In Vitro Bioactivation of Dihydrobenzoxathiin Selective Estrogen

Feb 24, 2005 - Estrogens and selective estrogen receptor modulators (SERMs) are prescribed widely in the clinic to alleviate symptoms in postmenopausa...
0 downloads 0 Views 253KB Size
Download PDF
Chem. Res. Toxicol. 2005, 18, 675-685

675

In Vitro Bioactivation of Dihydrobenzoxathiin Selective Estrogen Receptor Modulators by Cytochrome P450 3A4 in Human Liver Microsomes: Formation of Reactive Iminium and Quinone Type Metabolites Zhoupeng Zhang,*,† Qing Chen,† Ying Li,† George A. Doss,† Brian J. Dean,† Jason S. Ngui,† Maria Silva Elipe,†,‡ Seongkon Kim,§ Jane Y. Wu,§ Frank DiNinno,§ Milton L. Hammond,§ Ralph A. Stearns,† David C. Evans,† Thomas A. Baillie,† and Wei Tang† Departments of Drug Metabolism and Medicinal Chemistry, Merck Research Laboratories, Rahway, New Jersey 07065 Received November 17, 2004

Estrogens and selective estrogen receptor modulators (SERMs) are prescribed widely in the clinic to alleviate symptoms in postmenopausal women, and they are metabolized to reactive intermediates, which may elicit adverse effects. As part of our efforts to develop safer SERMs, in vitro covalent protein binding of (2S,3R)-(+)-3-(4-hydroxyphenyl)-2-[4-(2-piperidin-1-ylethoxy)phenyl]-2,3-dihydro-1,4-benzoxathiin-6-ol (I) was evaluated. Radioactivity from [3H]I became covalently bound to proteins in a fashion that was both time- and NADPH-dependent in human liver microsomes and reached a value of 1106 pmol equiv/mg protein following a 45 min incubation. At least three pathways are involved in the bioactivation of I, namely, oxidative cleavage of the dihydrobenzoxathiin moiety to give a hydroquinone/para-benzoquinone redox couple, hydroxylation at position 5 or 7 of the benzoxathiin moiety leading to an o-quinone intermediate, and metabolism of the piperidine ring to give an iminium ion. The latter reactive intermediate was identified as its bis-cyano adduct when human liver microsomal incubations were performed in the presence of sodium cyanide. Structural modification of I, including a replacement of the piperidine with a pyrrolidine group, led to (2S,3R)-(+)-3-(3-hydroxyphenyl)2-[4-(2-pyrrolidin-1-ylethoxy)phenyl]-2,3-dihydro-1,4-benzoxathiin-6-ol (II), which did not form a reactive iminium ion. Following the incubation of II with human liver microsomes, covalent binding to proteins was reduced (461 pmol equiv/mg protein), the residual level of binding apparently due to the formation of a rearranged biphenyl quinone type metabolite. Studies with inhibitory antibodies and chemical inhibitors showed that P450 3A4 was the primary enzyme responsible for oxidative bioactivation of I and II in human liver microsomes. These studies thus demonstrated that gaining an understanding of bioactivation mechanisms may be exploited in terms of guiding structural modifications of drug candidates to minimize covalent protein binding and, hopefully, to lower the potential for drug-mediated adverse effects.

Introduction Endogenous estrogens elicit their hormonal and reproductive effects through binding to estrogen receptors. Estrogens are agonists at their receptors and have been used in hormone replacement therapy to alleviate symptoms and urogential atrophy at the time of menopause in women (1). However, it has been shown that the use of estrogens leads to the development of cancer in the breast and endometrium (1). Although the carcinogenic mechanism remains obscure, reactive intermediates of estrogens have been hypothesized to play an etiological role in tumor formation (2). In recent years, several selective estrogen receptor modulators (SERMs)1 have * To whom correspondence should be addressed. Tel: 732-594-4433. Fax: 732-594-4820. E-mail: [email protected]. † Department of Drug Metabolism. ‡ Current address: Department of Analytical Sciences, Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320. § Department of Medicinal Chemistry.

been evaluated for their potential to antagonize the adverse effects of estrogens on uterine and breast tissues while producing beneficial estrogen-like effects on bone tissue and the cardiovascular system (3, 4). Tamoxifen and reloxifene are the two nonsteroidal SERMs approved by the Food and Drug Administration for the treatment of breast cancer and osteoporosis, respectively (4). It was reported that the metabolic activation of tamoxifen led to covalent binding to proteins of rat and human liver microsomes (5) and formation of DNA adducts in patients (6). Our recent studies showed that reloxifene caused irreversible inhibition of cytochrome P450 3A4, possibly through formation of an arene oxide intermediate (7). 1 Abbreviations: CID, collision-induced dissociation; P450, cytochrome P450; ER-R, estrogen receptor-R subtype; ESI, electrospray ionization; equiv, equivalent; GSH, glutathione; HMBC, heteronuclear multiple bond correlation; LC/MS/MS, liquid chromatography-tandem mass spectrometry; NAc, N-acetylcysteine; NOE, nuclear Overhauser effect; SERM, selective estrogen receptor modulator; TOCSY, total correlation spectroscopy.

10.1021/tx0496789 CCC: $30.25 © 2005 American Chemical Society Published on Web 02/24/2005

676

Chem. Res. Toxicol., Vol. 18, No. 4, 2005

Zhang et al.

Figure 1. Structures of compounds I-IV, M1, M10, M12, and M17. The sites of tritium incorporation in [3H]I and [3H]II are indicated. Lower case letters denote specific hydrogen atoms cited in the descriptions of NMR spectra and in Table 2.

Reactive metabolites have been hypothesized to be associated with organ toxicity. In this regard, studies showed that tamoxifen caused increased incidence of endometrial cancer in patients (8) and initiated hepatocellular carcinoma in rats (4). To minimize the potential risk associated with reactive intermediates, a low propensity to form reactive intermediates is considered to be a desirable characteristic of drug candidates (9, 10). In an attempt to develop a safer SERM with an estrogen receptor-R subtype (ER-R) selectivity for the treatment of osteoporosis (11-18), (2S,3R)-(+)-3-(4-hydroxyphenyl)-2-[4-(2-piperidin-1-ylethoxy)phenyl]-2,3-dihydro-1,4-benzoxathiin-6-ol (I; Figure 1) was synthesized as a lead compound with potent and selective ER-R antagonist activities in vitro and in vivo in animals. However, I appears to undergo bioactivation forming chemically reactive intermediates in human liver microsomes and hepatocytes. We report herein the identification of reactive metabolites that may be responsible for the covalent binding of I to proteins. Subsequent structural modification based on this information led to the selection of (2S,3R)-(+)-3-(3-hydroxyphenyl)-2-[4-(2pyrrolidin-1-ylethoxy)phenyl]-2,3-dihydro-1,4-benzoxathiin-

6-ol (II; Figure 1) as a new lead (12, 14), which exhibits optimal pharmacological activities with a reduced propensity for bioactivation.

Materials and Methods Materials. NADPH, glutathione (GSH), N-acetylcysteine (NAc), Coomassie R-250 Blue, and protein markers for molecular weights ranging from 30 000 to 200 000 (Sigma SDS-6H) were purchased from Sigma-Aldrich Co. (St. Louis, MO). NuPAGE LDS sample buffer, acrylamide NuPAGE tris-acetate gel, NOVEX tris-glycine native running buffer, and XCell Surelook electrophoresis apparatus were from Invitrogen (Carlsbad, CA). All other reagents and solvents were obtained from Fisher Scientific (Fair Lawn, NJ). Microsomes from baculovirus-insect cells coexpressing human P450 3A4 and NADPH-P450 oxidoreductase were prepared at Merck Research Laboratories (19). Human liver samples were obtained from Pennsylvania Regional Tissue Bank (Exton, PA). Liver microsomes were prepared from individual livers by differential centrifugation, and aliquots of each preparation were pooled on the basis of equivalent protein concentrations (20, 21). The syntheses of I, II, (6R)-6-{(S)-hydroxy[4-(2-pyrrolidin-1-ylethoxy)phenyl]methyl}6H-benzo[C]thiochromene-3,8-diol (M12), and M17 were previously reported (12, 14, 18, 22) (Figure 1). Reference samples of

In Vitro Bioactivation of Dihydrobenzoxathiin SERMs metabolites M1 and M10 were synthesized in Merck Research Laboratories (Supporting Information) (Figure 1). Tritiumlabeled I and II tracers were synthesized in Merck Research Laboratories with a radiochemical purity of >98.5% (by HPLC). The specific activity was 21.8 and 29.0 mCi/mg for I and II, respectively. Instrumentation. Liquid chromatography-tandem mass spectrometry (LC/MS/MS) was carried out on either a Finnigan LCQDeca XP Plus mass spectrometer (San Jose, CA) interfaced to a HPLC system consisting of two Shimadzu LC-10AD pumps (Kyoto, Japan), a Shimadzu SIL-10AD auto injector and a Finnigan UV6000LP photodiode array detector, or a PerkinElmer SCIEX API 3000 tandem mass spectrometer (Toronto, Canada) interfaced to an HPLC system consisting of a PerkinElmer Series 200 quaternary pump and a Series 200 autosampler (Norwalk, CT). Qualitative analyses using the LCQDeca XP Plus mass spectrometer employed electrospray ionization (ESI) in the positive ion mode. The heated capillary temperature was 220 °C, the normalized collision energy was 42%, the sheath gas flow rate was 60 units, and the auxiliary gas flow rate was 20 units. The ion spray voltage, the capillary voltage, and the tube lens offset were adjusted to achieve maximum sensitivity using the parent SERMs of interest. Corresponding experiments on the SCIEX API 3000 mass spectrometer employed a heated nebulizer interface (for quantitation of I and II) with positive ion detection. The source temperature was set at 400 °C, the ion spray voltage was set at 5 kV, the focusing potential was set at 120 V, and the entrance potential was set at -10 V. The collision gas was nitrogen. For metabolite identification by LC/MS/MS, samples (75 µL) were loaded onto an Agilent Zorbax Rx-C8 column (4.6 mm × 250 mm, 5 µm, Wilmington, DE). The flow rate was set at 1 mL/min with a 1:5 split to the mass spectrometer ion source and a Packard flow scintillation analyzer (model 500TR, Boston, MA), respectively. The mobile phase consisted of solvent A (5 mM ammonium acetate in water-acetonitrile-acetic acid, 95:5:0.05, v/v/v) and solvent B (5 mM ammonium acetate in acetonitrile-water-acetic acid, 95:5:0.05, v/v/v). The HPLC runs were programmed by a linear increase from 0 to 80% of solvent B during a 30 min period. The MS/MS spectra were recorded by collision-induced dissociation (CID) of MH+ species. For quantitation of I and II by LC/MS/MS, samples (20 µL) were loaded onto a Keystone Scientific Betasil C8 column (4.6 mm × 50 mm, 5 µm, Wilmington, DE) and eluted at a flow rate of 1 mL/min. The mobile phase consisted of 5 mM ammonium acetate in acetonitrile-H2O-acetic acid (95: 5:0.05, v/v/v). The HPLC run time was 3 min, and quantitation was based on multiple reaction monitoring (MRM) of the transitions of m/z 464.2 f 324.1 (I) and 450.2 f 310.1 (II). Standard curves of I and II were linear over a range of 5-2000 nM. NMR spectra of isolated metabolites dissolved in deuterated solvents were recorded with a Varian Inova 600 spectrometer operating at 600 MHz. Chemical shifts (δ) are expressed as parts per million (ppm) downfield relative to tetramethylsilane, and coupling constants (J) are expressed in Hertz (Hz). In Vitro Metabolism of [3H]I and [3H]II in Human Liver Microsomes and Recombinant P450 3A4. Human liver microsomes (1 mg protein/mL) or recombinant P450 3A4 (250 pmol/mL) were suspended in phosphate buffer (100 mM, pH 7.4) containing EDTA (1 mM) and MgCl2 (0.1 mM) in a total volume of 1 mL. Substrates ([3H]I or [3H]II at a final specific activity 10 µCi/mg) in methanol were added to a final concentration of 10 µM, such that the concentration of methanol in the incubation mixture did not exceed 0.2%. Incubations were performed in the presence of NADPH (1.2 mM) at 37 °C for 45 min. The reaction was quenched by adding 2 mL of acetonitrile. The suspension then was sonicated for 5 min and centrifuged at 20800g for 10 min. The supernatants were removed, and the pellets were extracted twice with 1 mL of methanol-water (3:1, v/v). The extracts were combined with the above supernatants and were evaporated to dryness under nitrogen at room temperature. The residues were dissolved in

Chem. Res. Toxicol., Vol. 18, No. 4, 2005 677 300 µL of solvent B, and an aliquot (75 µL) was loaded onto an HPLC column for LC/MS/MS analysis. Studies of Covalent Protein Binding in Human Liver Microsomes and Human Hepatocytes. Incubations performed to evaluate covalent protein binding in human liver microsomes followed a protocol similar to that described above. The specific activity of [3H]I and [3H]II in these experiments was 20 µCi/mg. Reactions were performed in duplicate and quenched with 5 mL of acetonitrile at 0, 15, 30, and 45 min. The duration of the incubations conducted in the presence of sodium cyanide (1 mM) and GSH (5 mM) was 45 min. Samples were centrifuged at 2500g to afford protein pellets, which then were suspended in aqueous ethanol (water-ethanol, 1:4, v/v), vortexed, and centrifuged. This procedure was repeated until radioactivity in the supernatant was less than 2-fold background. The protein pellet then was dissolved in 0.1 M sodium hydroxide (1 mL), 50% of which was neutralized with 0.1 M hydrochloric acid and analyzed by liquid scintillation counting. The protein concentration in the remaining aliquot was determined using a Pierce bicinchoninic (BCA) protein assay kit (Rockford, IL). Covalent protein binding was estimated based on residual radioactivity in the protein pellets. Control experiments were performed in the absence of NADPH. Cryopreserved human hepatocytes from three male and two female donors (lot numbers 70, 78, 88, 91, and 95) were obtained from In Vitro Technologies (Baltimore, MD), and their testosterone 6β-hydroxylation activity ranged from 18 to 105 pmol/ min/106 cells. The cells exhibited viabilities of 70-80% based on exclusion of trypan blue. Incubations were performed by suspending the hepatocytes in Krebs-biocarbonate buffer followed by the addition of I or II in methanol. The final substrate concentration in the suspension was 10 µM in a final volume of 1 mL (1 × 106 cells/mL). Incubations were allowed to proceed at 37 °C for 1 h and were quenched with acetonitrile (5 mL). The remaining procedures were the same as those described above for the measurement of covalent protein binding in human liver microsomes. SDS-PAGE Analysis of Protein Modification by [3H]I. Human liver microsomes were incubated with [3H]I (10 µM; specific activity, 250 µCi/mg) in the presence or absence of NADPH for 45 min. The protein concentration was 1 mg/mL in an incubation volume of 1 mL. The reaction was stopped by adding acetonitrile (3 mL), and proteins were precipitated and washed with aqueous ethanol (H2O/EtOH, 20:80, v/v). The wash was repeated until radioactivity in the supernatant was less than 2× background, at which point residual radioactivity in the protein pellet was determined and used as a measure of covalent binding. A portion of this protein pellet (10 µg) was solubilized with NuPAGE LDS sample buffer and was subject to electrophoresis using a 7% acrylamide NuPAGE tris-acetate gel at 125 V for 2 h. The gel dimensions were 8 cm × 8 cm × 0.15 cm, and the buffer was NOVEX tris-glycine native running buffer. Protein markers for molecular weights ranging from 30 000 to 200 000 (Sigma SDS-6H) were included in the assay. The gel was stained with Coomassie R-250 Blue and was cut into 2 mm strips. The gel fragments were solubilized with 30% hydrogen peroxide (1.5 mL), followed by liquid scintillation counting. Inhibition of Metabolism and Adduct Formation of I and II in Human Liver Microsomes by Monoclonal Antibodies and Chemical Inhibitors. In the inhibition studies of I and II, human liver microsomes were preincubated with monoclonal antibodies against P450 3A4 and 2D6 (0.25 and 2.5 mg IgG/nmol P450) or with ketoconazole (1 and 10 µM) and quinidine (5 and 25 µM), at room temperature for 15 min before adding substrate. The substrate concentration was 10 µM, and the reaction volume was 1.5 mL. Controls contained ascites from untreated animals or methanol. The remaining conditions and procedures were the same as those described above. Reactions were initiated by adding NADPH. Aliquots (100 µL) of the reaction mixture were removed at 5, 10, 20, 30, and 45 min and were added to solutions of aqueous urea (8 M, 1.5 mL) contain-

678

Chem. Res. Toxicol., Vol. 18, No. 4, 2005

ing 15 ng of internal standards. The internal standards for quantitation of I and II were II and I, respectively. The samples were applied to a 96 well Oasis MCX extraction plate, which was prewashed with methanol and water. The plate was washed consecutively with water (500 µL) and eluted with methanol (300 µL). Aliquots of the resulting elutes (20 µL) were injected into the LC/MS/MS system. In the inhibition studies of adduct formation of [3H]I and [3H]II, human liver microsomes were preincubated with monoclonal antibodies against P450 3A4 and 2D6 (0.25 and 2.5 mg IgG/nmol P450), or with ketoconazole (1 and 10 µM) and quinidine (5 and 25 µM), at room temperature for 15 min before adding substrates and NADPH in the presence of individual trapping agents (1 mM sodium cyanide for I or 5 mM NAc for II). The substrate concentration was 10 µM, and the reaction volume was 1 mL. Controls contained ascites from untreated animals. The incubation time was 45 min. The remaining conditions and procedures were the same as those described above for the in vitro metabolism of [3H]I and [3H]II. Isolation of the Bis-cyano Adduct III, the Metabolite M12, and Its NAc Adduct IV. Incubations of [3H]I in the presence of 1 mM sodium cyanide or [3H]II in the presence of 5 mM NAc with recombinant P450 3A4 (250 pmol/mL) were performed as described above. The substrate concentration was 50 µM, and the incubation volume was 50 mL. Reactions were quenched by the addition of two volumes of acetonitrile, and the resulting suspensions were sonicated for 5 min and centrifuged at 20800g for 10 min. The supernatants were evaporated to dryness under nitrogen, and the residues were dissolved in solvent B. Aliquots were loaded onto a semipreparative Waters YMC-AQ C18 column (10 mm × 250 mm, 5 µm, Milford, MA) and eluted at a flow rate of 3 mL/min with a 1:25 split to the mass spectrometer ion source and a fraction collector. The HPLC gradient was the same as that described above, and fractions containing radioactivity were collected. Fractions corresponding to metabolite M12 or the adducts of interest were pooled and were evaporated to dryness under nitrogen at room temperature. The purified samples were taken for NMR analysis.

Zhang et al.

Figure 2. Covalent protein binding of [3H]I and [3H]II in human liver microsomes in the presence or absence of NADPH, GSH, and sodium cyanide. Incubations were performed, as described under the Materials and Methods, for a period of 45 min.

Figure 3. Covalent protein binding of [3H]I and [3H]II in human hepatocytes at 0 and 60 min (n ) 5). Values represent means ( SD.

Results In Vitro Covalent Protein Binding of I and II in Human Liver Microsomes and Hepatocytes. Metabolic stability studies of I and II in human liver microsomes showed that turnover rates of I and II were linear over a period of 60 min under the condition used (data not shown). The rates of metabolism of I and II in human liver microsomes were 51 and 131 pmol/min/mg protein, respectively. Therefore, the incubations to determine the covalent protein binding of [3H]I and [3H]II were performed in human liver microsomes for a period of 45 min. Covalent protein binding was both NADPH- and timedependent (data not shown). In the presence of NADPH, the covalent binding for [3H]I was 1106 pmol equiv/mg protein and was attenuated significantly by GSH or sodium cyanide (Figure 2). The covalent binding of [3H]II was 461 pmol equiv/mg protein and also was significantly attenuated by GSH. However, sodium cyanide had little effect on the binding of II (Figure 2). In human hepatocytes (n ) 5), the covalent protein binding values of I and II were 170 ( 104 and 48 ( 20 pmol equiv/mg protein at 60 min, respectively (Figure 3). SDS-PAGE Analysis of Protein Modification by [3H]I. Upon SDS-PAGE analysis of microsomal incubation products, a significant amount of radioactivity (∼0.4 µCi) comigrated with microsomal proteins, notably with components that had a molecular mass of approximately 55 kDa (Figure 4). This radiolabeled band was absent in incubations lacking NADPH.

Figure 4. Covalent binding of [3H]I to human liver microsomal proteins as determined by SDS-PAGE. Lettered arrows indicate molecular mass markers: A, phosphorylase b (94 kDa); B, bovine serum albumin (67 kDa); C, ovalbumin (43 kDa); and D, carbonic anhydrase (30 kDa).

In Vitro Metabolism of [3H]I and [3H]II in Human Liver Microsomes and Recombinant P450 3A4. The oxidative metabolism of [3H]I and [3H]II was investigated in incubations (45 min, 37 °C) with human liver microsomes and recombinant P450 3A4. The results of the metabolic studies on I are summarized in Scheme 1. The structures of the corresponding metabolites (M1-M8) were based upon LC/MS/MS data (Table 1 and Supporting Information). During the study period, 23 and 24% of the initial substrate were consumed in incubations with human liver microsomes and recombinant P450 3A4, respectively, resulting in the formation of an S-oxide derivative M2 and the N-dealkylated primary amine derivative M4 as major metabolites. Other metabolites

In Vitro Bioactivation of Dihydrobenzoxathiin SERMs

Chem. Res. Toxicol., Vol. 18, No. 4, 2005 679

Scheme 1. Proposed Metabolic Pathways of Compound I in Incubations with Human Liver Microsomes and Recombinant P450 3A4a

a

The percentile in parenthesis represents the % of metabolism in human liver microsomes (HLM) and recombinant P450 3A4. Table 1. Molecular Ions (MH+) and MSn Fragmentations of Compounds I-IV and M1-M15

compound

MH+ (m/z)

MS2

I

464

324 (100%), 259, 239, 211

II

450

310 (100%), 259, 239, 211, 145

III IV M1 M2

531 (+NH3) 611 482 480

514, 487 (100%), 391, 347 593, 482, 464 (100%) 340, 324 (100%), 206 340, 324 (100%), 206

M3 M4 M5

480 396 480

462, 340 (100%), 206 302, 259, 256 (100%), 211 462, 379, 340 (100%), 322, 259

M6 M7

462 480

322 (100%), 259, 239 340, 324 (100%), 259

M8 M9 M10

478 220 468

324 (100%), 273, 245 192, 149, 121, 98 (100%), 84, 70 450, 326, 310 (100%)

M11

466

326, 310 (100%)

M12

450

432, 361, 259, 221 (100%), 220

M13 M14 M15

448 448 464

229 (100%), 220 308 (100%), 259 324, 259, 206 (100%)

included a ring-opened hydroquinone M1, mono-oxygenated derivatives M3, M5, and M7, and a dehydrogenated derivative M6 (Scheme 1). The structure of the metabolite M1 was proposed on the basis of LC/MS/MS data and was confirmed by comparison with a synthetic standard

MSn MS3

(464 f 324): 239 (100%), 211, 145, 112, 107, 98 MS4 (464 f 324 f 239): 221, 211, 145 (100%), 107 MS3 (450 f 310): 282, 239, 221, 211, 145, 107, 98 (100%) MS4 (450 f 310 f 239): 221, 211, 145 (100%) MS3 (531 f 487): 460 (100%), 347, 320, 259, 202 MS3 (611 f 464): 393, 273 (100%) MS3 (382 f 340): 255, 160, 112 MS3 (480 f 340): 255 (100%), 161, 137, 112 MS3 (480f 324): 239, 211, 112, 98 (100%) MS3 (480 f 340): 322, 255, 227, 206, 112 (100%), 98 MS3 (396 f 259): 231, 226, 165, 153, 137 MS3 (480 f 340): 322 (100%), 239, 145, 128, 114 MS4 (480 f 340 f 322): 239, 211, 145, 110, 96 MS3 (462 f 322): 239, 211, 145, 110, 96 MS3 (480 f 340): 239, 211, 195, 145, 112, 98 (100%) MS3 (480 f 324): 239, 211, 195, 145, 112, 98 MS3 (478 f 324): 239, 211, 145, 112, 98 MS3 (468 f 326): 255, 227, 98 (100%) MS3 (468 f 310): 239 (100%), 145, 98 MS3 (464 f 326): 255 (100%), 227, 209, 161, 145, 98 MS3 (464 f 310): 239, 221, 211, 145, 98 (100%) MS3 (450 f 259): 241, 217 (100%) MS3 (450 f 221): 190, 98, 84, 70 MS3 (448 f 229): 169, 141 MS3 (448 f 308): 239 (100%), 221, 211, 145, 119, 96 MS3 ( 464 f 206): 112 (100%)

(Table 1, Scheme 1, and Supporting Information). Upon LC/MS/MS analysis, the MH+ ion of M8 was observed at m/z 478, 14 mass units higher than that of the parent compound. The CID spectra of M8 indicated that the additional 14 mass units were associated with the A ring

680

Chem. Res. Toxicol., Vol. 18, No. 4, 2005

Zhang et al.

Scheme 2. Proposed Metabolic Pathways of Compound II in Incubations with Human Liver Microsomes and Recombinant P450 3A4a

a

The percentile in parenthesis represents the % of metabolism in human liver microsomes (HLM) and recombinant P450 3A4.

of the molecule (Table 1, Scheme 1, and Supporting Information), most likely in the form of an o-quinone structure. Theoretically, an initial oxidation may occur at any of the positions 5, 7, and 8 of the A ring of I (Figure 1). However, only the 5,6- or 6,7-catechol can be oxidized further to an o-quinone, whose molecular weight was consistent with that of the detected ion. The results of the metabolic studies on II are summarized in Scheme 2. The structures of the corresponding metabolites (M9-M15) were based upon LC/MS/MS data (Table 1 and Supporting Information). In this case, the metabolic turnover figures were 59 and 30% in human liver microsomes and recombinant P450 3A4, respectively. The major metabolite proved to be an extended biphenyl hydroquinone type derivative M12, the identity of which was established as outlined below. Other metabolites included the aldehyde M9, a ring-opened hydroquinone derivative M10, an S-oxide M11, a dehydrogenated derivative M14, and a ketone or lactam derivative M15. The structure of the minor hydroquinone metabolite M10 was proposed on the basis of LC/MS/MS data and was confirmed by comparison with a synthetic standard (Table 1 and Supporting Information). On the basis of LC/MS/MS analysis, M13 exhibited a molecular ion species (MH+) at m/z 448, two mass units less than that of II. Subsequent identification of M13 was based on comparison of its HPLC retention time and mass spectral fragmentation pattern with that of the aglycon of M17, which indicated that M13 contained a dienone moiety (Scheme 2 and Supporting Information). The

metabolite M17 (Figure 1) was observed in urine from monkeys treated with II and subsequently was prepared by synthesis (22). Structural Analysis of the Bis-cyano Adduct III. Upon incubation of I with human liver microsomes fortified with NaCN and analysis of the products by LC/MS/MS (Table 1 and Supporting Information), a biscyano adduct was detected (M + NH4+ at m/z 531). The MS/MS spectrum of this adduct, obtained by CID of the m/z 531 species, exhibited product ions at m/z 514 (loss of NH3) and m/z 487 (loss of NH3 and HCN) (Table 1 and Supporting Information). Two minor fragments (at m/z 347 and 391) associated with the right-hand side of the dihydrobenzoxathiin moiety also were detected. The MS3 spectrum, obtained by CID of m/z 487, showed loss of a second HCN molecule (27 Da) to give an ion at m/z 460. The other product ions are listed in Table 1. Collectively, these LC/MS/MS data suggest the presence of two cyano groups in the piperidine ring, the most logical positions being R- to the ring nitrogen (Figure 1). Results from the 1H NMR analysis of the bis-cyano adduct III are summarized in Table 2. As compared to I, the main changes in the spectrum of the adduct III were confined to the piperidine ring, which showed little dispersion suggesting a symmetric substitution. The protons R- to the piperidine ring nitrogen of I at position c (2.54 ppm, Figure 1) were shifted downfield to 4.30 ppm in the spectrum of III [confirmed by total correlation spectroscopy (TOCSY) and nuclear Overhauser effect (NOE) experiments], suggesting that each of the two

In Vitro Bioactivation of Dihydrobenzoxathiin SERMs

Chem. Res. Toxicol., Vol. 18, No. 4, 2005 681

Table 2. Proton NMR Data for Compounds I-IV and M12a Ib H2 H3 H5 H7 H8 H2′ H2′/6′ H3′/5′ H4′ H5′ H6′ H2′′/6′′ H3′′/5′′ H-a H-b H-c H-d H-e

IIb

5.37 d 4.37 d 6.56 d 6.50 dd 6.78 d

J ) 2.3 J ) 2.3 J ) 2.9 J ) 2.9, 8.8 J ) 8.8

1H 1H 1H 1H 1H

6.72 d 6.48 d

J ) 8.6 J ) 8.6

2H 2H

6.96 d 6.75 d 4.07 t 2.76 d 2.54 m 1.62 m 1.47 m

J ) 8.5 J ) 8.5 J ) 5.8 J ) 5.8

2H 2H 2H 2H 4H 4H 2H

5.44 d 4.44 d 6.61, d 6.54 dd 6.83 d 6.46 t 6.58 dd 6.91 dd 6.39 d 7.01 d 6.75 d 4.02 t 2.80 t 2.56 m 1.72 m

J ) 2.3 J ) 2.3 J ) 2.9 J ) 2.8, 8.8 J ) 8.6 J ) 2.5 J ) 2.6,8.0 J ) 8.0 J ) 7.7 J ) 8.7 J ) 8.7 J ) 5.8 J ) 5.8

III 1H 1H 1H 1H 1H 1H 1H 1H 1H 2H 2H 2H 2H 4H 4H

IV

5.38 d 4.38 d 6.56 d 6.50 dd 6.78 d

J ) 2.3 J ) 2.3 J ) 2.7 J ) 2.7, 8.8 J ) 8.8

1H 1H 1H 1H 1H

6.73 d 6.49 d

J ) 8.5 J ) 8.5

2H 2H

6.96 d 6.77 d 4.12 t 3.12 t 4.30 m 1.85 m c

J ) 8.5 J ) 8.5 J ) 4.9 J ) 4.9

2H 2H 2H 2H 2H 4H

M12

5.05 d 4.39 d 6.69 d 6.67 dd 7.54 d absent

J ) 8.8 J ) 8.8 J ) 2.6 J ) 2.6, 8.5 J ) 8.5

1H 1H 1H 1H 1H

6.97 d 7.55 d

J ) 8.6 J ) 8.5

1H 6.82 dd J ) 2.6, 8.5 1H 1H 7.48 d J ) 8.5 1H

7.15 d J ) 8.6 6.83 d J ) 8.6 4.31 m c c c

4.51 d 3.87 d 6.60 d 6.62 dd 7.43 d 6.77 d

J ) 7.4 J ) 7.4 J ) 2.6 J ) 2.6, 8.5 J ) 8.5 J ) 2.5

2H 7.03 d J ) 8.5 2H 6.71 d J ) 8.6 2H 4.22 m c c c

1H 1H 1H 1H 1H 1H

2H 2H 2H

a Spectra were acquired at 600 Hz in CD OD for I and III and in CD CN-D O (9:1) for II, IV, and M12. Chemical shifts (ppm), 3 3 2 multiplicity, coupling constants (Hz), and integral. Positions of the denoted hydrogen atoms are shown in Figure 1. b Free base. c Not assigned.

cyano groups were located at that position. Furthermore, an heteronuclear multiple bond correlation (HMBC) experiment with III showed two cross-peaks between the CN (δ 117.6 ppm) and the protons at 4.30 (c) and 1.85 (d) ppm (data not shown). The above observations suggest that two cyano groups were located at the c positions; therefore, the structure of the bis-cyano adduct III was deduced to be as shown in Figure 1. Structural Analysis of Metabolite M12 and its NAc Adduct IV. LC/MS/MS experiments with M12 revealed the parent (MH+) ion at m/z 450, indicating that M12 had the same molecular weight as II. The MS2 CID spectrum of m/z 450 showed fragments corresponding to MH+ - H2O (m/z 432) and MH+ - H2O - pyrrolidine (m/z 361). The facile loss of the elements of water (18 Da) in the spectrum of M12 was not seen in that of II (Table 1 and Supporting Information). The other product ions are listed in Table 1. The behavior of M12 under CID conditions differed appreciably from that of II (Table 1 and Supporting Information), suggesting that the structures of the two compounds were distinct. On the basis of LC/MS/MS data, a possible structure for M12 is depicted in Figure 1. Results from the 1H NMR analysis of the isolated sample of M12 are shown in Table 2. The spectrum of M12 indicated that one aromatic proton had been lost as compared to II and showed two sets of 1,2,4-trisubstituted benzene rings, in addition to the para-disubstituted ring. These two sets were distinguished by an NOE experiment wherein H-3 was irradiated and an enhancement of H-2′ (Figure 1) was observed (22). Furthermore, the oxathiine ring protons in II, H-2 (5.44 ppm) and H-3 (4.44 ppm), were shifted upfield to 4.51 and 3.87 ppm in M12, respectively, and their coupling constants changed significantly from 2.3 to 7.4 Hz (Table 2). The above observations are consistent with the proposed structure of M12 shown in Figure 1. This structure was confirmed by comparison in LC/MS/MS analysis with a synthetic standard (22) (Supporting Information). Four isomeric NAc adducts related to M12 were detected by LC/MS/MS analysis from incubations of II with recombinant P450 3A4 in the presence of NAc (data not shown). The major adduct IV was isolated for further structural characterization. LC/MS/MS analysis revealed

the presence of the MH+ ion at m/z 611, consistent with the addition of the elements of NAc to the parent (Table 1 and Supporting Information). The MS2 spectrum of m/z 611 revealed a loss of the elements of water (18 Da) to give a product ion at m/z 593. The other fragment ion at m/z 482 was ascribed to cleavage adjacent to the thioether moiety. The product ion at m/z 464 resulted from elimination of the elements of water from the fragment ion at m/z 482, similar to the case with M12 (Table 1 and Supporting Information). The LC/MS/MS data suggest that the structure of the adduct IV is closely related to that of M12, and a possible structure for IV is shown in Figure 1. Results from the 1H NMR analysis of the NAc adduct IV are shown in Table 2 and Supporting Information. The spectrum of IV showed that one aromatic proton had been lost as compared with M12. Moreover, the aromatic region of the spectrum indicated that the substitution was either at carbon 5 or 2′ since the remaining two protons of the substituted ring were ortho disposed (Figure 1, Table 2, and Supporting Information). The exact position of substitution remains uncertain but is most likely to be at carbon position 2′ based on chemical shift considerations and comparison with the NMR spectrum of M12 (22) (Table 2 and Supporting Information). Inhibition of Metabolism and Adduct Formation of I and II in Human Liver Microsomes by Antibodies and Chemical Inhibitors. In screening studies using a battery of recombinant P450 enzymes, only P450 2D6 and 3A4 were capable of catalyzing the turnover of I and II (unpublished results). At a substrate concentration of 10 µM, the metabolism of I and II in incubations with human liver microsomes was inhibited completely by a monoclonal antibody against P450 3A4 and a selective P450 3A4 inhibitor ketoconazole (Table 3). No inhibition was observed by a monoclonal antibody against P450 2D6 or by a selective P450 2D6 inhibitor quinidine. Similarly, the formation of the bis-cyano adduct of I and the NAc adducts of II in incubations with human liver microsomes also was completely inhibited by a monoclonal antibody against P450 3A4 and by ketoconazole; no inhibition was observed by a monoclonal antibody against P450 2D6 or by quinidine (Table 3).

682

Chem. Res. Toxicol., Vol. 18, No. 4, 2005

Zhang et al.

Table 3. Inhibitory Effects of Monoclonal Antibodies and Chemical Inhibitors on the Metabolism and Adduct Formation of I and II in Human Liver Microsomal Incubations % control of metabolisma

% control of adduct formation

antibody or inhibitor

concentration

I

II

Ib

IIc

none anti-P450 3A4 antibody anti-P450 3A4 antibody anti-P450 2D6 antibody anti-P450 2D6 antibody ketoconazole quinidine

0.25 mg IgG/nmol P450 2.5 mg IgG/nmol P450 0.25 mg IgG/nmol P450 2.5 mg IgG/nmol P450 1 and 10 µM 5 and 25 µM

100 0 0 100 100 0 100

100 0 0 100 100 0 100

100 0 0 100 100 0 100

100 8 0 100 100 0 100

a The control values for the rate of metabolism of I and II in human liver microsomes were 51 and 131 pmol/min/mg protein, respectively. The incubation time was 45 min. b For the formation of the bis-cyano adduct III of I. c For the formation of the NAc adducts of II.

Discussion Estrogens are known to be metabolized by P450 enzymes to form catechols, which may undergo further oxidation to o-quinones, which either react with DNA to form covalent adducts or induce DNA oxidation (2). The o-quinones also may react with cysteinyl sulfhydryl groups of proteins to form covalent adducts. Similarly, tamoxifen has been shown to be metabolized by P450 enzymes to form reactive tamoxifen-o-quinone and quinone methide intermediates (23, 24). Recently, we have shown that reloxifene is a time-dependent P450 3A4 inhibitor in human liver microsomes, and this inhibition possibly is due to the formation of an arene oxide or a quinone type intermediate (7). It is widely accepted that biotransformation of drugs to chemically reactive intermediates may play a critical role in idiosyncratic drug toxicity (25). One of the postulated mechanisms for such idiosyncratic reactions involves the formation of immunogenic conjugates from the reaction of an electrophilic intermediate with cellular proteins (26). Therefore, it is important to select drug candidates with a low potential for bioactivation early in the drug discovery stage in order to minimize the potential for serious adverse effects later in development (9, 10, 27, 28). Compound I is a SERM with potent in vitro and in vivo ER-R antagonist activities (12, 14). As a part of our strategy to develop safer drugs, I was evaluated for its propensity to covalently modify proteins in human liver microsomal incubations. Appreciable covalent protein binding of I was observed (>1000 pmol equiv/mg protein), which was both time- and NADPH-dependent. A separate SDS-PAGE analysis of proteins derived from the incubation of [3H]I with human liver microsomes confirmed that the association between the proteins and the [3H]I-related radioactivity was irreversible in nature and that the molecular mass of the modified protein(s) was ∼55 kDa. Several potential reactive metabolites of I were identified following incubations with human liver microsomes or recombinant P450 3A4. M1 is a hydroquinone metabolite resulting from oxidative cleavage of the dihydrobenzoxathiin moiety (Scheme 1). It is known that hydroquinones, after further oxidation to the corresponding quinones, alkylate proteins to form covalent adducts (2). M8 was identified tentatively as an o-quinone (Scheme 1), which could contribute to the observed covalent binding (2, 29). In addition, a bis-cyano adduct (III) was detected when incubations were performed in the presence of sodium cyanide (Figure 1). It is likely that this bis-cyano adduct III was formed by reaction of cyanide

with an iminium ion intermediate of I (10, 30). Together, these data suggest that I is subject to at least three bioactivation pathways, two of which appear to lead to quinone metabolites, while the third involves formation of an iminium ion. In support of this hypothesis, it was demonstrated that the binding of I was attenuated significantly by inclusion of either GSH or sodium cyanide in incubations of I with human liver microsomes. Although GSH adducts of hydroquinone and o-quinone derivatives were not detected in these experiments, it is possible that these benzoquinone-GSH adducts were unstable and underwent further cyclization and/or polymerization (31). Separate inhibition studies using monoclonal antibodies and chemical inhibitors showed that P450 3A4 is a major enzyme responsible for the formation of reactive intermediates of I in human liver microsomes. In light of the above findings, more than 150 compounds were synthesized and evaluated for pharmacological activity, pharmacokinetics, and their propensity to undergo metabolic activation (12-18). It appears that structural changes on the benzoxathiin moiety of I (Figure 1) led to either a loss of potency, a loss of selectivity, or poor pharmacokinetics properties. On the other hand, the 3′-hydroxy analogue of I exhibited superior potency to I in vivo in rats. Further substitution of the piperidine side chain with a pyrrolidine moiety led to II that showed optimal in vitro and in vivo potency and selectivity toward ER-R receptors (12, 14). Covalent protein binding of II in human liver microsomes was estimated to be 461 pmol equiv/mg protein, which represents a significant improvement as compared with I. The level of binding associated with II was not significantly attenuated by sodium cyanide but was reduced by approximately 78% by GSH. Subsequent metabolism studies showed that two hydroquinones (M10 and M12) were formed in incubations of II with human liver microsomes and P450 3A4. The major metabolite M12 proved to be a biphenyl hydroquinone derivative. M12 is structurally distinct from that of II, suggesting that II underwent structural rearrangement to form M12. The formation of two hydroquinone species suggests a potential for covalent binding to proteins (2, 29). In addition, trapping studies with sodium cyanide and NAc failed to reveal the formation of cyano adducts of II but demonstrated that four NAc adducts were generated in incubations of II with either human liver microsomes or recombinant P450 3A4. The major NAc adduct (IV) probably was formed by addition of the sulfur nucleophile to a reactive quinone intermediate, while the three minor NAc adducts are thought to represent positional isomers

In Vitro Bioactivation of Dihydrobenzoxathiin SERMs

Chem. Res. Toxicol., Vol. 18, No. 4, 2005 683

Scheme 3. Proposed Mechanism for the P450 3A4-Mediated Conversion of II to the Hydroquinone Metabolite M10, the Extended Biphenyl Hydroquinone Derivative M12, and the Dienone M13

Scheme 4. Proposed Mechanism for the P450 3A4-Mediated Conversion of II to the Aldehyde Metabolite M9 and the S-Oxide Metabolite M11

of IV. Structural characterization showed that IV is a NAc adduct of the biphenyl hydroquinone M12 (Figure 1). These results suggest that II also undergoes bioactivation to form reactive quinone species in human liver microsomes and that the bioactivation pathway is associated with the formation of M12. Separate inhibition studies using a monoclonal antibody against P450 3A4 suggested that P450 3A4 was a major enzyme responsible for the bioactivation of II in human liver microsomes. The formation of hydroquinones M10, M12, and the dienone metabolite M13 in human liver microsomes may be catalyzed by P450 3A4 through a two-electron oxidation process to afford the quinonium cation intermediate V (32) (Scheme 3). The para-quinone intermediate VI, in turn, likely is formed through hydrolysis of the quinonium cation V (32) (Scheme 3). Alternatively, VI could be formed through a two-electron oxidation of II with an initial oxidation on the sulfur atom to give the

radical cation XI (33-36) (Scheme 4). It may be proposed that the para carbon atom of the phenolic ring of VI undergoes intramolecular nucleophilic addition to the quinone carbonyl, followed by dehydration of the product VII to give the extended quinone intermediate VIII (Scheme 3). The intermediate VIII then is reduced to give M12 or is captured as the NAc adduct IV in the presence of the trapping agent. The reduction of VI affords the hydroquinone metabolite M10 (Scheme 3). The para carbon atom of the phenolic ring of the quinonium cation V also may undergo intramolecular nucleophilic addition to the quinonium carbonyl to give the dienone intermediate X. Tautomerization of X yields the dienone M13. A similar mechanism can be invoked to explain the formation of the aldehyde metabolite M9 and the S-oxide M11 (Scheme 4). In this case, the sulfur cation radical XI undergoes oxygen rebound with the P450 (FeO)2+ species to give the S-oxide M11. Alternatively, rearrangement

684

Chem. Res. Toxicol., Vol. 18, No. 4, 2005

of XI with heterocyclic ring scission to XII, followed by cleavage of the C-O bond, would afford M9 (Scheme 4). It is also possible that P450 3A4 may initiate the formation of the diradical intermediate IX through a twoelectron oxidation process, followed by a phenolic coupling to give the dienone intermediate X (Scheme 3). This reaction is analogous to the P450-catalyzed biotransformation of (R)-reticline to salutaridine in morphine biosynthesis (37, 38). Cleavage of the ether bond of X then would afford the extended quinone intermediate VIII (Scheme 3). Compound II was further evaluated for its propensity to covalently bind to proteins in rat in vivo, when it was shown that the values were less than 10 pmol equiv/mg protein in both plasma and liver (unpublished results). This lower level of covalent protein binding in rats may be explained by the fact that glucuronidation is the major mechanism of clearance of II in rats (unpublished results). In addition, II exhibited low covalent protein binding in human hepatocytes, as compared to I. In light of these findings, II was selected as a new lead compound in the SERM program (14, 18). In conclusion, our results indicate that I is subject to P450 3A4-catalyzed bioactivation via three pathways, including formation of quinone and iminium ion intermediates, all of which may participate in the covalent modification of human liver microsomal proteins. On the basis of this information, structural modification of I led to the identification of II as a new lead for further development. Compound II exhibited a reduced potential for bioactivation as compared with I and retained desirable pharmacological properties. The data presented in this report illustrate an example of a drug discovery program in which the risk of bioactivation, potentially leading to adverse effects, was minimized during the lead optimization phase prior to selection of a lead candidate for development.

Acknowledgment. We acknowledge Dr. Yui-Sing Tang, Dr. Ashok Chaudhary, Wensheng Liu, and Dr. Dennis Dean (Merck Research Laboratories) for the synthesis of tritium-labeled compounds. Supporting Information Available: Mass spectra and proposed fragmentation patterns of compounds I-IV and metabolites M1-M15, the NMR spectra of II, IV, M1, and M10, and the synthetic routes to metabolites M1 and M10. This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) Dutertre, M., and Smith, C. L. (2000) Molecular mechanisms of selective estrogen receptor modulator (SERM) action. J. Pharmacol. Exp. Ther. 295, 431-437. (2) Bolton, J. L., Trush, M. A., Penning, T. M., Dryhurst, G., and Monks, T. J. (2000) Role of quinones in toxicology. Chem. Res. Toxicol. 13, 135-160. (3) Jordan, V. C. (2003) Antiestrogens and selective estrogen receptor modulators as multifunctional medicines. 1. Receptor interactions. J. Med. Chem. 46, 883-908. (4) Jordan, V. C. (2003) Antiestrogens and selective estrogen receptor modulators as mutifunctional medicines. 2. Clinical considerations and new agents. J. Med. Chem. 46, 1081-1111. (5) Mani, C., and Kupfer, D. (1991) Cytochrome P-450-mediated activation and irreversible binding of the antiestrogen tamoxifen to proteins in rat and human liver: possible involvement of flavincontaining monooxygenases in tamoxifen activation. Cancer Res. 51, 6052-6058. (6) Shibutani, S., Suzuki, N., Terashima, I., Sugarman, S. M., Grollman, A. P., and Pearl, M. L. (1999) Tamoxifen-DNA adducts

Zhang et al. detected in the endometrium of women treated with tamoxifen. Chem. Res. Toxicol. 12, 646-653. (7) Chen, Q., Ngui, J. S., Doss, G. A., Wang, R. W., Cai, X., Dininno, F. P., Blizzard., T. A., Hammond, M. L., Stearns, R. A., Evans, D. C., Baillie, T. A., and Tang, W. (2002) Cytochrome P450 3A4mediated bioactivation of reloxifene: irreversible enzyme inhibition and thiol adduct formation. Chem. Res. Toxicol. 15, 907914. (8) Liu, X., Pisha, E., Tonetti, D. A., Yao, D., Li, Y., Yao, J., Burdette, J. E., and Bolton, J. L. (2003) Antiestrogenic and DNA damaging effects induced by tamoxifen and toremifene metabolites. Chem. Res. Toxicol. 16, 832-837. (9) Baillie, T. A., and Kassahun, K. (2001) Biological reactive intermediates in drug discovery and development: A perspective from the pharmaceutical industry. Adv. Exp. Med. Biol. 500, 4551. (10) Evans, D. C., Watt, A. P., Nicoll-Griffith, D. A., and Baillie, T. A. (2004) Drug-protein adducts: An industry perspective on minimizing the potential for drug bioactivation in drug discovery and development. Chem. Res. Toxicol. 17, 3-16. (11) Chen, H. Y., Dykstra, K. D., Birzin, E. T., Frisch, K., Chan, W., Yang, Y. T., Mosley, R. T., DiNinno, F., Rohrer, S. P., Schaeffe, J. M., and Hammond, M. L. (2004) Estrogen receptor ligands. Part 1: The discovery of flavanoids with subtype selectivity. Bioorg. Med. Chem. Lett. 14, 1417-1421. (12) Kim, S., Wu, J. Y., Birzin, E. T., Frisch, K., Chan, W., Pai, L., Yang, Y. T., Mosley, R. T., Fitzgerald, P. M. D., Sharma, N., Dahllund, J., Thorsell, A., Dininno, F., Rohrer, S. P., Schaeffer, J. M., and Hammond, M. L. (2004) Estrogen receptor ligands. II. Discovery of benzoxathiins as potent, selective estrogen receptor R modulators. J. Med. Chem. 47, 2171-2175. (13) Kim, S., Wu, J. Y., Chen, H. Y., and DiNinno, F. (2003) Dehydrativereduction: A highly diastereoselective synthesis of syn-bisaryl(or heteroaryl) dihydrobenzoxathiins and benzodioxane. Org. Lett. 5, 686-688. (14) Kim, S., Wu, J., Chen, H. Y., Birzin, E. T., Chan, W., Yang, Y. T., Colwell, L., Li, S., Dahllund, J., Dininno, F., Rohrer, S. P., Schaeffer, J. M., and Hammond, M. L. (2004) Estrogen receptor ligands. Part 4. The SAR of the syn-dihydrobenzoxathiin SERMs. Bioorg. Med. Chem. Lett. 14, 2741-2745. (15) Blizzard, T. A., DiNinno, F. P., Morgan, J. D., II, Chen, H. Y., Wu, J. Y., Gude, C., Kim, S., Chan, W., Birzin, E. T., Yang, Y. T., Pai, L.-Y., Zhang, Z., Hayes, E. C., DaSilva, C., Tang, W., Rohrer, S. P., Schaeffer, J. M., and Hammond, M. L. (2004) Estrogen receptor ligands: Part 7. Dihydrobenzoxathiin SERMs with bicyclic amine side chains. Bioorg. Med. Chem. Lett. 14, 38613864. (16) Blizzard, T. A., DiNinno, F. P., Morgan, J. D., II, Wu, J. Y., Chen, H. Y., Gude, C., Kim, S., Chan, W., Birzin, E. T., Yang, Y. T., Pai, L.-Y., Zhang, Z., Hayes, E. C., DaSilva, C., Tang, W., Rohrer, S. P., Schaeffer, J. M., and Hammond, M. L. (2004) Estrogen receptor ligands: Part 8. dihydrobenzoxathiin SERMs with heteroatom substituted side chains. Bioorg. Med. Chem. Lett. 14, 3865-3868. (17) Blizzard, T. A., DiNinno, F. P., Morgan, J. D., II, Chen, H. Y., Wu, J. Y., Kim, S., Chan, W., Birzin, E. T., Yang, Y. T., Pai, L.Y., Fitzgerald, P. M. D., Sharma, N., Li, Y., Zhang, Z., Hayes, E. C., DaSilva, C., Tang, W., Rohrer, S. P., Schaeffer, J. M., and Hammond, M. L. (2005) Estrogen receptor ligands: Part 9. dihydrobenzoxathiin SERMs with alkyl substituted pyrrolidine side chains and linkers. Bioorg. Med. Chem. Lett. 15, 107-113. (18) Song, Z. J., King, A. O., Waters, M. S., Lang, F., Zewge, D., Bio, M., Leazer, J. L., Jr., Javadi, G., Kassim, A., Tschaen, D. M., Reamer, R. A., Rosner, T., Chilenski, J. R., Mathre, D. J., Volante, R. P., and Tillyer, R. (2004) An efficient asymmetric synthesis of an estrogen receptor modulator by sulfoxide-directed borane reduction. Proc. Natl. Acad. Sci. U.S.A. 101, 5776-5781. (19) Mei, Q., Tang, C., Assang, C., Slaughter, D., Rodrigues, D. A., Rushmore, T. H., and Shou, M. (1999) Role of a potent inhibitory monoclonal antibody to cytochrome P-450 3A4 in assessment of human drug metabolism. J. Pharmacol. Exp. Ther. 291, 749759. (20) Tang, W., Stearns, R. A., Bandiera, S. M., Zhang, Y., Raab, C., Braun, M. P., Dean, D. C., Pang, J., Leung, K. H., Doss, G., Strauss, J. R., Kwei, G. Y., Rushmore, T. H., Chiu, S. H. L., and Baillie, T. A. (1999) Studies on cytochrome P-450-mediated bioactivation of diclofenac in rats and in human hepatocytes: Identification of glutathione conjugated metabolites. Drug Metab. Dispos. 27, 365-372.

In Vitro Bioactivation of Dihydrobenzoxathiin SERMs (21) Raucy, J. L., and Lasker, J. M. (1991) Isolation of P450 enzymes from human liver. Methods Enzymol. 206, 557-587. (22) Kim, S., Wu, J. Y., Zhang, Z., Tang, W., Doss, G. A., Dean, B. J., DiNinno, F. P., and Hammond, M. L. (2005) Estrogen receptor ligands. Part 12: Synthesis of the major metabolites of an ER R-selective dihydrobenzoxathinn antagonist for osteoporosis. Org. Lett. 7, 411-414. (23) Zhang, F., Fan, P. W., Liu, X., Shen, L., van Breeman, R. B., and Bolton, J. L. (2000) Synthesis and reactivity of a potential carcinogenic metabolite of tamoxifen: 3,4-Dihydroxytamoxifeno-quinone. Chem. Res. Toxicol. 13, 53-62. (24) Marques, M. M., and Beland, F. A. (1997) Identification of tamoxifen-DNA adducts formed by 4-hydroxytamoxifen quinone methide. Carcinogenesis 18, 1949-1954. (25) Williams, D. P., and Naisbitt, D. J. (2002) Toxicophores: Groups and metabolic routes associated with increased safety risk. Curr. Opin. Drug Discovery Dev. 5, 104-115. (26) Ju, C., and Uetrecht, J. P. (2002) Mechanism of idiosyncratic drug reactions: Reactive metabolite formation, protein binding and the regulation of the immune system. Curr. Drug Metab. 3, 367-377. (27) Baillie, T. A., Cayen, M. N., Fouda, H., Gerson, R. J., Green, J. D., Grossman, S. J., Klunk, L. J., LeBlanc, B., Perkins, D. G., and Shipley, L. A. (2002) Drug metabolites in safety testing. Toxicol. Appl. Pharmacol. 182, 188-196. (28) Uetrecht, J. (2003) Screening for the potential of a drug candidate to cause idiosyncratic drug reactions. Drug Discovery Today 8, 832-837. (29) Iverson, S. L., Shen, L., Anlar, N., and Bolton, J. L. (1996) Bioactivation of estrone and its catechol metabolites to quinoidglutathione conjugates in rat liver microsomes. Chem. Res. Toxicol. 9, 492-499. (30) Masumoto, H., Ohta, S., and Hirobe, M. (1991) Application of chemical cytochrome P-450 model systems to studies on drug

Chem. Res. Toxicol., Vol. 18, No. 4, 2005 685 metabolism. IV. Mechanism of piperidine metabolism pathways via an iminium intermediate. Drug Metab. Dispos. 19, 768-780. (31) Alt, C., and Eyer, P. (1998) Ring addition of the R-amino group of glutathione increases the reactivity of benzoqinone thioethers. Chem. Res. Toxicol. 11, 1223-1233. (32) He, K., Takaat, R. E., and Woolf, T. F. (2004) Incorporation of an oxygen from water into troglizazone quinone by cytochrome P450 and myeloperoxidase. Drug Metab. Dispos. 32, 442-446. (33) Ortiz de Montellano, P. R., and De Voss, J. J. (2002) Oxidizing species in the mechanism of cytochrome P450. Nat. Prod. Rep. 19, 477-493. (34) Kobayashi, S., Nakano, M., Kimura, T., and Schaap, A. P. (1987) On the mechanism of the peroxidase-catalyzed oxygen-transfer reaction. Biochemistry 26, 5019-5022. (35) Doerge, D. R., Pitz, G. L., and Root, D. P. (1987) Organosulfur oxygenation and suicide inactivation of lactoperoxidase. Biochem. Pharmacol. 36, 972-974. (36) Perez, U., and Dunford, H. B. (1990) Transient-state kinetics of the reactions of 1-methoxy-4-(methylthio)benzene with horseradish peroxidase compounds I and II. Biochemistry 29, 2757-2763. (37) Amann, T., and Zenk, M. H. (1991) Formation of the morphine precursor salutaridine is catalyzed by a cytochrome P-450 enzyme in mammalian liver. Tetrahedron Lett. 32, 3675-3678. (38) Amann, T., Roos, P. H., Huh, H., and Zenk, M. H. (1995) Purification and characterization of a cytochrome P450 enzyme from pig liver, catalyzing the phenol oxidative coupling of (R)reticuline to salutaridine, the critical step in porphine biosynthesis. Heterocycles 40, 425-440.

TX0496789