Immunochemical Identification of Hepatic Protein Adducts Derived

and Department of Molecular and Cellular Pathology, Ninewells Hospital and Medical School,. University of Dundee, Dundee DD1 9SY, U.K.. Received ...
0 downloads 0 Views 176KB Size
Chem. Res. Toxicol. 1998, 11, 863-872

863

Immunochemical Identification of Hepatic Protein Adducts Derived from Estragole Hiroshi Wakazono,‡ Iain Gardner, Erik Eliasson,§ Michael W. H. Coughtrie,† J. Gerald Kenna,* and John Caldwell Molecular Toxicology, Imperial College School of Medicine, Norfolk Place, London W2 1PG, U.K., and Department of Molecular and Cellular Pathology, Ninewells Hospital and Medical School, University of Dundee, Dundee DD1 9SY, U.K. Received December 10, 1997

Hepatic protein adducts derived from the allylbenzene food flavor estragole, which is hepatocarcinogenic when given to rodents at high doses, have been identified using immunochemical approaches. Male Fischer 344 rats were given estragole orally and hepatic protein adducts were detected by immunoblotting, using antisera raised by immunizing rabbits with 4-methoxycinnamic acid-modified rabbit serum albumin. A major 155-kDa adduct was expressed in livers of animals that had been treated with estragole at 100, 300, or 500 mg/kg. Levels of expression of the adduct increased disproportionately with respect to dose, and other adducts (170, 100, 44, and 35 kDa) were detected also in the high-dose group. Rats given estragole for 5 days, at 300 mg/kg/day, expressed predominantly 155- and 44-kDa adducts. The 155-, 100-, 44-, and 35-kDa adducts were detected in greatest abundance in liver microsomal fractions, while the 170-kDa adduct was most abundant in the nuclear fraction. Interestingly, whereas the 170-, 155-, 100-, and 35-kDa adducts were detected in cytosolic fractions, relatively low levels of the 44-kDa adduct were detected in nuclear fractions but not in cytosolic fractions. The various adducts were solubilized when microsomal fractions were extracted with sodium carbonate and were digested by trypsin. This implies that the target proteins are peripheral membrane proteins bound to the outer surface of microsomal membranes. Experiments undertaken with isolated rat hepatocytes and with V79 cells transfected with human monoamine phenol sulfotransferase cDNA revealed that adduct formation required 1′-hydroxylation of estragole, followed by sulfation. The pattern of adducts expressed when the transfected V79 cells were incubated with 1′-hydroxyestragole was very similar to that expressed in livers of estragole-treated rats. These cells should constitute a valuable in vitro model system for investigation of toxicological consequences arising from estragole-induced protein adduct formation.

Introduction The substituted anisole derivative estragole (1-allyl4-methoxybenzene) is present in the volatile oils of plants such as fennel, Chinese star anise, basil, and tarragon and is also used as an added flavoring agent in a variety of baked foods, sweets, and alcoholic and nonalcoholic beverages (1). In addition, estragole has been used in traditional medicines for thousands of years. Consequently, estragole is ingested in the human diet at doses estimated in the United States to be approximately 70 µg/day (2). This compound is of toxicological concern because, in common with the structurally related allylbenzenes safrole and methyleugenol, it has been shown to be hepatotoxic and hepatocarcinogenic when given to rodents at high doses (exceeding 500 mg/kg/day) (3, 4). * To whom correspondence should be addressed, at: Zeneca Central Toxicology Laboratory, Alderley Park, Maccclesfield, Cheshire SK10 4TJ, U.K. Tel: +44-1625 818558. Fax: +44-1625 586396. E-mail: [email protected]. † University of Dundee. ‡ Present address: Fukui Research Institute, Ono Pharmaceutical Co., Ltd., 50-10 Yamagishi, Mikuni-Cho, Sakai-gun, Fukui 913, Japan. § Present address: Division of Molecular Toxicology, IMM, Karolinska Institutet, S-171 77 Stockholm, Sweden.

Estragole and other allylbenzenes undergo metabolic bioactivation in livers of mice and rats to reactive species that bind covalently to DNA (5, 6). Furthermore, estragole has been shown to elicit unscheduled DNA synthesis in isolated hepatocytes (7). These findings indicate that genotoxic mechanisms are likely to play a major role in the rodent carcinogenicity of the compound. The metabolic pathway that is responsible for formation of DNA adducts in livers of rodents given genotoxic doses of the compound is initiated by 1′-hydroxylation of the allyl side chain (Scheme 1) (3, 5). Subsequently, 1′hydroxyestragole is sulfated by cytosolic sulfotransferases to an unstable sulfate ester that hydrolyzes spontaneously, producing electrophilic carbonium ions which bind covalently to DNA (primarily via the N2-position of guanine) (5, 6, 8). Estragole and 1′-hydroxyestragole are also metabolized at the 2′, 3′ allylic double bond to reactive epoxides (9) (Scheme 1), but these metabolites do not make a substantial contribution to DNA adduct formation in rodents (5, 6, 8). This appears to be because the epoxides are detoxified efficiently in vivo, by hepatic epoxide hydrolases and glutathione S-transferases (10). In addition to forming adducts with DNA, reactive metabolites of estragole and other carcinogenic allylben-

S0893-228x(97)00218-X CCC: $15.00 © 1998 American Chemical Society Published on Web 07/10/1998

864 Chem. Res. Toxicol., Vol. 11, No. 8, 1998

Wakazono et al.

Scheme 1. Metabolic Bioactivation of Estragolea

aAbbreviations: P450, cytochrome P450; ST, sulfotransferase. Unstable reactive metabolites, which have not been isolated and characterized, are enclosed in square brackets.

zenes have been shown to bind covalently to hepatic proteins. In fact, the levels of protein adducts detected radiochemically, in livers of mice given 1′-[3H]hydroxyestragole, were very similar to the levels of DNA adducts (5). Since protein adduct formation has been implicated in the mechanism of toxicity of other chemicals (11) and also in the organ toxicity caused by various drugs (12), it is conceivable that protein adducts contribute to the toxicity of the allylbenzenes. Recently, we have used immunochemical approaches to demonstrate dose- and time-dependent formation of a major 44-kDa microsomal protein adduct in livers of rats treated acutely or chronically with methyleugenol (13). In the present study, we have used similar methods to investigate the nature of the protein adducts expressed in livers of estragoletreated rats and to explore the role played by metabolic bioactivation in generation of the adducts.

Experimental Procedures Materials. Chemicals were obtained from the following sources: estragole (1-allyl-4-methoxybenzene), rabbit serum albumin (RSA), 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide hydrochloride (EDC),1 4-methoxycinnamic acid (MCA), Freund’s complete adjuvant, Freund’s incomplete adjuvant, Thimerosal, o-phenylenediamine, 30% hydrogen peroxide, NADP+, glucose-6-phosphate, glucose-6-phosphate dehydrogenase, sodium deoxycholate, bovine pancreatic trypsin (type 1), and soybean trypsin-chymotrypsin inhibitor from Sigma Chemical Co. (Poole, U.K.). Horseradish peroxidase-conjugated antiserum to rabbit IgG was obtained from Serotec, U.K. (code 6430). Electrophoresis chemicals, molecular weight markers, and horseradish peroxidase-conjugated protein A conjugate (blotting grade) were from Bio-Rad Laboratories Ltd., U.K. 1′Hydroxyestragole (HE) was synthesized according to Drinkwater et al. (3) and had UV, 1H NMR, and mass spectra consistent with those reported by these authors. 3,4-Dimethoxycinnamic

acid-modified RSA (DMCA-RSA) was synthesized as described previously (12). All other chemicals were of the highest grade commercially available. Caution: Estragole and HE are potential carcinogens. EDC is irritant to skin and eyes, while Thimerosal is toxic by ingestion and inhalation. These chemicals must be handled using appropriate safety precautions. Antisera to rat liver cytochrome P450 2E1 (which was a gift from Prof. M. Ingelman-Sundberg, Stockholm, Sweden) and to rat liver protein disulfide isomerase were prepared as described previously (14). Synthesis of RSA Conjugates. The methods used are illustrated schematically in Scheme 2. In the first method, 1′acetoxyestragole (AE) was synthesized as described by Drinkwater et al. (3), dissolved in acetonitrile, and incubated overnight with RSA, at room temperature with stirring. In the second method, MCA (an acidic congener of estragole) was covalently coupled to RSA by a two-step conjugation method (15). EDC (288 mg, 1.5 mmol) was dissolved in 7.5 mL of 20 mM sodium phosphate (pH 5.0), added to 2.5 mL of a solution of MCA in acetonitrile (53.5 mg, 0.3 mmol), and incubated for 2 min. The reaction mixture was added to RSA (40 mg in 20 mL of 160 mM sodium phosphate, pH 8.0) and incubated overnight at room temperature. After the reactions, the conjugates were dialyzed exhaustively against water (3 L) at 4 °C, then lyophilized, and stored at 4 °C. Estimates of the hapten densities of 1 Abbreviations: AE, 1′-acetoxyestragole; AE-RSA, protein conjugate prepared by coupling 1′-acetoxyestragole to rabbit serum albumin; DMCA, 3,4-dimethoxycinnamic acid; DMCA-RSA, 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; ELISA, enzymelinked immunosorbent assay; HE, 1′-hydroxyestragole; HRP, horseradish peroxidase; MCA, 4-methoxycinnamic acid; MCA-RSA, protein conjugate prepared by coupling 4-methoxycinnamic acid to rabbit serum albumin; MPST monoamine phenol sulfotransferase; PAPS, 3′phosphoadenosine 5′-phosphosulfate; PBS, phosphate-buffered saline (137 mM sodium chloride, 2.7 mM potassium chloride, 8 mM disodium hydrogen phosphate, 1.5 mM potassium dihydrogen phosphate, pH 7.4); RSA, rabbit serum albumin; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis.

Estragole-Protein Adducts in Rat Liver Scheme 2. Synthesis of Hapten-Modified RSA Conjugates

the conjugates were undertaken by UV spectrophotometry, assuming that the absorption spectra of the conjugated haptens were equivalent to those of the unconjugated reactants (which exhibited UV maxima at 259 nm for AE and 305 nm for MCA). These estimates indicated that the AE-RSA conjugate contained 14.6 mol of adducts/mol of RSA and the MCA-RSA conjugate contained 12.0 mol of adducts/mol of RSA. Preparation of Antisera. Two female New Zealand white rabbits (3 kg; from Froxfield Farms, U.K.) were immunized with the AE-RSA conjugate, and two rabbits were immunized with the MCA-RSA conjugate. Each animal was immunized initially into 6-8 sites on the back with an emulsion formed by dissolving the protein conjugates (250 µg) in 500 µL of sterile phosphatebuffered saline (PBS) and then vortex mixing with an equal volume of Freund’s complete adjuvant. After 4 weeks, rabbits were boosted in a similar manner with 250 µg of the protein conjugate in an emulsion prepared using Freund’s incomplete adjuvant. Three weeks later, the rabbits were bled and sera were prepared and stored at -20 °C. The immunoblotting experiments were undertaken using anti(MCA-RSA) antiserum which had been immunoadsorbed to deplete antibodies that bound to unmodified rat liver proteins. Immunoadsorption was achieved by incubation of diluted antiserum (final dilution 1:300 in PBS) with microsomes from livers of control rats (at 2.8 mg of protein/mL) for 6 h on ice, with shaking. The mixture was centrifuged at 100000g for 60 min at 4 °C; then the supernatant was stored at -80 °C in small aliquots. These were thawed and centrifuged immediately before use (for 5 min at 10000g and 4 °C) to remove aggregated antibodies. Treatment of Rats with Estragole and Preparation of Liver Subcellular Fractions. Male Fischer 344 rats (240270 g; from Harlan-Olac, Oxon, U.K.) were treated orally, by gavage, with either single (100, 300, or 500 mg/kg) or repeated (10, 100, or 300 mg/kg/day, for 5 days) doses of estragole in 0.5% (w/v) methylcellulose solution. Control animals received an equivalent volume of 0.5% methylcellulose solution. Animals had free access to food (LabSure CRM rat pellets from Special Diet Services, Witham, Essex, U.K.) and water throughout the experiments and were killed by cervical dislocation at 3, 12, or 24 h after the last dose. Livers were removed from rats immediately after sacrifice and homogenized in ice-cold buffer; then subcellular fractions were prepared by differential centrifugation at 4 °C, as described previously (13). In Vitro Incubations. Liver microsomal fractions were diluted to 2 mg of protein/mL in 50 mM Tris-HCl buffer, pH 7.4, containing 5 mM MgCl2, and preincubated with estragole [final concentration 50 and 500 µM, added as a stock solution in acetonitrile such that the final concentration of acetonitrile was 95% inhibition in the presence of 1.8 µM hapten), but not when the antiserum was preincubated with equivalent concentrations of unconjugated RSA or of a conjugate prepared by coupling 3,4dimethoxycinnamic acid to RSA (Figure 1, panel B). Unconjugated estragole was a much less potent inhibitor of antibody binding than MCA-RSA (maximal inhibition

Estragole-Protein Adducts in Rat Liver Chart 1. Structures of the Compounds Used in Antibody Inhibition Studies

observed was 38%, at 1 mM) but was a more potent inhibitor than the structurally related compounds eugenol (27% inhibition at 1 mM) and methyleugenol (8% inhibition at 1 mM) (Figure 1, panel B; see Chart 1 for chemical structures). Detection of Protein Adducts in Livers from Estragole-Treated Rats. Protein adducts expressed in livers of rats given estragole orally were identified by SDS-PAGE and immunoblotting, which was undertaken using anti(MCA-RSA) antiserum. The antiserum was immunoadsorbed with microsomes from livers of control rats, to remove antibodies that bound to unmodified rat hepatic proteins, before use in these studies. Several polypeptide antigens were detected which were expressed in livers from rats sacrificed after administration of single or multiple doses of estragole, but not in livers from untreated rats (Figures 2 and 3). The most abundant antigen detected in livers of rats given a single dose of estragole exhibited a molecular mass of 155 kDa, and expression of this antigen increased linearly with time up to 24 h in animals given the compound at 500 mg/kg (Figure 2). In addition, appreciable levels of several other antigens (most notably 170-, 100-, 44-, and 35-kDa adducts) were detected in livers of rats given this single dose of estragole. Both 44- and 155-kDa antigens were detected in livers of rats given estragole for 5 days at 300 mg/kg/day, as were relatively low levels of 100- and 35-kDa antigens (Figure 3). Levels of expression of the various antigens increased disproportionately with dose and were roughly 250-fold higher in livers of rats given a single dose of estragole at 500 mg/kg than in livers from animals given the compound at 100 mg/kg (Figure 2). To confirm that the various estragole-induced polypeptide antigens comprised protein adducts derived from the compound, a series of immunoblot inhibition studies was

Chem. Res. Toxicol., Vol. 11, No. 8, 1998 867

undertaken. Compounds which inhibited antibody binding, with varying potencies, were MCA, anethole, 4-hydroxycinnamic acid, 1′-hydroxyestragole, estragole, and acetaminophen (Table 1). All of these compounds contain aromatic rings substituted with allyl or other side chains para to hydroxyl or methoxy groups (see Chart 1 for chemical structures). In contrast, antibody binding was inhibited only very poorly by methyleugenol (which contains two adjacent bulky methoxy substituents that might sterically hinder recognition of the aromatic ring) and by allylbenzene (which contains no ring substituent other than the allyl side chain) (Table 1). Subcellular Distribution and Membrane Topography of Estragole Adducts Expressed in Rat Liver in Vivo. To investigate the subcellular distributions of the various protein adducts, subcellular fractions which had been prepared by differential centrifugation of liver homogenates were analyzed by immunoblotting. The 155- and 44-kDa adducts were present at greatest abundance in the microsomal fraction, as were the 100and 35-kDa adducts (Figures 2 and 3). In addition, the 155-, 100-, and 35-kDa adducts were detected in the cytosolic fraction at levels that were estimated, by densitometry, to be 20-30% of the levels expressed in the particulate fractions, while low levels of the 44-kDa adduct were detected in the nuclear fraction but not in the cytosolic fraction (Figures 2 and 3). The 170-kDa adduct was detected primarily in the nuclear fraction and in the cytosolic fraction and was only detected in the microsomal fraction at very low levels (Figure 2). To investigate the mode of association of the adducts with microsomal vesicles, the effect on the adducts of extraction of microsomal fractions with sodium carbonate and digestion with trypsin was investigated. Extraction with sodium carbonate has been shown to solubilize peripheral membrane-associated proteins, but not integral membrane proteins (18). Digestion with trypsin reveals whether proteins are exposed on the outer surface of the lipid bilayer or sequestered within membrane vesicles (19,20). The various estragole adducts were solubilized by sodium carbonate (Figure 4, panel A) and were degraded by trypsin (Figure 4, panel B). Control studies were undertaken using antibodies to two proteins which have well-defined topographies with respect to the microsomal membrane, namely, cytochrome P450 2E1 [an integral membrane protein of the endoplasmic reticulum (14)] and protein disulfide isomerase [a protein derived from the lumen of the endoplasmic reticulum, which resides within microsomal vesicles (21)]. Cytochrome P450 2E1 was not extracted by sodium carbonate, while protein disulfide isomerase was resistant to digestion by trypsin unless microsomal vesicles were permeabilized by addition of the detergent sodium deoxycholate (results not shown). Generation of Protein Adducts Derived from Estragole in Vitro. Novel protein adducts were not detected by immunoblotting in subcellular fractions prepared from isolated rat hepatocytes that had been incubated for 3 h with 50 or 500 µM estragole in vitro (Figure 5). Furthermore, no protein adducts were generated when liver microsomal fractions from untreated rats were incubated with 50 or 500 µM estragole in the presence of an NADPH-generating system, for up to 1 h, or when postmitochondrial supernatants that had been supplemented with cofactors for cytochromes P450 and sulfotransferases were incubated for up to 2 h with 50,

868 Chem. Res. Toxicol., Vol. 11, No. 8, 1998

Wakazono et al.

Figure 2. Detection by immunoblotting of hepatic protein adducts expressed in male Fischer 344 rats sacrificed after oral administration of single doses of estragole. Immunoblots were developed using anti(MCA-RSA) antiserum. Panel A: Subcellular fractions from livers of rats given various doses of estragole and then sacrificed after 2, 12, or 24 h were probed using anti(MCARSA) antiserum. The molecular masses of the major adducts are indicated. Panel B: Densitometric quantitation of time- and dosedependent expression of the 155-kDa liver microsomal adduct. The time course was investigated in livers of rats given estragole at 500 mg/kg, and the dose-response was studied in animals sacrificed after 24 h.

Figure 3. Detection by immunoblotting of hepatic protein adducts expressed in male Fischer 344 rats sacrificed after daily oral administration of estragole, for 5 days.

250, or 500 µM estragole (results not shown). However, many protein adducts were generated when isolated rat hepatocytes were incubated for 3 h with 500 µM 1′hydroxyestragole (Figure 5). These adducts ranged in molecular mass between 35 and >200 kDa and were detected in each of the hepatocyte subcellular fractions (combined nuclear/mitochondrial, microsomal, and cytosolic). It is notable that the pattern of adducts expressed in the hepatocyte subcellular fractions differed markedly from the pattern of adducts expressed in vivo in livers of

estragole-treated rats. In particular, a major 155-kDa adduct was not detected in the hepatocyte subcellular fractions, and although a 44-kDa adduct was generated in the hepatocytes, it was present in the cytosolic fraction, not the microsomal fraction (compare Figure 5 with Figures 2 and 3). Protein adducts that were recognized by anti(MCARSA) antiserum were also expressed when 1′-hydroxyestragole was incubated in vitro with a V79 cell line that stably expressed a transfected human phenol sulfotrans-

Estragole-Protein Adducts in Rat Liver

Chem. Res. Toxicol., Vol. 11, No. 8, 1998 869

Table 1. Summary of Immunoblot Inhibition Studiesa % inhibition of recognition of 155-kDa adduct inhibitor

33.3 µM inhibitor

333 µM inhibitor

4-methoxycinnamic acid anethole 4-hydroxycinnamic acid 1′-hydroxyestragole estragole acetaminophen methyleugenol allylbenzene

100.0 71.3 21.0 31.0 17.3 0 15.2 20.0

100.0 100.0 97.8 91.4 77.8 85.3 10.6 18.3

a Microsomal fractions from livers of rats given estragole orally for 5 days, at 500 mg/kg/day, were resolved by SDS-PAGE and transferred electrophoretically to sheets of nitrocellulose. These were incubated with anti(MCA-RSA) antiserum (1:9000 dilution) that had been preincubated for 30 min, in the presence or absence of the various test compounds at the final concentrations indicated. Quantitation was achieved by densitometric scanning of ECL film obtained after immunodevelopment of the nitrocellulose. Each value is the mean of a duplicate determination.

Figure 5. Expression of estragole adducts in vitro in isolated rat hepatocytes. Hepatocytes were incubated for 3 h at 37 °C with either estragole or 1′-hydroxyestragole, at either 5 M or 500 µM. The cells were disrupted by freeze-thawing and sonication, subcellular fractions were prepared by differential centrifugation, and these were analyzed by immunoblotting using anti(MCA-RSA) antiserum.

Figure 4. Solubility in 0.1 M sodium carbonate and accessibility to proteolytic digestion of protein adducts expressed in liver microsomes from rats sacrificed 24 h after oral administration of estragole, at 500 mg/kg. Immunoblots were developed using anti(MCA-RSA) antiserum. Panel A: Microsomes from estragole-treated (E) and control (C) rats were incubated with sodium carbonate and then centrifuged to yield solubilized (Extract) and unsolubilized (Residue) proteins. Panel B: Microsomes from setragole-treated rats were incubated for the indicated times, at 4 °C, in the presence or absense of trypsin.

ferase isozyme (MPST-V79). The pattern of protein adducts expressed in these cells was very similar to the pattern of adducts expressed in vivo, in livers of estragole-treated rats (Figure 6). Adducts derived from estragole were not expressed when the nontransfected parental V79 cell line was incubated with equivalent concentrations of 1′-hydroxyestragole (results not shown).

Discussion Our studies have shown that antisera which recognize hepatic protein adducts derived from estragole can be generated by immunizing rabbits with a synthetic MCARSA conjugate. The adducts recognized by these antisera were expressed in a dose-dependent manner in livers of rats given estragole, but not in livers of control rats (Figures 2 and 3), and recognition of the adducts could be inhibited by preincubation of the antisera with es-

Figure 6. Expression of estragole adducts in vitro in MPSTV79 cells. The cells were incubated for 24 h with various concentrations of 1′-hydroxyestragole; then combined microsomal/cytosolic fractions were prepared and analyzed by immunoblotting, using anti(MCA-RSA) antiserum. For comparative purposes, microsomes from rats sacrificed 18 h after oral administration of estragole at 500 mg/kg, or vehicle control, were analyzed also (in vivo).

tragole and various other structurally related compounds (MCA, anethole, 4-hydroxycinnamic acid, and 1′-hydroxyestragole). Marked (albeit less potent) inhibition of antibody binding to the adducts was even observed in the presence of 333 µM acetaminophen, whereas equivalent concentrations of allylbenzene and methyleugenol inhibited antibody binding very poorly (Table 1). Unlike allylbenzene, the various inhibitory compounds contain either hydroxy or methoxy substituents on the aromatic ring (see Chart 1 for structures). This implies that a polar substituent on the aromatic ring is required for efficient antibody binding. The very poor inhibition of antibody binding observed in the presence of methyleugenol could be a consequence of steric hindrance,

870 Chem. Res. Toxicol., Vol. 11, No. 8, 1998

arising from the presence of the two methoxy groups on the aromatic ring. It is unclear why rabbits immunized with a synthetic AE-RSA conjugate exhibited markedly lower titers of adduct-specific antibodies than the rabbits immunized with the MCA conjugate, especially since the two conjugates had very similar deduced hapten densities (14.6 and 12.0 mol hapten/mol of RSA, respectively). We found previously that higher titers of adduct-specific antibodies were generated in rabbits immunized with DMCA-RSA than in rabbits immunized with 1′-(acetoxymethyl)eugenol-modified RSA (13). These latter studies were undertaken as part of an immunochemical investigation of hepatic protein adducts derived from methyleugenol. Perhaps the different immunogenicities of the various conjugates are a consequence of differences in the degree of exposure of haptenic groups on the surface of the conjugates. The novel epitopes expressed on the MCARSA and DMCA-RSA conjugates may be more accessible to antibodies than the novel epitopes expressed on the AE-RSA and 1′-(acetoxymethyl)eugenol-RSA conjugates. Several estragole adducts which were expressed in a dose-dependent manner in livers of rats given estragole orally were identified by immunoblotting. The studies undertaken with livers from rats given single doses of estragole revealed that a 155-kDa adduct was by far the most abundant adduct expressed in animals given estragole at doses of 100 or 300 mg/kg and was also the major adduct expressed in animals given estragole at 500 mg/kg (Figure 2). When considered alongside the disproportionate relationship between dose of estragole and levels of adducts generated (Figure 2), this implies that the 155-kDa polypeptide has a markedly greater intrinsic reactivity with reactive metabolites of estragole than the other target proteins. Interestingly, in livers of rats treated with estragole at 300 mg/kg/day for 5 days, levels of expression of the 44-kDa adduct were very similar to levels of expression of the 155-kDa adduct. Perhaps this occurs because the rate of turnover of the 155-kDa adduct is markedly greater than the rate of turnover of the 44kDa adduct. The investigations of the subcellular distributions of the estragole adducts revealed that the 155-, 100-, 44-, and 35-kDa adducts were expressed in greatest abundance in the microsomal fraction. It should be noted that, although the microsomal fraction primarily contains sealed vesicles derived from the lumen of the endoplasmic reticulum, it also contains vesicles derived from various other organelles, including the Golgi apparatus, endosomal vesicles, and vesicles derived from the basolateral domain of the plasma membrane (22). Therefore, the precise organellar location of the estragole adducts remains to be determined. Since the 155-, 100-, 44-, and 35-kDa adducts were solubilized when microsomal fractions were extracted with sodium carbonate (18) and were degraded by trypsin (19, 20), it appears that the target proteins comprise peripheral membrane proteins exposed on the outer surface of microsomal vesicles. The appreciable levels of the 155-, 100-, and 35-kDa adducts that were detected in the cytosolic fraction (Figures 2 and 3) raise the possibility that the target proteins equilibrate between membrane-bound and soluble forms. However, it is conceivable that adduct formation perturbed the degree of membrane association of the target proteins and that one or more of them are exclusively membranebound or cytosolic unless adduct-modified. It is highly

Wakazono et al.

unlikely that the results obtained can be attributed to cross-contamination between microsomal and cytosolic components. This is because previous marker enzyme studies have verified that our fractionation scheme results in the expected distribution of subcellular components (23, 24). Clearly the subcellular location of the 170-kDa estragole adduct, which was detected in greatest abundance in nuclear and cytosolic fractions and was barely detectable in the microsomal fraction, differs from that of the other adducts. Whether the 170-kDa adduct is partly associated with the nucleus, or with the apical domain of the hepatocyte plasma membrane [which also sediments in the nuclear fraction, as prepared using our differential centrifugation strategy (24], remains to be determined. Recently we have demonstrated, by use of immunochemical approaches, that a single major 44-kDa hepatic protein adduct is expressed in livers from rats given methyleugenol (13). In common with the 44-kDa estragole adduct identified in the present studies, the 44kDa methyleugenol adduct was found to be concentrated in the microsomal fraction, to be detectable also in the nuclear fraction but not in the cytosolic fraction, and to be solubilized when microsomal fractions were extracted with sodium carbonate (13). In view of these findings, it is highly likely that the same 44-kDa target protein is covalently modified by reactive metabolites derived from these two structurally similar compounds. Perhaps the 44-kDa polypeptide will prove to be a common protein target that is modified covalently by reactive metabolites derived from other allylbenzenes. The studies undertaken in vitro, with isolated rat hepatocytes, liver subcellular fractions, and V79 cells transfected with human sulfotransferase isozyme MPST (17), have clearly implicated 1′-hydroxylation of estragole, followed by enzymic sulfation (see Scheme 1), in the mechanism of protein adduct formation in livers of rats treated with the compound. This is because adduct formation was observed when isolated rat hepatocytes and MPST-V79 cells were incubated with 1′-hydroxyestragole, and the pattern of adducts generated in the transfected V79 cells was remarkably similar to the pattern of adducts detected in livers from rats given a single dose of estragole at 500 mg/kg (Figures 5 and 6). Previous investigations have demonstrated that both 1′hydroxylation and sulfation are also responsible for generation of DNA adducts in livers of mice treated with estragole and other allylbenzenes (5, 6, 8), for induction of unscheduled DNA synthesis in isolated rat hepatocytes (7), and for formation of hepatic protein adducts in rats treated with methyleugenol (13). The lack of adduct formation that was observed when liver subcellular fractions were incubated with estragole, in the presence of an NADPH-generating system (in order to support cytochrome P450-mediated reactions but not sulfation), established that the reactive side-chain epoxides formed by cytochrome P450-mediated metabolism of estragole (9) (see Scheme 1) do not play a significant role in formation of the adducts detected by our antibodies. These latter experiments also exclude the possibility that formation of the adducts that were detected immunochemically might have involved additional potential routes of bioactivation, which are not illustrated in Scheme 1, namely, O-demethylation of estragole (25) followed by further

Estragole-Protein Adducts in Rat Liver

enzymic processing of the resulting 4-allylphenol or ring hydroxylation. The sulfotransferases that catalyze sulfation of xenobiotics are homo- or heterodimeric cytosolic proteins, which contain monomeric subunits that exhibit molecular masses of between 30 and 36 kDa (26-29). Consequently, the 35-kDa estragole adduct might comprise a reactive-metabolite-modified form of one or more sulfotransferase subunits (which then become partly membrane-associated, as discussed previously). However, the molecular masses of the other major estragole adducts (170, 155, 100, and 44 kDa) are inconsistent with the known molecular masses of these sulfotransferases (and of other drug-metabolizing enzymes, including cytochromes P450). It is unclear why protein adducts were not formed when isolated hepatocytes were incubated with estragole and why the pattern of adducts formed when hepatocytes were incubated with 1′-hydroxyestragole was so different from the pattern of adducts expressed in the MPST-V79 cells and in livers of rats. Although it is possible that the lack of adduct formation in hepatocytes was a consequence of loss of cellular viability, this is unlikely since previous studies have shown that the viability of rat hepatocytes that have been incubated with estragole at the highest concentration used in the present studies (500 µM) is approximately 80% (7, 30). Isolation of hepatocytes is known to disrupt intercellular contacts and cell polarity, a good example being the loss of zonal localization of bile canalicular plasma membrane proteins and their associated intracellular cytoskeletal proteins (31, 32). In addition, the activities of certain isozymes of cytochrome P450, and of other drug-metabolizing enzymes, are preserved poorly following hepatocyte isolation (33). Perhaps protein adduct formation in livers of estragole-treated rats in vivo, and in the MPST-V79 cells incubated with 1′-hydroxyestragole in vitro, requires metabolic processes and/or structural features that are not preserved in suspensions of isolated hepatocytes. The data we have obtained with the MPST-V79 cells demonstrate that a human sulfotransferase isozyme can catalyze metabolic bioactivation of estragole. This particular isozyme is expressed at high levels in the gastrointestinal tract (34). However, this does not mean that the very low levels of the compound that are ingested in the human diet necessarily pose a risk to human health. It is important to note that the carcinogenicity studies that have been undertaken on this and other allylbenzenes have involved administration to rodents of very high doses of the compounds (many milligrams per kilograms per day) (3-6). In contrast, the levels of estragole ingested in the human diet are of the order of 1 µg/kg/day (2). As is evident from our dose-response studies (Figure 2), administration of high doses of the compound to rodents results in a disproportionate increase in bioactivation and adduct formation. This can be attributed to dose-dependent variability in metabolism of estragole, which has been documented in rats and mice given single doses of the compound (25). O-Demethylation, which is a major pathway of detoxication of estragole, was shown to predominate in animals given the compound at doses of 0.05, 5, and 50 mg/kg, but not in animals given doses of 500 or 1000 mg/kg (25). Conversely, the extent of 1′-hydroxylation of estragole, which we have implicated in protein adduct formation, increased markedly in animals given the higher doses of

Chem. Res. Toxicol., Vol. 11, No. 8, 1998 871

the compound (25). Nonlinear dose-response relationships have also been observed in studies of the genotoxicity of estragole and other allylbenzenes in isolated rat hepatocytes in vitro, which were undertaken using the technique of unscheduled DNA synthesis (7, 30). An important goal of our future studies will be purification and characterization of the novel protein adducts derived from estragole and from methyleugenol (13). If the nature of the target proteins can be defined, it may be possible to understand why they are modified selectively in vivo by reactive metabolites of the compounds. Characterization of the target proteins may also cast light on the relevance of the protein adducts to the rodent carcinogenicity of the allylbenzenes, as may functional studies undertaken with the MPST-V79 cells (which, when incubated with 1′-hydroxyestragole in vitro, express a pattern of protein adducts similar to that expressed in rats given estragole in vivo). If adduct formation were to interfere with the normal cellular functions of the target proteins, this could be an important epigenetic mechanism of tumorigenesis.

Acknowledgment. H. Wakazono was supported by an Advanced Training Fellowship from Ono Pharmaceutical Co., Japan. This work was supported by grants from the Ministry of Agriculture, Fisheries and Food, U.K. (MAFF Contract 1A008) to J.G.K. and J.C. and the Commission of the European Communities (BMH1-CT920097) to M.W.H.C. E. Eliasson was supported by a Wellcome Trust Fellowship and by grant awards from the Swedish MRC and M. Bergvalls Stiftelse. We thank Dr. Mary Bollard (Molecular Toxicology Section, Imperial College School of Medicine) for the gift of rat isolated hepatocytes.

References (1) Leung, A. Y. (1980) Encyclopaedia of Common Natural Ingredients Used in Food, Drugs and Cosmetics, John Wiley & Sons, New York. (2) Flavor and Extract Manufacturer’s Association. (1978) Scientific Literature Review of Anisole and Derivatives in Flavor Usage, FEMA, Washington, DC. (3) Drinkwater, N. R., Miller, E. C., Miller, J. A., and Pilot, H. C. (1976) Hepatocarcinogenecity of estragole (1-allyl-4-methoxybenzene) and 1′- hydroxyestragole in the mouse and mutagenicity of 1′-acetoxyestragole in bacteria. J. Natl. Cancer Inst. 57, 13231331. (4) 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 and synthetic alkenylbenzene derivatives related to safrole and estragole. Cancer Res. 43, 1124-1134. (5) 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. (6) Randerath, K., Haglund, R. E., Phillips, D. H., and Reddy, M. V. (1984) 32P-Post-labeling analysis of DNA adducts formed in the livers of animals trreated with safrole, estragole and other naturally occurring alkenylbenzenes. I. Adult female CD-1 mice. Carcinogenesis 5, 1613-1622. (7) Chan, V. S. W., and Caldwell, J. (1992) Comparative induction of unscheduled DNA synthesis in cultured rat hepatocytes by allylbenzenes and their 1′-hydroxy metabolites. Food Chem. Toxicol. 30, 831-836. (8) 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. (9) Swanson, A. B., Miller, E. C., and Miller, J. A. (1981) The sidechain epoxidation and hydroxylation of the hepatocarcinogens

872 Chem. Res. Toxicol., Vol. 11, No. 8, 1998

(10) (11) (12) (13)

(14) (15)

(16) (17)

(18)

(19) (20) (21) (22) (23)

safrole and estragole and some related compounds by rat and mouse liver microsomes. Biochim. Biophys. Acta 673, 504-515. Luo, G., and Guenthner, 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. Hinson, D. H., and Roberts, D. W. (1992) Role of covalent and noncovalent interactions in cell toxicity: effects on proteins. Annu. Rev. Pharmacol. Toxicol. 32, 471-450. Bolesterli, U. A. (1993) Specific targets of covalent drug-protein interactions in hepatocytes and their toxicological significance in drug-induced liver injury. Drug Metab. Rev. 25, 395-451. Gardner, I., Bergin, P., Stening, P., Kenna, J. G., and Caldwell, J. (1996) Immunodetection of covalently modified protein adducts in livers of rats treated with methyleugenol. Chem. Res. Toxicol. 9, 713-721. Eliasson, E., and Kenna, J. G. (1996) Cytochrome P450 2E1 is a cell surface autoantigen in halothane hepatitis. Mol. Pharmacol. 50, 573-582. Davis, M. T., and Preston, J. F. (1981) A simple modified carbodiimide method for conjugation of small-molecular-weight compounds to immunoglobulin G with minimal protein crosslinking. Anal. Biochem. 116, 402-407. Moldeus, P., Hogberg, J., and Orrenius, S. (1978). Isolation and use of liver cells. Methods Enzymol. 52, 60-71. Jones, A. L., Hagen, M., Coughtrie, M. W. H., Roberts, R. C., and Glatt, H. (1995) Human platelet phnolsulfotransferase; cDNA cloning, stable expression in V79 cells and identification of a novel allelic variant of the phenol-sulfating form. Biochem. Biophys. Res. Commun. 208, 855-862. 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. Nilsson, O. S., and Dallner, G. (1977) Enzyme and phospholipid asymmetry in liver microsomal membranes. J. Cell Biol. 72, 568583. Ito, A., and Sato, R. (1969) Proteolytic dissection of smoothsurfaced vesicles of liver microsomes. J. Cell Biol. 40, 179-189. Freedman, R. (1989) Protein disulphide isomerase: multiple roles in the modification of nascent secretory proteins. Cell 57, 10691072. DePierre, J. W., and Ernster, L. (1977) Enzyme topology of intracellular membranes. Annu. Rev. Biochem. 26, 1201-1262. 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 hepatitis. J. Pharmacol. Exp. Ther. 242, 733-741.

Wakazono et al. (24) Wade, L. T., Kenna, J. G., and Caldwell, J. (1997) Immunochemical identification of mouse hepatic protein adducts derived from the nonsteroidal antiinflammatory drugs diclofenac, sulindac and ibuprofen. Chem. Res. Toxicol. 10, 546-555. (25) Anthony, A., Caldwell, J., Hutt, A. J., and Smith, R. L. (1987) Metabolism of estragole in rat and mouse and influence of dose size on excretion of the proximate carcinogen 1′-hydroxyestragole. Food Chem. Toxicol. 25, 799-806. (26) Falany, C. N. (1991) Molecular enzymology of human liver cytosolic sulfotransferases. Trends Pharmacol. Sci. 12, 255-259. (27) Weinshilboum, R. M., and Otterness, D. M. (1994) Sulfotransferase enzymes. In Conjugation-Deconjugation Reactions in Drug Metabolism and Toxicity (Kauffman, F. C., Ed.) Chapter 22, Vol. 112, pp 45-78, Springer-Verlag, Berlin, Heidelberg. (28) Kiehlbauch, C. C., Lam, Y. F., and Ringer, D. P. (1995) Homodimeric and heterodimeric aryl sulfotransferases catalyze the sulfuric acid esterification of N-hydroxy-2-acetylaminofluorene. J. Biol. Chem. 270, 18941-18947. (29) Weinshilboum, R. M., Otterness, D. M., Aksoy, I. A., Wood, T. C., Her, C., and Raftogianis, R. B. (1997) Sulfotransferase molecular biology: cDNAs and genes. FASEB J. 11, 3-14. (30) 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. (31) Oude Elferinck, R. P. J., Meijer, D. K. F., Kuipers, F., Jansen, P. L. M., Groen, A. K., and Groothuis, G. M. M. (1995) Hepatobiliary secretion of organic compounds; molecular mechanisms of membrane transport. Biochim. Biophys. Acta 1241, 215-268. (32) Gautam, A., Ng, O.-C., and Boyer, J. L. (1987) Isolated rat hepatocyte couplets in short-term culture: structural characteristics and plasma membrane reorganization. Hepatology 7, 216223. (33) Guillouzo, A. (1986) Use of isolated and cultured hepatocytes for xenobiotic metabolism and cytotoxicity studies. In Isolated and cultured hepatocytes (Guillouzo, A., and Guguen-Guillouzo, C., Eds.) Chapter 14, pp 313-332, John Libbey Eurotext Ltd./ INSERM, Paris. (34) Rubin, G. L., Sharp, S., Jones, A. L., Glatt, H., Mills, J. A., and Coughtrie, M. W. (1997) Design, production and characterization of antibodies discriminating between the phenol- and monoaminesulphating forms of human phenol sufltransferase. Xenobiotica 26, 1113-1119.

TX9702188