1842
Chem. Res. Toxicol. 2005, 18, 1842-1848
Stability and Comparative Metabolism of Selected Felbamate Metabolites and Postulated Fluorofelbamate Metabolites by Postmitochondrial Suspensions Robert J. Parker,† Neil R. Hartman,† Bryan A. Roecklein,‡ Henry Mortko,‡ Harvey J. Kupferberg,| James Stables,§ and John M. Strong*,† Laboratory of Clinical Pharmacology, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, Maryland 20993, MedPointe Pharmaceuticals, Somerset, New Jersey 08873, Potomac, Maryland 20854, National Institute of Neurological Diseases and Stroke, National Institutes of Health, Bethesda, Maryland 20892 Received May 17, 2005
Evidence has been presented suggesting that a reactive metabolite, 2-phenylpropenal (ATPAL), may be responsible for the toxicities observed during therapy with the antiepileptic drug felbamate (FBM). Formation of ATPAL from its unstable immediate precursor, 3-carbamoyl-2-phenylpropionaldedhyde (CBMA) requires the loss of the hydrogen atom at position 2 in the propane chain, and it has been postulated that substitution of this atom with fluorine would prevent the formation of ATPAL. On the basis of this hypothesis, 2-fluoro-2-phenyl1,3-propanediol dicarbamate (F-FBM) was synthesized and is presently undergoing drug development. To test this hypothesis, we compared the metabolism by human liver postmitochondrial suspensions (S9) in vitro of selected FBM and postulated F-FBM metabolites leading to formation of CBMA or 3-carbamoyl-2-fluoro-2-phenyl-propionaldehyde (F-CBMA). All S9 incubations included GSH as a trapping agent for any reactive metabolites formed. Our results indicated that, in phosphate buffer, pH 7.4, at 37 °C, the half-life for 4-hydroxy5-phenyltetrahydro-1,3-oxazin-2-one (CCMF) was 2.8 and 3.6 h in the presence or absence of GSH, respectively; compared to 4-hydroxy-5-fluoro-5-phenyl-tetrahydro-1,3-oxazin-2-one (FCCMF) which lost only 2.5% or 4.9% over 24 h under the same conditions. When incubated with S9 in the presence of the cofactor, NAD+, 2-phenyl-1,3-propanediol monocarbamate (MCF) was oxidized to CCMF which was further oxidized to 3-carbamoyl-2-phenylpropionic acid (CPPA). 2-Fluoro-2-phenyl-1,3-propanediol monocarbamate (F-MCF) under similar conditions was stable, and no metabolites were observed. When CCMF was incubated with S9 in the presence of NAD+ cofactor, oxidation to CPPA and reduction to MCF were observed. In addition, a new atropic acid GSH adduct (ATPA-GSH) was identified by mass spectrometry. When F-CCMF was incubated under the same conditions as CCMF, both reduced and oxidized metabolites, F-MCF and 3-carbamoyl-2-fluoro-2-phenylpropionic acid (F-CPPA), respectively, were formed but at significantly lower rates, and no GSH conjugates were identified. Our results support the hypothesis that F-FBM and F-CCMF are not metabolized by S9 in vitro to the known reactive FBM metabolite, ATPAL.
Introduction 1
Shortly after the antiepileptic drug, felbamate (FBM) was approved for marketing in 1993 and following 50 cases of aplastic anemia (1) or hepatotoxicity (2) the U.S. Food and Drug Administration restricted its use to patients already receiving the drug or to those refractory to other epilepsy treatments (3). Efforts to identify the mechanism leading to toxicities observed in patients receiving FBM have focused on the metabolism of FBM to a reactive intermediary metabolite, 2-phenylpropenal (ATPAL) (4-9). Recently, Popovic et al. (10) reported that * To whom correspondence should be addressed at U.S. Food and Drug Administration, White Oak Bldg. 64, Room 2026, 10903 New Hampshire Ave, Silver Spring MD 20993-0002. E-mail: strongj@ cder.fda.gov. Tel: 301-796-0121. Fax: 301-796-9818. † Laboratory of Clinical Pharmacology, Center for Drug Evaluation and Research, U.S. Food and Drug Administration. ‡ MedPointe Pharmaceuticals. | Potomac, Maryland. § National Institute of Neurological Diseases and Stroke, National Institutes of Health.
ATPAL is a very potent immunogen. Their findings support the possible role of this reactive metabolite in felbamate-induced hepatotoxicity and aplastic anemia (10). 1 Abbreviations: ATPA, R-methylenebenzeneacetic acid, 2-phenylacrylic acid, atropic acid; ATPA-GSH, atropic acid GSH adduct; ATPAL, atropaldehyde, 2-phenylpropenal; ATPAL-GSH, atropaldehyde GSH adduct; CBMA, 3-carbamoyl-2-phenylpropionaldehyde, 3-oxo-2phenylpropyl aminooate; CCMF, 4-hydroxy-5-phenyltetrahydro-1,3oxazin-2-one, 4-hydroxy-5-phenyl-1,3-oxazaperhydroin-2-one; CPPA, 3-carbamoyl-2-phenylpropionic acid, 3-(aminocarbomyloxy)-2-phenylpropionic, acid; FBM, felbamate, 3-(aminocarbonyloxy)-2-phenylpropyl aminooate, 2-phenyl-1,3-propanediol dicarbamate, Felbatol; F-CBMA, 3-carbamoyl-2-fluoro-2-phenyl-propionaldehyde, 3-oxo-2fluoro-2-phenylpropyl aminooate; F-CCMF, 4-hydroxy-5-fluoro-5-phenyl-tetrahydro-1,3-oxazin-2-one, 4-hydroxy-5-fluoro-5-phenyl-1,3-oxazaperhydroin-2-one; F-CPPA, 3-carbamoyl-2-fluoro-2-phenylpropionic acid, 3-(aminocarbomyloxy)-2-fluoro-2-phenylpropionic acid; F-FBM, fluorofelbamate, 3-(aminocarbonyloxy)-2-fluoro-2-phenylpropyl aminooate, 2-fluoro-2-phenyl-1,3-propanediol dicarbamate; F-MCF, 2-fluoro2-phenyl-1,3-propanediol monocarbamate, 3-hydroxy-2-fluoro-2-phenylpropyl aminooate; MCF, 2-phenyl-1,3-propanediol monocarbamate, 3-hydroxy-2-phenylpropyl aminooate.
10.1021/tx050130r CCC: $30.25 © 2005 American Chemical Society Published on Web 11/22/2005
Felbamate and Fluorofelbamate Metabolites in Vitro Scheme 1. Partial FBM and F-FBM Metabolic Pathwaysa
Chem. Res. Toxicol., Vol. 18, No. 12, 2005 1843
dro-1,3-oxazin-2-one as two disatereomers (F-CCMF) by S9 in vitro. The three specific goals of this study were (1) to compare the stability of CCMF and F-CCMF in S9 phosphate buffer (pH 7.4) at 37 °C, (2) to compare the metabolism of CCMF and F-CCMF by S9 with and without the presence of NAD+ cofactor, and (3) to compare the metabolism of MCF and F-MCF by S9 with and without the presence of NAD+ cofactor. All S9 incubations contained 2 mM GSH as a trapping agent for any reactive metabolites formed.
Experimental Methods
a Arrows with question marks indicate postulated pathways for FBM and F-FBM but not identified.
FBM human metabolic pathways leading to formation of ATPAL, a reactive metabolite, are shown in Scheme 1. Initially, one of the FBM carbamoyl moieties is hydrolyzed by an esterase or amidase leading to the formation of 2-phenyl-1,3-propanediol monocarbamate (MCF) which is then oxidized to form 3-carbamoyl-2phenylpropionaldehyde (CBMA). The fate of CBMA is dependent on three competing pathways: (1) formation of the cyclic compound, 4-hydroxy-5-phenyl-tetrahydro1,3-oxazin-2-one, as two diastereomers (CCMF); (2) further oxidation most likely by aldehyde dehydrogenase to the acid, 3-carbamoyl-2-phenylpropionic acid (CPPA); or (3) loss of carbamic acid resulting in formation of the reactive metabolite, ATPAL. Thompson et al. (4) have reported a ratio of CCMF/CBMA of 10:1, a halftime of CBMA under physiological conditions in the order of minutes, and, in the presence of GSH, a CBMA halftime that was too short to measure. Conversion of CBMA to ATPAL results in the loss of the hydrogen atom in position 2 of the CBMA propane chain, and it has been postulated that, if this hydrogen were substituted by a fluorine atom, CBMA would be stable to loss of carbamic acid and formation of the reactive ATPAL species. On the basis of this hypothesis, 2-fluoro-2-phenyl-1,3 propanediol dicarbamate (F-FBM) was synthesized and is presently undergoing development as a new antiepileptic drug with properties similar to FBM (11). The focus of this paper was to test the hypothesis that substitution of a fluorine atom for a hydrogen atom at position 2 in the FBM propane chain would eliminate conversion of F-CBMA to ATPAL. Interspecies differences between human and animal metabolism of FBM to ATPAL have been reported, indicating that animals may be of limited usefulness as an indicator of FBM toxicity in man. For example, Dieckhaus et al. (12), reported that, when rats were dosed with MCF, 84% of the compound was cleared by direct glucuronidation and, therefore, not available for further oxidation to CBMA leading to formation of the reactive metabolite, ATPAL. Since the enzyme(s) responsible for hydrolysis of FBM to MCF have not been identified and are not present in human liver postmitochondrial Suspension (S9), we decided to compare the metabolism of FBM and F-FBM intermediary metabolites, MCF, 2-fluro-2-phenyl-1,3-propanediol (FMCF), CCMF, and 4-hydroxy-5-fluoro-5-phenyl-tetrahy-
Chemicals and Materials. CCMF as two diastereomers (purity 99.84%) and F-CCMF as two diastereomers (purity, 95.1%) were purchased from Cardinal Health, Research Triangle Park, NC. MCF, F-MCF, and CPPA were provided by MedPointe Pharmaceuticals, Somerset, NJ. [Glycine-2-3H]GSH (4.5 Ci/ mmol) and Ultima-Flo M scintillation cocktail were obtained from Perkin-Elmer, Shelton, CT. All other chemicals were at least reagent grade, and solvents were HPLC grade. Pooled human liver S9 fractions from 17 human donor livers, obtained from In Vitro Technologies, Baltimore, MD, were maintained at -70 °C until used. CCMF and F-CCMF Stability in S9 Buffer. CCMF or F-CCMF, 100 mM in acetonitrile, was diluted to 500 µM with 100 mM potassium phosphate buffer (pH 7.4) with or without 2 mM reduced GSH. Aliquots of individual CCMF and F-CCMF solutions were placed in 1.5 mL HPLC sample vials. The autosampler was maintained at 37 °C, and 100 µL aliquots were injected into the HPLC every 2.8 h. Each incubation was carried out in triplicate. S9 Incubations. S9 (2 mg protein/mL) incubations were carried out in 2 mL of 100 mM potassium phosphate buffer (pH 7.4) containing 3.3 mM MgCl2, and 2 mM GSH, with or without 2 mM NAD+ cofactor. MCF, F-MCF, CCMF, or F-CCMF (100 mM) in acetonitrile were added to the incubation media to give a final concentration of 500 µM. Control samples with or without 2 mM NAD+ but no MCF, F-MCF, CCMF, or F-CCMF were also prepared. Incubations were carried out for 4 h at 37 °C on a shaking water bath. The reaction was terminated either by freezing or by the addition of 2 vol of acetonitrile. Samples were stored at -70 °C until analyzed by LC/MS. Individual samples were run in duplicate. For the radioactivity studies, S9 incubations were carried out as described for the nonradiolabeled experiments, except that 20 µCi of 3H-GSH was added to each sample, and the total incubation volume was 1 mL. As stated above, the reaction was terminated with acetonitrile, protein was precipitated, and the supernatant was removed and evaporated under N2. Residual water was evaporated on a Savant SpeedVac sample concentrator (Thermo Savant, Farmingdale, NY). The residue from a 1 mL sample was taken up in 0.5 mL of mobile phase, and 100 µL aliquots were analyzed by HPLC using both UV and radiochemical detection. HPLC Analysis for Stability Studies and Media from S9 Experiments. FBM and F-FBM intermediates and their metabolites were measured by HPLC using an Agilent model 1050 system (Agilent Technologies, Palo Alto, CA) equipped a with 1050 Series diode array detector, set at 214 nm. For the radioactive experiments, a Packard 150TR radioactivity detector (Packard Instrument Co., Meriden, CT) using a 0.5 mL flow cell and Ultima-Flo M scintillation cocktail at a 2:1 ratio was placed in series with the diode array detector. Following centrifugation at 12 000g for 10 min, 100 µL aliquots of supernatant from previously frozen S9 media or buffer from stability studies were applied to a Luna C18(2) column 250 mm × 4.6 mm, 5 µm particle size (Phenomenex, Torrance, CA). Samples were eluted on a linear gradient from 10 to 30% acetonitrile in 0.1% formic acid over 30 min, followed by a linear gradient from 30 to 70%
1844
Chem. Res. Toxicol., Vol. 18, No. 12, 2005
acetonitrile in 0.1% formic acid over 10 min at a flow rate of 0.7 mL/min. These conditions were held for 10 min, followed by reequilibration to starting values over the final 10 min of each run. Mass Spectral Analysis of Media from S9 Experiments. Aliquots (100 µL) of supernatant from S9 media were chromatographed as described above using an Alliance 2695 HPLC (Waters, Milford, MA). Compounds eluting from the column were detected on a model 2996 photodiode array detector at 214 nm (Waters, Milford, MA) coupled to a Quattro Micro mass spectrometer (Micromass, Manchester, U.K.) operating in the positive electrospray ionization mode. Eluate was diverted away from the mass spectrometer for the first 8 min of each run. Capillary and cone voltages were set at 4 kV and 20 V, respectively. Data were acquired as full scans from 100 to 700 amu. Product ion scans of ATPAL-GSH and ATPA-GSH protonated molecular ions at 440 and 456 m/z, respectively, were obtained using argon gas and a collision energy of 20 eV. Calibration Standards. Quantification of CCMF, F-CCMF, MCF, F-MCF, and CPPA was accomplished using an external standard method after preparing calibration standards by diluting 10 mM stock solutions of each compound in acetonitrile with 100 mM phosphate buffer (pH 7.4). Calibration standards were prepared in triplicate at concentrations of 0.1, 0.2, 0.5, 1.0, 2.0, 5.0, 10.0, 20.0, 50.0, and 100 µM. Aliquots (100 µL) of each solution were analyzed by LC/MS as described above. Quantification was performed by integration of the reconstructed MH+ ion at m/z 194 and 212 for CCMF and F-CCMF, respectively, and major fragment ions for MCF at m/z 117, F-MCF at m/z 133, and CPPA at m/z 149. For CCMF and F-CCMF, total area counts (sum of both diastereomers) were used for quantification. Calibration curves were constructed by fitting the calibration standard area counts to added concentrations for each standard concentration to a three-parameter power curve using SigmaPlot version 9.0 (Systat Software Inc., Richmond, CA). Errors observed in the expected values for calibration standards (n ) 3 determinations) ranged from 0.3% to 17% for MCF, 0.9% to 12.7% for F-MCF, 0.02% to 10.2% for CCMF, 0.5% to 10.2 for F-CCMF, and 0.03% to 10.7% for CPPA, for n ) 3 determinations.
Results Metabolism of the known felbamate intermediary metabolites MCF and CCMF were compared with the metabolism of the proposed F-FBM intermediary metabolites, F-MCF and F-CCMF, by S9. In addition, since it is known that CCMF is unstable at physiological pH, the stabilities of CCMF and F-CCMF were compared in S9 incubation phosphate buffer (pH 7.4) at 37 °C. CCMF and F-CCMF Stability in Phosphate Buffer. CCMF in phosphate buffer at pH 7.5 was reported to have a half-life of 4.6 h (13); however, the temperature was not specified in these studies, and no stability data is available for F-CCMF under the same conditions. Therefore, we compared the stability of CCMF and F-CCMF in S9 incubation buffer (pH 7.4), at 37 °C, in the presence or absence of GSH. The results are presented in Figure 1 as a percent of total added CCMF or F-CCMF (sum of both diastereomers) remaining over a period of 24 h. Half-lives of 2.8 and 3.6 h were calculated for CCMF in the presence and absence of GSH, respectively. Although the F-CCMF stability experiments were not carried out far enough to obtain an accurate halflife, the loss of F-CCMF following a 24 h incubation in the absence of GSH was 2.5% ( 0.7% SD, n ) 3, and 4.9% ( 2% SD, n ) 3, in the presence of GSH. LC/MS positive ion electrospray analysis of the media following incubation of CCMF revealed numerous breakdown products that were unstable over time; however, no
Parker et al.
Figure 1. Stability of CCMF (500 µM) and F-CCMF (500 µM) in phosphate buffer (pH 7.4) used for S9 incubations. Data points represent the total (sum of two diastereomers) for each compound remaining for (O) CCMF in the absence of GSH, (b) CCMF in the presence of GSH, (0) F-CCMF in the absence of GSH, and (9) in the presence of GSH. Table 1. Metabolism ((nmol/mg protein)/4 h) of CCMFand F-CCMF by S9a S9 incubation
MCF
F-MCF
CPPA
CCMF + NAD+
9.6 9.4
