6, which was in good agreement with the in vitro hepatocyte data (Marathe et al., 2010). In human hepatocytes, along with M1 and M3 formed through oxidative pathway, direct glucuronide metabolite M6 was also formed (Marathe et al., 2010), suggesting that these pathways are also likely to be important in vivo metabolic pathways in humans (Fig. 5). In addition, the presence of these metabolites in vivo in animals suggested that these toxicology species, rats, dogs, and rabbits, were exposed to human-relevant metabolites.
M2 was a monohydroxylated metabolite of BMS-690514, with the site of hydroxylation on the 3-methoxyaniline group in a position that is para to the amino group. Studies have suggested that some drugs containing a hydroxylaniline group may be bioactivated to form a quinone imine, which is capable of reacting with glutathione (GSH) or nucleophilic groups in proteins (Kalgutkar et al., 2005a,b; Madsen et al., 2008; Li et al., 2009). Furthermore, this cytochrome P450 (P450)-dependent bioactivation may also result in mechanism-based inactivation of P450 enzymes (Kalgutkar et al., 2005b; Li et al., 2009). Although a time-dependent inhibition study of P450 enzymes has not been conducted for BMS-690514, the current data showed that no GSH adducts, cysteine adducts, or quinone imine metabolites were identified in any samples collected in the in vivo studies. In addition, no GSH conjugates of BMS-690514 were found in the in vitro incubations of BMS-690514 with GSH-supplemented rat or human liver S9 fractions (data not shown). Good extraction recovery of drug-related radioactivity was observed from plasma samples from rats, rabbits, or dogs after oral administration of [14C]BMS-690514. Therefore, the mechanism-based inactivation of P450 enzymes by BMS-690514 resulting from the bioactivation of M2 through the quinone imine pathway is not anticipated. In addition, BMS-690514 showed little potential to inhibit P450 enzymes in competitive inhibition assays, which included an evaluation of CYP3A4 and CYP2D6, two of the key P450 enzymes involved in the metabolism of BMS-690514 (Marathe et al., 2010).
M1 and M37 were both hydroxylated metabolites with hydroxyl group on the 16 position of the molecule. M37 generated from deuterated BMS-690514 (labeled on the 16 position) completely lost the isotope label, whereas M1 generated from the same deuterated BMS-690514 retained the deuterium label on the original site (data not shown), suggesting that M37 was not an intermediate of M1. Detailed experiments conducted to elucidate the exact structure of M1 and the mechanism for the formation of M1 and M37 will be described in a separate manuscript. It should be noted that because of well characterized affects of isotope substitution at sites of P450-mediated metabolism (Nelson and Trager, 2003), the substitution of deuterium at that position would affect the biotransformation and possibly the kinetics of clearance of BMS-690514.
BMS-690514 contains an m-anisidine (3-methoxyaniline) moiety. Anisidine isomers are known to be toxic (National Toxicology Program, 1978a,b; Thompson and Eling, 1991). Therefore, we considered the possibility that the m-anisidine group may be released from BMS-690514 after oral administration. No m-anisidine was detected in the plasma or excreta samples from any of the species, suggesting that the toxicity through this pathway is unlikely to be seen in vivo.
In summary, this investigation showed that BMS-690514 was well absorbed and cleared primarily by metabolism in rats and dogs. Biliary secretion was the major route for elimination of drug-related radioactivity in BDC rats. The primary in vivo metabolic pathways for BMS-690514 in rats, rabbits, and dogs were hydroxylation to form M1, M2, and M37; O-demethylation to form M3; and direct glucuronidation to form M6, which was similar to the in vitro metabolism of BMS-690514 observed in hepatocyte incubations from various animal species and humans. The major metabolic and clearance pathways in the preclinical species studied here are likely to be the same as those in humans, thus confirming the suitability of these species for the toxicity assessments of BMS-690514.
Footnotes
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
doi:10.1124/dmd.110.032755.
-
ABBREVIATIONS:
- HER
- human epidermal growth factor receptor
- EGFR
- epidermal growth factor receptor
- VEGFR
- vascular endothelial growth factor receptor
- NSCLC
- non–small cell lung cancer
- BMS-690514
- (3R,4R)-4-amino-1-((4-((3-methoxyphenyl)amino)pyrrolo[2,1-f][1,2,4]triazin-5-yl)methyl)-3-piperidinol
- LC/MS/MS
- liquid chromatography/tandem mass spectrometry
- BDC
- bile duct-cannulated
- HPLC
- high-performance liquid chromatography
- AUC
- area under the curve
- HMBC
- heteronuclear multiple-bond correlation
- GSH
- glutathione
- P450
- cytochrome P450.
- Received February 16, 2010.
- Accepted April 2, 2010.
- Copyright © 2010 by The American Society for Pharmacology and Experimental Therapeutics