Document LJdJxw8VGnJybpDkvJxV1dY7

PSana*ofcThgr. Vot 71, No*. 1/2, pp. 7-28, 1990* Copyright 1996 Elsevier Science Inc ISSN SO 163-7253/96 532 00 Pll 50163-7258(96)00060-5 7-C--<?/- ^ ELSEVIER Vinyl Chloride Mechanistic Data and Risk Assessment: DNA Reactivity and Cross-Species Quantitative Risk Extrapolation John Whysner,* C. Clifford Conaway, Lynne Verna, and Gary M. Williamsf TOXICOLOGY AND RISK ASSESSMENT PROGRAM. AMERICAN HEALTH FOUNDATION, I DANA ROAD. VALHALLA, NY 10595-1599. USA ABSTRACT. Vinyl chloride produced several rumor types among species. Angiosarcoma of the liver is found in all tested species, including humans with occupational exposures. Vinyl chloride is biotransformed by CYP2E1 to DNA-reactive chloroethylene oxide producing cyclic etheno adducts, which are mutagenic. The dose-response for angiosarcoma of the liver formation in rodents is supralinear, which is consistent with saturation of metabolic activation, and the tumor rate in humans at occupational exposure levels is similar to that for equivalent exposures in rodents. Pharmacol, ther. 71(1/2): 7--28, 1996. KEYWORDS. Angiosarcoma, DNA adducts, genotoxicity, risk assessment, vinyl chloride. CONTENTS 1. Introduction.......................................... 2. Results of Animal Bioassays............... 2.1. Inhalation studies........................ 2.1.1. Sprague-Dawley doseresponse STUDIES................... 7 8 8 8 2.1.2. Studies in other strains and species........................... 2.1.3. Studies of exposure 9 PARAMETERS AND EXPOSURE AGE................................ 10 2.2. Dietary and gavage studies . . . 11 2.3. Effects of intraperitoneal and subcutaneous INJECTION...................................................... 11 3. Epidemiological Studies..................... 11 3.1. Studies in the united States and Canada...................................... 11 3.2. studies in Europe............................. 12 4. Biotransformation................................ 13 5. Identification of DNA Adducts . . . 15 6. Genetic alterations by Vinyl Chloride-DNa adducts.................... 16 6.1. Studies of oncogenes and tumor-suppressor genes .... 16 7. Genotoxicity Studies With Vinyl Chloride................................................. 17 8. Toxicity, Cell Proliferation and Altered Foci.........................................17 9. Summaries.............................................. 20 9.1. Rodent bioassays and human studies........................................... 20 9.2. Mechanistic information ... 20 9.3. Species-to-species extrapolation............................ 21 9.4. Dose-response extrapolation............................ 21 10. Use Of Mechanistic Data in Risk assessment.............................................. 23 acknowledgements..................................... 24 References .................................................... 24 ABBREVIATIONS. eA, I,N6-ethenoadenme; Ado, 1 .N^-ethenoadenosine; ASL, angiosarcoma of the liver; CAA, chloroacetaldehyde; CEO, chloroethylene oxide; Cl, confidence interval; eC, 3,N4-ethenocytosine; eCyo, 3,N4-ethenocyridine; edA, 1 ,N6-ethenodeoxyadenosine; eJC, 3,N4-ethenodeoxycytidine; edG, N:,3-ethenodeoxvguanosine; eG, N:,3-erhenoguanine; 7-OEJG, 7-(2-oxoethyl)-deoxyguanosine; PVC, polyvinyl chloride; SCE, sister chromatid exchange: SMR, standardized mortality race; VC, vinyl chloride. 1. INTRODUCTION Vinyl chloride (VC) is a colorless gas with a boiling point of -- \ 3.TC. Ic polymerizes in light or in the presence of a catalyst. It is soluble in ethanol, diethyl ether, carbon tetra chloride and benzene, and is slightly soluble in water. The structure of VC is shown in Fig, l. VC is used to produce polyvinyl chloride (PVC), which is formed into many prod ucts as well as copolymer products, such as films and resins. `Corresponding author, tThis manuscript was developed for the International Expert Panel on Carcinogen Risk Assessment and under the aegis of the Environmental Health and Safety Council of the American Health Foundation. The membership of the Panel is provided in the first paper of this senes. Human exposure to VC occurs mainly in occupational set tings, but environmental exposures may have occurred since VC is released into the air and waste water from the plastics industry. Tobacco smoke has also been found to contain 1.3-27.3 ng of VC per cigarette (Hoffmann et al , 1976). Also, due to VC impurities in PVC, small amounts of VC may enter the drinking water from PVC pipes or may dissolve in foods packaged in PVC plastic food wraps. Occupational exposure to VC has been associated with angiosarcoma of the liver (ASL), which is a rare tumor among the unexposed population. In some studies, expo sure to VC has also been associated with increases in cancer of the CNS, lung and lymphacic/hematopoietic system. Due to these effects, the International Agency for Research on Cancer has determined that VC is a human carcinogen ASI 000013990 8 H / H2C = C \ Cl FIGURE 1. Chemical structure of VC. (IARC, 1979). VC has been chosen for review by the Panel because it is an example of a DNA-reacttve chemical for which the same tumor type has been reported in humans as in several other animal species. Extensive dose-response bioassays and epidemiological studies are available, which include quantitative estimates of exposure, thereby present ing a unique opportunity to compare the relative potency of this genotoxic carcinogen in humans and animals. This review is one of a series, which documents bioassay, epidemiological and mechanistic studies for 10 chemicals that are carcinogenic in rodents. The purpose of these doc uments is to illustrate the use of mechanistic information that assists in the determination of cancer risk at human exposure levels. The data for each topic are first presented wtth a minimum of analysis. The data are summarized and analyzed in the last sections, which the reader may wish to consult first for an overview. 2. RESULTS OF ANIMAL BIOASSAYS Maltoni et al. (1984) investigated the carcinogenicity of VC using male and female Sprague-Dawley rats, Wistar rats, Swiss mice and Syrian golden hamsters. In the initial experiment, 15 different VC doses, between 0 and 30,000 ppm, were administered by inhalation for a 52-week period to male and female Sprague-Dawley rats. Other VC inhala tion studies were performed over 5-, 17- and 25-week peri ods. Studies were also conducted on pregnant rats and em bryos. In addition, Wistar rats were tested so that strain variability could be evaluated; Swiss mice and Syrian golden hamsters were also tested so that species differences could be determined. VC was administered by stomach tube (gavage) and by i.p. injection to determine route-spe cific effects. Although the study was the most extensive ar^alysis of carcinogenicity due to VC conducted to date, other investigators have provided significant additional in formation. All of these studies are discussed in more detail in the following sections. 2.1. Inhalation Studies 2.1.1. Sprague-Dawley dose-response studies. In the initial series of experiments, groups of male and female SpragueDawley rats, 11-17 weeks of age, were exposed via inhala tion to 15 different concentrations of VC ranging from 0 to 30,000 ppm 4 hr/day, 5 days/week for 52 weeks, followed by lifetime observation (Maltoni etal., 1984). Six types of ma lignancies occurred in statistically significant excess. ]. Whvsner et al Angiosarcoma of the liver: The frequency of ASl m Spra gue-Dawley male and female rats exposed to VC for 52 weeks is presented in Fig, 2. The display of incidence vs, the logarithm of air concentration is presented based upon the relationship of log (dose) to hepatic macromolecular bind ing (Watanabe et al., 1978). Dose-related increases m the proportion of rats developing ASL were observed in the range of exposures from 10 to 30,000 ppm. The incidence became statistically significant at doses of 200 ppm and greater for males and 50 ppm and greater for females. Mammary gland tumors: Significant increases in the inci dence of malignant mammary gland tumors in female rats occurred at airborne concentrations of 5 ppm and above. The dose-response relationship was unusual in that at con centrations above 150 ppm, the incidence of malignant mammary gland tumors was not significantly greater than the incidence observed in control rats. However, the de creased incidence may be at least partially explained by the decreased survival in these groups due to other tumor types in comparison with the latency for development of malig nant mammary gland tumors. Zymbal gland carcinoma: The proportions of animals devel oping carcinomas of the Zymbal gland, which is the seba ceous gland of the ear canal, were significantly increased at 10,000 ppm and 30,000 ppm. Nephroblastoma: Statistically significant increases in num bers of nephroblastomas occurred in males in the dosage range of 100-2500 ppm and in females at 250 ppm and 500 ppm. The decreased incidence of this tumor type at higher doses may have been due to increased mortality of the ani mals from other tumor types, precluding development of a neoplasm with long latency. Neuroblastoma: Various types of encephalic tumors of neu ronal cells, usually without metastasis, were observed m the higher-dosage groups. Only in females of the 10,000 ppm dose group was the incidence statistically significant m this FIGURE 2. Dose-response of VC-induced angiosarcoma of the liver in rats. () indicates data not used in regression analysis. Data from Maltoni et al. (1984). ASl 000013991 or 52 s. the n the bmdn the n the fence i and inci; rats hove, con,7nant than e deV the types -lalig- evet^ebaed at "age ! 500 igher amof a neuthe ppm this zcooo ysis. Vinyl Chloride Mechanistic Data dose-response study, although an elevated incidence ap peared beginning at 2500 ppm. Forestomach tumors: Papillomas and squamous cell carcino mas of the forestomach occurred in 18.3% of rats exposed to 30,000 ppm VC and m several of the lower-dose groups in nonsignificant proportions. Other tumors: The incidences of the following rumors asso ciated with VC exposure were increased, although the in creases reported were noc statistically significant using the Fisher exact probability test (P < 0.05): angiomas and fibroangiomas of the liver; hepatocellular carcinomas; an giosarcomas, angiomas and fibroangiomas at sites other diaii uW liver; lung adenomas; and cutaneous epithelial tu mors (squamous cell carcinomas). 2.1.2. Studies in other strains and species. Studies of Wistar Rats: In an early experiment, 26 male Ar/IRE Wistar rats, 3 months old, were exposed to 30,000 ppm VC 4 hr/day, 5 days/week, for a period of 12 months (Viola et at., 1971). In all 17 rats surviving 54 weeks, skin tumors in the submaxillary parotid region were reported. Lung tumors were re ported in 7 rats and osteochondroma in 5 rats. No tumors were reported in 25 control rats. Maltoni et al. (1984) later concluded that the skin tumors were carcinomas originat ing from the Zymbal gland and that the lung tumors were probably metastases from Zymbal gland carcinomas. Inhala tion studies on Wistar rats were also conducted by Feron et al. (1979), who exposed groups of males and females to 0 or 5000 ppm VC for 52 weeks; after 4, 13, 26 and 52 weeks, 10 rats/group were killed. In rats killed after 26 weeks, only foci of cellular alteration, consisting of clear cells (14/20) or basophilic cells (5/20), were observed in the livers, but an giosarcoma (9/19) and hepatocellular carcinoma (2/19) were observed after 52 weeks. Primary tumors in the brain, lung, Zymbal gland and nasal cavity were also reported in these rats (Feron and Kroes, 1979). Controls had few liver changes and no liver tumors or other tumors. Male Wistar and Sprague-Dawley rats were utilized in life-span inhalation studies to investigate the effect of strain differences on tumor incidence (Maltoni et al., 1984). Both strains of rats were exposed 4 hr/day, 5 days/ week for 52 weeks to VC concentrations of l, 50, 250, 500, 25400, 6000 and 10,000 ppm. The incidence of liver an giosarcomas was comparable between the two strains at doses up to 10,000 ppm. Fewer Zymbal gland tumors were recorded in male Wistar rats than in male Sprague-Dawley rats. For example, at 10,000 ppm the Zymbal gland tumor incidence was 2/18 (11.1%) for Wistar rats compared with 1CK30 (33.3%) for Sprague-Dawley rats. The incidence of many other tumor types that occurred in Wistar rats was comparable with the tumor incidence in Sprague-Dawley rats. Forestomach papillomas and squamous cell carcino mas, however, were not recorded in the Wistar rats (Mal toni etal., 1984). In another investigation, Bi et al. (1985) reported that 36.8% (7/19) adult male Wistar rats devel oped liver angiosarcoma and 10.5% (2/19) developed lung 9 angiosarcoma after exposure to 100 ppm VC 6 hr/day, 6 days/week for 12 months; killing was at 18 months. Studies Comparing Rats, Mice and Hamsters: When CD rats and CD-I mice, approximately 2 months of age, were ex posed to 0, 50, 250 and 1000 ppm VC 6 hr/day, 5 days/week for up to 12 months by Lee etal. (1978), ASL was the most prevalent neoplasm induced in the rats occurring in 17% and 30% at 250 and 1000 ppm VC, respectively. Hemangiosarcoma of the lung also occurred in rats at 250 ppm (3/ 70) and 1000 ppm (13/70). A few hemangiosarcomas were also found at other sites. The mice also developed bronchioalveolar adenomas and mammary gland tumors in high incidences. A subsequent study, in which both CD rats and CD-1 mice were exposed to 0, 50, 250 and 1000 ppm VC in an identical regimen for up to 10 months, but followed by a 12-month observation period (Hong et al., 1981), resulted in comparable values for ASL in racs and mice. The inci dence of mammary gland tumors in female mice, however, was somewhat reduced compared with the earlier study. In another study, groups of 5-week-old male CD1 Swiss/ChR mice were exposed to 2500 or 6000 ppm VC 5 hr/day, 5 days/week for 5 or 6 months by Suzuki (1978). Almost all of the dosed mice (26/27) killed 2-37 days after termination of exposures had developed lung neoplasia diagnosed as "alveogenic cancer"; none of 16 controls developed lung can cer. Light and electron microscopic examination of the tis sues suggested chat the neoplastic cells had arisen from Type U alveolar epithelium (Suzuki, 1978). A small propor tion of the mice had developed ASL. F344 rats and ICR mice were exposed by inhalation to 50-50,000 ppm VC for 1 hr (Hehir et al., 1981). The age range of these animals was 15-18 weeks, and che animals were observed for their lifetime, which was 18-24 months. VC produced no significant increases in pulmonary neopla sia in ICR mice until the chamber level attained 5000 ppm (1/143 carcinoma and 24/143 adenomas) or 50,000 ppm (3/137 carcinomas and 45/137 adenomas) compared with controls (0/120 carcinomas and 12/120 adenomas). Pneu monitis occurred in all groups of mice exposed to 500 ppm or more for 1 hr. None of the F344 rats developed tumors even at the highest dose. No dose-related increases in other types of tumors were found in this study. Studies by Drew et al. (1983) indicated that B6C3F[ and Swiss female mice had the highest incidence of ASL related to the lowest concentration of exposure for any of the stud ies. Mice exposed to 50 ppm VC for 12 months and ob served for an additional 12 months showed ASL incidence of 77% for B6C3F| and 64% for Swiss mice. No ocher doses were tested in this experiment. Syrian golden hamsters dosed at 200 ppm showed much lower ASL incidences, whereas female F344 tats dosed at 100 ppm showed a 20% incidence of these tumors. Swiss mice and Syrian golden hamsters were also utilized by Maltoni et al. (1984) in order to investigate species dif ferences m response to VC exposure In Swiss mice, ASL appeared in greater proportions at lower-dose levels after AS I 000013992 10 30-week exposures than in Sprague-Dawley rats exposed for 52 weeks; at higher doses, however, the incidences were comparable. The incidence of ASL in male Syrian golden hamsters was considerably less after 30-week exposures than the incidence in male Sprague-Dawley rats or in male Swiss mice at corresponding daily exposure levels, indicat ing a definite species difference for hamsters. At high expo sure levels, the incidence of mammary carcinomas in fe male Swiss mice was greater, but ar lower levels, ic was similar to the incidence in female Sprague-Dawley rats (43.3% vs. 10.0% at 2000 ppm, 40.0% vs. 35.0% at 50 ppm) exposed for 52 weeks. Other tumor types char were common in rats, e.g.. Zymbal gland carcinomas, neuroblas tomas and nephroblastomas, did not occur in Swiss mice or Syrian golden hamsters. Melanomas, lymphomas and leu kemias occurred only in Syrian golden hamsters; corre sponding neoplasms did nor occur in rars or mice. 2.1.3. Studies of exposure parameters and exposure age. In one protocol. 12-week-old Sprague-Dawley rats were ex posed to several concentrations of VC ranging from 50 to 10,000 ppm 4 hr/day, 5 days/week for 17 weeks. Animals were observed for their lifetime (156 weeks) ro determine if a reduction in the number of total hours of exposure influ enced the tumor yield. Compared with incidences in groups exposed for 52 weeks, dramatic decreases in the proportions of rars with ASL (1.7% vs. 22.0% for 6000 ppm and 0.0% vs. 11.7% for 10,000 ppm) were observed, but the effect of reducing the exposure duration by a factor of three on the incidence of Zymbal gland carcinomas in the 6000 and 10,000 ppm groups was equivocal. Some reductions m nephroblastomas were also recorded in the rats exposed for 17 weeks (Maltoni et ai., 1984). In another study, varia tions m the exposure regimen were tested, keeping rhe total exposure time constant. Sprague-Dawley rats, 11 weeks old, were exposed to 6000 and 10,000 ppm VC 4 hr/day, 5 days/ week for 5 weeks; 4 hr/day, 1 day/week for 25 weeks; or 1 hr/ day, 4 days/week for 25 weeks (100 hr total m each in stance), followed by lifetime observation. The incidence of ASL was very low; however, the incidences of mammary and Zymbal gland carcinomas in these rats did not differ significantly among the various dosage groups, suggesting that the total exposure time rather than dosage regimen is the most importanc determinant of tumortgeniaty (Mal toni et ai, 1984). In investigations on the effect of age, embryo and new born Sprague-Dawley rats appeared to be much more sensitive to VC exposures than 11- to 13-week-old rats tn developing brain neuroblastomas, ASL and hepatocellular carcinomas after exposure (Maltoni et ai, 1984; Maltoni and Cotti, 1988). Newborn Sprague-Dawley animals were exposed to 6000 and 10,000 ppm VC for 5 weeks according to the same protocol as adults with lifetime observation. The re sulting ASL incidences for newborns were 17/42 (40%) and 15/44 (34%) for 6000 and 10,000 ppm, respectively. Large numbers of adenomatous tumors of the lung were also associated with perinatal exposure to VC. Eleven-week-old J Whysner ;i ai rats exposed according ro the same doses and regimen had ASL incidences of 0/120 (0%) and 1/118 (1%), respec tively. The incidences of hepatomas showed similar differ ences. Anocher experiment compared tumor incidence be tween breeder animals and embryos for 2500 ppm VC exposure beginning ar 13 weeks of age for breeders and 12 days of gestation for embryos and continuing for 76 weeks (20-35 hr/week). In this experiment, the number of neuro blastomas and angiosarcomas were similar between breeders and offspring, but hepatocellular carcinoma incidences were 9,2% and 51.2% for breeders and offspring, respec tively. Further investigations on the role of exposure intensity compared with number of exposures, resulting in the same total VC dose, were conducted by Hehir et al (1981) in mice. When A/] mice had 10 I-hr exposures to 500 ppm VC, the incidence of pulmonary adenoma (124/166) and lung carcinoma (22/166) was significantly greater than matched controls (31/90 pulmonary adenomas and 3/9Q lung carcinomas, P < 0.001) and also considerably greater than when they received 100 1 -hr exposures to 50 ppm VC (65/158 pulmonary adenomas and 7/158 lung carcinomas). Consequently, ic appeared chat these mice were somewhat more sensitive to less frequent, more intense exposures; however, the differences were only 2- to 3-fold. An extensive study by Drew et al (1983) further demon strated that age of female Fisher 344 rats, C57BL/6N mice, CD-1 Swiss mice and Syrian golden hamsters influenced the number and type of tumors observed after VC adminis tration. For example, 4/76 (5.3%) of 5- to 6-week-old female Fisher 344 rats exposed to 100 ppm VC 7 hr/day, 5 days/week for 6 months developed a statistically significant number of hemangiosarcomas. (Only total hemangiosarcomas were reported in this study.) In contrast, 7-month-oU female Fischer rats exposed to VC by an identical experi mental protocol developed 2/52 (3.8%) hemangiosarcomas, and rats held for 12 months or 18 months prior to the daily exposures for 6 months failed to develop any hemangiosatcomas. When VC exposure was for a total of 12 months, deferring the period of exposure for 12 months reduced the ASL incidence front 11/55 (20%) to 2/49 (4%). In Swiss mice, 6 months' exposure to VC (50 ppm) produced signifi cant incidences of hemangiosarcomas, mammary gland car cinomas and lung carcinomas. After 12 months of expo sure, the incidence of ASL was 30/47 (64%). If initiation of exposures was deferred until the mice were 12 months of age, the incidence for 12 months of exposure was decreased to 3/50 (6%). The incidence of hemangiosarcomas and mammary gland carcinomas was also reduced m B6C3F, mice that were dosed at ages older than 5-6 weeks; the re duction of total hemangiosarcomas after a 12-month defer ral of dosing was from 69/90 (77%) co 29/48 (60%). Female Syrian golden hamsters developed hemangiosarcomas, mammary gland carcinomas, stomach adenomas and skin carcinomas after exposure to 200 ppm VC beginning at 2 months of age for as little as 6 months' duration. When the exposures were starred at 8 monchs of age, the incidence of ASI 000013993 -r etal. r i had ispecdiffer ed bea VC nd 12 weeks leuro.eders ences -'spec- |i ; I, znsity rame .1) in ' ppm ) and than 3/90 rearer n VC mas), what .ures; mon- unisk-old lay. 5 icant arcoh-old perimas, daily osarichs, I the ' WISS :mfiI carxpo>n of ns of used and C3F1 e re:eferinale mas, i the .e of Vinyl Chloride Mechanistic Data hemangtosarcomas and mammary gland carcinomas was again markedly reduced. 2.2. Dietary and Qavage Studies Male and female Sprague-Dawley rats were dosed 5 times weekly by gavage with 0, 0.03, 0.3, 1,0, 3.33, 16.6 and 50.0 mg/kg/day VC m olive oil for 52 weeks and observed for their life span by Maltoni et al. (1984). ASL were induced in males at dose levels as low as 0.3 mg/kg/day; the inci dence was 12% at 16.6 mg/kg/day and 21% at 50.0 mg/kg/ Jay. Dose-related increases m liver nodular hyperplasia and diffuse hyperplasia were also observed; however, the levels of neoplastic nodules did not achieve statistical signifi cance. Also, nephroblastomas (2% and 3.7% in respective dose groups) and forestomach papillomas and squamous cell carcinomas (2.5% and 1.2% in respective dose groups) oc curred in the 50.0 mg/kg/day and 16.6 mg/kg/day groups. Zymbal gland carcinomas occurred at 1.0 mg/kg/day (3.3%), 16.6 mg/kg/day (2.5%) and 50.0 mg/kg/day (1.2%). Extrahepatic angiosarcomas and angiomas were recorded in the 50.0 mg/kg/Jay (2.5% incidence, each type) and 3.33 mg/ kg/day gavage Jose groups (2.5% extrahepatic angiosarco mas and 1.2% angiomas). A lifetime oral toxicity study of VC monomer was per formed by incorporating PVC containing 0, 1.8, 5.6 and 17.0 mg VC/kg/day into the diet of Wtstar rats for 135 weeks in males or 144 weeks in females. VC was also ad ministered at 300 mg/kg/day, 5 days/week for 83 weeks by gastric intubation in soy bean oil (Feron et al., 1981). Sta tistically significant numbers of ASL were observed at 5.6 mg/kg/day in males and at 17.0 mg/kg/day in females. The tumor response demonstrated a shift from 0/116 ASL and 5/116 hepatocellular carcinomas at the 1.8 mg/kg/day di etary dose to a mixture of 8/114 ASL and 21/114 hepato cellular carcinomas in the 5.6 mg/kg/day dose group, 36/116 ASL and 37/116 hepatocellular carcinomas in the 17.0 mg/kg/day diet group, and finally to a preponderance of 56/109 ASL and 1/109 hepatocellular carcinomas in the 3C0 mg/kg gavage group. Angiosarcomas of the lung, extrahepatic abdominal angiosarcomas and tumors of the Zym bal gland appeared in rats at doses of 5.6 mg/kg/day or greater, A second study in the same laboratory was con ducted using male and female Wtstar rats, which were fed VC in powdered PVC in diet for 149 weeks. The doses used were 0, 0.014, 0.13 and 1.3 mg VC/kg/day (Til etal , 1991). On the basis of the results, it was concluded that induction of tumors, neoplastic nodules and hyperplastic foci in the liver were the most sensitive responses to dietary VC and that 0.13 mg/kg/day was the "no-observed-adverse-effecctevel" for the induction of liver tumors m Wistar rats. 2.3. Effects of lntraperitoneal and Subcutaneous Injection lntraperitoneal or subcutaneous injections of VC were not effective in causing various types of tumors (Maltoni et al., 1984); however, the cocal doses administered were small. 11 The incidence of ASL was 0 in 4 groups of 60 rats adminis tered 4-25 mg/kg VC i.p. either 4, 3, 2 or 1 time at 2-month intervals; 2/188 angiosarcomas occurred at other sites in the 4 dosage groups. In another study, a single nephroblastoma was diagnosed in 32 Sprague-Dawley rats surviving to 85 weeks after a single i.p. dose of 4.25 mg VC/kg; no nephro blastomas occurred m 49 control rats (Maltoni et al., 1984). 3. EPIDEMIOLOGICAL STUDIES In a number of studies, VC has been found to cause ASL, a rare tumor in humans. Some quantitative information is also available regarding the historical VC levels in the workplace. Barnes (1976) provided estimates that exposure concentrations for various years were the following; 19451955, 1000 ppm; 1955-1960, 400-500 ppm; 1960-1970, 300-400 ppm; mid-1973, 150 ppm; 1975, 5 ppm. Based upon measurements of VC and interviews with workers concerning VC odor detection, Heldaas et al. (1984) esti mated exposure levels of 2000, 1000, 500 and 100 ppm for the years 1950-1954, 1955-1959, 1960-1967 and 19681974, respectively. Thus, the estimates of Barnes (1976) and Heldaas etal. (1984) are within a factor of two. 3.1. Studies in the United States and Canada Monson et al. (1975) reported an epidemiological study among workers at two plants in Kentucky, USA involved in the production of VC and PVC. Death certificates were obtained for all 161 deceased workers who had been ac tively employed or pensioned. There were eight deaths due co cancer of the liver and biliary tract, and five of these were diagnosed as ASL. A proportional analysis revealed that the ratios of observed vs. expected mortality were 11.0 for liver and biliary tract, 1.6 for lung, 1.6 for digestive tract and 4.2 for brain. Heath et al. (1975) reported that seven workers at a plant in Louisville, KY, USA developed ASL after having been exposed for between 12-28 (average 17) years. This was among a work force of approximately 270 workers exposed to VC beginning in 1942. Wu etal. (1989) reported the most recent update of a co hort analysis utilizing death certificates and a nested casecontrol study of 3635 male workers exposed to VC and PVC. Workers were exposed between 1942 and 1974, with follow-up extending through 1986. Prior published reports of these workers have included Waxweiler et al. (1976) and Waxweiler et al. (1981). Only liver cancer incidence was elevated [standardized mortality rate (SMR) = 333; confi dence interval (Cl)* =* 202--521]- Estimates of cumulative dose were made, and a statistically significant association with VC and liver cancer was determined. Analyses for lung and brain cancer were also reported, and no associa tion with VC exposure was found. After review of medical records and death certificates, there were 12 ASL cases and *Cl is the 95% Cl unless otherwise stated. ASI 000013994 12 7 hepatocellular carcinoma cases. Only the ASL cases ex hibited a positive dose-response with VC. Theriault and Allard (1981) reported that 451 Canadian workers who were heavily exposed to VC for more than 5 years, employed from 1943 through 1972 and followed through 1977, had an excess of digestive system tumors. Histopathological confirmation of all cancer cases was de termined, and there were 10 total ASL cases found, result ing in an SMR of 5714. The exposures of these workers were thought to be quite heavy because several episodes of unconsciousness were reported. In this study, the exposure durations for both the exposed population and the ASL cases were reported. They were found to be comparable, al though there were no ASL cases among the 105 men "'"+1 the longest exposures. This study will be analyzed in Section 9 4. Rinsky et al. (1988) and Teta et al. (1990) reported re sults of exposure to VC at Union Carbide Corporation fa cilities in the Kanawha Valley of West Virginia, USA, be tween 1940 and 1978. The Rinsky et al. (1988) study reported on a particular set of facilities within the Teta et al. (1990) study. In the Rinsky et al. (1988) study, the re sults were specifically linked to VC exposure. Increased in cidences of liver cancer (SMR = 174; Cl = 102-280) and lympho- and reticulosarcomas (SMR = 140; Cl = 104-187) were reported. The highest SMRs occurred among persons who worked at least 25 years (SMR = 239) and who died 30 years or more after first employment (SMR = 301). The results for lympho- and reticulosarcomas did not show ef fects of latency or years of employment. A study of a cohort in which there were 1536 deaths out of 10,173 men who had at least 1 year of exposure to VC due to employment at 37 plants in the United States was completed by Wong et al. (1991). A previous report of this cohort was reported by Tabershaw and Gaffey (1974). The updated report included observed mortality from 1942 through 1982. Fifteen ASL cases were reported among the 37 deaths due to cancers of the liver and biliary tract (SMR = 641; Cl = 450-884). The excess mortality due to cancers of the liver and biliary tract (SMR = 386) was still present even after the removal of ASL cases. Also reported were increased cancers of the brain and central nervous sys tem (SMR = 180; Cl = 114-271). No excess mortality was found for cancers of the respiratory system, lymphatic and hematopoietic system or digestive system other than the liver. Analysis of mortality by latency for liver and biliary tract tumors (including ASL) showed a clear positive trend; however, for brain and CNS, this trend was not obvious. There was the expected inverse relationship of age of first exposure for liver and biliary tract tumors, but for brain and CNS tumors, there was an opposite relationship. Conse quently, the increased mortality due to cancers of the brain and CNS appears to be equivocal. 3.2. Studies in Europe Byren et al. (1976) reported three cases of angiosarcoma among 58 cancer cases in 771 workers who had worked in I- Whysner et al various chemical operations with exposure to VC in Swe den. These workers were compared with standardized mor tality statistics from the Swedish Central Bureau of Statis tics. Four cases of cancer of the liver/pancreas, which included two of the ASL cases, resulted in an SMR of 413 (P -- 0.017). One of the cases of ASL was added to the pa per as a note in proof. Also, these authors found two cases of brain tumors, yielding an SMR of 612 (P = 0.043). How ever, one of these cases had been employed for less than l year at the time of diagnosis. Fox and Collier (1977) re ported the results of a study of 393 deaths among 7000 men exposed to VC from 1940 to 1974 in the United Kingdom. There were four deaths due to liver cancer, and two of these deaths were ASL. The two cases of ASL had 8 and 25 years of high constant exposure, which was estimated to average 200-500 ppm. No other cancers were found to be at ele vated incidence. A follow-up of this study through 1984 was reported by Jones et al. (1988). The number of cases was restricted to the 5498 workers having exposure to VC for at least 1 year and during at least 25% of the working period. At this point, 11 cases of primary liver tumor had been identified (SMR = 567; Cl 3 283-1015) from the 780 deaths from all causes. Among autoclave workers, there were seven deaths due to liver tumors, and they were all identified as ASL. The mean latency period from time of first exposure was 25 years. One of the cases had an expo sure of 2-4 years, 2 had between 5-9 years and 4 had greater than 10 years. There were no increases in neoplasms ot the brain, lung, colon, lymph system or in malignant mela noma. An increase in urogenital cancers was not correlated with degree of VC exposure. Weber et al. (1981) reported on German workers exposed to VC or PVC who died be fore 1974- Death certificates of these workers were com pared with mortality rates of the West German population and also to a group of chemical workers not exposed to VC. The SMR for workers exposed to VC was 1523, compared with 401 for the reference group. VC exposure was also as sociated with statistically significant increased incidences of malignancies of the lymphatic and hematopoietic tissues (SMR = 214) and of the gastrointestinal tract and perito neum (SMR = 149). The SMRs for 13-60, 61-120 and >121 months were 874, 1525 and 2528, respectively. Heldaas et al. (1984) reported one ASL case among 23 can cer cases in 454 VC and PVC workers in Finland. The one case was in a person exposed for 22 years. During most ot this person's employment history, VC levels were between 500 and 2000 ppm. There were also four cases of malignant melanoma reported, resulting in an SMR of 510. Pirastu et al. (1990) investigated 253 deaths among persons who had been involved with using VC or PVC at plants in Italy. Death certificates and clinical and pathological data were reviewed, and seven cases each of ASL and hepatocellular carcinoma were verified. The mean duration of exposure was 16 and 18 years, respectively. It was not possible to compare these results with expected values. The mean latencies were 19 and 20 years, respectively. A large multicenter cohort mortality study of European VC workers j I j : .! 1 ASI 000013995 f eretdi. ^ i morStatiswhich of 413 he pa> cases Howhan 1 7) re0 men gdom. ' these > years 1 erage it ele- 1984 cases o VC irking >r had m the there re all me of expo- : melaelated orted d becomiatitm * VC. pared Uo asiences 'issues 'erito3 and ively. ) cane one 1st of ween tnant stu et i had Italy, were Ilular osure i J Vinyl Chloride Mechanistic Data found a 3-fold increase in cancer of the liver (Simonato et al, 1991). This study included 14,351 workers from Italy, Norway, Sweden, and the United Kingdom. Some of this data was included in other reported studies (Byren et al, 1976; Jones et al., 1988; Heldaas et al, 1984; Pirascu et al, 1990); however, this study included additional workers, and follow-up was extended. Only cancers of the liver and intrahepatic bile ducts were statistically significantly elevated (SMR = 286; Cl = 183-425). Fourteen cases of brain can cer vs, 13 expected were reported. 4. BIOTRANSFORMATION VC can be biotransformed to chloroerhylene oxide (CEO), which readily reacts with DNA or chloroacetaldehyde (CAA), which primarily reacts with protein (Fig. 3), The main enzymatic pathway for metabolism of VC requires mi crosomal mixed-function oxidase. Microsomal activation of VC has been confirmed both in vivo (Reynolds et al, 1975; Guengerich and Watanabe, 1979; Bartsch et al., 1979; Guengerich et al, 1981) and in vitro (Barbin et al., 1975; Guengerich et al., 1979; Guengerich, 1982); oxygen and NADPH are required as cofactors. When a mixture of VC and pure oxygen was passed through a mouse liver microso mal fraction fortified with an NADPH-generating system, volatile alkylating metabolites were trapped by reaction with excess 4-(4-mtroben:yl)pyndine. The absorption spectra of the adducts formed were identical to those ob tained by reaction of CEO with 4-(4-nitrobenzyl)pyridine, 13 indicating that CEO, the highly reactive epoxide of VC, was formed (Barbin et al., 1975; Bartsch et al, 1979). Microsomes prepared from human surgical specimens of liver exhibited a 9-fold variation in the capacity to convert VC into electrophiles mutagenic to Salmonella cyphimurium strain TA 1530 (Sabadie et al, 1980). The specific mixedfunction oxidase required for bioactivation of VC to its ep oxide CEO was identified as CYP2E1 by selective inhibi tion of the enzyme with diethyldithiocarbamace and bystudies with anti-CYP2El in liver microsomes (Guengerich et al, 1991). Additional evidence for the role of CEO will be described in Section 5. CEO (Fig. 4) has a half-life of 1.6 min in aqueous solution and can spontaneously rearrange to CAA (Barbin et al , 1975). In addition, CEO can be hydrolyzed by the enzyme epoxide hydrolase. When epoxide hydrolase was added to an aqueous solution of CEO, the ti/s for disappearance of CEO was decreased. When epoxide hydrolase was added to a rat liver microsomal preparation, NADPH-dependent co valent binding of 14C (from 14C-VC) to proteins was de creased by 50% (Guengerich et al., 1979). However, subse quent studies have shown almost a complete inhibition of DNA-adduct formation from CEO following the addition of epoxide hydrolase to the incubation (Guengerich, 1992). VC metabolism was also reported to be partially inhib ited by pyrazole and ethanol, both of which can inhibit al cohol dehydrogenase; this suggests that VC may be oxidized to 2-chloroethanol and then via alcohol dehydrogenase to CAA (Hefner et al, 1975; Guengerich et al, 1979). How ever, it is also possible that decreased VC metabolism may \ c=c x Cl /\ VC P450 0,, NADPH Cl "c-c" W o CEO REARRANGEMENT cich2-cho CAA CHO I 0 ch2 'tkers 7-OEdG 3,N4-Ethenodeoxycytidine 1,N6-Ethenodeoxyadenosine N2,4-Ethenodeoxyguanosine FIGURE 3. VC metabolism and adducts. Data from Ciroussel et al. (1990). AS I 000013996 14 CICHj-CHj-OH 2 chloroethanol ii i i' \: C!CH=CHj --^YP"~-- CIH-C-CH, VC CEO J- Whysner it ai DNA binding epoxide .hydrolaw HOCHj-CHO glycolaldehyde fcSjr + glutathione GS-CHj-COOH 3-carboxymethyl glutathione Further metabolism and excretion FIGURE 4. Proposed metabolic pathways for VC. have been due to competitive inhibition of CYP2E1 by eth anol, Oxidation of 2-chloroethanol in the presence of hy drogen peroxide and catalase has also been proposed (Hefner et ai., 1975), but evidence for this is not available. CAA can be oxidized by aldehyde dehydrogenase to 2-chloroacetic acid (Fig, 4), which has been found in the urine of VC-dosed rats. 2-Chloroacetic acid can also bind to glutathione to form S-carboxymethylglutachione, which can also be formed from S-formylmethylglutathione (Plugge and Safe, 1977). CEO and CAA can bind covalently to sulfhydryl-containing groups m protein (Kappus et ai., 1975) or react with glutathione via glutathione epoxide trans ferase (or other glutathione transferases) to form S-formylmethy[glutathione. Further metabolism (not shown in Ftg. 4) occurs with conversion of S-formylmethylglutathione to S-carboxymethylcysteine and S-(2-hydroxyethyl)-cysteine; these products have been identified in che urine of VC-dosed rats. N-acetylation of S-(2-hydroxyethyl)-cyst^ine then yields N-acetyl-S-(2-hydroxyethyl)-cysteine, thiodiglycolate is formed from S-carboxymethyl-cysteine by deamination, and subsequent decarboxylatton products have been identified as urinary metabolites of VC (Watanabe et ai., 1976a,b; Plugge and Safe, 1977). Several studies have indicated that the epoxide metabo lite of VC, CEO, is the ultimate carcinogen that reacts with DNA. Using in vitro studies, Guengerich etal. (1981) found that epoxide hydrolase (which reacts wich CEO), but not alcohol dehydrogenase (which reacts with 2-chloroethanol, a CAA precursor), inhibited binding of VC metabolites to calf thymus DNA (Fig. 4), proving that CEO was the DNA-reactive species. Gwinner et ai. (1983) determined m vivo that the likely proximal carcinogen was CEO rather than CAA. After exposing young rats to VC and to the precursors 2,2'-dichlorodiethyl ether and chloroethanol, which are metabolized only to CAA and not to CEO, only the VC-exposed rats developed preneoplastic hepatocellu lar ATPase-deficient foci, indicating that VC was DNA re active (presumably through a CEO intermediate) and that CAA was not DNA reactive in vivo. A single 1 mg (12.7 P-mol) s.c. dose of CEO to 8- to 10-week-old XVllnc/Z mice was administered followed by 2 p,g 12-O-n-tetradecanoylphorbol-13-acetate 3 times per week for 42 weeks (Zajdela et ai., 1980). After 590 days, 18/22 (82%) and 5/22 (23%) of animals had papillomas and carcinomas ot the skin, respectively. In comparison, the same dose of CAA produced papillomas in 3/20 (15%) of animals, which was not different from controls, and there were no carcinomas. The authors concluded from this study that CEO rather than CAA was the proximate carcinogen for VC. Protein binding of VC following inhalation exposure to 2, 100 or 1000 ppm for 5 hr occurred in the liver, kidney, lung and small intestine, with lesser amounts in other or gans and tissues of male Wistar rats (Bolt and Filser, 1977), Binding of VC or its metabolites to protein was partially blocked by reduced glutathione, but binding to DNA was not inhibited. By addition of radiolabeled CAA to the sys tem, it was determined that CAA, or its products of metab olism, readily bind to intracellular protein and that reaction of CAA with DNA was very slow (Guengerich etal., 1981) In rats, VC metabolism is a dose-dependent saturable process. Watanabe et ai. (1976a) investigated the fate ot ra diolabeled VC following oral exposure in male Sprague- ASI 000013997 isr et al. Hanoi, >, only 'celluJAred that (12.7 ! Inc/Z trade weeks ! 5/22 >f the CAA h was omas. rather ire to Jney, et or977). tially \ was e sysetabction ft of raigue- Vmyl Chloride Mechanistic Data Dawley rats. As the dose was increased from 1 to 100 mg/ kg, the proportion of the administered dose expired in the lungs as VC increased from 2 to 67%. Following exposure to air containing 10 ppm VC, urinary and expired radiola bel accounted for 68% and 2%, respectively. Exposure to 1000 ppm VC resulted in urinary and expired radiolabel amounts of 56% and 12%, respectively (Watanabe et al., 1976b), Several 6-hr doses of VC ranging from 1 to 4600 ppm were administered to male Sprague-Dawley rats by in halation (Watanabe et al., 1978; Gehring et al., 1978). The amount of metabolism and the formation of protein-bound merabolires in the liver were reported to be a sigmoidal function of the log of exposure concentration, which means that binding of UC-VC or metabolites was low at exposure concentrations below 50 ppm, linear with respect to log concentration in the range 50-500 ppm, and reached a pla teau at 500 ppm. This dose-response relationship was also found to obey apparent Michaelis-Menten kinetics. 5. IDENTIFICATION OF DNA ADDUCTS A number of studies have been performed with the goal of identification of DNA and RNA adducts formed after VC exposure. The adducts of nucleic acids that have been con sistently identified are shown in Fig. 3. As described below, in vitro adducts were prepared for adenosine, cytosine and guanosine. Also, m vivo DNA adducts have been found in mouse liver after i.p. injection and in the liver, lung, kidney and brain of rats after inhalation or drinking water expo sure. The major DNA adduct formed m vivo was 7-(2-oxoethyD-deoxyguanosine (7-OEdG); there were smaller amounts of etheno adducts formed from deoxyadenosme, deoxycytosine and deoxyguanosine (Fig. 3), Osterman-Golkar et al. (1977) reported the isolation of 14C-labeled 7-{2-hydroxyethyl)guanine from hydrolysates of DNA from the livers of BALB and CBA mice exposed to [1,2-I4C] VC. The authors concluded that the alkylation product originally present m the DNA was 7-OEdG. Green and Hathway (1978) found 3,N4-ethenodeoxycytidine (edC) and 1 ,N6-ethenodeoxyadenosine (edA) in liver DNA from male and female Wistar rats that had been ex posed to VC in their drinking water (250 ppm) for approxi mately 2 years, beginning on the 27th day of life. These adducts were separated by liquid chromatography and iden tified by mass spectrometry. Laib et al. (1981) assessed the potential of [1,2-14C] VC to form in vitro and in vivo DNA adducts. When calf thymus DNA was incubated with VC, rat liver microsomes and NADPH, edC, edA and 7-OEdG adducts were identified. In vivo, they found only the 7-OEdG adduct by liquid chromatography of rat liver DNA after a 5-hr inhalation exposure of the rats to VC. The au thors reported that l,N6-ethenoadenosine (eAdo) and 3,N4-ethenocytidine (eCyo) were the major alkylation products with RNA, in vivo and in vin-o. In a later study, Laib et al. (1985a), using similar procedures, were able to detect the NL3-ethenoguanine (eG) adduct in liver DNA of 12-day-old rats after 2 daily exposures to l4C-VC, Berg 15 man (1982) identified VC-DNA adducts in liver and de tectable amounts of radioactivity in the DNA of spleen, kidney, lung, pancreas and testis of male NMR1 mice after a single i.p. injection of 25 p.Ci 14C-VC. DNA was hydro lyzed and separated by liquid chromatography. In the liver, which had the greatest amount of VC radioactivity, edA adducts were possibly identified from DNA hydrolysates; a major unknown adduct, probably 7-OEdG, was also de tected. Binding to RNA was observed in the spleen, pan creas, liver, kidney, lung and testis, but not in brain. In RNA of liver and kidney, a large part of the radioactivity was incorporated as C.| fragments, but significant amounts of eCyo in the kidney and eA in liver and kidney were identified. Using specific monoclonal antibodies, Eberle et al, (1989) detected edA and edC adducts in liver and lung DNA from Sprague-Dawley rats exposed to 2000 ppm VC by inhalation for 10 days. The DNA was hydrolyzed, and the adducts were separated by HPLC. Cirousse! et al. (1990) exposed 7-day-old and 13-week-old BD VI rats to 500 ppm VC 7 hr/day, 7 days/week for 2 weeks. Using HPLC and competitive radioimmunoassay with murine monoclonal antibodies, edA and edC were identified in DNA hydrolysates of liver, lung and brain, but not in the kidneys of the young rats. In the older rats, only the liver was analyzed, and levels of each adduct were 6 times tower than for the young tats. The authors suggested that their data were in good agreement with the organotropism and age-related sensitivity of VC-induced carcinogenicity in ro dents. DNA adducts were identified in the liver, lung and kidney of preweanling Sprague-Dawley rats exposed to 600 ppm VC 4 hr/day for 5 days (Fedtke et al., 1990; Swenberg et al., 1992). HPLC and fluorescence detection were used to analyze 7-OEdG. eG was determined by isotope dilution mass spectrometry, and edC and edA were determined by competitive radioimmunoassay using monoclonal antibod ies, after separation by HPLC. At the end of the 5-day pe riod, the liver had 3- to 8-fold higher amounts of each of the adducts than lung or kidney. No DNA adducts were identified in brain or spleen. The 7-OEdG adduct repre sented 98% of the adducts detected immediately after dos ing; the eG and edC adducts were present at about 1% and 0.6%, respectively, and the edA adduct occurred at about 0.1%. The persistence of adducts was determined m livers 0,3,1 and 14 days poscexposure. The 7-OEdG adduct had a half-life of 62 hr, the eG adduct had a half-life of 30 days, and the edC and edA adducts were unrepaired. As a result, at the end of 14 days, all of the detectable adducts were of the etheno type. These data show that the etheno adducts, although formed much less readily, are poorly recognized by repair enzymes and thus accumulate. Reaction of calf thymus DNA with CEO was found to produce adducts in the order 7-OEdG > eAdo > eG by analysis of fluorescent properties (Guengerich, 1992). CEO was found to be much more effective than CAA in the pro duction of eAdo and eG, and no 7-OEdG was produced from CAA. Epoxide hydrolase almost completely blocked the formation of eAdo by VC, and either calf thymus DN A AS! 000013998 16 or adenosine, in the presence of rat liver microsomes and NADPH. Glutathione (10 mM) reduced the formation of eAdo in similar conditions by 76% and 61% from adenosine and DNA, respectively. This experiment shows that CEO (and not CAA) is the proximate DNA-reacnve species and that protective mechanisms can prevent adduct formation. 6. GENETIC ALTERATIONS BY VINYL CHLORIDE-DNA ADDUCTS Although 7-OEdG is the major DNA adduct formed, etheno adducts have been found to miscode during DNA replication and transcription, and they are presumed to be involved in VC-induced cancer. However, there is conflict ing evidence concerning the specific etheno adduct and mutation involved. In an attempt to deduce the importance of the 7-OEdG adduct, Politzer et al. (1986) suggested that it might be in equilibrium with a cyclic hemiacetal, and, if this were the case, that hydrogen bonding between guanine and cytosine would be affected. Computational studies of the energies of both the adduct and its hemiacetal sug gested a very low probability that hemiacetal formation was a factor in prevention of hydrogen bonding between the bases. Therefore, it was concluded that the 7-OEdG adduct probably would not interfere with G-C base pairing. Krzyzosiak et al. (1986) used conformational analyses to de termine the mutagenic potential of etheno adducts. The orientation of adenosine to its sugar was more greatly af fected by erheno-adduct formation than was the orienta tion of cytidine to its sugar. In the case of adenosine, eAdoadduct formation changed the sugar pucker from C3 endo to C2 endo and the glycosyl torsion angle from 10 to 26. eCyo-adduct formation resulted in essentially no change to the sugar pucker and a change of glycosyl torsion angle from 18 to 10. This is significant because base mispainng can occur as bases change their glycosyl torsion angles, making hydrogen-bonding sites unavailable. Singer et al. (1991) synthesized the N:,3-ethenodeoxyguanosine (edG) adduct and found that it was incorpo rated opposite T at a greater rate than opposite C during DNA replication of a synthetic oligonucleotide template. The frequency of eG to thymine mismatch was similar with all polymerases tested: Escherichia colt DNA polymerase 1 (Klenow fragment), exonuclease-free Klenow, Drosophila melanogaster polymerase a and human immunodeficiency virus-I reverse transcriptase. The authors concluded that the replicating enzymes apparently recognize the same structural features, and on replication, G to A transitions would occur, resulting in GC to AT mutations. The mu tagenic and genotoxic properties of l,N6-ethenoadenine (eA), 3,N4-ethenocystine (eC) and a product formed by ring opening of eA, 4-amino-5-(imidazol-2-yl)imidazole, were investigated by Basu et al. (1993), who inserted these adducts into acceptor oligonucleotides; they were then li gated at a known site into the genome of the bacteriophage M13-Nhel. After transfection of the adducted phage DNAs into E. coli, each of the adducts was found to be genotoxic. ? J. Whvsner st al The most toxic lesion was 4-amino-5-(imidazol-2-yl)imidazole, which reduced survival of the genome by 97%. eC and eA reduced survival of . coli by 90% and 65%, respec tively. In this system, the least mutagenic of the adducts was eA, which caused primarily A--G transitions in 0.1% of survivors. The eA rearrangement product 4-amino-5-(imida;ol-2-yl)imidazole induced mutations at a 20-fold higher frequency ("j2%). eC induced mutations at a frequency of 1.5-2%; the mutations were mainly C--sT transitions, al though targeted C--sA and -- I deletions were also de tected. The mutagenic potency of the exocyclic adduct edC in E. coli and in cultured simian kidney (COS) cells was compared by Moriya et al. (1994). A single-stranded shuttle vector containing an edC lesion was introduced; after replica tion, the DNA sequence corresponding to the adduct :,tc ..as analyzed. The mutation frequency for single-stranded DNA that contained edC was 2% in nonirradiated E. coli, 32% in E. coli that had been preirradiated with UV light, and 81% in COS cells, indicating that the same lesion in bacteria and mammalian celts causes strikingly different effects. The pri mary mutations generated in both E. coli and COS cells were edC--A and edC-->T base substitutions. Oligonucle otide duplexes of a defined sequence containing one eAdo were synthesized by Oesch et al. (1994) and used as substrates to study the repair of this DNA lesion in cell homogenates of peripheral mononuclear blood cells from workers exposed to VC. The rate of repair of this lesion in preparations from the exposed workers did not differ from the rate of repair of the identical eAdo lesion in cell homogenate preparations from lymphocytes of nonexposed control subjects. 6.1. Studies of Oncogenes and Tumor-Suppressor Qenes Investigations were performed at codons 12, 13 and 61 of c-Ha-ras, c-Ki-ras and N-ras in angiosarcoma samples taken from VC polymerization plant workers (Marion ec al , 1991). A GC to AT transition at base 2 of codon 13 in the c-Ki-ros oncogene was the only mutation detected in five of six samples. A mutation of this nature leads to substitution of aspartic acid for glycine in the resulting p21 protein, which is a common ammo acid substitution found in many human cancers with mutated codon 13 in the Ki-ros onco gene. In a subsequent study from this group, 8 of 9 (89%) VCexposed workers with ASL were found to have serum posi tive for immunoblotting with monoclonal antibodies to p2l mutant protein (DeVivo et al., 1994). Also, 22 of 45 (49%) of VC-exposed workers with no evidence of ASL were also positive for p2l mutant protein by this mono clonal antibody assay. Consequently, this particular VCinduced mutation appears to be an early event in the neo plastic process and may be associated with the initiation of carcinogenesis. An investigation of hepatocellular carcino mas and ASL induced in Sprague-Dawley rats by VC inha lation exposure did not have similar oncogene mutations (Froment et al., 1994). Female rats with their pups were ex posed 8 hr/day, 6 days/week to 500 ppm VC from days 3 through 42 postpartum. DNA of various oncogenes was an alyzed by allele-specific oligonucleocide hybridization, di- AS I 000013999 ner et al. idaeC and respecadducts n 0.1% `-5-(imhigher ency of ms, alUo deict edC 11s was shuttle replicasite was .! DNA 32% m 31% in rta and he pri'S cells mucte_ eAdo hstrates rates of xposed a^tom of rations /enes 1 61 of . taken et al., m the five of itution rotein, many onco.) VCi posilies to of 45 f ASL monor VCc neoion of rcinotnhaations Vinyl Chloride Mechanistic Data rect sequencing of polymerase chain reaction products and sequencing after cloning. No mutations in codons 12, 13 or 61 of the Ki-ros oncogene, in codon 12 of the Ha-ros onco gene or in codon 61 of the N-ras oncogene were found. However, 2 of 5 resulting ASLs contained codon 13 (GGC to GAC) and codon 36 (ATA to CTA) mutations of the N-ras A oncogene, and 2 of 2 hepatocellular carcinomas contained an AT co TA transversion at base 2 of codon 61 in the Ha-ras oncogene. Mutations in the p53 tumor suppressor gene from cancers of factory workers exposed to VC were reported by Hollstein et al. (1994). In 2/4 ASL and 1 hepatocellular carcinoma stud ied, AT to TA transversions were detected in the highly con served regions at codon 249 and codon 255, respectively. 7. GENOTOXICITY STUDIES WITH VINYL CHLORIDE Genotoxicity tests have shown mostly positive results; most in vitro tests required metabolic activation (Table 1). Addi tionally, some in vitro tests without activation were posi tive, indicating possible direct reactivity of VC with DNA or a low level of oxidation of VC in these systems. DNA damage was found with VC in the E. colt polymerase A as say by Rosenkrant: and Letfer (1980) and in the Bacillus subtilis assay by Elmore et al. (1976). Shimada et al. (1985) found that VC induced unscheduled DNA synthesis in pri mary hepatocytes derived from rats. In general, bacterial re verse mutation assays showed that VC was positive in Ames strains TA100, TA1530, and TA1535, but negative in strains TA1537 and TA1538, which respond only to agents causing frameshift mutations (Table 1). Some of the Ames tests were positive without activation. Other reverse mutation testing, with E. coli K1Z, with CHL/V79 cells at the HPRT and ouabain gene loci and with Chinese hamster ovary cells, was positive provided that there was metabolic activa tion (Greim et al., 1975; Drevon and Kuroki, 1979; Krahn et al., 1982). Gene conversion testing in Saccharomyces cerevisiae was positive with activation m tests that were reported by Loprieno et al. (1976) and by Eckardt et al. (1981). Recessive le thal testing of D meknogaster was positive in two sets of tests conducted by Verburgt and Vogel (1977) and by Magnusson and Ramel (197$). Negative results were found in dominant lethal in vivo testing in the rat and mouse (Anderson et al., 1976; Short etal., 1977; Himenoetal., 1983). Positive micronucleus results were reported in CBA and C57 Bl/6] mice (Jenssen and Ramel, 1980; Richardson et al., 1983) and in human lymphocytes (Fucic et al., 1990). There were nine positive and three negative tests for chromosomal ab errations in lymphocytes from humans exposed to VC, which are listed in Table 1. In the most recent study, Fucic et al. (1990) investigated VC-induced chromosomal damage among PVC factory workers who had been employed for an average of 15 years and were matched by age to controls. Chromosomal aberrations, micronuclei and sister chroma tid exchange (SCE) frequencies showed statistically signifi cant increases in workers exposed to VC compared with the 17 control group. If the workers were also smokers, SCE were elevated still further, but the micronucleus test and chro mosomal aberrations were not affected. Some evidence in dicated that the elevated SCE frequencies were present for up to several years after exposure (Fucic et al., 1992). Aneuploidy testing in D. mekmogaster was negative (Verburgt and Vogel, 1977; Ramel and Magnusson, 1979). SCE in human lymphocytes showed mixed results in vivo. Hansteen (1979) found only negative results, while Kucerova et al. (1979) reported a weak positive result. Anderson et al. (1981) obtained a negative result when the test was done in vivo and could get a positive test in vitro with activation. 8. TOXICITY, CELL PROLIFERATION AND ALTERED FOCI The sequence of events leading to the development of ASL after exposure to VC and the relationship between hepato cytes and sinusoidal cells have been carefully detailed (Feron et al., 1979; Wisniewska-Knypl et al., 1980; Sokal et al., 1980; Feron et al., 1981; Maltoni et al., 1984). Major proliferative changes occurred in both the hepatocytes and the endothelial cells of the liver after animals were exposed to VC. In the hepatocytes, proliferation of the smooth en doplasmic reticulum, the appearance of swollen, abnormal mitochondria in which the ctistae were arranged radially around the periphery of the organelle, or the formation of irregularly shaped mitochondria were the earliest indices of hepatocellular damage after administration of VC (Feron et al., 1979). Proliferation of the endoplasmic reticulum was noted after 3 months' exposure to 50 ppm VC, while swell ing of mitochondria with reduction of glycogen deposits in hepatocytes occurred after being exposed to 50 ppm VC 5 hr/day, 5 days/week for 10 months or after 3 months' expo sure to 500 ppm VC (Wisniewska-Knypl et al., 1980). The liver parenchymal cells typically became polymorphic, with hyperchromatic, pleomorphic nuclei. Small droplets of fat were deposited in the cytoplasm of parenchymal cells, and the rough endoplasmic reticulum degenerated with loss of ribosomes to the surrounding cytoplasm. Sokal et al. (1980) administered VC to male Wistar rats beginning at 2 months of age 5 hr/day, 5 days/week for 10 months. No sta tistically significant changes in liver cells were found after exposure at 50 ppm. However, at 500 and 20,000 ppm, 13/43 (38%) and 8/17 (47%) of animals, respectively, showed proliferation of the hepatic sinusoidal endothelium. At these levels there was also an increase in nuclear poly morphism of hepatocytes. Hyperplastic changes and dyspla sia accompanied the appearance of basophilic preneoplastic foci and liver nodules. The hepatocytes of foci exhibited re duced glucose-6-phosphatase histochemical staining, while the sinusoidal cells located at the periphery of hepatic lob ules had increased acid phosphatase (Feron et al., 1979; Sokal et al., 1980). Fatty changes in parenchymal cells and the formation of pockets of cellular necrosis further exacer bated the disruption of the normal cvtoarchitecture in tis sue sections of liver as sinusoidal cells infiltrated the paren- ASI 000014000 18 TABLE 1. Genotoxicity of Vinyl Chloride Species/Strain DNA Interaction DNA Adduce Formation Calf Mouse/NMRl Rac/Wistar Rat/W istar Rat Rat Rat/BD VI Rat/Sprague-Dawiey DNA Damage E. colt B. subtilis Unscheduled UNA Synthesis Rat Mutagenicity Assays Reverse Mutation Salmonella typhimunum Salmonella typhtmurium Salmonella typhimunum Salmonella typhtmurium Salmonella typhtmurium Salmonella typhtmurium Salmonella typhtmurium Salmonella typhtmurium Salmonella typhtmurium Salmonella typhtmurium Salmonella typhimunum Salmonella typhtmurium Salmonella typhimunum Salmonella typhimunum Salmonella typhimunum Salmonella typhtmurium E. coh Gene Mutation Saccharomyces cerevisiae Saccharomyces cerevisiae Hamster Hamster Hamster Dominant Lethal Mutation Mouse Mouse Rat Sex-Linked, Recessive Lerhal D melanogaster D, melanogaster Cytogenetic Assays Chromosomal Aberrations Human Human Human Human Human Human Human Cell/Strain Thymus DNA Liver Liver Liver Liver Liver, lung Liver, lung, brain Liter, lung, kidney Pol A Rec Primary hepatocyte TA1535 TA1537, 1538 TA1530 TA1535 TA100, 1535 TA1535 TA100 TA100 taioo TA98, 100, 1535, 1538 TA100, 1535 TA1C0 TA1530 TA1530 TA1530 TA1530 K 12 CHL/V79/HPRT CHL/V79/ouabain CHO Lymphocytes Lymphocytes Lymphocytes Lymphocytes Lymphocytes Lymphocytes Lymphocytes J. Whysner et al In Vivo test Results Without activation With activation Reference no1 ND; yes4 + yes + yes yes + yes -1yes 4" yes + no ND no no 4- no no ND no + no ND no + no + no no no no ND no ND no ND no ND no ( + ) no ND no no ND no no no no no - yes yes yes yes 4yes 4- yes yes 4* yes 4yes + yes yes 4yes 4- +' Laib etal,, 1981 ND Bergman, 1982 ND Ostertnan-Golkaretol,, 1977 ND Green and Hathway, 1978 ND Laih et al, 1981 ND Eberle et at , 1989 ND Ciroussel et al., 1990 ND Swenberg er al , 1992 4" Rosenkrant: and Leifer, 1980 ND Elmore et al , 1976 ND Shimada et al , 1985 + - + + 4~ 4" ND + ND - 4" ( + )6 + 4444- 4+ 44* 4- ND ND ND ND ND Rannug etal., 1974 Rannug et al., 1974 Bartschetal., 1975 Bartsch etal., 1975 McCann et al., 1975 Andrews et al-, 1976 Elmore et al,, 1976 Simmon et al., 1977 Laumbach etal., 1977 Anderson and Styles, 1978 Jones and Hathway, 1978 Bartsch et al., 1979 Bartsch et al., 1979 Poncelet et al.. 1980 Hallstrdm etal , 1981 Duverger-Van Bogaert etal , 1982 Greim et al., 1975 Loprieno et al., 1976 Eckhardr et al, 1981 Drevon and Kurokl, 1979 Drevon and Kuroki, 1979 Ktahnetal , 1982 Anderson et al., 1976 Htmeno et al . 1983 Short etal , 1977 Verburgt and Vogel, 1977 Magnusson and RameL 1978 ND Purchase etal., 1975 ND Ducatman et al., 1975 ND Funes-Cravioto et al , 1975 ND Kilian and Picciano, 1976 ND Kucerova, 1976 ND Srentesi etal., 1976 ND Purchase et al., 1978 (Li'mnut'J) AS I 000014001 Vinyl Chloride Mechanistic Data 19 TABLED Continued Species/Strain Cell/Strain Human Human Human Human Human Micronucleus Mouse Mouse Human SCEs Human Human Human Human Human Aneuploidv D. meianogaster D. metanogaster Lymphocytes Lymphocytes Lymphocytes Lymphocytes Lymphocytes Lymphocytes Lymphocytes Lymphocytes Lymphocytes Lymphocytes Lymphocytes 1 no indicates test was not done in amo 2ND indicates test was not done 3 + indicates a positive test result. yes indicates test was done in two. 5 - indicates a negative test result. 6 (+) indicates a weak positive test result. In Vivo test yes yes yes yes yes Results Without activation With activation ND ND -1" ND - ND + ND Reference Fleig and Theiss, 1978 Hansteen ctal., 1978 KucerovS ec al., 1979 Anderson ei al , 1980 Fucic ec al , 1990 yes 4* yes -f- yes + ND Jenssen and Rjmel, 1980 ND Richardson n al , 1983 ND Fucic eral., 1990 yes yes ( + ) no yes yes + ND Hansteen, 1979 ND Kucerova et al.. 1979 + Anderson at al , 1981 ND Anderson etal., 1981 ND Fucic et al , 1992 no - ND Verburgt and Vogel, 1977 no -- ND Ramel and Magnusson, 1979 chyma. The sequence of pathological changes in the endothelial cells lining the sinusoids has been described by Maltom et al. (1984). Initially, there was hyperplasia that may have been accompanied by dysplasta of the endothelial cells. Varying degrees of fibrosis and sinusoidal dilatation may accompany the hyperplasia and dysplasia of the sinuso idal lintng cells, leading to a cavernous appearance of some hemangiosarcomas. The angiosarcomas were usually de scribed as multicentric with various patterns, such as cav ernous, vascular or anaplastic. At low concentrations such as 1 and 5 ppm by inhalation in Sprague-Dawley rats, none of these proliferative changes were evident m either the pa renchymal or sinusoidal cells (Maltom et al., 1984). Ac 10 ppm and above the incidence of proliferative changes in creased, but it was not noted to be >25% in any of the groups. Because these results were not obtained early in the stuyly, they are difficult to interpret. Laib at al. (1985b) investigated the VC-induced devel opment of phenotypically altered foci in the liver. These authors exposed male and female Wistar rats to 2000 ppm VC 8 hr/day, 5 days/week by inhalation for varying time pe riods beginning m utero (exposure of pregnant females) or beginning at postnatal day 7 or 21. The livers of the rats at 4 months of age were evaluated for ATPase-deficient foci. The number of ATPase-deficient foci was not increased af ter exposure to VC in utero or during postnatal days 1-5, but highly significant increases in foci areas were observed when newborn rats were exposed for 11 or 17 days; the inci dence was not further enhanced by a 47- or 83-day expo sure. The authors concluded that the period of sensitivity to VC-induced altered foci was characterized by rapid liver growth. This may also indicate the requirement for P450 activation, which is absent in the very young rat. In an other study, Laib et al. (1985c) studied dose-response of the VC-induced, ATPase-deficient foci in rat liver. Newborn Wistar rats were exposed to 0, 10, 40, 70, 150, 500 and 2000 ppm VC 8 hr/day, 5 days/week for 10 weeks. Also, male and female Sprague-Dawley and Wistar rats were ex posed to 2.5-80 ppm VC starting at 3 days of age for 3 weeks and were then maintained for 10 weeks without fur ther dosing. In both sets, there was a linear relationship be tween dose and area of induced foci expressed as percent of liver area. The authors concluded that within the dose range investigated, there was no evidence for the threshold value of 50 ppm, which had been previously suggested in the literature (Gehring ec al., 1978). There was, however, evidence for saturable kinetics; this was seen when the area of foci increased linearly in the range 2.5-500 ppm, with increasing dose until it reached a plateau at 500 ppm in the 10-week-exposure regimen m female Wistar rats. There was a slight increase in the female/male foci-induction ratio in Wistar rats in the dose-response protocol, but there was a much greater increase in this ratio using the Sprague-Dawley rat. No-observed-effect-levels at 10 ppm for Sprague-Dawley rats and ac 5 ppm for Wistar rats were found. However, the results appeared to be consistent with a linear dose-respQnse at low doses due to the small number of foci and variability. The authors concluded that there was no threshold. The AS I 000014002 20 relevance of dose-response of liver foci in the pathogenesis of ASL needs co be clarified. 9. SUMMARIES 9.1. Rodent Bioassays and Human Studies VC offers one of few examples in which both roden: bioassays and human epidemiological studies find the same type of rumor, and this tumor type does not occur appreciably in unexposed rodents or humans. Numerous epidemiological studies have found relatively high incidences of ASL re lated to VC exposure compared with the rare occurrence of this tumor in the general population. SMRs for ASL in these studies ranged from 400 to several thousand. No other tumor types m humans have been convincingly demon strated to be associated with VC exposures (Doll, 1988). Brain tumors were found elevated in two studies, and other tumor types have been found in only one study each: malig nant melanomas, lympho- and reticulosarcomas, and malig nancies of the lymphatic and hematopoietic systems. As re viewed by Doll (1988), one of the studies reporting increased brain cancer incidence (Monson et al., 1975) was included in the larger analysis of the United States cohort, which found no clear trend for latency or the expected in verse relationship for age of first exposure. Also, the inclu sion of the study by Byren et al. (1976), which found an in crease in brain tumors, into the larger European cohort (SimonatoetaL, 1991) resulted in no increase in brain tumors. In contrast to humans, the large series of rodent bioassays by Maltoni demonstrated nine types of tumors that were consid ered by the authors to be correlated with VC dosing in the rat: ASL, extrahepatic angiosarcomas, mammary adenocarcino mas, Zymbal gland carcinomas, nephroblastomas, hepatomas, neuroblastomas, forestomach papillomas and squamous cell carcinomas. For certain types of tumors such as mammary gland, there appeared to be a decrease in incidences at higher doses. A possible explanation, based upon examina tion of the survival curves from those studies, is that at high doses, the animals succumbed to tumors with shorter laten cies such as ASL, thereby decreasing the chance of observing a tumor with a longer latency, such as a mammary tumor. As shown by Maltoni et al. (1984), the route of adminis tration appeared to be a significant factor in the organotro pism of VC. When ingested, VC appeared to affect the liver preferentially, producing both hepatocellular carcino mas and angiosarcomas. Hepatocellular carcinomas were only found when VC was incorporated in the diet; VC in duced hepatocellular carcinomas at lower doses and a mix ture of hepatocellular carcinomas and ASL at higher doses. When given by gavage, VC produced ASL and liver hyper plasia, but no significant increases in adenomas or hepato cellular carcinomas. When injected, VC showed only a borderline tumorigenic effect, probably due to the low con centrations of VC used. In comparison with adults, 12-dayold embryos and newborn rats were much more likely to de velop ASL, hepatocellular carcinomas and neuroblastomas (Maltoni et al., 1984; Maltoni andCotti, 1988). J. Whysner et al. 9,2. Mechanistic Information The weight of evidence suggests that CEO, the epoxide of VC formed m the presence of microsomal P450-dependent monooxygenase, is the primary DNA-reactive species (Guengerich, 1992). CEO rapidly rearranges nonenzymatically or enzymatically in the presence of epoxide hydrolase to form CAA. Conversion of CEO, before it reacts with nucleic ac ids, to CAA converts the DNA alkylating species to a less DNA-reactive compound; CAA also has a higher affinity for protein than for DNA or RNA. This evidence is consis tent with the finding that CAA is not tumorigenic (Guengerich, 1992). The detoxifying enzyme epoxide hy drolase and glutathione have substantial protective effects on the rates of etheno-adduct formation by VC (Guenger ich, 1992). The preferential reactivity of CEO with DNA and the reactivity of CAA with protein were also reported by Gwinner et al, (1983), and they found that precursors of CEO, but not CAA, were capable of forming preneoplastic foci in liver. Reticuloendothelial cells lining liver sinusoids are almost devoid of microsomal mixed-function oxidase, in contrast to parenchymal cells. Guengerich et al. (1981) demonstrated that both CEO and CAA leave rat hepatocytes after the cells have been incubated with VC. Guen gerich (1986), therefore, proposed that reactive intermedi ate metabolites of VC are produced in the hepatocytes, which then bind to critical molecules of the sinusoidal cells where they exert their effects, causing cellular transforma tion. The reason for primary involvement of the sinusoidal cell rather than the hepatocyte m the tumorigenic process may be due to a greater degree of sensitivity to adduct for mation or due to poor repair mechanisms of sinusoidal cells. The DNA adduct formed in the greatest amount during metabolism of VC is 7-OEdG, with minor amounts of edC, edA and edG also being detected after in vivo exposure to VC. The 7-OEdG adduct and all three etheno adducts have been identified by Swenberg et al. (1992) in rat liver, lung and kidney. Other investigators have also identified the edC and edA adducts in the liver, lung and brain (Ciroussel et al., 1990). One difficulty with these experimental studies in liver is that DNA adducts identified may be associated with hepatocytes, endothelial cells or both. Consequently, it is not possible to definitively link one of these adducts to ASL by information on specific DNA-adduct formation. Studies have implicated the edG adduct as the likely candidate for the particular mutations found in ASL in hu mans. The major adduct formed by VC, 7-OEdG, does not cause a direct mutagenic event because the adduct is posi tioned in the major groove of an undistorted B-DNA helix where no base pairing occurs. In contrast, cyclic etheno ad ducts require reaction with the exocyclic ammo group and an adjacent ring nitrogen on guanine, cytosine or adenine and thus, are hindered from forming when DNA is doublestranded and hydrogen bonded, resulting in less adduct for mation. However, these etheno adducts have mutagenic potential because, once formed, they prevent normal base pairing. GC to AT transitions, which have been reported as AS! 000014003 aer et al. >xide of endent (Guentally or to form leic ac0 a less affinity consisrigenic ide hyeffects lenger1 DNA .-ported rsors of plastic msoids lase, in (1981) tepatoGuenrmediocytes, il cells aa- dal rocess ct for1 cells, during t edC, ure to s have , lung id the roussel rudies ciated lently, icts to likely in hues not posihelix 10 adp and enine uble:t for- ft ! i i I | | | : ' j j i i ! ; 1i Vinyl Chloride Mechanistic Data VC-induced mutagenic changes, appear to result from mis coding of the eG adduct. This adduct is more likely to base pair with thymine than with its normal pairing partner, cy tosine, thus initiating the GC to AT transition. Such tran sitions in codon 13 of the c-Ki-ros oncogene have been found in ASL tumors of VC-exposed workers (Marion et ai, 1991). The sera of workers with ASL have been found to concain mutated p21 protein, which is the oncogene product. Also, workers with VC exposure, but no evidence of ASL, have sera containing this mutated oncogene product. Consequently, VC appears to initiate malignant transfor mations by a mechanism that involves eG-adduct formation, mutation and oncogene activation. Although oncogene ac tivation has also been found in ASL and hepatocellular car cinomas from rats exposed to VC, the types of mutations and oncogenes involved are different from ASL in humans. Evidence of VC-DNA-adduct formation and genetic al teration is reflected m the positive results of genotoxicity tests, as shown in Table 1. The result of a test for unsched uled DNA synthesis in rat liver was positive. VC-induced mutagenic effects in bacterial systems were found both with and without the presence of a metabolic activation system. This may represent a low level of direct reactivity of VC or a low level of nonenzymaric production of CEO in the test system. The results of strains TA1537 and TA1538 have been negative, which suggests that the VC adducts formed were unable to produce the large conformational changes required for frameshift mutagenesis. In mammalian cells with metabolic activation in vitro, positive tests for gene mutation were reported. Peripheral lymphocytes containing chromo somal aberrations, micronucleus and SCE were found among humans exposed in the workplace. However, dominant lethal rests were negative, and most tests for SCE were negative. Studies of VC-induced cytotoxicity and cell proliferation have shown that a reaettve proliferative response in the si nusoidal endothelial cell may play a role in the induction of ASL (Feron et ai, 1979; Wismewska-Knypl et aL, 1980; Sokal et al., 1980; Feron et ai, 1981; Maltoni et ai., 1984). The dose-response of proliferative change occurs within a similar range as the tumorigenic response, in chat no signif icant tumors or proliferative responses are induced up to 10 ppm by inhalation. Proliferative responses occur in the pa renchymal cell,,and the development of VC-related altered hepatic foci have also been described. However, no signifi cant cumongenic response occurs m rhe hepatocyte unless VC has been administered in the diet. In this case, the tumongenic response is primarily in the hepatocyte at low doses, but it involves the sinusoidal endothelial cells as well at higher doses. 9.3. Species-Co-Species Extrapolation Incidences of ASL were found by Maltoni et al. (1984) to be comparable between different strains of rats and mice, although mice appeared to be more sensitive at lower doses. The hamster was relatively insensitive to VC-induced ASL. Studies by Drew et al. (1983) in F344 rats and Swiss mice 21 gave incidences approximately 4-5 times higher than those of Maltoni et al. (1984). This may be partially explained by the use of younger animals by Drew et ai (1983), which were 8-9 weeks of age vs. the 13-week-old animals used in the experiments by Maltoni et al. (1984). Also, the Drew et al. (1983) experiments had a 50% greater daily exposure period. The experiments by Lee et al. (1978) in CD-I rats also produced much higher ASL incidences per unit dose than the Maltoni et al. (1984) experiments. Almost onehalf of these rats showed tumors at the end of 12 months of exposure to 1000 ppm VC. Consequently, the dose-response curve shown in Fig. 2 does not represent the most sensitive strain or species, and at a dose of 100-1000 ppm VC, other studies have shown incidences of ASL approximately 5 times higher. Neither the mouse nor the hamster had VC-induced Zymbal gland carcinomas, nephroblastomas or neuroblasto mas, and, in addition, the hamster did not develop mam mary carcinomas or hepatomas. The several mouse strains tested, i.e., Swiss, B6C3F! and CD-I (Lee et al., 1978; Hehir etal., 1981; Drew eta/., 1983) were much more prone to develop mammary tumors and lung tumors than were rats. Strain variability was also evident, as the Wistar rats had much lower rates of Zymbal gland carcinomas and some what lower rates of ASL than the Sprague-Dawtey racs. Such variability of tumor types other than ASL from spe cies to species would indicate that extrapolation from ani mal data to estimates for human populations could be sub ject to considerable uncertainty. However, ASL was found to be a major tumor type in all three species and in the var ious strains of these species studied. Most importantly, there appears to be a good correlation between the ex pected dose-response in rodents compared with humans. There are several possible methods for comparing the re sults of human and animal studies, which have been re ported in the literature. These correlations will be described in the next section. 9.4. Dose-Response Extrapolation Maltoni et al. (1984) demonstrated that VC was a dosedependent animal carcinogen, and in inhalation experi ments, exposure time, exposure route, animal age, species and strain affected ics carcinogenic potential. The ages of animals exposed co VC greatly affect the dose response for ASL. In studies reported by Maltoni etal. (1984), newborn animals exposed for only 5 weeks developed approximately the same incidence of ASL as did IT to 13-week-old ani mals exposed for 52 weeks. Drew et al. (1983) reported total hemangiosarcomas at all sites in rats and found chat if 12-month exposures were deferred for 6 or 12 months, the reductions in tumor incidences weie about 2- and 10-fold, respectively. In the Maltoni et al. (1984) Sprague-Dawley rat experiments, a reduction of total exposure time by twothirds decreased the incidence of ASL from 12-22% to less than 2%. These authors also found that decreasing the number of exposure periods lowered the incidence of sev- ASI 000014004 11 J. Whysner et al eral ocher VC-related tumors, but this apparently had little effect on the incidence of Zymbal gland cumors. The evaluation of dose-response information necessarily VC-reactive metabolite per unit surface area and that the exposure of rats for 12 months approximates exposure of humans for 35 years (half a lifetime). Presuming a time- focuses on ASL resulting from inhalation exposure because weighted average human exposure of 200 ppm, the 8-10 this is the tumor type and exposure route that occurs in both rodents and humans. Overall, the test results by Maltoni et al. (1984) demonstrated a dose-response relation ship between concentration and tumor response for ASL. cases predicted by the models compared favorably wtth the 5 recorded angiosarcoma cases in a cohort of 9677 exposed workers from an unpublished study quoted by these authors. Swaen et al. (1987) found that results by Maltoni et al. These data are presented in Fig. 2. Data have been plotted (1984) demonstrated that an incremental increase in ASL on this graph with the dose scale as the log(dose). This of 1% was associated with an increase of 29 ppm VC ad dose-response curve shows chat there is an apparent linear ministered for 52 weeks in Sprague-Dawley rats. These au relationship between log(dose) and ASL tumor incidence. thors also estimated that in the combined epidemiological Such a relationship corresponds co a supralinear dose-response experiments, an average SMR of 1300 for liver cancer was curve. At two doses, 1 and 5 ppm, no apparent ASL was associated wich an exposure duration of 8.7 years to 500 found; however, this is consistent with the low probability ppm VC. After accounting for differences in the period of predicted and the number of test animals. exposures, these authors concluded that the death rates for The studies examining the metabolism and protein bind comparable exposures were within a factor of three between ing of VC have been found to follow Michaelis-Menten ki the rodents and humans. Humans appeared to be somewhat i' netics, exhibiting a saturation at high doses (Gehring et ai, less sensitive to the effects of VC than rodents. A more 1978). The dose-response of metabolic activation, there elaborate comparative approach was used by Chen and ii fore, would be represented by a supralinear curve. This data Blancato (1989), which incorporated physiologically based was also found co correlate to the tumor incidence data pharmacokinetic models for humans and animals. For this from the dose-response studies in Sprague-Dawley rats. effort, conversion of the inhalation data to a metabolized However, below the saturation point, normal first-order ki dose was used. Metabolic constants for rodents were ob netics have also been found to apply. The saturation point tained from the literature, and those for humans were de has been reported to be approximately 250 ppm in male rived from primate data. The authors concluded from this Wistar rats (Filser and Bolt, 1979). At 250 ppm, che first or analysis that whereas the number of cases predicted by the der rate of clearance of VC in Wistar rats is 11.0 L/hr/kg animal model on a per surface area basis was within a factor body weight; the zero order Vmax at concentrations higher of two, the prediction would be 20-fold less on a per body than 250 ppm is 110 pmol/hr/kg body weight. Although weight basis. However, it should be noted that this analysis several mathematical mechods have been used to correlate was based upon only two cases of ASL reported by Fox and animal data to human data and to correlate metabolic data Collier (1977). to animal daca, their usefulness in predicting extrapolation Another approach is to use the inhalation dose-specific to low doses is limited. Because the relationship of adminis tumor rate in rodents for comparison with the human tu tered dose to delivered dose is linear and first-order below mor rate, based upon the tumor incidence among a derived the saturation of 250 ppm, administered dose provides a vir number of exposure years for exposed workers. The epide tually identical low-dose extrapolation compared with de miological study from which such estimates can be made livered dose (Krewski et al., 1987). Several different mathe was reported by Theriault and Allard (1981). In this study, matical models are capable of fitting the dose-response data there were 451 workers who had exposures for 5 or more in Sprague-Dawley rats, and all provide similar comparisons years, and during the period of study, 10/451 (2.3%) work to the human epidemiological data (Gehring et al., 1979). ers were diagnosed with ASL. Based upon the information The use of these various models provides, however, a wide reported in this study, using observation period and expo variation in che prediction of human risk at low doses (Pur sure years, the estimates of numbers of workers exposed in chase et al., 1987). various years can be made (Table 2). From the estimates of Sfveral approaches have been used to compare the dose- exposure by Barnes (1976) during those years, the total response for ASL from the Sprague-Dawley VC inhalation number of VC ppm-exposure years can be derived, which 1 studies with epidemiological results. Four models (probit %, when divided by the number of workers and by an exposure 'i linear %, linear forced through origin and one-hit model) duration of 35 years results in an average yearly exposure of were used to empirically fit the dose-related tumor inci 240 ppm/worker. The 35-year-exposure duration has been dence data (Gehring et al., 1979; Schumann et ai, 1982). selected to be one-half of a person's lifetime, which would Using the data of Maltoni et al. (1984), there was no clear correspond to the 52-week duration in the rat studies by benefit in using any one of the methods to model the doseresponse relationship for ASL. All four models were then Maltoni et al. (1984). Although the rat studies had only a 4-hr/day exposure period, this duration would be within the used to predict the human angiosarcoma incidence from rat angiosarcoma data. In their extrapolations from rat data to range of the workers' exposure, which would probably occur only during part of an 8-hr work day. The dose-response of predictions of human cases, che investigators assumed that ASL in these studies has been summarized in Section 2.1.1, rhe induction of angiosarcoma was related to the amount of and the Maltoni et al. (1984) experiments are shown in Fig- AS I 000014005 it ad. time: 8-10 th the < posed ithors. et al. i ASL C adse auogical er was 0 500 tod of tes for tween ewhar more a and based >r this olized e obre deii this 7 the m alysis x and ecihc n turived pidemade 'tudy, more vorkation xpozd in :es of total htch isure re of been ould s by ily a 1 the >ccur Vinyl Chloride Mechanistic Data 2. The ASL incidence that corresponds to 237 ppm in these studies is about 7%. As indicated previously, a 2.3% ASL incidence in the cohort of Canadian workers was found. However, the average age of the workers at the time of the study was about 54 or 55 years of age, and the lifetime ASL incidence among this group would be expected to be some what greater than 2.3%. Consequently, the risk estimates for humans, based upon the Maltoni et al. (1984) studies and using various exposure comparison methods, appear to be quite close to those found in human epidemiological studies, even without making adjustments for pharmacoki netic parameters. Examination of the dose-response of proliferative changes was done by Maltoni et al. (1984); however, in most cases, the autopsies were done months atter the cessa tion of VC administration. There was an increase m prolif erative changes of sinusoidal cells noted especially at 50 ppm and above in Sprague-Dawley rats. Similar findings were not reported for Wistar rats. However, Sokal et al(1980) examined Wistar rats after 10 months of VC admin istration and found no increase in proliferation of sinusoi dal cells at 50 ppm, but did find an increased incidence of 38% at 500 ppm. This dose-response corresponds to the dose-response of angiosarcomas in Wistar rats reported by Maltoni et al. (1984), since there were none at 50 ppm and 10.7% at 500 ppm. Consequently, increased cell prolifera tion may play a role in the shape of the dose-response for VC-induced ASL. The dose-response study by Laib et al. (1985c) demonstrated a plateau at 500 ppm for the forma tion of altered hepatic foci in hepatocytes, thereby suggest ing the operation of saturable metabolic activation of VC. 10. USE OF MECHANISTIC DATA IN RISK ASSESSMENT VC is a DNA-reacttve carcinogen that requires bioaccivation to the reactive epoxide CEO (Fig. 4). CEO may spon taneously rearrange to an aldehyde, or VC may be metabo lized and eliminated by pathways that compete with epoxide formation. Several DNA adducts have been identi fied in the target organ for ASL. However, the major ad duct 7-OEdG does not appear to be mutagenic and, there fore, may not be responsible for tumor formation. The lesser 23 amounts of cyclic etheno adducts, which are also present in the livers of VC-exposed animals, are capable of producing mutations and, therefore, would be involved in the mecha nism of tumor formation (Fig. 3). Results shown in Table 1 indicate that VC genotoxicity studies were largely positive. In vivo DNA adducts were found in mouse liver after i.p. in jection and in liver, lung and brain of rats after inhalation and drinking water exposure. Murations produced by VCDNA adducts, especially of genes involving cell cycle regu lation, would be expected to produce initiated cells of the sinusoidal endothelium. There is also some evidence that VC induces proliferation of the sinusoidal endothelium, al though this response could be the result of mutation rather than an epigenetic effect. The dose-response data in rats for ASL (Maltoni et al., 1984), the cumor attributed to VC exposure in humans, was used by the United States EPA (1985a) to calculate an up per bound qT of 2.95 X 1CM (mg/kg/day)_l for the cancer risk due to inhalation exposures. The data of Feron et al. (1981) from the lifetime feeding study in Wistar rats was utilized by the United States EPA for calculation of a q,* for oral exposure of 2.3 (mg/kg/day)_1 based on the liver and lung tumor incidence (EPA, 1985b). The development of cancer potency factors derived from animal experiments assume species-to-species extrapolation, which has been the basis for most of the regulation of environmentally oc curring carcinogens in the United States. In the case of VC, human mortality studies demonstrating high inci dences of ASL associated with inhalation exposure m the workplace provide a basis for comparison with the inci dence of ASL from animal studies. Noc only does VC expo sure produce the same tumor type in both rodents and hu mans, but as documented in Section 9.4, the dose-related incidences are similar. Compared with the experiments by Maltoni et al. (1984) in Sprague-Dawley rats, other rodent species and strains have shown relatively small differences in dose-response relationships for ASL. Some studies of hamsters, rats and mice demonstrated up to a 5-told, and possibly even a 10-fold, lesser tumortgenic response. On the other hand, studies by other investigators have found up to a 5-fold greater tumor incidence for rats or mice at exposure levels that approximate those of exposed workers in the past, i.e., up to 1000 ppm. These differences between spe- TABLE 2. Calculation of Average Exposure Air Level for VC Workers Exposure Years (Yr) 1943-1947 1948-1952 1953-1957 Number of workers (N) ppm N X ppm X Yr (in IQ4) 103 1000 51.5 195 1000 97.5 304 750 114 1958-1982 289 400 58 Calculation Workers' Average Adjusted Exposure Level 3,750,000 workers - ppm 451 workers x 35 yr yr Data from Theriaultand Allard (1931). 237 ppm. 1963-1967 285 300 42.5 1968-1972 169 150 11.5 Total -- -- 375 ASI 000014006 24 cies are relatively small compared with those found for epi genetic carcinogens. The similar DNA-reactive cancer mechanism among species is modulated by variations in various metabolic parameters. Such variations may be ac counted for through the use of pharmacologically based pharmacokinetic modeling techniques. Young animals have been found to be more susceptible to VC-induced tumors resulting from DNA adducts, presumably due to greater rates of cell proliferation that would result in higher muta tion rates. Although several other target organs for VC-induced cancer have been found m rodents, these have not been demonstrated in humans. Among rodents, the larger differences in species and strain susceptibility for these tu mor types may explain the lack of evidence for these other tumor types in humans. Some of the mathematical models described in Section 4 4 have attempted to calculate dose-response expressions for low doses of VC. However, the dose-response similari ties between animals and humans are all in the range of oc cupational exposures. In other words, although these com parisons show that human incidences of ASL are similar to rats at similar exposure levels, they do not validate the use of a particular mathematical model for low-dose extrapola tions. The choice of a particular model for low-dose extrap olation for humans is still based upon mechanistic assump tions that predict a similar dose-response relationship as that found in experimental data on tumor formation. The linear-at-low-dose model used by the United States EPA assumes that at low doses, the dose-response for the supralinear curve is essentially the same as a linear relationship. In summary, ASL induced by VC has provided an oppor tunity to examine the extrapolation of animal to human data using a DNA-reactive carcinogen for risk assessment purposes. It has been found that there is relatively good agreement between dose-specific tumor incidences for ro dents and humans. Mechanistic information is consistent with a supralinear dose-response relationship, which is ef fectively linear at low doses. Consequently, although the linear-at-low-dose extrapolation method cannot be vali dated based upon human data, the well-documented hu man carcinogenicity of VC justifies the use of this doseextrapolation method. Arknowieci^merur-This manuscript has been developed under the guid ance of the international Expert Panel on Carcinogen Risk Assessment of ttte American Health Foundation. Memhers of the Panel who provided extensive review of this manuscript are Dry Helmut Greim (lnstuut fur Toxikologie, Neuherberg, Germany), James A. Swenherg (University of North Carolina) and Richard R. Munson (Harvard School of Public Health). The authors gratefully acknowledge support for this work trom the National Cancer Institute Grant *CA53160 and wish to thank Janet Marino and Robert Steward for their editorial assistance. References Anderson, D. and Styles, J. A. (1978) The bacterial mutation test. Br. J. Cancer 37: 924-930. Anderson, D., Hodge, M. C. E. and Purchase, I. F. H. (1976) Vinyl chloride: dominant lethal studies in male CD-I mice. Mutat. Res. 40: 359-370. ] Whysner et al Anderson, D, Richardson, C. R., Weight, T. M., Purchase, 1, F. H. and Adams, W. G. F. (1980) Chromosomal analyses in vinyl chloride exposed workers. Results from analysis 18 and 42 months after an initial sampling. Mutat. Res. 79: 151-162. Anderson, D., Richardson, C. R., Purchase, I, F H., Evans, H. J and O'Rtordan, M. L. (1981) Chromosomal analysis in vinyl chloride exposed workers: comparison of the standard tech nique with the sister-chromatid exchange technique. Mutat. Res. 83: 137-144- Andrews, A. W., Zawiscowski, E. S. and Valentine, C. R. (1976) A comparison of the mutagenic properties of vinyl chloride and methyl chloride. Mutac. Res. 40: 273-276. Barhm, A., Bresil, H., Croisy, A., Jacquignon, P., Malaveiile, C., Montesano, R. and Bartsch, H. (1975) Liver-microsome-mediated formation of alkylating agents from vinyl bromide and vinyl chloride. Biochem. Biophys. Res. Commun. 67: 596-603. Barnes, A. W. (1976) Vinyl chloride and the production of PVC. Proc. R. Soc. Med. 69: 277-281. Bartsch, H., Malaveiile, C. and Montaesano, R. (1975) Human, rat and mouse hver-mediaced mutagenicity of vinyl chloride in S. cyphimunum strains. Br J. Cancer 15: 429-437. Bartsch, H., Malaveiile, C., Barbin, A. and Blanche, G (1979) Mutagenic and alkylating metabolites of halo-ethylenes, chlorobucadienes and dichlorobutenes produced by rodent or human liver tissues. Arch. Toxicol. 41 (Suppl.): 249-277. Basu, A. K., Wood, M. L., Niedemhofer, L. J,, Ramos, L. A. and Essigmann, J. M. (1993) Mutagenic and genotoxic effects of three vinyl chloride induced DNA lesions. l,N6-ethenoademne, 3,N4-ethenocytosine and 4-ammo-5-(imidazol-2-yl)imidazole. Biochemistry 32: 12793-12801. Bergman, K. (1982) Reactions of vinyl chloride with RNA and DNA of various mouse tissues in two. Arch. Toxicol 49 (Suppl ): 117-129. Bi, W,, Wang, Y., Huang, M. and Meng, D. (1985) Effect of vmvl chloride on testis in rats. Ecoroxicol. Environ. Safety 10: 281-289, Bolt, H. M, and Filser, J. G. (1977) Irreversible binding of chlori nated ethylenes to macromolecules. Environ. Health Pcrspect 21: 107-112. Byren, D., Engholm, G , Englund, A. and Westerholm, P (1976) Mortality and cancer morbidity in a group of Swedish VCM and PCV production workers. Environ. Health Perspect. 17: 167-170. Chen, C. W. and Blancato, J. N. (1989) Incorporation of biologi cal information in cancer risk assessment: example--vinyl chlo ride. Cell Biol, Toxicol. 5: 417--444. Ciroussel, F., Barbin, A., Eberle, G. and Bartsch, H. (1990) Inves tigations on the relationship between DNA ethenobase adduct levels in several organs of vinyl chloride-exposed rats and can cer susceptibility. Biochem. Pharmacol. 39: 1109-1113 DeVivo, I., Marion, M, J , Smith, S. J., Carney, W P. and BrandtRauf, P. W. (1994) Mutant c-Ki-ros p21 protein in chemical carcinogenesis in humans exposed to vinyl chloride. Cancer Causes Control 5: 273-278. Doll, R. (1988) Effects of exposure to vinyl chloride. Scand. J Work Environ. Health 14: 61-78. Drevon, C. and Kuroki, T. (1979) Mutagenicity of vinyl chloride, vinytidene chloride and chloroprene in V79 Chinese hamster cells. Mutat. Res. 67: 173-182. Drew, R, T-, Boorman, G. A,, Haseman, J. K., McConnell, E E., Busey, W. M. and Moore, J, A. (1983) The effect of age and exposure duration on cancer induction by a known carcinogen in rats, mice and hamsters. Toxicol. Appl. Pharmacol. 68 L2C130. ASI 000014007  T "ter u al. 1 and 42 62. ->s, H. J. in vinyl J tech- Mucat. (1976) ide and die, C., e-medide and 6-603. >t PVC. i liman, 'ride in (1979) i, chloenc or 7. A. and L'Ct5 Of noadeyl)imi- f vinyl 1-289. chlorirspect. 1976) Vt and "-170. tologiI chlo- Invesidduct J can- randtmtcal lancer nd. J. onde, inster E. E,, e and Vinyl Chloride Mechanistic Data Ducatman, A., Hitschhom, K. and Selikoff, I. J. (1975) Vinyl chloride exposure and human chromosome aberrations. Mutat. Res. 31: 163-168. Duverger-Van Bogaert, M., Lambotte-Vandepaer, M., De Meester, C., Merciec, M. and Poncelet, F. (1982) Vinyl chloride and acrylonitrile: activation mechanism and mutagenicity. Toxicol. Eur. Res. 4: 35-37. Eberle, G., Barbin, A., Laib, R, J., Ciroussel, F,, Thomale, J-, Bartsch, H. and Rajewsky, M. F. (1989) l-(N6-etheno-2')deoxyadenosine and 3-(N4-etheno-2')-deoxycytidine detected by monoclonal antibodies in lung and liver DNA of rats exposed to vinyl chlortde. Carcinogenesis 10: 209-212. Eckardt, F., Muliawan, H., de Ruiter, N. and Kappus, H. (1981) Rat hepatic vinyl chloride metabolites induced gene conversion in the yeast strain D7RAD in vitro and m vivo. Mutat. Res. 91: 381-390. Elmore, J. D., Wong, J. L., Laumbach, A. D. and Stretps, U. N. (1976) Vinyl chloride mutagenicity via the metabolites chlorooxirane and chloroacetaldehyde monomer hydrate. Biochim. Biophys. Acta 442: 405-419. EPA (Environmental Protection Agency) (1985a) Health and environmental effects profile for chloroethene. ECAO-CIN-PI55. Office of Environmental Criteria and Assessment, Cincinnati. EPA (Environmental Protection Agency) (1985b) Drinking water criteria document for vinyl chloride, NT1S PB86-118320. Office of Drinking Water, Washington, DC. Fedtke, N., Boucheron, J. A.. Walker, V. E. and Swenberg, J. A. (1990) Vinyl chloride-induced DNA adducts. II. Formation and persistence of 7-(2'-oxoethyl)guamne and N:,3-ethenoguanine m rat tissue DNA. Carcinogenesis 11: 1287-1292. Feron, V. J, and Kroes, R. (1979) One-year time-sequence inhala tion toxicity study of vinyl chloride in rats. 11. Morphological changes in the respiratory tract, ceruminous gland, brain, kid ney, heart and spleen. Toxicology 13: 131-141. Feron, V. J., Spit, B. J., Immel, H. L. and Kroes, R. (1979) One-year time-sequence inhalation toxicicy study of vinyl chlonde in rats. III. Morphological changes in the liver. Toxicology 13: 143-154. Feron, V. ]., Hendriksen, C. F. M., Speek, A. J-, Til, H. P. and Spit, B. J. (1981) Lifespan oral toxicity study of vinyl chloride in rats. Food Cosmet. Toxicol. 19: 317-333. Fiber, J. G. and Bolt, H. M. (1979) Pharmacokinetics of hatogenated ethylenes in rats. Arch. Toxicol. 42 (Suppl.): 123-136. Fletg, I. and Thiess, A. M. (1978) Mutagenicity of vinyl chloride. External chromosome studies on persons with and without VC illness and on VC exposed animals. ]. Occup. Med. 20: 557-561. Fox, A. J. and Collier, P. F. (1977) Mortality experience of work ers exposed to vinyl chlonde monomer in che manufacture of polyvinyl chloride in Great Britain. Br. J. Ind. Med. 34: 1-10. Foment, O., Boivin, S., Barbin, A., Bancel, B., Trepo, C. and Marion, M. J. (1994) Mutagenesis of ras proco-oncogenes in rat liver tumors induced by vinyl chloride. Cancer Res. 54: 5340-5345. Fucic, A., Horvat, D. and Dimitrovic, B. (1990) Mutagenicity of vinyl chloride in man: comparison of chromosome aberrations with micronucleus and sister-chromatid exchange frequencies. Mutat. Res. 242: 265-270. Fucic, A., Garaj-Vrhovac, V., Dimitrovic, B. and Skara, M. (1992) The persistence of sister-chromatid exchange frequen cies in men occupationally exposed to vinyl chloride monomer. Murat. Res. 281: 129-132. Funes-Cravioto, F., Lambert, B., Lindsten, Ehrenberg, L., Nacarajan, A. T. and Osterman-Golkar, S. (1975) Chromosome aberrations in workers exposed to vinyl chloride. Lancet v. 459. 25 Gehring, P. J., Watanabe, P. G. and Park, C. N. (1978) Resolution of dose-response toxicity data for chemicals requiring metabolic activation: example--vinyl chloride. Toxicol. Appl. Pharma col. 44: 581-591. Gehring, P, J., Watanabe, P. G. and Park, C. N. (1979) Risk of angiosarcoma in workers exposed to vinyl chlortde as predicted from studies in rats. Toxicol. Appl. Pharmacol, 49: 15-21. Green, T. and Hathway, D. E. (1978) Interactions of vinyl chlo ride with rat liver DNA in uvo. Chem. Biol. Interact. 22: 211-224 Griem, H., Bense, G., Radwan, Z., Reichert, D. and Henschler, D, (1975) Mutagenicity m vitro and potential carcinogenicity of chlorinated ethylenes as a function of metabolic oxirane func tion. Btochem. Pharmacol. 24: 2013-2017. Guengerich, F. P. (1982) Metabolism of vinyl halides: in vitro stud ies on roles of potential activated metabolites. In: Biological Reactive Intermediates II. Proceedings Second International Symposium on Chemical Mechanism and Biological Effects, Part A, pp. 685-692, Snyder, R. (ed.) Plenum Press, New York. Guengerich, F. P. (1986) Covalent binding to apoprotein is a major fate of heme in a variety of reactions in which cyto chrome P-450 is destroyed. Biochem. Biophys. Res. Commun, 138: 193-198. Guengerich, F. P. (1992) Roles of che vinyl chloride oxidation products 2-chlorooxirane and 2-chloroacetaldehyde in the m vitro formation of etheno adducts of nucleic acid bases. Chem. Res. Toxicol. 5: 2-5. Guengerich, F. P. and Watanabe, P. G. (1979) Metabolism of [14C]- and [,6CIl-labeled vinyl chloride in vivo and in vitro. Bio chem. Pharmacol. 28: 589-596. Guengerich, F. P., Crawford, W. M. and Watanabe, P. G. (1979) Activation of vinyl chloride to covalently bound metabolites: roles of 2-chloroacecaldehyde. Biochemistry 18: 5177-5182. Guengerich, F. P., Mason, P. S., Stott, W. T., Fox, T. R. and Watanabe, P. G. (1981) Roles of 2-haloethylene oxides and 2-haloacetaldehydes derived from vinyl bromide and vinyl chloride in irreversible binding to protein and DNA. Cancer Res. 41:4391-4398. Guengerich, F. P., Kim, D. H. and Iwasaki, M. (1991) Role of human cytochrome P-450 11E1 in the oxidation of many low molecular weight cancer suspects. Chem. Res. Toxicol. 4: 168179. Gwinner, L. M., Laib, R. J., Filser, ]. C. and Bolt, H. M. (1983) Evidence of chloroethylene oxide being the reactive metabolite of vinyl chloride towards DNA: comparative studies with 2,2'dichlorodiethylether. Carcinogenesis 4: 1483-1486. Hallstrom, L, Sundvall, A., Rannug, U., Gtafstriim, R. and Ramel, C. (1981) The metabolism of drugs and carcinogens in isolated subcellular fractions of Drosophila melanogaster. I. Activation of vinyl chloride, 2-aminoantheracene and beruofajpyrene as mea sured by mutagenic effects in Salmonella cyphimunum. Chem. Biol. Interact. 34: 129-143. Hansteen, 1. L. (1979) A follow-up study of PVC workers two years after exposure. Preliminary results using sister chromatid exchange frequency an assay of genetic damage. In: Genetic Damage is Caused by Environmental Agents, pp. 279-285, Berg, K. (ed.) Academic Press, New York. Hansteen, I. L., Hillestad, L., Thus-Evensen, E. and Heldaas, S. 5. (1978) Effects of vinyl chlortde in man: a cytogenetic follow-up study. Mutat. Res. 51: 271-278. Heath, C. W., Falk, H. and Creech, J. L., Jr. (1975) Characteristics of cases of angiosarcoma of the liver among vinyl chloride workers in the United States. Ann. NY Acad. Set. 246: 231-236. ASI 000014008 26 Hefner, R. E., Watanabe, P. G. and Gehring, P. J. (1975) Percuta neous absorption of vinyl chloride. Toxicol. Appl, Pharmacol. 34: 529-532. Hehir, R. M., McNamara, B. P , McLaughlin, J., Willigan, D. A., Bierbower, G. and Hardiscy, J. F. (1981) Cancer induction fol lowing single and multiple exposures to a constant amount of vinyl chloride monomer. Environ. Health Perspect. 41: 63-72. Heldaas, S. S., Langard, S. L. and Andersen. A. (1984) Incidence of cancer among vinsl chloride and polyvinyl chloride workers. Br.J. Ind. Med. 41: 25-30. Himeno, S., Okuda, H. and Suzuki, T. (1983) Lack of dominant lethal effects in male CD-1 mice after short-term and long-term exposures to vinyl chloride monomer. Toxicol. Lett. 16: 47-53. Hoffmann, D,, Patrianakos, C. and Brunnemann, K. D (1976) Chromatographic determination of vinyl chloride in tobacco smoke. Anal. Chem. 48: 47-50. Hollsrem, M., Marion, M. Lehman, T., Welsh, J., Hams, C. C., Martel-Planche, G., Kusters, 1. and Montesano, R. (1994) p53 mutations at A:T base pairs in angiosarcomas of vinyl chloride exposed factory workers. Carcinogenesis 15: 1-3. Hong, C. B.t Winston, J. M., Thornburg, L. P, and Lee, C. C. (1981) Follow-up study on the carcinogenicity of vinyl chloride and vinylidene chloride in rats and mice: tumor incidence and mortality subsequent to exposure. J. Toxicol. Environ. Health 7: 909-924. IARC (1979) IARC Monograph of the Evaluation of the Carcino genic Risk of Chemicals to Humans, Some Monomer, Plastics and Synthetic Elastomers and Acrolein, Vol. 19. IARC, Lyon. Jenssen, D and Ramel, C. \ 1980) The micronucleus test as part of a short-term mutagenicity test program for the prediction of carctnogenicity evaluated by 143 agents tested. Mutat. Res. 75: 191-202. Jones, B. and Hathwav, D. E. (1978) Tissue-mediated mutagenic ity of vinylidene ohlocide in Salmonella ryp/ummuim TA1535. Cancer Lett. 5' 1-6. Jones, R. D , Smith, D. M. and Thomas, P. G. (1988) A mortality study of vinyl chloride monomer workers employed in the United Kingdom in 1940-1974. Scand. J. Work Environ. Health 14: 153-160. Kappus, H., Bolt, H. M , Butcher, A. and Bolt, W. (1975) Rat liver microsomes catalyse covalent binding of uC-vinyl chlo ride to macromolecules. Nature 257: 134--135. Kilian, D. J. and Picciano, D. (1976) Cytogenetic surveillance of industrial populations. In: Principles and Methods forTheir Detec tion, Vol. 4, pp 321-339, Hollaender, A. (ed.) Plenum, New York. Kruhn, D. F., Baryky, F C. and McCooey, K. T. (1982) CH0/ HGPRT mucjeion assay: evaluation of gases and volatile liq uids. Environ. Sci. Res. 25: 91-103. Krewski, D., Murdoch, D. J. and Withey, J. R. (1987) The applica tion of pharmacokinetic data in carcinogenic risk assessment. * In: Pharmacokinetics in Risk Assessment, Drinking Water and Health, Vol. 8, pp. 441-468, National Academy Press, Wash ington, DC. Krzyzosiak, W J., Wiewidrowski, M. and Jaskolski, M. (1986) Chemical modification of adenine and cytosine residues with chloroacecaldehyde at the nucleoside and che tRNA level: the structural effect of chloroacetaldehyde modification. In: The Role of Cyclic Nucleic Acid Adducts in Carcinogenesis and Mutagenesis, pp 76-81, Singer, 8. and Battsch, H. J. (eds.) IARC Scientific Publication No. 70, Lyon. Kucerova, M. (1976) Cytogenetic analysis of human chromosomes and its value for the estimation of genetic risk. Mutat. Res. 41: 123-130. J. Whysneret al. Kucerova, M-, Polivkovd, Z. and Batora. J. (1979) Comparative evaluation of the frequency of chromosomal aberrations and the SCE numbers in peripheral lymphocytes of workers occupa tionally exposed to vinyl chloride monomer. Mutat. Res 6797-100. Laib, R. J., Gwinner, L. M. and Bolt, H. M. (1981) DNA alkyla tion hy vinyl chloride metabolites, erheno derivatives of 7-alky lation of guanine.' Chem. Biol. Interact. 37. 219-231. Laib, R. J., Doerjer, G. and Bolt, H. M. (1985a) Detection of N-,3ethenoguanme in liver DNA hydrolysates of young rats after exposure of the animals to uC-vinyl chloride j. Cancer Res Clin. Oncol. 109: A7. Laib, R. J., Klein, K. P. and Bolt, H. M. (1965b) The rat liver foci bioassay: I. Age-dependence of induction by vinyl chloride of ATPase-deficient foci- Carcinogenesis 6: 65-68. Laih, R. J., Pellio, T. Wunsrhel I I. M., Zimmerman, N. and Bolt, H. M. (1985c) The rat liver foci hioassay: II. Investigations on the dose-dependent induction of ATPase-dettcient foci hy vmyl chloride at very low doses Carcinogenesis 6: 69-72. Laumbach, A. D., Lee, S , Wong, J. and Streips, U. N. (1977) Studies on the mutagenicity of vinyl chloride metabolites and related chemicals. In: Proceedings of the 3rd International Symposium on Prevention and Detection of Cancer, pp 155170, Nieburgs, H. E. (ed.) Marcel Dekker, New York. Lee, C. C., Bhandarl, J. C., Winston, J. M. and House, W. B. (1978) Carcinogenicity of vinyl chloride and vinylidene chlo ride. J. Toxicol. Environ. Healrh 4: 15-30. Lopneno, N-, Barale, R., Baroncelli, S., Bauer, C., Bronzectt, G., Cammellim, A., Cercignam, C., Cotsi, C., Gervasi, G., Leportni, C., Nieri, R., Rossi, A. M , Stretn, G. and Turchi, G (1976) Evaluation of the genetic effects induced by vinvl chlo ride monomer (VCM) under mammalian metabolic actuation1 studies in vitro and in vivo. Mutat. Res. 40: 85-96, Magnusson, J. and Ramel, C. (1978) Mutagenic effects ot \iml chloride on Drosophila melanogaster with and without pretreat ment with sodium phenobarbiturate. Mutat. Res. 57: 307-312 Maltoni, C. and o-otti, G. (1988) Carcinogenicicy of vinyl chlo ride in Sprague-Dawley rats after prenatal and postnatal expo sure. Ann. NY Acad. Sci. 534: 145-159. Maltoni, C., Lefemine, G , Cihbertt, A., Cotti, G. and Carretti, D. (1984) Experimental Research on Vinyl Chloride Carcinogen esis, Vol. II. Princeton Scientific Publishers, Inc., Princeton, Marion, M. J., Froment, O. andTrepo, C. (1991) Activation of Ki-ras gene by point mutation in human liver angiosarcoma associates with vinyl chloride exposure. Mol. Carcinogen. 4- 450-454. McCann, J., Simmon. V., Streitwieser, D- and Ames, B. N. ( 1975) Mutagenicity of chloroacetaldehyde, a possible metabolic prod uct of 1.2-dichloroethane (ethylene dichlonde), chloroethanol (ethvlene chlorohydrin), vinyl chloride, and cyclophospha mide. Proc. Natl. Acad. Sci. USA 72: 3190-3193. Monson, R. R., Peters. J. M. and Johnson, M. N. ( 1975) Propor tional mortality among vinyl chloride workers. Environ, Health Perspect- 1 1: 75--77. Monya, M,, Zhang, W., Johnson, F and Grullamn, A. P. (1994) Mutagenic potency of exocychc DNA adducts: marked differ ences between Eschenchia coll and simian kidney cells. Proc. Natl. Acad. Sci. USA 91: 11899-11903. Oesch, F., Weib, C. M. and Klein, S. (1994) Use of oligonucle otides containing ethenoademne to study the repair of the DNA lesion. Archiv Toxicol. 68: 358-363. Osterman-Golkar, S., Hulrmark, D., Segerback, D-, Calleman, C. J., Goche, R., Ehrenberg, L. and Wachtmeister, C. A. (1977) AS I 000014009 eret al. * >ccupales. 67: alkyla7-alky- >r'N-.3fs after er Res. er trci ride of d Bolt, ons on v vinyl (1977) es and itional 155- W. B. 1 chlo- ri, G , , Lephi, G. vinyl treat-312. chloexpo- rri, D. ogenn Ki-ras ciates 1975) ('rodli.inol 'pha- iporVulth 1994) lilfer1 'n ic ande- I the Vinyl Chloride Mechanistic Data Alkylation of DNA and proteins in mice exposed to vinyl chlo ride. Biochem. Biophys. Res. Commun. 76: 259-266. Pirastu, R., Comba, P., Reggiani, A., Foa, V., Masina, A. and Maitoni, C. (1990) Mortality from liver disease among Italian vinyl chloride monomer/polyvinyl chloride manufacturers. Am. J. Ind.Med. 17: 155-161. Plugge, H. and Safe, S. (1977) Vinyl chloride metabolism--a review. Chemosphere 6: 309-325. Pohtzer, P., Bar-Adon, R. and Zilles, B. A. (1986) The structure and properties of 7-(2'-oxoethyl)guanme: a model for a key DNA alkylation product of vinyl chloride. In: The Role of Cyclic Nucleic Acid Adducts in Carcinogenesis and Mutagens, pp. 37-43, Singer, B. and Bartsch, H. (eds.) [ARC Scientific Publications No. 70, Lyon. Poncelet, F , De Meester, C., Duvetger-Van Bogaert, M., Lambotte-Vandepaer, M., Roberfroid, M. and Mercier, M. (1980) Influence of experimental factors on mutagenicity of vinylic monomers. Arch. Toxicol. 4 (Suppl.): 63-66. Purchase, 1. F. H., Richardson, C. R. and Anderson, D. (1975) Chromosomal and dominant lethal effects of vinyl chloride. Lancet ii: 4L0--411 Purchase, [. F. H-, Richardson, C. R , Anderson, D,, Paddle, G. M. and Adams, W. G. F. (1978) Chromosomal analyses in vinyl chloride-exposed workers. Murat. Res. 57: 325-334. Purchase, 1. F, H., Stafford, J. and Paddle, G. M. (1987) Vmyl chloride: an assessment of the risk of occupational exposure. Food Chem. Toxicol. 25: 187-202 Ramel, C. and Magnusson, J. (1979) Chemical induction of nondisjunction in Drosophila. Environ. Health Perspect. 31: 59-66. Rannug, U., Johansson, A., Ramel, C. and Wachtmeister, C. A. (1974) The mutagenicity of vmyl chloride after metabolic acti vation. AMBIO 3: 194-197. Reynolds, E. S., Moslen, M. T-, Szabo, S., Jaeger, R. J. and Mur phy, S D. (1975) Hepatotoxicity of vinyl chloride and 1,1dichloroethylene. Am. J. Pathol, 31 1: 219-235. Richardson, C. R., Styles, J. A. and Bennett, I P. (1983) Activity of vmyl chloride monomer in the mouse micronucleus assay, Mutat. Res. 122: 139-142. Rinskv, R. A., Ott, G., Ward, E., Greenberg, H., Halperin, W. and Leet, T. (1988) Study of mortality among chemical workers m the Kanawha Valley of Wesc Virginia. Am. J. Ind. Med. 13: 429-138. Rosenkranz, H. S. and Leifer, Z. (1980) Determining the DNAmodifymg activity of chemicals using DNA-poiymerase-dehcient Escherichia coh. In: Chemical Mutagens: Principles and Methods for their Detection, Vol. 6, pp. 109-147, de Serres. F. J. and Hollaender, A. (eds.) Plenum Press, New York. Sabadie, N., Malavedle, C., Camus, A. M. and Bartsch, H. (1980) Comparison of the hydroxylation of benzo(a)pyrene with the metabolism of vinyl chloride, N-nitrosomorpholine, and N-nitroso-N-methylpiperazine to mutagens by human and rat liver microsomal fractions. Cancer Res. 40: 119-126. Schumann, A. M., Wacanabe, P. G., Reitz, R. H. and Gehring, P. J. (1982) The importance of pharmacokinetic and macromolecular events as they relate to mechanisms of tumorigenicity and risk assessment. In; Toxicology of the Liver, pp. 311-331, Plaa, G. and Hewitt, W. R. (eds.) Raven Press, New York. Shimada, T, Swanson, A. F., Leber, P. and Williams, G. M. (1985) Activities of chlorinated ethane and ethylene com pounds in the Salmoneilo/rat microsome mutagenesis and rat hepatocyte/DNC repair assays under vapor phase exposure con ditions. Cell Biol. Toxicol. 1: 159-179. 2` Short, R. D., Minor, J. L., Wmsron, J. M- and Lee, C. C, (1977) A dominant lethal study in male rats after repeated exposures to vinyl chloride or vmyhdene chloride, j. Toxicol. Environ, Health 3: 965-968. Simmon, V. F., Kauhanen, K. and Tardiff, R. G. (1977) Mutagenic activity of chemicals identified in drinking water. In: Progress in Genetic Toxicology, Vol. 2, pp. 249-258, Scott, D., Bridges, B. A, and Sobels, F. H. (eds.) Elsevier/North-Holland Biomedi cal Press, Amsterdam. Simonato, L., L'Abbe, K. A., Andersen, A., Belli, S., Comba, P., Engholm, G., Ferro, G., Hagmar, L., Langord, S , Lundberg, I., Pirastu, R., Thomas, P., Winkelmann, R. and Saracci, R. (1991) A collab orative study of cancer incidence and mortality among s myI chlo ride workers. Scand. J. Work Enuron. Health 17. 159-169 Singer, B., Kusmierek, J. T., Folkman, W., Chavez, F. and Dosanjh, M K. (1991) Evidence for the mutagenic potential of the vinyl chloride induced adduct, N:,3-etheno-deoxyguanosine, using a site-directed kinetic assay. Carcinogenesis 12: 745-747. Sokal, J, A., Barinski, B., Majka, J., Rolecki, R., Stetkiewicz. J., Ivanova-Chemishanska, L., Vergteva, T., Antonov, G-, Mirkova, E-, Kolakowski, J., Szendzikowski, S. and Wroblewska, K. (1980) Experimental studies on the chronic toxic effects of vinyl chloride in rats. J. Hyg. Epidemiol. Microbiol. Immunol. 24: 285-294. Suzuki, Y. (1978) Pulmonary tumors induced in mice by vinvl chloride monomer. Environ. Res. 16: 285-301. Swaen, G M. H., de Hollander, A. E. M., Kroes, R , den Engelse, L-, Feron, V. ]., Mulder, G. J., Verbeek, A. L, M., Verschuuren, H. G., Vogel, E. W. and van der Wielen, A. W. (L987) A sci entific basis for the risk assessment of vmyl chloride. RegulToxicol. Pharmacol. 7: 120-127. Swenberg, J- A., Fedtke, N., Ciroussel, F., Barbm, A. and Bartsch, H. (1992) Erheno adducts formed in DNA of vinyl chloride-exposed rats are highly persistent in liver. Carcinogenesis 13 727-729. Szentesi, L, Homyak, E., Ungvary, G., Czeizel, A., Bogndr, Z and Timar, M. (1976) High rate of chromosomal aberration in PVC workers. Mutat. Res. 37: 313-316. Tabershaw, 1. R. and Gaffey, W. R. (1974) Mortality study of workers in the manufacture of vinyl chloride and its polymers. J. Occup. Med. 16: 509-518. Teta, M. j., Schnatter, A R., Ott, M. G. and Pell, $. (1990) Mortality surveillance in a large chemical company: the Union Carbide Cor poration experience, 1974--1983. Am. J. Ind. Med. 17:435--147. Theriault, G. and Allard, P (1981) Cancer mortality of a group of Canadian workers exposed to vmyl chloride monomer. J Occup. Med, 23: 671-676. Til, H. P, Feron, V. J. and Immel, H. R. (1991) Lifetime (149week) oral carcinogenicity study of vinyl chloride in rats- Food Chem. Toxicol. 29: 713-718. Verburgt, F. G. and Vogel, E. (1977) Vinyl chloride mutagenesis in Drosophila meianogaster, Mucat. Res. 48: 327-336. Viola, P- L., Bigotti, A. and Caputo, A. (1971) Oncogenic response of rat skin, lungs and bones to vinyl chloride. Cancer Res. 31; 516-522. Watanabe, P. G., McGowan, G. R. and Gehring, P. J. (1976a) Fate of [UC1 vinyl chloride after single oral administration m rats. Toxicol. Appl. Pharmacol. 36: 339-352. Watanabe, P. G., McGowan, G. R., Madrid, E. O. and Gehring, P. J. (1976b) Face of [l4C] vinyl chloride following inhalation exposure m rats. Toxicol. Appl. Pharmacol. 37: 49-59. Watanabe, P. G., Zempel, J. A., Pegg, D. G. and Gehring, P J (1978) Hepatic macromolecular binding following exposure to vinyl chloride. Toxicol. Appl. Pharmacol. 44: 571-579. AS I 000014010 28 Waxweder, R. J., Stringer, W., Wagoner, J. K., Joners, J., Falk, H. and Carter, C. (1976) Neoplastic risk among workers exposed to vinyl chloride. Ann. NY Acad. Sci. 271: 40-A8, Waxweiler, R. J., Smith, A. H., Falk, H. and Tyrolet, H. A. (1981) Excess lung cancer risk in a synthetic chemicals plant. Environ. Health Perspect. 41: 159-165. Weber, H., Reinl, W. and Greiser, E. (1981) German investiga tions on morbidity and mortality of workers exposed to vinyl chloride. Environ. Health Perspect. 31: 95-99. Wisniewska-Knypl. J., Khmccak, J. and Kolakowski, J. (1980) Monooxygenase activity and ultrastructural changes of liver in the course of chronic exposure of rats to vinyl chloride. Int. Arch. Occup. Environ. Health 46: 241-249, J. Whysner ecd. Wong, O., Whorton, M. D., Foliart, D. E. and Ragland, D (199|) An industry-wide epidemiologic study of vinyl chloride work ers, 1942-1982. Am.J.Ind. Med. 20:317-334. Wu, W., Steenland, K., Brown, D., Wells, V., Jones, J., Schulte, P, and Halperin, W. (1989) Cohort and case-control analyses of workers exposed to vinyl chloride: an update. J, Occup. Med 31.518-523. Za|dela, F., Croisy, A., Barhin, A., Malaveille, C , Tomans, L. and Bartsch, H. (1980) Carcinogenicity of chloroethylene oxide, an ultimate reactive metabolite of vinyl chloride, and bis(chloromethyl)ether alter subcutaneous administration and in initiation-promotion experiments in mice Cancer Res. -Kv 352-356 ) ASI 000014011