Abstract
The metabolism and disposition of (1S,5R)-2,3,4,5-tetrahydro-7-(trifluoromethyl)-1,5-methano-1H-3-benzazepine (1), an α4β2 nicotinic acetylcholine receptor partial agonist, was determined in Sprague-Dawley rats after oral administration of [14C]1. In intact animals, mass balance was achieved within 48 h, with 5 times more radioactivity excreted in urine than in feces. Compound 1 underwent renal and metabolic clearance equally and exhibited a very long half-life attributable to a secondary peak occurring 8 h postdose in its serum concentration-time curve. In bile duct-cannulated (BDC) rats, mass balance was also achieved within 48 h with 73.7, 23.4, and 5.5% of the dose detected in bile, urine, and feces, respectively. Rats metabolized 1 by two primary routes: four-electron oxidation to either four amino acids or a lactam and formation of an N-carbamoyl glucuronide (M6), which was only detected in bile. The presence of M6 solely in bile and the double-humped serum concentration-time curve of 1 suggested the indirect enterohepatic cycling of 1 via M6 after oral administration. To explore this mechanistic hypothesis further, intravenous studies were conducted with 1 in both intact and BDC rats to determine the extent of 1 undergoing indirect enterohepatic cycling via M6. Compared with the pharmacokinetics in intact rats, total serum clearance was higher (1.7-fold) and volume of distribution was lower (1.6-fold) in BDC rats, resulting in a correspondingly shorter (2.5-fold) half-life, with 56% of administered 1 undergoing recirculation, an amount consistent with that (68% of dose) of M6 observed in bile from rats dosed orally with [14C]1.
Enterohepatic cycling (EHC) is a distributional event in which a compound is absorbed from the intestine, excreted into bile, and then reabsorbed into the portal vein after bile secretion into the intestine (Rowland and Tozer, 1995). This type of recirculation may occur to the administered compound directly and/or indirectly via its metabolite(s). For metabolite-mediated indirect EHC of a compound, one of the most common pathways is via a glucuronide conjugate (Scheline, 1973; Dickinson et al., 1979; Horton and Pollack, 1991; Bullingham et al., 1996; Skonberg et al., 2008). In these instances, the dosed compound (aglycone) experiences enzyme-mediated glucuronidation and the resulting glucuronide undergoes biliary clearance, ultimately entering the intestine upon bile secretion. Once in the intestinal lumen, the glucuronide is hydrolyzed, typically enzymatically (Scheline, 1973; Parkinson, 2001), to the aglycone, which reenters the portal circulation. This process often results in a “double-humped” plasma concentration-time curve after oral dosing (Dickinson et al., 1979; Bullingham et al., 1996; Wajima et al., 2002), with the second peak arising from the originally ingested compound entering the systemic circulation after its EHC, causing increased exposure and a prolonged half-life.
In this study, we report the absorption, metabolism, and excretion of (1S,5R)-2,3,4,5-tetrahydro-7-(trifluoromethyl)-1,5-methano-1H-3-benzazepine (1), an α4β2 nicotinic acetylcholine receptor partial agonist (Coe et al., 2005) potentially useful for smoking cessation, in Sprague-Dawley rats after a single oral dose of [14C]1 (Fig. 1). During this radiolabeled mass balance study, specific phenomena were observed that suggested enterohepatic recirculation of 1 via an N-carbamoyl glucuronide, its major metabolite. To investigate this particular disposition further, a subsequent intravenous bolus study was conducted with 1 in both intact and BDC rats. A combination of the results from both the radiolabeled oral study and the nonradiolabeled intravenous pharmacokinetics study provided an integrated data set to define mechanistically the indirect EHC of 1 via its N-carbamoyl glucuronide in Sprague-Dawley rats.
Materials and Methods
Chemicals and Reagents. Compounds 1, 2 (lactam metabolite), and 3 (formamide metabolite) were prepared by the Synthesis Group at Pfizer Global Research and Development (PGRD) (Groton, CT); [14C]1 (52.0 mCi/mmol, 96.7% radiochemical purity, 98.5% enantiomeric excess) was synthesized by the Radiochemical Synthesis Group at PGRD. The chemical purity of all synthetic compounds was >99%. Male Sprague-Dawley rat liver microsomes (RLMs) (21 mg of protein/ml, 0.61 nmol P450/mg protein) were prepared by the Candidate Enhancement Group at PGRD. β-Glucuronidase (EC 3.2.1.31, type IX-A, 1–5 × 106 units/g protein) was obtained from Sigma-Aldrich (St. Louis, MO). Jugular vein-cannulated/BDC male and female Sprague-Dawley rats were procured from Charles River Laboratories, Inc. (Wilmington, MA). Chemicals and solvents of reagent or HPLC grade were supplied by Aldrich Chemical Co. (Milwaukee, WI), Thermo Fisher Scientific (Waltham, MA), and Mallinckrodt Baker (Phillipsburg, NJ). Deuterated solvents were purchased from Cambridge Isotope Laboratories, Inc. (Cambridge, MA). All rat excreta, bile, sera, and whole brains were collected gravimetrically and stored at –20°C until analysis.
In Vivo Studies with [14C]1. The in-life portion of the study was conducted at Charles River Laboratories International, Inc. (Worcester, MA) in accordance with the Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, 1996). Sprague-Dawley rats were fasted overnight before compound administration and for 2 h postdose. A single dose (10 mg/kg, 40 μCi) of [14C]1 in deionized H2O (5 mg/ml) was administered via oral gavage to each rat. Individual animal doses were calculated based on respective pretreatment body weights and a dose volume of 2 ml/kg. The actual amount of dose solution administered to each animal was determined by weighing the loaded dosing syringe before and after it was dispensed. The study included four groups of rats.
Group 1 (three per gender). From intact animals, urine and feces were collected separately into metabolism cage containers surrounded by dry ice predose and in 24-h intervals from 0 to 168 h postdose for the assessment of cumulative excretion of total radioactivity, mass balance, and metabolite profiling and identification.
Group 2 (two per gender). From BDC animals, which received an infusion of a bile salts replacement solution (44 mmol of cholic acid and 13 mmol of NaHCO3/l of 0.9% NaCl, pH 7.0–7.4) during the collection period, bile was collected into containers surrounded by cold packs predose and from 0 to 4, 4 to 8, 8 to 12, 12 to 24, and 24 to 48 h postdose for metabolite profiling and identification. Urine and feces were also collected from these animals in 24-h periods from 0 to 48 h postdose.
Group 3 (three per gender). From jugular vein-cannulated rats, 0.5-ml blood samples were collected just before dosing and at 0.25, 0.5, 1, 2, 4, 6, 8, 12, and 24 h postdose for the pharmacokinetic evaluation of 1 and total radioactivity. After collection of the 1 and 8 h samples, 2 ml of whole blood from control donor animals was administered to each rat via the jugular vein cannula to maintain blood volume.
Group 4 (two per gender per time point). From intact animals, blood samples were collected by cardiac puncture after animal euthanasia by CO2 asphyxiation at 1, 4, 8, and 24 h postdose. Subsequently, the brain was extracted from each animal, rinsed of excess blood with saline, placed into a preweighed vial, weighed, and snap-frozen in liquid nitrogen.
All blood samples were collected into anticoagulant-free tubes and processed immediately to obtain serum. Control serum was harvested from blood collected from untreated animals that were not sacrificed.
Determination of Radioactivity within Biological Matrices. Triplicate gravimetric aliquots of each sample of urine (0.1 g), bile (0.05–0.1 g), and serum (0.05–0.1 g) were mixed with liquid scintillation cocktail (10 ml for bile and 5 ml for all other matrices) and counted for 2 min in a model LS 6000 or LS 6500 liquid scintillation counter (Beckman Coulter, Fullerton, CA).
Fecal samples were homogenized with deionized H2O (20% w/w, feces-H2O) using a probe-type homogenizer. Cage debris/rinse samples (collected in 50% reagent alcohol in H2O) were homogenized directly with a probe-type homogenizer. Triplicate gravimetric aliquots (0.1–0.5 g for either fecal homogenate or cage debris/rinse homogenate) were transferred into tared cones and pads, weighed, and combusted before radioanalysis. Sample combustion was performed using a model 307 or 387 sample oxidizer (PerkinElmer Life and Analytical Sciences, Waltham, MA). Combustion efficiency using a 14C standard was determined daily before the combustion of study samples, and the measured radioactivity content in feces and cage debris/rinse was adjusted using daily combustion efficiency values. Liberated [14C]CO2 was trapped in Carbo-Sorb E (PerkinElmer Life and Analytical Sciences) and mixed with Permafluor E+ scintillation fluid (PerkinElmer Life and Analytical Sciences), and the samples were analyzed for total radioactivity in a model LS 6000 or LS 6500 liquid scintillation counter for 2 min. Scintillation counter data were automatically corrected for counting efficiency using an external standardization technique and an instrument-stored quench curve generated from a series of sealed quench standards.
Thawed brains were weighed, diluted with H2O (3 ml of H2O/g brain), and homogenized with a probe-type homogenizer. After liquefaction, aliquots (0.2 g) of brain homogenate in scintillation cocktail (7 ml) were analyzed by LSC for 5 min.
Quantitative Analysis of 1 in Serum from Radiolabeled Studies. Serum concentrations of 1 were quantified using a characterized LC-MS/MS assay. To ensure that serum concentrations of 1 were within the dynamic range of the analytical assay, all samples, except the predose samples, were diluted accordingly before analysis. To each serum aliquot (100 μl) contained in a 15-ml glass centrifuge tube was added 20% MeCN in H2O (10 μl) containing an internal standard. Each sample was basified (100 μl, 0.1 M NaOH) and extracted with methyl tert-butyl ether (4 ml) via vortex mixing and centrifugation (1811 rcf for 10 min). The organic supernatant was transferred to a clean 15-ml glass centrifuge tube, concentrated to dryness under N2 at 35°C, and reconstituted in 20% MeCN in 10 mM ammonium formate, pH 3.0 (250 μl, solvent A). Samples were analyzed by an LC-MS/MS comprising a PE Sciex API-3000 tandem quadrupole mass spectrometer with a TurboIonSpray interface (PerkinElmer Life and Analytical Sciences), two Shimadzu LC-10A HPLC pumps (Shimadzu USA, Columbia, MD), and a CTC PAL Autosampler (LEAP Technologies, Carrboro, NC). Analytes within sample aliquots (20 μl) were eluted at 0.3 ml/min on a Phenomenex Synergi Max-RP analytical column (4 μ, 2.0 × 50 mm) using solvent A and MeCN (solvent B). The following one-step gradient was used: 0 to 4 min, 20% solvent B in solvent A; 4 to 5 min, 20 to 95% B in A; and 5 to 7 min, 95% B in A. After elution of 1 and the internal standard, the column was returned over 1 min to 20% B in A where it remained for 4 min before the next injection. Mass spectral data were collected in positive ionization mode using multiple-reaction monitoring following the m/z 228.1 → 191.1 fragmentation for 1. Instrument settings and potentials were adjusted to provide optimal data. The assay dynamic range was 0.25 to 50.0 ng/ml for 1.
Pharmacokinetic Calculations for Radiolabeled Studies. Pharmacokinetic parameters were calculated for each rat by noncompartmental analyses using WinNonlin Professional Version 3.2 (Pharsight, Mountain View, CA). Values used to determine the pharmacokinetic parameters of total radioactivity were calculated by converting the raw data generated by LSC to concentrations (nanogram-equivalents per milliliter) using the specific activity (3.6 mCi/mmol) of administered [14C]1. The AUC0–24 was calculated using the linear trapezoidal method, the elimination rate constant (kel) was determined by linear regression of the log concentration versus time data during the last observable elimination phase, half-life (t1/2) was calculated as 0.693/kel and AUC0–∞ was calculated as the sum of AUC0–24 and AUC24–∞, which was determined by dividing the plasma concentration at 24 h postdose by kel. Both maximal serum concentration (Cmax) and its first time of occurrence (Tmax) were taken directly from the serum concentration versus time data. Means and SDs were calculated when half or greater of the values exceeded the lower limit of quantification for 1 (0.25 ng/ml) or total radioactivity (2 times background radioactivity, which equated to 48 ng-Eq/ml). A value of 0 was used when a measured value was less than the lower limit of quantification.
Sample Preparation for Metabolite Profiling and Identification. At each step during the sample preparation of all biological matrices, total radioactivity levels were determined by LSC for recovery calculations. After preparation, all samples were analyzed as described later by LC-MS/MS with radiometric detection. Predose and blank samples served as controls for determining background radioactivity and endogenous, noncompound-related ions observed within respective matrices or their extracts by LC-MS/MS.
Urine. Urine samples from each rat collected from 0 to 72 h postdose representing ≥93% of total urinary radioactivity were pooled proportional to the amount of urine in each sampling period. Pooled samples were centrifuged (1811 rcf for 5 min) to remove particulate materials, and the resulting supernatants were analyzed directly.
Bile. Bile samples collected from 0 to 24 h postdose representing >89% of total biliary radioactivity were pooled proportional to the amount of bile in each sampling period. Pooled samples were centrifuged (1811 rcf for 5 min) to sediment particulate materials producing the analytical supernatant.
Feces. Fecal homogenates (approximately 1.0–7.7 g) from each rat collected from 0 to 96 h postdose, representing ≥95% of total fecal radioactivity, were pooled proportional to the amount of feces in each sampling period. After mixing at 37°C for 14 h in a reciprocal water bath, pooled fecal homogenates were diluted with MeCN (3 ml/g of homogenate), vortex-mixed, and centrifuged (1811 rcf for 10 min), and the resulting supernatants were isolated. If necessary, the remaining fecal pellets were extracted further with 33% H2Oin MeCN (6 ml) until ≥90% of the radioactivity, as determined by LSC analysis of the combined supernatants, from each pooled sample was extracted. The supernatants were concentrated to dryness by a N2 stream at 35°C and reconstituted in 10% MeCN in 10 mM ammonium formate, pH 3.4 (1 ml), for analysis.
Serum. Serum from blood samples collected by terminal bleeding of rats at 1, 4, 8, and 24 h postdose were used for circulatory metabolite profiling and identification. Serum samples were pooled within gender according to the method of Hamilton et al. (1981); i.e., 1.0, 1.8, 5.0, and 4.0 ml, respectively, of serum from each time point sample were combined to produce 11.8 ml of pooled serum for each gender profile. Proteins were precipitated from the serum pools by the addition of MeCN (24 ml), and the samples were vortex-mixed and centrifuged (1811 rcf for 10 min). The remaining serum protein pellets were extracted one additional time with 4 ml MeCN to ensure that >90% of radioactivity, as determined by LSC analysis of the pooled supernatants, from each pooled serum sample was extracted. The combined supernatants were concentrated to dryness, and the resulting residues were reconstituted in 10 mM ammonium formate, pH 3.4 (0.6 ml), and analyzed by LC-MS/MS. Quantitative analysis was performed by isolating HPLC effluent by a fraction collector set at 30-s intervals, and each respective fraction was mixed with 7 ml of TruCount scintillation fluid and subjected to LSC for 5 min. Individual serum radiochromatograms were generated from respective liquid scintillation data using Microsoft Excel (Microsoft Office 2000 9.0.7619 SP-3; Microsoft, Redmond, WA).
Brain. Thawed brains (two per time point per gender), isolated from the same rats sacrificed at 1, 4, 8, and 24 h postdose for the collection of blood for circulatory metabolite profiling and identification, were weighed, diluted with H2O (3 ml of H2O/g brain), and homogenized with a probe-type homogenizer. Brain homogenates were pooled within gender according to the method of Hamilton et al. (1981); i.e., 1.0, 1.8, 5.0, and 4.0 g, respectively, of homogenate were combined from each time point sample, which included two brains per gender. To precipitate proteins, the pooled homogenates (11.8 ml each) were diluted with 23.6 ml MeCN, vortex-mixed, and centrifuged (1811 rcf for 10 min). The protein sediments were not extracted further because >93% of radioactivity, as determined by LSC analysis of the supernatants, from each pooled sample was isolated by just one extraction. The supernatants were removed and concentrated to near dryness at 35°C by a N2 stream, and the resulting residues were reconstituted in 10% MeCN in 10 mM ammonium formate, pH 3.4 (1 ml) for analysis.
Metabolite Profiling, Identification, and Quantification. Samples were analyzed by LC-MS/MS (described previously) on an Agilent Zorbax SB-C18 analytical column (5 μ, 4.6 × 150 mm) in series with a β-RAM radiometric detector (IN/US Systems, Inc., Tampa, FL) containing a liquid scintillant cell (500 μl). Analytes within sample aliquots (25–100 μl) were eluted at 1 ml/min with 10 mM ammonium formate, pH 3.4 (solvent C), and MeCN (solvent B). The following gradient was used: 0 to 20 min, 10% solvent B in solvent C; 20 to 55 min, 10 to 50% B in C; 55 to 58 min, 50 to 80% B in C; and 58 to 60 min, 80% C in B. After the elution of 1 and its metabolites, the column was returned over 1 min to 10% B in C where it remained for 4 min before the next injection. Because of the coelution of two biliary metabolites (i.e., M5 and M6) using the aforementioned gradient, bile samples were injected a second time using the same gradient profile but substituting 10 mM ammonium acetate, pH 7.0, as solvent C to achieve baseline resolution of M5 and M6. For each matrix, >99% of the radioactivity injected onto the column eluted during the 65-min gradient programs. The HPLC effluent was split 1:9 between the mass spectrometer and the radiometric flow detector; liquid scintillation cocktail flowed at 3 ml/min to the radiometric detector. Mass spectral data were collected, using electrospray ionization in the positive ion mode and in full, precursor ion, neutral loss, product ion, and multiple-reaction monitoring scanning modes. Instrument settings and potentials were adjusted optimally in each mode. Analyst version 1.3 (PerkinElmer Life and Analytical Sciences) and Laura version 3.1.3.39 (LabLogic Systems Limited, Sheffield, UK) software were used for the acquisition and processing of mass spectral and radiochromatographic data, respectively. Metabolites were quantified by peak integration within generated radiochromatograms.
In Vitro Generation ofN-Carbamoyl Glucuronide M6. Incubations (2.5 ml) were performed in a capped but vented 15-ml Erlenmeyer flask under a CO2-saturated atmosphere at 37°C in a shaking water bath. Each incubation contained male RLMs [2 mg of protein/ml NaHCO3 (0.1 M) buffer, pH 7.5], 25 μmol of MgCl2, 0.063 μmol of 1, 13 μmol of uridine 5′-diphosphoglucuronic acid trisodium salt, 125 μg of alamethicin, and 12 μmol of d-saccharolactone. After 2 h, the incubation was quenched with 2.5 ml of MeCN, vortex-mixed, and centrifuged (1811 rcf for 5 min), and the resulting supernatant was concentrated and reconstituted in 20% H2O in MeCN (200 μl) for LC-MS/MS analysis.
Isolation of [14C]M6 from Bile and Its NMR Analysis. From the 0 to 4 h bile sample collected from one rat, [14C]M6 was isolated for NMR analysis by preparative HPLC using the aforementioned LC-MS/MS system equipped with a Chromolith SpeedRod (4.6 × 50 mm; Merck KGaA, Darmstadt, Germany) connected to a Chromolith Performance analytical column (4.6 × 100 mm). Three successive HPLC reversed-phase separations were performed using three different solvent gradients and manual fraction collection of column effluent. This entire isolation procedure and the LC-MS-NMR analysis of 1 and partially purified [14C]M6 are described in the Supplemental Data.
The highly purified and concentrated bile sample containing [14C]M6 (approximately 330 μg) in DMSO-d6 was analyzed on a Bruker BioSpin 600 MHz Avance DRX spectrophotometer equipped with a 2.5-mm broadband inverse probe. Proton chemical shifts are reported in parts per million (δ) relative to tetramethylsilane as referenced from the shift of residual protons in DMSO-d6 (2.49 ppm). 1H, correlation spectroscopy, and total correlation spectroscopy spectra were obtained at 298 K using double presaturation of solvent NMR resonances. A heteronuclear single quantum coherence spectrum was obtained at 298 K using WET suppression and a heteronuclear multiple bond correlation spectrum was obtained at 298 K. 1H NMR spectra were also obtained at 298 and 373 K.
Treatment of Purified [14C]M5 with β-Glucuronidase. Rat bile extract HPLC effluent (0.2 ml, 113 nCi), composed of approximately 90% [14C]M5 (approximately 27 nmol) and 10% [14C]M6 (approximately 3 nmol) and obtained as a side product of the purification of [14C]M6 for NMR analysis (see Supplemental Data), was diluted to 0.5 ml with buffer (0.1 M KH2PO4, pH 7.4) or β-glucuronidase (1000 units in 0.3 ml of buffer) in a 15-ml glass conical tube open to air. The samples were briefly vortex-mixed and then incubated at 37°C for 45 h in a shaking water bath. Aliquots (100 μl) were removed from each vial after 0, 2, 22, and 45 h of incubation, quenched with 100 μl of MeCN, diluted with 200 μl of H2O, vortex-mixed, and analyzed directly (50–100-μl analytical aliquots) by LC-MS/MS with radiometric detection.
In Vivo Studies with Nonradiolabeled 1. The in-life portion of the study was conducted at PGRD in accordance with the Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, 1996). Sprague-Dawley rats were fasted overnight before compound administration and for 7 h postdose. A single intravenous bolus dose (0.2 mg/kg) of 1 in deionized H2O (0.2 mg/ml) was delivered via a jugular vein catheter to each rat. Individual animal doses were calculated based on respective pretreatment body weights and a dose volume of 1 ml/kg. The study included two groups of rats.
Group 1 (three per gender). From jugular vein-cannulated (intact) rats, blood samples (0.3 ml) were collected at 0.083, 0.17, 0.33, 0.5, 0.67, 1, 1.5, 2, 3, 4, 7, 10, 14, 24, 32, 48, and 72 h postdose for the pharmacokinetic evaluation of 1. Whole-blood replacement (approximately 3 ml/rat) from donor rats was performed at 10 h postdose to replace the approximate amount of blood loss during the study. Serial blood samples were collected in Microtainer tubes (BD Biosciences, Franklin Lakes, NJ) and placed on ice until centrifugation (13,400 rcf for 10 min) to prepare serum. Serum samples were transferred to 1.2-ml polypropylene tubes and stored frozen at –20°C until analysis. Urine was collected from the metabolic cages over the following time intervals: 0 to 7, 7 to 24, 24 to 48, and 48 to 72 h postdose. The urine volume was measured at the end of each collection interval, transferred to a polypropylene tube, and stored at –20°C until analysis.
Group 2 (three per gender). From jugular vein-cannulated/BDC rats, blood samples (0.3 ml) were collected at 0.083, 0.25, 0.75, 1.5, 3, 7, 10, 24, 32, 48, and 72 h postdose for the pharmacokinetic evaluation of 1. Serial blood samples were collected in red-capped Microtainer tubes and placed on ice until centrifugation (13,400 rcf for 10 min) to prepare serum. Serum samples were transferred to 1.2-ml polypropylene tubes and stored frozen at –20°C until analysis. Urine was collected from the metabolic cages over the following time intervals: 0 to 10, 10 to 24, 24 to 48, and 48 to 72 h postdose. The urine volume was measured at the end of each collection interval, transferred to a 1.2-ml polypropylene tube, and stored at –20°C until analysis. Bile was collected via the bile duct catheter over the same time intervals as urine collection and stored as described for urine samples.
Quantitative Analysis of 1 in Serum, Urine, and Bile from Nonradio-labeled Studies. Serum, urine, and bile concentrations of 1 were quantified using the characterized LC-MS/MS assay described previously with a dynamic range of 0.25 to 50.0 ng/ml for 1. To ensure that serum concentrations of 1 were within the dynamic range of the analytical assay, all samples, except the predose samples, were diluted accordingly before analysis. The serum workup was the same as discussed previously. For urine and bile, both 20% MeCN in H2O (10 μl) containing an internal standard and 2 ml of MeCN were added to each sample aliquot (100 μl) contained in a 15-ml glass centrifuge tube. The diluted sample was vortex-mixed and centrifuged (600 rcf for 10 min). The organic supernatant (2 ml) was transferred to a clean 15-ml glass centrifuge tube, concentrated to dryness under N2 at 35°C, and reconstituted in 20% MeCN in solvent A (200 μl). Sample analysis was identical to that already described.
Pharmacokinetic Calculations for Nonradiolabeled Studies. For intravenous studies with 1, the estimated initial concentration [C0(est)], area under the serum concentration-time curve from time 0 to the last measurable serum concentration (AUC0–Tlast), AUC0–∞, kel, t1/2, total serum clearance (CL), and volume of distribution at steady-state (Vss) were calculated from the serum concentration versus time data in individual animals by noncompartmental analysis as described previously. The cumulative percentage of the dose recirculated was calculated as follows (Horton and Pollack, 1991), where AUCI and AUCBDC are the mean AUC0–∞ values determined from intact (I) and BDC rats, respectively: Renal clearance (CLR) or biliary clearance (CLbil) of 1 was calculated from the total amount of excreted 1 (Ae) detected within urine or bile, respectively, over 72 h and the serum AUC0–∞ of the intravenously dosed animals:
Results
Excretion of Total Radioactivity After Oral Administration of [14C]1. The excretion of total drug-related material was quite rapid in both genders; on average, >86% of the administered radioactivity was excreted within the first 48 h. Although there were no readily apparent gender-related differences in overall excretory routes or mass recoveries, male and female recovery values (mean ± S.D.) for both intact and BDC animals are listed in Table 1. Overall in intact rats, the mean recovery across genders of drug-related material within excreta and cage rinse was 101.2 ± 0.5%, with 84.2 ± 5.7% and 17.0 ± 6.1% of the dose detected within urine and feces, respectively. On average in BDC animals, 73.6 ± 22.4% of the dose was detected in bile from 0 to 48 h postdose, with the 0 to 24 h bile sample comprising >89% of the 0 to 48 h biliary radioactivity (Table 1). Urine and feces collected from BDC animals over the 48-h interval accounted for, on average across genders, 23.3 ± 15.5% and 5.5 ± 7.3% of the administered dose, respectively.
Pharmacokinetics of 1 and Total Radioactivity After Oral Administration of [14C]1. As with total radioactivity excretion patterns, no obvious gender-related differences in systemic exposure to 1 or total radioactivity were observed. For 1, its average serum concentration versus time is plotted in Fig. 2, and its gender-specific pharmacokinetic parameters (mean ± S.D.) are listed in Table 2. On average in males and females, serum concentrations of 1 rose to a mean Cmax of 416 ± 74 ng/ml at a mean Tmax of 5.0 ± 2.4 h postdose with a mean t1/2 of 16.2 ± 5.5 h.
For total radioactivity, its average serum concentration-time curve and pharmacokinetic parameters (mean ± S.D.) are also in Fig. 2 and Table 2, respectively. On average in males and females, the mean Cmax and Tmax values of total radioactivity were 733 ± 85 ng-Eq/ml and 7.0 ± 3.0 h, respectively, whereas the mean t1/2 was 51.1 ± 38.3 h. It is important to note that the reported elimination half-lives for both 1 and total radioactivity are considered estimates because the span ratio was less than 2, and this has direct consequences on the degree of confidence in CL/F and AUC values, which at best also are estimates. On average in both genders, the AUC0–∞ for 1 represented 25.3% of the AUC0–∞ for total drug-related material. It is interesting to note that serum concentration-time curves for both 1 and total 14C had a double-humped profile with the second upward leg occurring at approximately 6 h postdose, suggestive of the EHC of 1.
Structural Rationalization of 1 and Its Metabolites. Compound 1 had a protonated molecular ion of m/z 228 and an LC tR of approximately 31.8 min. The CID product ion spectrum of m/z 228 contained fragment ions with m/z 211, 199, 196, 191, 179, 177, 159, and 141 (Table 3); structures for the majority of these fragment ions were mechanistically rationalized to facilitate metabolite structure elucidation (Fig. 3). For metabolite identification purposes, precursor ion scanning of two diagnostic fragment ions m/z 191 (the base peak within the CID spectrum of 1) and m/z 177 (a very minor ion within the CID spectrum of 1) was undertaken. Fragmentation patterns of all detected metabolites and available metabolite synthetic standards (i.e., 2 and 3) demonstrated that biotransformation on or directly adjacent to the nitrogen within 1 dramatically shifts the base peak of the modified compound from m/z 191 to m/z 177 because of a change in molecular electronics that alters the CID bond-breaking scenario proposed for 1 (Fig. 3). Therefore, any metabolites detected by a precursor ion scan of m/z 191, but not m/z 177, would be predicted to be enzymatically modified in a region removed from the alicyclic secondary nitrogen. A summary of all metabolite LC-MS/MS data is found in Table 3. In addition, neutral loss of 176 scanning was used throughout MS analyses to detect specifically potential glucuronide metabolites. The identification of a metabolite as a fully characterized standard was determined by the indistinguishable CID spectra and LC tR values of the compounds, as well as by an increase in metabolite MS peak height upon addition of the authentic standard to the analytical sample.
A key interest of the metabolite identification work was the characterization of the suspected N-O-glucuronide (M5) and N-carbamoyl glucuronide (M6) metabolites, in particular, the chemical fate of both M5 and M6 when treated with β-glucuronidase and the definitive structure elucidation of the major bile-specific metabolite M6. Treatment of bile extracts containing M5 (m/z 420, LC tR approximately 38.5 min) with β-glucuronidase resulted in its full conversion to a new single radioactive peak (LC tR approximately 37.7 min). Mass spectral analysis of the new radioactive peak detected two compound-related ions: m/z 244, 16 amu greater than 1, and m/z 242, which had an LC tR and CID spectrum identical to those of putative nitrone M7. The compound-related molecule with m/z 244 is believed to be N-hydroxyl-1, which could itself be oxidized further, either in vitro, upon sample workup and/or within the MS source (Huizing and Beckett, 1980; Lindeke, 1982), to M7, which was the only one of these two compounds detected in vivo. Based on the enzymatic hydrolysis of M5 to a compound with m/z 244 (Miller et al., 2004), the fact that no carbon-hydroxylated metabolite of 1 was observed in any biological matrix, and the structural rationalization of each fragment ion observed in its CID product ion spectrum (especially the m/z 244 and 226 fragments, which are believed to correspond to N-hydroxyl-1 and its dehydrated iminium, respectively), M5 was tentatively identified as an N-O-glucuronide of 1. Dissimilarly, M6 was converted quantitatively to 1 by β-glucuronidase, paralleling the hydrolytic behavior of N-carbamoyl glucuronides (Tremaine et al., 1989; Schaefer, 1992; Shaffer et al., 2005). Furthermore, an N-carbamoyl glucuronide of 1 was generated in vitro; the CID spectra, whose fragmentation cascades mirrored those of known N-carbamoyl glucuronides (Shaffer et al., 2005; Schaefer, 2006; Suzuki and Kamimura, 2007), and LC tR values of the biosynthesized N-carbamoyl glucuronide and M6 were identical. NMR analyses of M6 purified from rat bile extracts unequivocally identified it as an N-carbamoyl glucuronide of 1. Specifically, the heteronuclear multiple bond correlation spectrum indicated a long-range correlation between the anomeric proton and the carbamate carbon. In addition, in LC-MS-NMR experiments, the [M + D]+ detected by mass spectrometry for M6 in deuterated solvents was 5 amu greater than its [M + H]+ in nondeuterated solvents, consistent with four exchangeable protons within M6. A summary of all 1H NMR data with full spectral assignments for both 1 and M6 is found in the Supplemental Data.
Quantitative Profile of [14C]1 and Its Metabolites in Excreta.Urine. In addition to 1, five urinary metabolites were observed for both genders (Table 1). Four metabolites (M1, M2, M3, and M4) were proposed to be amino acids, whereas the fifth (M5) was tentatively identified as an N-O-glucuronide of 1. The structural assignments of M1, M2, M3, and M4 as amino acids arose from both structural deciphering of ion fragments within their respective CID spectra and the quantitative base-catalyzed hydrolysis of an authentic standard of 2 (4 μM in 1 M NaOH, 42 h, 25°C) to two compounds, with LC tR values and CID spectra identical to M2 and M4, in a ratio of 1:14 based on total ion intensity peak height ratios. Heating the basic aqueous solution to 70°C for another 1 h (or 24 h) increased the ratio, presumably via epimerization, of M2/M4 to 1:2, suggesting that M4 was both the kinetically and thermodynamically favored diastereomeric amino acid and was consistent with the metabolite ratios observed in vivo. Identical chemical behavior occurred during the base-catalyzed hydrolysis of the other regioisomeric lactam to M1 and M3. On average in urine, 55.2 ± 6.8% of the administered dose was unchanged 1, which equated to a renal clearance of 3 times GFR (Davies and Morris, 1993).
Feces. In addition to 1, only the urinary amino acid metabolites (M1, M2, M3, and M4) were also observed in male feces, with all but M1 detected in female feces. On average in feces, 5.5 ± 2.1% of the administered dose was unchanged 1 (Table 1).
Bile. In both genders, two glucuronide metabolites (M5 and M6) were detected in bile, with M6 corresponding on average to 68% of administered 1 (Table 1). A trace amount (<1%) of 1 was detected within male bile only.
Quantitative Profile of [14C]1 and Its Metabolites in Serum and Brain. In addition to 1, six radioactive peaks were identified in both male and female serum: amino acids M1, M2, M3, and M4, N-O-glucuronide M5, nitrone M7, lactam 2, and formamide 3 (Table 4). On average in serum, 1 comprised 14.0% of total serum radioactivity whereas the two most significant circulatory metabolites M3/M4 and 2 accounted for 21.1 and 14.7%, respectively. In both genders, 1 represented 93% of total radioactivity in brain accompanied by only minor amounts of 2 and 3.
Pharmacokinetics of 1 in Intact and BDC Rats After Intravenous Administration of 1. Upon completion of the mass balance study with [14C]1, an intravenous bolus pharmacokinetic study was undertaken with nonradiolabeled 1 in both intact and BDC Sprague-Dawley rats to assess quantitatively its extent of EHC. Gender-specific and mean (male and female) pharmacokinetic parameters (mean ± S.D.) for this study are summarized in Table 5. No gender differences in intravenous pharmacokinetics were obvious in either intact or BDC rats; thus, mean serum concentration-time curves are plotted (Fig. 4), and combined data sets were used for the statistical analysis of pharmacokinetic parameters (Table 5).
Compared with the pharmacokinetics in intact rats, CL in BDC rats was higher (1.7 times; t test p < 0.01) and Vss was lower (1.6 times; t test p < 0.01), resulting in a correspondingly shorter (2.5 times; t test p < 0.01) t1/2. These results suggest that the disposition of 1 in rats involves EHC. The cumulative percentage of the intravenous dose of 1 undergoing recirculation in rats was calculated to be 56% based on mean AUC0–∞ in intact and BDC animals. The CLR of 1 in BDC rats was similar to that determined in intact animals, with the changes in CL resulting from higher (2.4 times; t test p < 0.01) nonrenal CL (CLnr). In BDC rats CLbil accounted for only 0.52% of the CL of 1. Therefore, biliary excretion of hepatic biotransformation product(s) of 1 and their back-conversion to 1 and its reabsorption in the intestine (rather than direct recycling of 1 via biliary secretion) is likely to be the predominant underlying mechanism of EHC of 1 in Sprague-Dawley rats. Because of the lack of a synthetic standard of M6, only qualitative analyses of bile collected from intravenous studies could be conducted to determine the presence of M6 within these samples. Thus, aliquots (100 μl) from bile samples (0–10 h and 10–24 h collection interval samples) underwent the developed metabolite identification LC-MS/MS methodology (i.e., neutral loss, precursor ion, and product ion scanning modes) described previously and were determined to contain M6.
Discussion
No obvious gender-related differences were observed for the pharmacokinetics, excretion patterns, or metabolic profiles in Sprague-Dawley rats after a 10 mg/kg oral dose of [14C]1. On average in intact animals, 101% of administered radioactivity was recovered over the study period with 84.2% in urine and 17.0% in feces. In BDC rats, mass balance (105%) was achieved within 48 h postdose with 73.7, 23.4, and 5.5% of the dose detected in bile, urine, and feces, respectively. A relatively short Tmax and large amounts of radioactivity excreted in bile and urine from 0 to 24 h postdose suggested that [14C]1 was readily and substantially absorbed in rats. In intact animals, 1 underwent renal and metabolic clearance equally, and exhibited a very long t1/2. However, it is important to note that the reported elimination half-lives are considered estimates because the span ratio was less than 2, and this has direct consequences on the degree of confidence in presented CL/F and AUC values, which too are at best estimates. Active renal secretion of 1 was observed as its unbound CLR was 3-fold greater than the GFR. An average brain-to-serum AUC0–24 ratio of 28 suggests that 1 has extensive brain penetration in rats, consistent with its physicochemical properties and the fact that it is not a multidrug resistance 1 P-glycoprotein substrate.
A schematic overview of the metabolism of 1 in Sprague-Dawley rats is presented in Fig. 5. Qualitatively identical and quantitatively similar metabolite profiles for 1 in all biological matrices were observed in both genders. Based on the identification of all major metabolites in excreta, bile, serum, and brain, 1 undergoes two primary routes of metabolism within rats: chronological N-carbamoylation and glucuronidation to generate an N-carbamoyl glucuronide (M6) and four-electron oxidation to either four amino acids (M1, M2, M3, and M4) or a lactam (2). Minor biotransformation pathways of 1 were oxidation to a putative hydroxylamine to produce a nitrone (M7) or N-O-glucuronide (M5) and N-formylation (3). It is interesting to note that certain metabolites were only observed in specific matrices. For example, amino acid metabolites (M1, M2, M3, and M4) were detected in urine, feces, and serum; their absence in brain and high percentages of total circulating radioactivity with only minor urinary quantities are probably attributable to their low volumes of distribution as expected by their polar, zwitterionic nature. Alternatively, the lactam 2 was only observed in serum and brain. Its presence in circulation but absence in excreta suggest the possibility that it may be hydrolyzed (possibly enzymatically within the kidney) to its corresponding amino acids M2 and M4 before excretion. Although an authentic standard of 2 was readily converted to M2 (minor) and M4 (major) under aqueous basic conditions, 2 was completely stable in both RLMs (over 1 h, ±NADPH) and rat hepatocytes (over 4 h). This was also true for an authentic standard of the regioisomeric lactam 2, which produced M1 (minor) and M3 (major) when treated with aqueous base.
The most intriguing metabolic pathway for 1 was its conversion to N-carbamoyl glucuronide (M6), which was isolated for its structural confirmation by NMR analysis; NMR and MS spectra of N-carbamoyl glucuronides are well characterized in the literature (Shaffer et al., 2005; Schaefer, 2006). In addition, the stability of N-carbamoyl glucuronides in buffer (Thomas et al., 2008), urine (Elvin et al., 1980; Straub et al., 1988), plasma (Tremaine et al., 1989), and serum (Thomas et al., 2008) is also well documented. More interesting than purely the identification of M6 as an N-carbamoyl glucuronide was its influence over the disposition of 1 (and total radioactivity) in rats after an oral dose of [14C]1. Preliminary mass balance study observations of ≥80% of administered radioactivity excreted in urine and 1 being the major urinary component (≥63%) suggested that 1 was well absorbed and underwent only moderate metabolism in vivo. This initial theory was supported by BDC animal data in which large amounts (≥65%) of orally dosed radioactivity were observed in bile with a concomitant drop in urinary and fecal radioactivity. These excretion patterns coupled with the double-humped serum concentration-time curves suggested that 1 was probably undergoing direct EHC. However, metabolite profiling of biliary radioactivity provided a much different interpretation, in that the predominant (>95%) biliary radioactive component was M6, a metabolite not detected in any other biological matrix, and not 1 (<1%). Together, these data clearly pointed to the indirect EHC of 1 via M6.
To define this suspicion more definitively, intravenous studies were conducted with 1 in both intact and BDC Sprague-Dawley rats to determine empirically the extent of 1 undergoing indirect EHC via M6. Compared with the pharmacokinetics in intact rats, CL in BDC rats was higher (1.7-fold) and Vss was lower (1.6-fold), resulting in a correspondingly shorter (2.5-fold) t1/2. These results suggested that the disposition of 1 in rats involves EHC, with 56% of the intravenous dose of 1 undergoing recirculation, an amount consistent with that (68% of dose) of M6 observed in bile from rats orally dosed with [14C]1. The CLR of 1 in BDC rats was similar to that in intact animals, with the changes in CL resulting from a 2.4-fold higher CLnr. In addition, CLbil accounted for only 0.5% of the net CL of 1 in BDC rats, again consistent with <1% of 1 detected in bile from the oral radiolabeled study. Therefore, biliary clearance of a hepatic biotransformation product (i.e., M6) of 1, its back-conversion to 1 and the reabsorption of 1 in the intestine (rather than direct recycling of 1 via biliary excretion) is proposed as the principal underlying mechanism of EHC of 1 in Sprague-Dawley rats.
Based on the reported data, a molecular mechanism rationalizing the EHC of 1 is proposed in Fig. 6. After oral dosing and absorption, 1 undergoes hepatic-mediated metabolic conversion to M6, which itself undergoes biliary clearance. Bile excretion into the upper gastrointestinal tract exposes M6 to intestinal microflora rich in β-glucuronidase (Scheline, 1973; Parkinson, 2001), the enzyme responsible for the quantitative conversion of M6 to 1 in vitro, resulting in glucuronide hydrolysis and spontaneous decarboxylation to 1, which is predominately reabsorbed and ultimately enters systemic circulation (Fig. 6A). [Hypothetically, M6, which was converted directly to 1 by strong aqueous base (1 N NaOH, pH 14), may also be hydrolyzed nonenzymatically within the intestine (pH 6.8). However, this seems less likely because M6 was completely stable in buffer (pH 7.4) at 37°C for 45 h, consistent with other N-carbamoyl glucuronides (Thomas et al., 2008).] This circulatory reentry of 1 manifests in the secondary peak in its serum concentration-time curve. A combination of the intact and BDC radiolabeled study data suggest that this EHC phenomenon occurs over a fairly short period of time, as total biliary radioactivity crests within the first 4 h postdose and then declines over the next 4 to 8 h (Fig. 6B) as the serum concentration-time curves within intact animals show a second peak at approximately 8 h postdose (Fig. 6C), consistent with the timing of the drop in biliary radioactivity in BDC animals.
The suggestion that the conversion of 1 to M6 is mainly (if not exclusively) metabolically mediated hepatically arises from hepatocyte incubations conducted with [14C]1 in which 80% of 1 was metabolized to M6 over 4 h. Over a 1-h period, RLMs (±NADPH, 2 mg of protein/ml, 10 μM 1) failed to metabolize 1, but they did convert 1 to M6 (approximately 15% conversion at 25 μM) when optimized for N-carbamoyl glucuronidation. These results are supported by a recent data set (Cerny et al., 2008) reporting that under similar conditions hepatic microsomes had a 25-fold greater relative rate than intestinal microsomes for the formation of an N-carbamoyl glucuronide of the structurally similar secondary alicyclic amine lorcaserin.
It is worthy of note that two glucuronide metabolites were observed in rats: the predominant N-carbamoyl glucuronide (M6) and the minor N-O-glucuronide (M5). Whereas M6 was only in bile, M5 was observed in bile (minimally), serum, and urine. This finding suggests that once formed hepatically, M6 undergoes exclusive biliary clearance. Alternatively, M5 does not solely undergo biliary clearance but rather it reaches systemic circulation and is ultimately excreted renally. This yet to be defined selectivity may be due to a glucuronidecentric hepatic transporter, such as the organic ion transporter mrp2 (Klaassen and Watkins, 1984; Keppler and Arias, 1997; Seitz et al., 1998; Xiong et al., 2000; Westley et al., 2006), within the canalicular membrane that results in the preferential transport of M6 over M5 into bile. Conversely, it may be hypothesized that M5 undergoes preferential (versus M6) transport into blood, such as that mediated by mrp3 (Soroka et al., 2001; Villaneuva et al., 2008). This mrp2 hypothesis is currently being investigated in our laboratory using either probenecid-treated wild-type rats or Wistar TR– rats lacking mrp2 (Jansen et al., 1985; Xiong et al., 2000).
Acknowledgments
We acknowledge Dr. Klaas Schildknegt (Radiosynthesis Group at PGRD, Groton, CT) for the synthesis andpurification of [14C]1, Dr. Jotham Coe for providing 1, Jason McKinley for synthesizing 2, and Nga Do for supplying 3.
Footnotes
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.109.027037.
-
ABBREVIATIONS: EHC, enterohepatic cycling; 1, (1S,5R)-2,3,4,5-tetrahydro-7-(trifluoromethyl)-1,5-methano-1H-3-benzazepine hydrochloride; [14C]1, (1S,5R)-2,3,4,5-tetrahydro-7-(trifluoromethyl)-1,5-methano-1H-4[14C]-3-benzazepine hydrochloride; BDC, bile duct-cannulated; 2, (1S,5R)-2,3,4,5-tetrahydro-7-(trifluoromethyl)-1,5-methano-4-oxo-1H-3-benzazepine; 3, (1S,5R)-2,3,4,5-tetrahydro-7-(trifluoromethyl)-1,5-methano-1H-3-benzazepine-3-carboxaldehyde; PGRD, Pfizer Global Research and Development; RLM, rat liver microsome; HPLC, high-performance liquid chromatography; LSC, liquid scintillation counting; LC, liquid chromatography; MS/MS, tandem mass spectrometry; MeCN, acetonitrile; rcf, relative centrifugal force; AUC, area under the concentration-time curve; MS, mass spectrometry; CID, collision-induced dissociation; amu, atomic mass units; GFR, glomerular filtration rate; mrp, multidrug resistance-associated protein.
-
↵ The online version of this article (available at http://dmd.aspetjournals.org) contains supplemental material.
-
↵1 Current affiliation: Millennium Pharmaceuticals Inc., Department of Clinical Pharmacology, Cambridge, Massachusetts.
-
↵2 Current affiliation: Gilead Sciences, Inc., Drug Safety and Evaluation, Foster City, California.
- Accepted March 27, 2009.
- Received February 2, 2009.
- The American Society for Pharmacology and Experimental Therapeutics