Abstract
Ezlopitant is metabolized by cytochrome P450 primarily to two metabolites: a benzyl alcohol and a corresponding alkene. The alkene arises as a direct product of metabolism of ezlopitant rather than through dehydration of the benzyl alcohol. The mechanism of this cytochrome P450 (P450)-catalyzed dehydrogenation reaction was probed with five different deuterium-labeled analogs of ezlopitant. At saturating ezlopitant concentrations, deuterium substitution resulted in small differences in reaction velocity. When deuterium was incorporated into the benzylic position ([d1]ezlopitant and [d7]ezlopitant), low isotope effects on the formation of both the benzyl alcohol and alkene were observed (1.25–1.55 for CYP3A4 and 1.48–2.61 for CYP2D6), suggesting that abstraction of the benzylic hydrogen is obligatory in the formation of both metabolites. A small amount of metabolic switching occurred because isotope effects were slightly higher for alkene and alcohol formation than for ezlopitant consumption. Intramolecular deuterium isotope effects of the dehydrogenation reaction for tri- and tetradeuterated analogs were very low (1.13–1.15) for both CYP3A4 and CYP2D6, whereas intramolecular isotope effects for the chemical dehydration of correspondingly deuterated ezlopitant benzyl alcohol (CJ-12,764) were 3.8 to 5.9. Thus, dehydrogenation does not appear to occur via enzyme-mediated general acid catalysis of the benzyl alcohol. A mechanism for the dehydrogenation of ezlopitant is proposed in consideration of the data presented.
Ezlopitant, (2S,3S,4S)-2-diphenylmethyl-3-(5-isopropyl-2-methoxy benzylamino-1-azabicyclo[2.2.2]octane (Fig.1), is a novel nonpeptide antagonist of the substance P receptor. It represents a potential therapeutic agent for pathologies that involve the substance P receptor, such as inflammation, pain, and emesis. In the course of the characterization of the metabolism of this compound, it was discovered that a major metabolite in human liver microsomes (Obach, 2000) and the systemic circulation of human study subjects was a dehydrogenated analog (CJ-12,4581; Fig. 1). The other major metabolite of ezlopitant observed in vitro and in vivo was CJ-12,764, a benzyl alcohol analog (Fig. 1). In previous work, evidence was obtained that suggested that CJ-12,458 did not arise by a chemical- or enzyme-catalyzed dehydration of the benzyl alcohol (Obach, 2000). Thus, CJ-12,458 appears to arise via the direct action of cytochrome P450 (P450), and as such, ezlopitant represents another example substrate that undergoes the unusual reaction of P450-catalyzed dehydrogenation.
Examples of other P450-catalyzed dehydrogenation reactions (Ortiz de Montellano, 1995) include the formation of alkene metabolites of valproic acid (Rettie et al., 1987, 1988; Fisher et al., 1998), the dehydrogenation of the 6 and 7 positions of testosterone (Nagata et al., 1986), dehydrogenation of lauric acid (Guan et al., 1998), 2-ethylhexanoic acid (Pennanen et al., 1996), warfarin (Fasco et al., 1978), and the 3-hydroxy-3-methylglutaryl-CoA reductase inhibitors lovastatin (Vyas et al., 1990) and simvastatin (Vickers and Duncan, 1991). The metabolism of valproic acid to the Δ4(5) alkene metabolite represents a dehydrogenation reaction of an sp3-hybridized center, which is not adjacent to any sp2-hybridized carbon center. Experiments byRettie et al. (1988) used deuterated analogs of valproic acid to obtain evidence that the P450-catalyzed dehydrogenation reaction proceeded via initial hydrogen atom abstraction of the 4 position. The alkene arises by either abstraction of a second hydrogen atom by the Fe(IV)-OH intermediate and subsequent collapse of the diradical or via abstraction of an electron and deprotonation of the cationic intermediate to yield the alkene (Ortiz de Montellano, 1995). In contrast to valproic acid, the formation of the Δ6(7) alkene testosterone metabolite represents a P450-catalyzed dehydrogenation occurring at a position adjacent to an sp2-hybridized center, ultimately leading to a diene conjugated to a carbonyl group (Nagata et al., 1986). Deuterium isotope effects were also used in this case to gain mechanistic insight for this reaction (Korzekwa et al., 1990). The data supported a mechanism of sequential hydrogen atom abstraction reactions from the 6 and 7 positions.
Ezlopitant represents another type of substrate that undergoes dehydrogenation: a benzylic compound. Due to the fact that the dehydrogenation reaction represents a substantial component of ezlopitant metabolism by CYP3A4, CYP3A5, and CYP2D6 enzymes (Obach, 2000) and that formation of the corresponding benzylic alcohol represents the remaining amount of metabolism of ezlopitant by these enzymes, ezlopitant represented an appealing probe through which the mechanism of dehydrogenation of a benzylic position could be explored with deuterium isotope effects. Furthermore, the fact that the substituent that undergoes dehydrogenation in ezlopitant is symmetrical (isopropyl group) opened the possibility of the exploration of intramolecular isotope effects for this reaction. Since the rate-limiting step in the P450-catalytic process occurs before the C-H bond-breaking step, intramolecular isotope effects can provide mechanistic information in the event that isotope effects are masked by a high “commitment to catalysis” of the enzyme.
Thus, the objective of the experiments described here was to gather evidence to determine the mechanism of P450-catalyzed dehydrogenation of benzylic positions. Five deuterated analogs of ezlopitant, each targeted to address a specific mechanistic aspect, were synthesized to accomplish this objective. Furthermore, two human P450 enzymes of different evolutionary families (CYP3A4 and CYP2D6) that both catalyze the dehydrogenation of ezlopitant were compared with regard to deuterium isotope effects to determine whether the dehydrogenation mechanism was consistent among P450 enzymes or whether differences were possible among the enzymes.
Experimental Procedures
Materials.
Toluenesulfonyl chloride, hexadeuterated acetone, trideuterated methylmagnesium iodide, zinc iodide, LiAlD4, LiAlH4, NaBH3CN, NaBH(OAc)3, 4-methoxyacetophenone, diethyl carbonate, sodium metal, 4-methoxyacetic acid, dichloromethyl methyl ether, TiCl4, and 4-bromoanisole were from Aldrich Chemical Co. (Milwaukee, WI). Monodeuterated and octadeuterated 2-propanol was from Cambridge Isotope Labs (Woburn, MA). 2-(4-Methoxyphenyl)diethylmalonate was prepared from the ethyl ester of 4-methoxyphenyl acetic acid, according to a published procedure (Zvilichovsky and Fotador, 1974). Nonlabeled ezlopitant and (2S,3S,4S)-2-diphenylmethyl-3-amino-1-azabicyclo[2.2.2]octane were kindly provided by the Process Research and Development Department (Pfizer, Groton, CT). Microsomal preparations containing recombinant human CYP3A4 and CYP2D6 heterologously expressed in Sf9 cells were obtained from the Molecular Genetics Department (Pfizer). The preparations contained coexpressed cytochrome P450 NADPH oxidoreductase. All glassware used in enzyme incubations and analytical procedures was subjected to gas-phase silylation using hexamethyldisilazane according to a published method (Fenimore et al., 1976).
Synthesis of Deuterated Ezlopitant Analogs.
Monodeuterated 4-isopropyl anisole
To 6.4 ml of 2-deuteroisopropanol in 80 ml of dry pyridine was added 15.8 g of toluenesulfonyl chloride, portion-wise, while stirring. The reaction mixture was stirred overnight, after which it was poured into 500 ml of ice water. The mixture was stirred 30 min, followed by extraction with ether (200 ml). The ether was washed with 5% aqueous acetic acid until the washes were acidic. The ether was then dried (MgSO4), filtered, and evaporated under N2 until the acetic acid odor was gone to yield 8.2 g of monodeuterated isopropyl toluene sulfonate as a yellow oil. This was used in the subsequent reaction below.
To 0.93 g of Mg turnings (previously acid-washed) in 5 ml of ether, under N2, was added 4.7 ml of 4-bromoanisole in 20 ml of ether. The mixture was gently heated to effect synthesis of the Grignard intermediate over 1 h. The monodeuterated isopropyl toluene sulfonate (8.2 g) in 30 ml of ether was added to the reaction in a dropwise manner, and a precipitate was observed. The reaction was stirred for 1 h, after which it was poured onto 300 ml of ice water. NH4Cl (975 ml, 26% solution) was added; the mixture was extracted with ether (200 ml); the ether washed with 0.1 M NaOH (2 × 100 ml), dried (MgSO4), filtered, and evaporated under N2 to yield a pale yellow oil. The oil was subjected to chromatography on Silica (40 g) using hexane, and the fractions containing the monodeuterated isopropyl anisole product were pooled and evaporated under N2 to yield 920 mg of clear oil. A minor contaminant was observed on TLC and NMR, but this material was subsequently used as such in the preparation of 2-methoxy-5-(2-(2-deuteropropyl))benzaldehyde. All subsequent products using this material as a starting material were devoid of impurities.
Trideuterated 2-(4-methoxyphenyl)-2-propanol.
To a three-neck flask fitted with nitrogen line, condenser, and delivery funnel was added 50 ml of a 1 M solution of CD3MgI in ether. The flask was cooled in an ice bath, and 7.5 g of 4′-methoxyacetophenone in 40 ml of ether was slowly added from the delivery funnel. The reaction was allowed to come to ambient temperature. After stirring for 4 h, it was cooled on ice, and 50 ml of saturated NH4Cl solution was slowly added. The mixture was extracted with ether (2 × 50 ml); the organic fraction was dried (MgSO4), filtered, and evaporated under N2 to yield 7.1 g of yellow oil. This material was purified on Silica (100 g) by first washing with hexane, followed by hexane/ethyl acetate (25:1), and the product eluted with hexane/ethyl acetate (9:1). The fractions were pooled and the solvent removed in vacuo to yield 2.7 g of oil product that gave one band on TLC.
Trideuterated 4-isopropyl anisole.
To 2.7 g of trideuterated 2-(4-methoxyphenyl)-2-propanol in 80 ml dichloroethane was added 7.6 g of ZnI2, followed by 7.4 g of NaBH3CN. The reaction mixture was stirred for 4 h, after which it was filtered through celite. The solvent was removed under a stream of N2 to yield a clear oil. The oil was applied to a Silica column (18 g), and the trideuterated 4-isopropyl anisole was eluted with hexane. The solvent was evaporated under N2 to yield 0.85 g of product that gave one band on TLC.
1,1,3,3-Tetradeutero-2-(4-methoxyphenyl)-propane-1,3-diol.
A solution of 2-(4-methoxyphenyl)diethylmalonate (30 g) in ether (120 ml) was slowly added to a suspension of LiAlD4(5.0 g) in ether under N2. Ten hours later, an additional 1.5 g of LiAlD4 was added followed by an additional 0.5 g 6 h later. Four hours after the final addition of reagent, 100 ml of 10% HCl was added dropwise, while stirring in an ice bath under N2. The precipitate was dissolved in an additional 300 ml of 10% HCl, and the entire mixture was extracted with ethyl acetate. The organic fraction was washed with 5% NaHCO3, dried (MgSO4), filtered, and the solvent was removed under N2. The resulting oil (8 g) was subjected to Silica chromatography (14 g). The column was washed with hexane/ethyl acetate (1:1), and the product eluted with ethyl acetate. The fractions containing the product were pooled, and the solvent was evaporated under N2 to yield 3 g of white solid, pure by TLC.
Tetradeuterated 4-isopropyl anisole.
A solution of 1,1,3,3-tetradeutero-2-(4-methoxyphenyl)-propane-1,3-diol (3.0 g) in dry pyridine (15 ml) was added to a solution of toluenesulfonyl chloride (6.8 g) in dry pyridine (15 ml), and the reaction was stirred overnight. The reaction mixture was poured into 300 ml of ice water and extracted with ethyl acetate (300 ml). The ethyl acetate was washed with 2% HCl until the washes were acidic. The ethyl acetate fraction was dried (MgSO4), filtered, and the solvent was evaporated under a stream of nitrogen to yield 6.3 g of oil that was pure by TLC. The ditoluenesulfonate was dissolved in tetrahydrofuran (60 ml) and added to a 1 M solution of LiAlH4 in tetrahydrofuran under N2. The reaction was refluxed for 1 h, cooled, and slowly added to water (300 ml). The mixture was vacuum filtered, and the filtered solids were stirred with ethyl acetate overnight. The mixture was vacuum filtered, the filtrate was dried (MgSO4), filtered, and the solvent was removed under N2 to yield 870 mg of tetradeuterated 4-isopropyl anisole as an oil, pure by TLC.
Hexadeuterated 2-(4-methoxyphenyl)-2-propanol.
To 1.22 g of Mg turnings was slowly added 6.24 ml of 4-bromoanisole in 25 ml of ether with gentle heating to generate the Grignard intermediate. The reaction was cooled on ice, after which 3.7 ml of hexadeuterated acetone in 5 ml of ether was added, during which the reaction was observed to generate heat. After stirring for 30 min, 26% NH4Cl was added, the mixture was filtered, and the solids were washed with ether. The filtrate was dried down under N2 to generate a yellow oil. This procedure was repeated in 2-fold scale, and the materials were combined. The combined material was subjected to vacuum distillation (5 mm of Hg), and the fraction between 125–130°C was collected to yield 5 g of hexadeuterated 2-(4-methoxyphenyl)-2-propanol as a colorless oil. The product was pure by TLC and [1H]NMR.
Hexadeuterated 4-Isopropyl anisole.
To 3.4 g of hexadeuterated 2-(4-methoxyphenyl)-2-propanol in 100 ml of 1,1-dichloroethane was added 9.6 g of ZnI2 and 9.4 g of NaBH3CN. The reaction mixture was stirred for 2 h, after which it was filtered through celite. The filtrate was evaporated under N2 to obtain 2.8 g of oil. The oil was subjected to Silica chromatography using hexane. The fractions containing the product were evaporated under N2 to yield 1.0 g of pure hexadeuterated 4-isopropyl anisole as a colorless oil.
Heptadeuterated 4-isopropyl anisole.
To 12.8 ml of octadeuterated isopropanol in 80 ml of pyridine was added 31.6 g of toluenesulfonyl chloride, portion-wise, while stirring. The reaction mixture was stirred overnight, after which it was poured into 500 ml of ice water. The mixture was stirred 30 min, followed by extraction with ether (200 ml). The ether was washed with 5% aqueous acetic acid until the washes were acidic (4 × 125 ml). The ether was then dried (MgSO4), filtered, and evaporated under N2 until the acetic acid odor was gone to yield 25.8 g of heptadeuterated isopropyl toluenesulfonate as a yellow oil. This was used in the subsequent reaction below.
To 1.86 g of Mg turnings in 5 ml of ether under N2 was added 9.4 ml of 4-bromoanisole in 40 ml of ether. The mixture was gently heated to trigger formation of the Grignard intermediate for 1 h. The heptadeuterated isopropyl toluenesulfonate (16.5 g) in 60 ml of ether was added to the reaction in a dropwise manner, and the reaction was stirred. After 1 h, it was poured onto 600 ml of ice water. NH4Cl (150 ml, 26% solution) was added, the mixture was extracted with ether (200 ml), and the ether washed with 0.1 M NaOH (2 × 100 ml), dried (MgSO4), filtered, and evaporated under N2 to yield a pale yellow oil. The product was purified by chromatography on Silica (80 g) using hexane to yield 1.0 g of clear oil.
Preparation of deuterated 2-methoxy-5-(2-propyl)benzaldehydes.
A similar method was used to prepare all five deuterated analogs of 2-methoxy-5-(2-propyl)benzaldehydes. As an example, 850 mg of deuterated 4-isopropyl anisole was dissolved in 11 ml of CH2Cl2 in a 50 ml flask fitted with a rubber septum. Nitrogen was flushed through the flask through a bubbler containing silicon oil. The flask was cooled on an ice/salt bath. TiCl4 (0.85 ml) was slowly and cautiously added to the stirring reaction solution using a glass syringe through the rubber septum. The reaction was allowed to stir for 45 min on the ice/salt bath, after which dichloromethyl methyl ether (0.48 ml) was slowly added using a glass syringe through the septum. The reaction was stirred for 2 h, after which the system was opened and the solution was slowly poured into 100 ml of saturated NaHCO3. The mixture was vacuum filtered through celite, and the material washed with CH2Cl2. The organic portion of the filtrate was dried (MgSO4), filtered, and the solvent was removed to yield a deep blue oil. The oil was applied to a Silica column (8 g) and eluted with hexane/ethyl acetate (40:1). The solvent was removed to yield 450 mg of a clear oil, pure by TLC. [1H]NMR spectra were consistent with the addition of the carbonyl group to the position ortho to the methoxy group.
Preparation of deuterated ezlopitant analogs.
Deuterated ezlopitant analogs were all prepared by reductive amination of the corresponding benzaldehydes with NaBH(OAc)3 and (2S,3S,4S)-2-diphenylmethyl-3-amino-1-azabicyclo[2.2.2]octane. A typical procedure is as follows. 2-Methoxy-5-isopropylbenzaldehyde (450 mg) and (2S,3S,4S)-2-diphenylmethyl-3-amino-1-azabicyclo [2.2.2]octane (382 mg) were condensed by refluxing in toluene (80 ml) overnight in a Dean-Stark apparatus. The solvent was removed in vacuo, and the oily residue was dissolved in 1,1-dichloroethane (40 ml). NaBH(OAc)3 (725 mg) was added, and the reaction mixture was stirred overnight at ambient temperature. The solvent was removed under a stream of N2, the residue was taken up in water (50 ml), NaOH was added (20 ml, 1 M), and the aqueous mixture was extracted with ether (2 × 100 ml). The organic fraction was back-extracted with 50 ml of 1 M HCl; the aqueous fraction was rebasified with 1 M NaOH and re-extracted with ether. The ether fraction was dried (MgSO4), filtered, and the solvent was removed under N2 to yield 460 mg of white crystals. The deuterated ezlopitant products were pure by TLC, and structure was confirmed by [1H]NMR and HPLC-MS (APCI interface; positive ion mode). The extent of deuterium incorporation for each of the analogs is as follows: [d1 ]ezlopitant, >99.5%; [d3 ]ezlopitant, 98.2%; [d4 ]ezlopitant, 99.4%; [d6 ]ezlopitant, 96.7%; [d7 ]ezlopitant, 93.0% with 6.1% as pentadeuterated contaminant. No observable contamination by [d0 ]ezlopitant was found in any of the analogs.
Biosynthesis of Deuterated CJ-12,764 Analogs.
The synthesis of [d3 ]CJ-12,764 and [d4 ]CJ-12,764 (benzyl alcohol metabolite of ezlopitant) was accomplished using recombinant CYP3A4. Deuterated ezlopitant (50 μM) was incubated with CYP3A4 microsomes (0.5 mg/ml), NADPH (1.9 mM), and MgCl2 (3.3 mM) in a total volume of 25 ml of KH2PO4, pH 7.5, in a 125-ml flask open to the atmosphere in a shaking water bath set at 37°C. The reaction was carried out for 60 min, followed by the addition of 0.1 ml of NaOH (10 M). The mixture was extracted with methyl t-butyl ether (30 ml), the organic fraction evaporated under N2, and the residue was reconstituted in 0.1 ml of H2O/CH3CN (1:1). The deuterated CJ-12,764 was purified by reverse phase HPLC containing a Waters Symmetry C18 column (3.9 × 150 mm; Milford, MA) using an injection volume of 85 μl. The initial mobile phase composition was 36.5% CH3CN in 20 mM HOAc, pH 4 (with NH4OH), at a flow rate of 0.8 ml/min. This composition was held for 5 min, followed by a linear gradient to 63.5% CH3CN at 10 min, and a second gradient to 95% CH3CN at 20 min. Fractions (1 min) were collected into vials containing 0.05 ml of 1 M NaHCO3 and analyzed for product. CJ-12,764 eluted in the sixth fraction (5–6 min) while CJ-12,458 and ezlopitant eluted at 10 and 11 to 14 min. A dihydroxy metabolite eluted just after the void volume. To the CJ-12,764 fractions was added 0.01 ml of NaOH (1 M), which was extracted with methyl t-butyl ether (4 ml). The organic fraction was evaporated under N2and reconstituted in CH3CN for use in chemical dehydration experiments.
Deuterium Isotope Effect Experiments.
Before conducting isotope effect experiments, incubation conditions were established to ensure initial rate linearity. Initially, enzyme kinetics for each of the deuterated analogs were examined in triplicate to determine whether isotope effects were onVmax or onVmax/KM. For measurement of isotope effects, unlabeled ezlopitant and the five deuterated analogs of ezlopitant were incubated with Sf9 cell microsomes containing heterologously expressed human CYP3A4 (0.052 nmol of P450/mg of protein; 0.5 mg of protein/ml incubation) or CYP2D6 (0.012 nmol/mg of protein; 0.4 mg of protein/ml incubation), in a total volume of 0.2 ml of 25 mM KH2PO4, pH 7.5, containing 3.3 mM MgCl2 and 1.3 mM NADPH (n= 6). The substrate concentrations were 50 μM for CYP3A4 and 10 μM for CYP2D6. After a 2-min preincubation period at 37°C in which all components were mixed except for NADPH, the reactions were commenced by the addition of the cofactor and shaken in a 37°C water bath open to the air. The reactions were terminated after 5 min by vortex mixing with 3 ml of methyl t-butyl ether. Analysis was done by HPLC-MS, as described below.
To determine the isotope effects on overall ezlopitant consumption, incubations were conducted using 1.0 mg of protein/ml for CYP3A4 and 1.6 mg of protein/ml for CYP2D6. Ezlopitant substrate concentrations were 0.42 μM for CYP3A4 and 0.042 μM for CYP2D6. Incubations (n = 6) were carried out for 20 min, followed by termination and analysis as before. Ezlopitant concentrations were compared with identical incubations, which lacked NADPH (n = 6).
Deuterium Isotope Effects on the Chemical Dehydration of CJ-12,764.
Solutions of [d3 ]CJ-12,764 or [d4 ]CJ-12,764 (1.25 μg/ml) were prepared in CH3CN containing FeCl3 (1 mM) or trifluoroacetic acid (0.01%, v/v). Initial experiments using nonlabeled CJ-12,764 demonstrated that the rate of dehydration to CJ-12,458 was linear for greater than 16 h. The incubations containing [d3 ]CJ-12,764 or [d4 ]CJ-12,764 (n = 3) were allowed to react for 16 h (in the dark), followed by direct injection (10 μl) onto the analytical HPLC system described below.
Analysis of Ezlopitant, CJ-12,458, and CJ-12,764.
Measurement of concentrations of ezlopitant, CJ-12,458 and CJ-12,764, and their deuterated counterparts was accomplished by liquid extraction followed by HPLC-MS analysis. To terminated incubation mixtures (already containing 3 ml of methyl tertiary butyl ether) was added 100 ng of internal standard CJ-11,957 in 0.1 ml of water (CJ-11,957 is identical to ezlopitant with the exception that the isopropyl group in the latter is replaced with an ethyl group). The mixtures were vortex mixed for 1 min, followed by separation of the layers by spinning at 3000 rpm at ambient temperature in a Jouan model CT422 swinging bucket tabletop centrifuge (Jouan, Inc., Winchester, VA). The samples were placed in a dry-ice acetone bath to effect freezing of the aqueous layer, and the organic layer was decanted into a fresh, silylated glass test tube. The solvent was removed under N2 at 30°C in a Zymark TurboVap (Zymark Corporation, Hopkinton, MA), and the residue reconstituted in 0.1 ml of HPLC mobile phase.
The HPLC-MS system consisted of a Hewlett-Packard 1100 HPLC system (Palo Alto, CA) coupled to a PE Sciex API 100 single quadrupole mass spectrometer (Toronto, ON, Canada) containing an APCI interface. The column was a Waters Symmetry C18 (3.9 × 150 mm; 5-μm particle size packing), and the initial mobile phase consisted of 45.5% CH3CN in 20 mM acetic acid, adjusted to pH 4 with NH4OH at a flow rate of 0.8 ml/min. Samples (75 μl) were injected, and the initial mobile phase composition was maintained for 2 min, after which a linear gradient was applied resulting in 95% CH3CN at 6 min.
The entire flow was introduced into the APCI source operated in the positive ion mode. The orifice voltage was 45 V, and the nebulizer temperature was set at 500°C. For each analyte, protonated molecular ions were followed (m/z 471.2 for CJ-12,764;m/z 453.2 for CJ-12,458;m/z 455.2 for ezlopitant;m/z 441.0 for CJ-11,957 internal standard) with a dwell time for each ion of 150 ms (corresponding ions for deuterated analogs were followed, as appropriate). The retention times were 1.8 min for CJ-12,764, 5.2 min for CJ-12,458, 5.6 min for ezlopitant, and 5.0 min for CJ-11,957 ISTD. Quantitation was accomplished by extrapolation from a standard curve with linear dynamic range from 0.1 to 100 ng/ml (1/x weighting). Quantitation of deuterated analytes was done after correction for natural isotope abundance, when appropriate.
Results
Synthesis of Deuterated Ezlopitant Analogs.
A general scheme of preparing variously deuterated 4-isopropyl anisoles (Fig. 2) followed by formylation and reductive amination (Fig. 3) was successful in generating deuterated analogs of ezlopitant. The use of deuterium-labeled precursors of high isotopic purity resulted in high isotopic purity of the 4-isopropyl anisoles, as assessed via [1H]NMR spectra (Table1). By [1H]NMR, contamination with undesired hydrogen atoms was not detectable. Deuterated ezlopitant analogs were also of high isotopic purity, as assessed through mass spectrometry. Again, contamination with hydrogen was not detectable. [1H]NMR spectra for ezlopitant are very complicated; therefore, the extent of deuterium incorporation assessed by NMR was better accomplished by examination of the spectra of the 4-isopropyl anisole and 5-isopropyl-2-methoxybenzaldehyde precursors.
Deuterium Isotope Effects on the Formation of CJ-12,458 and CJ-12,764 Catalyzed by CYP3A4 and CYP2D6.
In initial experiments, the enzyme kinetics were determined for both CYP3A4 and CYP2D6 for ezlopitant and all five deuterated analogs. For each enzyme, the KM values were nearly identical regardless of the positions and extent of deuteration, but reaction velocities differed for the different isotopically labeled analogs. KM values ranged between 0.8 and 1.4 μM and 8.2 and 11 μM for CYP2D6 and CYP3A4, respectively. Thus, deuterium isotope effects were equivalent forVmax andVmax/KM, and subsequent isotope effect measurements were conducted at saturating substrate concentrations. Mean deuterium isotope effects for all five deuterated ezlopitant analogs are listed in Table2. For CYP3A4, deuterium isotope effects of 1.25 to 1.55 were observed for the formation of both CJ-12,458 and CJ-12,764 when the benzylic hydrogen was replaced with deuterium (i.e., [d1 ]ezlopitant and [d7 ]ezlopitant). When the benzylic position possessed a hydrogen, but the other positions of the isopropyl substituent were variously substituted with deuterium, isotope effects were barely detected for formation of the alkene (1.06–1.12), and small inverse isotope effects were observed for the formation of the benzylic alcohol (0.84–0.87; Table 2).
For CYP2D6, a similar pattern of isotope effects was observed, but the magnitude differed. For analogs possessing deuterium at the benzylic position, isotope effects of 1.48 to 2.61 were observed. Barely detectable effects were observed for alkene formation (1.07–1.12) when the other isopropyl positions were deuterated (Table 2). As with CYP3A4, small inverse isotope effects were observed for benzylic hydroxylation (0.78–0.90).
The ratios of benzyl alcohol to alkene formed by CYP3A4 and CYP2D6 for ezlopitant and all five deuterated analogs are listed in Table3. For CYP3A4, the ratio was 1.08 (±0.04) for nondeuterated ezlopitant. When the benzylic position was deuterated but the other isopropyl carbons possessed only hydrogen atoms ([d1 ]ezlopitant), the alcohol to alkene ratio was fairly similar at 0.94 ± 0.01. When the 1° carbons had 3, 4, or 6 hydrogens replaced with deuterium, the alcohol became somewhat favored over the alkene with ratios ranging from 1.32 to 1.43. Increased deuterium incorporation (from three to seven deuterium atoms) did not yield greater ratios. The corresponding ratio of alcohol to alkene for unlabeled ezlopitant was 1.27 (±0.03) for CYP2D6. Ratios for [d1 ]ezlopitant, [d3 ]ezlopitant, [d4 ]ezlopitant, and [d6 ]ezlopitant were similar and ranged between 1.58 and 1.75, again with no trend toward higher ratios with increasing deuterium incorporation. For CYP2D6, the ratio of alcohol to alkene formation was 2.12 (±0.09) for heptadeuterated ezlopitant.
Deuterium Isotope Effects on the Overall Consumption of Ezlopitant Catalyzed by CYP3A4 and CYP2D6.
In P450-catalyzed reactions, incorporation of deuterium can result in the switching of the position of metabolism to a distant site. To measure this, the rate of consumption of ezlopitant and its deuterated analogs at [S] < KM were determined and compared (it should be noted that since no effect of deuteration onKM had been observed, measurement of reaction velocity at [S] < KM provides a meaningful estimate of isotope effect). The overall rates of metabolism of ezlopitant and deuterated analogs were measured at a substrate concentration well below KM. For CYP3A4, isotope effects on overall consumption of ezlopitant were not observed (Table 4). However for CYP2D6, overall isotope effects of 1.39 and 1.53 were observed for [d1 ]ezlopitant and [d7 ]ezlopitant, respectively, representing both compounds that possess deuterium substitution at the benzylic position. When the benzylic position possessed a hydrogen ([d3 ]ezlopitant, [d4 ]ezlopitant, and [d6 ]ezlopitant), no isotope effects on overall consumption of ezlopitant were observed (Table 4).
Intramolecular Isotope Effects of CJ-12,458 Formation.
Intramolecular isotope effects were measured for dehydrogenation by examining the ratio of [d3 ]CJ-12,458 versus [d2 ]CJ-12,458 formed from [d3 ]ezlopitant and the ratio of [d4 ]CJ-12,458 versus [d3 ]CJ-12,458 formed from [d4 ]ezlopitant. Isotope effects were very low at 1.13 to 1.15 for each deuterated substrate for both P450 isoforms (Table 5).
Isotope Effects on Chemical Dehydration of CJ-12,764 (Benzyl Alcohol) to CJ-12,458 (Alkene).
In the chemical dehydration experiments, no detectable CJ-12,458 was observed when CJ-12,764 was incubated in CH3CN (control). In incubations containing acetonitrile solutions of FeCl3 and TFA, dehydration of the alcohol to the alkene was observed. For [d3 ]CJ-12,764, intrinsic isotope effects of 5.90 and 4.47 were observed for the FeCl3- and TFA-catalyzed dehydration reactions, respectively (Table 6). For [d4 ]CJ-12,764, the isotope effects were 4.51 and 3.79, respectively.
Discussion
The use of deuterium-containing analogs of substrates in the measurement of deuterium isotope effects on reaction rates/kinetics is a standard approach in the examination of mechanisms of reactions in organic and bio-organic chemistry. Such an approach has been extensively and successfully applied to the P450 enzymes to understand the mechanisms of aliphatic hydroxylation and heteroatom dealkylation reactions (Ortiz de Montellano, 1995). An interesting aspect of the application of deuterium isotope effects to P450-catalyzed reactions resides in the fact that the enzyme exhibits a high “commitment to catalysis”. That is, in the multistep reaction cycle for many of these enzymes, a step or steps before the isotopically sensitive step is essentially rate limiting and irreversible. This can result in the masking of isotope effects; little change is observed in the overall consumption of substrate when hydrogen is replaced with deuterium despite there being a substantial slowing down of the C-H bond-breaking step. Many examinations of deuterium isotope effects of P450-catalyzed reactions have relied upon the observation of intramolecular isotope effects. In this approach, identical, chemically equivalent positions simultaneously possess both deuterium and hydrogen atoms, and the ratio of product possessing deuterium versus that containing hydrogen is measured. This results in an “unmasking” of the isotope effect and generates useful data for mechanistic interpretation. A second possibility in the application of isotope effects to study P450-catalyzed reactions is the observation of metabolic switching. In this case, replacement of hydrogen atoms with deuterium atoms results in a diminished rate of metabolism at the site where the deuterium atoms are present and a concomitant acceleration of the rate of metabolism at chemically unequivalent sites in which hydrogen atoms were maintained with little or no overall change in the rate of substrate consumption.
In addition to its use in understanding the mechanisms of the common P450-catalyzed hydroxylation and heteroatom dealkylation reactions, the approach of measuring deuterium isotope effects has also been applied to the study of P450-catalyzed dehydrogenation reactions. Two noteworthy deuterium isotope effect studies have been conducted on the P450-catalyzed dehydrogenation of valproic acid (Rettie et al., 1988) and testosterone (Korzekwa et al., 1990). In the former case, the dehydrogenation reaction occurs at an sp3-hybridized center of an alkane chain and is catalyzed by liver microsomes of phenobarbital-induced rabbits. The intramolecular isotope effect observed for the dehydrogenation reaction was 5.58, which was nearly identical to that observed for the corresponding hydroxylation reaction (5.05). The authors used this as evidence that the two products arise from a common radical intermediate. The enzyme bound radical intermediate can partition between two subsequent reactions: rebound of the hydroxyl radical to yield the alcohol or abstraction of a second hydrogen atom followed by collapse of the diradical to form the alkene. For testosterone, the CYP2A1-catalyzed dehydrogenation occurs at a position adjacent to an sp2-hybridized center. Measurement of intrinsic isotope effects supported a mechanism of formation of a common radical intermediate that partitions between formation of the alcohol and the alkene (Korzekwa et al., 1990).
Unlike valproic acid and testosterone, the dehydrogenation of ezlopitant occurs at a benzylic position. The possibility that the alcohol metabolite gives rise to the alkene via either a spontaneous dehydration or dehydration catalyzed by an unknown component in the incubation mixtures was considered and explored. Several lines of direct and indirect evidence had been previously obtained that support the notion that CJ-12,458 does not arise via a simple dehydration of CJ-12,764: 1) no CJ-12,458 was observed when radiolabeled CJ-12,764 was incubated under conditions in which ezlopitant gave rise to CJ-12,458 (Obach, 2000); 2) no CJ-12,458 was observed when CJ-12,764 was incubated under conditions in which ezlopitant gave rise to CJ-12,458 using a sensitive HPLC-MS assay; 3) in other metabolic systems, including liver microsomes from several preclinical species and heterologously expressed recombinant CYP3A5, the ratio of CJ-12,764 to CJ-12,458 was markedly different from metabolite ratios observed for CYP3A4, CYP2D6, and human liver microsomes; and 4) incubation of CJ-12,764 and CJ-12,458 with human liver microsomes yields distinctly different product profiles. Thus, the conversion of ezlopitant to CJ-12,458 appears to be another example of direct P450-catalyzed dehydrogenation of a saturated alkane.
For ezlopitant, isotope effects were observed onVmax, which is common for P450 reactions since the rate-limiting step in the catalytic cycle for many of these enzymes occurs before the first C-H bond-breaking step. The low isotope effects for the metabolism of ezlopitant to the alkene suggest that there is a very high commitment to catalysis for this reaction. Only small isotope effects were observed on the overall consumption of ezlopitant, which is typical for P450 enzymes (Table 4). However, there was also little effect on the formation of the two metabolites (Table2) indicating that metabolic switching did not occur to a great extent. When the benzylic position possessed a deuterium atom ([d1 ]ezlopitant and [d7 ]ezlopitant), isotope effects were observed for formation of both the alcohol and alkene. However, when the benzylic position possessed a hydrogen atom while the adjacent positions possessed three or more deuterium atoms, isotope effects were not observed for alkene formation, with small inverse isotope effects observed for benzyl alcohol formation. These data support the notion that the formation of the alkene requires initial abstraction of the benzylic hydrogen. Thus, the benzylic radical is an intermediate that partitions between formation of the benzylic alcohol and alkene. This is also what has been proposed for the alkene formation reactions for valproic acid and testosterone (Rettie et al., 1988; Korzekwa et al., 1990) and is consistent with the accepted hypothesis of P450 mechanism.
The P450-catalyzed dehydrogenation of ezlopitant exhibits almost no intramolecular isotope effect when chemically equivalent positions are partially deuterium substituted (i.e., [d3 ]ezlopitant and [d4 ]ezlopitant; Table 5). In Fig.4, possible mechanisms for ezlopitant dehydrogenation are depicted in consideration of these data. All require the initial abstraction of the benzylic hydrogen to yield a common benzylic radical intermediate. In pathway A, the Fe(IV)-OH intermediate abstracts a second hydrogen atom from the terminal position, water is released, and the diradical collapses to yield the alkene. In pathway B, the benzylic alcohol is formed but undergoes dehydration to the alkene (possibly via a carbocation intermediate), which must occur before release of the alcohol (since benzyl alcohol + P450 does not yield alkene). The dehydration could be catalyzed by the heme Fe(III) or by an amino acid in the active site acting as a general acid catalyst. These possibilities were tested by measuring the intrinsic isotope effect of dehydration of [d3 ]CJ-12,764 and [d4 ]CJ-12,764 catalyzed by both Lewis (FeCl3) and protic (TFA) acids. These isotope effects were substantially higher than those observed for P450-catalyzed CJ-12,458 formation (Tables 5 and 6), which is evidence against a general acid-catalyzed dehydration of CJ-12,764. Furthermore, incubation of ezlopitant with CYP3A4 and CYP2D6 in [18O]H2O did not result in any incorporation of oxygen-18 into CJ-12,764, indicating that a benzylic carbocation is not formed as an intermediate in the metabolism of ezlopitant (data not shown).
The mechanism in pathway A is that which has been proposed for valproic acid (Rettie et al., 1988). The data obtained with [d3 ]ezlopitant, a very small isotope effect of 1.15, is similar to that observed for the CYP2B1-catalyzed dehydrogenation of [5,5,5-2H3]valproic acid (Rettie et al., 1995). However, the distinction of [d3 ]ezlopitant versus [5,5,5-2H3]valproic acid is that the competition between hydrogen versus deuterium for the second H-atom abstraction would occur between exactly equivalent positions that are directly bonded to the carbon from which the initial hydrogen was abstracted. The very low isotope effect for [d3 ]ezlopitant is consistent with pathways A or C, and with the previous observations for valproic acid.
A mechanism is proposed in pathway C, which is consistent with but not unequivocally proven by the data. In this proposal, the benzylic radical is partially stabilized by the p-methoxy substituent leading to a delocalization of the electron density throughout the aromatic ring and isopropyl group, resulting in restricted rotation around the benzylic bond. The P450 Fe(IV)-OH intermediate could then abstract a hydrogen atom from the terminal position but due to the restricted rotation could not selectively abstract a hydrogen atom over a deuterium atom in [d3 ]ezlopitant. Thus, intrinsic isotope effects would be masked for the trideuterated analog. However, no intramolecular isotope effects were observed for the [d4 ]ezlopitant analog either. In this case, one would expect that the C-C bond that would orient H or D to the active Fe(IV)-OH center of the enzyme should be able to freely rotate and not suppress intrinsic isotope effects. The observation of virtually no intramolecular isotope effects for the dehydrogenation of [d4 ]ezlopitant suggests one of two possibilities. One possibility is that abstraction of the second hydrogen atom from the terminal position may not be rate limiting (i.e., the isotope effect is masked by a very high rate of abstraction of the second hydrogen atom relative to the rate of rotation of the C-C bond). This could be due to a high drive to re-aromatize the ring, which would be effected by abstraction of the second hydrogen atom. A second possibility is that the restricted rotation proposed for the benzylic bond due to resonance stabilization from thep-methoxy group could extend to the adjacent C-C bond, thereby reducing selectivity for removal of a hydrogen atom versus a deuterium atom (as in pathway C of Fig. 4). This proposed mechanism requires an important role for the p-methoxy substituent, and further testing of this hypothesis requires the examination of the P450-catalyzed metabolism of close-in analogs that possess electronically different substituents at this position (i.e., measurement of the propensity of other analogs to undergo dehydrogenation). However, if the low intramolecular isotope effect observed for [d3 ] and [d4 ] is merely due to a rapid rate/high commitment to catalysis for the second hydrogen atom abstraction, then the data do not distinguish between pathways A and C.
In conclusion, the P450-catalyzed dehydrogenation of ezlopitant represents a mechanistically interesting reaction. The isotope effect data obtained appear to distinguish this reaction from other P450-catalyzed dehydrogenation reactions. Whether, this distinction is due to the fact that the dehydrogenation of ezlopitant occurs at a benzylic position or whether this is due to some unique characteristic of this particular substrate remains to be determined.
Acknowledgments
The author thanks Kevin Shenk and Harry R. Howard for technical advice on the formylation of isopropyl anisole and Drs. Deepak Dalvie and Jae Lee of Pfizer for discussions and advice concerning this work. Special thanks to Dr. Alfin Vaz for extended discussions of the interpretation of these data.
Footnotes
- Abbreviations used are::
- CJ-12,458
- ezlopitant alkene
- CJ-12,764
- ezlopitant benzyl alcohol
- P450
- cytochrome P450
- TLC
- thin layer chromatography
- HPLC-MS
- high-pressure liquid chromatography-mass spectrometry
- APCI
- atmospheric pressure chemical ionization
- Received July 19, 2001.
- Accepted September 20, 2001.
- The American Society for Pharmacology and Experimental Therapeutics