Document OzwxXywo0m7O9KpZexvE4LJVw

SYMPOSIUM Design of Carcinogenicity Studies: Considerations in Pathology interpretation. vIVt/ fi TOXICOLOGIC Pa.thcm.ocv Vof. 11, No. 2, J983 ^Non-Reversibility of Vinyl Chloride ^ Carcinogenesis in Rodents C.Jagdish Bkandari SterJing-Winthrop Research-Institute, Rensselaer, New-York 12144 Abstract A review of the data obtained from various studies on carcinogenicity of vinyl chloride (VC) in rodents, particularly on the effect of dose, age, duration of exposure and potential reversibility of lesions, revealed that vinyl chlorideinduced carcinogenicity in rodents was dose and time related; no recovery occurred in mice even after only 1 month of VC exposures or~m rats after 6rflOMh exposures. In Addition, younger animals (2 months old) were msA sdscepiibte to viMnduced carcinogenicity than animals held for 6 or 12 months prior to exposure. Initial 6 or 12 month exposures were adequate to detect the carcinogenic potential of VC. The above information was used as a basis for discussion on design of carcinogenicity studies. Possibility of determining the carcinogenic potential of a compound in a shorter period than the traditional 2 year studies in rodents was discussed in consideration with appropriate doses, species, age and expo sure duration. Although this approach may be applicable to a strong carcinogen, it was not considered practicable in case of weak or unknown carcinogens. Y *j. r- ^ Introduction The data I am going to present today comes from three different studies. Two of these studies in which I was involved, were con ducted at Mid-West Research Institute, (1, 2, 3) for the National Institute of Environmental Health Sciences (NIEHS) and the third study was conducted by NIEHS at Research Trian gle Park in North Carolina by Drew et aJ. (4) in 1983. The objective of this presentation is not really to report the data on the carcino genicity of vinyl chloride, which is well pub lished, but the primary objective is to review the data obtained from these studies, from the point of view of study design. As we have seen, from the papers presented here yester day and also from this morning, the current carcinogenicity bioassay model is the con ventional two year studies in two species of " Presented et the Second International Symposium of the Society of Toxicology Pathologists. Session III: Morphologic Considerations in Protocol Design, Msy 9-11,1963, Arlington, Virginia. This Symposium section Is continued from Volume n. Number 1.1983. rodents. Can we now think of a different design in which we can do these same car cinogenicity studies to-determine the carci nogenic potential of the reference compound hi a shorter period, possibly-six-months vrr a ~yeaT? Other considerations include selection of appropriate doses, species, appropriate age, and duration of exposure. I would hope that we can look at these data from this point of view and see if we can have some discussion on these areas and how we can best utilize this information. Vinyl chloride is a very simple chemical (H2C = CHC1). It is a monomer used mainly to make polyvinyl chloride (PVC) which is an essential material in our daily life. Interest in the carcinogenicity of the vinyl chloride was raised by the reports of hepatic angio sarcomas in workers associated with PVC manufacturing (5, 6). Vinyl chloride, like many other carcinogens, is not carcinogenic without metabolic activation. It is activated by the cytochrome P-450 system into an epoxide (chloroethylene oxide) which then can alkylate the macromolecules of cells and leads to neoplastic transformation (7). AP00013501 182 BHANDAR1 Toxicologic Pathology First Study In the /irst study (1, 2). GDniiw and SfttflgwPwiey ^ats, were used. The age at FIG. 1--Mortality in mice (top) and in rats (bot tom)exposed to vinyl chloride for 12 months. initiation of VC exposure was two months. Thirty-six animals of each sex were used in each of controller), 8gQ*ad-40QQ parts per million (ppm) groups. The exposure was by inhalation, six hours per day, five days per week. The housing was four to six mice per cage, and two rats per cage. They were fed pulverized or pelleted commercial diet at all times except during the exposure. Water was available at all times. Inhalation was done in cubic stainless steel chambers with a volume of 3.5 m3. Vinyl chloride gas 99.8% purity was obtained from the Matheson Products, New Jersey. The controls were exposed to the fil tered air in similar chambers. The param eters were clinical observation, hematology and blood chemistry, cytogenicity, gross and microscopic pathology and other effects. For the purposes of this presentation we will con centrate on clinical observations, particularly the mortality or survival and the pathology. Four animals of each sex per group were terminated atwnqt two<4hree.,,eLM, etmBTind at the final termination at42<ownths. General Observations. In rats, clinical signs observed after five months were lethargy, roughened hair coat, anorexia, rapid weight loss, distended abdomens and death. The mortality in mice exposed to VC for 12 months is shown in Fig. 1. Significant dose- related mortality occurred after 6 months. All high dose mice died by the ninth month. Many deaths occurred during 10-12 months in the mid and low dose groups. Only 2 male FIG. 2--Incidence of bronchiolo-alveolar adenoma in mice exposed to vinyl chloride (/eft); 72 animals per group. At center, the incidence of hemangiosarcoma in mice. Graphs at center and right depict similar groups and treatments as in the graph at /eft. (Solid tine: liver; dashed tine: other organs). At right, mammary gland tumor incidence (solid tine) and lung metastasis (dashed line). ( = 1000 ppm; A = 250 ppm; O * 50 ppm; = Oppm). AP00013502 Vol. 11, No. 2. 1983 VINYL CHLORIDE CARCINOGENESIS 183 mice died in the control group which were attributed to injuries from fighting, females TTwirp susceptible to the carcinogenicity of VC. There were no significant toxic effects other than carcinogenic effects in this study, so, all the mortality was, in fact, related to neoplasia. Rats were relatively resistant to the effects of VC. Most of the deaths in the high and mid dose groups occurred during 8-12 months. No mortality occurred in controls or low dose males. Females were somewhat more suscep tible than males (Fig. 1). Pathology. The primary neoplasms in duced by VC in mice were bronchiolo-alveo lar adenoma, hemangiosarcoma of the liver, and mammary gland carcinoma with metastases to lungs. These neoplasms occurred in a dose and time related fashion as shown in Fig. 2 and included bronchiolo-alveolar ade noma, hemangiosarcoma and mammary gland tumor and metastases to lung. In rats, hemangiosarcoma and hepatocel- MONTH 10 MONTH ae4awc>. >o z U3J o UJ a: ii_ 0 50 250 1000 0 50 250 1000 OOSE (ppm) FIG. 4--Mortality in male (top) and female (bottom) mice exposed to vinyl chloride for 1, 3 or 6 months followed by recovery for 12 months. 5 MONTH 10 MONTH 100 90 ~ 60 o & 70 &. 60 - 50 - 3 40 - 50 1tr11. ii_ 20 to 0 0 50 250 1000 0 50 250 1000 DOSE (ppm) FIG. 5--Mortality in male (fop) and female (bottom) rats exposed to vinyl chloride for 1, 3, 6 and 10 months followed by recovery for 12 months. Exposures of 1 and 3 months are not plotted Because or'no significant mortality* AP00013503 164 BHANDARI Toxicologic Pathologv Iular carcinoma with metastases to the lungs occurred in a dose and time related fashion (Fig. 3). Second Study The second (recovery) study (3) was essen tially the same design as the first one. The same strains of mice and rats were used. The exposure, all methods and parameters were similar. Eight to 28 mice of each sex were gxposedior-e period of lr^-end 6 months and four to 16 KtS of each sex per group were exposed for4i h 6 and 10 months. At the end of the exposure periods, all animals were mgmteimdwn their respective rooms for a 1 ?*monlh follow-up observation (recovery) period. All results including clinical observations, and pathologic findings were in line with the first study. The incidence of mortality in mice is summarized in Fig. 4. The incidence of mortality and early terminations increased with increased dose and exposure duration. Most of the mice exposed to VC died during the recovery period. Notice that even those that were just exposed for one month and then allowed to recover, did have a doserelated mortality. However, after 3 months exposure the mortality increased and in those that were exposed for 6 months, 100% mor- 100 90 -- 0 - TO - 0 ! Mcnth O 50 *30 IOOO 0 50 *80 1000 oose (ppm) 0 50 *50 WOO FIG. 6--Incidence of bronchiolo-alveolar adenomas in male {left) and female {right) mice, exposed to vinyl chloride for 1,3 or 6 months and followed by 12 months of recovery period. /> p> 'OL 'U \tM <c/ XL 0 96 BO COO k_d o Ktw xeo c m iweoo _d 0 MOD KBD n o mm coo jO O SOW TO FIG. 7--Hemangiosarcoma incidence in all tissues of male (/eft) and female (center) mice, exposed for 1,3 or 6 months to vinyl chloride and followed by 12 months of recovery. The graph at right displays the incidence of mammary gland carcinoma in female mice included in the center graph. AP00013504 Vol. 11, No. 2, 1983 VINYL ,ORIDE CARCINOGENESIS 185 tality occurred in mid and high^toseTinice of both sexes and low dose females. Rats were relatively resistant, however, and dose- as well as exposure-duration-re lated mortality occurred during the recovery period (Fig. 5). AfterA months of exposure, only the high dose rats showed sipuiicant mortality. however. atteT in months expb- sure, dose-related mortality occurred. Pathology. The incidence of tumors in mice was related to the dose and duration of ex posure. The incidence of various tumors is summarized in Fig. 6 and 7, including bron- chiolo-alveolar adenoma hemangiosarcoma and mammary gland carcinoma. A high in cidence of bronchiolo-alveolar adenoma oc curred in mid and high dose groups, just after TABLE I--Mean Survival of Female Animals Throughout the Study Exposure Period (Months) Fch*r-344 Hits Time (Days) Golden Syrian Hamsters B6C3F1 Mice Swiw CD-I Mice 0 703 463 780 474 4-6 682 390 316 340 0-12 634 3S5 301 347 4-16 375 342 304 321 0-24 622 347 ___ -- 4-6 682 390 316 340 6-12 703 468 480 472 12-16 668 456 695 521 16-24 708 499 ___ ___ 0-12 634 355 301 347 6-18 659 455 479 443 12-24 717 424 632 472 T *0 - FIG. 8--Graph indicating the incidence of he mangiosarcoma in tissues from male and female rats, exposed for 6 or 10 months and observed during a 12 month recovery period. Mr 4! - MONTH! O! VC EXPOtum 0 tz is 24 MONTHS ---- EXPOSURE PERIOD N54: P-544 Halt, Homtltft -- HOLDING PERIOD NS4i BSC3FI, CO-1 Swiw Mice RATS AND HAMSTERS ONLY FIG. 9--Protocol design schematic of the vinyl chloride regimen used in the third study. Exposures were 16 hours/day, 5 days/week; mice received 50 ppm, rats 100 ppm and hamsters 200 ppm of vinyl chloride. O o o- o-it o-ia a-i< -it >i-i -* i*-m ia-t< month! or vc cxroiuK FIG. 10--Histogram describing the incidence of hemangiosarcoma in tissues from Swiss (top) and B6C3F1 (bottom) mice after vinyl chloride expo sures in the third study. AP0dbt3565 186 BHANDARI Toxicologic Pathology one month exposure. In rats, the primary tumor seen was hemangiosarcoma in Liver and other tissues, which occurred in mid and high dose groups after 10 months exposure (Fig. 8). Third Study The third study (4) that was conducted at NIEHS, Research Triangle Park, North Caro lina, was designed to determine the effect of age and exposure duration on cancer induc tion. The study design is shown in Fig. 9. Female rs were exposed toJWriOO OP'200^pm of VC, respectively, for &r44?"48>r 44-mouths and allowed to live their life span thereafter. Other groups of animals were held'fort "months and then exposed for 6-or 12 months. The mean survival is shown in Table I. Mean survival was shortened with initial ex posure at early age or longer exposure dura tion. The incidence of VC induced neoplasms in various species of rodents, age and expo sure duration groups is shown in Fig. 10-12. The incidence of various neoplasms in creased with early initial exposure and longer duration. In ail three rodent species an initial 12 month exposure to VC was ade quate to detect its carcinogenic potential. Conclusions. The overall evaluation of these studies indicated that vinyl chloride- induced carcinogenicity in rodents was dose and time related. No recovery occurred in mice even after only 1-month VC exposures orin rats after 6-months exposures. Younger apimals (2 months) were more susceptible to FIG. 11--Mammary gland carcinoma incidence in Swiss (left) and B6C3F1(right) in the third study. FIG. 12--Time distribution of liver hemangio sarcoma (/e/t) and mammary gland carcinoma (right) in Fischer 344 rats exposed to vinyl chloride in the third study. the VC-induced carcinogenicity than animals held for 6 or 12 months prior to exposure, and initial 6 or 12 month exposures were adequate to detect the carcinogenic potential of the compound. References 1. Lea CC, Bhandari JC, Winston JM, House WB, Dixon RL. Woods JS (1978): Carcinogenicity ofvinyl chloride and vinylidine chloride. J Tax Cnv Health 4:15-30. 2. Lee CC. Bhandari IC. Winston JM. House WB. Pe ters PJ, Dixon RL. Woods JS (1977): inhalation tox icity of vinyl chloride and vinylidine chloride. Env Health Perspect 21:25-27. 3. Hong CB, Winston JM. Thornburg LP, Lee CC (1981): Foilow-up study on the carcinogenicity of vinyl chloride and vinylidine chloride in rats and mice: Tumor incidence and mortality subsequent to exposure. I Tox Env Health 7:909-924. 4. Drew RT, Boorman GA. Haseman )K. McConnell EE, Busey WM. Moore JA (1983): The effect of age and exposure duration on cancer induction by a known carcinogen in rats, mice and hamsters. Tox AppI Pharmacol 63:120-130. 5. Falk H. Creech JL )r, Heath CW Jr, Johnson MN, Key MM (1974): Hepatic disease among workers at a vinyl chloride polymerization plant. J Am Med Assn 230:59-63. 6. Heath CW Jr, Falk J. Creech JL Jr (1975): Character istics of cases of angiosarcoma of the liver among vinyl chloride workers in the United States. Ann NY Acad Sci 246:231-236. 7. Malaveille C, Bartsch H, Barbin A. Camus AM. Montesano R. Croisy A. Jacquingnon P (1975): Mu tagenicity of vinyl chloride, chloroethyleneoxide. chloracetaldehyde and chloroethanol. Biochem Biophys Res Commun 63:363-370. AP00013506 Vol. 11, No. 2, 1983 VINYL CHLORIDE CARCINOGENESIS 187 Discussion of the Paper DR. DIENER: You have given us quite a '`slug" of information and also your conclu sions, but I wonder if you would tell us how this relates to any possible protocols that you would devise for a drug or chemical that, let's say, is a weak carcinogen? 1 don't quite know what you're driving at. What are you advo cating? DR. BHANDARI: Maybe I'll change the compound and call it a hypolipidemic com pound (from the last presentation), which induces liver tumors in rodents. The only thing that mattered was the dose. At lower doses, it may take a year or more to produce tumors. But if that dose was, say, ten times the lower dose, which qualifies maximum tolerated dose or minimum toxic dose, be cause animals will not die because of toxicity. At this high dose if you can produce adeno mas by 6 months and, hepatocellular carci nomas by 12 months, why do we have to wait for two years? Of course it is only true if we have a positive finding. Say, if we did a study--we went with the high doses and we don't see anything by one year. Does it mean that we don't need to go anymore? There we do have a problem, and we do have to go up to two years to express the full carcinogenic potential of the test compound. On the other hand if we did have a positive result as early as six months or you were convinced at that time, that there is no sense going on to two years to prove that the compound is a carcin ogen, you can save the time and the re sources. DR. VESSELINOVITCH: Once you know the potency of a carcinogen, this is an impor tant carcinogen, but if you don't know, how can you guess the time and the dose? DR. BHANDARI: True. That brings us to my question of yesterday to Dr. Diener, where we talked about the maximum toler ated dose or the minimum toxic dose. If we can think of changing some of those criteria for MTD, such as a dose that does not effect the survival because of the toxicity for at least one year. As I mentioned in my presen tation, the animals did not die of toxicity, but the survival was affected due to carcinogenic ity. Generally, industry-wise, we try to stay as low as possible, because we don't want to have a carcinogenic compound. I'm advocat ing that if we're going to have a carcinogen, why not find out earlier? Why do we have to wait a year or two at lower dosing? AP00013507 [CANCER RESEARCH 45, 786-193, January 1985] Formation of Glutathione Conjugates by Reactive Metabolites of Vinylidene Chloride in Microsomes and Isolated Hepatocytes1 Daniel C. Llebler* Michael J. Meredith, and F. Peter Guengerich* Department* et Pharmacology (D. C. L] end Biochemistry (U. J. M,, F. P. Q.} ond Corner in Molecular Toxicology (D. C. L. M. J. M, F. P. Q.j, VantJo/biit Unh/trshy School of Medletoe, NoshviDo, Tvnotoeo 37232 ABSTRACT * Oxidation of the vinyl halide carcinogen and hepatotoxin vi nylidene chloride (VDC) by microsomal cytochrome P-450 yields 2,2-dichloroacetaJdehyde, 2-chloroacetyi chloride, 2-chtoroacetic acid, and 1,1-dlchioroethyiene oxide. The roles of these metab olites in covalent modification of proteins and reduced glutathione (GSH) were examined. 2-Chloroacetyl chloride reacted with model thiols at least t OMoW faster than did 1,1-dichloroethylene oxide and at (east 10*-fold faster than did 2,2-dlchloroacetaldehyde or 2-chioroacetic add. Microsomal covalent binding of P4C]VDC was inhibited by GSH but not by lysine, suggesting that protein thiols, rather than amino groups, are major targets. Liver microsomes catalyzed the formation of three GSH:VDC metabolite conjugates, identified as S-(2.2-dichlro-1-hydroxy)ethylglutathione, 2-(S-giutathionyl)ecetate1 and S-(2-gli)t&thionyljacetylgiutathione, a novel conjugate containing both sta ble (thioether) and labile (thioester) linkages. The fatter two conjugates also were formed In Isolated rat hepatocytes and measurable amounts of 2-(S-glutathlonyl)acetate were released into the incubation medium. Both 2-{S-glutathiony!)acetate and S-(2-g!utathionyl)acetyfgiutathione were formed wKh ["SJGSH added to the hepatic medium, indicating that reactive VDC metabolites are capable of crossing the plasma membrane to react with extracellular targets. Unlabeled S-(2-glutathionyi)acetylglutathlone underwent carbonyl substitution with added [3*S]GSH, suggesting that this conjugate may participate in modification of protein thiols. This conjugate also underwent hydrolysis with a half-life of approximately 3 hr. QSH:VDC me tabolite conjugates may serve as accessible models for labile covalent adducts formed between VDC metabolites and protein thiols. INTRODUCTION Vinyl halide monomers have afforded investigators an oppor tunity to study the consequences of procarcinogen bioactivation using relatively simple compounds. Early etudes demonstrated ' that vmyt chloride is oxidized to chtoroethylene oxide by micro somal cytochrome P-450 (1, 2, 6), and ft was inferred that concomitant production of 2-chtoroacetatdehyde was the result of epoxide rearrangement (6,11). These and other observations contributed to the widely held assumption that epoxides are the 1 Thi* wort? was supported by USPH3 Grants ES 02205 tnd ES 00267 and by Vanderbilt University Research Councfl Grant 361024. 'Recipient ol a Pharmaceutical Manufacturer'* Association Foundation Ad vanced Predoc(oral Fellowship. Present address: Department of Blochoiretry and Biophysics, Oregon Stale University, Corvallis. OR 97331. * Burroughs WeOooma Scholar In Toaiooiogy (1963 to 1966). To whom requests for rtprint* should be addressed, at Department of Biochemistry, VanderMt University School of Medicine. Nashvile. TN 37232. Receives May. 21,1984; accepted September 27.19S4. principal products of microsomal vinyl halide oxidation, which, as such, occupied a position of central importance in covalent modification of cellular macromolecules and production of more stable metabolites (3). Recent mechanistic studies in this labo ratory have challenged this view. Epoxides are not obligate Intermediates in vinyl halide biotransformation but are formed togetherwith carbonyl products (hatoacetaldehydes and haloacyi halides) via partitioning of a common catalytic intermediate (17, 20). Selectivity of individual vinyl halide metabolites for reaction with particular cellular targets has been observed. Studies with vinyl chloride Indicated that ehioroethylene oxide is primarily responsible for modification of DNA fn vivo and in vitro, while 2cNoroacetakfehyde binds primarily to protein (10, 37). The mi crosomal oxidation of VDC1 yields VDC oxide, 2-chloroacetyl chloride, 2-chioroacetic acid, and 2.2-dlchloroacetaidehyde (4, 17). which may react with a variety of cellular nucleophiles. The variable reactivity of these metabolites may be expected to Influence their target selectivity. Relatively low levels of radiolabel are bound to DNA isolated from rats given p4C]VDC, and Reitz et a/. (26) suggested that VDC may initiate a tumorigenic re sponse via Interaction with cellular components other than DNA. Alternative targets of possible importance fn this epigenetic mechanism may include cellular proteins that regiriate DNA tran scription, chromatin structure, or cellular metabolic status (19). Information concerning the disposition of VDC metabolites would therefore be required to establish their roles in the initiation ct carcinogenic or toxic lesions. Several reports have suggested that cellular thiols exert * protective influence in VDC intoxication and serve to detoxicate VDC metaboBtes (13,16,27). Accordingly, depletion of hepatic `GSH by starvation or pretreatment with diethytmaleate enhanced VDC hepatotoxicity In rats (13. 27). Animals administered (14CJVDC excrete rediolabel In the urine as AZ-ecetylcartooxymethyicy8telne, thiogJyeoHrc acid, thiodiglycolBc acid, and unidenti fied sulfur-containing metabolites (16, 16). The importance of Cellular GSH as a detoxicating nucleophile for VDC metabolites implies that protein thiols may themselves be major targets for covalent modification. This suggestion has been verified for other compounds which interact with GSH in a srnifar manner (22.31. 32). Because urinary adducts reflect significant renal as well e* hepatic biotransformation, their utility as indicators of intracellular metabolite disposition is limited. In this work, VDC metabolite* were compared on the basis of relative reactivities towards sulfhydryt compounds, relative contributions to protein covalent binding in vitro, and ability to form GSH conjugates in bepaw* `The abbraviabant used an: VDC, vtnjdldaoa chloride (l.t-dtehtorcethytarirt VDC oxide. M-tScrtoroethytane oxide: GSH. reduced glutathione: GSSG. oxidlz*^ glutathione; HPLC. high-performance Equid chromatography; FDNB. 1 -fhjoro-2.^ dfttrobenzeno; DNP. 2,4-dMtraphenyt. CANCER RESEARCH VOL. 45 JANUARY 1985 186 APOOOI3508 VOC ADDUCTS tea and microsomal systems. The formation of extracellular t* 5H conjugates was used as an index of the ability of VDC i -rtaboiites to cross hepatoeyte membranes. The data indicate it multiple VDC metabolites participate in covalent modification oroteins and GSH. iTERIALS AND METHODS Ihemiesl*. [U-,4C|VDC wa* synthesized from 1,1,2-thchloro-{U-,4C] ane (Amersham, Arlington Heights, IL; 10 mCi/mmol). The labeled ting material was diluted with unlabeled 1,1,2-triehloroethane to a l specific activity of approximately 0.2 mCi/mmo! and treated with iaDicycio(S.4.0)unde<>7*ene for 1 min at 60*. A stream of nitrogen > passed through the mixture to sweep VDC into a tube immersed in y ice:isopropyl alcohol bath. The f(AMC}VDC thus collected was of radiochemical purity as assessed by gas chromatography (Tenax, .*). ("SjMethionine (1170 Ci/mmol) and ["SJGSH (69 Ci/mmol) were i New England Nuclear (Boston, MA). DC and thiophenol were purchased from Aldrich (htfwaukee, Wl); 0 was distilled before use. Chioroaeeryf chloride was from Eastman Chester, NY) and was also distilled before use. GSH, 2-(S-cysteiacetate, and glutathione-S-trensferase were purchased from Sigma Louis, MO). This oommerciel glutathions-S-transferase preparation imposed largely of isozyme B and lesser amounts of other isozymes. $e liver alcohol dehydrogenase was purchased from Boehringerinheim (Indianapolis, IN). Fetal calf serum was from Grand Island ogical Co. (Grand Island, NY). Jichloroacetaldehyde and VDC oxide were synthesized as described jwhere (17). AH other chemicals were of the highest purity commery available. lepatoeyle and Microsomal Incubations. Rat liver and human Ever Tosomes wet prepared as described previously (9, 36). Rat hepa.ytes were prepared by collagenase perfusion (29). and viability was sessed by trypan bfue exclusion. Fischer's medium, supplemented n 15% fetal calf serum but deficient in sulfur amino acids, was the sdium used for an experiments with hepatocytes. Hepatocytes (2 to 3 x 10* cells In a volume of 1 rrri) were incubated sealed 15-ml glass scintillation vials containing Fischer's medium. VDC is added to the suspension from a 1 m stock solution in acetone to a .'al concentration of 5 mM. [*S]GSH was added to cell suspensions om a 0.4 m stock solution prepared immediately before use. Hepatovtes were Incubated for 3 hr In Fischer's medium supplemented with j 5 mM (MS]methk>nine at a specific activity of 2 Cl/mmol to label sulfur cools. The cells were then washed twice and resuspended In fresh medium containing 0.5 mM unlabeted methionine. The specific activity of the hepatocefiiriar ["SJGSH produced was 89 mCl/mmol and was deter mined from fractions collected from HPLG analyses (23). After incubation with VOC for 60 min at 37*. the cells were rapidly separated from the medium by centrifugation at 1000 x g for 30 sec, and the medium was removed for analysis of GSH adducts. The pellet wee resuspended fri 1 mi on o% HCIO4 and recentrifuged. The resulting supernatant contained intracellularOSH and GSH:VDC metabolite conjugates and was analyzed by Ion-exchange HPLC. Microsomal incubations contained 100 mM potassium phosphate. pH 7.7, an NADPH generating system consisting of 0.35 IU giucose-6phosphate dehydrogenase per ml. 10 mu glucose 6-phosphate, and 0.5 mM NADP*, and microsomal protein et a final concentration of 5 to 10 mg/mi. GSH was added from a 40 mM aqueous stock solution; this solution was prepared immediately before use. VDC was addsd from a 1 m stock solution in acetone or methanol, incubations were terminated after 60 mtn at 37s by addition of ZnSO to a final concentration of 1% (w/v) or HCIO4 to a final concentration of 3.5% (w/v) and centrifuged at 3500 rpm for 5 min. The supernatants were then neutralized with KHCO* for analysis of GSH conjugates. Assays. Depletion of GSH during aqueous incubations with VDC metabolites was monitored using S,5'-dithiobis-2,2'-dinitrobenzoic add (28). VDC metabolites were added directly to solutions containing 0.5 mM GSH in 200 mM potassium phosphate. pH 8.0. at 37". with an initial metabolite concentration of 50 mu Aliquots were then removed at various times for assay of residual thiol content. 2-Chloroacetic acid was added as its sodium salt to minimize pH changes. Leas than 5% of the GSH was oxidizsd to GSSG during these incubations. Depletion of thiophenol in CHCIj by VDC metabolites was monitored using 6.5'dithiobls-2,2-dinitrobenzoie add in aqueous acetone. Thirty >imot of each VOC metabolite were added to 30 ymol thiophenol in 6 ml CHCIj at 25*. Twenty-<il aliquots were then transferred to tubes containing 1 ml of acetone plus 200 xl of 10 mM SrS'-dithtobls-a^'-dlnttrebenzoic add In 85% acetone:l5% HjO. Twenty ill of triethyfamine were added, and the absorbance was Immedately recorded versus a blank at 490 nm. Sec ond-order rate constants for the reaction of thiophenol with VDC metabofites were determined from plots of reciprocal absorbance versus time. Covalent binding of [,4C]VDC radiolabel to mierosomal proteins was assayed by the method of WalBn et at. (35). incubations contained 3 mg microsomal protein (from untreated rats) per ml and 10 mM [,4CjVDC (0.03 *tCi) and were for 30 min at 37". Proteins were pretipitated on 2.4cm glass fiber fillers (Fisher G6). Following 3 washes each with ethanol, methanol, and acetone, the filters were counted using ACS scintillation cocktail (Amersham). Separation and Characterization of Glutathione Conjugate* of VDC. Glutathione and cysteine conjugates were analyzed using ionexchange HPLC as described by Reed ef /. (23). Aminopropyl silica was prepared from 5 pm Spherisorb sica (Rainln, Woburn, MA), and 0.5- x 30-cm columns were slurry packed as described previously (24). Super natants from ZnSOi- or HCtO-quenched microeomai incubations were neutralized with KHCO* and treated with ethanollc FDNB for 30 to 60 min before injection. Elution Solvent A contained 60% methanol and 20% HjO (v/v), and Solvent B contained 3 m sodium aoetate. pH 4.5, In 64% methanol. The column was loaded Isocratfcally at 95% Solvent A for 10 min. Conjugates were eluted during a 30-min Inear gradient from 5 to 99% Solvent B at a flow rate of 1 ml/imin. Fractions (0.5 ml) were collected and counted using ACS scintillation cocktail, while elution of DNP derivatives was monitored at 360 rm For preparative and some analytical chromatography, ammonium acetate was substituted for so dium acetate in Solvent B. Effluent fractions containing 3 GSH conjugates (Conjugates A. B, and C) were collected, and several collections were pooled. Methanol was removed from the pooled fractions in vacuo, and the fractions were then lyophtlized. The residue corresponding to Conjugate A was treated with 2 n HCl for 12 hr at 25". The solution was then extracted with 3 portions of ether; the extracts were concentrated under a stream of nitrogen and then analyzed for 22-dcHoroacetaldehyde by gas chromatography (17). A portion of the residue corresponding to conjugate B was dissolved in n HCl and heated at 110* tor 30 min. A second portion was dissolved in 6 m HCl and heated under nitrogen at 110* for 30 hr. Both samples were then neutralized with KHCOj and gently gassed with oxygen overnight to oxidize (berated thiols to disulfides. The mixtures were then analyzed by ion-exchange HPLC as described above. Residue centshing Conju gate C was hydrolyzed with 6 n HCl at 110* for 24 hr. The hydrolysate was reanalyzed for 2-(S-cysteinyl)acetate by ton-exchange HPLC. RESULTS Reaction of VDC Metabolites with Thiols. Thiols reacted with major VDC metabolites at measurable rates in aqueous and nonaqueous solutions (Table 1). In aqueous buffer at pH 8, GSH reacted with 2,2-dichloroacetaldehyde and 2-chloror.ceiate at moderate rates. Despite rapid hydrolysis, 2-ehloroacety " foride reacted with GSH at least 4 orders of magnitude faster. did 2-cMoroacetie acid or 2,2-dichloroacetaldehyde. The rate re action of GSH with 2-chloroacetyl chloride was calculated ,ing CANCER RESEARCH VOL. 45 JANUARY 1965 187 1 lI AP00013509 VOC ADDUCTS Tatttl ftwta constants for raacBon at thiols wttfi VDCmataboikaa Reaction* were started by adding VOC metabolites to solution* of the thiols. ABquot* were then analyzed tor residual thiol content as described In "Materials ana Methods.' initial YDS metaboitrihiol molar rates were 1:1 to the chci, system and 100:1 to the aqueous syitam. Pseudo-first-order rata constants for ths OSH reaction were determined ham ptora Of the logarithm of absorbance versus am*. Second-ofOsr rate oonstama tar the thiophenol reaction were determined frornplotaof reciprocal absorbanca versus tima. System Metabolite TWo- pheno):CHC1j, 25*. k (M-' irtin "') VDC oxda 2-Ctiloroacetyl chloride 2.2-DfchtofoaeetaIdehyde 2-OloroacetJc add 90 * 20* >7X10* <6 <6 * Mean S.D, from 3 individual experiments. *--, not determined. GSH.20O mM potassium phos phate, pH 8.0,37*, kow <mto") __t >3.6 X 1D* 0.024 * 0.002 0008 0.000 a hydrolysis rata constant of 4 s'1 for the acyl chloride. This value was estimated from earlier studies comparing the hy drolyses of acetyl chloride and 2-chloroacetyi chloride in aqueous acetone (33) end from the reported hydrolysis rate constant for acetyl chloride in water at 0" (7). The assumptions made were than the ratios of hydrolysis rate constants of the 2 acyl chlorides were similar in water and aqueous acetone, and that the hydrol ysis rate constant doubled with each 10s hcrease In tempera ture. It also was assumed that the phosphate buffer would not significantly affect the sofvotysie rate. VOC oxide was not studied in the aqueous system, because it can be produced only as a dilute solution in chloroform. Addition of the required volume of epoxide solution thus produced a 2-phase mixture unsuitable for the experiment, and attempts to concentrate the epoxide solu tion were unsuccessful. To compare the relative reactivities of VOC oxide and 2chloroacetyl chloride with thiols in the absence of competing hydrolysis, the reaction of each metabolite with an equimolar concentration of thiophenol in chloroform was studied. The acyt chloride reacted with thiophenol at least 800-foid faster than did the epoxide, while neither 2,2-dfchtoroacetaidehyde nor 2-chioroacetic acid depleted the thiol at a measurable rate in this system (Table 1). The slow reaction of 2-chJoroacetlc acid relative to 2-chloroacetyl chloride suggests that acyfation, rather than alkylation, is the major reaction between 2-chloroacetyl chloride and the thiol. The rate constants presented describe bimolecular reactions, although pseudo-first-order conditions were selected for the aqueous system. High concentrations of VDC metabolites were used In order to achieve rates of GSH conjugation signifi cantly in excess of that of GSH oxidation. Covalent Binding of ['*CJVDC Radiofabe! in Mieroeome*. Rat and human liver mfcrosomes catalyzed covalent binding of [14CJVDC radiolabel to microsomal proteins (Table 2). No radio activity was bound when NADPH was omitted from the incuba tion mixtures. Inclusion of 5 rriM GSH in the incubation mixtures inhibited roughly half of the binding, but 5 mM lysine inhibited the binding only slightly. These results suggest that protein thiols, rather than protein amino groups, are major targets lor covalent modification by VDC metabolites. Inhibition of covalent binding by GSH was not increased when glutathione-5-transferase (0.3 mg/ml) was added to the incubations (data not shown). Covalent Tbt*a Mtofosomai covalent landing& l **G]VOO Complete incubation mixtures contained 100 mu potassium phosphate. pH 7.7 3 rat or human ver microsomal proton per ml. 0.35 IU yeast gluttse-6phosphate dehydrogenase per mi, 10 mM glucose fkphosphet#, 0-5 mM NADP* and to mu (,4C)voc (0.03 pC<). Rat wer microsomes were from untreated erimais Covalent binding at (MC)VDC radtolabel to microsomal protein was essayed as described to 'Material* and Methods." System nmol Bourtd/mg/30 % of complete min system Complete -NADPH +5 mu GSH Alcohol dehydrogenise (0.5 mg/ml): 0.S mM NADH +Alcohol dehydrogenase (0.5 mg/ml) (bOtied):0.S mM NADH +0.5 mM NADH +5 mu lyshe Human liver 31 Human tiver BO 22 r3* 2 1 12*1 91 23 s 3 28 3 18 * 1 11.10 20.22 100 9 55 41 106 127 82 * Mean 1 S.D. from 3 individual experiments. Results of duplicate experiments. Chart 1. loo-exchange HPLC of GSH^'CiVDC metabolite conjugates from rat liver microsome* (10 mg protein permq incubatedwith("CJVDC and GSH. Samples were oertvaeized with TONS as described in 'Materials and Methods.' Analyses of iresWy prepared ncuMOcn products (?)and products after 48 hr pH 8 (2). The Identities of the 3 initial peaks m 2 are unknown. binding of [,4CJVDC radioiabel also was decreased In micro somes supplemented with 0.3 mg alcohol dehydrogenase per ml plus 0.5 mM NADH. Substitution of heat-denatured dehydrogen ase returned binding to control levels, while addition of 0.5 mM NADH alone Increased binding by 25%. Microsomes prepared from 2 human liver samples catalyzed covalent binding at levels comparable to rat microsomes. Production of GSHtVDC Metabolite Conjugates in Micro somes. Rat Sver microsomes supplemented with GSH and [14C]VDC produced 3 GSH conjugates which were analyzed as V-DNP derivatives using ion-exchange HPLC. NADPH and GSH were required lor formation of all 3 conjugates. Representation chromatograms of both rat and human Nver microsomal products are shown in Charts 1 and 2. Three GSH-.VDC metabolite con jugates were designated A, 6. and C, in order of elution. Conju gate A (23 min) was present In higher levels in the 2 human samples than in the rat samples, while the refatlve amounts of Conjugates B (36 min) and C (39 min) were similar in both rat CANCER RESEARCH VOL. 45 JANUARY 1985 185 APOOO135(0 VDC ADDUCTS art 2. toMxchange HPLC of GSH:['*C]VDC metaboil* conjugates from .n liver microsomes (10 mg protein per ml) incubated wftn [,aC]V0G and GSH. lies were derivatlzed with FDNB as described in `Materials and Methods.* .see of incubation products from Samples 80 (T) and 31 (2) tom human Bv*r nown. Table 3 Formethn of GSH:VOC moteOoHt* eon/offatot by itolattdhtpatecytea totaled rat hepatocytee were Incubated at a concentration of 2 lo 3 x 10* cels/ i Fischer's medium containing S rim VOC- After 1 hr ol incubation, cells ware dly separated from the medium by centrifugation, and GSHrVDC metabolta jugates in cells and mediumware analyzedseparately as described in 'Materials Methods.* nmol conjugate/10* oaia/60 min Jonjugata ceis* Medum* Medium: Medium 5 mw GSH6 10 mM OSH* A <0.4" <0.4* <0.4* <0.4* B 4.8 0.8" <0,3* 0.fl 3.9* C 12.5*3.0" 1.5 *1.1" T1.1* 45.7* * Hepatocytes were prefncirisated with ["Sjmethioolne to tsbel hepatocyte GSH .% described in 'Materials and Methods.* 0 Unlabeled hepatocytes were incubated fn medium containing S or 10 mu ''SjGSH (2.5 mO/mmol). c Mean (a S.O.) of 3 Individual experiments. e Values from a single experiment. * Mean of duplicate experiments. and human samples. Microsomes from Human Liver 80, which produced greater levels of covalent protein binding than microsomes from Human Liver 31 (Table 2), also produced more of each GSH conjugate than microsomes from Human Liver 31. Both the total amount of conjugates formed and the relative proportions of Individual conjugates were similar in several In cubations with rat liver microsomes and wsre unaffected by the addition of 0.3 mg of purified gkitathicne-S-transferase per ml to the incubation mixtures. Production of GSH:VDC Metabolite Conjugate* in Isolated Hepatocytes. Isolated rat hepatocytes preincubated with [,sS]metNonine to label ceiular GSH (24) produced [*S]GSH at a specific activity sufficiently high (89 mCi/mmoi) to permit de tection of GSH;VDC metabolite conjugates both within the ceils and in the surrounding medium. Detectabie levels of Conjugates B and C. but not of Conjugate A, were found in "S-prelabeled cells (Table 3). Conjugate C, the major intracellular conjugate, was found at levels approximately 3-fokj higher than Conjugate B. Conjugate C also was found in the Incubation medium at approximately 10% of intracellular levels, but Conjugates A and B were nor present fn measuratiie amounts. The presence ol measurable levels of a GSH:VDC metabolite conjugate in the hepatocyt* incubation medium indicates that conjugates may be released from the hepatocyte. Alternatively, the extracellular conjugates may be formed outside the ceils by VOC metabolites which cross cell membranes to react with extracellular GSH. In order to determine if one or both of these processes occur, unlabeled rat hepatocytes were incubated with 5 mM VDC in medium containing 10 mM |MS]GSH. Cells Incubated in this manner accumulated 0.28 nmol (MS]GSH per 10s cells during the incubations, a level which corresponds to approximately 0.5% of the total cellular GSH content and which may represent GSH leakage into a smalt fraction of nonviable cells. Thus, labeled conjugates would be formed only by VDC metabolites which cross the plasma membrane to react with extracellular GSH. The major conjugate formed extraceiluiarty was Conjugate C (Table 3), although lesser amounts of Conjugate B were also detected. Conjugate A was not formed outside the ceils at detectable levels. Conjugate B was formed In the medium at approximately 10% of Conjugate C levels, but the amount of each conjugate formed increased with increasing medium GSH content (Table 3). Chemical Characterization of GSKVDC Metabolite Conju gates. Chemical characterization of GSHiVDC metabolite con jugates was based on chromatographic properties and chemical degradation of the conjugates. All 3 conjugates required deiivatbation with FDNB to elute as shown in Chart 1. Conjugate A, which was produced in microsomal preparations but not In isolated hepatocytes. disappeared from FDNB-treated samples after standing 48 hr at pH 8. Conjugate A was collected from Ion-exchange HPLC and subjected to hydrolysis with 2 w HCI at 25 for 12 hr. Gas chromatography (electron capture) (17) of ether extracts of the hydrolysate indicated that 2.2-dichloroacetaldehyde was released during hydrolysis. Further, addition of 2,2-dichloroacetaldehyde to neutral aqueous solutions of GSH, followed by S-carboxymethylation of unreacted GSH with lodoacetate and treatment with FDNB, produced (in addition to the expected GSH and GSSG derivatives) a product with a retention time identical to Conjugate A. Both GSH and /V-acetyicysteine reacted at a slmfl&r rate with 2,2-dichloroacetalclehyde, suggesting that a thlohemlacetal rather than a Schiff base was the reaction product. 'H nuclear magnetic resonance of the Nacetylcysteine adduct in DaO was also consistent with this interpretation (S.33 5, muitipiet, 1H [CljCjj--]; 4.42 i. triplet, 1H [--CtKNHCOCHaJCOaH]; 3.32 i, muitipiet, 1H (_Qj(OH)S--]; 3.12 6, doublet, 2H [--SCtf*--}; 2.03 *, singlet, 3H [--NHCOCfcUJ). Conjugate C coeluted with authentic N-DNP2-{S-glutathionyl)acetate. Add hydrolysis of Conjugate C, prepared using microsomes, GSH, and (14CjVDC, followed by analysis of Hie hydrolysate by HPLC, indicated lhat Conjugate C was cleaved to 2-{S-cySteinyt)[14C}acetate (Chart 3). The carboxymethylated thiol is an expected VDC metabolite and may be formed by conjugation of GSH with VDC oxide, 2-chforoacetyi chloride, or 2-chloroacetate. This conjugate was stable for at least 48 hr at pH 8. Conjugate B, the major adduct formed In microsomal incuba tions, was only moderately stable at pH 6 and had disappeared completely from derivatized samples within 48 hr. Virtually all of the radioactivity initially appearing as Conjugate B eluted as CANCER RESEARCH VOL. 45 JANUARY 1985 189 AP00013511 VDC ADDUCTS Chart 3. lon-exchang* HPLC of 2<<S-cya)BlnyfH'<C]acatate ratnaad by add hydrolysis of M&on]ugata C. '*C-Corjug*t C was produced using mierasom**. GSH, and [,4C]VDC and coflaciw usng (On-exchange HPLC, following derivatfz*- ton with PDNB. The collected, derivabzed MC-Conjugate C was subjected to add hydrolysis (6 n HO, 110*. 24 nr), and me products wars analyzed using tonexchange HPLC. Arrow* ), 2, 3, and 4 merit the retention timet of M<2,4- dinitrophenyl) derivatives of cystine, 2-{S-cystfinyl)acetate, 2<S-gfutatfilo- nyDacetate, and glutathione cSsuffide, respectively. Chart 4. Substitution of (**S}GSH into untabetod Conjugate a. Untabdad Con jugate a was praparad by Incubating rat over microsoma* with VOC anti urfabeteti GSH, After precipitation of fnfcroaomaf protein and neutralization to pH 7.5. f*$)GSH (S nC0 was added. After incubation tor 90 nvn. the samples were derivatized with FDN8 and analyzed by ion-exchange HPLC as dsscnbed in Materials and Methods' Arrowa mark the retention times of Conjugates a and C. DISCUSSION Conjugate C [A/-DNP:2-($-glutathfonyf)acetate] in samples treated with FDNB and left standing at pH 8 for 48 hr (Chart 1). Lyophilized samples of Conjugate B collected from Ion-exchange HPLC were subjected to acid hydrolysis and oxidative work-up to convert liberated thiols to disulfides. Treatment of Conjugate B with n HCf at 110 for 1 hr produced 2-<S-glutathionyi)acetate and GSSG in a 2:1 molar ratio. Hydrolysis of a second sample In 6 n HQ at 110 for 24 hr cleaved Conjugate B to 2-{Sglutathionyt)acetate, cystine, and glutamate. Glycine was not retained on the ion-exchange HPLC column used for these analyses. These results suggest that Conjugate B contains 2 GSH molecules bridged by a stable (thioether) linkage and a labile (thioester) linkage. Such a bisgiutathlonyl conjugate would be expected to undergo both transesteriflcatton and hydrolysis reactions. To determine if Conjugate 5 would transacyfate ex ogenous thiols, untabeled Conjugate B was generated In microsomes incubated with VDC and unlabeled GSH. After precipita tion of microsomal protein and neutralization to pH 7.5 with KHCO:KOH, ["S]G$H was added, and the mixture was allowed to stand 15 min. The mixture was then deiivatized with FDNB and analyzed by HPLC. Although soma radiolabeled Conjugate C was detected, most of the radiolabel appearing in this region of the chromatogram eluted as Conjugate B (Chart 4). The appearance of radiofabei In Conjugate C may be attributed to Scarboxymethyiation of ["SjGSH by chtoroaeetate present In the neutralized supernatant. The presence of radiolabel in Conjugate B indicates that the ["GjGSH substituted for unlabeled GSH in the thioester. Further, when unlabeled GSH was omitted from the original microsomal incubation, ("SJGSH radiolabel did not elute as Conjugate B. In a separate experiment, "S-labeled Conjugate B was derivatlzed with FDNB. tire pH was adjusted to 7.5. and aliquots of the mixture were analyzed periodically by HPLC. Conjugate B underwent a pseudo-first-order decay with a half-life of approximately 3 hr (data not shown). Previous investigators assigned a central role to VOC oxide in the formation of covalent adducts with cellular nucleophiles (12. 16, 25). The data in Table 1 show that 2-chloroacetyl chloride was considerably more reactive toward thiols than was the epoxide or other metabolites. Whfle the epoxide was not studied in aqueous buffer, it is most likely that it would also react with GSH. These data Indicate that both the epoxide and acyl chloride may modify thiols despite rapid hydrolysis. 2-ChIoroacetic acid and 2,2-dfchloroacetaldehyde ere considerably more stable in aqueous solution, and while they react more slowly with thiols, their stability may enhance their overall contribution to covalent modification of target macromolecules in vfvo. The importance of protein thiols as major targets for covalent modification by VOC metabolites is suggested by the data pre sented in Table 2. Inhibition of binding by GSH and not by lysine would suggest that protein thiols are major targets and that protein amino groups are minor ones. The possible toxicological significance of modifying even a small population of amino targets cannot be discounted. The comparative chemical properties of thiol and amine nucleophiles may account for the predominance of thiol adducts. Protonation of amines at physiological pH would largely restrict the population of available nitrogen nucleophiles to those lodged In hydrophobic regions of proteins or mem branes. These targets may themselves be exposed to lower concentrations of VOC metabolites. The selectivity cf VDC me tabolites for thiols over other nucleophiles may reflect an overall preference of VDC metabolites for `soft* nucleophiles. Three of the VDC metabolites, 2-chloroacettc add, 2-chloroacetyi chloride (a-carbon), and VDC oxide, would most likely alkylate via en S2 mechanism and would be expected to prefer the softer nucleo philes (thiols) to the somewhat harder amines. The data in Table 2 support a role for 2,2-dichloroacetaidehyde in microsomal covalent oinding. Approximately 60% of the bind ing was inhibited by added alcohol dehydrogenase plus NADH. indiating that the aldehyde and not the corresponding alcohol CANCER RESEARCH VOL. 45 JANUARY 1985 190 T|f^^B^'taafciwea* .-J:- AP00013512 VDC ADDUCTS terms adducts. These observations also suggest that cytosolic and mitochondrial dehydrogenases capable of reducing or oxic`z>ng 2,2-dichloroacetaidehyde may afford partial protection * 'iinst cellular damage due to aldehyde production. he reactivity of VDC metabolites toward thiols is indicated by only the influence of GSH on covalent binding but also the ity of VDC metabolites to form 3 different GSH conjugates, vious workers have shown that the hepatotoxidty and co?nt binding of VDC are potentiated In vivo when hepatic GSH jls are reduced by fasting or diethylmaleate pretreatment (13, . Other studies have shown that p4C]VDC radiolabel is exed in the urine as sulfur-containing metabolites apparently 'ived from GSH conjugates (16.16). The urinary metabolites itain secondary modifications due to renal biotransformation d enterohepatic recirculation which may mask the identities of 3 initially formed products. The use of microsomes and isolated oatocytes to study the formation of GSH conjugates was ended to minimize such secondary modifications. Conjugate which can be synthesized by adding 2,2-dichloroacetaidehyde neutral aqueous solutions of GSH, exhibits properties ex acted of S-(2,2-dich!oro-1 -hydroxy)ethylglutathione, a thlohemcetal of the aldehyde and GSH. The assignment of a thiohem;cstal structure, rather than a Schiff base (with the a-amino roup of the 7-glutamyi residue of GSH) is based on several onsiderations. Both GSH and /v-acetylcysteine (which cannot crm a Schiff base) react with 2,2-dlchloroaeetaldehydeat similar ates in aqueous buffer (data not shown). 'H nudear magnetic 'esonance of the N-acetylcysteine conjugate also supported a thiohemiacetal structure. Mild acid hydrolysis of Conjugate A released 2.2-dichloroacetafdehyde. While aldehydes do not gen erally form stable thiohemiacetals, Introduction of electron-with drawing o-substituents stabilizes thiohemiacetals formed with aldehydes and thlohemiketals formed with ketones (6). Conjugate B is identified as S-<2-giutathionyf)acetyigiutathione, a bisglutathionyi conjugate formally derived from one molecule of 2-chloroaeetyl chloride and 2 of GSH. This conjugate is appropriately labile for a Ihioester and undergoes carbonyl sub stitution with added ("SjGSH at neutral pH (Chart 4). The characteristic reactions of the acyl moiety of Conjugate B are essentially the same as those for 2-chloroacetyt chloride, /.a., acylation and hydrolysis. However, the conjugate hydrolyzes much more slowly than the highly reactive acyl chloride and is thus available to a greater number of targets for a greater period of time. In this case, conjugation serves to prolong the Bfe of a reactive VDC metabolite, albeit as a metabolite of somewhat lesser reactivity. While Conjugate B formally derives from 2chloroacetyi chloride, attack of GSH on the methylene carbon of VDC oxide, followed by acylation of a second thiol by the resulting intermediate, would also yield Conjugate B. However, previous studies of VDC oxidation demonstrated that VDC oxide is formed in much smaller amounts than 2-chloroacetyt chloride (17) and lead to the conclusion that the epoxide is only a minor contributor to the production of Conjugate B. The identity of Conjugate C with 2-{S-glutathionyl)acetate was almost immediately evident. This product is an expected metab olite and would, upon renal biotransformation, give rise to several of the reported urinary metabolites, including 2-(S-(/V-acetyl)cysteinyl]acetate, thiodiglycollic add. dithloglycoitic acid, thiogtycollic acid, and possibly N-acetyt-S-(2-hydraxy)ethylcysteine (16. 18). Conjugate C is also the most stable of the 3 conjugates, due most likely to the stable thioether linkage between the VDC-derived carbon and cysteine sulfur. Conjugate C may be formed by the direct carboxymethytation of GSH by 2-chloroacetate, or by alkylation of GSH by 2-chloroacetyl chlo ride or VDC oxide, followed by hydrolysis of the intermediate product. Conjugate B hydrolysis also may provide a significant source of Conjugate C. The relative amounts of GSHtVDC metabolite conjugates dif fered in microsomal and hepatocyte incubations. The production of conjugates in the microsomal system is probably limited only by the amounts of metabolites generated, their stability, and the concentration of GSH. In contrast to the microsomal system, the formation of GSH:VDC metabolite conjugates in isolated hepatocytes is subject to additional factors present in whole cells. Conjugate C was the major GSH conjugate found in the hepatocytes, and Conjugate B was found at approximately one-fourth the level of Conjugate C (Table 3). Conjugate A was not found in the hepatocytes, an indication that 2,2-dichloroacetaldehyde produced In the cells Is so efficiently oxidized or reduced by cellular dehydrogenases that only negligible amounts are conju gated. The relative ratio of Conjugate B to Conjugate C In hepatocytes is nearly the opposite of that seen in microsomes (Chart 1; Table 3). The reduced levels of Conjugate B in hepa tocytes may be due to reduced availability of 2-chloroacety! chloride or VDC oxide. A more attractive alternative explanation for the inversion of Conjugate B:C ratios In microsomes versus hepatocytes is a possible contribution of enzymatic hydrolysis of Conjugate B in the hepatocytes. However, the ability of Conju gate B to serve as a substrate for hepatocyte esterases is entirely speculatory. Hepatocytes that were incubated with VDC released measur able amounts of Conjugate C but not Conjugates A or B into the medium. The extracellular conjugate in "S-prelabeled cells may represent conjugate efflux from the cells or conjugate formed outside the ceils by VDC metabolites that cross the plasma membrane to reaet with labeled GSH released into the medium. To determine the extent to which the latter process occurred, unlabeled ceils and VDC were incubated In medium containing [MC]GSH. Because hepatocytes cannot accumulate GSH from the medium, only extracellularly formed conjugates contain ra diolabel. Labeled Conjugates B and C were recovered from the medium (Table 3), demonstrating extraceiular formation of both conjugates. Formation of extracellular conjugates increased with increasing medium GSH content. Cells that were incubated without added GSH would release approximately 2.5 nmol/hr/ 10* cells into the medium (5). Assuming all glutathione efflux as GSH. the maximum amount of GSH present outside the cells during a 1-hr incubation would be S to 10 jim. In view of the fact that the levels of extracellularty formed conjugates with 5 mu medium ["SJGSH were approaching the Bmits of detection, it is likely that levels of extracellularly formed conjugates fn 3SSprelabeled ceOa were substantially below detection limits. The conjugates observed in the medium of the 95S-prelabeled celfs therefore represented efflux of Conjugate C from the cells. Extracellular formation of Conjugate B demonstrates that 2chloroacetyi chloride and VDC oxide are stable enough to mi grate a considerable distance through ceils and perhaps within a tissue. This phenomenon has been demonstrated previously for metabolites of benzo(a)pyrene(30). dimethylnitrosamine (34), vinyl chloride (10), and trichloroethylene (21). The phenomenon CANCER RESEARCH VOL 45 JANUARY 1985 191 AP00013513 VDC ADDUCTS Charts. General scheme for oxidative and oon}mjattve metabolismof VDC. See text tor discussion. of metabolite ceU-to-ceU migration has been offered as an expla nation for formation of DNAninyi chloride metabolite adducts in hepatic nonparertchymal oells, which do not efficiently metabolize vinyl chloride (10). The covalent modification of macromol&cules within target cells which do not produce VDC metabolites may be a consequence of metabolite migration. An integrated scheme for the major oxidative and conjugative pathways of VDC metabolism is presented in Chart 5. The pathways presented represent major reactions related to cova lent modification of GSH (and protein thiols). Secondary oxidative and reductive pathways for 2,2-dichloroacetaidehyde have been omitted as have possible reactions of VDC metabolites with ONA, which wifi be a subject of further study in this laboratory. The data presented here argue that multiple VDC metabolites participate in the covalent modification of ceBular proteins and that protein thiols are major targets. Formation of GSH:VDC metabolite conjugates offers a model for VDC metabolite-protein thiol interaction that presents advantages over the isolation of amino acid adducts via protein hydrolysis. The GSH conjugates may be isolated and quantitated rapidly, and individual conju gates contain characteristic metabofite:thfol linkages. Stability of these conjugates may reasonably approximate the stability of protein thiol:metabotite adducts. Interestingly, Jaeger et a/. (14) observed that hepatic p'cjVDC covalent binding, measured as trichloroacetic acid-insoluble radioactivity, decayed with a halflife of 2 to 3 hr. This figure closely approximates the half-life of Conjugate B. in contrast, hepatic covalent bineflng of acetami nophen, which apparently forms stable phenylthioether adducts with protein thiols (32), decays with a half-fife of approximately 12 hr (15). These data suggest that a significant fraction of the covalent binding of VDC metabolites consists of labile adducts which are subsequently lost to hydrolysis or dissociation. The consequences of formation of such adducts have not been addressed, but it may be hypothesized that such metastable adducts play a significant role In the Initiation of tumorigenic and toxic processes. REFERENCES 1. Ba/bin, A.. BresB. H.. Crolsy. A., Jacouigncn, P., Malaveile. G-. Montesarw. R,, and Bartaeh, H. Liver nticrosomfrmednrted formation of alkylating agent* from vinyl cfnande and vinyl bromide. Bfochem. Biophys, Res. Commun.. 07: $96-603,1975. 2, Bartsch, H,, MateveNle. C., Barbln, A.,and Planche.G. Mutagane and alkylating metabolites ot hamethylenes, cfWrcOutadienes. and dicWotobutenes produced by rodent or human liver tissues: evidence forotcirane formation by cytochrome P*4S04nked monooxygenasea. Arch. Toxicol.. 41:249-277,1979. 3. Soft. H. M~ Laid. R. J , and Finer, J. G. Reactive metabolites and eartinognv Ictty of halOgenated sthyfsnea. Bfochem. Pharmacol. 31: i 1962. 4. Cotta, A. K,, and tvanetich,K.M. WtyflCenechfohder its melabofem by hepatic microsomal cytochrome P-450 to vitro. Bicehem. Pha/mawl, 3t: 2083-2092. T962. 5. Fariss. M. W.. and Reed. D. J. Measurement of gfutalNone and giutalhtons dsuKde efflux from isofatad rat hepetoeytea. tor R. A. Hams and N. W. Cornell (eds.). Isolation. Characterization, end Use of Hepstocytea. pp. 349-366. New York: Elsevier North-HoHand Excerpts Medici, 1983. 9.Reid. L,, and Sweetman, B. BfolcgicaDy oriented organic sulfur ehtmiatry. f. Reaction* of thiol* with highly reactive carbonyl compounds, j. Org. Chem 34; 1788-1803.1969. 7. Goto, V., and Hilton, J. The hydrolysis of acetic anhydride. Part IV. Catalysia by hydrochloric acidand the hydrolysis of aeatyt chloride. J. Chem. Soc., 838642.1955. 8. Guengerich, F. P.. Crawford. W. M . Jr., and Watanebe. P. G. Activation of vinyl chforide to covaientry bound metabolite*: roles of 2-chtoroetftylene oxide and 2-chioroacataldehyde. Bfochemistry, 76:5177-5182,1979. 9. Guengerich, P. P., end Martin, M. W. Purification of cytochrome P*450,NADPH cytochromeP450 reductase, and epoxide hydratase from t single preparation of rat liver microsomes. Arch. Biochem. Biophys.. 203; 365-379.1989. 10. Guengerich, F. P., Mason, P. 3.. Stott, W. T.. Fox. T. R.. and Wetanaba, P. G. Roies of 2-hakMthylene oxides and 2-hakaectaidehydes derived from vinyl bromide and vinyl chloride In irreversible bindfog to protein and DNA. Cancer Res., 47:4391-4390.1981. 11. Guengerich, F. P,, and Watanabe. P. G. Metaboiiam of [MC|* and ("Cfytabeled vinyl chloride in vivo and In vitro. Biochem. Pharmacol., 2S: 589-596.1679. 12. Henschler, D.. and Bonse. G. Metabolic activation of chlorinated ethylene*: dependence of mutegenfe effect on efoctrophnc reactivity of the metabofioafly formed epoxides. Arch. Toxicol., 39; 7-12.1977. 13. Jaeger. R, J., Conolly, R. B., and Murphy, S. D- Effect of 18 hr tost and glutathione depletion on 7.1 -dkhtoroethylene-indueed hepatotoxlcity and le thality In rats. Exp. Mol. Pathol., 20:167-198,1974. 14. Jaeger, R. J., Shoner, L. G.. and Coffman, L 1,1-Ofchtoroethylefte hapatotoxieity: proposed mechanism of action and distribution and binding of "C radioactivity loiloweig Inhalation expoaure hi rats. Environ, Health Perspect,, 27:113-119,1977. 15. JaNaw. 0. J,, Mftthel, J. R. Potter. W. Z.. Davis. D. C,. Gillette, J. ft., and Brodi*. B. B. Acetaminophen-induced hepatic necrosis, tl. Rcie of covalent binding in Wvo. J. Pharmacol Exp.Ther., 197:195-202,1973. 18. Jones, B. K,, and Hathway. D. E. The bioiogicaHata of vinylidene chloride in rata. Cham^BW. Interact., 20; 27-41.1978. 17. Uebltr, O. C., and Guengerich, F. P. Olefin oxidation by cytochrome P-4SQ: evidence for group migration in catalytic intermediates formed with vinylidene chloride and irans-1-phenyt-t-butene. Bfochemistry, 22: 5482-5489,1983. 18. McKenna. M. J., Zempei, J. A., Madrid, E. O.. Braun. W. R, and Gehring, P. J. Metabolism and pharmeeokinetio profile of vinylidene chlonae hi rats followfog oral administration. Toxicol. Appl. Pharmacol, 45.- 621-935.1978, 19. Miller, e. c.. ana Miner, j. A. Searches for ultimate chemical carcinogens and their (sections with ceilLiar macromolecules. Cancer (Phla,), 4?: 2327-2345, 1981. 20. Mnier. R. E,, and Guengerich, F. P. Oxidation of trichloroethylene by liver microsomal cytochrome P-450; evidenoefor chlorine migration In a transition state not fovofvfng trfchloroechyfene oxide. Btochemiatry, 21: 1090-1097, 1982. 21. Miliar, R. E.. end Guengerich. F. P. Metabolism of triehiorosthylone In Isolated hepatocyte*. mrnosomaa, and reconstituted enzyme system* containing oytoehrom# p-450. Cancer Re*.. 4X1145-1152,1863. 22. Miwa, G. T., West. S. S.. Welsh, J. 8.. Wolf. F. J.. and Lu. A. Y. H. Drug residue formation from rtxKdazoie. a s-nflrohrtiduoie. lit. Studies on the mechanism of protein alkylation in vitro, Chem.-Biol Interact., 41: 297-312, 1962. 23. Reed. D. J.. Babson. J. R,, Beatty, P. W,, Srodfo, A. E,, Em*. W. W., and Potter, 0. W. High-perfoananoe liquid chromatography analysis of nanomole levels of glutathione, glutathione disulfide, end rested trtot* and Gsuifldss. Anal. Bioctiem., f06: 55-62.1980. 24. Reed, d. J.. and Orreniua. S. The roteofmethfontoa In glutathione bioiynthesis by isolated hapatoeytes. Bfochem. Biephys. Res. Commun., 77: 1257-1284, 1977. 25. Refohen. D., Werner, H. W,, Metzier. M.. end Henschiar. O. Molecular meeh. arism of 1,1-dlcWoroethyiena toxicity; excreted metabolitea reveal different pathway* of reactive intermediates. A/oh, Toxicol.. 42:159-169.1979. 26. Reitz, R. H- Walanabe, P. G., McKenna. M. J.. Quast. J. F.. and Gerirfog. P. J. Effects of vinylkiena chloride on ONA synthesis and ONA repair in the rat and mouse; a comparative study with dimetnyf nltrosamfo*. Toxicol. Appl. Pharmacol. 32' 357-370,1980. 27. Reynolds. E. 3.. Mosten, M. T., Boor. P. J., and Jaeger. R. G. i.i-Cicrtcroetriyiene hepetotoxidty: time course of GSH cnanoes and biochemical aber rations Am, J. Pathol., 101:331-344,1980. 28. Sedlack. J.. and Lindsay, R. H. estimation of total, protein-bound, and nonpro tein suiihydryt groups n tisaue with Elhran's reagent. Anal Bioctiem.. 25: 195-20S, 1968. CANCER RESEARCH VOL 45 JANUARY 1985 192 Uto. i. u gES3S3SSE5saaE9Kwn*w AP00013514 *paifc '2098, 'tMon# 2oroa# 'New ffy. f, hem., alysi* 838- on of OXIde DPHation . P. wtyl ineer Mted '8. net: caffy and J ie ato "C JCt., and effl ! in SO; n# P. no S. or XI 7, d yg e d i VDC ADDUCTS 1 29 Segfen, P. O. Preparation of rat Ivar oe(s. II. Effects of iocs and chelators on tissue dispersion. Exp. Celt Re*., 7ft- 25-30,1973. 3T Shan, A. L, Pahl. W. 6.. and Jetcoate, C. R. Metabolism of benzo(a)pyrene by isolated hepatocytsa and tactore effecting covalent binding of benzo4a)pyrena metabolttes to DNA in htpatocytes and microsomal systems. Arch. Btochem. etop#iYJ..2M: 571-523. 198a 3- Smith. C. V., Todd. E. L, Hughes, H. L., end Mitchell, J. R. Isolation and identification of specific products of alkylation of hepatic protein In vko by reactive metabottes of acetaminophen and bromobenzene. Fed. Proc., 4X 381,1984. Streeter, A. J.. Dahlifi, D. C-, Nelson, s. D.. end BaiNe, T. A. Mechanistic studies on the covalent binding 6f acetaminophen to protein: evidence for cysteine residues as major sites of arytation m vitro. Chsm.-Bioi. Interact., J4i 349-368.1984. Ugi, k, and Beck, F. Reakdon von CafbonsaurahalOQenlden mit Wasser und AmJnen. Cham. Bar., 04-1639-1850.1901. 34. Umhenhauer, 0. R.. and Pegg, A. E. Alkylation of Intraoeifuiarand extraceNuiar ONA by dmethylnltrosamlria foiowlng activation by isolated rat hepatecytes. Cancer Rea.. 41: 3471-8474,1061. 95. Wslfci, H.. Schelrt C.. Tunek. A., and Jergl. B. A tepid snd sensitive method for determination of eovusnt binding of benzo(a)pyrsne to proteins. Chem.Biol, interact.. 38; 109-118.1981. 38. Wang, P. P., Beeuna. P.. Kamineky, L. S.. Dannan. Q. A., Kadlubar. F, F., Urry, 0.. and Guangerleh. P. P. Purification and characterization of six cytochrome P-450 isozymes from human Ever rncrosomes. Biochemistry. 22: 8375-5383,1083. 37. Zaldeis. F.. Crotsy, A., Bsrtin, A, Malaveflte. C., Tomatfs. L,, end Bartscft. H. Carcinogenicity of chbroethyleneoxide. an Uttmate reactive metabolite of vinyl chloride, and bisfchtoromethytyethar after subcutaneous administration and in initiation-promotion experiments in mice. Career Re*., 40: 352-356,1980. CANCER RESEARCH VOL. 45 JANUARY 1985 193 J AP000135I5