Formation and reactivity of alternative quinone methides from

The phenolic antioxidant BHT' is used widely as a preservative of foods, drugs, and other consumer products and also has anticarcinogenic properties (...
0 downloads 0 Views 696KB Size
Chem. Res. Toxicol. 1990,3,65-70

65

Formation and Reactivity of Alternative Quinone Methides from Butylated Hydroxytoluene: Possible Explanation for Species-Specific Pneumotoxicity Judy

L. Bolton, Hubert Sevestre, Basil 0. Ibe, and John A. Thompson*

Molecular Toxicology and Environmental Health Sciences Program, School of Pharmacy, University of Colorado, Boulder, Colorado 80309-0297 Received October 20, 1989

Previous work has shown t h a t butylated hydroxytoluene [2,6-di-tert-butyl-4-methylphenol (BHT)] undergoes 7-oxidation in liver microsomes to form the quinone methide 2,6-di-tertbutyl-4-methylene-2,5-cyclohexadienone (QM). This electrophilic species binds covalently t o glutathione and protein thiols and is believed to initiate pulmonary toxicity in mice. I n the present investigation, we identified another quinone methide metabolite of BHT, 6 - t e r t - b ~ tyl-2-(hydroxy-tert-butyl)-4-methylene-2,5-cyclohexadienone (QM-OH), formed subsequent t o the microsomal hydroxylation of B H T a t a tert-butyl group. Mouse liver and lung microsomes generate the two quinone methides, and evidence was obtained that both metabolites also are formed in vivo. In contrast, rat microsomes produce QM almost exclusively, with only traces of QM-OH formed in liver and none in lung. Studies of the chemical reactivities of the two quinone methides with GSH demonstrated that QM-OH reacts about 6-fold faster than QM. Infrared spectra, lH NMR spectra, and electrochemical measurements all support the proposal that the enhanced electrophilicity of QM-OH is due to intramolecular hydrogen bonding of the ring oxygen with the side-chain hydroxyl. The results provide evidence, therefore, that the previous metabolic scheme for bioactivation of B H T to a pulmonary toxin should be amended t o include tert-butyl hydroxylation and subsequent 7-oxidation to the activated electrophile QM-OH. This scheme is consistent with published data concerning BHT-induced pulmonary toxicity and provides a n explanation for the species specificity of this effect.

in rat bile (11). Work with rat liver microsomes has demonstrated that the formation of QM is catalyzed by cytoThe phenolic antioxidant BHT' is used widely as a chrome P-450 and that adduct formation with GSH occurs preservative of foods, drugs, and other consumer products nonenzymatically (12). Covalent binding of [14C]BHTand also has anticarcinogenic properties (1, 2). Several derived radioactivity to liver microsomal protein has been adverse effects of BHT have been documented, however, explained by attack of the metabolite QM on thiol groups which include hemorrhagic death in rats (3), acute (reof cysteine residues (13). Following administration of versible) lung damage in mice (4, 5), and lung tumor [14C]BHT,lung protein contains the highest ratio of copromotion in mice (5, 6 ) . The mouse lung, therefore, is valently bound to free radioactivity compared to other a valuable system for investigating mechanisms associated tissues, and the amount of binding in the lung is second with BHT-induced toxicity, and the information derived only to that in the liver (on a per milligram of protein from such work could prove to be useful in designing safer basis) (14). BHT treatment decreases GSH levels in lung, antioxidant and cancer chemopreventive agents. and treatment with cysteine prevents BHT-induced pulSeveral studies indicate that the pulmonary toxicity of monary toxicity in mice; conversely, the depletion of GSH BHT is related to its oxidation by cytochrome P-450 (5, enhances toxicity (15, 16). 7). As shown in Figure 1, the compound is metabolized The results summarized above strongly implicate QM along three main pathways; benzylic hydroxylation proas the metabolite responsible for lung toxicity. This duces the 4-hydroxymethyl derivative, tert-butyl hyproposal, however, does not explain the fact that, although droxylation yields BHT-OH, and oxidation of the 7-elecboth mice and rats convert BHT to QM, only mice are tron system generates several nonaromatic products, inaffected (5). Although it is possible that mouse lung is cluding the quinone methide QM (8). The preponderance inherently more susceptible to chemical damage, an alof data supports QM formation as the initiating event in ternative explanation is that the relationship between pulmonary toxicity. QM has been identified in the livers metabolic activation of BHT and pulmonary toxicity has (9) and bile of rats treated with BHT (10). BHT-SG, not been elucidated fully. formed by nucleophilic attack of GSH on the electrophilic In detailed comparisons of BHT metabolism by liver and 4-methylene group of QM (Figure 11, has been identified lung microsomes from rats and mice, the only major difference was in production of BHT-OH; mice produced relatively large quantities, but rats formed only trace Abbreviations: butylated hydroxytoluene or BHT, 2,6-di-tert-buamounts (8). When this compound is administered to tyl-4-methylphenok BHT-OH, 6-tert-butyl-2-(hydroxy-tert-butyl)-4methylphenol;QM, 2,6-di-tert-butyl-4-methylene-2,5-cyclohexadienone; mice, the same changes in lung morphology occur as with QM-OH, 6-tert-butyl-2-(hydroxy-tert-butyl)-4-methylene-2,5-cyclo- BHT, but at a 40-fold lower dose (17). In contrast, several hexadienone; GSH, glutathione; BHT-SG, S-(3,5-di-tert-butyl-4hydroxybenzy1)glutathione; BHT-OH-SG,S-[3-tert-butyl-4-hydroxy-5- other metabolites of BHT, including 2,6-di-tert-butyl-4(hydroxy-tert-butyl)benzyl]glutathione. (hydroxymethyl)phenol, 2,6-di-tert-butyl-l,4-benzo-

Introduction

0893-228~ f 90f 2703-0065$Q2.50f 0 0 1990 American Chemical Society

66 Chem. Res. Toxicol., Vol. 3, No. 1, 1990

Bolton et al.

(thioglycerol) m / z 526 (55%) (M + l)', 308 (93%) (GSH + l)', 220 (100%) BHT+. BHT-OH-SG was obtained biosynthetically from incubations of phenobarbital-treated GRS/N mouse liver microsomes and the ion substrate BHT-OH as described below. The metabolite was isolated from the aqueous phase on C-18 extraction cartridges and eluted with methanol. The eluates were concentrated and subjected to semipreparative HPLC on an Altex Ultrasphere ODS column (10 X 250 mm) with a flow rate of 3.5 mL/min and the same mobile-phase composition described below for the analytical BHT-OH QY-OH work. By this procedure, 5 mg of pure BHT-OH-SG was isolated from the column effluent with a (2-18 extraction cartridge. The QU \ OSH following spectral data were obtained: 'H NMR (DzO)(in addition to the glutathionyl protons) 6 1.18 (s, 6 H , CH3CO), 1.20 (s, 9 H, \ I tert-butyl), 3.38 (s, 2 H, CH,O), 3.52 (s, 2 H, benzylic CHz),6.88 (s,1 H, ArH), 6.92 (s, 1 H, ArH); UV (CH3OH) A, 238,282 nm; negative ion FAB-MS (dithiothreitol/dithioerythritol,3:l) m / z BHT-OH-SC 541 (12%) M-, 307 (100%) GSH-. so BHT-SC Kinetic Experiments. A freshly prepared solution of QM or Figure 1. Partial scheme for the microsomal oxidative metabQM-OH in acetonitrile (100 r L ) was added to 1 mL of either olism of BHT and formation of glutathione adducts. methanol or 1.0 mM aqueous GSH in a quartz cuvette, producing a quinone methide concentration of 0.1 mM. After vigorous quinone, and 2,6-di-tert-butyl-4-hydroxy-4-methyl-2,5- mixing, the first-order decay was followed by the change in abcyclohexadieneone, do not affect the lungs at doses sorbance a t 287 nm for at least three half-lives with a Hewlett Packard Model 8452 diode array UV spectrophotometer. Rate equivalent to toxic doses of BHT. constants were determined in triplicate. In order to reconcile data supporting the role of quinone Incubations. Male Sprague-Dawley rats (120-160 g) were methide formation in BHT-induced pulmonary toxicity obtained from Sasco Inc. (Omaha, NE); male C3H/2IBG and with the apparent role of tert-butyl hydroxylation in the C57BL/6IBG mice (20-25 g) were obtained from the Institute bioactivation process, we investigated the conversion of of Behavioral Genetics, University of Colorado; and male GRS/N BHT to QM-OH via BHT-OH (Figure 1)in liver and lung mice (20-25 g) were bred a t the University of Colorado from microsomes from rats and mice. The results demonstrate breeding pairs provided by the National Institutes of Health. that mice produce QM-OH readily, but rats produce only Microsomes were prepared from the livers or lungs of untreated trace amounts in liver and no detectable levels in lung. or phenobarbital-treated rats or mice, and protein and cytochrome QM-OH is substantially more reactive than QM due to P-450 concentrations were determined as described previously (8). Incubations containing 1 mg/mL microsomal protein were intramolecular hydrogen bonding that increases the elecconducted for 20 min a t 37 "C in 50 mM phosphate buffer (pH trophilicity of the 4-methylene group. It is proposed that 7.4, 1 mL total volume). Substrates, utilized a t a final concenthe species-specific formation of the hydroxylated quinone tration of 0.5 mM, were added as solutions in dimethyl sulfoxide methide, which is activated toward covalent binding with (5 pL/mL of incubate), and [3H]GSH (specific activity 1.0 cellular nucleophiles, is responsible for the selective toxicity mCi/mmol) was added in phosphate buffer to achieve a final of BHT in mouse versus rat lung. concentration of 5.0 mM. An NADPH-generating system consisting of 0.4 mM NADP+, 7.5 mM glucose &phosphate, and 1 Experimental Section unit/mL glucose-6-phosphate dehydrogenase was used together with 5.0 mM MgCl,. For control incubations, NADP+ was Chemicals. All chemicals were purchased from Aldrich unless omitted. The reactions were initiated by the addition of NADP+ stated otherwise. [3H]GSH (gl~cine-2-~H) was obtained from NEN and terminated by chilling in an ice bath. and diluted to a specific activity of 1.0 mCi/mmol. The synthesis Metabolite Quantitation. Perchloric acid (50 KL)was added of BHT-OH was described previously (8). Quinone methides were to precipitate microsomal protein, and the incubates were cenprepared by adding 380 mg of PbOz to 25 mg of either BHT or trifuged at 13000 rpm for 6 min. Aliquots of the supernatant were BHT-OH in 100 mL of pentane and stirring the mixture for 2 injected directly onto a Beckman Model 332 HPLC system with h a t 25 "C. The mixture was filtered, acetonitrile added, and the a 4.6 x 150 mm Ultrasphere C-18 column and either a Hitachi pentane removed by evaporation. Quinone methides were stored Model 100-40 UV detector operated a t 280 nm or a Hewlett a t -20 "C in acetonitrile solution for a maximum of 24 h and the Packard Model 1040M diode array UV detector. The mobile concentrations determined by UV spectroscopy a t 287 nm with phase consisted of 40% methanol in 0.1 M NH4H2P04a t 1.0 e 2.82 X lo4 (18). mL/min for 5 min, followed by a linear increase to 90% methanol BHT-SG was prepared from 2,6-di-tert-butyl-4-(chloroover 15 min. For quantitation, 0.3-mL aliquots of the column methy1)phenol. A solution of 200 mg of 2,6-di-tert-butyl-4-(hyeffluent were collected and radioactivity was measured with a droxymethy1)phenol in 20 mL of hexane was mixed with 10 mL Beckman Model LS 8000 liquid scintillation counter. Concenof concentrated HCI under an inert atmosphere for 20 h a t 25 trations of the GSH conjugates were calculated by summing the "C. The hexane fraction was washed with water and dried over radioactivity associated with each peak and converting the data MgS04 and the solvent evaporated. To 1.18 g of crude product to nanomolar amounts according to the known specific activity was added 1.42 g of GSH dissolved in 10 mL of a 2:l (v/v) mixture of the [3H]GSH. of ethanol/water, followed by 1.94 mL of triethylamine and 5 mg of 4-(N,N-dimethylamino)pyridinewith stirring a t 25 "C for 3 Glutathione Adducts in Vivo. A group of 19 phenoh. The ethanol was removed, the aqueous layer containing barbital-treated C57 mice were treated by intraperitoneal injection with BHT (1500 mg/kg) dissolved in corn oil. Four hours later, BHT-SG was filtered and extracted with chloroform, and the product was isolated with a C-18 extraction cartridge (Baker). the mice were killed and bile was removed from the gall bladder BHT-SG was eluted from the cartridge with methanol, the solvent by syringe. Bile collections were pooled, the conjugates isolated evaporated, and the product crystallized from a mixture of ethwith a C-18 extraction cartridge, and products analyzed by HPLC anol/water. BHT-SG was purified further by flash chromatogas described above. A portion of the bile extract was incubated with 90 units of P-glucuronidase (Sigma) in pH 7.4 phosphate raphy on silica gel using 5% acetic acid in methanol. The following buffer at 37 "C for 10 min, and an aliquot was analyzed by HPLC. 6 1.21 (s, 18 H, spectral data were obtained: 'H NMR (DzO) tert-butyl), 7.06 (s, 2 H, ArH) [the ArCH,S protons were obscured Instrumentation. 'H NMR spectra were obtained with a by glutathione protons (signals associated with the glutathionyl Varian Gemini 300-MHz spectrometer and FAB mass spectra moiety were consistent with published spectra of other GSH recorded with a VG 7070 EQ instrument. GC/MS analysis of adducts, ref 19)]; UV (CH,OH) A,, 230, 278 nm; FAB-MS quinone methides was conducted on a Hewlett Packard 5988

\

\

PH

+y-

Chem. Res. Toxicol., Vol. 3, No. 1, 1990 67

Quinone Methides from BHT

R H b

h S

+@ +CH,

Figure 3. Structure of QM-OH showing intramolecularhydrogen bonding.

. f

0)

0 C

O

Lll I.

mass/charge

am

Y

sn

(.a

fZ

n

. f

0

a29

Figure 2. Mass spectra of quinone methides. QM m/z 218 (M)+, 203 (M - CH3)+,189 (M - CzH&+,175 (M - C3H7)+,161 (M C4H9)+.QM-OH: m / z 234 (M)', 216 (M - HzO)+,201 (M - HzO - CH3)+,173 (M - Hz0 - C3H7)+,159 (M - HzO - C4H9)+.

1

4000

6.8

3 008

2000

I 000

J.

P t

Table I. 'H NMR Data for Quinone Methides compd chemical shift 8, ppma QM QM-OH

1.27 ( 8 , 18 H) 5.75 (s, 2 H) 6.92 (s, 2 H) 1.24 (9, 6 H) 1.26 (s, 9 H) 3.68 (a, 1 H), 3.69 (s, 1 H) 5.84 (s, 1 H), 5.85 (s, 1 H) 6.98 (d, 1 H, J 2.1 Hz) 7.01 (d, 1 H, J 2.1 Hz)

Compounds were synthesized and spectra recorded in hexadeuteriobenzene. equipped with a 12-m capillary column that was programmed from 100 to 180 "C at 8 OC/min. FT-IR analysis was carried out on a Hewlett Packard 5890 GC equipped with a 59656 FT-IR detector, and solution infrared work was conducted in CC14 with a Mattson Polaris FT-IR. Electrochemical determinations were performed in acetonitrile solutions of the phenols (2 mg/mL) containing tetrabutylammonium perchlorate (0.3 M) with a Bioanalytical Systems CV-27 voltammograph, a glassy carbon working electrode, and a Ag/AgCl reference electrode. The scan rate was 200 mV/s.

Results Chemical Characteristics of QM and QM-OH.

Oxidation of BHT or BHT-OH in pentane with lead dioxide yielded solutions of the corresponding quinone methides with UV absorption maxima at 287 and 289 nm and the mass spectra shown in Figure 2. Both compounds yielded molecular ions and ions corresponding to losses of hydrocarbon fragments. QM-OH also lost water during the fragmentation process. The 'H NMR data are presented in Table I. It is noteworthy that the vinyl protons of QM-OH are shifted slightly downfield relative to those of QM, revealing the deshielding influence of the hydroxyl group on these remote protons. This effect could be due

4808

3000

aai

1008

wavenumber ( c m ' ) Figure 4. Gas-phase infrared spectra of (A) QM and (B)QM-OH

from GC/FT-IR analysis. The broad 0-H peak at 3516 cm-' in the latter spectrum is assigned to the fraction that is hydrogen bonded and the peak at 3671 cm-l to the free fraction (20).

to intramolecular hydrogen bonding as shown in Figure 3, which would increase the positive charge density on the methylene group. Infrared analysis of QM-OH in the gas phase (Figure 4) supports such an interaction; the spectrum includes two absorption bands corresponding to 0-H stretching vibrations in the region 3500-3700 cm-', which is characteristic of a mixture of hydrogen-bonded and free hydroxyl groups (20). Hydrogen bonding also produces a shift in the carbonyl absorption of QM-OH to slightly lower frequency relative to QM. Similar infrared results were obtained with solutions of the quinone methides at concentrations that were sufficiently low to eliminate intermolecular hydrogen bonding (20). In order to obtain additional evidence for an interaction between the side-chain hydroxyl and the ring oxygen of the aromatic resonance form shown in Figure 3, the electrochemical oxidation of BHT and BHT-OH was inves-

68 Chem. Res. Toxicol., Vol. 3, No. 1, 1990 Table 11. Rate Constants for Reactions of Quinone Methides with NucleoDhiles" rate constants, sW1 nucleophile QM QM-OH 1.2 x 104 2.9 x 10-3 methanol glutathione 3.5 x 10-3 2.0 x 10-2

" Reactions were conducted either in methanol or in an aqueous solution of glutathione as described under Experimental Section. Pseudo-first-order rate constants were calculated from the disappearance of the UV absorbance corresponding to the quinone methides.

0

2

4

6

8

10

12

Bolton et al. Table 111. Formation of Quinone Methides by Rat and Mouse Microsomes source of nmol/(nmol of P450.min) microsomes" substrate QM QM-OH rat liver BHT 1.3 f 0.2* 0.09 f 0.02 BHT-OH 3.5 f 0.6 BHT 0.36 f 0.03 nd' rat lung BHT-OH 0.38 f 0.05 mouse liver BHT 0.30 f 0.03d 1.25 f 0.15d BHT 0.27 f 0.02e 0.85 f 0.06e BHT-OH 3.4 f 0.6d mouse lung BHT 0.39 f 0.06e 0.38 f 0.12O BHT-OH 2.4 f O.le

" Incubations were conducted for 20 min with 0.5 mM substrate and microsomes prepared from rats or mice that had been treated with phenobarbital as described under Experimental Section. This treatment had no discernible effect on quinone methide formation by lung microsomes, but the oxidations of BHT and BHT-OH to the respective quinone methides by liver microsomes were enhanced 3- to 7-fold compared with untreated animals of both species (manuscript in preparation),consistent with previously reported data for QM formation (15). *Resultsare the means f SE of three determinations. It should be noted that the conversion of BHT to QM-OH requires two oxidative reactions, so the rates of QM and QM-OH production from their immediate precursors BHT or BHT-OH, respectively, provide the best comparison of rates of quinone methide formation. Not detected. Microsomes from male GRS/N mice. eMicrosomes from male C3H/2IBG mice. A.

B.

TIME ( M I N I

Figure 5. HPLC analysis of adducts produced from BHT by rat and mouse liver microsomes in the presence of an NADPHgenerating system and 5.0 mM [3H]GSH. Radioactivity eluting from the HPLC column was measured in fractions collected at 15-9 intervals.

tigated. Cyclic voltammetry yielded peak potentials of 1.93 and 1.74 V, respectively, for these phenols. The lowering of the oxidation potential of BHT by a peripheral hydroxyl substituent is consistent with the proposed intramolecular hydrogen bonding. Reactivities of the quinone methides were determined under pseudo-first-order conditions in the presence of either methanol or GSH. The rate constants (Table 11)demonstrate that QM-OH is considerably more reactive than QM, with half-lives of about 35 and 200 s, respectively, with excess GSH. Enzymatic Formation of Quinone Methides. Direct analysis by GC/MS and HPLC revealed the formation of small amounts of QM in incubates of BHT with rat or mouse liver microsomes but failed to confirm the formation of QM-OH. The greater electrophilicity of QM-OH compared to QM could account for the absence of the former in postincubate fractions due to more efficient scavenging by nucleophiles. The in vitro formation of QM has been assayed previously by conducting reactions in the presence of GSH and measuring the formation of the adduct BHT-SG (Figure 1)(12). In order to utilize this approach in the present work, the GSH adducts of both quinone methides were prepared and characterized by UV, 'H NMR, and FAB MS, and a HPLC method utilizing 13H]GSH was developed to quantitate these adducts (Figure 5 ) . The data summarized in Table 111clearly demonstrate a species-related difference; although liver microsomes

I 12 16 20 12 16 20 TIME (MINI

Figure 6. HPLC analysis (with UV detection at 280 nm) of conjugates extracted from the bile of BHT-treated mice (A) before and (B) after incubation with @-glucuronidase.The retention times and UV spectra of peaks 3 and 4 corresponded to those of BHT-OH-SGand BHT-SG, respectively. The remaining peaks were not identified. from both species produce similar total amounts of quinone methides, rats preferentially oxidize BHT to QM but mice predominantly form QM-OH. Because liver microsomes from both species readily oxidize BHT-OH to QM-OH, the previously reported inability of rats to hydroxylate a tert-butyl group of BHT (8) explains the very low levels of QM-OH produced by that species. Rat lung microsomes do not generate detectable quantities of QM-OH from BHT and are much less active than mouse microsomes in oxidizing of BHT-OH to QM-OH. In addition, the total quinone methide production from BHT in lung microsomes is about 2-fold greater in mouse than in rat tissue.

Conversion of BHT to Quinone Methides in Vivo. After the administration of BHT to mice, bile was removed and conjugates were extracted and analyzed by HPLC. The chromatograms in Figure 6 indicate the presence of both BHT-SG (peak 4) and BHT-OH-SG (peak 3), because the retention times and W spectra of the metabolite peaks were indistinguishable from those of the authentic GSH adducts, and the peaks were unaffected by incubation with @-glucuronidase.Peaks 1 and 2, however, disappeared with

Quinone Methides from BHT &glucuronidase treatment, and two new compounds were formed (peaks 5 and 6), suggesting that the former two peaks were due to glucuronide conjugates and the latter two to the corresponding aglycons.

Discussion Liver and lung microsomes from both rats and mice catalyze the two-electron oxidation of BHT to the corresponding quinone methide. This electrophilic metabolite, formed by a cytochrome P-450 dependent process (12), binds covalently to GSH and protein thiol groups (13). A large body of evidence links this pathway to the pulmonary toxicity elicited by BHT in mice. In addition to the supporting data already cited, Mizutani et al. (21) investigated isotope effects associated with the metabolism and toxicity of BHT after deuteration of the 4-methyl group. This labeled analogue yielded substantially less QM both in vitro and in vivo, with a concomitant reduction in toxicity. Although there is little doubt that the pulmonary toxicity of BHT is intimately related to quinone methide formation, the fact that toxicity is expressed in mice but not rats has never been explained. Recent findings that BHT-OH is a major metabolite in mouse lung but not in rat lung (8),and that BHT-OH is a substantially more potent mouse lung toxin than BHT (In,suggest that tert-butyl hydroxylation is a critical event in the bioactivation process. A study of relationships between pulmonary toxicity and chemical structure for a series of alkylphenols demonstrated the requirement for at least one tert-butyl group ortho to the phenolic hydroxyl and a methyl or ethyl group in the para position of the ring (22). These results also are consistent with the proposal that an important step in the bioactivation of BHT involves hydroxylation followed by oxidation to the ultimate toxin QM-OH (17). Results of the present investigation provide evidence to support a role for QM-OH in the mouse lung toxicity exhibited by BHT. Quinone methides in general are strongly electrophilic with the center of positive charge density on the exocyclic methylene group (23). This property was enhanced by hydroxylation of QM; the half-life for reaction of GSH with QM-OH was only about one-sixth of that for QM. Evidence was provided by IR, NMR, and electrochemical measurements that this increased reactivity with nucleophiles is due to a hydrogen-bonding interaction between the hydroxyl group and the ring oxygen, which stabilizes the resonance form shown in Figure 3. Measurements of quinone methide formation from BHT in microsomal incubations demonstrated that although both rats and mice produce QM, only the latter species generates QM-OH readily; rats produce trace amounts of QM-OH in liver and none in lung. Also, rat lung microsomes were only one-sixth as active as mouse lung in forming QM-OH with BHT-OH as the substrate. These results indicate that attempts to induce pulmonary toxicity in rats by administering BHT-OH to circumvent the deficiency in tert-butyl hydroxylation would not be likely to succeed. The organ specificity of BHT toxicity in mice could be due the fact that the concentrations of GSH in mouse liver are relatively high compared to lung; BHT causes hepatic necrosis only after GSH is depleted by administering buthionine sulfoximine (24). In conclusion, our evidence suggests that the scheme advanced previously for metabolic activation of BHT to a pulmonary toxin in mice should be modified. Rather than direct oxidation to QM by cytochrome P-450 (12) and/or peroxidases (25),and subsequent covalent binding to cellular nucleophiles, a two-step metabolic process in-

Chem. Res. Toxicol., Vol. 3, No. 1, 1990 69 volving tert-butyl hydroxylation and conversion of BHT-OH to the reactive electrophile QM-OH is more consistent with the published data. The occurrence of this pathway in vivo was confirmed when the glutathione conjugate of BHT-OH was detected in bile from mice injected with BHT.

Acknowledgment. This work was supported by USPHS Grant CA 33497. Registry No. BHT, 128-37-0;QM, 2607-52-5;QM-OH, 124755-19-7;BHT-OH, 112700-14-8;BHT-SG, 88332-40-5; BHT-OH-SG, 124755-20-0; 2,6-di-tert-butyl-4-(hydroxymethyl)phenol, 88-26-6; 2,6-di-tert-butyl-4-(chloromethyl)phenol, 955-01-1;glutathione, 70-18-8.

References (1) Wattenberg, L.W. (1978)Inhibition of chemical carcinogenesis. Adu. Cancer Res. 26,197-226. (2) Kensler, T. W., Egner, P. A., Trush, M. A., Bueding, E., and Groopman, J. D. (1985)Modification of aflatoxin B1 binding to DNA in uiuo in rats fed phenolic antioxidants, ethoxyquin and a dithiothione. Carcinogenesis 6, 759-763. (3) Takahashi, O.,and Hiraga, K. (1978)Dose-response study of haemorrhagic death by dietary butylated hydroxytoluene (BHT) in male rats. Toxicol. Appl. Pharmacol. 43, 399-406. (4) Marino, A. A,, and Mitchell, J. T. (1972)Lung damage in mice following intraperitoneal injection of butylated hydroxytoluene. Exp. Biol. Med. 140,122-125. Proc. SOC. (5) Witschi, H. P., Malkinson, A. M., and Thompson, J. A. (1989) Metabolism and pulmonary toxicity of butylated hydroxytoluene (BHT). Pharmacol. Ther. 42,89-113. (6)Witschi, H. P., Williamson, D., and Lock, S. (1977)Enhancement of urethan tumorigenesis in mouse lung by butylated hydroxytoluene. J. Natl. Cancer Inst. 58,301-305. (7) Kehrer, J. P., and Witschi, H. P. (1980)Effects of drug metabolism inhibitors on butylated hydroxytoluene induced pulmonary toxicity in mice. Toxicol. Appl. Pharmacol. 53,333-342. (8) Thompson, J. A,, Malkinson, A. M., Wand, M. D., Mastovich, S. L., Mead, E. W., Schullek, K. M., and Laudenschlager, W. G. (1987)Oxidative metabolism of butylated hydroxytoluene by hepatic and pulmonary microsomes from rats and mice. Drug Metab. Dispos. 15, 833-840. (9) Takahashi, O., and Hiraga, K. (1979) 2,6-Di-tert-butyl-4methylene-2,5-cyclohexadienone:a hepatic metabolite of butylated hydroxytoluene in rats. Food Cosmet. Toxicol. 17,451-454. (10) Tajima, K., Yamamoto, K., and Mizutani, T. (1981)Biotransformation of butylated hydroxytoluene (BHT) to BHT-quinone methide in rats. Chem. Pharm. Bull. 29, 3738-3741. (11) Tajima, K., Yamamoto, K., and Mizutani, T. (1983)Identification and determination of glutathione and glucuronide conjugates formed from butylated hydroxytoluene in rats. Chem. Pharm. Bull. 31,3671-3677. (12) Tajima, K., Yamamoto, K., and Mizutani, T. (1985)Formation of a glutathione conjugate from butylated hydroxytoluene by rat liver microsomes. Biochem. Pharmacol. 34,2109-2114. (13) Nakagawa, Y., Hiraga, K., and Suga, T. (1983)On the mechanism of covalent binding of butylated hydroxytoluene to microsomal protein. Biochem. Pharmacol. 32, 1417-1421. (14) Nakagawa, Y., Hiraga, K., and Suga, T. (1979)Biological fate of butylated hydroxytoluene (BHT); binding in uiuo of BHT to macromolecules of rat liver. Chem. Pharm. Bull. 27, 442-446. (15) Mizutani, T., Nomura, H., Yamamoto, K., and Tajima, K. (1984)Modification of butylated hydroxytoluene-induced pulmonary toxicity in mice by diethyl maleate, buthionine sulfoximine, and cysteine. Toxicol. Lett. 23,327-331. (16) Nakagawa, Y.,Suga, T., and Hiraga K. (1984)Preventive effect of cysteine on butylated hydroxytoluene-induced pulmonary toxicity in mice. Biochem. Pharmacol. 33, 502-505. (17) Malkinson, A. M., Thaete, L. G., Blumenthal, E. J., and Thompson, J. A. (1989)Evidence for a role of tert-butyl hydroxylation in the induction of pneumotoxicity in mice by butylated hydroxyltoluene. Toxicol. Appl. Pharmacol. 101,196-204. (18) Filar, L. J., and Winstein, S. (1960)Preparation and behavior of simple quinone methides. Tetrahedron Lett. 25,9-16. (19) Rashed, M. S.,and Nelson, S. D. (1989)Characterization of glutathione conjugates of reactive metabolites of 3’-hydroxyacetanilide, a nonhepatotoxic positional isomer of acetaminophen. Chem. Res. Toxicol. 2,41-45.

70 Chem. Res. Toxicol., Vol. 3, No. 1, 1990 (20) Tichy, M. (1965) The determination of intramolecular hydrogen

bonding by infrared spectroscopy and its applications in stereochemistry. Adu. Org. Chem. 5 , 115-298. (21) Mizutani, T., Yamamoto, K., and Tajima, K. (1983) Isotope effects on the metabolism and pulmonary toxicity of butylated hydroxytoluene in mice by deuteration of the 4-methyl group. Toxicol. Appl. Pharmacol. 69, 283-290. (22) Mizutani, T., Ishida, I., Yamamoto, K., and Tajima, K. (1982) Pulmonary toxicity of butylated hydroxytoluene and related alkylphenols: structural requirements for toxic potency in mice. Toxicol. Appl. Pharmacol. 62, 273-281.

Bolton et al. (23) Wagner, H.-U., and Gomper, R. (1974) Quinone methides. In The Chemistry of the Quinonoid Compounds, Part 2 (Patai, S., Ed.) pp 1145-1175, Wiley, New York. (24) Mizutani, T., Nomura, H., Nakanishi, K., and Fujita, S. (1987) Hepatotoxicity of butylated hydroxytoluene and its analogs in mice depleted of hepatic glutathione. Toxicol. Appl. Pharmacol. 87, 166-176. (25) Thompson, D. C., and Trush, M. A. (1986) The toxicological

implications of the interaction of butylated hydroxytoluene with other anitoxidants and phenolic chemicals. Food Chem. Toxicol. 24, 1189-1195.