In Vitro Metabolism of N - American Chemical Society

Chem. Res. Toxicol. 1997, 10, 733-741

733

Articles In Vitro Metabolism of N-(5-Chloro-2-methylphenyl)-N′-(2-methylpropyl)thiourea: Species Comparison and Identification of a Novel Thiocarbamide-Glutathione Adduct Gregory J. Stevens, Karen Hitchcock, Y. Karen Wang, Gary M. Coppola, Richard W. Versace, Jefferson A. Chin, Michael Shapiro, Somchai Suwanrumpha, and Bonnie L. K. Mangold* Preclinical Safety and Research Departments, Novartis Pharmaceuticals Corporation, 59 Rt 10, East Hanover, New Jersey 07936 Received February 17, 1997X

The in vitro metabolism of SDZ HDL 376, a thiocarbamide developed for the treatment of atherosclerosis, was investigated in rat, dog, monkey, and human liver microsomes, as well as in rat and human liver slices. [14C]SDZ HDL 376 was extensively metabolized in all the species except human. In rat liver microsomes an S-oxide was the major metabolite. In human and monkey microsomes, carbon hydroxylation was favored. The NADPH-dependent oxidation of SDZ HDL 376 resulted in covalent binding to microsomal protein. Addition of GSH to the incubations decreased protein binding in a concentration-dependent manner and resulted in a novel SDZ HDL 376-GSH adduct. Adduct formation required NADPH and was mediated predominately by cytochrome P450. Inhibition of cytochrome P450 by 1-aminobenzotriazole resulted in a 95% decrease in adduct formation, while heat inactivation of flavin-containing monooxygenases resulted in a 10% decrease. Unlike other thiocarbamides which form disulfide adducts with GSH, the SDZ HDL 376 adduct contained a thioether linkage as characterized by LC/MS/MS and reference to a synthetic standard. Reactions performed with [35S]GSH resulted in a [35S]SDZ HDL 376-GSH adduct, demonstrating the sulfur was derived from GSH. Adduct formation was faster in rat microsomal reactions compared to human microsomes. Other structurally unrelated thiocarbamides (phenylthiourea, methimazole, 2-mercaptobenzimidazole, 2-mercaptoquinazoline, and 2-propyl-6-thiouracil) did not form similar adducts in rat liver microsomes supplemented with GSH. Therefore, the GSH adduct of SDZ HDL 376 is unique for this type of thiocarbamide. These results suggest that the bioactivation and detoxification of SDZ HDL 376 differ significantly from other thiocarbamides. Furthermore, the in vitro formation of S-oxides and GSH adducts in rat hepatic tissue, and ring hydroxylation and glucuronidation in human hepatic tissue, suggests rats may be more susceptible to the toxicity of SDZ HDL 376 compared to humans.

Introduction Thiocarbamides are widely used as insecticides, as therapeutic agents, and in the manufacturing of plastics. The toxicological effects of thiocarbamides include thyroid depression, pulmonary edema, and liver necrosis (1). Many of these adverse effects are attributed to the formation of sulfoxide intermediates catalyzed primarily by microsomal cytochrome P450 (P450)1 and flavincontaining monooxygenases (FMO) (1-6). Initial oxidation results in the formation of electrophilic sulfenic acids (S-oxide), which can be further oxidized to form sulfinic (S-dioxide) and sulfonic (S-trioxide) acids. These result* To whom correspondence and reprint requests should be addressed at Dept. of Drug Metabolism, Bldg. 405, Novartis Pharmaceuticals Corp., 59 Rt 10, East Hanover, NJ 07936. Phone: (201) 503-6751. Fax: (201) 503-6076. X Abstract published in Advance ACS Abstracts, June 15, 1997. 1 Abbreviations: ETU, ethylenethiourea; ABT, 1-aminobenzotriazole; FMO, flavin-containing monooxygenase; SDZ HDL 376, N-(5chloro-2-methylphenyl)-N′-(2-methylpropyl)thiourea; P450, cytochrome P450.

S0893-228x(97)00023-4 CCC: $14.00

ing S-oxides can be hydrolyzed to the corresponding urea (2, 3). Many of the S-oxide intermediates are extremely labile and difficult to isolate and identify. Based on their reactivity, the sulfur oxidation products are believed to be the reactive intermediates responsible for thiocarbamide-induced toxicity (1-6). Metabolic activation of thiocarbamides results in reactive intermediates capable of protein binding and enzyme inhibition (7, 8). In rat hepatic microsomes, the thiocarbamide ethylenethiourea (ETU) was shown to bind to microsomal protein and inactivate P450 (9). These studies also demonstrated that the addition of GSH completely inhibited the enzyme inactivation and resulted in the formation of GSSG. It is reported that GSH reduces S-oxides via an intermediate glutathionyldisulfide adduct (9, 10). In the presence of GSH, oxidized glutathione is formed and the S-oxides are reduced to the parent thiocarbamide (2, 9, 10). These results are supported by reports of increased biliary GSSG efflux from rat livers perfused with thiourea, phenylthiourea, © 1997 American Chemical Society

734 Chem. Res. Toxicol., Vol. 10, No. 7, 1997

and methimazole (10). This “futile cycling” may result in rapid depletion of free GSH and ultimately of NADPH. The depletion of cellular GSH and protein adduct formation may be partly responsible for thiocarbamide-induced liver necrosis (8-10). SDZ HDL 376 (N-(5-chloro-2-methylphenyl)-N′-(2methylpropyl)thiourea) is a thiocarbamide originally developed for the treatment of atherosclerosis. The present study was undertaken to determine the extent of SDZ HDL 376 metabolism in microsomal incubations, as well as in precision cut liver slices. Oxidation of SDZ HDL 376 was evaluated in microsomes prepared from rat, dog, monkey, and human liver. Protein binding was assessed in rat and human microsomal incubations with and without the addition of GSH. In addition to identifying the major oxidative metabolites of SDZ HDL 376, the structural elucidation of a novel thiocarbamide-GSH adduct is described.

Experimental Procedures Chemicals. SDZ HDL 376 was provided by the Research Department of Novartis Pharmaceuticals (East Hanover, NJ) and was 98% pure as determined by HPLC. β-Glucuronidase, sulfatase, and NADPH were obtained from Sigma Chemical Co. (St. Louis, MO). All other reagents were of the highest grade available. All synthetic procedures should be carried out in a well-ventilated hood. Gloves should also be worn as a precaution for exposure to solutions of organic compounds in DMSO. Synthesis of GSH-Adduct. The synthesis of the SDZ HDL 376-GSH adduct was accomplished by initial formation of a carbodiimide intermediate followed by reacting the intermediate with GSH. The carbodiimide intermediate was prepared by adding 2.3 g of methanesulfonyl chloride dropwise to a solution containing SDZ HDL 376 (2.57 g), triethylamine (3.0 g), and 4-(dimethylamino)pyridine (50 mg) in 100 mL of methylene chloride. After stirring for 5 min at room temperature, the solvent was removed under reduced pressure and the residual oil was filtered twice through a pad of silica gel using methylene chloride to elute the product. The solvent was removed under reduced pressure to give 2.3 g (100%) of the desired carbodiimide: IR (neat) 2144 cm-1; 1H NMR (CDCl3 δ 7.10-6.92 (m, 3H), 3.27 (d, 2H), 2.23 (s, 3H), 1.92 (m, 1H), 1.00 (d, 6H). Synthesis of the GSH adduct was accomplished by adding carbodiimide (200 mg) to triethylamine (180 mg) and GSH (262 mg) in 8.0 mL of Me2SO. The suspension was stirred for 5 min at room temperature and then heated until a solution formed. Stirring was continued at room temperature for an additional 2 h. Methyl tert-butyl ether was added to the solution until a gum was formed. The solvent was decanted from the gum to afford 400 mg of crude product. The crude product (200 mg) was purified using a preparative Hamilton PRP1 column (21.5 mm × 25 cm) and eluted with acetonitrile and water using a linear gradient from 0% to 20% water for 30 min. Pure product (45 mg) was analyzed by MS/MS and NMR: 1H NMR (Me2SO) δ 0.9 (d, isobutyl-CH3), 1.95 (m, -CH), 1.9-2.0 (m, -CH2), 2 (s, Ar-CH3), 2.3-2.4 (m, -CH2), 2.9-3.4 (m, -CH2), 3.1-3.15 (m, -CH2), 3.35 (m, Glu-CH), 3.63-3.67 (m, Gly-CH2), 4.4 (m, CysCH), 7.05 (d, Ar-H), 6.8 (d, Ar-H), 6.6 (s, Ar-H). 13C NMR (D2O) δ 15.6, 20.4, 27.4, 28.9, 32.7, 33.9, 44.6, 52.9, 53.5, 55.3, 127.6, 132.9, 133.7, 134.9, 171.6, 174.9, 176, 177.2. Hepatic Microsome Preparation. Fresh livers from four male Sprague-Dawley rats were not frozen but homogenized immediately upon harvest as described below. Liver pieces (6 g) from three female Beagle dogs and liver pieces (10 g) from three male Cynomolgus monkeys were frozen in liquid nitrogen and stored at -80 °C until use. The tissues were thawed and homogenized in buffer (50 mM Tris-HCl, pH 7.4, 0.25 M sucrose, 154 mM KCl, 1.0 M EDTA) at 4 °C. All subsequent steps were performed at 4 °C. Homogenate from each of the corresponding species was pooled and centrifuged at 9000g for 20 min. The resulting supernatant was centrifuged at 105000g for 1 h. The

Stevens et al. microsomal pellets were washed in homogenization buffer containing 0.1 M sodium pyrophosphate at pH 7.4 and centrifuged at 105000g for 1 h. Microsomal pellets were resuspended in storage buffer (0.1 M phosphate, pH 7.4, containing 20% (v/ v) glycerol) and stored at -80 °C. Human liver tissue was obtained from two separate donors and provided by The Stanford Research Institute (Palo Alto, CA). Human liver-15 (HL-15) was obtained from a healthy Caucasian male, aged 17, who died in a motor vehicle accident. HL-22 was obtained from a 29 year old Caucasian female who died from a gun shot wound to the head. Medical history of this individual indicated alcohol abuse, iv drug use, a history of psychological problems, and smoking. Upon receipt, human liver specimens were stored at -80 °C until use. Total P450 was determined spectrophotometrically by the method of Omura and Sato (11) using 91 cm2/mmol as the extinction coefficient of P450. Protein was determined by the method of Bradford (12) using bovine serum albumin as the standard. Microsomal Incubations with SDZ HDL 376 and Other Thiocarbamides. Microsomal incubations (total volume 1.0 mL) consisted of 50 mM phosphate buffer (pH 7.4), 75 mM KCl, 0.75 mM EDTA, 2.5 mM MgCl2, 0.1 mM [14C]SDZ HDL 376 (33.2 µCi/mg) in 0.2% (v/v) Me2SO and microsomes (1-2 nmol of P450/ mL). Reactions were initiated with addition of NADPH (1 mM) and allowed to incubate at 37 °C for 30-45 min in a shaking water bath. In some incubations, either unlabeled GSH or [35S]GSH was present at 100 or 500 µM (3.4 or 6.8 µM [35S]GSH at 413 Ci/mmol diluted with unlabeled GSH). For rate experiments, adduct formation was assessed at 3, 6, 9, 12, and 15 min. Only the linear portion was used to calculate the rate (rat 3-9 min; other species were linear over the entire 15 min). Adduct formation was less than 20% for rat and less than 10% for primate and human for the linear portion of the curve. For heat inactivation studies, microsomes were incubated at 37 °C for 1 h in the absence of NADPH prior to incubation with SDZ HDL 376. Microsomal reactions with 1.0 mM 1-aminobenzotriazole (ABT) were preincubated with an NADPH-regenerating system (0.5 mM NADP+, 10 mM glucose-6-phosphate, 1 unit of glucose6-phosphate dehydrogenase, 50 mM HEPES buffer (pH 7.6), 0.1 mM EDTA, 15 mM MgCl2) for 10 min prior to the addition of SDZ HDL 376. Additional incubations were performed with [35S]GSH (100 µM) and 100 µM methimazole, 2-mercaptobenzimidazole, phenylthiourea, 6-propyl-2-thiouracil, or 2-mercaptoquinazoline as the thiourea substrates. Reactions were terminated with the addition of acetonitrile (250 µL). The reaction mixtures were centrifuged at 15 000 rpm for 5 min in an Eppendorf centrifuge at room temperature. The supernatants were stored at -80 °C until analysis by HPLC and/or LC/ MS/MS. Protein Binding. Protein pellets from microsomal incubations were washed with 1.0 mL of acetonitrile and twice with 1.0 mL of methanol. The washed pellets were resuspended in water, and protein concentration was determined as described above. A 1.0 µg sample was mixed with loading buffer (12.5 mM Tris-HCl, pH 6.8, 2% (w/v) sodium dodecyl sulfate, 8% (v/ v) glycerol, 2.5% (v/v) 2-mercaptoethanol, 0.003% (w/v) bromophenol blue) and boiled for 6 min. The samples were loaded on a 10% (w/v) polyacrylamide minigel (Bio-Rad, Cambridge, MA) and run at 180 V for 1 h. The gel was dried under vacuum at 80 °C for 45 min and exposed to a phosphoimager screen (Molecular Dynamics, Sunnyvale, CA) for 3 days. The nmol of [14C]SDZ HDL 376 bound/mg of protein was based on the amount of radioactivity present in each well as determined by image analysis. Liver Slice Preparation. Human liver slices were obtained from the Keystone Skin Bank (Exton, PA) from donor #1024931, a 22 year old Caucasian female who died from dural-sinus thrombosis. There was no history of substance abuse, diabetes, smoking, or liver disease. The liver slices were shipped in cold Beltzers solution at 4 °C. Upon receipt, slices were preincubated in 1.7 mL of Waymouth's MB 752/1 medium (phenol red free; Gibco BRL, Gaithersburg, MD) supplemented with 10 mM

Metabolism of SDZ HDL 376 HEPES, 5 mM sodium bicarbonate, 10 mM NaCl, 84 µg/mL gentamycin sulfate, 0.35 mg/mL L-glutamine, and 10% fetal calf serum. Tissue slices were incubated on screens in roller cultures under an oxygenated atmosphere (95:5 O2:CO2) at 37 °C for 1 h. The medium was replaced with warm medium containing [14C]SDZ HDL 376 (0.1 mM) and incubated for 24 or 48 h. Following incubation, the media from three liver slices were combined, filtered through a 0.2 µm filter, and stored at -80 °C until analysis by HPLC. Incubations without slices were conducted to assess the stability of the parent drug under identical incubation conditions. Rat liver slices were prepared from male Sprague-Dawley rats (300-450 g) as previously described (13). Slices (20 ( 5 mg) were incubated with [14C]SDZ HDL 376 as described above except there was no preincubation of the slices. Incubations without slices were also conducted to assess the stability of the parent drug under similar conditions. Prior to analysis, conjugated metabolites in the medium were hydrolyzed. β-Glucuronidase (60 units, 200 units/mL) or sulfatase (2.5 mg in 100 µL) was incubated with 300 or 250 µL, respectively, of the 24 h rat liver slice medium. For human liver slices, β-glucuronidase (80 units, 200 units/mL) or sulfatase (4 mg in 150 µL) was incubated with 400 µL of medium from 24 h slice incubations. Incubations were performed at 37 °C in a shaking water bath for approximately 16 h. The mixtures were centrifuged at 14 000 rpm for 5 min and the supernatants stored at -80 °C. HPLC Analysis. Metabolites were separated using a Waters HPLC system (Milford, MA) equipped with a model 625 pump, 600E controller, Wisp autoinjector (model 715), and a model 991 photodiode array detector. The system was controlled by a computer equipped with Waters Powerline Control software. Radioactivity was monitored with an IN/US β-RAM flowthrough radioisotope detector (IN/US Systems, Tampa, FL). Floscint II (Packard Instrument Co., Meriden, CT), at a flow rate of 3.0 mL/min, was used as the scintillant. Samples were analyzed on a Zorbax RX-C18 5 µm column (4.6 mm × 25 cm) at a flow rate of 1.0 mL/min. A gradient of 0.1 M ammonium acetate and acetonitrile was used. Initial conditions were 20% acetonitrile and 80% 0.1 M ammonium acetate. Upon injection of sample, a 30 min linear gradient to 100% acetonitrile was performed. Analysis of liver slice incubations was performed with initial conditions of 100% 0.1 M ammonium acetate for 10 min. This was followed by a linear gradient to 30% acetonitrile in 20 min and then to 100% acetonitrile for 10 min. MS Analysis. Thermospray LC/MS and LC/MS/MS experiments were performed on a Finnigan MAT TSQ 70 triple-stage quadrupole spectrometer (Finnigan MAT, San Jose, CA) with a Finnigan thermospray LC/MS interface. A typical vaporizer temperature was 105 °C, and a source block temperature was 220 °C. Argon was used as the collision gas in the MS/MS experiments. The collision cell pressure was 1.8 mTorr and the collision energy was -28 to -30 eV. The HPLC conditions are described above but at a flow rate of 0.8 mL/min. The mass spectra of the GSH adducts were taken on a Sciex API III instrument (Thornhill, Canada) equipped with a pneumatically assisted electrospray interface. Argon was used as the collision gas in the collision-induced dissociation experiments. The argon density was maintained at 5 × 1014 atoms/ cm2, while the collision energy was 80 eV. NMR Analysis. NMR data were measured using a Bruker AM500 spectrometer operating at 500.13 MHz for 1H NMR and at 125.77 MHz for 13C NMR with a 5 mm dual tuned probe (1H, 13C). The temperature of the sample was regulated at 305 K. The 90° 1H pulse for the probe was 13.3 µs, the 90° 13C pulse was 6.6 µs, and CPD decoupling pulse was 106 µs. Proton spectra were recorded with a 45° observation pulse and spectral width of 2911 Hz. For dual-tube experiments, half the sample was dissolved into CH3OD and the other half into CH3OH. The sample was placed into a coaxial NMR tube with the CH3OD solution placed in the interior compartment, and carbon data were collected. The carbon spectrum was recorded using a 45° observation pulse

Chem. Res. Toxicol., Vol. 10, No. 7, 1997 735

Figure 1. HPLC radiochromatograms obtained from [14C]SDZ HDL 376 metabolism in (A) rat, (B) dog, (C) monkey, and (D) human (HL-15) hepatic microsomes. Reactions were performed at 2 nmol of P450/mL for 30 min at 37 °C with a substrate concentration of 0.1 mM. Individual peaks were designated a-h based on order of elution. and a spectral width of 29 411 Hz. Data points (64K) were collected and Fourier transformed using an exponential multiplication of 1 Hz. CH3OH (49 ppm) was used as the internal reference.

Results Microsomal Metabolism of SDZ HDL 376. The metabolism of SDZ HDL 376 was assessed in hepatic microsomes from rats, dogs, monkeys, and humans (Figure 1). Extensive metabolism of SDZ HDL 376 was observed in all species tested except for human. In the absence of NADPH, essentially no metabolites were formed (data not shown). The individual peaks (a-h) were similar across species; however, the amount of metabolite generated differed. Scheme 1 shows the proposed in vitro metabolic pathway for SDZ HDL 376 and the structures of the proposed metabolites. Structures for seven of the metabolites were proposed based on LC/MS and LC/MS/MS data (Table 1). For all the following protonated molecules, the m/z is given for the 35Cl isotope. The 37Cl isotope was present in all cases. Under the MS conditions used for metabolite identification, SDZ HDL 376 produced a protonated molecule (MH+) at m/z 257. Following collision of the precursor ion at m/z 257, product ions at m/z 184 (MH+ - C4H11N), 142 (MH+ - C5H9NS), 116 (C5H10NS+), 74 (C4H12N+), and

736 Chem. Res. Toxicol., Vol. 10, No. 7, 1997 Scheme 1. Proposed in Vitro Metabolism of SDZ HDL 376a

a The metabolite designation (a-h) is the same as used for Figure 1 and Table 1. The position of 14C is marked with an asterisk

Table 1. Protonated Molecules (Precursor ions) and Product Ions of SDZ HDL 376 and in Vitro Metabolites +2H 201

35Cl

H+ 257

184 +2H 142 S NH

NH

CH3 +2H 74

57

116 35Cl-protonated

metabolite designationa

molecules (m/z)

product ions (m/z)

SDZ HDL 376 a b c d e f g h

257 225 257 273 273 273 239 241 225

184, 142, 116, 74, 57 169, 57 239, 90, 72 255, 184, 142, 114, 90, 72 164, 140, 57 200, 158, 74, 57 168, 72, 55 142, 107, 72, 57 169, 57

a Metabolites were designated a-h based on chromatographic retention time as shown in Figure 1.

57 (C4H9+) were observed. Metabolite a, which was formed in all the species, gave a protonated molecule (MH+) at m/z 225. The product ion spectra for metabolite a was identical to that for metabolite h. A structure for metabolite a was not proposed due to its potential instability under the MS/MS conditions. Metabolite b

Stevens et al.

was a minor product and postulated to be the urea of SDZ HDL 376 with a hydroxyisobutyl side chain. Peak c was identified as the N-(2-methyl-2-hydroxypropyl) metabolite of SDZ HDL 376 (MH+, m/z 273). This metabolite produced similar product ions and had an identical retention time to that of a synthetic standard (data not shown). Of interest was metabolite d, which was a major metabolite in the rat and minor in the other species. This metabolite, with a m/z at 273 (MH+ 257 + 16), was postulated to be the S-oxide of HDL 376. Similar to the S-oxidation of other thiocarbonyl compounds (14), the maximum absorbance of the metabolite was shifted from 240 to 265 nm. In addition, this metabolite was identical in HPLC retention time and product ion mass spectra to that of a product generated by oxidation of SDZ HDL 376 with H2O2 (data not shown). Consistent with the instability of thiourea S-oxides (15), this metabolite peak decomposed upon standing to SDZ HDL 376 and the urea of SDZ HDL 376 (data not shown). Although not readily apparent from the chromatograms, mass spectral evidence was obtained for the formation of peak d in monkey and human liver microsomes. Peak e, a product generated in monkey and human microsomal incubations, was proposed to be the ring-hydroxylated metabolite of SDZ HDL 376. This was consistent with the product ions observed at m/z 200 and 158, from a precursor ion at m/z 273. Metabolite f was minor in all four species and proposed to be the urea of SDZ HDL 376 with an isobutylene side chain. The urea of SDZ HDL 376 was identified as peak g with m/z of 239. Retention times and MS/MS spectra of peak g were identical to a synthetic standard. Peak h (MH+, m/z 225), a nonpolar metabolite, was proposed as an amidine. Identification of SDZ HDL 376-GSH Adduct. Addition of GSH to [14C]SDZ HDL 376 microsomal incubations led to a large early eluting metabolite peak and a corresponding decrease in S-oxide and urea formation (Figure 2). This polar metabolite peak was formed in greater quantity in rat microsomal incubations compared to human microsomes. Microsomal incubations with nonlabeled SDZ HDL 376 and [35S]GSH formed a product with the same HPLC retention time (Figure 3). Both the 14 C- and 35S-labeled metabolite peaks were individually subjected to LC/MS and LC/MS/MS analysis. A protonated molecule (MH+) at m/z 530 was observed. Upon collision-induced dissociation of m/z 530, major product ions at m/z 257 and 145 were generated, which could be assigned as the result of bond breakage of the glutathione-thioether bond followed by cleavage of glutamic acid, respectively (Figure 4). The synthetically prepared SDZ HDL 376-GSH adduct had identical mass, MS/MS spectra, and HPLC retention time to those obtained from microsomal incubations (Figures 3, 4). Structural identification was further confirmed by NMR analysis. To identify the position of the thiocarbamide double bond in the adduct, double-tube NMR experiments were performed. Assignment of the thiocarbamide-NH was based upon protio/deuterio isotopic chemical shift differences observed in the carbon spectrum (Figure 5). The incorporation of deuterium within the molecule via exchangeable protons induces an upfield shift for the neighboring carbon resonance. Doubling of the carbon resonance was observed for the aromatic ring system indicating that the thiocarbamide -NH group was adjacent to the aromatic ring. Doubling of the isobutyl carbon resonance was not observed which further supported the position of the thiourea proton.

Metabolism of SDZ HDL 376


Figure 2. HPLC 14C radiochromatograms obtained from (A) rat and (B) human (HL-22) hepatic microsomal incubations in either the presence or absence of GSH. Microsomal incubations were performed with 0.1 mM [14C]SDZ HDL 376 and compared to reactions supplemented with 500 µM GSH. An asterisk represents the [14C]SDZ HDL 376-GSH adduct which coeluted with synthetic standard.

Figure 3. 35S radiochromatograms from (A) rat, (B) monkey, (C) human HL-15, and (D) human HL-22 microsomal incubations with SDZ HDL 376 and [35S]GSH. Microsomes (2 nmol/mL) were incubated for 30 min with 0.1 mM SDZ HDL 376 and 500 µM GSH which contained 6.8 µM [35S]GSH. The retention time for the 35S adduct was the same as that observed in Figure 2.

SDZ HDL 376-GSH Adduct Formation. The [35S]SDZ HDL 376-GSH adduct formation was used to compare the potential of SDZ HDL 376 electrophile generation across species (Figure 2). Under the conditions of the assay, rat microsomes formed a greater quantity of adduct than the primates. When the rate of adduct formation was compared across species, rat microsomes formed the adduct (3.8 nmol min-1 mg-1) at a 4-8-fold faster rate compared to monkey (0.97 nmol min-1 mg-1) or HL-22 human (0.4 nmol min-1 mg-1) microsomes. Assays were performed to determine if other structurally unrelated thiocarbamides form similar GSH adducts.

None of the five thiocarbamides tested (phenylthiourea, 6-propyl-2-thiouracil, 2-mercaptobenzimidazole, methimazole, 2-mercaptoquinazoline) formed detectable GSH adducts under the conditions described for SDZ HDL 376. In addition, reactions were also performed at pH 8.5 which favors FMO (3, 5); however, no GSH adducts were observed with the other thiocarbamides. The enzyme system responsible for the electrophile that ultimately results in the GSH adduct was assessed. Assays were performed in rat liver microsomes to determine if FMO or P450 catalyzed SDZ HDL 376 to intermediates capable of forming GSH adducts. Heating microsomes for 1 h at 37 °C in the absence of NADPH


Stevens et al.

Figure 5. 13C NMR spectra from double-tube probe experiment to determine the position of the SDZ HDL 376-GSH -NH bond. Deuterium caused an upfield resonance shift relative to the proton spectra of carbons one to two bonds away from a deuterium in the 13C NMR spectra. This results in doubling of carbon resonance, such as the aromatic carbons (insert), while unaffected carbon resonances remain singlets as shown for the isobutyl side chain carbons.

Figure 4. Collision-induced LC/MS/MS analysis of SDZ HDL 376-GSH adduct obtained from (A) microsomal incubations, (B) rat liver slices, and (C) synthetic standard. The adduct had a corresponding m/z at 530 and major product ions at m/z 145 and 257.

has been shown to decrease FMO activity by 80-90% without affecting P450 activity (3, 5). Heat inactivation of FMO had little effect, inhibiting only 10% of the SDZ HDL 376-GSH adduct. Inhibiting P450 with the addition of the suicide substrate ABT (16, 17), resulted in a 95% decrease in GSH adduct formation. Reactions performed to optimize FMO activity (pH 8.5) in the presence of ABT also produced an inhibition of GSH adduct formation. SDZ HDL 376 Protein Binding. On the basis of the evidence that thiocarbamides undergo protein binding following NADPH-dependent bioactivation (2, 8), assays were initiated to determine if SDZ HDL 376 was capable of binding microsomal proteins. Incubation of rat and human liver microsomes with 100 µM [14C]SDZ HDL 376 resulted in covalent binding of radioactivity to microsomal protein (Figure 6). Several proteins were labeled with [14C]SDZ HDL 376 from rat and human microsomal incubations. Proteins at 340, 54, and 37 kDa were the major adducts from rat microsomal incubations. The binding was dependent on NADPH and was inhibited by the addition of 100 or 500 µM GSH. Quantitative differences in the binding between rat and human were not determined since the amount of microsomal protein (nucleophile) was about 2-fold higher in the HL-22 reactions. Metabolism of SDZ HDL 376 in Rat and Human Liver Slices. Precision cut liver slices have been used

successfully in identifying the metabolism of many pharmaceutical agents (18-21). In efforts to predict the formation of SDZ HDL 376-GSH adducts in vivo, rat and human liver slices were exposed to 500 µM [14C]SDZ HDL 376 for 24 h. Metabolism of SDZ HDL 376 was more extensive in rat liver slices compared to human (Figure 7). Incubating the medium with β-glucuronidase resulted in the disappearance of two peaks from rat slice medium (peaks 1 and 2) and one peak from human (peak 1). The peaks that disappeared are thought to be glucuronide conjugates. Cleavage of the glucuronide from the peak 1 metabolite resulted in a polar species that coeluted with the ring-hydroxylated metabolite (metabolite e). In addition, incubating the medium with sulfatase resulted in the loss of peak 4 from rat liver slice medium. Of interest was the presence of a possible GSH adduct (peak 3). The peak coeluted with the synthetic SDZ HDL 376-GSH adduct (data not shown) and gave the same MS/MS spectrum. The adduct was not detected in human liver slice incubations.

Discussion The metabolism of thiocarbamides is complex, but it is believed that sulfur oxidation is an obligatory step in the bioactivation of thiocarbamides resulting in protein binding, enzyme inactivation, and hepatotoxicity (1-6). Metabolism of SDZ HDL 376 was investigated in a variety of species in vitro in efforts to assess S-oxidation and the formation of electrophilic intermediates. SDZ HDL 376 was extensively metabolized in rat, dog, and monkey and moderately metabolized in human liver microsomes. Carbon (aromatic and aliphatic) hydroxylation was the major pathway for SDZ HDL 376 metabolism in human and monkey microsomes, whereas sulfur


Figure 6. Binding of [14C]SDZ HDL 376 to (A) rat and (B) human HL-22 microsomal protein. The amount of radioactivity bound was determined by SDS-PAGE of microsomal protein following a 30 min reaction. Addition of 1.0 mM NADPH to the reaction resulted in binding of SDZ HDL 376 to microsomal protein. Binding was inhibited by the addition of 100 and 500 µM GSH. Data represent the analysis of a single gel.

oxidation was the major pathway in rat microsomal reactions. Similar to other thiocarbamides, SDZ HDL 376 is metabolized in vitro to reactive intermediates capable of irreversibly binding microsomal protein. The microsomal metabolism and protein binding of SDZ HDL 376 required NADPH. Addition of GSH to the microsomal reactions decreased protein binding and resulted in a novel SDZ HDL 376-GSH adduct (Scheme 2). The adduct does not form in microsomal incubations containing GSH and the urea of SDZ HDL 376 (data not shown). Reactions performed with [35S]GSH demonstrate that the sulfur present in the SDZ HDL 376-GSH adduct is derived from GSH and not the parent thiourea. The molecular weight of the adduct, labeling by 35S, and requirement of NADPH for adduct formation are consistent with GSH displacing an oxidized form of the thiourea sulfur. The SDZ HDL 376-GSH adduct linkage consists of a thioether bond instead of a disulfide. To our knowledge this is the first report of a thioether GSH adduct formed by microsomal oxidation of thiocarbamides. Similar adducts have been prepared by reacting propyluracil-2sulfonate with N-acetylcysteine (22), but such an adduct was not shown to form in an enzymatic system. Furthermore, amine nucleophiles were shown to react with sulfinic acids by displacement of the sulfinic acid group (23, 30). Such an adduct was shown to form in rats following exposure to thioacetamide (24). These data suggest protein amino and thiol substituents may be targets for electrophilic S-oxides resulting in a thioether


Figure 7. Metabolism of [14C]SDZ HDL 376 in (A) rat and (B) human liver slices. Slices were exposed to 100 µM drug for 24 h. Metabolites 1 and 2 are proposed glucuronide conjugates. This was accomplished by incubating the culture medium with glucuronidase which resulted in the loss of these peaks. Metabolite 3 coeluted with synthetic SDZ HDL 376-GSH. A possible sulfate conjugate (4) was observed in rat liver slice incubations. The peak was absent when sulfatase treatment of the medium was performed.

Scheme 2. Proposed Formation of the GSH Adduct from SDZ HDL 376a

a Mechanism proposed analogous to that of Maryanoff et al. (30) for the formation of guanidines from substituted thioureas.

protein adduct. The protein binding and GSH adduct formation suggest SDZ HDL 376 forms a diffusible intermediate capable of reacting with protein thiols. The NADPHdependent oxidation of a number of thiocarbamides, including ETU and methimazole, results in a loss of P450 activity which can be prevented by the coincubation with GSH or dithiothreitol (9, 25). Metabolism of ETU in the presence of GSH results in the formation of GSSG, which was suggested to occur via an initial GSH-ETU disulfide adduct. The thiosteroid, spironolactone, was also shown to form a GSH disulfide adduct in microsomal incuba-


tions (26). Such an adduct was not identified for SDZ HDL 376. Preliminary data suggest the formation of an electrophilic species resulting in the SDZ HDL 376-GSH adduct is catalyzed by P450. Since many thiocarbamides are catalyzed by both P450 and FMO systems (1-6), assays performed with purified FMO would provide evidence for this enzyme’s involvement in SDZ HDL 376 metabolism. Although the participation of FMO can not be excluded from the bioactiviation of SDZ HDL 376, the results differ with respect to other thiocarbamides. Inactivation of P450 by ABT was shown either to have no effect or to enhance thiocarbamide protein binding (8, 9). ETU metabolism to reactive intermediates capable of covalently binding protein is catalyzed by FMO, whereas other thiocarbamides such as methimazole are catalyzed by FMO and P450 (8, 9, 25, 27). The spironolactone-GSH disulfide adduct is catalyzed primarily by FMO (26). In addition, heat inactivation of FMO attenuates spironolactone-GSH adduct formation (25), decreases ETU protein binding, and preserves P450 enzyme activity (8). In the present study, heat inactivation of FMO had little effect on SDZ HDL 376-GSH adduct formation. The inhibition of P450, which in many instances enhances FMO activity, decreased SDZ HDL 376-GSH adduct formation. The P450 suicide inhibitor ABT decreased GSH adduct production by over 95% relative to controls. Inhibition of SDZ HDL 376-GSH formation by ABT is not due to ABT interaction with GSH since control reactions with [35S]GSH did not form a GSH adduct (data not shown). In addition, ABT does not conjugate with, or alter, GSH levels following in vivo exposure to male rats (28, 29). Thus, preliminary information suggests the involvement of P450 in the bioactivation of SDZ HDL 376. Efforts were made to determine if other structurally diverse thiocarbamides were capable of forming a similar GSH adduct. Many reports demonstrate that the addition of GSH to thiocarbamide microsomal reactions prevents covalent protein binding; however, in many instances GSH adducts were not identified (9, 25). Microsomal reactions with phenylthiourea, 6-propyl-2thiouracil, 2-mercaptobenzimidazole, methimazole, and 2-mercaptoquinazoline failed to form a similar GSH adduct. Therefore, the GSH adduct with SDZ HDL 376 appears to be unique for this particular thiocarbamide. The molecular basis for this selectivity is unknown, although stability and therefore reactivity of sulfenic and sulfinic acids are influenced by the R group(s) attached to the thiourea nitrogens (15, 30). The metabolism of SDZ HDL 376 in liver slices results in the formation of the GSH adduct in rat but not human liver slices. In addition to GSH adduct formation, SDZ HDL 376 is metabolized to glucuronide and sulfate conjugates. Metabolism of SDZ HDL 376 in the human liver slices was low and limited to the identification of a glucuronide conjugate. The low turnover of SDZ HDL 376 in human liver slices prevented the detection of the SDZ HDL 376-GSH adduct and complicates the prediction of GSH adduct formation following in vivo exposure. However, the in vitro studies suggest ring hydroxylation followed by glucuronidation as a major route of metabolism in humans. In summary, the in vitro metabolism of SDZ HDL 376 in the presence of glutathione differs significantly with respect to other thiocarbamides. SDZ HDL 376 oxidation results in protein binding, and the binding is diminished

Stevens et al.

by the addition of GSH. Incubations containing GSH result in a novel SDZ HDL 376-GSH adduct. Other structurally diverse thiocarbamides which form S-oxides and bind microsomal protein failed to form similar adducts. Formation of SDZ HDL 376-GSH was catalyzed by P450. These data suggest SDZ HDL 376 may form reactive S-oxides and react with cellular protein sulfhydryls; however, toxicity induced by reactive Soxides may be dependent on GSH status. Furthermore, the high level of S-oxide formation in rat microsomal reactions and the faster rate of SDZ HDL 376-GSH adduct formation compared to the other species suggest that the rat may be more sensitive to adverse effects of SDZ HDL 376 in liver under diminished GSH levels. Studies addressing the role of S-oxidation and GSH status in the toxicity of SDZ HDL 376 are currently in progress.

References (1) Neal, R. A., and Halpert, J. (1982) Toxicology of thiono-sulfur compounds. Annu. Rev. Pharmacol. Toxicol. 22, 321-339. (2) Ziegler, D. M. (1982) Functional groups bearing sulfur. In Metabolism of Functional Groups (Jakoby, W. B., Bend, J. R., and Caldwell, J., Eds.) pp 171-184, Academic Press, New York. (3) Ziegler, D. M. (1988) Flavin-Containing monooxygenases: catalytic mechanism and substrate specificities. Drug Metab. Rev. 19, 1-32. (4) Poulsen, L. L., Hyslop, R. M., and Ziegler, D. M. (1979) SOxygenation of N-substituted thioureas catalyzed by the pig liver microsomal FAD-containing monooxygenase. Arch. Biochem. Biophys. 198, 78-88. (5) Cashman, J. R., and Hanzlik, R. P. (1981) Microsomal oxidation of thiobenzamide, a photometric assay for the flavin-containing monooxygenase. Biochem. Biophys. Res. Commun. 98, 147-153. (6) Cashman, J. R. (1995) Structural and catalytic properties of the mammalian flavin-containing monooxygenase. Chem. Res. Toxicol. 8, 165-181. (7) Hunter, A. L., and Neal, R. A. (1975) Inhibition of hepatic mixedfunction oxidase activity in vitro and in vivo by various thionosulfur-containing compounds. Biochem. Pharmacol. 24, 21992205. (8) Decker, C. J., and Doerge, D. R. (1992) Covalent binding of 14Cand 35S-labeled thiocarbamides in rat hepatic microsomes. Biochem. Pharmacol. 43, 881-888. (9) Decker, C. J., and Doerge, D. R. (1991) Rat hepatic microsomal metabolism of ethylenethiourea. Contributions of the flavincontaining monooxygenase and cytochrome P450 isozymes. Chem. Res. Toxicol. 4, 482-489. (10) Krieter, P. A., Ziegler, D. M., Hill, K. E., and Burk, R. F. (1984) Increased biliary GSSG efflux from rat livers perfused with thiocarbamide substrates for the flavin-containing monooxygenase. Mol. Pharmacol. 26, 122-127. (11) Omura, T., and Sato, R. (1964) The carbon monoxide-binding pigment of liver microsomes. I. Evidence for its hemoprotein nature. J. Biol. Chem. 239, 2370-2378. (12) Bradford, M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248-254. (13) Smith, P. F., Fisher, R., Shubat, P. J., Gandolfi, A. J., Krumdieck, C. L., and Brendel, K. (1987) In vitro cytotoxicity of allyl alcohol and bromobenzene in a novel organ culture system. Toxicol. Appl. Pharmacol. 87, 509-522. (14) Cashman, J. R., and Hanzlik, R. P. (1981) Microsomal oxidation of thiobenzamide. A photometric assay for the flavin-containing monooxygenase. Biochem. Biophys. Res. Commun. 98, 147-153. (15) Walter, W., and Randau, G. (1969) Thioharnstoff-S-monoxide. Leibigs Ann. Chem. 722, 52-79. (16) Ortiz de Montellano, P. R., and Mathews, J. M. (1981) Autocatalytic alkylation of the cytochrome P450 prosthetic haem group by 1-aminobenzotriazole. Biochem. J. 195, 761-764. (17) Mathews, J. M., and Bend, J. R. (1986) N-Alkylaminobenzotriazoles as isozyme-selective suicide inhibitors of rabbit pulmonary microsomal cytochrome P450. Mol. Pharmacol. 30, 25-32. (18) Barr, J., Weir, A. J., Brendel, K., and Sipes, I. G. (1991) Liver slices in dynamic organ culture. I. An alternative in vitro technique for the study of rat hepatic drug metabolism. Xenobiotica 21, 331-339. (19) Barr, J., Weir, A. J., Brendel, K., and Sipes, I. G. (1991) Liver slices in dynamic organ culture. II. An in vitro cellular technique


(20)

(21) (22) (23) (24) (25)

for the study of integrated drug metabolism using human tissue. Xenobiotica 21, 331-339. Vickers, A. E. M., Fischer, V., Connors, S., Fisher, R. L., Baldeck, J. P., Maurer, G., and Brendel, K. (1992) Cyclosporin A metabolism in human liver, kidney, and intestine slices: comparison to rat and dog slices and human cell lines. Drug Metab. Dispos. 20, 802-809. Vickers, A. E. M. (1994) Use of human organ slices to evaluate the biotransformation and drug-induced side-effects of pharmaceuticals. Cell Biol. Toxicol. 10, 407-414. Waldhauser, L., and Uetrecht, J. (1991) Oxidation of propylthiouracil to reactive metabolites by activated neutrophils: Implication for agranulocytosis. Drug Metab. Dispos. 19, 354-359. Sterling, C. J. M. (1971) The sulfinic acids and their derivatives. Int. J. Sulfur Chem. 64, 277-320. Dyroff, M. C., and Neal, R. A. (1981) Identification of the major protein adduct formed in rat liver after thioacetamide administration. Cancer Res. 41, 3430-3435. Kedderis, G. L., and Rickert, D. E. (1985) Loss of rat liver microsomal P-450 during methimazole metabolism; role of flavincontaining monooxygenase. Drug Metab. Dispos. 13, 58-61.

Chem. Res. Toxicol., Vol. 10, No. 7, 1997 741 (26) Decker, C. J., Cashman, J. R., Sugiyama, K., Maltby, D., and Correia, M. A. (1991) Formation of glutathionyl-spironolactone disulfide by rat liver cytochromes P450 or hog liver flavincontaining monooxygenases: a functional probe of two-electron oxidation of the thiosteroid. Chem. Res. Toxicol. 4, 669-677. (27) Lee, P. W., and Neal, R. A. (1978) Metabolism of methimazole by rat liver cytochrome P450-containing monooxygenases. Drug Metab. Dispos. 6, 591-600. (28) Mugford, C. A., Mortillo, M., Mico, B. A., and Tarlogg, J. B. (1992) 1-Aminobenzotriazole-induced destruction of hepatic and renal cytochromes P450 in male Sprague-Dawley rats. Fund. Appl. Toxicol. 19, 43-49. (29) Town, C., Henderson, L., Chang, D., Mortillo, M., and Garland, W. (1993) Distribution of 1-aminobenzotriazole in male rats after administration of an oral dose. Xenobiotica 23, 383-390. (30) Maryanoff, C. A., Stanzione, R. C., and Plampin, J. N. (1986) Reactions of oxidized thioureas with amine nucleophiles. Phosphorus Sulfur 27, 221-232.

TX9700230

In Vitro Metabolism of N - American Chemical Society

Recommend Documents