Evidence for the Involvement of N-Methylthiourea, a Ring Cleavage

H. Eng , R. Sharma , A. Wolford , L. Di , R. B. Ruggeri , L. Buckbinder , E. L. Conn , D. K. Dalvie , A. S. Kalgutkar ... Shivakumar Chitturi , Geoffr...
0 downloads 0 Views 127KB Size
Download PDF
170

Chem. Res. Toxicol. 2000, 13, 170-176

Evidence for the Involvement of N-Methylthiourea, a Ring Cleavage Metabolite, in the Hepatotoxicity of Methimazole in Glutathione-Depleted Mice: Structure-Toxicity and Metabolic Studies Tamio Mizutani,* Kaoru Yoshida, Mihoko Murakami, Mutsuko Shirai, and Sadahiro Kawazoe Department of Food Sciences and Nutritional Health, Kyoto Prefectural University, Shimogamo, Kyoto 606-8522, Japan Received August 31, 1999

In mice depleted of GSH by treatment with buthionine sulfoximine (BSO), methimazole (2mercapto-1-methylimidazole, MMI) causes liver injury characterized by centrilobular necrosis of hepatocytes and an increase in serum alanine transaminase (SALT) activity. MMI requires metabolic activation by both P450 monooxygenase and flavin-containing monooxygenase (FMO) before it produces the hepatotoxicity. MMI and its analogues were examined for the ability to increase SALT activity in GSH-depleted mice. Saturation of the C-4,5 double bond in MMI resulted in a complete loss of hepatotoxicity. Similarly, ring fusion of a benzene nucleus to the C-4,5 double bond, forming 2-mercapto-1-methylbenzimidazole, abolished the toxic potency. As for MMI, 2-mercapto-1,4,5-trimethylimidazole, and 2-mercapto-1-methyl-4,5-di-n-propylimidazole, the toxic potency decreased with the increasing bulk of the 4- and 5-alkyl substituents. Furthermore, methylation of the thiol group of MMI totally reduced its toxicity. These structural requirements and the known toxicity of thiono-sulfur compounds led us to the hypothesis that MMI would undergo epoxidation of the C-4,5 double bond by P450 enzymes and, after being hydrolyzed, the resulting epoxide would be then decomposed to form N-methylthiourea, a proximate toxicant. Before N-methylthiourea would produce toxicity, it would be further biotransformed to its S-oxidized metabolites mainly by FMO. Evidence for this hypothesis was provided by the facts that N-methylthiourea and glyoxal as the accompanying fragment were identified as urinary metabolites in mice treated with MMI and that N-methylthiourea caused a marked increase in SALT activity when administered to mice in combination with BSO.

Introduction 1

Methimazole (2-mercapto-1-methylimidazole, MMI ) is widely used clinically in the treatment of hyperthyroidism. The adverse effects caused by this drug include dermal and gastrointestinal disorders, loss of taste, lupoid-like syndrome, and bone marrow depression, with agranulocytosis as the most serious complication (1-3). Hepatic injury is another reaction that occurs during MMI therapy, although its frequency is considered low (1, 4). In most cases of MMI-induced liver injury, cholestatic jaundice, without evidence of hepatic necrosis on liver biopsy, has been reported (5). However, Floreani et al. (6) have observed a case in which anicteric hepatocellular damage developed and suggested that the liver damage induced by MMI therapy may be more frequent than so far documented on the basis of jaundice cases. Our previous study (7) has shown that MMI causes severe liver injury in mice depleted of GSH by pretreatment with buthionine sulfoximine (BSO), an inhibitor of * To whom correspondence should be addressed: Department of Food Sciences and Nutritional Health, Kyoto Prefectural University, Shimogamo, Kyoto 606-8522, Japan. Telephone: +81-75-703-5408. Fax: +81-75-721-8023. E-mail: [email protected]. 1 Abbreviations: BSO, buthionine sulfoximine; FMO, flavin-containing monooxygenase; MMI, methimazole; SALT, serum alanine transaminase.

GSH synthesis. The injury is characterized by centrilobular necrosis of hepatocytes and an increase in serum alanine transaminase (SALT) activity. Moreover, the previous results have suggested that MMI is activated by reactions mediated by both P450 monooxygenase and flavin-containing monooxygenase (FMO) to reactive metabolite(s) and that the inadequate rates of detoxification of the metabolite(s) are responsible for the hepatotoxicity in GSH-depleted mice. A variety of drugs, such as acetaminophen (8), adrenergic agonists (9), adriamycin (10), and morphine (11), and conditions, such as dietary protein deficiency (12) and exercise (13), depress hepatic GSH levels in animals. In particular, hyperthyroidism itself also significantly reduces hepatic GSH levels as shown in the hyperthyroid rat model (14). Although these data cannot necessarily be extrapolated to humans, interactions between MMI metabolite(s) and cell constituents following hepatic GSH depletion could also be a mechanism for the liver damage observed in humans during MMI therapy. We designed this study to elucidate the structural requirements of MMI analogues for the hepatotoxicity in GSH-depleted mice, to suggest the chemical identity of the postulated reactive metabolite, and then to provide positive support for the suggestion produced by metabolic studies.

10.1021/tx990155o CCC: $19.00 © 2000 American Chemical Society Published on Web 03/02/2000

N-Methylthiourea: A Toxic Species of Methimazole Chart 1. Structures of Compounds Examined for Hepatotoxicity

Experimental Procedures Chemicals. The compounds examined for hepatotoxicity are shown in Chart 1. All the compounds were chromatographically pure. MMI (1a), thiourea (5a), N-methylthiourea (5b), and N-ethylthiourea (5c) were purchased from Aldrich Chemical Co. (Milwaukee, WI). 1-Methyl-2-methylthioimidazole (2) (15), 2-mercapto-1-methylbenzimidazole (4) (16), and BSO (17) were synthesized according to published procedures. Synthesis of 2-Mercapto-1,4,5-trimethylimidazole (1b). A mixture of acetoin (3.52 g, 40 mmol), N-methylthiourea (5b) (2.89 g, 32 mmol), and EtOH (3 mL) was heated to 150 °C in a sealed glass tube for 9 h, then allowed to cool, and diluted with ether (20 mL). A white solid separated, which was washed with ether and recrystallized from EtOH to give 1.4 g (31%) of 1b: mp 222-224 °C; 1H NMR (60 MHz, CDCl3) δ 2.06 (s, 3H, 4-CH3 or 5-CH3), 2.08 (s, 3H, 5-CH3 or 4-CH3), 3.50 (s, 3H, NCH3). Anal. Calcd for C6H10N2S: C, 50.67; H, 7.09; N, 19.70. Found: C, 50.70; H, 7.07; N, 19.87. Synthesis of 2-Mercapto-1-methyl-4,5-di-n-propylimidazole (1c). This compound was prepared according to the procedure (18) described previously for 2-mercapto-4,5-di-npropylimidazole. A mixture of butyroin (3.61 g, 25 mmol), N-methylthiourea (5b) (1.80 g, 20 mmol), and EtOH (3 mL) was heated to 150 °C in a sealed glass tube for 9 h, then allowed to cool, and diluted with ether (20 mL), causing a white solid to separate. The solid was collected by filtration, washed with ether, and recrystallized from EtOH to give 1.54 g (39%) of 1c: mp 140-142 °C; 1H NMR (60 MHz, CDCl3) δ 0.77-1.10 (m, 6H, two CH3CH2CH2), 1.13-1.83 (m, 4H, two CH3CH2CH2), 2.272.53 (m, 4H, two CH3CH2CH2), 3.53 (s, 3H, NCH3). Anal. Calcd for C10H18N2S: C, 60.56; H, 9.15; N, 14.13. Found: C, 60.29; H, 9.07; N, 14.29. Synthesis of 2-Mercapto-1-methyl-4,5-dihydroimidazole (3). This compound was prepared according to the procedure (19) described earlier for 2-mercapto-4,5-dihydroimidazole. To a solution of N-methylethylenediamine (8.9 g, 120 mmol) in EtOH/water (1:1) (40 mL) was added carbon disulfide (80 mL) by portions over a period of 1 h. The mixture was heated to reflux for 1 h and then treated with HCl (1.0 mL). The mixture was heated under reflux for an additional 8 h and allowed to cool in a refrigerator. The white solid which separated was collected by filtration, washed with acetone, and recrystallized from EtOH to give 5.9 g (42%) of 3: mp 128-129 °C [lit. mp (20) 128 °C]. Synthesis of 2-(Methylamino)-4-(p-nitrophenyl)thiazole (11). A solution of N-methylthiourea (5b) (1.26 g, 14 mmol) in EtOH (40 mL) was added to a solution of p-nitrophenacyl bromide (2.44 g, 10 mmol) in EtOH (40 mL), and the mixture was heated to reflux for 2 h. EtOH was removed under reduced pressure, and the residue was partitioned between CHCl3 (100 mL) and H2O (100 mL). The CHCl3 extract was dried over

Chem. Res. Toxicol., Vol. 13, No. 3, 2000 171 Na2SO4 and concentrated. The residue was recrystallized from EtOH to give 1.69 g (72%) of 11 as a yellow solid: mp 187-189 °C; 1H NMR (60 MHz, Me2SO-d6) δ 2.40 (br s, 1H, NH), 2.85 (d, J ) 4.3 Hz, 3H, NHCH3), 7.30 (s, 1H, thiazolyl H-5), 7.90 (d, J ) 8.8 Hz, 2H, phenyl H), 8.11 (d, J ) 8.8 Hz, 2H, phenyl H). Anal. Calcd for C10H9O2N3S: C, 51.05; H, 3.86; N, 17.86. Found: C, 51.09; H, 3.80; N, 17.96. Animals. Six-week-old male ICR mice were obtained from Charles River Japan, Inc. (Kanagawa, Japan) and acclimatized to our laboratory conditions for 1 week before being used. Mice were housed in aluminum boxes on a wood-chip bedding (White Flake, Charles River Japan, Inc.) at a constant temperature (23 ( 2 °C) and under a 12 h light/dark cycle. Except when stated otherwise, mice received food (Funabashi F-2, Funabashi Farms, Chiba, Japan) and water ad libitum. Assessment of Hepatotoxicity. Mice were treated ip with BSO (3 mmol/kg of body weight) in water (20 mL/kg of body weight). One hour later, the treated animals received po doses of MMI analogues (1a-c and 2-4) or thiourea analogues (5a-c). MMI, 3, and thioureas 5a-c were dissolved in water (5 mL/kg of body weight) and 1b, 1c, 2, and 4 in olive oil (5 mL/kg of body weight). The animals were fasted for 16 h before dosing with test compounds and then for 2 h after administration. Blood was collected by cardiac puncture under pentobarbital anesthesia. The blood was allowed to clot at 37 °C for 1 h, and serum was prepared by centrifugation (900g). SALT activity was measured with a commercial kit, GPT-UV test Wako (Wako Pure Chemical Industries, Osaka, Japan). Study on the Urinary Metabolites of MMI. Four mice received MMI (0.2 mmol/kg of body weight) in water (5 mL/kg of body weight) by po intubation and were placed collectively in a metabolic cage. The urine was collected for 32 h. The cage was rinsed with a small volume of water; the wash and collected urine were combined, and the total volume was adjusted to 30 mL with water. After removal of particulate materials by centrifugation, the urine sample was passed through an octadecylsilica cartridge (Bond Elute LRC C18, Varian) to remove hydrophobic materials and then subjected to analysis for the presence of N-methylthiourea (5b) and glyoxal (8). (A) Identification of N-Methylthiourea (5b). A portion (13 mL) of the urine sample was added to a solution of p-nitrophenacyl bromide (10 mg) in EtOH (15 mL), and the mixture was heated for 2 h under reflux. The reaction mixture was diluted with water (20 mL), made alkaline with aqueous ammonia, and extracted with ether (2 × 50 mL). The extract was dried over Na2SO4 and evaporated to dryness by using a Kuderna-Danish concentrator and finally with a nitrogen stream. The residue was dissolved in benzene (1 mL) and applied to a silica gel dry column [Silica Gel 60, 70-230 mesh, Merck, deactivated with 5% (w/w) water; 3 g, 7.5 mm i.d.]. The column was washed with 10 mL of hexane, followed by 14 mL of hexane/ether (6:4), and the derivatized metabolite was eluted with an additional 10 mL of hexane/ether (6:4). The eluate was analyzed by GC/MS for the presence of 2-(methylamino)-4-(pnitrophenyl)thiazole (11) derivatized from N-methylthiourea (5b). (B) Identification of Glyoxal (8). To a portion (13 mL) of the urine sample was added o-phenylenediamine (50 mg) in 1 mL of EtOH, and the mixture was heated at 80 °C for 30 min. The mixture was extracted with ether (2 × 40 mL). The organic phase was washed with 0.08 N HCl (2 × 10 mL), dried over Na2SO4, and evaporated to a volume of 0.5 mL. The residue was diluted with hexane (2 mL) and applied to a silica gel dry column [Silica Gel 60, 70-230 mesh, Merck, deactivated with 5% (w/w) water; 3 g, 7.5 mm i.d.]. The column was washed with 10 mL of hexane, followed by 6 mL of hexane/ether (8:2), and the derivatized metabolite was eluted with an additional 2 mL of hexane/ether (8:2). The eluate was analyzed by GC/MS for the presence of quinoxaline (12) derivatized from glyoxal (8). Quantitation of N-Methylthiourea (5b), a Urinary Metabolite of MMI and 1c. MMI (dissolved in water, 5 mL/ kg of body weight) and 1c (in olive oil, 5 mL/kg of body weight)

172

Chem. Res. Toxicol., Vol. 13, No. 3, 2000

were each administered by po intubation to four groups of four mice at a dose of 0.2 mmol/kg of body weight. Mice were placed in groups of four in metabolic cages, and the urine was collected for 32 h. For metabolite quantitation, pooled urine samples from four sets of four mice were analyzed individually. Each urine sample was derivatized with p-nitrophenacyl bromide according to essentially the same procedure as that used for the identification of the MMI metabolite. HPLC analysis of the derivatized metabolite was performed on a Shim-pack CLC-ODS column (6 mm × 150 mm) (Shimadzu, Kyoto, Japan) on a Shimadzu LC-6A solvent delivery system. The column was eluted with CH3CN/water [7:3 (v/v)] at a flow rate of 1 mL/min. Under these conditions, an authentic sample of 11 was eluted with a retention time of 5.7 min. The samples were monitored by UV absorbance at 275 nm using a Shimadzu SPD-6A detector. The peak area responses were linearly related to the amounts of the reference compound in the range 0.02-0.1 µg. GC/MS. Analyses of the derivatized metabolites were carried out by GC/MS using a model 5890 series II gas chromatograph (Hewlett-Packard, Palo Alto, CA) coupled to a model JMSSX102A mass spectrometer (JEOL Ltd., Tokyo, Japan). A 0.2 mm i.d. × 25 m fused silica capillary column (HP-5, HewlettPackard) was used for GC separations. The injector and the GC/ MS interface temperatures were set at 210 and 230 °C, respectively. The samples were injected in the splitless mode with an initial column temperature of 80 °C for 1 min. The column temperature was programmed at a rate of 15 °C/min to 210 °C and then kept at 210 °C for 18 min. The carrier gas was helium with a flow rate of 0.7 mL/min. The mass spectrometer was operated in the electron impact ionization mode and scanned from 50 to 500 amu every 0.5 s with an ionization energy of 70 eV. Statistical Analysis. Data from the SALT activity assay were analyzed by Kruskal-Wallis nonparametric analysis of variance followed by a Mann-Whitney’s U test. For data of urinary N-methylthiourea, comparison was made with a Student’s t test. Differences were considered significant if p < 0.05.

Results Hepatotoxicity of MMI and Structurally Related Compounds in Mice Depleted of GSH by BSO. MMI and several structurally related compounds were evaluated for their hepatotoxicity in mice depleted of GSH by pretreatment with BSO. Our previous study has shown that MMI in combination with BSO produced injury in the liver, as evidenced by centrilobular necrosis of hepatocytes and an increase in SALT activity (7). Therefore, changes in SALT activity at 6 h after the administration of test compounds (0.2 and/or 0.6 mmol/kg of body weight) were used as an indication of hepatotoxicity. Mice were pretreated with BSO according to the same regimen that was used in the previous study (7). We have reported in detail on the tissue GSH depletion produced by the BSO pretreatment regimen (7). Briefly, the pretreatment with BSO reduces hepatic GSH concentrations to 33 and 37% of the control after 2 and 4 h, respectively. As shown in Table 1, MMI (1a) caused a severe increase (300-fold) in SALT activity at a dose of 0.2 mmol/kg of body weight. At the lower dose (0.2 mmol/kg of body weight), compound 1b resulted in a minor, but statistically significant, increase (5-fold) in SALT activity, whereas it produced a marked increase (100-fold) in SALT activity at the higher dose (0.6 mmol/kg of body weight). Compound 1c produced only a marginal increase (6-fold) in SALT activity even at the higher dose. In contrast with compounds 1a-c, compounds 2-4 (up to 0.6 mmol/kg of body weight) were totally ineffective in inducing hepatotoxicity

Mizutani et al. Table 1. SALT Activity 6 h after Administration of MMI Analogues in Mice Depleted of GSH by Treatment with BSOa SALT (Karmen units/mL) compound

0.2 mmol/kg

0.6 mmol/kg

control 20.8 ( 2.7 MMI (1a) 5980 ( 3120b 2-mercapto-1,4,5-trimethyl103 ( 19.1b 2050 ( 266b imidazole (1b) 2-mercapto-1-methyl-4,5-di-n-propyl58.9 ( 8.6b 132 ( 22.6b imidazole (1c) 1-methyl-2-methylthioimidazole (2) 29.9 ( 4.6 2-mercapto-1-methyl-4,5-dihydro30.5 ( 4.8 imidazole (3) 2-mercapto-1-methylbenzimidazole (4) 20.0 ( 7.7 a Mice were treated with BSO (3 mmol/kg of body weight, ip) and received MMI analogues (po) 1 h later. Values are means ( SE from 5-12 mice. b Significantly different from control group (p < 0.05).

Scheme 1. Derivatization of N-Methylthiourea (5b) with p-Nitrophenacyl Bromide

as judged by SALT activity. None of the compounds (1a-c and 2-4) produced a sign of hepatotoxicity in the absence of BSO pretreatment (data not shown). Identification of N-Methylthiourea (5b) as a Urinary Metabolite of MMI. The presence of N-methylthiourea (5b) in the urine of mice treated with MMI (1a) was determined by condensing 5b with p-nitrophenacyl bromide, forming 2-(methylamino)-4-(p-nitrophenyl)thiazole (11) (Scheme 1). This procedure is based on Hantzsch’s method for the synthesis of thiazoles (21) and, as previously reported (22, 23), is effective for the isolation and identification of unstable and/or nonvolatile thioamide metabolites. Panel A in Figure 1 shows the reconstructed ion chromatograms for the urinary sample derivatized with p-nitrophenacyl bromide. The selected ion chromatograms at m/z 235 and 189 equally showed a peak (M - 1) with a retention time (19.2 min) identical with that of the authentic standard, 2-(methylamino)-4(p-nitrophenyl)thiazole (11). The mass spectrum of peak M - 1 (Figure 1B) was very similar to that of the authentic sample (Figure 1C), with a molecular ion at m/z 235, an M - NO peak at m/z 205, an M - NO2 peak at m/z 189, an M - NO2 - CH3 peak at m/z 174, and an M - NO - CH3NHCN peak at m/z 149. These results thus unambiguously demonstrate the presence of Nmethylthiourea (5b) in the urine as a metabolite of MMI (1a). Identification of Glyoxal (8) as a Urinary Metabolite of MMI. The production of glyoxal (8) as a urinary metabolite of MMI (1a) was examined by GC/MS after derivatization to its quinoxaline derivative (12) (Scheme 2). Panel A in Figure 2 shows the reconstructed ion chromatograms for the urinary extract obtained after derivatization with o-phenylenediamine. The selected ion traces at m/z 130 and 103 both exhibited a peak (M - 2) with a retention time (7.6 min) identical with that of

N-Methylthiourea: A Toxic Species of Methimazole

Chem. Res. Toxicol., Vol. 13, No. 3, 2000 173

Figure 1. (A) GC/MS-reconstructed ion chromatogram of the derivatized urinary extract from mice treated with MMI (1a). The arrow indicates the tR of authentic 8. (B and C) Electron impact mass spectra of (B) the derivatized metabolite M - 1 and (C) the 4-(p-nitrophenyl)thiazole derivative (11) of authentic N-methylthiourea (5b).

Figure 2. (A) GC/MS-reconstructed ion chromatogram of the derivatized urinary extract from mice treated with MMI (1a). The arrow indicates the tR of authentic 12. (B and C) Electron impact mass spectra of (B) the derivatized metabolite M - 2 and (C) the quinoxaline derivative (12) of authentic glyoxal (8).

Scheme 2. Derivatization of Glyoxal (8) with o-Phenylenediamine

Table 2. SALT Activity 6 h after Administration of Thiourea Analogues in Mice Depleted of GSH by Treatment with BSOa compound control N-methylthiourea (5b)

authentic 12. The mass spectra of peak M - 2 (Figure 2B) and the authentic sample (Figure 2C) were in good agreement with respect to typical peaks, a molecular ion at m/z 130, an M - HCN peak at m/z 103, and an M 2HCN peak at m/z 76. Thus, the formation of glyoxal (8) as a urinary metabolite of MMI was confirmed. Urinary Excretion of N-Methylthiourea as a Metabolite of MMI and 2-Mercapto-1-methyl-4,5di-n-propylimidazole (1c). After administration of equimolar doses (0.2 mmol/kg of body weight) of MMI (1a) and 2-mercapto-1-methyl-4,5-di-n-propylimidazole (1c), N-methylthiourea (5b) excreted in the urine was derivatized to its thiazole derivative for quantitation by HPLC. The derivatized metabolite (11) was eluted with a retention time of 5.7 min under conditions described in Experimental Procedures. The levels of urinary excretion of 5b 32 h after treatment with MMI and 1c were 6.3 ( 2.1 and 1.3 ( 0.3% (mean ( SD of four experiments), respectively, of the dose; these values were

thiourea (5a) N-ethylthiourea (5c)

dose (mmol/kg of body weight)

SALT (Karmen units/mL)

0.2 0.6 1.0 1.0 1.0

28.6 ( 3.7 241 ( 11.1b 1840 ( 1570b 3870 ( 1620b 1050 ( 146b 11900 ( 6130b

a Mice were treated with BSO (3 mmol/kg of body weight, ip) and received thiourea analogues (po) 1 h later. Values are means ( SE from 5-12 mice. b Significantly different from control group (p < 0.05).

significantly different (p < 0.005). This indicates that 1c was metabolized to 5b to a smaller extent than MMI. Hepatotoxicity of Thiourea Homologues in Mice Depleted of GSH by BSO. Thiourea (5a), N-methylthiourea (5b), and N-ethylthiourea (5c) were examined for their ability to increase SALT activity in mice depleted of GSH by treatment with BSO. The results are shown in Table 2. Treatment with 5b produced dose-dependent increases in SALT activity at doses up to 1.0 mmol/kg of body weight. At a dose of 1.0 mmol/kg of body weight, 5a and 5c also produced marked increases in SALT activity; the rank order of hepatotoxicity

174

Chem. Res. Toxicol., Vol. 13, No. 3, 2000

Mizutani et al.

Scheme 3. Proposed Pathway of Metabolic Activation of MMI in Mice

was 5c > 5b > 5a as judged by SALT activity. Treatment with 5b (up to at least 1.0 mmol/kg) in the absence of BSO produced no significant changes in SALT activity (data not shown).

Discussion To understand the chemical nature of the postulated toxic metabolite of MMI, we studied structure-toxicity relations for a number of MMI derivatives in mice depleted of GSH by BSO. The results presented in Table 1 are informative for considering the structural requirements for the hepatotoxic effect of MMI in GSH-depleted mice. The saturation of the C-4,5 double bond in MMI, forming 2-mercapto-1-methyl-4,5-dihydroimidazole (3), resulted in a complete loss of the hepatotoxicity. Similarly, ring fusion of a benzene nucleus to the 4 and 5 positions of MMI to give 2-mercapto-1-methylbenzimidazole (4) abolished the toxic potency. As for MMI analogues 1a-c, the hepatotoxic potency appears to decrease with the increasing bulk of the 4- and 5-alkyl substituents (1a > 1b > 1c). Collectively, these results suggest that the presence of a C-4,5 double bond is critically important for obligatory metabolic activation of MMI and that the rate of activation may be affected by the type of substitution at the 4 and 5 positions. Furthermore, the total loss of toxicity observed by methylation of the thiol group of MMI, forming 1-methyl-2methylthioimidazole (2), clearly shows that the free thiol also has a critical role in hepatotoxicity caused by MMI. These structural requirements for the hepatotoxicity can satisfactorily be explained by assuming that before the development of toxicity MMI undergoes a ring cleavage reaction involving the 4 and 5 positions to yield some thiono-sulfur-containing species including the C-2 unit of parent MMI. The well-known toxic properties of a number of thiono-sulfur compounds such as thioamides and thioureas (24) lead us to suggest this idea. A possible mechanism for the postulated ring cleavage is shown in Scheme 3. Initially, MMI (Ia) may be oxidized by P450

enzymes to form the 4,5-epoxide (6). The enzymatic or nonenzymatic hydrolysis of the epoxide (6) would produce an unstable hemiketal-like intermediate (7), which is expected to undergo spontaneous ring cleavage to form glyoxal (8) and N-methylthiourea (5b). The metabolism of thioureas is complex, but it is believed that sulfur oxidation, mediated mainly by FMO, primarily to the sulfenic acids and then possibly to the sulfinic acids is a necessary step in the bioactivation of thioureas resulting in protein binding, enzyme inactivation, and organ toxicity (24-26). Therefore, 5b, if formed, would be further biotransformed to its S-oxidized metabolites (9 and 10) before it produces hepatotoxicity. In an earlier paper, we reported the breakdown of hepatotoxic thiazoles, such as thiabendazole, in a manner similar to that proposed here for MMI which yields thioamide metabolites as proximate toxicants (22, 23). The metabolic pathways shown in Scheme 3 may suggest another possibility that MMIinduced hepatotoxicity is mediated by epoxide 6 as a reactive intermediate. However, this hypothesis seems incompatible with the result of structure-toxicity study whereby the free thiol substituent of MMI plays an important role in the development of hepatotoxicity. Our previous study has shown that P450-dependent monooxygenase inhibitors such as methoxsalen, piperonyl butoxide, and isosafrole prevent the hepatotoxicity caused by MMI in combination with BSO (7); these observations are consistent with suggesting the involvement of epoxidation by P450 in the toxification processes of MMI. In addition, the substitution effect observed in the hepatotoxic potency of 4- and 5-substituted MMI analogues can be explained on the basis of the steric effect of the substituents on the epoxidation of the C-4,5 double bond. The increasing bulk of the substituents in the 4 and 5 positions is likely to reduce the rate of epoxidation catalyzed by P450 enzymes (27). Furthermore, the previous results that competitive substrates of FMO, N,N-dimethylaniline and ethyl methyl sulfide, result in remarkable suppression of the hepatotoxicity

N-Methylthiourea: A Toxic Species of Methimazole

are consistent with the suggestion that N-methylthiourea (5b) as the proximate toxicant is further biotransformed mainly by FMO to its S-oxidized metabolites (9 and 10) before it produces hepatotoxicity. In contrast, the effectiveness of the FMO inhibitors in preventing MMIinduced toxicity cannot reasonably be understood on the basis of the alternative hypothesis that epoxide 6 itself is the toxic species of MMI. The structure-toxicity study presented here showed requirement of the free thiol group of MMI for toxicity; this is also in accord with the structural feature that is essential for the existence of S-oxidized intermediates (9 and 10). On the other hand, previous studies have shown that MMI is ultimately metabolized to N-methylimidazole and sulfite both in vivo (28) and in vitro (29). S-Oxidation products of MMI, the sulfenic (13) and sulfinic (14) acids, are believed to be reactive metabolites produced during the metabolism of MMI to N-methylimidazole (29); several events such as the loss of rat liver microsomal P450 during MMI metabolism (30) and the olfactory toxicity induced by MMI in rats (31) are attributed to these intermediates. These results may suggest the possibility that metabolic activation by direct oxidation of the thiol group of MMI is responsible for the hepatotoxicity induced by MMI in combination with BSO. Seemingly, requirement of the free thiol group for the toxicity is also likely to be consistent with this mechanism. However, the major metabolic pathways of a variety of cyclic thiocarbamides other than MMI, including 2-mercapto-4,5-dihydroimidazole (32) and 2mercaptobenzimidazole (33), compounds structurally closely related to 3 and 4, respectively, and 4 itself (33), are also known to involve oxidation at their thiol groups, giving the corresponding sulfenates. Nevertheless, 3 and 4 were totally ineffective in inducing the hepatotoxicity (Table 1); this strongly suggests that the direct Soxidation pathway is not responsible for the hepatotoxicity of MMI in GSH-depleted mice. Moreover, the inhibitory effect of substituents at the 4 and 5 positions on the toxic potency cannot reasonably be explained by the direct S-oxidation mechanism. To provide experimental evidence for the postulated metabolic pathways, we studied the metabolism of MMI. When MMI was administered to mice, a ring cleavage product, N-methylthiourea (5b), and the accompanying fragment, glyoxal (8), were identified as urinary metabolites after being trapped as 2-(methylamino)-4-(p-nitrophenyl)thiazole (11) and quinoxaline (12), respectively (Figures 1 and 2). A 4,5-disubstituted analogue of MMI, 2-mercapto-1-methyl-4,5-di-n-propylimidazole (1c), also formed a ring cleavage product 5b as a urinary metabolite; however, the extent of formation of 5b from 1c was considerably lower than that from MMI. On the other hand, the structure-toxicity study clearly showed that 1c is much less hepatotoxic than MMI (Table 1). Collectively, these findings support the idea that the hepatotoxicity caused by MMI is mediated by 5b as a proximate toxicant. Furthermore, the observation that when externally administered, 5b is hepatotoxic to mice depleted of GSH by BSO (Table 2) strengthens the hypothesis (Scheme 3) that this product is a proximate metabolite responsible for the toxicity of MMI. Unexpectedly, 5b, a putative proximate hepatotoxicant, was found to be less toxic than parent MMI (Tables 1 and 2). However, this discrepancy may be ascribed to possible differences between 5b and MMI in the extent of absorption and

Chem. Res. Toxicol., Vol. 13, No. 3, 2000 175

distribution to the target organ because 5b is much less lipophilic than MMI. This idea may be supported by the present observation that there are marked differences in the hepatotoxicity of a series of thiourea homologues (5c > 5b > 5a) (Table 2) because this result suggests that the lipophilicity of compound plays an important part in determining the toxic potency of thioureas externally administered. In the study presented here, N-methylthiourea (5b), as well as MMI, required GSH depletion to produce hepatotoxicity when administered to mice. This suggests that GSH plays a protective role toward MMI by inactivating reactive metabolite(s) generated by further metabolism of 5b; this metabolite can attack the cellular constituents which results in the hepatotoxicity in GSHdepleted mice. As mentioned above, for many thioureas the sulfenic and/or sulfinic acid metabolites have been postulated as the reactive intermediates responsible for their protein binding, enzyme inactivation, and organ toxicity (24-26). In metabolic studies with microsomes, a number of thioureas were shown to bind to microsomes and to inactivate P450 enzymes; the addition of GSH completely inhibited the enzyme inactivation with the concomitant formation of GSSG. Similarly, Krieter et al. (34) have observed an increase in the biliary efflux of GSSG upon perfusion of rat livers with thiourea, phenylthiourea, and MMI. It is reported that the sulfenic acids generated from the thioureas oxidize GSH to GSSG nonenzymatically. Thus, GSH can be involved in the detoxification of thioureas as a cellular reductant for their reactive intermediates. In agreement with such a role of GSH, GSH conjugates were not identified in many metabolic studies on thioureas, although many studies demonstrated that the addition of GSH to thiourea microsomal reactions prevents covalent protein binding. Recently, the existence of a novel GSH conjugate formed from N-(5-chloro-2-methylphenyl)-N′-(2-methylpropyl)thiourea, a thiourea developed for the treatment of atherosclerosis, has been demonstrated in rat microsomal incubations (35). In rat liver microsomes, the major metabolite of this thiourea is an S-oxide. The NADPHdependent oxidation of the thiourea resulted in covalent binding to microsomal protein; addition of GSH to the incubations decreased the extent of protein binding and resulted in the formation of a novel GSH conjugate. These observations suggest another possibility that GSH plays a role as a nucleophile in the detoxication of reactive intermediates formed from thioureas. In the case presented here, however, the exact nature of the ultimate toxic species formed via 5b (Scheme 3) and the chemical mechanism of its interaction with GSH remain to be elucidated.

Acknowledgment. We thank Rika Shimizu, Yumi Nagura, Maki Tanaka, and Kaori Tanaka for their excellent technical assistance.

References (1) Cooper, D. S. (1984) Antithyroid drugs. N. Engl. J. Med. 311, 1353-1362. (2) Wing, S., and Fantus, I. (1987) Adverse immunologic effects of antithyroid drugs. Can. Med. Assoc. J. 136, 121-127. (3) Bartalena, L., Bogazzi, F., and Martino, E. (1996) Adverse effects of thyroid hormone preparations and antithyroid drugs. Drug Saf. 15, 53-63. (4) Martinez-Lopez, J. I., Greenberg, S. E., and Kling, R. R. (1962) Drug-induced hepatic injury during methimazole therapy. Gastroenterology 43, 84-87.

176

Chem. Res. Toxicol., Vol. 13, No. 3, 2000

(5) Vitug, A. C., and Goldman, J. M. (1985) Hepatotoxicity from antithyroid drugs. Horm. Res. 21, 229-234. (6) Floreani, A., Chiaramonte, M., Scandola, C., and Naccarato, R. (1983) Methimazole-induced liver damage: Case report. Eur. Rev. Med. Pharmacol. Sci. 5, 19-20. (7) Mizutani, T., Murakami, M., Shirai, M., Tanaka, M., and Nakanishi, K. (1999) Metabolism-dependent hepatotoxicity of methimazole in mice depleted of glutathione. J. Appl. Toxicol. 19, 193-198. (8) Mitchell, J. R., Jollow, D. J., Potter, W. Z., Gillette, J. R., and Brodie, B. B. (1973) Acetaminophen-induced hepatic necrosis IV. Protective role of glutathione. J. Pharmacol. Exp. Ther. 187, 211-217. (9) James, R. C., Roberts, S. M., and Harbison, R. D. (1983) The perturbation of hepatic glutathione by R2-adrenergic agonists. Fundam. Appl. Toxicol. 3, 303-308. (10) Olson, R. D., MacDonald, J. S., vanBoxtel, C. J., Boerth, R. C., Harbison, R. D., Slonim, A. E., Freeman, R. W., and Oates, J. A. (1980) Regulatory role of glutathione and soluble sulfhydryl groups in the toxicity of adriamycin. J. Pharmacol. Exp. Ther. 215, 450-454. (11) James, R. C., Goodman, D. R., and Harbison, R. D. (1982) Hepatic glutathione and hepatotoxicity: changes induced by selected narcotics. J. Pharmacol. Exp. Ther. 221, 708-714. (12) McLean, A. E. M., and Day, P. A. (1975) The effect of diet on the toxicity of paracetamol and the safety of paracetamol-methionine mixtures. Biochem. Pharmacol. 24, 37-42. (13) Lew, H., Pyke, S., and Quintanilha, A. (1985) Changes in the glutathione status of plasma, liver, and muscle following exhaustive exercise in rats. FEBS Lett. 185, 262-266. (14) Smith, A. C., Berman, M. L., James, R. C., and Harbison, R. D. (1983) Characterization of hyperthyroidism enhancement of halothane-induced hepatotoxicity. Biochem. Pharmacol. 32, 35313539. (15) Wohl, A., and Marckwald, W. (1889) Ueber Condensationsproducte aus Amidoacetal. (II). Ber. Dtsch. Chem. Ges. 22, 13531362. (16) Doerge, D. R. (1988) Synthesis of 14C- and 35S-labeled 2mercaptobenzimidazoles. J. Labelled Compd. Radiopharm. 25, 985-990. (17) Griffith, O. W., and Meister, A. (1979) Potent and specific inhibition of glutathione synthesis by buthionine sulfoximine (S-n-butyl homocysteine sulfoximine). J Biol. Chem. 254, 75587560. (18) Basse, A., and Klinger, H. (1898) Zur Kenntniss der Butyroine und des Isovaleroins. Ber. Dtsch. Chem. Ges. 31, 1217-1225. (19) Allen, C. F. H., Edens, C. O., and VanAllan, J. (1955) Ethylene thiourea. Organic Syntheses, Collect. Vol. 3, pp 394-395, Wiley, New York. (20) Kister, J., Assef, G., Dou, H. J.-M., and Metzger, J. (1976) Hydrolysis of 1-methyl-2-methylthio-∆2-imidazoline. Tetrahedron 32, 1395-1398. (21) Vernin, G. (1979) General synthetic methods for thiazole and thiazolium salts. In The Chemistry of Heterocyclic Compounds (Weissberger, A., and Taylor, E. C., Eds.) Vol. 34, Part 1, pp 175-180, Wiley, New York.

Mizutani et al. (22) Mizutani, T., Yoshida, K., and Kawazoe, S. (1993) Possible role of thioformamide as a proximate toxicant in the nephrotoxicity of thiabendazole and related thiazoles in glutathione-depleted mice: structure-toxicity and metabolic studies. Chem. Res. Toxicol. 6, 174-179. (23) Mizutani, T., Yoshida, K., and Kawazoe, S. (1994) Formation of toxic metabolites from thiabendazole and other thiazoles in mice. Identification of thioamides as ring cleavage products. Drug Metab. Dispos. 22, 750-755. (24) Neal, R. A., and Halpert, J. (1982) Toxicology of thiono-sulfur compounds. Annu. Rev. Pharmacol. Toxicol. 22, 321-339. (25) 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. (26) Cashman, J. R. (1995) Structural and catalytic properties of the mammalian flavin-containing monooxygenase. Chem. Res. Toxicol. 8, 165-181. (27) Ortiz de Montellano, P. R. (1985) Alkenes and alkynes. In Bioactivation of Foreign Compounds (Anders, M. W., Ed.) pp 126131, Academic Press, Orlando, FL. (28) Skellern, G. G., and Steer, S. T. (1981) The metabolism of [2-14C]methimazole in the rat. Xenobiotica 11, 627-634. (29) Poulsen, L. L., Hyslop, R. M., and Ziegler, D. M. (1974) SOxidation of thioureylenes catalyzed by a microsomal flavoprotein mixed-function oxidase. Biochem. Pharmacol. 23, 3431-3440. (30) Kedderis, G. L., and Rickert, D. E. (1985) Loss of rat liver microsomal cytochrome P-450 during methimazole metabolism. Role of flavin-containing monooxygenase. Drug Metab. Dispos. 13, 58-61. (31) Genter, M. B., Deamer, N. J., Blake, B. L., Wesley, D. S., and Levi, P. E. (1995) Olfactory toxicity of methimazole: doseresponse and structure-activity studies and characterization of flavin-containing monooxygenase activity in the Long-Evans rat olfactory mucosa. Toxicol. Pathol. 23, 477-486. (32) Savolainen, K., and Pyysalo, H. (1979) Identification of the main metabolite of ethylenethiourea in mice. J. Agric. Food Chem. 27, 1177-1181. (33) Decker, C. J., Doerge, D. R., and Cashman, J. R. (1992) Metabolism of benzimidazoline-2-thiones by rat hepatic microsomes and hog liver flavin-containing monooxygenase. Chem. Res. Toxicol. 5, 726-733. (34) 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. (35) Stevens, G. J., Hitchcock, K., Wang, Y. K., Coppola, G. M., Versace, R. W., Chin, J. A., Shapiro, M., Suwanrumpha, S., and Mangold, B. L. K. (1997) In vitro metabolism of N-(5-chloro-2methylphenyl)-N′-(2-methylpropyl)thiourea: species comparison and identification of a novel thiocarbamide-glutathione adduct. Chem. Res. Toxicol. 10, 733-741.

TX990155O