Document vBoGRmLgR3QQqqQ0pKagR18E

JNSon, A., Campbell, M. A., Polybrominated biphenyls as ixylase inducers: Structure-acn.-Biol. Interact. 42, 53-66. . R. (1980). Covalent binding icnyls to rat liver microsomes tetive metabolites and target tl. Appl. Pharmacol. 55, 490- iBE, Y, (1983). Activation of :nyl to protein-bound metab:rosomal cytochrome P-448nase system. Toxicol. Appl. TOXICOLOGY AND APPLIED PHARMACOLOGY 78, 105-122 (1985) Delineation of the Role of Metabolism in Trichloroethylene and Perchloroethylene: ^ uooc-cnea atuay John A. Buben1 and Ellen J. O'Flaherty Department ofEnvironmental Health. University ofCincinnati. Cincinnati, Ohio 45267 Received April 30. 1984; accepted November 14. 1984 Delineation of the Role of Metabolism in the Hepatotoxicity of Trichloroethylene and Perchloroethylene: A Dose-Effect Study. Buben, J. A., and O'Flaherty, E. J. (1985). Toxicol. Appl. Pharmacol. 78, 105-122. The relationships among dose, metabolism and hepatotoxicity in mice which resulted from subchronic exposure to the chlorinated solvents trichloroethylene (TRI) and perchloroethylene (PER) were examined. Male Swiss-Cox mice received either TR1 (0 to 3200 mg/kg/day) or PER (0 to 2000 mg/kg/day) in com oil by gavage for 6 weeks. Urinary metabolites from individual mice were quantified to estimate the extent to which each compound was metabolized. Four parameters of hepatotoxicity were assessed: liver weight, triglycerides, glucose-6-phosphatase (G6P) activity, and SGPT activity. TRI significantly affected liver weight and G6P activity: PER affected all four parameters. The metabolism of TRI was linearly related to dose through 1600 mg/kg, but then became saturated. The metabolism of PER was saturable. The dose-effect curves of the affected hepatotoxicity parameters of both compounds were nonlinear and resembled the dose-metabolism graph of the corresponding solvent. Plots of the hepatotoxicity data of each compound against total urinary metabolites were linear in all cases, suggesting that the hepatotoxicity of both PER and TRI in mice is directly related to the extent of their metabolism. This pattern is consistent with formation of ihe toxic intermediate in the primary metabolic pathway of each compound. 1985 Academic Press. Inc. Vudies by the National Cancer Institute pound at high and low doses, and the rela lNCI) have demonstrated that mice treated tionship between dose-related pharmacoki ^ironically with high doses of the widely netic factors and toxic response. Insights into llSt`d industrial solvents trichloroethylene both of these relationships can be gained 1 RD and perchloroethylene (PER) had high from an examination of the pharmacokinetic incidences of hepatocellular carcinoma (Na- properties of such compounds. The impor nonal Cancer Institute, 1976, 1977). Attempts tance of pharmacokinetic analysis in inter " assess the carcinogenic risk to man of preting toxic response has been reviewed by -sposure to these and other compounds Andersen (1981). 'and to cause tumors in animals are being In most carcinogenicity bioassays only a aadc. However, such assessments have been few very high doses are used. The percentage nallenged because of lack of information of animals developing tumors from chronic i ncorning two important relationships: the treatment at these doses is determined. These ; aimnship between disposition of the com- data form the basis for assessment of the carcinogenic risk of the compound to man. vm address: Department of Medicinal Chemistry, 1 11 "i Pharmacv. University of Kansas. Lawrence, I'M I4S An extrapolation is made from the animal response data at high doses to the low doses more characteristic of human exposure. Such SL 032749 105 004I-008X/85 $3.00 Copyright |9H5 hv Academic Press, Inc. AH rights of reproduction in any form reserved jnson. A., Campbell, M. a., Polybrominated biphenyls as ixylase inducers: Structure-acn.-Biol. Interact. 42, 53--66. . R. (1980). Covalent binding enyls to rat liver microsomes active metabolites and target il Appl. Pharmacol. 55, 490- iBE, Y. (1983). Activation of nyl to protein-bound metab.rosomal cytochrome P~448nase system, Toxicol. Appl. TOXICOLOGY AND APPLIED PHARMACOLOGY 78, 105-122 (1985) Delineation of the Role of Metabolism in the Hepatotoxicity of Trichloroethylene and Perchloroethylene: A Dose-Effect Study John A. Buben1 and Ellen J. O'Flaherty Department of Environmental Health. University of Cincinnati. Cincinnati, Ohio 45267 Received April 30. 1984; accepted November 14. 1984 Delineation of the Role of Metabolism in the Hepatotoxicity of Trichloroethylene and Perchloroethylene: A Dose-Effect Study. Buben, J. A., AND O'Flaherty, E. J. (1985). Toxicol. Appl Pharmacol. 78, 105-122. The relationships among dose, metabolism and hepatotoxicity m mice which resulted from subchronic exposure to the chlorinated solvents trichloroethylene (TRI) and perchloroethylene (PER) were examined. Male Swiss-Cox mice received either TRI (0 to 3200 mg/kg/day) or PER (0 to 2000 mg/kg/day) in corn oil by gavage for 6 weeks. Urinary metabolites from individual mice were quantified to estimate the extent to which each compound was metabolized. Four parameters of hepatotoxicity were assessed: liver weight, triglycerides, glucose-6-phosphatase (G6P) activity, and SGPT activity. TRI significantly affected liver weight and G6P activity; PER affected all four parameters. The metabolism of TRI was linearly related to dose through 1600 mg/kg, but then became saturated. The metabolism of PER was saturable. The dose-effect curves of the affected hepatotoxicity parameters of both compounds were nonlinear and resembled the dose-metabolism graph of the corresponding solvent. Plots of the hepatotoxicity data of each compound against total urinary metabolites were linear in all cases, suggesting that the hepatotoxicity of both PER and TRI in mice is directly related to the extent of their metabolism. This pattern is consistent with formation of the toxic intermediate in the primary metabolic pathway of each compound. iss Academic Press Inc Judies by the National Cancer Institute 1VI) have demonstrated that mice treated vironically with high doses of the widely used industrial solvents trichloroethylene I I RD and perchloroethylene (PER) had high incidences of hepatocellular carcinoma (Na"Dnal Cancer Institute, 1976, 1977). Attempts II assess the carcinogenic risk to man of -\nobure to these and other compounds umd to cause tumors in animals are being ;'ade. However, such assessments have been 'ullenged because of lack of information morning two important relationships: the '`iionshtp between disposition of the com- nt address: Department of Medicinal Chemistry, `i Pharmacy. University of Kansas. Lawrence, I'M MS pound at high and low doses, and the rela tionship between dose-related pharmacoki netic factors and toxic response. Insights into both of these relationships can be gained from an examination of the pharmacokinetic properties of such compounds. The impor tance of pharmacokinetic analysis in inter preting toxic response has been reviewed by Andersen (1981). In most carcinogenicity bioassays only a few very high doses are used. The percentage of animais developing tumors from chronic treatment at these doses is determined. These data form the basis for assessment of the carcinogenic risk of the compound to man. An extrapolation is made from the animal response data at high doses to the low doses more characteristic of human exposure. Such 032750 105 004I-008X/X5 $3.00 Copyright 1985 by Academic Press, Enc All fights of reproduction in any torm reserved 106 BUBEN AND O'FLAHERTY an extrapolation may utilize any of a number of mathematical models (Fishbein. 1980), but each assumes that the pharmacokinetics of the compound in question obeys firstorder kinetics over the range of the extrapo lation. Such an assumption is rarely valid. A number of processes involved in the disposition of a compound, particularly the metabolic processes, are often dose depen dent. This dose dependency has particular importance since it is now well established that many chemicals known to be carcinogens are metabolized to reactive intermediates which can interact with cellular components to initiate the tumorigenic process (Miller, 1978; Weisburger and Williams, 1980). At high doses, the metabolism of many chemi cals is likely to be saturated, a result which can lead to shifts in the relative importance of metabolic pathways and to relative in creases or decreases in the toxic response. For this reason, the need to examine the pharmacokinetics of a compound when risk assessment or other evaluation of dose-re sponse relationships is to be undertaken has been stressed (Munro, 1977; Watanabe et al.. 1977; Andersen. 1981). Despite the problems associated with highdose data, little has been done to examine the effect of high doses on both metabolism and toxic response. TRI and PER are good models for such a study. Not only have both been shown to cause hepatic tumors in mice in NCI carcinogenicity bioassays, but both are believed to be metabolized through an epoxide intermediate, an intermediate often involved in initiation of carcinogenesis by other chemicals (Henschler, 1977: Van Duuren, 1975; Weisburger and Williams, 1980). Both compounds can also be eliminated un changed via the lungs, a nontoxic pathway (Daniel. 1963). However, despite these simi larities. there is evidence that the kinetics of TRI and PER metabolism differ significantly. The metabolism of TRI has been reported to be linearly related to dose, while that of PER is saturable (Ikeda et al.. 1972). This difference, together with their similar meta bolic pathways and their common site of action, makes them ideal candidates for studying the effect of metabolism on toxicity. The purpose of this study was to investigate the relationships among TRI and PER dose, metabolism, and hepatotoxicity in mice after subchronic exposure to these solvents, with particular interest in how these relationships are affected at high doses. A dosing format similar to those used in the NCI carcinoge nicity bioassays was followed. Studies by Schumann et al. (1980) on PER and by Stott et al. (1982) on TRI have suggested that the carcinogenicity of these solvents to mice oc curs via an epigenetic mechanism, involving repeated toxic insult to the liver as a prereq uisite to tumor formation. An understanding of the above relationships, therefore, could have significant bearing on an assessment of the carcinogenic potential of these solvents. METHODS Chemicals. Trichloroethylene and perchloroethylene (tetrachloroethylene) were obtained from the Aldrich Chemical Company. Trichloroethylene was distilled be fore use. Perchloroethylene had a purity greater than 99% and was used without further purification. Animals Male Swiss-Cox mice (outbred) were obtained by breeding from a colony at the Kettering Laboratory. When the mice were between 3 and 5 months of age, males from each litter were randomly assigned to each dose group and were tail marked. Three to five mice were generally housed together in stainless-steel cages with wire-grid bottoms. The mice had access to Punna Rodent Laboratory Chow at all times. Water was available ad libitum through an automatic watenng system. The room was maintained on a 12-hr light/dark cycle at 21 2C. Dosing. Dose mixtures were prepared fresh two or three times weekly by dissolving calculated amounts of the solvents in com oil. The mice were weighed three times each week; the weight of each mouse determined the amount of dose mixture it received. Mouse weights ranged from 34 to 45 g; dose volumes ranged from 0.18 to 0.24 ml. The mice were dosed by gavage five times each week for 6 weeks, a total of either 29 or 30 doses. Doses for PER were 0. 20. 100, 200. 500. 1000. 1500, and 2000 mg/kg/day. Those for TRI were 0. 100. 200, 400, 800, 1600, 2400, and 3200 mg/kg/day. Controls received com oil. Twelve to fifteen mice were used in most dose groups; the 100- and 3200-mg TRI/kg/day groups and the 1500- and 2000-mg PER/kg/day groups contain of 24 a- Prna hepatic liver gr liver tr pvruvat by cen blood s blood v centnfu inferior and the via the washed homogc 6.2). A. centrati homoge (1957) phosph. amount homoge M Tns20 min acid w, measure and is c gram ot bv the n albumin on 0.1 Frankel Burton Portu neutral The set stained histolog were pi. weight v Hepa varianct respecti Keuls p Urinl vidua) i into mi Manufa aval lab! lion wa; per moi Prelu to-week were a, housed and dai site of tes for )xicity. ,'stigate l dose, e after s, with nships format inogeies by y Stott iat the ce ocolving rereqnding could ent of ents. t- \ldnch led ber than named ratory. of age, 0 each mice cages Punna ailable i. The at 21 vo or nts of three nmed eights : 0.18 times loses. 1 MX), ZOO, ttrols ~d m /day METABOLISM-EFFECT STUDIES OF TRI AND PER 107 contained 4 to 6 mice. The two control groups consisted >f 24 and 26 mice. Procedure. Four parameters were chosen to assess nepattc toxicity: increases in liver weight, decreases tn uver glucose-6-phosphatase (G6P) activity, increases in iver triglycerides, and increases in serum glutamateivruvate transaminase (SGPT) activity. Mice were killed iy cervical fracture the day following their last dose. A 'lood sample was obtained by cardiac puncture. The 'lood was allowed to clot and the serum, obtained after entnfugation, was frozen until its assay for SGPT. The nfenor vena cava on the underside of the liver was cut, ,nd the liver was perfused with 5 ml of ice-cold saline ia the left ventricle of the heart. The liver was removed, cashed in cold saline, blotted dry, and weighed. It was uimogenized in 9 vol of 0.1 m Tris-maleate buffer (pH - 21. Aliquots were frozen until assay. Triglyceride con.-nitrations were determined on 1-ml aliquots of hepatic aimogenate by the method of Van Handel and Zilversmit 1957) as modified by Butler et at. (1961). Glucose-6'hosphatase activity was determined by measuring the imount of phosphate released after incubation of 0.1 ml Homogenate with 25 amol of glucose 6-phosphate in 0.1 q Tris-maleate buffer (pH 6.2. total volume = 1 ml) for 20 min at 37C. Four milliliters of 10% trichloroacetic .icid was used to stop the reaction. Phosphate was measured by the method of Fiske and Subbarow (1925) and is expressed as micrograms of phosphate per milli gram of protein. Protein concentration was determined bv the method of Lowry ei al. (1951) using bovine serum albumin as the standard. SGPT activity was determined on 0.1 ml of serum by the method of Reitman and Krankel (1957). DNA was measured by the method of Burton (1968). Portions of freshly removed liver were fixed in 10% neutral Formalin, embedded in paraffin, and sectioned. The sections were processed by standard procedures, stained with hematoxylin and eosin. and submitted for histological evaluation. Weighed portions of some livers were placed in an oven at 75C until a constant dry weight was obtained. Hepatotoxicity data were analyzed by analysis of variance, and comparisons between each dose and its respective control were done by the Student-NewmanK.euls procedure (Sokal and Rohlf. 1969). Unnary metabolites. During the dosing period, indi vidual mice from each of the dose groups were placed into mouse metabolism cages supplied by Wahmann Manufacturing Company. Feed and water were always available. After an acclimation day, a 24-hr urine collec tion was taken. Two to four unne samples were collected per mouse throughout the 6-week dosing period. Preliminary studies to investigate day-to-day and weekto-week variations in the pattern of urinary metabolites were also earned out. For these studies, mice were housed in metabolism cages throughout the dosing period and daily unne collections were made. Quantification of the urinary metabolites tnchloroethanol (TCE) and tnchloroacetic acid (TCA) was by the method of Humbert and Fernandez (1976). A Hew lett-Packard Model 7620A gas chromatograph equipped with a tritium electron capture detector was used for the analysis. The column, containing 5% OV-I7, was heated at 92C. Oxalic acid in urine was determined by the method of Hodgkinson and Williams (1972). A method reported by Rajagopal and Ramakrishnan (1975) was adapted to determine whether ethylene glycol was a unnary metab olite of PER. Data analysis. The hepatotoxicity data were plotted against both dose and total unnary metabolites. The data points were fit using the Nonhn computer program provided by the Upjohn Company (Metzler et al.. 1974), or by linear regression analysis, as appropnate. RESULTS Hepatotoxicity Mice tolerated the 6-week po dosing with either compound. Few deaths occurred except at the highest dose of each compound. These deaths appeared to be the result of central nervous system depression. Mice in all dose groups continued to gain weight throughout the 6-week dosing period. Both solvents caused dose-related increases in the liver weight to body weight ratio (Table 1). Since the body weights of the mice were generally unaffected by any of the treat ments, these increases represent true liver weight increases. Doses as low as 100 mg/ kg/day of each compound were sufficient to cause statistically significant increases in the liver weight/body weight ratio. The increases in liver size were attributable to hypertrophy of the liver cells as revealed by histological examination and by a decrease in the DNA concentration of the livers (Table 2). Mice in the highest dose group of each solvent displayed liver weight/body weight ratios which were about 75% greater than those of controls. PER caused a marked dose-related accu mulation of triglycerides in the liver (Table 1). Mice in the high-dose groups had six times as much triglyceride per gram of liver SL 032752 108 BUBEN AND O'FLAHERTY TABLE l Effect of Trichloroethylene and Perchloroethylene on the Relative Liver Weight, Hepatic Tri glyceride ConcentrariON. Glucose-6-Phosphatase (G6P) Activity, and Serum Glutamate Pyruvate Transaminase (SGPT) Activity of Mice after a 6-Weeks Exposure" Dose (mg/kg/day) Liver weight/body weight (%) Liver triglycerides (mg/g liver) G6P (Rg phosphate/mg protein/20 mm) SGPT (unns/0.1 ml serum) Trtchloroethylene 0 100 200 400 800 1600 2400 3200 Perchloroethylene 0 20 100 200 500 1000 1500 2000 5.22 0.09 (24) 5.84 0.20 (5)** 5.99 0.13 (12)*** 6.51 0.12 (12)*** 7.12 0.12 (12)*** 8.51 0.20 (12)*** 8.82 0.15 (12)*** 9.12 0.15 (4)*** 5.21 0.09 (26) 5 51 0.11 (13) 5.97 0.11 (13)*** 6.45 0.12 (15)*** 7.35 0.16 (15)*** 7.89 0.16 (19)*** 8.10 0.27 (6)*** 9 00 0 11 (6)*** 3.08 0.29 (24) 3.12 0.49 (5) 4.41 0.76 (12) 4.53 1.05 (12) 5.76 0.85 (12) 5.82 0.93 (12) 6.89 1.40(12)** 7.02 0.69 (4) 3.23 0.29 (26) 2.72 0.27 (13) 7.66 1.78 (13)* 13.21 2.28 (15)*** 22.69 1.61 (15)'** 25.68 1.54 (19)*** 24.52 2.59 (5)*** 20.10 2.08 (6)*** 125.5 117.8 *+- 116.4 117.3 111.7 -b 89.9 83.8 83.0 ~b 3.2 (12) 6.0(5) 2.8 (9) 4.6 (9) 3.3 (9)** 1.7 (9)*** 2.1 (8)*** 7.0 (3)*** 136.5 4- 3.1 (15) 132.0 7.6 (6) 120.4 7.7 (5) (18.4 12.4 (5) 110.4 -b 7.5 (5)** 101.5 4.1 (8)*** 94.5 2.1 (4)*** 99.2 4.0 (4)*** 11.8 1.3 (15) -- 11.3 2.3 (7) 11.7 2.5 (6) IU 1.3 (7) 14.1 2.1 (7) 31.5 6.8 (11)** 26.8 3.1 (4)* 11.8 1.0 (18) 11.3 0.7 (10) 11.8 1.1 (10) 15.3 1.9 (9) 28.4 2.7 (10)*** 44.6 5.3 (12)*** 44.1 5.3 (6)*** 46.4 5.6 (4)*** " Values are x t SE determined in (AO mice. * p < 0.05 from control group. " p < 0.01 from control group. *** p < 0.00) from control group. TABLE 2 DNA Content, Histopathological Evaluation, and Wet Weight/Dry Weight Ratios of Livers from Treated Mice Group DNA" (mg/g liver) Control 400 mg TRl/kg 1600 mg TRI/kg 200 mg PER/kg 1000 mg PER/kg 2.83 0 17 2.57 0.14* 2.15 0,08*** 2.52 0.27 2.36 0.21** Degeneration (r ++ +++ +++ Karyorrhexis + Necrosis _ + () + Polyploidy _ + + (i) + Wet wt/dry wt ratio* 3.45 0.12 3.49 0.07 3.48 0.15 3.23 0.13* 3.18 0.20* " x SD (n = 4 or 5). \\ SD (n = 5). ` Histological evalualions are graded: --. negative: + to ++++, increasing severitv of observed pathology. * p < 0.05. *' p < 0.01. *** p < 0 001. as cc little mice trigh valut of th of cc from sister all m PER pane of T nonr Be decrc (Tab abou TRI/ the h activ value SC treatt (Tab! half cause at 5C exist doses trans times mild parisi or th (160( totox sever j TRI j 70 ui i j Hixti Th were man 2. Lt hepa SL 032753 J L i ratic TriPyruvate -GPT !v'0.l ml rum) l 3 (IS) :.3 (7) 15 (6) 1.5 (7) 1! (7) s.S (11)** ' 1 (4)* 1.0 (18) > 7(10) :.i not : (9) ' ' (10)*** M M/drv i ratio* ' r 0.12 1 ^ 007 * 0.15 ; 0.13* - 0.30* METABOLISM-EFFECT STUDIES OF TRI AND PER 109 as control mice. In contrast. TRI had very little effect on hepatic triglycerides. Although mice in the highest dose groups had mean triglyceride concentrations twice the control value, the triglyceride concentrations in some of these mice were not elevated over those of controls. The gross appearances of livers from PER- and TRI-treated mice were con sistent with these results. The livers of almost all mice receiving 200 mg/kg/day or more of PER were pale and mottled in appearance, a Pattern characteristic of fatty liver. The livers of TRI mice, although enlarged, appeared normal. Both PER and TRI caused dose-related decreases in glucose-6-phosphatase activity (Table 1). These decreases were small, only about 10%, until the dose reached 800 mg TRl/kg or 500 mg PER/kg. Mice receiving the highest doses of TRI and PER had G6P activities which were 66 to 70% of the control value. SGPT activity was not increased in TRItreated mice except at the two highest doses (Table 1). Even at the 2400-mg/kg/day dose, half of the mice had normal values. PER caused significant increases in SGPT activity at 500 mg/kg/day. A threshold appeared to exist at 100 mg/kg/day. At the highest PER doses, SGPT activity averaged about 45 transaminase units per 100 u\ serum, four times the control value. This increase reflects mild to moderate hepatic damage. As a com parison, several mice were treated with two or three doses of CHC13 (400 mg/kg) or CC14 (1600 mg/kg), enough to cause severe hepatotoxicity; their SGPT activities ranged up to several hundred units per 100 pi. With either TRI or PER, SGPT activity rarely exceeded 70 units per 100 u\. Ilistopathology The livers of mice from several dose groups were histopathologically examined. A sum mary of the observations is shown in Table 2. Liver degeneration, manifested by swollen hepatocytes, was common in all four treat ment groups. Cells had indistinct borders: their cytoplasm was clumped and a vesicular pattern was apparent. The swelling was not simply due to edema, as wet weight/dry weight ratios did not increase (Table 2). Evidence of karvorrhexis, the disintegration of the nucleus, was present in nearly all specimens and suggested impending cell death. Central lobular necrosis was present in some specimens. Polyploidy was also char acteristic in the central lobular region. Hepatic cells had two or more nuclei or had enlarged nuclei containing increased amounts of chro matin. suggesting that a regenerative process was ongoing. Both solvents caused similar pathology, but it was generally more severe in the PER-treated mice. Also, fine droplets of lipid were commonly present in the cyto plasm of the hepatocytes of PER-treated, but not TRI-treated, mice. Preliminary Metabolism Studies Trichloroethanol and trichloroacetic acid were found in the urine of mice treated with TRI. Verification of the metabolites was ac complished by comparisons with known standards. TCE was the predominant metab olite, and much of it was conjugated with glucuronic acid. TCA generally accounted for between 15 and 30% of the total urinary metabolite. Urine from several mice was collected daily throughout the 6-week dosing period. The amount of urinary metabolite did not show significant day-to-day variability or week-to-week trends. Gas chromatographic analysis of urine from PER-treated mice identified TCA as the only metabolite. The possible presence of dechlorinated metabolites such as oxalic acid or ethylene glycol, reported as PER metabolites by other investigators (Pegg et al.. 1979; Daniel, 1963; Dmitrieva, 1967j, was investigated. Six control mice excreted 0.327 0.085 mg of urinary oxalic acid/day. Nine mice receiving PER doses ranging from 200 to 2000 mg/kg/day had 0.351 0.081 mg of oxalic acid/day in their urine. Therefore 032754 SL I 10 BUBEN AND O'FLAHERTY oxalic acid was eliminated as a kineticallv significant metabolite. Ethylene glycol also was not found. There were day-to-day trends in the amount of TCA excreted by mice dosed with PER. The amount of TCA excreted each day tended to increase, before leveling otf toward the end of the week. This pattern occurred because the metabolism and/or excretion of PER was slower than that of TR1. and 10 to 15% of the metabolized PER was not excreted within 24 hr. but was carried over to the following day. No week-to-week trends in TCA urinary excretion were observed. The influence of the day-to-day trend in urinary excretion of TCA was minimized by consis tently collecting urine from the mice on specific days late in the week. The contribution of fecal elimination of metabolites in mice dosed with either TRI or PER was examined. The amount of me tabolites in the feces of mice treated with either compound was less than 5% of the amount found in urine, and its contribution was neglected. TRI Metabolism A biphasic relationship between amount of urinary metabolite and TRI dose was observed (Fig. 1). The initial portion of this curve was linear through 1600 mg TRt/kg, before an abrupt plateau was reached. The linearity of this initial portion of the curve was verified by the fractional metabolism occurring at each dose. With the exception of the 200-mg/kg dose, 27.5% of each dose in the range 100 to 1600 mg/kg/day was converted into urinary metabolite. Above 1600 mg/kg the fraction of each dose metab olized decreased, suggesting that TRI metab olism approached saturation beyond this dose. PER .U, The hsm an` points gous to V = (.Wr pacitv-li the amo the max and exc case, tin tabolite apparen expressu 136 mg/ The decrease with cap of a vc but onb TRI Dose (mg/kg) I k;. I. Relationship between the TRI dose and the amount of total unnary metabolite excreted per dav by mice in each group. Values represent ,v SE. /V = 7 to 4 mice per group except lor the 100- and 1200-mg TRI/kg groups where A = 4 and 3. respectively. The slope a and rl value for the linear regression ht ol the 100- to 1600-mg/kg data points are given. 032755 > F. day and valu METABOLISM-EFFECT STUDIES OF TRI AND PER 111 / 27? Metabolism Effect of TRI Metabolism on Hepatotoxicity The relationship between PER metabom and dose is shown in Fig. 2. The data mts were fit with an equation analo gs to the Michaelis-Menten expression. " + X), which describes ca vity-limited kinetics. A' is the dose. }' is amount of metabolite resulting. A/ma, is -Hvt maximum amount of metabolite formed a Ad excreted in 24 hr. and K,,, is. in this Cji-x. the dose at which the amount of me tabolite excreted in a 24-hr period is half the Apparent maximum amount. The data tit the e<nression very well: r was 0.996. Mm3X was (36 mg/kg/day and K,, was 660 mg PER/kg. The fraction of each dose metabolized -creased with increasing dose, consistent nh capacity-limited metabolism. About 25% t a very low PER dose was metabolized, but only 5% of a very high dose. TRI significantly affected only two of the four hepatotoxicity parameters, liver weight and G6P activity. When dose was plotted on a linear scale against the percentage increase in liver weight, or against the percentage inhibition of G6P activity, biphasic curves exactly analogous to the TRI metabolism curve resulted (Figs. 3A and B). In each case, the relationship between dose and the toxic effect was linear up through 1600 mg TRI/ kg/day, reaching a plateau at high doses. However, when these same hepatotoxicity data points were plotted against the amount of metabolism (i.e., total urinary metabolites) at each respective dose, linear relationships were observed throughout the entire dose range (Figs. 4A and B). The points corre sponding to the plateau region of the doseeffect curves fall around the regression line PER Dose (mg/Xq) Fit.. 2. Relationship between the PER dose and the amount of total unnary metabolite excreted per day by mice in each group. Values represent v SE. A' = 9 to 11 mice per group except for the 1500and 2000-mg/kg doses where A' = 4 or 5. The points were fit by the Michaelis-Menten equation, and the values lor A/m,, and K,, are given, together with the r- value. SL 032756 TRI Dose (mg/kg) TRI Dose (mg/kg) Fig. 3. Dose-effect relationships between TRI dose and the hepatotoxicity parameters of (A) liver weight increases and (B) G6P inhibition. Indicated values are the slope a, intercept h, and r value for the linear regression of the 100- to 1600-mg/kg data points. in the metabolism-effect graphs. This indi cates that the nonlinearity seen in the doseeffect relationships can be explained by a change in the kinetics of TRI metabolism. Effect of PER Metabolism on Hepatotoxicity Graphs relating PER dose to each of the four hepatotoxicity parameters resulted in | i ii li hyper metat The . 03*1^1 sv METABOLISM-EFFECT STUDIES OF TRI AND PER 113 -t 3200 Total Urinary Metabolite (mg/kg) XT 3200 \) liver due for itotoxicitv ich of the Ated in Total Urinary Metabolite (mg/kg) Fig. 4. Relationships between the hepatotoxicitv parameters of (A) liver weight increases and (B) G6P inhibition and total urinary metabolite excreted per day by mice from the vanous TRI dose groups. Indicated values are the slope a, intercept h. and r- value of the linear regression of all points. hyperbolic curves analogous to the PER metabolism graph in Fig. 2 (Figs. 5A-D). The data points in each case were tit by the expression Y = {EmaxX)HK,, + A'), where Emax is the maximum effect, and the resulting Emax and K,, values, as well as the r1 values. SL 032758 114 BUBEN AND OFLAHERTY 30 r 25 ,2> Ol PER Dose (mg/kg) 4 00 40 (/> O 3C > <D aicno> o co CinO CcO 30 70 CL cao 10 0 I PER Dose (mg/kg) 1 Fig. 5. Dose-effect relationships between PER dose and the hepatotoxicity parameters: (A) liver weight increases: (B) G6P inhibition: (C) triglyceride increases; (D) SGPT activity increases. The points were fit with the Michaelis-Menten equation, and the values for m,, and A,,. and the r- value are given. The fit for C was based only on the 0- to 1000-mg/kg data points. (Figs. 6A-D). that the toxici'. amount of rr are presented on each graph. The Km values of three of the four hepatotoxicity parameters are similar to the K,,, of PER metabolism, a result which is expected if a relationship exists between metabolism and hepatotox icity. Graphs of the four hepatotoxicity param eters against metabolism result in linear tits etfect graph ot deviates from kg/day dose f the result of effect. Triglyc 03^759 Sl^ L METABOLISM-EFFECT STUDIES OF TRI AND PER 115 PER Dose (mg/kg) PER Dose (mg/hq) Fig. 5--Continued (Figs. 6A-D). These results strongly suggest that the toxicity of PER is also related to the amount of metabolism. The metabolismetfcct graph of hepatic triglycerides (Fig. 6C) deviates from linearity above the 1000-mg/ kg/day dose point. This deviation may be the result of achievement of a maximum effect. Triglyceride concentrations increase with PER dose up to 1000 mg/kg/day, but decrease at higher dose rates (Table 1). If a maximum in the total amount of hepatic triglycerides had been reached by the 1000mg/kg/day dose, the larger liver weights of the higher dose groups would result in a relative decrease in triglyceride concentration per gram of liver at these higher doses. 032760 SL 116 BUBEN AND O'FLAHERTY Total Urinary Metabolite (mg/kg) Total Urinary Metabolite (mg/kg) F:i(i. 6. Relationships between the hepatotoxicitv parameters: (A) liver weight increases. (B) G6P inhibition. (C) triglyceride increases, and (D) SGPT activity increases, and total urinary metabolite excreted per day by mice from the vanous PER dose groups. Data points were tit by linear regression (only the 0 to 1000 data points were used for C). and the slopes a, intercepts b. and r: values are indicated. SL 032761 Two ol a triglyce lipids (t; weight) di highest PI METABOLISM-EFFECT STUDIES OF TRI AND PER i 17 Total Urinary Metabolite (mg/kg) iB) G6P iK-tabolite regression llues are Total Urinary Metabolite (mg/kg) Fir,. b--Continued. Two observations support the concept of a triglyceride maximum: (1) total hepatic lipids (triglyceride concentrations X liver weight) did not greatly differ among the three highest PER groups; (2) mice receiving two or three doses of CC14 or CHC1, (1600 and 400 mg/kg/day, respectively), hepatotoxicants known to induce fatty liver, had hepatic triglyceride concentrations of approximately 24 1 mg/g liver, virtually identical to those SL 032762 118 BUBEN AND O'FLAHERTY seen in the mice receiving 1000 and 1500 mg PER/kg/day, suggesting that this value may be the upper limit of triglyceride accu mulation in this strain of mice. DISCUSSION The objective of this study was to examine the relationships among dose, metabolism, and hepatotoxicity in mice after exposure to either TRI or PER for 6 weeks to determine the pan that metabolism plays in the toxicity of these solvents. The linearity seen in all six metabolism-effect graphs, which relate the various hepatotoxicity parameters to the. amount of metabolism occurring at each dose of TRI and PER, indicates that the hepatotoxicity of both these compounds is directly proportional to the extent to which the solvents are metabolized. Such a result is consistent with the pro posed metabolic scheme of both compounds. Both TRI and PER are believed to be me tabolized by the mixed-function oxidase sys tem to epoxide intermediates, and then to stable terminal metabolites excreted mainly in the urine (Daniel, 1963; Henschier, 1977; Leibman and Ortiz, 1977; Yllner, 1961). TRI epoxide is converted into chloral hydrate, which is further metabolized to TCE and TCA by alcohol dehydrogenase and chloral hydrate dehydrogenase, respectively. PER epoxide is believed to convert spontaneously to trichloroacetyl chloride, which is rapidly hydrolyzed to TCA, The epoxide is believed to be the toxic intermediate in each case, reactive enough to bind to cellular macromolecules (Henschier, 1977; Van Duuren. 1975; Leibman and Ortiz. 1977; Allemand et al.. 1978). Covalent binding of both com pounds has been demonstrated (Van Duuren and Banerjee, 1976; Banerjee and Van Duuren. 1978; Bolt et al.. 1977; Uehleke and Poplawski-Tabarelli, 1977; Bolt and Filser, 1977; Allemand et al., 1978; Pegg et al.. 1979). Toxicity should parallel metabolism in any compound for which the reactive intermedi ate responsible for the toxic effect is part of the primary metabolic pathway, as is the case with TRI and PER. Although the data are consistent with the view that the metabolic pathways of both compounds involve the epoxide as the toxic intermediate, they are not restricted by it. Recently, Miller and Guengerich (1982, 1983) have proposed that the epoxide is not an obligate intermediate in TRI metabolism. The results of the present study demonstrate that formation of the re active intermediate, whatever its structure, must be proportional to the overall amount of metabolism, since when metabolism reaches a constant maximum value the tox icity also does not increase further. It should be pointed out that use of total urinary metabolites as the index of metabo lism is an approximation. Most of the TRI or PER which is metabolized through reactive intermediates is converted to either TCA or TCE and excreted in the urine. Very little is excreted in the feces. Some metabolized compound (both TRI and PER) is bound to macromolecules, and some is metabolized to carbon dioxide (Daniel. 1963; Pegg et al.. 1979; Schumann et al., 1980; Stott et al.. 1982; Parchman and Magee, 1982). To the extent to which these factors are involved, total urinary metabolite is an underestimate of the actual amount of metabolism. How ever, at the doses used, urinary excretion is by far the major route of disposition of metabolized compound. Therefore, total uri nary metabolites should be a reasonable ap proximation of the amount of metabolism. The relationship between the metabolism of these compounds and their toxicity is not an unexpected finding. It has been shown that the toxicity of numerous compounds is a result of their metabolism to reactive inter mediates (Mitchell et al.. 1976: Gillette et al., 1974). and Moslen et al. (1977) have demonstrated an association between TRI metabolism and hepatotoxicity. However, the implications of such a relationship, particu larly the quantitative association demon strated here, are important. SL 032763 the toxic effect is part of >lic pathway, as is the case Although the data are view that the metabolic compounds involve the ic intermediate, they are it. Recently, Miller and 1983) have proposed that an obligate intermediate The results of the present that formation of the re, whatever its structure, al to the overall amount ince when metabolism maximum value the toxncrease further, ited out that use of total as the index of metabolation. Most of the TRI abolized through reactive iverted to either TCA or e urine. Very little is Some metabolized <1 and PER) is bound to d some is metabolized to niel. 1963; Pegg et al.. ' al.. 1980; Stott el al., d Magee, 1982). To the se factors are involved. >lite is an underestimate u of metabolism. How led, urinary excretion is route of disposition of md. Therefore, total uri>uld be a reasonable ap.mount of metabolism. >etween the metabolism and their toxicity is not ng. It has been shown umerous compounds is holism to reactive inter7 al.. 1976; Gillette el ten et at. (1977) have .ociation between TRI totoxiciiv. However, the a relationship, particue association demon- metabolism-effect studies of tri and per 119 First. it is noteworthy that the relationship -etween toxicity and metabolism is demon- trated by both TRI and PER despite differnces in the kinetics and rates of their met.bolic processes. The metabolism of TRI is near with dose until high doses are reached. ,'hen it abruptly reaches saturation and levels >tf (Fig. 1). This type of behavior has recently men described in inhalation uptake studies >f a number of small halogenated com pounds. including TRI (Filser and Bolt, 1979; \ndersen et al.. 1980; Andersen, 1981), and ins been termed "flip-flop" kinetics. It occurs vhen the hepatic enzymes are able to metab olize the substrate readily (i.e,, the K,, value s small). At low and intermediate doses, the major factor limiting the rate of metabolism is the blood flow to the liver (Andersen, 1981). If the concentration of the compound in the blood is proportional to the dose, the amount of metabolism would also be pro portional to dose. As long as perfusion re mains rate limiting, first-order kinetics will be obeyed. However, as dose is increased, at some point the capacity of the metabolizing enzymes is reached and they become satu rated. At that point the kinetics shift from first order to zero order, and a plateau in the dose-metabolism relationship occurs. This kinetic behavior seen with TRI indicates that metabolism of TRI by the liver is a high affinity, high capacity process. In contrast, the metabolism of PER (Fig. 2) displays classical saturable kinetics. This pattern suggests, especially with its high K,,, value of 660 mg PER/kg, that extraction and/or metabolism of PER by the hepatic enzymes is not efficient. The extent of PER metabolism is dose dependent, even at low concentrations. In addition, the liver has a low capacity for metabolizing PER; the Mm of 136 mg TCA/kg/day is less than one-fifth (20%) of the extent to which TRI can be metabolized. This limited capacity for PER metabolism has been observed in numerous species (Ikeda et al., 1972; Daniel, 1963; Monster, 1979; Pegg et al., 1979). Despite the contrasts in the kinetics of TRI and PER metabolism, the toxicity of each compound is proportional to the extent of its metabolism over the entire dose range studied. A second inference, regarding the relative toxicities of the reactive metabolites, can be drawn from the results. A comparison of the hepatotoxicity elicited by TRI and PER shows that there were both quantitative and quali tative differences. Both compounds caused increases in liver weight, decreases in glucose6-phosphatase activity, and similar histopathology. Comparison of the doses which result in equivalent effect suggest that PER is at least twice as potent as TRI on a molar basis with respect to these toxic indices. The dif ference in potency is considerably greater with regard to SGPT and triglyceride in creases. Since at any given dose from two to five times more TRI than PER is metabolized, it follows that the toxic metabolite of PER is considerably more potent than that of TRI. Third, the dose-effect graphs (Figs. 3 and 5) illustrate that a maximum occurs in each effect as dose is increased. With the exception of the effect of PER on hepatic triglycerides, each maximum results from a change in the kinetics of metabolism. This result illustrates one way by which a nonlinearity between dose and effect can arise. Finally, the findings shed light on several factors involved in risk assessment. First, they establish that the amount of metabolism of both TRI and PER is directly related to their hepatotoxicity. It is the internal param eter upon which assessment of their toxicity should be based. Since there is strong evidence supporting the proposal that TRI and PER cause hepatocellular tumors in mice as a result of an epigenetic mechanism involving repeated liver damage (Schumann et al., 1980; Stott et al.. 1982), these results involving the effect of metabolism of TRI and PER on hepatic toxicity have implications with respect to their carcinogenicity as well. Risk extrapolation estimates to date have related dose to response. It would be much more fruitful to base the extrapolation of risk on metabolism rather than on dose for those 0&T6,` 120 BUBEN AND O'FLAHERTY compounds whose toxicity or carcinogenicity has been demonstrated to be dependent on metabolic activation. Other authors have dis cussed the merits of using internal parameters to explain toxic response or estimate carci nogenic potential (Gillette, 1974a,b: Gehring and Blau, 1977; Andersen, 1981). For risk assessment purposes, preliminary pharma cokinetic studies should be undertaken to relate the administered dose to the significant internal parameter of metabolism. This prin ciple may apply to many genetic carcinogens as well. Many of these agents are first metab olized to toxic intermediates which react with cellular nucleic acids. The extent of their metabolism would likely be more directly related to their carcinogenic activity than would dose. Second, the results of this study demon strate that use of high doses in studies such as the NCI carcinogenicity bioassays often lead to saturation of metabolism. The metabolism of PER deviated from linearity at doses above 100 mg/kg, considerably lower than the doses utilized in the NCI study (National Cancer Institute, 1977). Even the metabolism of TRI, a compound which is readily metabolized, approached saturation at 2400 mg/kg, a dose equivalent to the maximum tolerated dose used in the NCI bioassay (National Cancer Institute, 1976). As pointed out earlier, such saturated metab olism was associated with the occurrence of maxima in measured effects. This saturated metabolism may explain the absence of a dose-related response in the hepatic tumorigenicity data of mice in the NCI study of PER, as well as the rather small increase observed among male mice despite a doubling of the TRI dose (National Cancer Institute, 1976, 1977). Extrapolation for purposes of risk assessment of such high-dose data without correction for the pharmacokinetics of the metabolic process may lead to erroneous results (Watanabe et al., 1977) and may greatly overestimate the actual risk. It is unlikely that the pharmacokinetics of TRI and PER metabolism are unusual. The me tabolism of many compounds will likely re semble the kinetics of one or the other. Thus, implications, drawn from the results of this study, concerning the use of high doses in risk assessment have general applicability. The nonlineanties of the dose-effect curves which result from saturation of metabolic processes demonstrate the need to include at least one dose significantly lower than the maximum tolerated dose in carcinogenicity bioassays. ACKNOWLEDGMENTS The authors thank Dr. KJaus Stemmer for his help with the histological evaluations and for the gas chro matograph used in this study. This research was supported by National Institutes of Health Grant ES 07073 and a grant from the Monsanto Company. REFERENCES Allemand, H.. Pessayre. D.. Descatoire, V.. Degott, C.. Feldmann. G.. and Benhamou. J. (1978). Metabolic activation of trichloroethylene into a chem ically reactive metabolite toxic to the liver. J. Phar macol. Exp. Ther. 204, 715-723. Andersen. M. E. (1981). Saturable metabolism and its relationship to toxicity. CRC Cnt. Rev. Toxicol 9, 105-149, Andersen, M. E., Gargas. M. L,, Jones, R. a., and Jenkins. L. J. (1980). Determination of the kinetic constants for metabolism of inhaled toxicants in vivo using gas uptake measurements. Toxicol. Appl. Phar macol. 54, 100-116. Banerjee, S., and Van Duuren, B. L. (1978). Covalent binding of the carcinogen trichloroethylene to hepatic microsomal proteins and to exogenous DNA in vitro. Cancer Res. 38, 776-780. Bolt. H. M.. Buchter, a,. Wolowski, L., Gil. D. L., and Bolt. W. (1977). Incubation of uC-trichlorocthvlene vapor with rat liver microsomes: Uptake of radioactivity and covalent protein binding of metab olites. Ini. Arch Occup. Environ Health 39, 103-111. Bolt, H. M.. and Filser. J. G. (1977). Irreversible binding of chlorinated ethylenes to macromolecules. Environ. Health Perspccl. 21, 107-112. Burton, K. (1968). Determination of DNA concentra tion with diphenvlamine. In Methods of Enzvmologv (L. Grossman and K. Moldave, eds.), Vol 12B. pp. 163-166. Academic Press. New York. OS*165 sv METABOLISM-EFFECT STUDIES OF TRI AND PER 121 Butler. W. m., Maling, H. M.. Horning, M. G.. and Brodie, B. B. (1961). The direct determination of liver tngiycendes. J. Lipid Res. 2. 95-96. Daniel, J. W. (1963). The metabolism of "Cl-labelled trichloroethylene and tetrachloroethylene in the rat. Biochem. Pharmacol. 12, 795-S02. Dmitrieva, N. V. (1967). Tetrachloroethylene metabo lism. Gig. Tr Prof Zabol 11, 54-56 (Chem Abst 66: 93533b). Filser, J. G., and Bolt, H. M. (1979). Pharmacokinetics of halogenated ethylenes in rats. Arch. Toxicol. 42, 123-136. Fishbein, L. (1980). Overview of some aspects of quan titative risk assessment. J Toxicol. Environ Health 6, 1275-1296. Fiske, C. H.. and Subbarow, Y. (1925). The colori metric determination of phosphorous. J. Biol. Chem. 66, 375-400. Gehring, P. J., and Blau, G. E. (1977). Mechanisms of carcinogenesis: Dose response. J. Environ. Pathol. Toxicol. 1, 163-179. Gillette, J. R. (1974a). A perspective on the role of chemically reactive metabolites of foreign compounds in toxicity. I. Correlation of changes in covalent binding of reactive metabolites with changes in the incidence and seventy of toxicity. Biochem. Pharmacol. 23, 2785-2795.' Gillette, J. R. (1974b). A perspective on the role of chemically reactive metabolites of foreign compounds m toxicity. II. Alterations in the kinetics of covalent binding. Biochem. Pharmacol. 23, 2927-2938. Gillette. J. R.. Mitchell. J. R,, and Brodie, B. B. 11974). Biochemical mechanisms of drug toxicity. ,3nm< Rev Pharmacol 14, 271-288, Henschler, D. (1977). Metabolism and mutagenicity of halogenated oletins: A comparison of structure and activity. Environ. Health Perspect. 21, 61-64. Hodgkinson, A., and Williams, A. (1972). An im proved colon metric procedure for urine oxalate, dm. Chim Acta 36, 127-132. Humbert. B. E.. and Fernandez, J. G. (1976). Si multaneous determination of tnchloroacetic acid and tnchloroethanol by gas chromatography. Int. Arch. Oictip. Environ Health 36, 235-241. Ikeda. m,, Ohtsuji, H.. Imamura, T,, and Komoike, Y. (1972). Urinary excretion of total tnchloro-compounds. tnchloroethanol. and tnchloroacetic acid as a measure of exposure to trichloroethylene and tetra chloroethylene, Brit J. Ind. Med. 29, 328-333. I.eibman, k. C., and Ortiz. E. (1977). Metabolism of halogenated ethylenes. Environ. Health Perspect. 21, 91-97, Lowry, O. H . Rosebrough. N. J., Farr, L., and Rwdall. R J. (1951). Protein measurement with the Fulin phenol reagent. J Biol. Chem 193, 265275. Metzler, C. M,, Elfring. G. L.. and McEwen. a. j. (1974). A package of computer programs for phar macokinetic modeling. Biometrics 30, 562-563. Miller, E. C. (1978). Some current perspectives on chemical carcinogenesis in humans and expenmental animals: Presidential address. Cancer Res. 38. 1479-- 1496. Miller, R. E., and Guengerich, F. P. (1982). Oxi dation of trichloroethylene by liver microsomal cyto chrome P-450: evidence for chlorine migration in a transition state not involving trichloroethylene oxide. Biochemistry 21, 1090-1097. Miller, R. E,, and Guengerich. F. P. (1983). Metab olism of trichloroethylene in isolated hepatocytes. microsomes. and reconstituted enzyme systems containing cytochrome P-450. Cancer Res. 43, 1145-1152. Mitchell. J. R.. Snodgrass. W. R.. and Gillette, J. R. (1976). The role of biotransformations in chem ical-induced liver injury. Environ. Health Perspect. 15, 27-35. Monster, A. C. (1979). Differences in uptake, elimi nation. and metabolism in exposure to trichloroeth ylene, 1,1.1-tnchloroethane. and tetrachloroethylene. Int. Arch. Occup. Environ. Health 42, 311-317. Moslen, M. T,, Reynolds, E. S., and Szabo. S. (1977). Enhancement of the metabolism and hepatotoxicity of trichloroethylene and perchloroethylene. Biochem. Pharmacol 26. 369-375. Munro, 1. C. (1977). Considerations in chronic toxicity testing: The chemical, the dose, the design, J. Environ. Pathol Toxicol. 1, 183-197. National Cancer Institute (1976). Carcinogenesis Bioassay o/ Trichloroethvlene, DHEW Pub!. No. (N1H) 76-802. National Cancer Institute (1977). Bioassay oj Tetrachlo roethylene for Possible Carcinogenesis. DHEW Publ. No. (NIH) 77-813. Parchman. L. G., and Magee. P. N. (1982). Metabo lism of (1JC)trichloroethylene to IJCO, and interaction of a metabolite with liver DNA in rats and mice. J. Toxicol Environ. Health 9, 797-813. Pegg. D. G., Zempel, J. A.. Braun. W. H,, and WaTanabe, P. G. (1979). Disposition of tetrachloro[1JC]ethylene following oral and inhalation exposure in rats. Toxicol. Appl Pharmacol. 51, 465-474. Rajagopal, G., and Ramakrishnan, S. (1975). A new method for estimation of ethylene glycol in biological material. Anal. Biochem. 65, 132-136. Reitman, S.. and Frankel, S. (1957). A colorimetric method for the determination of serum glutamic ox aloacetic and glutamic pyruvic transaminases. Amer ./. Clin. Pathol. 28, 56-63. Schumann. A. M.. Quast. J. F.. and Watanabe, P. G. (1980). The pharmacokinetics and macromolecular interactions of perchloroethylene in mice and rats as related to oncogenicity. Toxicol. Appl. Pharmacol. 55. 207-219. SL 032766 122 BUBEN AND O FLAHERTY SokaL, R., and Rohlf, J. (1969). Biometrics, pp. 235246. Freeman. San Francisco. Stott. W. t., Quast. J, F.. and Watanabe, P. G. (1982). The pharmacokinetics and macromolecular interactions of trichloroethylene in mice and rats. Toxicol. Appl. Pharmacol 62. 137-151. Uehleke. H.. and Poplawskj-Tabarelli, S. (1977). Irreversible binding of uC-labelled trichloroethylene to mice liver constituents m vivo and in vitro. Arch. Toxicol. 37, 95-105. Van Duuren, B. L. f 1975). On the possible mechanism of carcinogenic action of vinyl chloride. Ann N. Y. Acad. Sci. 246, 258-267. Van Duuren, B. L.. and Banerjee. S. (1976). Covalent interaction of metabolites of the carcinogen trichloro ethylene in rat hepatic microsomes. Cancer Res 36, 2419-2422. Van FIandel. E,, and Zilversmit, D. B. (1957). Micromethod for the direct determination of serum triglycerides. J. Lab Clin Med. 50, 152-157. Watanabe. P. G., Young, J. D., and Gehring, P. J. (1977). The importance of nonlinear (dose-dependent) pharmacokinetics in hazard assessment. J Environ. Pathol. Toxicol. 1, 147-159. Weisburger, J. H., and Williams. G. M. (1980). Chemical Carcinogens. In Toxicology: The Basic Sci ence of Poisons (J. Doull, C. D. KJaassen, and M. O. Amdur, eds.), pp. 84-138. Macmillan Co., New York. Yllner. S. (1961). Urinary metabolites of IJC-tetrachloroethvlene in mice. Nature <Londont 191, 820. SL 032767