Immunochemical Detection of Covalently Modified Protein Adducts in

Chem. Res. Toxicol. 1996, 9, 713-721

713

Immunochemical Detection of Covalently Modified Protein Adducts in Livers of Rats Treated with Methyleugenol Iain Gardner,* Pauline Bergin, Peter Stening, J. Gerald Kenna, and John Caldwell Department of Pharmacology and Toxicology, Imperial College School of Medicine at St. Mary’s, Norfolk Place, London W2 1PG, United Kingdom Received December 18, 1995X

Methyleugenol is an allylbenzene food flavoring which has been shown to form DNA and protein adducts, and to cause hepatotoxicity and carcinogenicity in rodents. In order to investigate the nature of the protein adducts, specific antisera were raised by immunizing rabbits with conjugates prepared by coupling 1′-acetoxymethyleugenol, or its acidic congener 3,4-dimethoxycinnamic acid, to rabbit serum albumin (RSA). These polyclonal antisera were shown by enzyme linked immunosorbent assay (ELISA) to contain antibodies which recognized the 3,4-dimethoxyphenyl ring portion of methyleugenol. Analysis of livers from rats given methyleugenol ip for 5 days, at doses between 10 and 300 mg/kg/day, revealed dose-dependent formation of novel protein adducts which were recognized by the antisera. The adducts were detected by ELISA and by immunoblotting and were concentrated in the microsomal fraction, and were shown in inhibition studies to be derived from methyleugenol. A 44 kDa adduct was the only protein adduct detected in livers of rats given low doses of methyleugenol (10 or 30 mg/kg/day) and was the major adduct detected in rats given high doses of the compound (100 and 300 mg/kg/day). This adduct was solubilized when microsomal fractions were extracted using 0.1 M sodium carbonate, implying that it is a peripheral membrane protein. A pattern of protein adducts which mirrored the in vivo situation was generated when rat hepatocytes were incubated with 1′-hydroxymethyleugenol in vitro, but could not be reproduced in experiments undertaken using liver microsomes or postmitochondrial supernatants. These findings imply that generation of protein adducts in livers of rats given methyleugenol in vivo proceeds via the 1′-hydroxy metabolite and requires crucial cofactors, and/or structural features, which are present in intact hepatocytes but not in broken cell preparations and which remain to be defined.

Introduction Methyleugenol (3,4-dimethoxyallylbenzene) is a naturally occurring food flavoring which is of toxicological concern because it is a high-dose rodent hepatotoxin and carcinogen (1, 2) to which the human population is exposed at low levels. It is present in the diet as a component of a number of spices, including sweet bay, clove oil, nutmeg, and allspice. In addition, methyleugenol is a permitted food additive in the United States, is used as a fragrance material, and has been employed as a lure for the control of the male oriental fruit fly (3). Although the maximal safe daily intake of methyleugenol in humans is unknown, the Council of Europe (1981) has set a temporary acceptable daily intake of 2.5 mg/kg/day (4). The toxicity of methyleugenol and other allylbenzenes, such as safrole and estragole, is believed to involve bioactivation to electrophilic intermediates, which react with DNA and other cellular macromolecules, including proteins (1). Three distinct pathways of bioactivation have been described (Scheme 1). The first pathway is initiated by hydroxylation of the 1′-carbon atom of the allyl side chain by cytochromes P450 (5). The resulting alcohol is sulfated enzymatically, producing an unstable * Author to whom correspondence should be addressed. Tel: (+44) 171 594 3876; Fax: (+44) 171 723 7535. X Abstract published in Advance ACS Abstracts, April 15, 1996.

S0893-228x(95)00211-6 CCC: $12.00

intermediate which decomposes spontaneously to an electrophilic carbonium ion which binds covalently to hepatic DNA and proteins (1, 6-10). The second pathway proceeds via cytochrome P450 catalyzed epoxidation of the side chain (5, 11), while the third pathway, which has been identified recently for eugenol (12-14) and safrole (15), involves cytochrome P450 bioactivation to electrophilic and cytotoxic quinone methide metabolites. Whereas the nature, formation, and repair of DNA adducts derived from methyleugenol and other allylbenzenes has been the focus of intensive investigations (7, 8, 16, 17), the nature of the protein adducts and/or their possible toxicological significance have yet to be evaluated. In the present study, this potentially important issue has been investigated using an immunochemical approach. Specific polyclonal antisera were produced by immunization of rabbits with conjugates prepared by coupling analogues of methyleugenol to rabbit serum albumin (RSA) (Scheme 2). These antisera were used to identify protein adducts formed in livers of rats given methyleugenol in vivo and to investigate the mechanisms responsible for protein adduct formation.

Experimental Procedures Caution: Methyleugenol, 1′-hydroxymethyleugenol, and acetoxymethyleugenol are potential carcinogens and must be handled with the appropriate safety precautions. Materials. Methyleugenol, safrole, eugenol, isomethyleugenol, and estragole were purchased from Aldrich Chemical Co.

© 1996 American Chemical Society

714 Chem. Res. Toxicol., Vol. 9, No. 4, 1996

Gardner et al.

Scheme 1. Pathways of Bioactivation of Methyleugenol

Scheme 2. Proposed Structures of the Conjugates Formed by Reacting Acetoxymethyleugenol (left panel) or Dimethoxycinnamic Acid (right panel) with Rabbit Serum Albumin (RSA)

(Dorset, U.K.). Acrylamide (40%), N,N,N′,N′,-tetramethylethylenediamine, ammonium persulfate and dithiothreitol were purchased from Bio-Rad Laboratories Ltd. (Herts, U.K.). Unless otherwise stated, all other biochemicals were from Sigma Chemical Co. (Dorset, U.K.). 1′-Hydroxymethyleugenol was synthesized as described by Solheim and Scheline (18). 1′Acetoxymethyleugenol (AME)1 and 3′-hydroxyisomethyleugenol were synthesized from 1′-hydroxymethyleugenol according to Borchert et al. (19). These compounds were purified by column chromatography and were shown by TLC and capillary gas chromatography to be >99% pure. The compounds had 1H NMR spectra which were consistent with the proposed structures and

gave the following characteristic fragments when analyzed by mass spectrometry: 1′-hydroxymethyleugenol, m/z 194, 163, 139; 1′-acetoxymethyleugenol, m/z 236, 194, 177; 3′-hydroxyisomethyleugenol, m/z 194, 177. Preparation of Hapten-RSA Conjugates. The methods used to produce the two different hapten RSA conjugates are shown schematically in Scheme 2. In the first method, 1′acetoxymethyleugenol (500 mg) was dissolved in dimethylformamide (2 mL), mixed with RSA (40 mg in 20 mL of 160 mM sodium phosphate, pH 8.0), and incubated overnight at room temperature with stirring. In the second method, 1-ethyl3-[3-(dimethylamino)propyl]carbodiimide hydrochloride (EDC)

Methyleugenol-Protein Adducts in Rats (288 mg, 1.5 mmol) was dissolved in 7.5 mL of 20 mM sodium phosphate buffer (pH 5.0) and added to 2.5 mL of a methanolic solution of 3,4-dimethoxycinnamic acid (DMCA) (62.5 mg, 0.3 mmol). After 2 min, this reaction mixture was added to RSA (40 mg in 20 mL of 160 mM sodium phosphate, pH 8.0), and incubation was allowed to proceed overnight at room temperature, with stirring. Conjugates prepared by both methods were dialyzed against three changes of distilled water (3 L) and then lyophilized. An estimate of the hapten density of the DMCARSA conjugate, obtained by UV-visible difference spectroscopy, using RSA and DMCA as standards, indicated that 29 mol of DMCA was bound per mol of RSA. As albumin has 61 free amino groups (20), this implies that 48% of available free amino groups had been modified by this approach. The same approach could not be used to estimate the hapten density of the AMERSA conjugate because this conjugate was colored yellow. Preparation of Polyclonal Antisera. Two female New Zealand White rabbits (2.5-3 kg), obtained from Froxfield Farms (U.K.), were immunized with the DMCA-RSA conjugate, and two rabbits were immunized with the AME-RSA conjugate. For the primary immunizations, the hapten-RSA conjugates (1 mg) were emulsified in a mixture of 1 mL of filtered phosphate buffered saline (PBS) (which comprised 0.15 M NaCl, 10 mM potassium phosphate, pH 7.4) and 1 mL of Freund’s complete adjuvant. Emulsions (500 µL; 250 µg) were injected subcutaneously at 6-8 sites on the back of each rabbit. After 3 weeks, rabbits were boosted in a similar manner with the hapten-RSA conjugates (500 µg) emulsified in an equivalent volume of Freund’s incomplete adjuvant. Blood was collected from ear veins into glass tubes after a further 2 weeks. After clotting, serum was removed and stored at -20 °C. Administration of Methyleugenol to Rats. Male Fischer 344 rats (220-250 g) were obtained from Harlan Olac Ltd. (Bicester, Oxon, U.K.) and were allowed access to food and water ad libitum. For the single dose study, groups of 4 animals received methyleugenol dissolved in tricaprylin, by ip injection, at a dose of either 10 or 100 mg/kg. Control animals received an equivalent dose of tricaprylin vehicle alone. These animals were sacrificed by cervical dislocation 4 h after dosing. In the repeated dose study, groups of 5 animals were given daily ip injections of methyleugenol dissolved in tricaprylin, at doses of 10, 30, 100, or 300 mg/kg/day, for each of 5 days, while control animals received equivalent volumes of tricaprylin alone. These rats were sacrificed by cervical dislocation 24 h after the final dose. Livers from each dose group were removed, pooled, then placed in ice-cold sucrose buffer (0.25 M sucrose, 15 mM TrisHCl, 0.1 mM EDTA, pH 6.8), and processed as described below. Subcellular Fractionation. All steps were performed at 0-4 °C. Pooled livers were blotted to remove excess buffer, weighed, and then minced thoroughly with scissors. Five volumes of fresh sucrose buffer were added, and homogenates were prepared using eight strokes of a Potter homogenizer at 1000 rpm. The homogenates were strained through muslin and centrifuged at 1000g for 10 min. The resulting pellets were resuspended in fresh sucrose buffer and then centrifuged again. This procedure was repeated once more, and then the final pellets (nuclear fractions) were resuspended in sucrose buffer. The combined supernatants were centrifuged at 10000g for 20 min, and then the pellets were resuspended in fresh sucrose buffer and re-centrifuged. This yielded the pelleted mitochondrial/lysosomal fractions. The resulting supernatants were combined and were centrifuged at 100000g for 1 h to yield 1 Abbreviations: AME, 1′-acetoxymethyleugenol; AME-RSA, the protein conjugate prepared by coupling 1′-acetoxymethyleugenol to rabbit serum albumin; DMCA, 3,4-dimethoxycinnamic acid; DMCARSA, the protein conjugate prepared by coupling 3,4-dimethoxycinnamic acid to rabbit serum albumin; ECL, enhanced chemiluminescence; EDC, 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide hydrochloride; EGTA, ethylenebis(oxyethylenenitrilo)tetraacetic acid; ELISA, enzyme-linked immunosorbent assay; HRP, horseradish peroxidase; PAPS, 3′-phosphoadenosine 5′-phosphosulfate; PBS, phosphate buffered saline (10 mM phosphate: 0.15 M NaCl, pH 7.4); RSA, rabbit serum albumin; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

Chem. Res. Toxicol., Vol. 9, No. 4, 1996 715 cytosolic fractions (the supernatants) and microsomal fractions (the pellets). The microsomal fractions were resuspended in buffer comprising 100 mM potassium phosphate (pH 7.4)/0.1 mM EDTA. All of the subcellular fractions were stored at -80 °C in 1.5 mL aliquots. ELISA. Aliquots (100 µL) of the test antigens (hapten-RSA conjugate or liver subcellular fractions) were incubated for 18 h at 4 °C in wells of 96 well microtiter plates (Immulon 4, Dynatech Laboratories Inc.), at a concentration of 15 µg of protein/mL in PBS. The plates were washed with 4 cycles of wash buffer (casein (0.05% w/v), Tris (10 mM), NaCl (0.15 M), thimerosal (0.5 mM), pH 7.4) using an automated plate washer (Microplate washer S8/12, Titertek). Primary antiserum (100 µL of a 1:1000 dilution in PBS) was added to each well, and plates were incubated at room temperature for 3 h. Plates were washed with wash buffer (4 cycles); then 100 µL of secondary antibody (goat anti-rabbit IgG, horseradish peroxidase (HRP) conjugate, from Serotec Ltd., U.K.), at a dilution of 1/1000 in PBS, was added to the wells. After incubation for a further 2 h at room temperature, the plates were washed with 4 cycles of wash buffer followed by 2 cycles of PBS. Color development was started by addition of 100 µL/well of o-phenylenediamine solution (0.4 mg/mL in 24 mM citrate, 50 mM sodium phosphate buffer, pH 5.0, containing 0.4 µL/mL of 30% H2O2) and was stopped (after 3-5 min) by addition of 50 µL/well of 4M H2SO4. Finally, the A492 of each well was determined using a Titertek Multiscan Plus II automated microtiter plate reader (from Flow Laboratories Ltd, U.K.). When ELISA inhibition studies were undertaken, the primary antisera were incubated for 30 min at room temperature with various concentrations of the inhibitor compounds and then added to the ELISA plates. Where necessary, the inhibitors were dissolved in methanol before addition to the diluted antisera. Control incubations revealed that the low volumes of methanolic solutions used in the studies (always 90%, as assessed by trypan blue exclusion) were diluted in PBS to a final concentration of 6 × 106 cells/mL (final volume 1 mL). Methyleugenol (0, 50, 500 µM) or 1′-hydroxymethyleugenol (0, 50, 500 µM) was added in 5 µL of DMSO vehicle, and cells were incubated for 3 h at 37 °C. The hepatocytes were centrifuged at 2000g for 10 min, and the supernatants were discarded. The pelleted cells were resuspended in sucrose buffer (0.25 M sucrose, 15 mM TrisHCl, 0.1 mM EDTA, pH 6.8) and were snap frozen in liquid nitrogen and then stored at -80 °C. Subsequently, the frozen cells were thawed, sonicated for 30 short pulses (10-15 s per pulse) using a probe sonicator, and then centrifuged at 10000g for 10 min. The supernatants were centrifuged at 100000g for 50 min, at 4 °C, using a Beckmann TL-100 bench top ultracentrifuge and a TLA45 rotor. The 10000g pellets (nuclear and mitochondrial fractions) and the 100000g pellets (microsomal fractions) were resuspended in sucrose buffer and analyzed by immunoblotting. Protein Assay. Protein concentration of subcellular fractions were determined using a Bicinchoninic acid assay kit (Pierce and Warriner, Chester, U.K.).

Results Production and Characterization of Rabbit Antisera. Conjugates were prepared by allowing AME to react spontaneously with RSA and also by carbodiimidemediated coupling of DMCA to RSA. Each conjugate was used to immunize two rabbits. Characterization of the antisera was undertaken by ELISA.2 These studies 2 Data relating to the characterization of the antisera have been submitted as Supporting Information to this paper and can be accessed on the Internet through the American Chemical Society Home Page. These are results from ELISA studies, which involved investigation of antibody binding to DMCA-RSA, AME-RSA, and RSA alone, and also competitive inhibition studies performed using methyleugenol and various structural analogues. The results obtained from the ELISA inhibition studies are summarized in Table 1.

Gardner et al. Table 1. Summary of ELISA Inhibition Studiesa inhibitor

% inhibition % inhibition anti-(AME-RSA) anti-(DMCA-RSA)

methyleugenol DMCA 1′-hydroxymethyleugenol isomethyleugenol

86 ( 1.0 95 ( 1.2 88 ( 1.4 85 ( 1.2

40 ( 1.0 68 ( 1.0 69 ( 2.9 42 ( 1.6

eugenol isoeugenol estragole 1′-hydroxyestragole anethole safrole 1′-hydroxysafrole

0 ( 0.6 4 ( 1.6 1 ( 4.9 3 ( 1.7 4 ( 2.2 7 ( 2.2 10 ( 5.6

12 ( 0.4 1 ( 1.9 10 ( 2.9 13 ( 1.4 14 ( 1.4 1 ( 1.6 6 ( 0.9

a Antisera from rabbits immunized with either AME-RSA (dilution 1:5000) or DMCA-RSA (dilution 1:50 000) were preincubated with methyleugenol or various structural analogues at a concentration of 1 mM and then tested for recognition of AMERSA or DMCA-RSA, respectively. Each data point is the mean ( SD of 3 replicate determinations of the absorbance at 492 nm.

revealed that the rabbit antisera contained high titers of antibodies which recognized the respective haptenRSA conjugates, but not RSA alone, and that antisera from all four rabbits recognized the DMCA-RSA conjugate more strongly than the AME-RSA conjugate (results not shown).2 In addition, inhibition studies2 (summarized in Table 1) showed that preincubation of the various antisera with methyleugenol, isomethyleugenol, 1′-hydroxymethyleugenol, or 3,4-dimethoxycinnnamic acid, prior to addition to the ELISA wells, resulted in inhibition of recognition of the corresponding hapten-RSA conjugates by the antisera. Recognition of the AME conjugate by the anti-(AME-RSA) antisera was inhibited more efficiently than was recognition of the DMCA conjugate by the anti-(DMCA-RSA) antisera (Table 1). Recognition of the conjugates by the antisera was not inhibited by the closely related allylbenzene compounds safrole, eugenol, estragole, anethole, 1′-hydroxyestragole, and 1′-hydroxysafrole. Comparison of the structures of the inhibitory and non-inhibitory compounds (Chart 1) revealed that the inhibitory compounds all contained the 3,4-dimethoxyphenyl ring, while the noninhibitory compounds did not. Detection of Methyleugenol-Protein Adducts Expressed in Rats in Vivo. ELISA studies revealed dose-dependent generation of antigens recognized by anti-(DMCA-RSA) antiserum in subcellular fractions prepared by differential centrifugation from livers of rats treated ip with methyleugenol, for 5 days. The highest levels of antigen expression were detected in the microsomal fraction, although extensive antigen formation was also evident in nuclear and mitochondrial fractions (Figure 1). A similar pattern of results was obtained when the ELISA analysis was repeated using anti(AME-RSA) antiserum (data not shown). Analysis of the same subcellular fractions by immunoblotting revealed dose-dependent expression of a range of novel protein antigens which were expressed in livers of methyleugenol-treated rats, but not in livers of control rats (Figure 2). Although data obtained with only one anti-(AME-RSA) antiserum are presented, essentially identical results were obtained with antisera from each of the four rabbits. The major novel methyleugenolinduced protein antigen exhibited an apparent molecular mass of 44 kDa, as determined by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) under either reducing conditions (Figure 2) or non-reducing conditions (not

Methyleugenol-Protein Adducts in Rats Chart 1. Chemical Structures of Compounds Which (A) Inhibited Binding of Antibodies to the Hapten-RSA Conjugates and (B) Did Not Inhibit Antibody Binding, As Determined by ELISA

shown). This was the only novel protein antigen detected in livers of rats given low doses of methyleugenol (10 or 30 mg/kg/day) and was the major novel antigen expressed in animals given high doses of methyleugenol (>100 mg/ kg/day), although a range of other novel antigens which exhibited higher apparent molecular masses were evident also in livers of rats given the higher doses (Figure 2C). The 44 kDa adduct was detected at highest concentrations in the microsomal fractions, with lower levels evident in nuclear and mitochondrial fractions (Figure 2A). The novel 44 kDa methyleugenol-induced antigen was also expressed in a dose-dependent manner in livers of rats sacrificed 4 h after ip administration of a single dose of methyleugenol (Figure 3). The level of expression of the antigen in these livers was roughly 10-fold lower than the level of antigen expression in livers of rats given equivalent doses of methyleugenol for 5 days, and so its detection required more stringent development of immunoblots. Under these conditions, the anti-(AME-RSA) antiserum was found to recognize a 42 kDa polypeptide

Chem. Res. Toxicol., Vol. 9, No. 4, 1996 717

which was expressed in mitochondrial and nuclear fractions prepared from livers of control animals, that had not received methyleugenol (Figure 3). This nuclear/ mitochondrial polypeptide comigrated with the 44 kDa methyleugenol-induced antigen when the tissue fractions were analyzed using minigels (which have 5 cm resolving gels). However, the 42 kDa polypeptide was clearly resolved from the 44 kDa methyleugenol adduct, which was concentrated in the microsomal fraction but also detectable in the mitochondrial fractions, when analysis was undertaken using longer (12 cm) resolving gels (Figure 3). Furthermore, recognition of the 42 kDa nuclear/mitochondrial polypeptide was not inhibited when the antiserum was preincubated for 30 min with methyleugenol, before application to the nitrocellulose membrane (Figure 4B), or when the antiserum was preincubated with AME-RSA conjugate (data not shown). However, preincubation of the antiserum with methyleugenol (Figure 4), or with the AME-RSA conjugate (not shown), efficiently inhibited recognition of the methyleugenolinduced 44 kDa protein antigen by the antiserum. In view of this finding, it is clear that the polyclonal rabbit antiserum contains at least two distinct populations of antibodies: first, antibodies which can be inhibited by methyleugenol or by methyleugenol-RSA conjugate, and which recognize protein adducts comprising metabolite(s) of methyleugenol bound covalently to a 44 kDa liver microsomal protein and to other proteins; and second, antibodies which are not inhibited by either methyleugenol or methyleugenol-RSA conjugate, and which recognize a non-methyleugenol-modified 42 kDa polypeptide present at low levels in rat liver nuclear and mitochondrial fractions. Solubilization of the 44 kDa Methyleugenol Adduct with Sodium Carbonate. The 44 kDa methyleugenol adduct was solubilized when microsomal fractions from livers of methyleugenol-treated rats were extracted with 0.1 M sodium carbonate. This treatment was seen, by Coomassie Blue staining of gels, to solubilize a selected subset of total microsomal proteins (data not shown). Protein-Adduct Formation in Vitro in Liver Subcellular Fractions. Concentration- and timedependent formation of methyleugenol-protein adducts was achieved by incubation of microsomal suspensions prepared from the livers of untreated rats with methyleugenol, in the presence of an NADPH generating system. No methyleugenol-protein adducts were detectable in incubations in which either methyleugenol or the NADPH generating system was omitted (data not shown). However, the molecular masses of the polypeptide adducts generated in vitro differed markedly from those of the adducts generated in vivo in livers of methyleugenoltreated rats (Figure 5). In particular, the major adduct generated in vitro had a molecular mass of about 52 kDa, while prolonged development of immunoblots revealed lower levels of many other adducts, but only extremely low levels of the 44 kDa adduct (Figure 5). Adduct formation in postmitochondrial supernatants, which comprise a mixture of microsomes and cytosol, and which therefore contain both cytochromes P450 and sulfotransferases, was also investigated. Significant adduct formation was evident after incubation of these preparations with methyleugenol plus an NADPH generating system, plus or minus the sulfation cofactor PAPS, while only very low levels of adducts were gener-

718 Chem. Res. Toxicol., Vol. 9, No. 4, 1996

Gardner et al.

Figure 1. Detection by ELISA of methyleugenol adducts expressed in subcellular fractions prepared from livers of rats treated ip with methyleugenol. Groups of rats (n ) 5 per group) were treated ip with methyleugenol, at doses of (A) 0, (B) 10, (C) 30, (D) 100, and (E) 300 mg/kg/day, for 5 days. The primary antiserum was anti-(DMCA-RSA) antiserum (dilution 1:1000). Each data point is the mean ( SD of triplicate determinations.

Figure 2. Detection by immunoblotting of methyleugenol adducts in subcellular fractions from livers of rats treated with methyleugenol for 5 days. All blots were developed with anti-(AME-RSA) antiserum (dilution 1:2000). (A) Analysis of nuclear (nuc), mitochondrial (mit), microsomal (mic), and cytosolic (cyt) fractions prepared from the livers of rats (n ) 5 per group) treated ip with 0, 30, or 300 mg/kg/day methyleugenol for 5 days. Protein loading was 20 µg/lane, and exposure of X-ray film was for 10 s. (B) Analysis of microsomal fractions prepared from the livers of rats (n ) 5 per group) treated with 0, 10, 30, 100, or 300 mg/kg/day methyleugenol ip for 5 days. Protein loading was 20 µg/lane, and exposure of X-ray film was for 20 s. (C) Analysis of nuclear (nuc), mitochondrial (mit), microsomal (mic), and cytosolic (cyt) fractions from the livers of rats (n ) 5 per group) treated with 0 (-) or 300 (+) mg/kg/day methyleugenol ip for 5 days. Protein loading was 40 µg/lane, and exposure of X-ray film was for 2 min.

Figure 3. Detection by immunoblotting of methyleugenol adducts in subcellular fractions from the livers of rats treated with a single dose of methyleugenol. Subcellular fractions were prepared from groups of rats (n ) 4 per group) that were sacrificed 4 h after ip treatment with methyleugenol, at doses of 0, 10, or 100 mg/kg. Nuc ) nuclear, mit ) mitochondrial, mic ) microsomal, cyt ) cytosolic. Anti-(AME-RSA) was used as the primary antiserum (1:2000), protein loading was 20 µg/lane, and X-ray film was exposed for 2 min.

ated when the preparations were incubated with methyleugenol and PAPS alone (Figure 6A). The major protein adducts generated in these experiments were again of higher molecular mass than the 44 kDa adduct formed in vivo (Figure 6A). Significant levels of protein

adducts were generated when the postmitochondrial supernatants were incubated with 1′-hydroxymethyleugenol in the presence of PAPS, and formation of a 44 kDa methyleugenol adduct was observed in these studies, although this was not the major adduct produced (Figure 6B). No adducts were formed when the preparations were incubated with 1′-hydroxymethyleugenol without PAPS, with 3′-hydroxymethyleugenol with or without PAPS, or with eugenol in the presence of an NADPH generating system (Figure 6B). Protein Adduct Formation in Vitro in Isolated Hepatocytes. No methyleugenol-protein adducts could be detected following incubation of isolated rat hepatocytes with methyleugenol (Figure 7), although the anti(AME-RSA) antiserum recognized a 42 kDa polypeptide which was expressed in control hepatocytes which had not been exposed to the compound. In contrast, significant levels of methyleugenol adducts were generated when hepatocytes were incubated with 1′-hydroxymethyleugenol (Figure 7). Moreover, a 44 kDa adduct was the only major methyleugenol adduct generated when hepatocytes were incubated with a low concentration of 1′-hydroxymethyleugenol (50 µM). At this concentration of 1′-hydroxymethyleugenol, the 44 kDa adduct was detectable in the 100000g pellet (Figure 7A) but not in

Methyleugenol-Protein Adducts in Rats

Chem. Res. Toxicol., Vol. 9, No. 4, 1996 719

Figure 4. Effect of preincubation with methyleugenol on recognition by anti-(AME-RSA) of protein antigens expressed in subcellular fractions from livers of rats given a single dose of methyleugenol. Groups of rats (n ) 4 per group) received methyleugenol ip at doses of 0 mg/kg (C) or 100 mg/kg (ME) 4 h before sacrifice. (A) Microsomal fractions; (B) mitochondrial fractions. The anti-(AMERSA) antiserum (final dilution 1:500) was preincubated for 30 min with 0 µM, 100 µM, or 1 mM methyleugenol prior to incubation with the immunoblots. Protein loading was 40 µg/lane, and exposure of X-ray film was for 2 min.

Figure 5. Comparison of the pattern of protein adducts formed in vitro, in rat liver microsomes incubated with methyleugenol (ME) (200 µM), with the pattern of protein adducts formed in vivo in microsomes from livers of rats given methyleugenol (ME) ip, at a dose of 300 mg/kg/day for 5 days. The primary antiserum was anti-(AME-RSA) antiserum (dilution of 1:500). Protein loading was 40 µg/lane for in vitro samples and between 0 and 10 µg/lane for in vivo samples as indicated. Exposure of X-ray film was for 1 min.

the 10000g pellet (Figure 7B). At higher concentrations of 1′-hydroxymethyleugenol, while a number of methyleugenol protein adducts were detectable in both the 10000g and 100000g pellets, the 44 kDa adduct was one of the major adducts formed.

Discussion Immunization of rabbits with two synthetic protein conjugates, prepared by coupling structural analogues of methyleugenol to RSA (AME-RSA or DMCA-RSA), resulted in generation of high titers of antibodies which were shown by ELISA to recognize epitopes expressed on both of the conjugates, but not on RSA alone.2 Higher levels of antibody binding were evident when the DMCARSA conjugate was used as test antigen in ELISA’s than when the AME-RSA antigen was used as test antigen, regardless of whether the antibodies were obtained from rabbits immunized with AME-RSA or DMCA-RSA.2 Furthermore, inhibition of antibody binding by methyleugenol, and by related compounds, was markedly more efficient for the two antisera raised against AME-RSA than for the two antisera raised against DMCA-RSA

(Table 1). Whether these differences are attributable to a higher content of haptenic groups on DMCA-RSA, and/ or to important differences between the nature of the haptenic groups expressed on the two conjugates, is unclear. The ELISA inhibition studies revealed that, regardless of which hapten-RSA conjugate was used for immunization, the antisera recognized primarily the 3,4dimethoxyphenyl moiety of methyleugenol and related compounds but not the allyl side-chain region (Chart 1). The antisera were used to detect, by immunoblotting, a wide range of protein adducts derived from methyleugenol. The adducts were expressed in vivo in livers of rats dosed ip with methyleugenol, and in vitro when rat liver microsomal fractions, postmitochondrial supernatants, and isolated hepatocytes were incubated with methyleugenol or 1′-hydroxymethyleugenol, in the presence of appropriate cofactors. The primary objective of the in vivo studies was to identify adducts which might be involved in the hepatotoxicity of methyleugenol, while the in vitro studies were undertaken to investigate possible mechanisms of adduct formation. A single major 44 kDa protein adduct was detected in livers of rats treated ip with methyleugenol for 5 days at doses of 10 or 30 mg/kg/day, and also in livers of animals given a single dose of the compound (Figures 2 and 3). Moreover, the 44kDa adduct was the major protein adduct detected in livers of rats given higher doses of methyleugenol (100 or 300 mg/kg/day) for each of 5 days. Generation of a similar pattern of adducts was achieved in vitro (Figure 7) by incubation of isolated hepatocytes with 1′-hydroxymethyleugenol. This implies that the 1′-hydroxy metabolite, formed by cytochrome P450-mediated oxidation of the allyl side chain of methyleugenol, plays a crucial role in the mechanism of formation of the adducts generated in vivo. 1′-Hydroxyallylbenzene metabolites have been implicated previously in formation of DNA adducts from allylbenzene compounds (6-9). Significant levels of many different protein adducts were generated in vitro when rat liver microsomes or postmitochondrial supernatants were incubated with methyleugenol in the presence, but not in the absence, of an NADPH generating system (Figures 5 and 6). Although the nature of the reactive species was not investigated in the present study, allylbenzenes have been demonstrated to undergo cytochrome P450-medi-

720 Chem. Res. Toxicol., Vol. 9, No. 4, 1996

Gardner et al.

Figure 6. Generation of methyleugenol-protein adducts in rat liver postmitochondrial supernatants in vitro. Anti-(AME-RSA) antiserum was used as primary antiserum (dilution 1:500). Protein loading was 40 µg/lane, and exposure of X-ray film was for 30 s. (A) Rat liver postmitochondrial supernatants were incubated for 120 min with 0 or 500 µM methyleugenol (ME) in the presence of an NADPH generating system and/or PAPS, as indicated. (B) Rat liver postmitochondrial supernatants were incubated for 120 min with 1′-hydroxymethyleugenol (1HM) (0 or 500 µM), 3′-hydroxymethyleugenol (3HM) (500 µM), or eugenol (E) (500 µM), in the presence (+) or absence (-) of an NADPH generating system or PAPS, as indicated. Microsomal fractions prepared from livers of rats (n ) 5) treated ip with methyleugenol at 10 mg/kg/day for 5 days were also analyzed, for comparative purposes.

Figure 7. Generation of methyleugenol-protein adducts in rat hepatocyte suspensions in vitro. Suspensions of isolated hepatocytes were incubated for 180 min with methyleugenol or 1′hydroxymethyleugenol at 0, 50, or 500 µM. The 100000g pellet (A) and the 10000g pellet (B) was analyzed by immunoblotting, as were microsomes from livers of rats (n ) 4) treated ip with a single dose of methyleugenol (0 or 10 mg/kg). Anti-(AMERSA) was used as primary antiserum (dilution 1:500). Protein loading was 40 µg/lane, and exposure of X-ray film was for 30 s.

ated bioactivation to both quinone methide and epoxide metabolites (11-15, 24-27). Methyleugenol cannot directly form quinone methide species but is metabolized in vivo (18) and in vitro (14, 28) to eugenol which can be converted to a protein-reactive quinone methide species (12). However, when eugenol was incubated with postmitochondrial supernatants in the presence of an NADPH generating system, no detectable protein adducts were generated (Figure 6B). This implies that the quinone methide pathway is not involved in production of methyleugenol-modified proteins. Many methyleugenol-protein adducts were also generated when postmitochondrial supernatants were incubated in vitro with 1′-hydroxymethyleugenol plus PAPS, which is the es-

sential cofactor required by cytosolic sulfotransferases (29). However, the 44 kDa adduct either was not generated, or was generated only at extremely low levels, in the in vitro experiments with microsomal preparations and postmitochondrial supernatants (Figures 5 and 6). The reasons for this interesting and important discrepancy are unclear. Perhaps the 44 kDa target protein is located in very close proximity to the site of generation of the reactive metabolite(s) responsible for adduct formation in intact hepatocytes, but this close association is lost during preparation of liver subcellular fractions. Alternatively, transport mechanisms might operate in intact cells, but not in the broken cell preparations, which bring reactive metabolites of methyleugenol in close proximity to the 44 kDa protein. It is also conceivable that covalent modification of the 44 kDa protein requires crucial enzymes other than cytochromes P450 or sulfotransferases, which are inactive in the broken cell preparations, perhaps because of a lack of essential cofactors. Analysis of liver subcellular fractions prepared by differential centrifugation revealed that the 44 kDa adduct was concentrated in the microsomal fraction. This is a complex membrane fraction which contains membrane vesicles derived from the endoplasmic reticulum, the plasma membrane, and various endocytic vesicles (30). In view of this, further investigations will be required in order to determine the true organellar location of the adduct. Although the nature of the 44 kDa carrier protein remains to be determined, its solubility in 0.1 M sodium carbonate is of interest. Extraction of liver microsomal fractions with 0.1 M sodium carbonate has been shown to solubilize peripheral membrane proteins, but not integral membrane proteins (22). Thus, it appears that the 44 kDa protein is not bound to the microsomal membrane via hydrophobic interactions. Other examples of peripheral membrane proteins which are important targets of reactive metabolites are a group of trifluoroacetylated proteins, which are expressed in livers of halothane-treated animals and humans and which have been implicated as target antigens in the mechanism of halothane hepatitis (31). The doses of methyleugenol which were used in the in vivo studies of protein adduct formation in rats covered the range of doses which have been administered to rats previously in rat toxicity studies (2) and also the range of doses being used in an ongoing NTP carcinogenicity

Methyleugenol-Protein Adducts in Rats

bioassay (32). Formation of protein adducts could contribute to toxicity and/or carcinogenicity in several ways. If methyleugenol forms adducts with proteins involved in control of cell growth (e.g. growth factors, growth factor receptors, intracellular signalling molecules, and nuclear transcription factors), then normal cell growth may be disrupted. Given that a recent short term study of the subchronic toxicity of methyleugenol demonstrated that exposure to methyleugenol was associated with liver enlargement (32), it is possible that this compound has the potential to cause cell proliferation, and hence to cause carcinogenesis, by both epigenetic and genotoxic mechanisms. In future studies, it will be important to investigate whether generation of the methyleugenol adducts identified in the present study, and especially the 44 kDa adduct, could contribute to the hepatotoxicity and/or carcinogenicity of methyleugenol, and of other allylbenzenes.

Acknowledgment. The authors would like to thank Julie Spencer, Pharmacology and Toxicology, Imperial College School of Medicine at St. Mary’s, who provided the isolated hepatocytes. This work was supported by MAFF Contract 1A008. Supporting Information Available: Data relating to the characterization of the antisera (2 pages) are available as Supporting Information. Ordering information can be found on any current masthead.

References (1) Miller, J. A., Miller, E. C., and Phillips, D. H. (1982) The metabolic activation and carcinogenicity of alkenylbenzenes that occur naturally in many spices. Carcinog. Mutagens Environ. 1, 8397. (2) Miller, E. C., Swanson, A. B., Phillips, D. H., Fletcher, T. L., Liem, A., and Miller, J. A. (1983) Structure-activity studies of the carcinogenicities in the mouse and rat of some naturally occurring synthetic alkenylbenzene derivatives related to safrole and estragole. Cancer Res. 43, 1124-1134. (3) Shaver, T. N., and Bull, D. L. (1980) Environmental fate of methyleugenol. Bull. Environ. Contam. Toxicol. 24, 619. (4) Council of Europe (1981) Flavouring substances and natural sources of flavourings. 3rd ed. Strasbourg. (5) Swanson, A. B., Miller, E. C., and Miller, J. A. (1981) The sidechain epoxidation and hydroxylation of the hepatocarcinogens safrole and estragole and some related compounds by rat and mouse liver microsomes. Biochim. Biophys. Acta 673, 504-515. (6) Boberg, E. W., Miller, E. C, Miller, J. A., Poland, A., and Liem, A. (1983) Strong evidence from studies with brachymorphic mice and pentachlorophenol that 1′-sulfooxysafrole is the major ultimate electrophilic and carcinogenic metabolite of 1′-hydroxysafrole in mouse liver. Cancer Res. 43, 5163-5173. (7) Phillips, D. H., Miller, J. A., Miller, E. C., and Adams, B. (1981) Structures of the DNA adducts formed in mouse liver after administration of the proximate hepatocarcinogen 1′-hydroxyestragole. Cancer Res. 41, 176-186. (8) Phillips, D. H., Miller, J. A., Miller, E. C., and Adams, B. (1981) N2 atom of guanine and N6 atom of adenine residues as sites for covalent binding of metabolically activated 1′-hydroxysafrole to mouse liver DNA in vivo. Cancer Res. 41, 2664-2671. (9) Wislocki, P. G., Borchert, P., Miller, J. A., and Miller, E. C. (1976) The metabolic activation of the carcinogen 1′-hydroxysafrole in vivo and in vitro and the electrophilic reactivities of possible ultimate carcinogens. Cancer Res. 36, 1686-1695. (10) Wislocki, P. G., Miller, E. C., Miller, J. A., McCoy, E. C., and Rosenkranz, H. S. (1977) Carcinogenic and mutagenic activities of safrole, 1′-hydroxysafrole and some known or possible metabolites. Cancer Res. 37, 1883-1891. (11) Delaforge, M., Janiaud, P., Levi, P., and Morizot, J. P. (1980) Biotransformation of allylbenzene analogues in vivo and in vitro through the epoxide-diol pathway. Xenobiotica 10, 737-744.

Chem. Res. Toxicol., Vol. 9, No. 4, 1996 721 (12) Thompson, D. C., Teodosiu, D., Egestad, B., Mickos, H., and Moldeus, P. (1990) Formation of glutathione conjugates during oxidation of eugenol by microsomal fractions of rat liver and lung. Biochem. Pharmacol. 39, 1587-1595. (13) Thompson, D. C., Constantin, T. D., and Moldeus, P. (1991) Metabolism and cytotoxicity of eugenol in isolated rat hepatocytes. Chem.-Biol. Interact. 77, 137-147. (14) Thompson, D. C., Perera, K., Krol, E. S., and Bolton, J. (1995) o-Methoxy-4-alkylphenols that form quinone methides of intermediate reactivity are the most toxic in rat liver slices. Chem. Res. Toxicol. 8, 323-327. (15) Bolton, J. L., Acay, N. M., and Vukomanovic, V. (1994) Evidence that 4-allyl-o-quinones spontaneously rearrange to their more electrophilic quinone methides: potential bioactivation mechanism for the hepatocarcinogen safrole. Chem. Res. Toxicol. 7, 443450. (16) Chan, V. S. W., and Caldwell, J. (1992) Comparative induction of unscheduled DNA synthesis in cultured rat hepatocytes by allylbenzenes and their 1′-hydroxymetabolites. Food Chem. Toxicol. 30, 831-836. (17) Howes, A. J., Chan, V. S. W., and Caldwell, J. (1990) Structurespecificity of the genotoxicity of some naturally occurring alkenylbenzenes determined by the unscheduled DNA synthesis assay in rat hepatocytes. Food Chem. Toxicol. 28, 537-542. (18) Solheim, E. and Scheline, R. (1976) Metabolism of alkenylbenzene derivatives in the rat II eugenol and isoeugenolmethylether. Xenobiotica 6, 137-150. (19) Borchert, P., Wislocki, P. G., Miller, J. A., and Miller, E. C. (1973) The metabolism of the naturally occurring hepatocarcinogen safrole to 1′-hydroxysafrole and the electrophilic reactivity of 1′acetoxysafrole. Cancer Res. 33, 575-589. (20) Habeeb, A. F. S. A. (1966) Determination of free amino groups in proteins by trinitrobenzene sulfonic acid. Anal. Biochem. 14, 328336. (21) Kenna, J. G., Neuberger, J., and Williams, R. (1987) Identification by immunoblotting of three halothane-induced liver microsomal polypeptide antigens recognized by antibodies in sera from patients with halothane associated hepatitis. J. Pharmacol. Exp. Ther. 242, 733-740. (22) Fujiki, Y., Hubbard, A. L., Fowler, S., and Lazarow, P. B. (1982) Isolation of intracellular membranes by means of sodium carbonate treatment: application to endoplasmic reticulum. J. Cell Biol. 93, 97 -102. (23) Moldeus, P., Hogberg J., and Orrenius, S. (1978) Isolation and use of liver cells. Methods Enzymol. 52, 60-71. (24) Luo, G., Qato, M. K., and Guenther, T. M. (1992) Hydrolysis of the 2′,3′-allylic epoxides of allylbenzene, estragole, eugenol and safrole by both microsomal and cytosolic epoxide hydrolases. Drug Metab. Dispos. 20, 440-445. (25) Luo, G., and Guenther, T. M. (1994) Detoxication of the 2′,3′epoxide metabolites of allylbenzene and estragole: conjugation with glutathione. Drug Metab. Dispos. 22, 731-737. (26) Luo, G., and Guenther, T. M. (1995) Metabolism of allylbenzene 2′,3′-oxide and estragole 2′,3′-oxide in the isolated perfused rat liver. J. Pharmacol. Exp. Ther. 272, 588-596. (27) Marshall, A. D., and Caldwell, J. (1992) Influence of modulators of epoxide metabolism on the cytotoxicity of trans-anethole in freshly isolated rat hepatocytes. Food Chem. Toxicol. 30, 467473. (28) Scheline, R. (1991) Metabolism of phenols and ethers. In CRC Handbook of mammalian metabolism of plant compounds (Scheline, R., Ed.) pp 49-84, CRC Press, Boca Raton, FL. (29) Cappiello, M., Frandii, M., Giulani, L., and Pacifici, G. M. (1989) Distribution of 2-naphthol sulphotransferase and its endogenous substrate adenosine 3′-phosphate 5′-phosphosulphate in human tissues. Eur. J. Clin. Pharmacol. 37, 317-320. (30) Evans, W. H. (1992) Isolation and characterization of membranes and cell organelles. In Preparative centrifugation: a practical approach (Rickwood, D., Hames, and B. D., Eds.) pp 233-270, IRL Press, Oxford. (31) Kenna, J. G., Martin, J. L., and Pohl, L. R. (1992) The topography of trifluoroacetylated protein antigens in liver microsomal fractions from halothane treated rats. Biochem. Pharmacol. 44, 621629. (32) National Toxicology Program (1989) Subchronic Study of Methyleugenol (C60991), Contract No. NO1-ES-6-5158.

TX950211V