Inhibitor: Dehydrogenation and Inactivation of CYP3A4 - American

Dec 21, 2007 - TNF-r Inhibitor: Dehydrogenation and Inactivation of CYP3A4. Hao Sun and Garold S. Yost*. Department of Pharmacology and Toxicology, ...
0 downloads 0 Views 667KB Size
374

Chem. Res. Toxicol. 2008, 21, 374–385

Metabolic Activation of a Novel 3-Substituted Indole-Containing TNF-r Inhibitor: Dehydrogenation and Inactivation of CYP3A4 Hao Sun and Garold S. Yost* Department of Pharmacology and Toxicology, UniVersity of Utah, 30 South 2000 East, Room 201, Salt Lake City, Utah 84112-5820 ReceiVed August 18, 2007

SPD-304 is a recently discovered small-molecule TNF-R antagonist. However, SPD-304 contains a potentially toxic 3-alkylindole moiety. Previous studies on 3-methylindole and the 3-alkylindole-containing drugs zafirlukast and MK-0524 structural analogues found that they were bioactivated by cytochrome P450s through a dehydrogenation process to form 3-methyleneindolenine intermediates that are electrophilic R,β-unsaturated iminium species. These electrophiles could react with protein and/or DNA nucleophilic residues to cause toxicities. In the present study, we found that SPD-304 was bioactivated through a similar dehydrogenation mechanism to produce a similar electrophilic 3-methyleneindolenine intermediate. The electrophile was trapped with nucleophilic glutathione and identified by LC/MS/MS. The iminium or another reactive intermediate also was a mechanism-based inactivator of CYP3A4. The inactivation parameters were KI ) 29 µM and kinact ) 0.047 min-1. In addition, SPD-304 was metabolized through hydroxylation, N-dealkylation, and epoxidation pathways, and several metabolites and glutathione adducts were characterized by tandem mass spectrometry. The metabolism profile was also evaluated by in silico molecular docking of SPD-304 into the active site of CYP3A4, which predicted that the dehydrogenation reaction was initiated by 3-methylene C-H atom abstraction at the trifluoromethylphenyl-1H-indol-3ylmethyl portion of SPD-304. Hydroxylation of the 6′-methyl of the dimethylchromone portion of SPD304 was the other major predicted metabolic pathway. The molecular models correlated precisely with experimental metabolic results. In summary, dehydrogenation of SPD-304 may cause toxicities through the formation of electrophilic intermediates and cause drug-drug interactions through CYP3A4 inactivation. Introduction Tumor necrosis factor-R (TNF-R)1 plays an insidious role in the pathogenesis of rheumatoid arthritis (1). The recently introduced anti-TNF drugs etanercept, infliximab, and adalimumab are injected proteins, administered intravenously or subcutaneously (2). They block the interaction of TNF-R with the TNF receptor and hence the downstream pro-inflammatory cascade. Although the success rates of these anti-TNF therapies are close to 60% in clinical trials (1), developing a parenterally dosed small-molecule TNF-R antagonist remains a vital objective for the pharmaceutical industry. In a recent effort to identify small-molecule TNF-R inhibitors using fragment-based screening, SPD-304 was discovered to be a promising TNF-R antagonist that promotes the dissociation of TNF-R trimers and, therefore, blocks the interaction of TNF-R and its receptor (3). SPD-304 is 6,7-dimethyl-3-{[methyl-(2-{methyl-[1-(3-trifluoromethyl-phenyl)-1H-indol-3-ylmethyl]amino}ethyl)amino]methyl}chromen-4-one, which contains a 3-substituted indole moiety (Figure 1). Previous studies on 3-methylindole, which is produced in animal and human digestive systems through tryptophan degradation and is found in cigarette smoke, demonstrated that it is a potent pneumotoxin that can cause severe lung injury (4, 5). 3-Methylindole is bioactivated by * To whom correspondence should be addressed. Tel: 801-581-7956. Fax: 801-585-3945. E-mail: [email protected]. 1 Abbreviations: CID, collision-induced dissociation; ESI, electrospray ionization; GSH, glutathione; NADPH, nicotinamide adenine dinucleotide phosphate, reduced form; P450, cytochrome P450; TNF-R, tumor necrosis factor-R.

Figure 1. Structures of 3-methylindole (A) and its structural analogues zafirlukast (B), MK-0524 (C), and SPD-304 (D). 3-Substituted indole moieties within each chemical are emphasized with a box.

cytochrome P450s (P450s) to form a reactive electrophilic intermediate through the dehydrogenation pathway, which can subsequently react with protein and/or DNA nucleophilic residues to cause toxicities (6–8). 3-Alkylindole-containing drugs (see Figure 1 for structures) such as zafirlukast (9) and structural analogues of MK-0524 (10, 11) were also bioactivated by P450s through a similar dehydrogenation pathway. Zafirlukast, a leukotriene receptor antagonist for mild to moderate asthma, has been reported to cause hepatotoxicity in some patients (12). MK-0524 was recently identified as a prostaglandin D2 receptor antagonist for the treatment of seasonal allergic rhinitis and niacin-induced flushing (10, 11). In addition, 3-methylindole and zafirlukast inactivate the P450 enzymes that catalyze their dehydrogenation, presumably by mechanism-based inactivation

10.1021/tx700294g CCC: $40.75  2008 American Chemical Society Published on Web 12/21/2007

P450-Catalyzed Dehydrogenation of a TNF-R Inhibitor

by the 3-methyleneindolenine intermediates. Structural analogues of MK-0524 were efficiently dehydrogenated to electrophilic methyleneindolenine intermediates that were trapped with glutathione (GSH), bound covalently to protein, and inactivated their P450 catalysts via suicide mechanisms. However, MK-0524 itself was not a suitable substrate for dehydrogenation, because the structure incorporated electron-withdrawing substituents that presumably decreased the electron density of the indole enough to destabilize the electron-deficient methyleneindolenine intermediates. Considering the structural similarities of SPD-304 with 3-methylindole, zafirlukast and MK-0524 (Figure 1), we speculated that SPD-304 could also be dehydrogenated by P450s and is highly likely to cause toxicities and inactivate P450 enzymes. The formation of the electrophilic 3-methyleneindolenine, the cause of toxicities from P450-catalyzed dehydrogenation mechanism, was proposed to be a two-step process, in which dehydrogenation was initiated by either hydrogen atom abstraction from the 3-methyl or 3-methylene carbon, or electron abstraction from the indole nitrogen atom, and followed by a second one-electron oxidation (7, 9, 11). The hydrogen atom abstraction step was believed to be the rate-limiting step of 3-methylindole dehydrogenation, but either hydrogen atom abstraction or nitrogen oxidation was the potential initiating step for zafirlukast and analogues of MK-0524 dehydrogenation. SPD-304 contains a structural difference from zafirlukast and MK-0524 analogues that could change its dehydrogenation profile. The atom connected to the 3-methylene group of SPD304 is a nitrogen atom, but it is a carbon in both zafirlukast and MK-0524 analogues (Figure 1). Consequently, the electrondonating effect of an amine group in SPD-304 would facilitate hydrogen atom abstraction at the 3-methylene carbon. Another metabolic difference of SPD-304 from the other three is that the 3-methyleneindolenine electrophilic intermediate can add water to form reasonably stable alcohols, such as indole-3carbinol from 3-methylindole (7) and the secondary alcohols from zafirlukast and MK-0524 analogues (9, 11) However, the methyleneindolenine of SPD-304 would hydrate to form a carbinolamine that would be unstable and would eliminate the adjacent nitrogen to form an amine with concomitant formation of an aldehyde at the 3-indolyl position. The goals of these studies were to confirm the hypothesis that SPD-304 is efficiently dehydrogenated by P450 enzyme(s) to an electrophilic 3-methyleneindolenine that could cause toxicities and/or inactivate P450 enzyme(s) to produce adverse drug-drug interactions.

Experimental Procedures Materials. SPD-304 was purchased from Calbiochem (San Diego, CA). Nicotinamide adenine dinucleotide phosphate, reduced form (NADPH), GSH, ammonium acetate, and formic acid were obtained from Sigma-Aldrich, Inc. (St. Louis, MO). Solvents used for HPLC and LC/MS analysis were HPLC grade. Incubations with Human Liver Microsomes and cDNAExpressed Recombinant P450s. Pooled human liver microsomes (20 mg/mL protein) as well as individual P450s including CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4 that contain both P450 reductase and cytochrome b5 were obtained from BD Biosciences (San Jose, CA). Each incubation contained SPD-304 (∼200 µM), human liver microsomes, or individual P450s (25 pmol of P450), NADPH (2 mM), and potassium phosphate buffer (0.1 M, pH 7.4) in a final volume of 100 µL. The reaction was initiated by the addition of NADPH. The mixtures were incubated in a 37 °C shaking water bath for 30 min and then terminated by adding 100 µL of ice-cold

Chem. Res. Toxicol., Vol. 21, No. 2, 2008 375 acetonitrile, followed by centrifugation at 21000g for 30 min, with the supernatant collected and analyzed by LC/MS. Incubations without NADPH were used as the negative controls. Incubations with GSH. To characterize the reactive electrophilic intermediates that were formed, GSH (4 mM) was added to the incubations containing either pooled human liver microsomes or CYP3A4 (25 pmol of P450), with the same sample conditions as mentioned above, except that the incubation time was 45 min. The supernatant collected after centrifugation was further concentrated to around 50 µL using a constant nitrogen gas flow at room temperature, which was the final preparation before analysis by LC/MS. Identification of Metabolites by LC/MS, LC/MS/MS, and LC/MS3. The mass spectra of all NADPH-dependent metabolites from incubations with human liver microsomes and individual P450s were determined by LC/MS analysis using a Finnigan LCQ LC/MS system (Thermo Electron Corp., Waltham, MA). Collected incubation samples after centrifugation were chromatographed over a Phenomenex Luna 5 µm C18 (150 mm × 2.00 mm, 5 µm) reverse phase column (Phenomenex Inc., Torrance, CA) with a mobile phase that consisted of acetonitrile (solvent A) and 0.1% aqueous formic acid (solvent B) and eluted at a flow rate of 0.25 mL/min. The gradient system was set as 0 min, 10% A; 5 min, 10% A; 35 min, 90% A; 40 min, 95% A; 45 min, 95% A; and 50 min, 10% A. The LC effluent was monitored with a Finnigan Surveyor PDA detector (Thermo Electron Corp.) using a UV wavelength range from 200 to 400 nm before they were analyzed by the Finnigan LCQ Advantage MAX mass spectrometer (Thermo Electron Corp.). An electrospray ionization (ESI) source was used for the identification of metabolites by scanning the positive ions at the range of m/z from 200 to 600. The corresponding instrumental parameters were set as capillary temperature at 225 °C, source voltage at 5000 V, capillary voltage at 21 V, sheath gas flow rate at 50 units, and aux/sweep gas flow at 18 units, which were determined to be the optimal condition for the detection of pure SPD-304. MS/MS experiments were carried out under the following conditions. Each metabolite was numbered consecutively from M1 to M10 according to their order of elution from the HPLC column. Precursor ions that were derived from LC/MS experiments were m/z 548 (SPD-304), m/z 218 (M1), m/z 275 (M2), m/z 578 (M3), m/z 305 (M4), m/z 580 (M5), m/z 348 (M6), m/z 362 (M7), m/z 564 (M8), and m/z 535 (M9 and M10), which were selected individually for the CID. MS/MS experiments were then accomplished using collision energies set at 25% for SPD-304, M3, M5, M8, M9, and M10 or 28% for M1, M2, M4, M6, and M7, with activation Q at 0.25 and an activation time of 30 ms. The product ions were scanned at m/z 250–600 for SPD-304, M3, M5, M8, M9, and M10 as well as m/z 150–400 for M1, M2, M4, M6, and M7. MS3 experiments were conducted for SPD-304 and several metabolites, which focused on the complete identification of M9 and M10, as well as verification of the m/z 274 ion from MS/MS experiments. The peak of m/z 274 from SPD-304, M3, M4, and M6–10, m/z 287 from SPD-304, m/z 303 from M8, and m/z 273 from M9 and M10 derived from MS/MS experiment above were selected for MS3 analysis. MS3 experiments were then accomplished with collision energies set at 40% for m/z 274 or 35% for m/z 273, m/z 287, and m/z 303, with activation Q at 0.25 and activation time of 30 ms. The product ions were scanned at m/z 150–300 for m/z 273, m/z 274, and m/z 287 and 150–350 for m/z 303. Characterization of GSH Adducts by LC/MS/MS. LC/MS and LC/MS/MS experiments of GSH adducts were performed using the same Finnigan LCQ system, C18 column (150 mm × 2.00 mm) and mobile phase and gradient system as described above for the identification of metabolites. Normal mode MS/MS, data-dependent MS/MS, and neutral loss scan MS/MS were used to identify GSH adducts. Normal mode LC/MS/MS experiments were carried out with the following conditions: Precursor ions were set as m/z 853 (G1), m/z 667 (G2), m/z 610 (G3), and m/z 839 (G4), which were identified by LC/MS and selected individually for collision-induced dissociation (CID) with collision energies set at 25%, activation Q

376 Chem. Res. Toxicol., Vol. 21, No. 2, 2008

Sun and Yost

Table 1. Chromatographic and Spectral Characterization of SPD-304 and Metabolites compound tR (min) UV λmax (nm) mass

MS/MS

MS3

287, 244, 273, 187 205 218, 257 317, 303

274 f 205, 254 287 f 244, 242, 256

SPD-304 M1 M2 M3 M4 M5 M6 M7 M8 M9

37.3 16.1 19.3 24.3 24.8 26.7 27.5 28.2 30.5 35.7

210, 224, 218, 216, 258, 258, 224, 260, 250, 252,

252, 298 246, 308 268 256, 292 294 292 292 300 300 300

548 218 275 578 305 580 348 362 564 534

274, 187, 244, 274, 274 274, 274 274 274, 273,

M10

36.8

208, 252, 298

534

273, 274, 259, 228

a

274 f 205, 274 f 205,

319, 276, 305, 260, 219 303, 260, 289, 203 274, 259, 228

274 274 274 273

f f f f

205, 205, 205, 228,

273 f 228,

P450 enzymes

CYP3A4, CYP3A4, 254 CYP3A4 254 CYP3A4 CYP3A4 254 CYP3A4 254 CYP3A4 254 303 f 260, 258, 272 CYP3A4 216, 230 274 f 205, 254 CYP1A2, (13%) 230, 244 274 f 205, 254 CYP3A4, (7%)

CYP1A2 (64%)a CYP1A2 (50%)

CYP3A4 (10%), CYP2D6 CYP2D6 (10%), CYP1A2

Relative amounts of metabolites, normalized to the percentage of the most efficient one (CYP3A4 for all metabolites, except CYP1A2 for M9).

at 0.25, and activation time at 30 ms. The product ions were scanned at m/z 500–900 for G1 and G4 or m/z 200–700 for G2 and G3. For the data-dependent MS/MS scan, the most abundant, along with the second most and third most abundant ions in each parent ion scan were selected for MS/MS, with the normalized collision energies set at 25% and default isolation width set at 3. For the neutral loss mode MS/MS, both a neutral loss mass of 129 amu corresponding to the pyroglutamate residue of the GSH moiety (13) and of 274 amu corresponding to a typical trifluoromethylphenyl1H-indol-3-ylmethyl group loss in SPD-304 were used to discover GSH adducts, with the normalized collision energies set at 25% and the default isolation width set at 3. Inactivation of CYP3A4 by SPD-304. Primary incubations included SPD-304 (0–50 µM), 2 mM NADPH, 100 pmol of CYP3A4, and 0.1 M potassium phosphate buffer (pH 7.4). The mixture was incubated in a 37 °C shaking water bath for various time points (0, 2, 5, 8, 10, 15, and 20 min). At each preincubation time point, aliquots (5 µL) of the primary incubation mixtures containing 10 pmol of CYP3A4 were transferred to a secondary incubation to a final volume of 100 µL, which included 0.2 mM testosterone, 2 mM NADPH, and 0.1 M potassium phosphate (pH 7.4). The testosterone mixture was incubated in a 37 °C shaking water bath for 20 min and stopped by the addition of ice-cold acetonitrile in a 1:1 ratio (v/v), and then, the mixture was centrifuged at 21000g for 30 min. The supernatant was collected and analyzed by HPLC that was performed on an Agilent 1100 system (Agilent Technologies, Inc., Palo Alto, CA) including an autosampler and a diode array UV/vis detector. 6β-Hydroxytestosterone was separated using a Phenomenex Luna 5 µm C18 (250 mm × 4.60 mm, 5 µm) reverse phase column at a flow rate of 1 mL/min. The mobile phase consisted of acetonitrile (solvent A) and 1 mM ammonium acetate (solvent B) with the gradient solvent program that was set as 0 min, 10% A; 7 min, 35% A; 14 min, 50% A; 17 min, 55% A; 21 min, 95% A; 24 min, 95% A; and 32 min 10% A. All samples were stored at 4 °C in the autosampler before injection. The CYP3A4 activity was monitored by analysis of 6β-hydroxytestosterone ultraviolet absorption at 250 nm. In addition, reduced GSH (4 mM) or testosterone (0.2 mM) was added to certain primary incubations to study the effect of alternate substrates and exogenous nucleophiles. Molecular Docking. Three-dimensional coordinates of the X-ray crystal structure of the human CYP3A4 enzyme at a resolution of 2.05 were obtained from the Research Collaboratory for Structural Bioinformatics (RCSB) Protein Data Bank (PDB, code 1TQN) (14). It was used as the template for automatic docking of SPD-304 using the AutoDock program version 3.05 (Scripps Research Institute, La Jolla, CA) (15), which was run on a Dell Precision 690 workstation with two 64-bit Dual-Core Intel Xeon processors and Red Hat Enterprise Linux WS4 operating system. A threedimensional structure of SPD-304 was constructed using Chem3D Ultra10 (CambridgeSoft Corp., Cambridge, MA), and its energy was minimized using the molecular mechanics method (MM2). AutoDockTools (Scripps Research Institute) was used to prepare all docking parameter files for both SPD-304 and CYP3A4. In

general, Gasteiger atomic charges were assigned and flexible torsions were defined for SPD-304. Polar hydrogens, partial charges, and solvation parameters were added to the template of CYP3A4. A grid box with sufficient space to cover the whole active site of CYP3A4 was defined using AutoGrid 3.06 (Scripps Research Institute), in which AutoDock searches the optimal SPD-304 conformation and orientation using Lamarckian genetic algorithm (LGA), a hybrid of genetic algorithm (GA) with an adaptive local search (LC) method. Eight thousand LGA runs (from 40 independent experiments with 200 LGA runs each) were executed. The lowest-energy poses (from a pool of 8000 searched poses) were selected for analysis.

Results Metabolic Activation of SPD-304 by P450s. Incubations of the substrate SPD-304 with pooled human liver microsomes and individual P450 enzymes produced at least 10 metabolites. Ten specific major metabolites (M1-M10) were identified using LC/ MS/MS and/or LC/MS3 (Table 1 and Figure 2). All of these metabolites eluted earlier than the substrate in a reverse phase HPLC analysis and were named according to their elution order (M1 is the first eluted, and M10 is the last before SPD-304). They were produced in an NADPH-dependent manner by human liver microsomes. P450 enzymes CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4 were also used to evaluate their ability to form these metabolites. CYP3A4 was the only enzyme that produced all of these metabolites. CYP1A2 and CYP2D6 produced several metabolites, but SPD-304 was not a substrate for the rest of P450 enzymes tested. Therefore, both human liver microsomes and CYP3A4 were used for structural characterization of the metabolites and GSH adducts. Identification of Metabolites by LC/MS, LC/MS/MS, and LC/MS3. The base peaks of SPD-304 (MH+ ) 548) and its metabolites were analyzed by LC/MS. The metabolites were formed through several different pathways: (i) hydroxylation (MH+ ) 564, M8) of one methyl group in the dimethyl chromone ring and subsequent oxidation by CYP3A4 to form an acid (MH+ ) 578, M3) or a second hydroxylation of the other methyl group in the dimethyl chromone ring (MH+ ) 580, M5); (ii) N-demethylation of either methyl group of the dimethylamine spacer (MH+ ) 534, M9 or M10); (iii) Ndealkylation of the dimethylamine spacer to form several smaller amines (MH+ ) 218, M1; MH+ ) 275 M2; MH+ ) 305, M4; MH+ ) 348, M6; and MH+ ) 362, M7); and (iv) dehydrogenation of the 3-substituted indole to form the methyleneindolenine-reactive intermediate, followed by hydration to form the benzylic alcohol, a carbinolamine, which likely was not stable and subsequently fragmented to an aldehyde and an amine

P450-Catalyzed Dehydrogenation of a TNF-R Inhibitor

Chem. Res. Toxicol., Vol. 21, No. 2, 2008 377

Figure 2. Proposed SPD-304 metabolic pathways catalyzed by individual P450s. All metabolites were formed by CYP3A4; CYP1A2 produced M1, M2, M9, and M10, and CYP2D6 produced M9 and M10.

(MH+ ) 275, M2). Alternatively, M2 could also have been formed by direct hydroxylation of the 3-methylene group. Surprisingly, no corresponding aldehydes or acids were observed from cleavage of the indole portion. The protonated molecular ions for SPD-304 or the metabolites were used for CID to identify their structures by LC/MS/MS and LC/MS3. Several other metabolites with oxygen incorporation (+16 amu) were produced in amounts that were too small to be identified by LC/MS/MS. Even though metabolites with masses corresponding to the corresponding alcohols, aldehydes, or acids of metabolites such as M1, M2, M4, and M7 were observed, the MS/MS and MS3 experiments did not produce diagnostic fragments (with signal/noise ratios of at least 4/1) that could be used to confirm the structures of these products. Tandem mass spectrometry of SPD-304 by CID of MH+ showed two major product ions at m/z 274 and 287 (Figure 3A). The product ion at m/z 274 was assigned to the trifluoromethylphenyl-1H-indol-3-ylmethyl portion of SPD-304 resulting from the cleavage of the bond between the 3-methylene carbon and the adjacent nitrogen of the dimethylamine spacer, and the product ion at m/z 287 was assigned to the rest of the SPD-304 molecule without the trifluoromethylphenyl indole portion. Three minor product ions were also found at m/z 244 (cleavage of the dimethylamine spacer), m/z 273 (counterpart of the major ion at m/z 274), and m/z 187 (the dimethylchromone portion). The identity of two major product ions was further confirmed by MS3 analysis (data not shown). The m/z 274 ion

was further cleaved into two product ions: m/z 205 (loss of trifluoromethyl) and m/z 254 (loss of hydrofluoric acid). The m/z 287 ion fragmented into several product ions: m/z 242 and m/z 256 (loss of COOH or OCH3 from the chromone ring, respectively) and m/z 244 (from the cleavage within the dimethylamine spacer). Inspection of the MS/MS spectrum of M8 confirmed the putative hydroxylation of a methyl group on the dimethylchromone portion of SPD-304. Two major fragment ions at m/z 274 and m/z 303 corresponded to the two major ones of SPD304, which also suggested the hydroxylation position was either at the dimethylamine spacer or at the dimethyl chromone ring. Theoretically, hydroxylation products at the dimethylamine spacer would be unstable, so hydroxylation at the dimethyl chromone ring was the logical choice. This conclusion was confirmed by the presence of three minor product ions at m/z 260, m/z 289, and m/z 203 (Figure 3D) that corresponded to three of the minor ions of SPD-304. Both product ions at m/z 274 and 303 were selected for an additional round of fragmentation (MS3), which confirmed the presence of a primary alcohol on the dimethyl chromone ring for this metabolite. The CID of the m/z 303 ion formed product ions at m/z 260, 258, and 272, which corresponded to an addition of 16 amu to each fragment ion from the m/z 287 ion of SPD-304. The MS/MS spectra of M5 showed two similar major fragments (m/z 274 and m/z 319) and three minor fragments at m/z 276, 305, and 219, which confirmed the hydroxylation at both methyl groups on the

378 Chem. Res. Toxicol., Vol. 21, No. 2, 2008

Sun and Yost

Figure 3. LC/MS/MS product ion spectra of SPD-304 and its metabolites by human liver microsomes or CYP3A4, obtained by CID of the MH+ ions corresponding to (A) m/z 548, SPD-304; (B) m/z 578, M3; (C) m/z 580, M5; (D) m/z 564, M8; and (E) m/z 534, M9/M10. The putative assignments of characteristic fragment ions are shown.

dimethyl chromone ring (Figure 3C). The MS/MS results on metabolite M3 (Figure 3B with major peaks m/z 274 and m/z 317) were consistent with a pathway wherein SPD-304 was hydroxylated first to M8 and then oxidized to an acid (M3) by CYP3A4 (Figure 2), although air oxidation could not be excluded as a mechanism. MS3 experiments provided vital supporting data to identify the metabolites. Multiple daughter ions enabled the identification of M9 and M10, when their MS/MS spectra were not sufficient. Metabolites M9 and M10 had the same mass (loss of 14 amu

from SPD-304), and they eluted closely to each other, just before SPD-304. They also had a similar profile of MS/MS fragmentation, product ions at m/z 273, 274, 259, and 228. The ions at m/z 274 and 273 corresponded to two of the major fragment ions of SPD-304. This information suggested that M9 and M10 were two N-demethylation metabolites, losing either of the methyl groups on the dimethylamine spacer. To establish the identity of each, MS3 analysis on ions m/z 274 and 273 was required. The product ions from the m/z 274 ion of M9 and M10 showed the same pattern as that of SPD304. However,

P450-Catalyzed Dehydrogenation of a TNF-R Inhibitor

Figure 4. LC/MS/MS product ion spectra of SPD-304 metabolites by human liver microsomes or CYP3A4, obtained by CID of the MH+ ions corresponding to (A) m/z 218, M1; (B) m/z 275, M2; (C) m/z 305, M4; (D) m/z 348, M6; and (E) m/z 362, M7. The putative assignments of characteristic fragment ions are shown.

the product ions from the m/z 273 ion of M9 and M10 provided crucial information. The cleavage of the dimethylamine spacer at the 3-methylene indole end formed ions at m/z 230 for M9 but m/z 244 for M10. At the same time, the cleavage of the dimethylamine spacer in the middle formed ions at m/z 216 for M9 but m/z 230 for M10. These cleavage patterns confirmed the structures of M9 and M10, as shown in Figure 2. Five other metabolites with smaller masses (m/z 218, m/z 275, m/z 305, m/z 348, and m/z 362) suggested that these metabolites were formed by fragmentation of the whole molecule. We opined that they were amine fragments of SPD-304, formed by dealkylation reactions. MS/MS patterns of these metabolites are shown in Figure 4. M1 appeared to contain a typical dimethyl chromone ion at m/z 187, which was commonly present in a MS3 fragment of the parent molecule and several metabolites. The M4, M6, and M7 metabolites possessed a fragment at m/z 274, which was confirmed by MS3 to be the same trifluoromethyl-phenyl-1H-indol-3-ylmethyl portion of the parent drug (Table 1). These metabolites fit nicely into a metabolic scheme

Chem. Res. Toxicol., Vol. 21, No. 2, 2008 379

(Figure 2), where P450 catalyzes multiple N-demethylation and N-dealkylation reactions that are centered on the dimethylamine spacer. Detection of GSH Adducts. Six GSH adducts (G1-G6) were detected. The LC/MS/MS product ion spectrum of G1 (MH+ ) 853, Figure 6A) strongly supported the identification of a GSH adduct of SPD-304, which was presumably produced by nucleophilic conjugation of GSH to the 3-methyleneindolenine electrophilic intermediate, formed by dehydrogenation of SPD304 (Figure 5). The GSH adduct (Figure 6A) showed a characteristic loss of 129 amu at m/z 724, corresponding to the loss of pyroglutamate residue of the GSH moiety (13). Additional major fragments indicated the loss of water to form ions at m/z 835; the loss of glycine to form ions at m/z 778 as well as loss of the whole GSH moiety to form product ions at m/z 546. Figure 5 illustrated a proposed scheme, wherein G2, G3, and G4 were GSH adducts of three metabolites of SPD304, M7, M4, and M9, respectively. These metabolites had not lost the indole-to-dimethylamine spacer structural feature, which suggested that the GSH adducts were formed through conjugation of GSH with the 3-methyleneindolenine electrophilic intermediates that were produced from these metabolites by dehydrogenation (Figure 5 and 6B-D). G2 (MH+ ) 667) was identified as a GSH adduct of M7 (MH+ ) 362). The CID fragmentation (Figure 6B) of m/z 667 generated ions at m/z 621, 537, 408, and 274. The product ion at m/z 621 was formed by the loss of COOH, and m/z 537 was attributed to a cleavage of the peptide bond between pyroglutamate and cysteine of GSH. Two product ions at m/z 274 and 408 could have been produced from the typical cleavage at either side of the 3-methylene group as observed in SPD-304 and most of its metabolites, but m/z 274 could also have been produced from the cleavage of the bond between the R-carbon and sulfur of GSH, as shown in Figure 6B. G3 (MH+ ) 610, Figure 6C) was characterized by LC/MS/MS as a GSH adduct of dehydrogenated M4 (MH+ ) 305). Apparently, the fragment ion at m/z 481 was produced via a neutral loss of 129 amu (the pyroglutamate moiety), and water loss from the protonated molecular ion formed an ion at m/z 592. The product ion at m/z 465 was attributed to a cleavage of the trifluoromethylphenyl group from SPD-304, from which an additional loss of glycine, water, and NH3 produced a fragment ion at m/z 355 (Figure 6C). The tandem MS spectrum of G4 (MH+ ) 839) demonstrated it was the GSH adduct of dehydrogenated metabolite M10 (MH+ ) 534). The fragment ions detected and fragmentation patterns are shown in Figure 6D, which include the loss of 129 amu to form product ions at m/z 710, the loss of water to form an ion m/z 821, the cleavage between 3-methylene group and indole moiety of SPD-304 to form an ion at m/z 578, the cleavage of the dimethyl chromone moiety of SPD-304 to form an ion at m/z 653, or an additional loss of CO2 to form an ion at m/z 609. A diagnostic loss of the 129 amu pyroglutamate residue of GSH is a highly consistent feature of GSH adducts (13). The direct neutral loss LC/MS/MS method as well as a similar datadependent neutral loss method were used to identify additional GSH adducts from the metabolism of SPD-304. Two GSH adducts (G5 and G6) were found by the neutral loss scan of 129 amu or 274 Da. Specifically, when a fragment peak with an m/z value corresponding to the neutral loss of 129 amu was found, it suggested that this metabolite was a GSH adduct. Likewise, when a fragment peak with an m/z value corresponding to the neutral loss of 274 amu, it suggested that the metabolite was a SPD-304 adduct. These two neutral loss fragments were indeed found in the CID spectra of G5 and G6

380 Chem. Res. Toxicol., Vol. 21, No. 2, 2008

Sun and Yost

Figure 5. Proposed pathways for GSH adduct formation from the metabolism of SPD-304 by human liver microsomes or CYP3A4. Brackets denote putative intermediates.

(Figure 7). The fragment ion at m/z 274 was also confirmed by MS3 to be the trifluoromethyl-phenyl-1H-indol-3-yl portion of SPD-304. As a result, the combination of both neutral loss fragments in G5 and G6 confirmed the identity of the SPD-304 GSH adducts. The information also suggested that the conjugation position should be within the chromone ring portion of SPD-304. If the GSH attack occurred at the C-3′ position, the epoxide ring opens, and the hydroxyl group in the C-2′ position would be very unstable. The chromone ring would then open to form the aldehyde structure shown in Figure 7A. This aldehyde would probably be unstable. In fact, instead of loss of the entire intact chromone ring upon CID, a portion of the chromone ring (shaded area) was cleaved within the ESI source to form an ion at m/z 721, which was the highest observable ion (MH+) of G5. Thus, we were not able to detect the G5 adduct as an intact molecule by MS, probably because of the unstable nature of the aldehyde. The CID spectrum of the m/z 721 ion showed the loss of CO2 to form an ion at m/z 678, the cleavage between the indole ring and the 3-methylene to form ions at m/z 460 (also diagnostic), m/z 274, and m/z 592 (-129 amu). On the other hand, if GSH attacked the 2′-position of the epoxide, an additional loss of a water molecule would form G6 (Figure 7B). However, this molecule also seemed to be unstable in the ESI source, because it fractured (shaded area) to leave

the highest observable ion at MH+ ) 705. The resulting MS/ MS fragments from CID of the ions at m/z 705 showed the loss of water at m/z 687, the loss of GSH at m/z 398, and the cleavage between the indole ring and the 3-methylene portion to form ions at m/z 444, 274, and 576 (-129 amu). CYP3A4 Inactivation Kinetics of SPD-304. Preincubation of SPD-304 with CYP3A4 showed that CYP3A4 was inactivated in a time- and concentration-dependent manner (Figure 8). The observed first-order rate constants (kobs) of the inactivation reaction, at a specific SPD-304 concentration, were calculated from the slopes of these lines. CYP3A4 lost approximately 78% activity over the 20 min preincubation period at the highest concentration. The inset of Figure 8 shows a plot of the reciprocal of kobs against the reciprocal of several SPD304 concentrations, from which rate constants were obtained. The kinact was determined to be 0.047 min-1 and KI was 29 µM. The partition ratio, which measures the inactivation efficiency, that is, the number of SPD-304 molecules metabolized per molecule of CYP3A4 inactivated, was 120. In addition, the effects of an alternate substrate (testosterone) or a nucleophile (GSH) were also evaluated. The rate of inactivation was decreased by approximately 50% when testosterone was included in the preincubation. Minor protection (15%) was observed when GSH was included, which suggested that GSH could also enter into the active site of CYP3A4 and react with

P450-Catalyzed Dehydrogenation of a TNF-R Inhibitor

Chem. Res. Toxicol., Vol. 21, No. 2, 2008 381

Figure 6. LC/MS/MS product ion spectra of GSH adducts of SPD-304 from dehydrogenation, obtained by CID of the MH+ ions (A) m/z 853, G1; (B) m/z 667, G2; (C) m/z 610, G3; and (D) m/z 839, G4. The putative assignments of characteristic fragment ions are shown.

electrophilic intermediates before they inactivated the enzyme. Alternatively, reactive intermediates could have been released from the active site and been trapped with GSH before they could inactivate the enzyme, but this mechanism would require extraordinary stability of the intermediates. Molecular Docking. Six different CYP3A4 crystal structures have been resolved and deposited in RCSB PDB with codes 1W0E (unbound; resolution, 2.80 Å), 1W0F (complex with progesterone; resolution, 2.65 Å), 1W0G (complex with metyrapone; resolution, 2.73 Å), 1TQN (unbound; resolution, 2.05 Å), 2J0C (complex with ketoconazole; resolution, 2.80 Å),

and 2J0D (complex with erythromycin; resolution, 2.75 Å). The 1TQN structure was selected for docking studies because it had the highest resolution, and we wished to utilize a structure without the perturbations associated with a bound substrate. Docking energies were calculated with the AutoDock program. The 10 lowest energy poses from 8000 docked conformations were selected for analysis. The binding free energies of these 10 poses were reasonably low (from -13.45 to -14.88 kcal/mol), as compared to the whole range of energies (from -1.82 to -14.88 kcal/mol). The poses with the lowest free energies are shown in Figure 9. These conformations had free

382 Chem. Res. Toxicol., Vol. 21, No. 2, 2008

Sun and Yost

Figure 7. LC/MS/MS product ion spectra of GSH adducts of SPD-304 from epoxidation, obtained by neutral loss scan MS/MS of fragment ions at 129 and 274 amu; (A) MH+ ) 721, G5; and (B) MH+ ) 705, G6. The putative assignments of characteristic fragment ions are shown.

energies of -14.88 (pose A) and -14.28 (pose B) kcal/mol. The benzylic methylene at the 3-position of the trifluoromethylphenyl indole was the closest residue (4.69 Å) to the heme iron (Figure 9A). In Figure 9B, the distances between each of the two methyl groups in the dimethyl chromone ring and the heme iron were 3.13 and 5.07 Å, respectively. From the selected 10 poses, three were similar to pose A, with the 3-methylene nearest to the heme; another was like pose B, with the 6′-methyl nearest to the heme; two poses had the benzene of the indole ring nearest to the heme; and one had the nitrogen in the dimethylamine spacer that is attached to the chromone ring closest to the heme iron.

Discussion This study established the biotransformation of an intriguing new small molecule TNF-R antagonist, SPD-304, by human liver microsomes, CYP3A4, and several other P450s, through pathways of dehydrogenation, hydroxylation, N-dealkylation (including N-demethylation), or epoxidation. Furthermore, several GSH adducts were identified that verified the production of several electrophilic intermediates. These adducts were

formed through dehydrogenation or epoxidation pathways. Additionally, we showed that SPD-304 inactivated CYP3A4 through a mechanism-based process. Previous studies from our laboratory have demonstrated that multiple substrates, such as 3-methylindole, zafirlukast, indolines, and capsaicin, were dehydrogenated by various P450 enzymes (8, 9, 16, 17), and the current results have extended our research on P450-catalyzed dehydrogenation. Moreover, the in silico docking results of SPD304 within the active site of CYP3A4 were strikingly predictive of the in vitro metabolic profiles. However, it should be noted that multiple enzyme/substrate complexes would be expected for this highly dynamic enzyme. Taken together, the overall results showed that the 3-substituted indole-containing drug, SPD-304, is indeed an excellent substrate for dehydrogenation as well as a mechanism-based inactivator of CYP3A4. Therefore, it may elicit drug-drug interactions and be toxic. Unfortunately, the toxicities and poor pharmacokinetics of lead compounds are frequent causes of drug development failures (18). Compounds can be bioactivated by P450s to form electrophilic intermediates, which will covalently bind to

P450-Catalyzed Dehydrogenation of a TNF-R Inhibitor

Figure 8. Time- and concentration-dependent inactivation of CYP3A4 by SPD-304. Key: 0 µM (×), 10 µM (0), 20 µM (4), 35 µM (O), and 50 µM (]). Each point was determined from three replicates. The inset is a Kitz-Wilson double reciprocal plot (KI ) 29 µM and kinact ) 0.047 min-1).

Figure 9. Molecular model of SPD-304 in the CYP3A4 active site, using AutoDock to generate 8000 conformations. The conformations of SPD-304 with the two lowest docking energies are illustrated as follows: (A) conformation showing dehydrogenation at the 3-methylene position at the indole (-14.88 kcal/mol) and (B) conformation showing oxidation of the methyl hydrogens on the chromone ring (-14.28 kcal/ mol). CYP3A4 is shown in a ribbon loop format (green), heme in sticks (pink), and iron as a sphere (orange). The substrate is shown with colorcoded sticks: nitrogen, blue; carbon, yellow; oxygen, red; and fluorine, cyan.

nucleophilic moieties in biological macromolecules in human tissues that can ultimately cause toxicities. The biotransformation of SPD-304 by P450s demonstrated here confirmed that 3-substituted indoles are potential toxicophores through the production of electrophilic 3-methylenindolenines by dehydrogenation reactions. Thus, 3-substitued indole moieties may not be suitable aromatic templates for the development of new drugs in the future. Molecular modeling and docking approaches have been widely used in today’s “hit to lead” drug discovery process, which generally includes hit confirmation as well as lead generation and optimization (19). Docking small molecules into the active site or binding site of proteins remains a key step in virtual screening of small-molecule drugs. After a lead compound is found, such as SPD-304 as a potential TNF-R antagonist, additional lead optimization remains, utilizing in vitro or in vivo ADME tools. In silico ADME tools have also become a crucial component in lead optimization in drug development. Recently, the structures of the major human P450 enzymes, CYP3A4 (14), CYP1A2 (20), CYP2A6 (21), CYP2D6 (22), and CYP2C8/9 (23, 24), have been determined by X-ray crystallography, which provide reliable templates for molecular docking to study drug metabolism in silico.

Chem. Res. Toxicol., Vol. 21, No. 2, 2008 383

A molecular docking method was employed to study the metabolic preference of SPD-304 within the active site of CYP3A4. The docking free energies that were calculated by AutoDock were the summation of van der Waals potential energy, average energy of hydrogen bonding, electrostatic potential energy, energy restriction of internal rotors of global rotation and translation, and desolvation energy change upon binding. These parameters provided a reliable in silico technique to predict substrate binding (15). The searching method used in this study was a Lamarckian genetic algorithm that successfully combined a genetic algorithm for global searching and a local search to perform energy minimization. As shown in Figure 9, incorporation of both molecular energies and space factors predicted that dehydrogenation of the 3-substituted indole and the hydroxylation of the methyl group on the chromone ring of SPD-304 are major metabolic pathways. These two lowest energy conformers predict that the major metabolites would be the dehydrogenated indole and hydroxylated methyl products. This prediction was consistent with the experimental results. The precise mechanisms that initiate and control the dehydrogenation reaction by P450s still have not been adequately elucidated. Our previous work showed that the rate-limiting step of dehydrogenation of 3-methylindole was hydrogen atom abstraction from the 3-methyl position (7). In this mechanism, the P450 iron-oxo (compound I) oxygen abstracts a hydrogen atom from the methyl and, subsequently, a second electron to directly produce the protonated 3-methyleneindolenine electrophile (methylene iminium). This dehydrogenation mechanism generally competes with the hydroxyl rebound mechanism that produces the corresponding alcohol. However, metabolism of 3-methylindole by the unique lung CYP2F1 and CYP2F3 enzymes produces the dehydrogenation product as the sole observable metabolite (25, 26). The dehydrogenated metabolite and corresponding alcohol metabolite of SPD-304 are shown as intermediates in the formation of G1 and M2 in Figure 5. For zafirlukast and analogues of MK-524, it is still uncertain if dehydrogenation is initiated from hydrogen atom abstraction at the 3-methylene carbon or by an electron abstraction from the indole nitrogen (9, 11). From molecular docking results, it appeared that the hydrogen atom abstraction at the 3-methylene carbon should be the rate-limiting step of SPD-304 dehydrogenation, because energetically and spatially it was identified as the most favorable conformation of P450 turnover. We did not find any conformation that showed the indole nitrogen within a reasonable distance of the iron, which decreases enthusiasm for the nitrogen oxidation mechanism. Many clinically important compounds have been identified as mechanism-based CYP3A4 inactivators. The total numbers of these CYP3A4 inhibitors have dramatically increased in recent years. Examples (along with several of their KI and kinact values) include the grapefruit juice component, bergamottin (KI ) 7.7 µM and kinact ) 0.3 min-1) (27); the selective peptide leukotriene receptor antagonist, zafirlukast (KI ) 13 µM and kinact ) 0.026 min-1) (9); the selective estrogen receptor modulator, tamoxifen (28); the natural cytotoxin, 4-ipomeanol (KI ) 20 µM and kinact ) 0.15 min-1) (29); a major constituent of oral contraceptives, 17R-ethynylestradiol (KI ) 18 µM and kinact ) 0.04 min-1) (30); and a potential HIV protease inhibitor (KI ) 7.5 µM and kinact ) 1.62 min-1) L-754,394 (31). The kinetic constants of the growing list of putative CYP3A4 inactivators can be compared to those of SPD-304 (KI ) 29 µM and kinact ) 0.047 min-1) to see that this new drug is a reasonably potent, that is, low KI, but relatively slow inactivator.

384 Chem. Res. Toxicol., Vol. 21, No. 2, 2008

In a classical mechanism-based enzyme inactivation, the inactivation should be both time- and concentration-dependent. Moreover, the presence of an alternate substrate should slow the inactivation or protect the enzyme. All of these characteristics were observed in our kinetic studies of SPD-304-mediated inactivation of CYP3A4. However, minor protection (15%) by GSH was an unexpected result. Theoretically, inactivation is believed to occur before the release of reactive electrophilic intermediates from the active site of the enzyme, and hence, GSH should not protect the enzyme. However, it is possible that GSH could have been present in the active site of CYP3A4 and could have trapped a small portion of the dehydrogenation imine or the epoxide of SPD-304, before these intermediates were released from the active site, or GSH could have reacted with the electrophiles after release from the enzyme. The epoxidation pathway is another major metabolic pathway that is considered to produce reactive intermediates and hence cause toxicities (32). In our previous studies, we found that epoxides were formed on the benzene ring of indoline (17) and 3-methylindole (33), and their ring-opened products and GSH adducts could be identified by LC/MS. Although our molecular docking studies suggested that this benzene ring in the 3-substituted indole portion of SPD-304 is also an energetically favored position for oxygenation, we did not detect any phenols. However, several very small hydroxylation peaks (+16 amu or m/z 564) that eluted near M8 were found by LC/MS, but the mass spectral fragmentation patterns did not permit us to establish the location of hydroxylation (either on the trifluoromethylphenyl ring or the indole ring of SPD-304). The abundance of these ions from LC/MS/MS was too low to permit MS3 identification. In addition, two GSH adducts (G5 and G6) that could have been produced by GSH addition to epoxides of the chromone ring of SPD-304 were also identified in this study but were not stable under ESI source and degraded to the two ions at m/z 721and m/z 705, respectively. Water could have added to the epoxides to form diols, but we did not observe any products from hydrolysis. If water attacked at either the C2′ or the C-3′ positions of the chromone ring epoxide, the products would be very unstable (with a hydroxyl at C-3′), and the chromone ring would then open to form an unstable aldehyde that probably could not be detected with the current LC/MS instrumental conditions. In conclusion, our data demonstrated that the novel smallmolecule TNF-R antagonist SPD-304 was dehydrogenated by CYP3A4 to form a reactive electrophilic R,β-unsaturated iminium intermediate, which was trapped by nucleophilic GSH. The electrophilic methyleneindolenine could possibly cause adverse effects by covalently binding to nucleophilic residues of protein and/or DNA. Other evidence for the production of potentially harmful reactive intermediates was confirmed by CYP3A4 inactivation studies that showed that SPD-304 is a mechanism-based inhibitor. Metabolites from dehydrogenation, hydroxylation, dealkylation, or epoxidation of SPD-304 were identified directly or through the characterization of their GSH adducts by LC/MS/MS, which were also confirmed by molecular docking studies. Future studies will identify the spatial and electronic parameters within the active site of CYP3A4 to elucidate the dehydrogenation mechanism, using both molecular modeling and site-directed mutagenesis studies. The identification of GSH adducts from in vitro studies only provides putative biomarker type evidence that electrophiles are generated, not that they are toxic. The toxicities of the intermediates must be assessed through other methods.

Sun and Yost

The results of the studies with SPD-304, coupled to several reports with other 3-substituted indole-containing drugs, provide convincing evidence that this aromatic entity should only be used with great caution to develop new therapeutic agents. The studies with MK-0524 demonstrated that sufficient electronwithdrawing substituents on the indole moiety inhibited the dehydrogenation mechanism and led to a potentially safe new agent. Thus, it may be possible to incorporate the 3-substituted indole moiety in new therapeutic entities. Acknowledgment. We thank Dr. Eric Johnson, Scripps Research Institute, for helpful instructions on the use of the AutoDock software. This research was supported by NIH Grants GM074249 from the National Institute of General Medical Sciences and HL013645 from the National Heart, Lung, and Blood Institute.

References (1) Feldmann, M. (2002) Development of anti-TNF therapy for rheumatoid arthritis. Nat. ReV. Immunol. 2, 364–371. (2) Schwartzman, S., and Morgan, G. J. (2004) Does route of administration affect the outcome of TNF antagonist therapy? Arthritis Res. Ther. 6 (Suppl. 2), S19–S23. (3) He, M. M., Smith, A. S., Oslob, J. D., Flanagan, W. M., Braisted, A. C., Whitty, A., Cancilla, M. T., Wang, J., Lugovskoy, A. A., Yoburn, J. C., Fung, A. D., Farrington, G., Eldredge, J. K., Day, E. S., Cruz, L. A., Cachero, T. G., Miller, S. K., Friedman, J. E., Choong, I. C., and Cunningham, B. C. (2005) Small-molecule inhibition of TNF-alpha. Science 310, 1022–1025. (4) Yost, G. S. (1989) Mechanisms of 3-methylindole pneumotoxicity. Chem. Res. Toxicol. 2, 273–279. (5) Nocerini, M. R., Carlson, J. R., and Yost, G. S. (1984) Electrophilic metabolites of 3-methylindole as toxic intermediates in pulmonary oedema. Xenobiotica 14, 561–564. (6) Regal, K. A., Laws, G. M., Yuan, C., Yost, G. S., and Skiles, G. L. (2001) Detection and characterization of DNA adducts of 3-methylindole. Chem. Res. Toxicol. 14, 1014–1024. (7) Skiles, G. L., and Yost, G. S. (1996) Mechanistic studies on the cytochrome P450-catalyzed dehydrogenation of 3-methylindole. Chem. Res. Toxicol. 9, 291–297. (8) Thornton-Manning, J., Appleton, M. L., Gonzalez, F. J., and Yost, G. S. (1996) Metabolism of 3-methylindole by vaccinia-expressed P450 enzymes: Correlation of 3-methyleneindolenine formation and protein-binding. J. Pharmacol. Exp. Ther. 276, 21–29. (9) Kassahun, K., Skordos, K., McIntosh, I., Slaughter, D., Doss, G. A., Baillie, T. A., and Yost, G. S. (2005) Zafirlukast metabolism by cytochrome P450 3A4 produces an electrophilic R,β-unsaturated iminium species that results in the selective mechanism-based inactivation of the enzyme. Chem. Res. Toxicol. 18, 1427–1437. (10) Dean, B. J., Chang, S., Silva Elipe, M. V., Xia, Y. Q., Braun, M., Soli, E., Zhao, Y., Franklin, R. B., and Karanam, B. (2007) Metabolism of MK-0524, a prostaglandin D2 receptor 1 antagonist, in microsomes and hepatocytes from preclinical species and humans. Drug Metab. Dispos. 35, 283–292. (11) Levesque, J. F., Day, S. H., Chauret, N., Seto, C., Trimble, L., Bateman, K. P., Silva, J. M., Berthelette, C., Lachance, N., Boyd, M., Li, L., Sturino, C. F., Wang, Z., Zamboni, R., Young, R. N., and Nicoll-Griffith, D. A. (2007) Metabolic activation of indole-containing prostaglandin D2 receptor 1 antagonists: Impacts of glutathione trapping and glucuronide conjugation on covalent binding. Bioorg. Med. Chem. Lett. 17, 3038–3043. (12) Danese, S., De Vitis, I., and Gasbarrini, A. (2001) Severe liver injury associated with zafirlukast. Ann. Intern. Med. 135, 930. (13) Baillie, T. A., and Davis, M. R. (1993) Mass spectrometry in the analysis of glutathione conjugates. Biol. Mass. Spectrom. 22, 319– 325. (14) Yano, J. K., Wester, M. R., Schoch, G. A., Griffin, K. J., Stout, C. D., and Johnson, E. F. (2004) The structure of human microsomal cytochrome P450 3A4 determined by X-ray crystallography to 2.05-Å resolution. J. Biol. Chem. 279, 38091–38094. (15) Goodsell, D. S., Morris, G. M., and Olson, A. J. (1996) Automated docking of flexible ligands: Applications of AutoDock. J. Mol. Recognit. 9, 1–5. (16) Reilly, C. A., and Yost, G. S. (2005) Structural and enzymatic parameters that determine alkyl dehydrogenation/hydroxylation of capsaicinoids by cytochrome p450 enzymes. Drug Metab. Dispos. 33, 530–536.

P450-Catalyzed Dehydrogenation of a TNF-R Inhibitor (17) Sun, H., Ehlhardt, W. J., Kulanthaivel, P., Lanza, D. L., Reilly, C. A., and Yost, G. S. (2007) Dehydrogenation of indoline by cytochrome P450 enzymes: A novel ”aromatase” process. J. Pharmacol. Exp. Ther. 322, 843–851. (18) Guengerich, F. P. (2006) Cytochrome P450s and other enzymes in drug metabolism and toxicity. Aaps J. 8, E101–E111. (19) Kitchen, D. B., Decornez, H., Furr, J. R., and Bajorath, J. (2004) Docking and scoring in virtual screening for drug discovery: Methods and applications. Nat. ReV. Drug DiscoVery 3, 935–949. (20) Sansen, S., Yano, J. K., Reynald, R. L., Schoch, G. A., Griffin, K. J., Stout, C. D., and Johnson, E. F. (2007) Adaptations for the oxidation of polycyclic aromatic hydrocarbons exhibited by the structure of human P450 1A2. J. Biol. Chem. 282, 14348–14355. (21) Yano, J. K., Hsu, M. H., Griffin, K. J., Stout, C. D., and Johnson, E. F. (2005) Structures of human microsomal cytochrome P450 2A6 complexed with coumarin and methoxsalen. Nat. Struct. Mol. Biol. 12, 822–823. (22) Rowland, P., Blaney, F. E., Smyth, M. G., Jones, J. J., Leydon, V. R., Oxbrow, A. K., Lewis, C. J., Tennant, M. G., Modi, S., Eggleston, D. S., Chenery, R. J., and Bridges, A. M. (2006) Crystal structure of human cytochrome P450 2D6. J. Biol. Chem. 281, 7614–7622. (23) Schoch, G. A., Yano, J. K., Wester, M. R., Griffin, K. J., Stout, C. D., and Johnson, E. F. (2004) Structure of human microsomal cytochrome P450 2C8. Evidence for a peripheral fatty acid binding site. J. Biol. Chem. 279, 9497–9503. (24) Wester, M. R., Yano, J. K., Schoch, G. A., Yang, C., Griffin, K. J., Stout, C. D., and Johnson, E. F. (2004) The structure of human cytochrome P450 2C9 complexed with flurbiprofen at 2.0-A resolution. J. Biol. Chem. 279, 35630–35637. (25) Lanza, D. L., Code, E., Crespi, C. L., Gonzalez, F. J., and Yost, G. S. (1999) Specific dehydrogenation of 3-methylindole and epoxidation

Chem. Res. Toxicol., Vol. 21, No. 2, 2008 385

(26) (27)

(28) (29) (30)

(31)

(32) (33)

of naphthalene by recombinant human CYP2F1 expressed in lymphoblastoid cells. Drug Metab. Dispos. 27, 798–803. Lanza, D. L., and Yost, G. S. (2001) Selective dehydrogenation/ oxygenation of 3-methylindole by cytochrome p450 enzymes. Drug Metab. Dispos. 29, 950–953. He, K., Iyer, K. R., Hayes, R. N., Sinz, M. W., Woolf, T. F., and Hollenberg, P. F. (1998) Inactivation of cytochrome P450 3A4 by bergamottin, a component of grapefruit juice. Chem. Res. Toxicol. 11, 252–259. Zhao, X. J., Jones, D. R., Wang, Y. H., Grimm, S. W., and Hall, S. D. (2002) Reversible and irreversible inhibition of CYP3A enzymes by tamoxifen and metabolites. Xenobiotica 32, 863–878. Alvarez-Diez, T. M., and Zheng, J. (2004) Mechanism-based inactivation of cytochrome P450 3A4 by 4-ipomeanol. Chem. Res. Toxicol. 17, 150–157. Lin, H. L., Kent, U. M., and Hollenberg, P. F. (2002) Mechanismbased inactivation of cytochrome P450 3A4 by 17 alpha-ethynylestradiol: Evidence for heme destruction and covalent binding to protein. J. Pharmacol. Exp. Ther. 301, 160–167. Lightning, L. K., Jones, J. P., Friedberg, T., Pritchard, M. P., Shou, M., Rushmore, T. H., and Trager, W. F. (2000) Mechanism-based inactivation of cytochrome P450 3A4 by L-754,394. Biochemistry 39, 4276–4287. Guengerich, F. P. (2003) Cytochrome P450 oxidations in the generation of reactive electrophiles: Epoxidation and related reactions. Arch. Biochem. Biophys. 409, 59–71. Yan, Z., Easterwood, L. M., Maher, N., Torres, R., Huebert, N., and Yost, G. S. (2007) Metabolism and bioactivation of 3-methylindole by human liver microsomes. Chem. Res. Toxicol. 20, 140–148.

TX700294G