Document 7O6LG0LB74mvwDDba9b6zLmRV

D^APr^V 1990 frfz. -fvyy (Xf^UiXL BIOLOGICAL THRESHOLD FOR VINYL CHLORIDE EXPOSURE-INDUCED HEPATIC ANGIOSARCOMA IN HUMANS introduction In most epidemiological studies, it has not been possible to determine the dose of the agents to which humans have been exposed. Occasionally, crude retrospective dost-response curves have been developed. A summary was made by the Meselson Committee regarding the dose levels of several known human carcinogens which appear to be carcinogenic in certain human populations.1 The characteristics of exposure and the spectrum of effect comes together in a correlated relationship customarily referred to as a dose-response relationship. This relationship is a fundamental and pervasive concept in toxicology. Indeed, an understanding of this relationship and its facile use is the essence of the study of toxic materials. Although a full understanding of the intricacies of both dose and response entails many complexities, only a few assumptions form the skeleton of the relationship. The first is that an implicity assumption of causality is made. To arrive at a quantitative and precise statement of the relationship between a toxic material and an observation effect or response, one must known with reasonable certainty that the relationship is indeed a causal one. This is true for vinyl chloride and angiosarcoma. In many epidemiological studies, results show an association between a response and the disease, and one or more impinging variables. Not infrequently the data are amenable to CMA 121491 Biological Threshold. , ., continued 2 presentation in terms similar to those employed in experimental use of dose response in pharmacology and toxicology. In strict usage, the dose-response relationship is firmly based on the knowledge or a reasonable presumption that the effect is a result of the toxic agent. The second assumption is simply and obviously that the response is, in fact, related to the dose. The simplicity of this assumption is often a source of misunderstanding. This assumption is really a composite of three others that will recur frequently: A. There is a molecular or receptor site with which the chemicals interact to produce the response; B. The production of response and the degree of response are related to the concentration of the agent at reactive sites; C. The concentration at the site is, in turn, related to the administered dose. Thus, the numerical and graphic dimensions of a dose-response relationship can include the assumptions that: (1) the response is a function of concentration at the site; (2) the concentration at the site is a function of the dose; and (3) the response and dose are causally related.2, 3 In toxicology, when assessing the safety of a substance, it is necessary to have both a quantifiable method of measurement and a variety of criteria in points of toxicity that can be used. The ideal criteria should be closely associated with the molecular events resulting from exposure to the toxic agent. This idea is usually considered unapproachable in clinical CMA 121492 Biological Threshold. . ., continued 3 settings, especially with regard to effects with great latency, such as malignancy. Vinyl chloride-induced hepatic angiosarcome among plastic workers represents such a clinical setting. The selection of a toxic end-point for measurement is also not always so straightforward. In vinyl chloride exposure, other end points (non-malignant) have been quantitatively demonstrated to be precise indicators of the acute and chronic phases of hepatotoxicity.4,5> 6 In addition, prospective clinical data has been accumulated and developed over the past ten years through extensive screening and medical surveillance of vinyl monomer-exposed workers which fulfill most of the above criteria. This paper provides the clinical and epidemiological evidence for both a dose-response and biological threshold response to an occupational hazard.7 Materials and Methods Vinyl chloride-associated hepatic angiosarcoma (HAS) was studied in two populations. One consisted of 1,200 vinyl chloride polymerization workers from the index plant, i.e., where the original human vinyl chloride-associated HAS observations were made. This index plant also is the single largest contributor of VC-related HAS cases worldwide. This cohort has been followed respectively from 174 to the present via a medical surveillance program. The medical surveillance program consisted of annual medical examinations, clinical laboratory screening on an annual/semi-annual basis which included over 50 biochemical studies of blood and urine, x-ray examinations of the chest, abdomen and hands, CMA 121493 Biological Threshold. . ., continued 4 radioisotopic studies of the liver, spleen and brain, and angiographic-hemodynamic examinations of the portosplenic system. Detailed follow-up clinical investigations were performed on all biochemical, radiological, or clinical abnormalities. Approximately 110 cases of hepatic disease constitute a sub-cohort of this population, 15% of whom have vinyl chloride-associated non-malignant liver injury.8,9 The second population of cases consisted of all VC-associated HAS cases reported to the Angiosarcoma of the Liver Registry of the Imperial Chemical Industries Corporation in England.10 This Registry contains all histologically-documented VC-associated HAS cases reported in the scientific literature, by government agencies and/or chemical industries, from North America, Europe, South America and Asia.11'23 All cases have been histologically confirmed as liver angiosarcomas. Each case has a date of birth, company or factory of employment, date of clinical diagnosis, age at diagnosis, date of first exposure, years of exposure (i.e., years worked in a vinyl chloride polymerization manufacturing plant), years from first exposure to time of diagnosis (latency), date of death, estimated exposure levels, occupational (job) activities, and some information with regards to prior occupation, nationality or country of origin. From these two cohorts, 108 histologically-confirmed cases of angiosarcoma constitute this study population group. Each case spent from three to 37 years working in vinyl chloride polymerization production plants in 12 different countries from 1938 through 1984. Forty-three (43) cases of HAS are from North America, 2 from Japan, 6 from CMA 121494 Biological Threshold. . ., continued 5 Eastern Europe, and the remaining 57 from Western Europe and the United Kingdom. All cases have been histologically confirmed as liver angiosarcomas. All histological identifications were made or confirmed by internationally-recognized and experienced pathologists before this analysis and independent of this study's author. No histologicallyconfirmed case has been excluded. Forty (40) cases were retrospectively identified or identified before the beginning of this study in 1975. The remaining 68 cases developed during the 10-year follow-up interval since the beginning of this study. Data and Results Worldwide HAS case occurrences demonstrate a poisson-like distribution with the peak occurring about 1976 (Figure 1). The earliest cases identified in North America occurred in 1962 with a peak and median occurrence between 1974-1975. The open circles identify the North American cases, including the index plant first identified in 1967, with the median in 1976 and peak in 1978. The eight remaining cases were identified between 196578 with the median and peak occurring in 1973. The annual incidence of HAS by date of diagnosis and geographic location for four leading countries is shown in Figure 2. The distribution by year of first exposure for the entire cohort begins in 1940 and finishes in 1968 (Figure 3). The initial 38 cases (triangle-based circles) retrospectively identified (between 1955-1974), began their vinyl chloride exposure before 1968. Since 1974, all 68 prospectively identified cases (closed circles) also began their year of first exposure prior to 1968. Figure 4 illustrates the age at which diagnosis was first made among CMA 121495 Biological Threshold. , ., continued 6 the VC-associated HAS. The frequency of HAS latency periods has a skewed distribution, peaking at approximately 20 years (Figure 5). The range of the latency periods is from 9-38 years. The mean latency period for each five-year interval, starting with 1940-44, demonstrates a progressive shortening from 35 to 14.6 years with an overall average latency Tyf/b of 22.6 years (Talnv4). The latency periods point out a discrepancy between the reduction in occupational exposure levels and the sudden decrease in HAS occurrence. On initial observation, the reduction of exposure levels in 1974 appears to have produced a rapid and profound reduction in the incidence of angiosarcoma (Figure 1). Had the reduction in exposure levels instituted in 1974 been a major factor in reducing the occurrence of HAS, its effect would not be expected for at least nine years, i.e., shortest latency. In fact, study of the latency period in other cancers in which the carcinogen, its dose and duration are known, have shown a prolongation of the latency period which a decrease in dosage or duration of exposure. A review of the index plant cases (13 cases) illustrates a reversed exposure-latency relationship between the total (or first year of) exposure and the onset of disease. This relationship holds for all the North American (United States and Canada) cases and the various European countries (western Germany, France, United Kingdom, Sweden and Yugoslavia). In contrast, these HAS cases' latency periods decrease with shorter duration of exposure (Figures 6 and 7). CMA 121496 Biological Threshold. . ., continued 7 The total years of exposure for each of the cases is shown in Figure 8. Ninety-three cases (93%) had 10-43 years of exposure (median 20 years); seven cases (7%) had only 3'A6 years of exposure (median 5 years). Continued exposure during the latency period could affect the length of latency. Therefore, the relationship between total years of exposure and latency was studied among those with concomitant exposure and latency periods versus the smaller groups with limited short total exposures. This latter group is not significantly affected by continued exposure which might of itself act as a promoter. These eight individuals (short-exposure group [SEG]) had left the chemical industry for other occupations with no or very limited exposure to other known carcinogens. Figure 9 correlates the year of first exposure with latency in SEG. In regards to total years of exposure, the same reverse relationship was found im . ,, oWutuWU in the long cases iexposure The Effect of Weighted Exposures A review of the literature and industrial records indicates a progressive linear decrease in exposure levels from 1940 (1-3 ppms) to 1974 (50-250 ppms). Data reported in the literature and extrapolated from the various industrial data were used to construct estimated exposure levels.7, 24 Two exposure curves were plotted from these reported exposure levels covering the 1940-74 years (Figure 10). CMA 121497 Biological Threshold, . ., continued 8 Utilizing this data, each year of exposure was weighted on the basis of the estimated level of exposure. Each case's total exposure was redetermined based on the exact year of exposure. Replotting of the latency versus total cumulative exposure (weighted exposure) continues to demonstrate the same linear relationship in an even more highly correlated manner. This is seen in index plant cases (Figure 11), USA cases (Figure 12), the Canadian cases (Figure 13), and holds true for the European and world-wide cases. It is most dramatically shown in the short-exposure group (Figure 14). Discussion These data provide, for the first time, clinical (human) evidence for a biological threshold for VC toxicity including carcinogenicity. The retrospective and prospective demonstration that all cases of VC-associated HAS began their first exposure before 1968 in plants which began operation before 1967 (majority before 1960), indicate that a common pattern of environmental working conditions were present during the induction and subsequent promoting periods. Vinyl chloride's carcinogenic effect is related to a particular time and environment. This pattern is confirmed in seven countries around the world. If, as traditionally understood, the risk of developing VC-associated HAS is related to the different levels of environmental exposure during the earlier years of employment, then the gradual reduction in exposure levels should be reflected in the changing incidence rates among those workers exposed at different times and durations.25 Exposure during early cmm21498 Biological Threshold. . ., continued 9 years (1940-1950) was at a higher intensity. The cytotoxic effect of VC was shown by the severity of the hepatocellular injury seen among workers exposed during those years. In these early workers, far more pamechymal injury, hepatocellular necrosis, fibrosis, and portal hypertension were found, concurrent with clinical development of HAS.26,27 It is highly probable that in these situations the liver and endothelial cells were unable to undergo cancer transformation between the VC cytotoxic injury did not allow the cells to survive. Therefore, fewer total induced cells were available for further transformation by the action of promoters or repeated exposure.28 Those individuals whose exposure occurred in later years (1951-60) were exposed to lower vinyl chloride levels. Although these levels were less cytotoxic, they were still able to induce the carcinogenic effect. Therefore, a larger proportion of surviving cells were initiated and available for the promoting effect of either continuing exposure or other environmental agents. This group had the more rapid induction (shorter latency) of HAS development. As the environmental levels were further decreased (after 1968), the liver was able to maintain adequate detoxification to prevent both the cytotoxic effect and the carcinogenic induction or initiating phase. This level of exposure appears not to be affected by other promoting agents or continued low-level VC exposure. In support of this hypothesis are the findings in later exposed workers (1960-1975), and evidence of histological hepatocellular adaptation, e.g., focal hepatocellular hyperplasia and mild increase in sinusoidal fibrosis.6 CMA 121499 Biological Threshold, . ., continued 10 These findings reflect the hepatocyte's adaptation and very mild cytotoxic injury to endothelial cells, which in turn stimulates collagen forming cells. The latter scenario is further supported by the fact that after five to ten years of follow-up, these histological lesions (which are highly correlated with the individual's total exposure to vinyl chloride) have shown no evidence of progression even with continued low-level VC exposure (< lOppm).20 In addition, all biochemical abnormalities initially seen have regressed and the histological picture on subsequent biopsies has been stable and unchanging.28,30 Therefore, these data support the concept that high leves (> 750ppm) of vinyl chloride produce severe liver cell (hepatic and sinusoidal) injury and reduce or significantly delay the malignant expression or transformation of initiated endothelial cells. At lower levels (or range of levels) which appear to be somewhere between 250-700 ppm, there is a maximal carcinogenic effect with less cytotoxic injury. Whether this latter phase can be totally accounted for by the VC level of exposure and not contributed to by confounding or enacting secondary agents, requires further study. Finally, low levels (1-50 ppm) appear to be within a biological threshold. Neither significant cytotoxic nor carcinogenic effects occur at these exposure levels. In addition, this level also does not appear to act as a promoter nor cause progression of already existing hepatocellular lesions. The data also suggests that there may be a co-factor or interaction of other chemicals occurring which may play a role in the shorter latency periods and reverse dose-response with regards to malignant transformation. CMA 121500 Biological Threshold. . ., continued 11 Greenberg and Tamburro,31,32 utilizing a rank-order method for estimating chemical exposure, illustrated this method's ability to identify the causal relationship between VC and other chemicals with the development of HAS. Their study suggests that catalysts which biochemically inhibit the liver's major detoxifying mechanism responsible for the removal of active vinyl chloride metabolite might be a confounding or interacting factor. Industrial plants performing different operations might have environments which provided dual exposures. One, like vinyl chloride, which is hepatotoxic and carcinogenic, and another, like diethyl maliete, which might decrease or impede the liver's detoxifying capabilities, thus allowing for a more rapid induction of carcinogenesis -- hence, a shorter latency. In contrast, those plants not utilizing these detoxifying inhibitors may account for the usual and more prolonged course of cancer induction, i.e., increased exposure, decreased latency. This hypothesis is supported by review of HAS cases in the various countries and their plants' start-up dates. As shown in Figure 15, all HAS cases developed in PCB plants whose start-up began before 1967 -- in eastern Europe before 1953, in Canada before 1941, in the United States before 1953, and in western Europe all but three plants before 1954. This indicates a high probability that either the level of exposure and/or type of exposure during these periods play an important role in type of injury seen and the way in which the tumor developed. These clinicqal data provide a clear demonstration of a biological threshold for a known carcinogenic non-radiation agent in a human population. It demonstrates a uniquely- CMA 121501 Biological Threshold. . ., continued 12 different pattern of latency in chemically-induced cancer, even though its distribution characteristics are mathematically the same as reported by Doll and Hill33 and fit the concepts of latency initially illustrated by Sartwell34 and later expanded by Armenian and Lilienfeld.35 The reverse dose-response to latency relationship, however, is quite different from that reviewed by Polednak, Cobb and others, and reflects a different mechanism or model for chemical carcinogenesis.36 This study provides carefully-observed human experience data, prospectively acquired, which can be used to guide the development of future environmental policy with regards to vinyl chloride environmental exposure. CMA121502 Biological Threshold. . ., continued 13 REFERENCES 1. Contemporary Pest Control Practices and Prospective. Reported of the Executive Committee, Vol. I. National Academy of Sciences. Washington DC, 1975, p. 75. 2. Albert A (1965). Fundamental aspects of selective toxicity. Annals of New York Academy of Science 123, pp. 5-18. 3. Loomis TA (1978). Essentials of Toxicology, 3rd Edition. Lea & Febiger, Philadelphia. 4. Popper H, Thomas LB (1975). Alterations of liver and spleen among workers exposed to vinyl chloride. Annals of New York Academy of Sciences 246, pp. 172-194. 5. Thomas LB, Popper H, Burke P, et al. (1975). Vinyl chloride induced liver disease; from idiopathic portal hypertension (banti syndrome) to angiosarcoma. New Engl J of Medicine 292, pp. 17-22. CMA 121503 Biological Threshold. . ., continued 14 6. Tamburro CH, Makk L, Popper H (1984). Early hepatic histological alterations among chemical (vinyl monomer) workers. Hepatology 4, pp. 413-418. 7. Tamburro CH, Greenberg RA (1982). Safe carcinogenic and toxic exposure levels in vinyl chloride induced hepatic angiosarcoma. Clinical Research 30, p. 307a. 8. Greenberg RA, Tamburro CH, Kupchella CE (1978). Prospective medical surveillance program for detection and prevention of industrially-related cancer. |n Prevention and Detection of Cancer, Niebeurgs H, Ed. New York: Marcel Dekker Inc., Part 11(2), pp. 1921-1928. 9. Dannaher CL, Tamburro CH, Yam LT (1981). Occupational carcinogenesis: the Louisville experience with vinyl chloride associated hepatic angiosarcoma. American J of Medicine 70, pp. 279-287. 10. Bennett B (1983). Angiosarcoma registry. Works Medical Officer, ICI Hill House Works, Cleveleys, Blackpool, Lancastershire, England.11 11. Marsteller HI, Ledback WK, Muller R, Juhe S, Lange CE, Rohnerf HG, Veltman G (1973). Chronisch-Toxische Leberschaedenbe, Arbeitem in der PVC Produktion. Deutsch Med Worchenschr 98, pp. 2311-2314. CMA 121504 Biological Threshold. . ., continued 12. Delorme F, Theriault G (1978). Ten cases of angiosarcoma of the liver in Shawinigan, Quebec. J of Occupational Med 20, pp. 338-340. 15 13. Spirtas R, Kaminski R (1978). Angiosarcoma of the liver in vinyl chloride/polyvinyl chloride workers: 1977 update. J of Occupational Med 20, pp. 427-429. r 14. Bonneton G, Champetier J, Foumet J, Guidicelli H, Legrand J, Dupre A, Hostein M, Marty F, and Pahn M (1977). Angiosarcome hepatique et fibrose portale chez les travailleurs due chlorure de vinyl. Deux observations. La Nouvelle Presse Medicate 6, pp. 735-742. 15. Pilichowski P, Faurd C, Aubert M, Pahn M, Latreille R, and Barrie J (1977). Angiosarcome costal lie a une intoxication au chlorure de polyvinyl. La Nouvelle Presse Medicate 8, pp. 2485-2486. 16. Falk H, Heath CW Jr, Carter CD, Wagoner JK, Waxweiler RJ, and Stringer WT (1974). Mortality among vinyl chloride workers. Lancet 2, pp. 784-785. 17. Lee Fi, and Hany DS (1974). Angiosarcoma of the liver in a vinyl chloride worker. Lancet 2, pp. 1316-1318. CMA 121505 Biological Threshold. . continued 16 18. Block JB (1974). Angiosarcoma of the liver following vinyl chloride exposure. J Amer Med Assoc 229, pp. 53-54. 19. Dannaher CL, Tamburro CH, and Yam LT (1981). Occupational carcinogenesis: The Louisville experience with vinyl chloride associated hepatic angiosarcoma. Amer J Med 70, pp. 279-287. 20. Creech JL and Johnson MN (1974). Angiosarcoma of the liver in the manufacture of polyvinyl chloride. J of Occupational Med 16, p. 150. 21. Saric M, Ulcar Z, Zorica J and Gelic I (1976). Malignant tumors of the liver and lungs in an area with a PVC industry. Environmental Health Perspectives 17, pp. 189-192. 22. Monson RR, Peters JM, and Johnson MN (1974). Proportional mortality among vinyl chloride workers. Lancet 2, pp. 397-398. 23. Heath CW Jr, Falk H, and Creech JL Jr. (1975). Characteristics of cases in angiosarcoma of the liver among vinyl chloride workers in the United States. Annals New York Academy of Sciences 246, pp. 231-236. CMA 121506 Biological Threshold. . ., continued 17 24. Binns CHB (1979). Vinyl chloride: a review. J of Society Occupational Med 29, pp. 134-141. 25. Tamburro CH and Creech JL (1983). The identification of hepatic injury and hepatic angiosarcoma among vinyl chloride workers: The epidemiological approach. J of Occupational and Environmental Health 5, pp. 37-48. 26. Makk L, Delmore F, Creech JL, et al. (1976). Clinical and morphological features of hepatic angiosarcoma in vinyl chloride workers. Cancer 37, pp. 149-163. 27. Berk PD, Martin JF, Young RS, et al. (1976). Vinyl chloride associated liver disease - NIH Conference. Annals oflnt Med 84, pp. 717-731. 28. Tamburro CH (1979). Chemical hepatitis, pathogenesis, detection and management.a Med Clinics of North America 63, pp. 545-566. 29. Tamburro CH (1984). Relationship of vinyl monomers and liver cancers: angiosarcoma and hepatocellular carcinoma. Seminars in Liver Disease 4, pp. 159-169. CMA 121507 Biological Threshold. . ., continued 18 30. Tamburro CH and Greenberg RA (1980). Identification of human toxicity and carcinogenicity by ethylene derivatives. In Mechanisms of Toxicology and Hazard Evaluation, Holmstedt B, et al., Eds. New York:Elsevier/North Holland Biomedical Press, pp. 319-334. 31. Greenberg RA and Tamburro CH (1981). Monitoring exposure to hazardous chemicals in an industrial setting: a method of demonstrated utility. In System Sciences in Health Care, Tilquin C, Ed. Toronto:Pergamon Press Ltd., pp. 1143-1151. 32. Greenberg RA and Tamburro CH (1981). Exposure indices for epidemiological surveillance of carcinogenic agents in an industrial environment. J of Occupational Med 23, pp. 353-358. 33. Doll R and Hill AB (1956). Lung cancer and other causes of death related to smoking: A second report on motality of British doctors. British Med J 2, pp. 1017-1081. 34. Sartwell PE (1950). The distribution of incubation periods of infectious diseases. Amer J of Hygiene 51, pp. 310-318. CMA 121508 Biological Threshold. . ., continued 19 35. Armenian HK and Lilienfeld AN (1974). The distribution of incubation periods of neoplastic diseases. Am J of Epidemiology 99, pp. 92-100. 36. Polednak AP (1974). Latency periods in neoplastic disease. Amer J of Epidemiology 100, pp. 354-356. CMA12^509 _ Biological Threshold. . ., continued LIST OF FIGURES 20 Figure 1: Worldwide HAS case occurrences, with the peak occurring about 1976. Figure 2: The annual incidence of HAS by date of diagnosis and geographic location for four leading countries. Figure 3: The distribution by year of first exposure for the entire cohort (1940-68). Figure 4: Illustrates the age at which diagnosis was first made among the VC-associated HAS. Figure 5: The frequency of HAS latency periods (skewed distribution), peaking at approximately 20 years. Figure 6: I. HAS cases^ latency periods decrease with shorter duration of exposure. Figure 7: II. HAS cases' latency periods decrease with shorter duration of exposure. A Figure 8: The total years of exposure for each case. Figure 9: Correlates the year of first exposure with latency in the short exposure group. CMA 121510 Biological Threshold. . ,, continued 21 Figure 10: Two exposure curves illustrating reported exposure levels covering the 1940-74 years. figure 11: Latencyvs. total cumulative exposure (weighted exposure) in index plant cases, figure 12: Latencyvs. total cumulative exposure (weighted exposure) in USA cases. figure 13: Latencyvs. total cumulative exposure (weighted exposure) in Canadian cases. Figure 14: Latencyvs. total cumulative exposure (weighted exposure) in the short exposure group. figure IS: All HAS cases developed in PCB plants whose start-up began before 1967 (in eastern Europe before 1953, in Canada before 1941, in the United States before 1953, and in western Europe all but three plants before 1954). CMA121511 Biological Threshold. . ., continued 22 LIST OF TABLES f^ fic The range of the latency periods is from 9-38 years. The mean latency period for each five-year interval, starting with 1940-44, demonstrates a progressive shortening from 35 to 14.6 years with an overall average latency of 22.6 years. CMA 121512 VINYL CHLORIDE ANGIOSARCOMA CASES WORLD-WIDE YEAR OF DIAGNOSIS -k> U VINYL CHLORIDE ASSOCIATED ANGIOSARCOMA IN VARIOUS COUNTRIES CANADA '41 I U.S.A. '39-'46 O^A *nuiOcuto Hf4*wy NUMBER OF CASES 5 FRANCE '41 -`57 3 5 GERMANY '52-'60 3 0 f J--------- 1______ i- i i i i i j______ i______ i______ i i i --l 01 1952 '56 '60 '64 '68 '72 '76 '80 YEAR op FIGURE 2 VINYL CHLORIDE ANGIOSARCOMA YEAR EXPOSURE BEGAN V* ppm 2000 2000 44 9r 200-500 500 44 50-200 LO LU co < u u. O tr IxJ m I- I o > 1940 M cn CJ1 19 74 f >M 974 f? ? ** ' * * ' ** !!!!!!! i i i l i i i i i i i i i i i i i t i i i i i i i i i i i '45 '50 '55 '60 '65 '70 YEAR FIGURE 3 mp VINYL CHLORIDE ANGIOSARCOMA AGE AT TIME OF DIAGNOSIS 0 9 f - CO LJ 00 7- < o i ll O 5 + Or: LU El P m t 't P*P i # < *i* P 0 * 0 P **' *P *0 * P P P -'pip 0 ' Z) Z 3 - *f` * * * ,* - **** f ft ** * M > 1 *f __1__1__!__1__I__l__1_1__i_1__1_1__1__1__1__1__1__1_J__1__1__!__L_1__!_1__1__1_1 1 [ l t l 1 l f 35 40 45 50 55 60 65 70 VIAIC CI) FIGURE U TOTAL NUMBER OF CASES VINYL CHLORIDE ANGIOSARCOMA TIME FROM FIRST EXPOSURE TO DIAGNOSIS - - - - - - - 11 1 ii1iiii1iii i1i iii1i iii1iiti 1iii 5 10 15 20 25 30 35 YEARS (Lfrr&fK'/) RELATIONSHIP OF EXPOSURE TO LATENCY IN VINYL CHLORIDE ASSOCIATED ANGIOSARCOMA {In North America ) I960 r 1955 SOOppm USA A CANADA FIRST YEAR OF EXPOSURE 1950 2000ppm 1945 - A A A A A 1940 10 15 20 25 30 LATENCY IN YEARS (first exposure to diagnosis) o CO FIGURE 6 FIRST YEAR OF EXPOSURE RELATIONSHIP OF EXPOSURE TO LATENCY IN VINYL CHLORIDE ASSOCIATED ANGIOSARCOMA 1965 ( In Europe ) 250ppm I960 AML A FRANCE A GERMANY ~500ppm 1955 a A A A AA AA 1950 ~ 2000ppm 1945 10 15 20 25 LATENCY IN YEARS {first exposure to diagnosis ) 30 (1941) ____ 9 O CD FIGURE 7 TOTAL YEARS OF EXPOSURE IN CASES OF HEPATIC ANGIOSARCOMA IN VINYL CHLORIDE WORKERS ' .. 9 i CO LlJ CO 7 < o Ll_ o q: LU CD I I 1^1 t t t* t t' ' ft ' ffi + i 1 i <, 3 -> - * 1 I I 1 .1 I 1 1.. 1 1 I I I 1 I 1 I ! I I I I I I 1 1 I I I I I I I I l I o 0 5 10 15 20 25 30 YEARS cron o sf GoI'.ti-if r %t> FIGURE 8 YEAR OF FIRST EXPOSURE VINYL CHLORIDE ASSOCIATED ANGIOSARCOMA (Short Exposure Cases ) 1965 250ppm 3.5 I960 ~500ppm 1955 INDEX CASE AND NUMBER OF YEAR EXPOSURE 1950 ~2000ppm 1945 5 10 15 20 25 LATENCY (first exposure to diagnosis ) 30 FIGURE 9 300 0- ESTIMATED VINYL CHLWtlDE EXPOSURE 1000 500 - EXPOSURE PPM 200 - 100 - ^ 50 - > 25 - o > ro a BINNS TAMBURRO A t -J-------------------------l-------------------------- 1-------------------------- 1_________________ i__ '30' 'UO 50 '60 '70 YEARS ESTIMATED CUMULATIVE EXPOSURE VINYL CHLORIDE ASSOCI^D ANGIOSARCOMA C BFG Exposure Cases) FIGURE II zz$\,zvw \o VINYL CHLORIDE ASSOCIATED ANGIOSACOMA USA CASES WEIGHTED ESTIMATE OF TOTAL EXPOSURE latency (first exposure to diagnosis) (YRS.) CMA 121524 FIGURE 12 W EIGHTED CUMULATIVE EXPOSURE CANADA LATENCY (YRS.) CASES VINYL CHLORIDE ASSOCIATED ANGIOSARCOMA (Short Exposure Cases) CMA 121526 * POLYVINYL CHLORIDE PRODUCTION PLANTS CMA 121527 e figure i5" CO FIGURE I6 YEAR EXPOSURE BEGAN OCT 20 U-l: rynpl-i hL'I'iH' lo'-i- fc:Ll"j. 11, 6-t- ltC^ P.2/2 Questions For Consideration With Respect To The Proposed University of Louisville Brain Cancer/VCM Case-Control Study Carlo: 1. What is the current status of your relationship with each of the three companies formerly comprising B.F. Goodrich in Louisville? When you met with the CMA panel earlier this year you said something about11re establishing a working network with all three companies." 2. What data would you need to get from all three units? Do you anticipate any problems obtaining the necessary data? Are there any other sources of the necessary data that could be accessed if needed? 3. You have described the brain cancer study in terms of "20-year prospective follow-up" and "25-year combined retrospective" evaluation. What periods of time do these cover? How are the data alike and/or different between the two periods of time? 4. Do you still have the peer reviewers' comments on your 1981 paper on ASL and vinyl chloride? If so, could you make them available to us? This would help us understand how other researchers have viewed the methodology that you employed for that paper and would employ for the brain cancer study. 5. What is your best estimate of the maximum number of brain cancer cases that could be available for study, assuming 1) that you could only get tissue for Goodrich employee cases that were hospitalized in the Louisville area; and 2) that you could get tissue from Goodrich employees hospitalized outside the Louisville area as well? This is really a question about what the maximum statistical power of the study is likely to be. 6. Can we develop a firm timetable for initiation and completion of the study? 7. Can we develop line-item breakdowns and prepare a more detailed study budget? 8. Would you consider working with someone like Kenneth Mundt as a ooinvestigator/co-author, to add some additional epidemiologic expertise to the brain cancer study and to take full advantage of the follow-up data that he will be generating in the industry-wide cohort study? ' CMA 121529 / Specific Antisera to Unique Antigenic Determinants in Human Hemoglobin-Acrylonitrile Adducts1 John L. Wong,^ Chun Ming An, Xiu Zhen Yang, Yu Ting Zheng, and Carlo H. Tamburro Department of Chemistry [J.L.W., C.M.A., X.Z.Y., Y.T.Z.], and Department of Medicine and of Pharmacology and Toxicology [C.H.T.], University of Louisville, Louisville, KY 40292 Key Words: antisera, immunoassay, hemoglobin, acrylonitrile, adducts Running Title: polyclonal antibodies to assay human hemoglobin-acrylonitrile adducts as biomarkers Footnotes: 1 Supported in part by a grant from the National Institute of Environmental Health Sciences ES 05353 and by a postdoctoral fellowship from Graduate Programs and Research of the Univ. of Louisville. 2 To whom requests for reprints should be addressed at Department of Chemistiy, University of Louisville, Louisville, KY 40292. CMA 121530 2 ABSTRACT Specific immunogenic hemoglobin adducts were used to develop an immunoassay for dosimetry of exposure to potential human carcinogens. Pathological human hemoglobin (Hb) conjugates of acrylonitrile (AN) were isolated as Hb-ANl and Hb-AN2. Structural modification of Hb was probed by comparing the kinetics of S-alkylation of blood thiols by the hydrophobic AN and hydrophilic chloroacetaldehyde hydrate, a vinyl chloride metabolite. This and^C counting for up to 3 AN per Hb and gel analysis of the adducts appear to be compatible with increasing levels of cyanoethylation in the distal heme pocket region to yield two antigenic adducts. The antisera raised in Balb/c mice to Hb-ANl and Hb-AN2, which should recognize various Hb epitopes, showed in ELISA a binding specificity for the adducts 2.7 to 3.3 times over Hb before immunopurification. After the antisera were passed through a Hb-Sepharose 4B column to retain antibodies to normal Hb, the polyclonal antibody filtrate showed 16.5 times specificity towards Hb-AN in blood with a sensitivity of 5 ng The lack of cross-reactivity with confounders such as Hb-chloroacetaldehyde and Hb-glycolaldehyde was demonstrated. Immunoassay of human blood exposed to AN in vitro gave a dose-response, illustrating the clinical applicability of specific immunoassay for the detection of hemoglobin adducts. To our knowledge, this represents the first example of an unique approach to the development of immunoassays of Hb adducts as biomarkers of human exposure to chemical carcinogens. CMA 121531 3 INTRODUCTION Acrylonitrile (AN) represents an example of an important industrial chemical of high U.S. production [2.65 billion lb in 1991 (1)] whose carcinogenic risk to humans remains in serious dispute. AN is used mainly as a starting material for synthetic fibers, resins and rubber, and is released to the environment as fugitive emissions and in wastewater during its production and use. It has been classified by the International Agency for Research on Cancer to be a Group 2A agent that is probably carcinogenic to humans (2). A risk level of E-4 (1 in 10,000) was estimated for respiratory cancer in workers exposed to an air concentration of 1 pg of AN per cubic meter (0.47 ppb) (3). However, two recent epidemiological studies of workers exposed to AN, one a cohort of 2,671 men who had worked at American Cyanamid (4), and the other consisted of 6,803 workers in the Netherlands (5), have found no indications that AN has a carcinogenic effect. If specific biomarkers had been available for measuring AN exposure in these populations, the human studies would have provided a more preferred and accurate assessment of risk than the default assumption based on animal carcinogenesis data of AN (6). Considerable progress has been made in the dosimetry of chemical carcinogens from adducts of human hemoglobin A (Hb) (7). Regarding the Hb-AN adducts, gas chromatography-mass spectrometry analysis of N-(2-cyanoethyI)-valine, an adduct formed by addition of AN to the Nterminal valine of Hb, has been reported (8). The adduct levels among 41 chemical workers were 0.02-66 nmol/g Hb or 1.29-4260 ppm. However, none of this developed methodology is readily applicable to large-scale epidemiological screening or routine monitoring. A specific and sensitive immunoassay to determine human exposure to AN and other chemicals would be ideal for human health assessment AN is known to be metabolized by direct addition to the double bond, a major pathway, and epoxidation (9) We have been interested in developing immunoanalysis of the direct conjugation products of AN. In our preliminary study of the glutathione-AN adduct (10), a synthetic immunogen was prepared by coupling glutathione-AN to hemoglobin as a carrier protein, which was CMA 121532 4 used to immunize Balb/c mice to raise hybridoma. After excluding those which exhibited anti-carrier or cross-reactivity, two monoclonal antibodies were found to show specificity to the glutathione-AN adduct, but only at moderate sensitivity. We have since focused on developing antibodies to intact Hb-AN. These adducts were implicated by the effects of AN on rat erythrocytes and the slow clearance of ^C-AN from red cells (half-life 825 h) (11). The Hb-AN adducts apparently involve modifying the distal heme pocket region to yield new antigenic sites in Hb. Herein we describe the kinetics, level of alkylation, and electrophoresis of Hb-AN adducts, the specificity of two Hb-AN antisera, and immunoassay of the pathological Hb-AN conjugates in human blood exposed to AN in vitro. The antisera are the first of its kind which provide a firm basis for the development of immunoassays of hemoglobin-carcinogen adducts to serve as biomarkers of human exposure where dosimetry has relied mainly on chromatography and mass spectrometry. MATERIALS AND METHODS Chemicals and Animals. Chemical reagents were purchased from Aldrich Chemical Co. (Milwaukee, WI), including AN which contained 35-45 ppm. of hydroquinone monomethyl ether as a polymerization inhibitor. The radiotracer, [2,3-^C]-AN at 0.3 mCi/ml in ethanokwater 1:1, 5,3 mCi/mmol, and purity verified by nuclear magnetic resonance, was obtained from Sigma Chemical Co. (St. Louis, MO), as were other biochemicals including human hemoglobin A (Hb), rabbit anti-Hb antiserum, human serum albumin (HSA), and glutathione (GSH) Another rabbit antiserum, antiHbF, was purchased from CalBiochem (La Jolla, CA). Balb/c mice, male, 6-8 weeks old, were purchased from Charles River Breeding Laboratories (Raleigh, NC), Hemoglobin-Acrylonitrile Adducts in Human Blood. To 2 ml of human blood collected in an EDTA-containing tube (Vacutainer, Becton Dickinson, Rutherford, NJ) was added 40 pi of AN. The reaction was carried out in a shaker-water bath at 37 C for 15 to 180 min. Aliquots of treated blood, 0.2 ml each, were removed and worked up to collect Hb according to a standard procedure (12): centrifuged to pellet RBC, washed in 0.9% NaCl, lysed in water, added Sephadex G25 for 10 CMA 121533 5 min, and centrifuged at 12,000 x g to obtain Hb in the supernatant. The Hb concentration was determined at OD 307 nm. The Hb-AN adducts were separated by PAGE as described below to compare with Hb from a control experiment with no AN added Five human blood samples were used to show reproducibility of the AN exposure study. Also, Hb isolated from fresh human blood was treated with AN in the same manner, yielding the same adducts as shown by PAGE. Kinetics of Alkylation of Blood Thiols and Adducts Obtained Rate studies were carried out in a deaerated solution of 0.5 M Tris-HCl (pH 7.4): methanol 9 : 1 in a shaker-water bath at 37C in triplicates. Three blood thiols: hemoglobin (Hb) and serum albumin (HSA), both 0.6 mM, and glutathione (GSH) 5 mM, were alkylated with either acrylonitrile (AN) or chloroacetaldehyde hydrate (CAA), which was used at a concentration of 150 mM for reaction with Hb and 60 mM for HSA or GSH. The Hb reaction was monitored for up to 3 h by removing 50 pi aliquots to determine the free SH with 4,4'-dipyridinedisulfide (4-PDS) according to the method of Morell et al. (13). Briefly, 4thiopyridone liberated was determined at 324 nm and Hb concentration at 307 nm. The GSH and HSA reactions were followed for up to 1 h by titrating the free SH with Ellman's reagent, 5,5'dithiobis-(2-nitrobenzoic acid) (DTNB), and monitoring the liberated chromophore, 2-nitro-5mercaptobenzoic acid at 412 nm (14). The optical density values were graphed as log C(/C0 vs. time, where Ct is the SH concentration at time t and CQ the initial concentration. Statistical analysis with linear regression was carried out by using standard plotting software. At the completion of alkylation of Hb and HSA the reaction mixtures were centrifuged in a Centricon microconcentrator (Amicon, MW cut-off 10,000) to isolate Hb-AN, Hb-CAA, and HSA-AN. In addition, glycolaldehyde (GA) and D-glucose were allowed to react with Hb under the same conditions used for CAA to prepare the Hb-GA and Hb-glucose adducts. AN Level by Radiotracer in Hb-AN. To 0.5 ml of deaerated Tris-HCl (pH 7.4) - methanol 9:1 was added Hb (19.4 mg, 0.6 mM) and the solution was kept under nitrogen. Triplicate solutions were prepared, two for counting controls, and the third to react with AN: 1.5 pi of [2,3-'4C]-AN (4.6 pg) and 5 pi of AN (4.03 mg) for 150 mM with 878 fold radiolabel dilution. These solutions were capped under nitrogen, incubated at 37C for 0 5 or 3 h, and 50 pi (30 nmol protein) was CMA 1Z1S34 6 transferee! into a vial containing 1.2 ml of 0.14 M of phosphate buffer (PB, pH 7.4) and 0.8 ml of 25% trichloroacetic acid (TCA) to precipitate Hb. To remove any nonspecifically bound AN, Hb was centrifuged, washed, dissolved in phosphate buffer, precipitated by adding TCA, and the process repeated from 3-6 times until the supernatant gave only background counts. The dark-colored Hb preparation was decolorized by adding 30% H2O2. To convert cpm to dpm, one control vial was used to determine the protein background cpm by duplicating the above work-up. To the second control vial was added 1.5 pi of U-^C-D-glucose (0.02 pCi/pl, ICN Biochemicals, Inc., Costa Mesa, CA), and 50 pi (6,660 dpm theoretical) of this solution was added to 2 ml of PB, followed by 8 ml of a complete counting cocktail, which yielded a counting efficiency of 16.4% for Hb. Electrophoresis of Protein Adducts. Electrophoresis was carried out on vertical polyacrylamide gel: 10% T resolving gel in Tris-HCl buffer (pH 9.15) and 4% T stacking gel in Tris-HCl buffer (pH 7.4) in 5 mM Tris-glycine running buffer (pH 8.3) for 6 h at a constant current of 15 mA through the stack and 20 mA through the resolving gel.. Each sample well was loaded with about 25 pg of protein containing 0 002% bromophenol blue tracking dye. For preparative gel separation, the visible Hb-AN bands after electrophoresis were cut out and the protein recovered by electroelution. For separation of the a and 3 chains of Hb and Hb adducts, SDS-PAGE with 12.5% T resolving gel was applied. The Hb adducts were dissolved in 0.02 M Tris-phosphoric acid buffer, pH 6.8, containing 8 M urea, 1 % SDS, 0.02 % 2-mercaptoethanol, and 0,008 % bromophenol blue tracking dye, and electrophoresed in 0.1 M Tris-phosphoric acid, pH 6.8, containing 0.1 % SDS by applying 50 V overnight Immunization of Balb/c Mice and Antisera. The mice were treated as follows. Week 1: intraperitoneal (i.p ) injection of 100 pg of an immunogen (Hb-ANl or Hb-AN2 from preparative gel) in complete Freund's adjuvant (FA). Week 3: 100 pg of immunogen in incomplete FA (i.p.). Week 5: 75 pg of immunogen in incomplete FA (i.p.) Normally at week 6 but sometimes 16.5-20, the immunized mice were bled and assayed by ELISA for antibody activity. The antisera were purified by a standard procedure followed by immunoadsorption on a Hb-Sepharose 4B affinity column prepared according to manufacturer's instructions. Briefly, 0.5 g of CNBr-activated Sepharose 4B gel coupled CMA 121535 7 with 10 mg of Hb at 4C for 20 h was used for 0.6 ml of Hb-AN antiserum. The filtrate enriched in anti-HbAN antibodies was assayed for protein content by calibrated absorbance at 280 nm. ELISA. The ELISA methodology used was adapted after Coleman et al. (15) and Santella et al. (16). Briefly, Nunc-immuno maxisorp plate was coated at 4C overnight with 0.1 ml of 0 (blank control), 1 or 10 jig of antigen per ml of 0.1M sodium carbonate. The plate was washed according to a standard procedure, 0.15 ml of 10% calf serum was added at 25 for 1 h to block nonspecific binding, and the wash procedure was repeated. Sera collected from immunized mice were serially diluted to 50, 250, 1250, 6250, 31250 and 156250 fold in PBS-T containing 0.05% calf serum, and 0.1 ml of these dilutions was added to the microwells, and incubated at 37C for 2 h. After washing, goat anti-mouse (or goat anti-rabbit for rabbit antiserum) IgG-alkaline phosphatase (diluted 1:4000, 0.1 ml) was added and the plates incubated for 2 h at 37C. The wash was repeated, and 0.1 ml of pnitrophenyl phosphate in diethanolamine buffer at pH 9.8 (1 mg/ml) was added, left at 25C for up to 1 h, and the color development was stopped by adding 50 pi of 3N NaOH. The absorbance was read at 410 nm by a microplate reader in triplicate wells for statistical analysis. For competitive ELISA, wells were coated with 0.1 ml of an antigen as above. The antiserum was diluted 1 : 1250, 50 pi of which was mixed with 50 pi of an inhibitor solution which was serially diluted from 1000 to 0.1 p g/ml, and the mixture was added to the wells. Subsequent steps were as those for direct ELISA. RESULTS The structural modification of Hb by AN was investigated as follows. It is already known that AN formed adducts to N-terminal valine of Hb in humans (8). Moreover, among the modified amino acids isolated from hemoglobin of mice treated with ^C-AN, the predominant radioactivity peaks were found to correspond to products formed by addition of AN to cysteine and to histidine (17). Although there are 6 SH groups in the human Hb tetramer, cysteines a 104 and 0112 are "masked" and only 093 has a "free" sulfhydryl group that reacts with non-mercurial thiol reagents (18). Hence, Cys 093 can be singled out for monitoring the major Hb alkylation. On the other hand, there are 38 CMA 121536 8 histidines in Hb, 10 in a- and 9 in (3-chain Locating these cyanoethylated residues in Hb-AN would require a lengthy sequence analysis which may not elucidate its antigenic determinants. Therefore, we chose to compare the rates of S-alkylation of hemoglobin, glutathione, and human serum albumin by AN and CAA in order to probe the structural change in Hb-AN. At time = 0, the mole ratio of free SH of Hb was titrated to be 1.89, or 2.30 when determined after treatment with 50 mM of 2mercaptoethanol for 2 h at 25C. All the rate studies yielded first order plots and the rate constants are compared in Table 1. Since polyacrylamide gel electrophoresis has been used to detect hemoglobin hybrids of a and (3 variants (19), we have worked out a PAGE technique to separate Hb from Hb-AN Figure 1 depicts the electrophoretic behavior of the Hb adducts The Hb-AN mixtures, derived either from human blood exposed to AN for 3 h or from a synthetic reaction with Hb, gave rise to two adduct bands (1A). The first one, identified as Hb-AN 1, was detected in reactions from 10-120 min, but the second band, Hb-AN2, appeared only in the 120 min reaction (IB). The latter was dominant in the Hb-AN 4 h mixture, but no additional band could be discerned from prolonged exposure. Thus, Hb-AN 1 and Hb-AN2 represent two identifiable levels of increasing cyanoethylation of Hb. Under denaturing SDS-PAGE conditions in 8M urea, both Hb and its adducts gave the same a- and [3-chain bands (1C). The number of AN molecules conjugated to one Hb was determined by using [2,3-14C]-AN After exhaustive removal of non-covalently bound AN, the products gave a ratio of 0.7 and 3.1 AN per Hb after 0 5 h and 3 h, respectively. It appears likely that Hb-AN 1 may have about 1 AN per Hb tetramer whereas Hb-AN2 may have 2-3 AN. The specificity and sensitivity of the antisera raised to the two levels of Hb-AN adducts were determined.by performing ELISA. The means of triplicate ELISA data, coefficient of variation < 7%, are used for the following plots. In Fig. 2 are shown the binding curves for various antisera in serial dilutions against 0.1 pg of coated antigen on the microwell: anti-HbAN] vs. Hb and Hb-ANl (2A), anti-HbAN2 vs. Hb and Hb-AN2 (2B), anti-Hb vs. HSA, Hb, Hb-ANl, and Hb-AN2 (2C), and antiHbF vs. fetal Hb-F, Hb, Hb-ANl, and Hb-AN2 (2D). The cross-reactivity of anti-HbAN 1 (2E) and of anti-HbAN2 (2F) with the AN adducts of Hb and HSA are also shown. Fig.3 depicts the competitive ELISA results: anti-HbAN! antiserum in 1:1250 dilution against 1 pg of Hb-ANl coated CMA 121537 9 antigen in the presence of increasing amounts of Hb-ANl, Hb-AN2, Hb-chloroacetaldehyde, and Hb as inhibitors (3 A), a similar competitive inhibition plot for anti-HbAN2 vs. 1 pg of Hb-AN2 coated antigen (3B), a similar competitive inhibition plot for anti-Hb vs, 0.1 pg of Hb plated (3C), and a similar plot for anti-HbANl vs. 0.1 pg of Hb-ANl, with Hb-ANl, Hb-glucose, Hb-glycolaldehyde, and Hb as inhibitors (3D). Fig.4 compares the detection of Hb-ANl over Hb by the Hb-ANl antisera before and after immunopurification on Hb-Sepharose 4B. Two antisera, collected 16.5-20 weeks after immunization, were purified: one from mouse A which was re-immunized 2 weeks before bleeding (4A) and the other from mouse B was not so treated (4B). The immunopurified antiserum from mouse A was used in 1:1250 dilution in a blood immunoassay where Hb-AN adducts were serially diluted as shown in Fig 5A to find the limiting signal-to-noise level for detecting Hb-AN over Hb. In Fig.SB, serial dilutions of the antiserum showed dose response to blood which was exposed to AN from 0 to 180 min. DISCUSSION Antigenic structure of Hb-AN. In this study, we have raised polyclonal antibodies which can distinguish between hemoglobin, Mr 64,500, and its adducts with 1-3 AN conjugated. The structural modification due to size must be minimal since the cyanoethyl group is smaller than even glycine. We surmise that AN has modified Hb antigenicity by presenting an unique determinant in the pathological hemoglobin adduct. There is no information on the antigenicity of any carcinogen-hemoglobin adduct, but human hemoglobin has been shown to possess 10 continuous antigenic sites, five each on the a and the P chain (20) Although none was found for the heme pocket region of P91-105, it was suggested that there should be an antigenic site at (393-98 by extrapolating the conformational location of a myoglobin antigenic site. It is likely that the quaternary tetrameric structure of Hb may have obstructed access to this site. Indeed, the SH group of Cys (393 in deoxy-Hb is known to be shielded by the 3-C-terminal salt bridge between Asp (394 and His pi46, but NMR shows that this is broken upon modification with thiol reagents (21). Thus, the Hb-AN adducts may lead to an CMA 121538 10 accessible antigenic site due to Cys P93 S-cyanoethylation. Furthermore, a comparison of sulfhydryl titration with *4C counting reveals the extent of N-alkylation: 0.3 SH alkylated vs. 0.7 14C-AN conjugated in 0.5 h in forming Hb-ANl, and 1.4 SH vs. 3.1 AN in 3 h in forming Hb-AN2. Since only a weak 4% binding, relative to the p-chain as 100%, was shown by the N-terminus 1-15 in the study of P-chain antigenic sites (22), the terminal Val-AN adduct should not cause substantial change in the antigenicity of Hb. On the other hand, the strongest binding affinity of the p-chain antigenic sites was mapped to the C-terminus peptide 131-146 with 55% of binding (22). In this connection, tandem mass spectrometric analysis of tryptic peptides derived from the in vitro reaction of Hb with styrene-7,8-oxide showed prominent alkylation at the externally accessible His P143 in addition to Cys P93 (23). In light of these observations, we speculate that His pi 43 Nr-cyanoethylation may make significant contribution to the development of unique antigenic determinants in Hb-ANl and 2. This is illustrated in Fig. 6 with the structure of the distal heme pocket region near the P-C-terminus (24), which includes Cys 93 and His 143, for comparing the fate of Hb alkylations by AN and CAA. Our premise is that the rate of S-alkylation is indicative of the stereoelectronic environment of the nucleophilic reaction We have derived structural information from the first order kinetics of blood thiol alkylation shown in Table 1 as follows. A mean value of 2.2 free SH groups per Hb tetramer, including two Cys P93, has been reported (24), which is similar to our titration of Hb-SH. Using glutathione as the standard thiol, the reaction of AN with Cys p93 in Hb was slower by about 20 times, but the corresponding reaction rates with CAA as the substrate were practically the same The difference between AN and CAA in S-alkylation may be attributed to their different physical and chemical properties. Referring to Fig. 6, reaction of the hydrophobic AN with Cys P93 would perturb both Tyr pi45 and the Asp P94 salt bridge, hence a slower reaction relative to glutathione. With the hydrophilic and bisalkylating CAA, it may crosslink the P93 SH group with the side chain of Ser P89 or Lys at P95 or p 144 with less strain. In this context, the fat-transport protein HSA, with the only free SH group at Cys 34, reacted with AN about 41 times faster than the corresponding Hb-AN reaction, but only 5 times faster when CAA was the alkylating agent. The faster rate of HSA cyanoethylation could be explained by the binding of AN to a hydrophobic pocket in the transport CMA 121539 11 protein wherein ready access to nucleophilic attack by the SH group occurred. Concurrent with Scyanoethylation is that Hb also undergoes His Nr-conjugation, possibly at the nearby P143 as shown in Fig.6 analogous to styrene oxide, yielding a less protonated imidazole ring due to the inductive effect of the cyanoethyl group. Thus, increasing levels of cyanoethylation of Cys P93 and His P143 may increasingly disrupt the ionic and hydrogen bonds in the distal heme pocket region of Hb, thereby creating unique antigenic determinants in Hb-AN. For substantiation, gel analysis of the Hb adducts was performed as shown in Fig. 1. Two bands were obtained for the AN conjugates, identical ones being formed from AN reacting with either purified Hb or with Hb in the red blood cell (1 A). Both bands were faster moving than that of Hb from the negative to positive electrode in non-denaturing PAGE. They represent consecutive levels of modification of Hb according to the time study shown in IB The formation of a faster moving component in non-denaturing electrophoresis implies the consumption of a positively charged group, presumably a quartemary nitrogen, hence freeing a carboxylate anion to facilitate migration from the negative to the positive electrode. The combined effects of Cys p93 S- and His P143 Nr- cyanoethylation, progressing with time from Hb-AN 1 (about 1 AN) to Hb-AN2 (about 2-3 AN), are compatible with such migration pattern. That no crosslinking of chains or chain cleavage of the Hb a and p chains has taken place can be seen by detecting the same a and p bands by SDS-PAGE for both the adducts and parent Hb in a denaturing gel analysis (1C) Antisera Binding of Hb Antigens Our hypothesis that increasing levels of cyanoethylation at Cys P93 and at His P143 generate the two isolable antigenic adducts, Hb-AN 1 and Hb-AN2, is supported by the antisera binding studies. The antisera to Hb-AN 1 and Hb-AN2, which are polyclonal in nature and hence should recognize various Hb epitopes, did not bind to Hb as well as to the AN adducts (Fig.2A and 2B). At 1250 fold dilution of anti-HbANl antiserum and 0.1 pg of plated antigen, the ELISA absorbance for Hb-ANl (0.74) was 2.7 times that for Hb (0.27). A similar pattern was given by anti-HbAN2 antiserum with 3.3 times the specificity between Hb-AN2 and Hb (0.63 vs. 0.19) under the same assay conditions. The control mouse sera gave only baseline values for the two AN adducts showing that the Hb-AN antisera were specifically raised to the immunogens inoculated. On CMA 121540 12 the other hand, the commercial anti-Hb rabbit antiserum could recognize Hb, Hb-ANl, and Hb-AN2 with the same absorbance, which was about 2 times that of the unrelated protein HSA (Fig.2C), indicating common Hb epitopes among the three Hb antigens. Another rabbit antiserum to human fetal hemoglobin, anti-HbF, was able to discern some differences among these Hb antigens (Fig. 2D). At 1250 fold dilution of anti-HbF, there seems to be increasing binding in the order of Hb (ct-2&2 tetramer) < Hb-AN2 < Hb-ANl < Hb-F (01272)- It seems that the AN-modified 3 chain in Hb-AN may at certain point resemble the y chain in Hb-F, e.g., His 3143 when cyanoethylated may become uncharged like Ser yl43, inspite of 39 different amino acids between the 3 and y chain (24). The similarity between Hb-AN 1 and Hb-AN2 is indicated by cross-reacting the antisera. The binding specificity of Hb-ANl and Hb-AN2 displayed no substantive difference when assayed with either antiHbAN 1 (Fig.2E) or anti-HbAN2 (Fig.2F), although each antiserum did recognize the corresponding antigen slightly better in all serum dilutions. In respect to binding unrelated proteins, both antisera showed little affinity for either HSA or its AN conjugate, HSA-AN. The cross-reactivities of the Hb-AN antisera with other hemoglobin adducts are shown in Fig. 3. The two adducts of particular concern are Hb-CAA and Hb-GA, which are derived from chloroacetaldehyde hydrate (CAA) and glycolaldehyde (GA), respectively. Both CAA and GA are metabolites of vinyl chloride (26) and have multiple environmental origins as well. Mice exposed to vinyl chloride resulted in alkylation of cysteine and histidine of hemoglobin to yield the 2-oxoethyl derivatives (27). It is therefore gratifying that competitive ELISA with anti-HbANl (Fig.3A) and anti-HbAN2 (Fig 3B) found no specific inhibition by Hb-CAA, which behaved like Hb, showing only low level non-specific inhibition In contrast, the Hb-AN adducts were highly inhibiting to their own antibodies. The Hb-ANl antiserum was inhibited by both adducts equally well, but the Hb-AN2 antiserum was more sensitive to Hb-AN2 than to Hb-ANl, consistent with Hb-AN2 having a higher level of similar antigenic determinants than Hb-ANl. In the presence of anti-Hb, all the adducts of AN and CAA were just as competitive which were better than Hb itself as inhibitor (Fig.3C), indicating the prevalence of the normal Hb epitopes in these adducts When either Hb-GA or Hbglucose was used to inhibit the binding of Hb-ANl by anti-HbANl (Fig.3D), neither was found to be CMA 121541 13 inhibiting. This is not surprising since both of these adducts are formed by a mechanism which is different from that for Hb-AN, involving Schiffbase formation, followed by Amadori rearrangement to carbonyl, and possibly a second addition by a neighboring nucleophile (28) This lack of cross-reactivity with Hb and other Hb adducts suggests that an unique antigenic determinant may have been formed in Hb-AN, provoking specific antibodies. Accordingly, immunopurification of the mouse antisera was illustrated with anti-HbANl using an affinity column of Sepharose 4B covalently linked to Hb. As shown in Fig.4, the binding specificity of the two antisera reflects the immunization history of their sources. The protein concentrations of antiserum A before and after the column treatment were 49.4 and 8.4 mg/ml, respectively In ELISA, the Hb-AN : Hb specificity ratio increased dramatically from 1.02 (before Hb-column) to 16.5 (after) at 20 |ig of antiserum A per well For antiserum B, the protein concentrations were 64.0 mg/ml before and 4.9 mg/ml after the Hb-column Here the specificity ratio increased only moderately from 1.0 (before) to 6.2 (after) even at 121.5 pg of antiserum B per well. Since mouse B was not re-immunized 2 weeks before being bled as did mouse A, the immune system memory effect would favor producing more of anti-Hb in the long run against the more prevalent Hb epitopes, hence less anti-HbANl antibodies were present in antiserum B This explanation is also consistent with the low initial specificity ratio of - 1 for these late antisera, collected in 16.5-20 weeks, instead of a ratio of ~3 for the antisera collected 6 weeks post-immunization The immunopurified anti-HbANl antiserum A displayed useful specificity and sensitivity in immunoassay of human blood as shown in Fig. 5. At 1250 times dilution of the antiserum and serially-diluted Hb antigens, the Fib isolated from human blood given AN for 15 h was clearly discernible from that of the control blood up to 5 ng of antigen plated (Fig. 5A). The positive control in this assay, a 4 h reaction product of Hb with AN, showed slightly better sensitivity which could be differentiated from the negative control Hb at 2 ng plated. This antiserum sensitivity exceeds the lowest level of Hb-AN adduct of 0.02 nmol/g Hb (1.29 ppm) detected in occupationally exposed workers by gas chromatography-mass spectrometry (8). A dose response for blood exposed to AN for 180, 120, 60, 15, and 0 min (control) was obtained at various dilutions of antiserum A (Fig.SB). Thus, the immunopurified antiserum is capable of determining the levels of Hb-AN CMA 121542 14 biomarkers in human blood, making it feasible to scale-up to rabbit antisera as well as to screen for monoclonal antibodies for validation studies. The present investigation has shown the specific immunogenicity of the Hb-AN adducts, characterized their likely antigenic modifications, and verified the sensitivity and specificity of their antibodies. Furthermore, they have illustrated the clinical applicability of specific immunoassay for the detection of carcinogen-hemoglobin adducts for the study of human dosimetry and environmental risk. To our knowledge, this represents the first example of an unique approach to the development of immunoassays for the detection of carcinogenhemoglobin adducts in exposed humans. REFERENCES 1. Chem. & Engineering News 70 (15): 17, 1992. 2. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Man. World > Health Organization, International Agency for Research on Cancer, S7 56, 1987. 3. Health Assessment Document:Acrylonitrile, EPA 600/8-82-007F, U S.Environmental Protection Agency, 1983. 4. Collins, J J, Page, L. C., Caporossi, J. C., Utidjian, H. M., and Saipher, J. Mortality patterns among employees exposed to acrylonitrile. J. Occup. Med. 37: 368-371, 1989. 5. Swaen, G. M., Bloemen, L. J., Twisk, J., Scheffers, T., Slangen, J. J., and Sturmans, F. Mortality of workers exposed to acrylonitrile. J. Occup. Med., 34: 801-809, 1992. 6. Environmental Health Criteria 28 Acrylonitrile. World Health Organization, Geneva, 1983. 7. Skipper, P. L., and Tannenbaum, S. R. Protein adducts in the molecular dosimetry of chemical carcinogens Carcinogenesis, 11. 507-518, 1990. 8. Bergmark, E., Calleman, C. J., He, F., and Costa, L G. Determination of hemoglobin adducts in humans occupationally exposed to acrylamide. Toxicol. Appl. Pharmacol., 120. 45-54, 1993 CMA 121543 15 9. Fennel, R. F., Kedderis, G. L., and Sumner, S. C. J. Urinary metabolites of [1,2,3-^C] Acrylonitrile in rats and mice detected by nuclear magnetic resonance spectroscopy. Chem. Res. Toxicol., 4: 678-687, 1991. 10. Wong, J. L., Ma, F. F., and Zhang, Y. Antibodies to acrylonitrile-glutathione conjugate Antibody Immunoconjug. Radiopharmaceut. 3: 89, 1990. 11. Farooqui, M. Y. H., Mumtaz, M. M., Ghanayem, B. I., and Ahmed, A. E. Hemoglobin degradation, lipid peroxidation, and inhibition of Na+/K+-ATPase in rat erythrocytes exposed to acrylonitrile. J. Biochem Toxicol., 5. 221-227, 1990, and previous papers. 12. Tomqvist, M., Mowrer, J., Jensen, S., and Ehrenberg L. Monitoring of environmental cancer initiators through hemoglobin adducts by a modified Edman degradation method. Anal. Biochem., 154. 255-266, 1986. 13. Ampulski, R. S., Ayers, V. E., and Morell, S. A. Determination of the reactive sulfhydryl groups in heme proteins with 4,4'-dipyridinedisulfide. Anal.Biochem., 32: 163-169, 1969. 14. Sedlak, J., and Lindsay, R. H. Estimation of total, protein-bound, and nonprotein sulfhydryl groups in tissue with Ellman's reagent. Anal.Biochem , 25: 192-205, 1968. 15 Coleman, J. W., Yeung, J. H. K., Tingle, M. D., and Park, B. K. Enzyme-linked immunosorbent assay (ELISA) for detection of antibodies to protein-reactive drugs and metabolites: criteria for identification of antibody activity. J. Immunol. Methods, 88: 37-44, 1986 16. Santella, R. M., Chen, D. L., and Dharmaraja, N. Monoclonal antidbodies to a benzo[a]pyrene diolepoxide modified protein. Carcinogenesis, 7: 441-444, 1986. 17. Bryant, M S ., and Osterman-Golkar, S. M. Hemoglobin adducts as dosimeters of exposure to DNA-reactive chemicals CUT Activities, 11: 1-9, 1991 18. Okonjo, K., Taiwo, A., Balogun, M and Ekisola, O. B. Reactivities of the sulfhydryl groups of dog hemoglobin. Biochim. Biophys. Acta, 576: 30-38, 1979. 19. Bernstein, S. C., and Bowman, J. E. The demonstration of asymmetric hemoglobin hybrids by polyacrylamide electrophoresis, Biochim. Biophys. Acta, 427: 512-519, 1976. CMA 121544 16 20. Oshima, M., and Atassi, M. Z. Generation of species-specific antihemoglobin antibodies by immunization with synthetic peptides of human hemoglobin. J. Protein Chem., 8. 767-778, 1989. 21. Garel, M. C., Caburi-Martin, J., Domenget, C., Kister, J., Craescu, C. T., Poyart, C., and Beuzard, Y. Changes of polymerization and conformation of hemoglobin S induced by thiol reagents. Biochim. Biophys. Acta, 1041. 133-140, 1990, and previous papers. 22. Yoshioka, N., and Atassi, M. Z. Antigenic structure of human hemoglobin. Localization of the antigenic sites of the b-chain in three host species by synthetic overlapping peptides representing the entire chain. Biochem. J., 234: 441-447, 1986. 23. Kaur, S., Hollander, D., Haas, R., and Burlingame, A. L. Characterization of structural xenobiotic modifications in proteins by high sensitivity tandem mass spectrometry. J. Biol. Chem., 264: 16981-16984, 1989. 24. Dickerson, R. E., and Geis, I. Hemoglobin: structure, function, evolution, and pathology, pp. 32-69. Menlo Park, Ca., Benjamin/Cummings, 1983. 25. Evelo, C. T. A, and Henderson, P. T. H Influence of glutathione on the formation of cysteine alkylation products in human hemoglobin. Toxicol. 52:177-186, 1988. 26. O'Neill, I., Barbin, A., Friesen, M., and Bartsch, H. Reaction kinetics and cytosine adducts of chloroethylene oxide and chloroacetaldehyde: direct observation of intermediates by FT-NMR and GC-MS In: B Singer and H. Bartsch (eds ), The role of cyclic nucleic acid adducts in carcinogenesis and mutagenesis, IARC Sci. Publ. No. 70, pp. 57-73, Lyon, IARC, 1986. 27. Osterman-Golkar, S., Hultmark, D,, Segerback, D., Calleman, C. J., Gothe, R., Ehrenberg, L., and Wachtmeister, C, A. Alkylation of DNA and proteins in mice exposed to vinyl chloride. Biochem. Biophys. Res. Commun., 76, 259-266, 1977. 28. Nacharaju, P., and Acharya, S Amadori rearrangement potential of hemoglobin at its glycation sites is dependent on the three-dimensional structure of protein. Biochem , 31 12673-12679, 1992. CMA 121545 Table I. Pseudo 1st order rate constants of S-alkylation of thiols at pH 7.4 and 37C by linear regression of SH group remaining. Thiols CH2=CH-CN (AN) k/10^s_1 (r2) Cl-CH2-CH(OH)2 (CAA) k/lO'V1 (r2) Hba GSHb HSAC 1.0 (0.98) 19.8 (0.98) 41.4 (0.98) 17.8 (0.97) 18.1 (0.99) 89.8 (0.97) a[Hb] = 0.6 mM, [AN or CAA] = 150 mM, reaction time 120 min b [GSH ] = 5 mM, [AN or CAA] = 60 mM, reaction time 12 min c [HSA ] = 0.6 mM, [AN or CAA] = 60 mM, reaction time 6 min CMA 121546 Legends of Figures Fig. 1 Comparison of non-denatured PAGE and SDS-PAGE of hemoglobin- acrylonitrile adducts. (A) non-denatured PAGE with 10% T resolving gel. Lane 1 Hb isolated from RBC, Lanes 2 and 3 replicates of Hb isolated from blood pre-treated with AN, 3h, 37, containing Hb-AN adducts, Lane 4 Hb standard. Lane 5 Hb treated with AN, 3 h, 37 contained Hb-ANl (band ahead of Hb) and Hb-AN2 (fastest-moving band). (B) PAGE of the time course of Hb-AN adduct formation with 10% T resolving gel. Lane 1 BSA standard, Lane 2 Hb standard, Lanes 3,4, and 5 Hb-AN reaction at 37 for 10, 30, and 120 min, respectively. (C) SDS-PAGE with 12.5% T resolving gel. Lane 1 HbANl isolated from preparative PAGE, Lane 2 Hb-AN2 isolated from preparative PAGE, Lane 3 Hb pretreated with CH3CHO as control, Lane 4 Hb standard. Fig. 2 Antisera binding of Hb and related antigens: anti-HbANl (A), antiHbAN2 (B), anti-Hb (C), anti-HbF (D), cross-reactivity of anti-HbANl (E) and anti-HbAN2 (F). Fig. 3 Competitive inhibitions of anti-HbANl (A), anti-HbAN2 (B), anti-Hb (C), and anti-HbANl (D) antisera in binding to Hb and adducts. Fig. 4 Immunopurification of anti-HbANl antisera on Hb-Sepharose 4B column. (A) Balb/c mouse A was bled 20 weeks after 1st immunization, re-immunized with HbANl in incomplete FA 2 weeks before and tail injection 4 days before blood collection . (B) Balb/c mouse B was bled 16.5 weeks after 1st immunization, tail injection 4 days before blood collection (B). CMA 121547 Fig. 5 Immunoassay of Hb-AN in human blood with immunopurified Hb-ANl antiserum. (A) Serial dilutions of antigen. (B) Serial dilutions of antiserum against blood exposed to AN for various lengths of time. Fig. 6 Nucleophilic additions of AN and CAA in the distal heme pocket region near the p-C-terminus of human hemoglobin. CMA 121548 Fig. 1 Comparison of non-denatured PAGE and SDS-PAGE of hemoglobinacrylonitrile adducts. (A) non-denatured PAGE with 10% T resolving gel. Lane 1 Hb isolated from RBC, Lanes 2 and 3 replicates of Hb isolated from blood pre-treated with AN, 3h, 37, containing Hb-AN adducts, Lane 4 Hb standard, Lane 5 Hb treated with AN, 3 h, 37 contained Hb-ANl (band ahead of Hb) and Hb-AN2 (fastest-moving band). (B) PAGE of the time course of Hb-AN adduct formation with 10% T resolving gel. Lane 1 BSA standard, Lane 2 Hb standard, Lanes 3,4, and 5 Hb-AN reaction at 37 for 10, 30, and 120 min, respectively. (C) SDS-PAGE with 12.5% T resolving gel. Lane 1 HbANl isolated from preparative PAGE, Lane 2 Hb-AN2 isolated from preparative PAGE, Lane 3 Hb pretreated with CH3CHO as control, Lane 4 Hb standard. AB 12345 12345 1234 CMA 121549 u HbANf A 13 B Ag 1 ug/ml -log 1 /serum dilution Control: normal mouse serum vs HbANI Ag: 1 ug/ml Hog 1 /asrum dilution Control: normal mouse serum vs HbAN2 Ag 1 ug/ml log 1 /serum dlutlon O Fig. 2. Antisera binding of Hb and related antigens byanti-HbANI (A), anti-HbAN2 (B), > to anti-Hb (C), anti-HbF (D), and cross-reactivity of anti-HbANI (E) and anti-HbAN2 (F). CCoJJll % inhibition % inhibition ^ irHbANl lOug/ml log Inhibitor concn, ug/ml Ag: HbAN21 Oug/ml log Inhibitor concn, ug/ml Ag: Hb 1 ug/ml log inhibitor concn, ug/ml Ag:HbAfsM lug/ml log Inhibitor concn,ug/ml Fig. 3. Competitive inhibition of anti-HbANI (A), anti-HbAN2 (B), anti-Hb (C), and anti-HbANI (D) antisera in binding to Hb and adducts. ELISA O.D. Fig. 4. Immunopurification of anti-HbANI antisera on Hb-Sepharose 4Bcolumn (A) Balb/c mouse A was bled 20wk after 1 st immunization, re-immunized with HbANI in incomplete FA 2 wk before and tail injection 4 days before blood collection. (A) Baib/c mouse B was bled 16.5 wk after 1 st Immunization, tali Injection 4 days before blood collection. 2.0 r Q.1.5 wi.o 111 0.5 0.0 1.5 2.5 Ag: 0.5ug/ml B Ag: Hb-AN In Blood 180 min exposed to AN 120 60 15 control blood 3l5 4JS 5.5 log 1/serum dilution Fig. 5. immunoassay of Hb-AN in human blood with immunopurified Hb-ANI antiserum. (A) Serial dilutions of antigen. (B) Serial dilutions of antiserum against blood exposed to AN for various lengths of time. AN CAA CH2--CH-CN hydrophobic 98--SH -------- 143--Nr ------- Cl--CH2-CH(OH)2 98--SH ------- hydrophilic 143--Nr ------- 98--S--CH2CH2CN 143--Nr--CH2CH2CN OH 98--S-CH2--CH-X X I 143--Nr--CH2CH--OH X = 89- 0; 95-, 143- NH Fig. 6 Nucleophilic additions of AN and CAA in the distal heme pocket region near the P-C-terminus of human hemoglobin CMA 121554 20 Practical Applications of Biomarkers in the Study of Environmental Liver Disease Carlo H, Tamburro John L. Wong Essential Background Introduction The concern over adverse health effects caused by exposure to environmental toxins and unique role of the liver in the metabolism of these toxicants makes biomarkers of liver metabolism and disease clinically very important. He patic biomarkers are needed to assess exposure, identify subclinical hepatic injury, monitor for chronic disease, assess long-term risk, and allow for pre ventive intervention before liver injury progresses to an irreversible stage. Examination of these hepatic biomarkers in the clinical, occupational, and environmental settings will verify their ability to detect specific exposures or adverse health effects. Also, under proper conditions, they can provide the means to reassure exposed individuals of no future adverse health risk. Liver Structure Hepatic biomarkers, especially enzymatic ones, are biologically related to the anatomical architecture of the liver. The architectural unit of the liver has been described classically as subunits of hexagonal lobules, 1 -2 mm in diameter, situated about a central vein. The boundaries of these lobules are demarcated by portal tracts, composed of two blood supplies (arterial and venous) and the biliary excretory system. These portal tracts approximately Molecular Epidemiology: Principles and Practices Copyright 1993 by Academic Prcw, Inc. All rights o( reproduction m any form reaerved, 517 CMA 121555 518 Carlo H. Tamburro and John L. Wong follow the angles of the hexagons (Figure 20.1A). Blood flows to the paren chymal tissue from the portal triads at the periphery of the classic lobule and exits via the central veins. About one-third of the blood supply to the lobules is provided by the hepatic artery; the remainder is supplied by the portal vein. Because hepatic blood flow is such an essential component of hepatic func tion, the portal triads are considered the center of the functional unit, called the asinus. The parenchymal cells surrounding this vascular distribution are divided into three zones: Zone 1, closest to the arterial and portal blood sup plies; Zone 2, between Zones 1 and 3 in the center of the parenchyma; Zone 3, surrounding the central vein region (Figure 20.1B). Function The liver has multiple functions. The primary role is metabolic, involv ing uptake of substrates for storage, metabolism, and distribution via the blood and bile. Its second major role is conversion of xenobiotic agents and endogenous materials into excretable compounds. This second metabolic function can sometimes convert otherwise harmless compounds into toxic ones. Third, the liver is a major site for clearance of bacterial and other ma- A. - Portal Triads O > Central Vein Midzone or Lobular Area Pericentral Area ,,0 >S'Zone 3 Ascinus Blood Flow From v*` Portal Area To Control VBins FIGURE 20.1 Normal architecture of the liver illustrating the structural concept of hepatic lobule (A) and the functional concept of hepatic ascinus (B). CMA 121556 20 Biomarkers in the Study of Environmental Liver Disease TABLE 20.1 Useful Biochemical Markers of Liver Function and Injury1 Enzymes Alanine aminotransferase (ALT) Aspartate aminotransferase (AST) Gammaglutamyl transpeptidase (GGT) Lactic acid dehydrogenase (LDH) Alkaline phosphatase (AP) Proteins Albumin (Alb) Prothrombin (PT) Bile acids (BA) Cholyglycine Total bile adds Bilirubin (TB) Conjugated/direct (DB) Unconjugated/indirect (IB) Metabolic/Physiologic Aminopyrine breath test (ABT) Indocyanine green clearance (ICG) ' Referred to as biochemical liver tests (BLTs), S19 terials by its phagocytic activity via the reticular endothelial system (RES), which also encompasses parts of the immune system. The metabolic role of the liver makes it vulnerable to toxic injury from exposure to a variety of metabolic and xenobiotic insults. Hepatic biomarkers can identify this injury and, in many cases, characterize the location, severity, and nature of the dam age. Toxic exposure may manifest itself in three forms: enzyme induction, hepatocellular damage, and cholestasis. Currently useful biochemical mark ers (clinical tests) that identify these liver responses to toxins are shown in Table 20.1. Biomarkers Currently in Use or under Consideration Exposure Detection Cellular Enzymes Xenobiotics are known to undergo hepatic biotransformation and pro duce bioactive, rather than detoxified, metabolites. Although most bioacti vation has been demonstrated in animals, acetaminophen, aflatoxin Bi, ar senic, carbon tetrachloride, halothane, isoniazide, and vinyl chloride have been shown to produce acute or chronic disease and malignant transforma tion in humans. The striking characteristic of biotransformation is the en hancement of cellular enzyme activity. This activity may be used to determine acute and chronic exposure to xenobiotics that exceed the background en zyme induction caused by natural products in the diet and inhaled air. These CMA 121557 520 Carlo H. Tamburro and John L. Wong cellular enzyme markers, as presently characterized, are generic in response to xenobiotic exposure. Oxidative induction: MFO-P450 Cytochrome P450, the monooxygenase system, is a family of mixed-function oxidase (MFO) enzymes with unusual versatility because of the multiplicity of forms. In humans, Wang et al. (198 3) have identified six P450s shown to metabolize different compounds. As shown in Table 20.2, these human P450s have different but overlapping broad substrate specificities. Chemical induction of P450 has been divided arbitrarily into two classes, the PB type (phenobarbitol-induced) and the 3MC type (3-methylcholanthrene-induced), on the basis of the induction of characteristic P450 isozymes and the mechanism of induction. For example, a dioxin derivative, TCDD, belongs to the 3MC class (Le Provost et al., 1983). This type of cellular enzyme biomarker may be used to identify spe cific xenobiotic injury for which organ tissue is available. Indirect measure ment of these types of enzyme biomarkers can be done by metabolic clear ance tests that are surrogate measures of oxidation. Surrogate measure of oxidation; aminopyrine breath test More applicable means of indirect measurements of the hepatic P450 oxidation system are metabolic clearance tests. There are a number of such tests, such as the ami nopyrine breath test (ABT) or caffeine clearance (Baker et al., 1983). ABT has had the most extensive use in xenotoxic assessment. Aminopyrine is ad ministered orally and oxidized primarily in the liver, liberating formalde hyde, which undergoes subsequent metabolism to C02. This carbon atom is labeled with either 14 C or 13 C and recovered in the breath, allowing an indi rect measurement of P450 activity as a reflection of the functional liver mass. TABLE 20.2 Human Hepatic Microsomal P450 Induction in Xenobiotic Metabolisin' Xenobiotic P4S0 -2 P4S0 -3 P4S0 -4 P450 -S P4S0 -7 P4J0 -8 Acetanilide Benzo(a)pyrene ^Benzphetamine Trichloroethylene 1-Naphthylamine Z-Naphthylamine L L M L L L ML L MMM HHH LLL -- UM -- L "" LL MM HH LL -- L/M --L Source: Adapted from Wang et al. (1983). * Relative rates of metabolism (nmol/min/nmol P4J0): L, low (s 0.09); M, medium (0.1- 0.99); H, high (a 1.0). CMA 121558 20 Biomarkers in the Study of Environmental Liver Disease 521 Lidocaine clearance Like ABT, lidocaine clearance is an induced measurement of P450 activity which, in turn, reflects the functional hepatic mass. Lidocaine is aminoethylacetanilide, which undergoes rapid N-deacylation via the hepatic cytochrome P450 system to yield several metabolites, princi pally monoethyglycinexylidide (MEGX). The concentration of MEGX in se rum before and 15 min after iv administration of 1 mg/kg lidocaine can be determined by a fluorescence polarization immunoassay system. Detoxification induction: glutathione Glutathione (GSH) is the major en dogenous protective substance that participates in covalent binding of reac tive electrophilic metabolites and in reducing peroxides. The enzymes (trans ferase) involved in catalyzing the glutathione detoxification effects are also potential biomarkers. These transferases comprise a family of enzymes with overlapping but distinct substrate specificities (Vander Jagt etal., 1985). For example, glutathione S-transferase (GST) is an enzyme involved in catalyzing the detoxification of potential diol-epoxide carcinogenic or mutagenic me tabolites of polycyclic aromatic hydrocarbons (Glatt et al., 1983). Metabolic and Physiologic Tests Clearance tests: indocyanine green clearance Hepatic function and reserve are related to the ability of the liver to clear substances from the blood. He patic extraction (intrinsic clearance) is a determinant of bioavailability that can be measured by the systemic clearance of liver specific substances, for example, iodocyanine green (ICG), galactose, or bile acids. There is good correlation between systemic clearance of ICG and early xenotoxic liver in jury. The measurement of hepatic extraction is highly correlated to hepatic blood flow. These substances are given intravenously and their clearance rate is determined by small serial blood samples over 10-15 min (Tamburro and Liss, 1986). Lower clearance of these substances results from lower extrac tion due to intrahepatic blood flow changes secondary to toxic liver injury. Bile acids Bile acids are naturally produced hepatic substrates, cleared solely by the hepatocytes, and have shown good correlation with early xenobiotic liver injury. The advantage of this study is that no injection of any substance is required. Bile acids are measurable in serum by radio immunoassays. Proteins: Antigens and Antibodies Some environmental hazards are biologic, for example, viruses that can occur concomitantly with xenobiotic exposure and act as confounders or cotoxins to the liver. Specific antigen-antibody markers are available for identification of exposure and active hepatocellular injury due to a major hepatic virus (Tamburro, 1991). The human virus (e.g., hepatitis B and C) CMA 121559 522 Carlo H. Tamburro and John L. Wong plays a very important role in the causation of liver cancers associated with natural and synthetic xenobiotics (e.g., aflatoxin). Hepatic fibrosis is the alternative repair mechanism (as opposed to re generation) for hepatic injury. It is the key indicator of serious hepatic injury after xenobiotic exposure. The detection of hepatic fibrosis is especially im portant in low-level chronic and subclinical exposure. Noninvasive serum markers of hepatic fibrosis showing early promise include N-terminal pro peptides of Type III procollagen and Type IV collagen fragments. These markers of serum concentration correlate well with gene expression (messen ger RNA levels) in dimethylnitrosamine- and carbon tetrachloride-induced hepatic fibrosis (Hayasaka et al., 1988; Salvolainen et al, 1988). Further studies are needed to establish baseline variations and to characterize their course in various forms of human liver injury. Adducts Assays for xenobiotic binding to GSH, various proteins, and DNA are under development and field application. Antibodies, polyclonal and mono clonal, have been in development for a number of hepatic xenobiotics, for example, aflatoxin B, (Sabbioni et al, 1990) and acrylonitrile (Wong et al., 1990). Immunoassays for these xenobiotic antibodies are being developed to detect adducts to GSH, albumin, hemoglobin, and DNA. Clinical trials for each type of adduct provides different information with respect to degree, duration, and dose of exposure. Monoclonal antibody assay for specific hepatic xenobiotics can be applied in many ways, as shown in Table 20.3 (Perera and Weinstein, 1982; Perera et al, 1986). Genetic Markers Restriction fragment length polymorphism Gene susceptibility for the de velopment of alcohol liver injury is suggested, since only a minority of alco holics develop cirrhosis. Using restriction fragment length polymorphism (RFLP) testing of the Type 1 collagen gene, the collagen type most prevalent TABLE 20.3 Useful Molecular Monitoring of Hepatic Xenobiotics by Adduct Formation Biologic site Adduct Half-life Metabolic Protein Cell Nucleus GSH Conjugates Albumin Hemoglobin (RBC) DNA Hours Days Weeks Months/Years (ifno repair) CMA121560 20 Biomarkers in the Study of Environmental Liver Disease S23 in human cirrhosis (Winer etal, 1988), specific agent identification can be made. A RFLP exists when two different but normal nucleotide patterns exist at the same site in the genomic DNA. This difference can be recognized by digestion with a bacterial restriction endonuclease. White blood cell DNA is obtained from exposed individuals, digested with two restriction enzymes, and hydridized with two Type I collagen DNA probes to reveal an RFLP. Six haplotypes or patterns of polymorphisms have been found in alcoholexposed individuals, based on the presence or absence of these two poly morphisms. One haplotype is found more frequently in cirrhotic alcoholics than in alcoholics without cirrhosis or in controls. If family studies confirm such a linkage, this type of identified polymorphism could provide a means of identifying individuals at risk of developing liver disease (cirrhosis) when exposed to alcohol. The ability to identify various capabilities of metabolism of xenobiotics, as well as the propensity for formation of collagen after in jury, can be used as a generic molecular marker for low-dose xenobiotic hepatic exposure or injury. Activated proto-oncogenes and inactivated tumor suppressor genes Acti vated transforming genes (oncogenes) have been found in a number of hu man tumors by use of assays in which transformed foci result from transfec tion of tumor DNA into NIH3T3 cells. One striking fact that has emerged from screening transfecting DNA is that, for both human and rodent tumor DNA, the transforming genes are virtually all related to the ras oncogene family. Activation of ras proto-oncogenes by a carcinogenic agent often in volves base substitutions at codons 12 and 61. Some examples of activated ras oncogenes found in liver tumors are given in a review by Harris (1991): aflatoxin B,-induced GM --> T and Gw A mutations in Ki-ras of rat; benzidine-induced C1'1 --> A mutation in Ha-ras of mouse; and urethaneinduced A182 --> T mutation in Ha-ras of mouse. In addition, Wogan and coworkers (McMahon et al., 1990) reported the presence of Ki-ras oncogenes (G C --> A-T or G-C --> T*A in codon 12) in the liver of flounders with hepato cellular carcinomas that were taken from a contaminated site in Boston Har bor, DNA samples from histologically normal liver of flounder from a less polluted site showed only wild-type DNA sequences at codon 12 of Ki-ras. Since the ras oncogene is involved in early, late, and metastatic stages of car cinogenesis, determination of ras mutation in a liver biopsy sample may be used as a biomarker of susceptibility (in the absence of liver impairment) or of effect (in conjunction with liver tumor). In contrast to proto-oncogenes, tumor suppressor genes are cellular genes that regulate cell growth, induce apoptosis (programmed cell death), and maintain genomic stability. Inactivating the normal allele will cause dysregulation of growth and differentiation pathways, enhancing cell transfor CMA 121561 S24 Carlo H. Tamburro and John L. Wong mation. In this sense, it is far more likely to disable a gene than to activate a proto-oncogene by point mutation. For example, the ras gene is activated by mutation in a few specific codons only. Further, some mutated forms of p53 are transforming oncogenes. In sum, the pS3 tumor suppressor gene has shown the best association with human liver cancers (Harris, 1991). The ma jority of the mutations in human tumors occur in exons 5-8 of the pS3 gene, where the hot spots are grouped in the coding region 248-282. A pS3 hot spot mutation in hepatocellular carcinoma has been linked to aflatoxin ex posure and hepatitis B virus. Codon 249 mutation of the pS3 gene is strongly associated with high aflatoxin exposure and identifies an endemic form of hepatocellular cancer (Ozturk et al., 1991). Detections of mutations in ras and p53 in small tissue samples are now made possible by polymerase chain reaction (PCR) technology. PCR rapidly is becoming the preeminent area of diagnostic hepatology with respect to environmental hazards, including vi ruses. Often, exposure to environmental xenobiotics is complicated by latent hepatotoxic agents, especially hepatitis types B, C, and D. Especially relevant is the ability of hepatitis B to become integrated into human DNA and no longer be identifiable by standard immunologic markers. PCR application of specific hepatic viral RNA and DNA species allows detection of such latent confounders, provides more accurate classification of the exposed individual, and has the ability to determine whether any synergistic effect may occur due to the dual hepatotoxin exposure. PCR allows vast amplification of specific DNA species of interest (10*-fold increases are routine). DNA can be de tected from nucleotide samples of less than 10 pg; therefore, the technique is applicable to human liver biopsy samples. Analytical Techniques Metabolites in body fluids Oxygenated derivatives of environmental carcinogens, such as benzo[a]pyrenes and aflatoxin in blood and urine, are direct exposure markers. Analytical techniques involving gas chromatogra phy, mass spectrometry (GC-MS), high-performance liquid chromatography (HPLC), and thin-layer chromatography (TLC) with fluorescence detection are sensitive to nanogram levels. However, blood and urinary metabolites of volatile chemicals (e.g., vinyl chloride and acrylonitrile yielding watersoluble thio acids) are not readily quantitated under low-level exposure con ditions. Direct analysis of hepatotoxin metabolites is most useful in heavy metal (e.g., iron, arsenic, copper) dose-response study. Although atomic ab sorption spectroscopy is widely used to determine metals at ppb concentra tion in biologic specimens, this method docs not provide species information because the analyte is determined at the atomic state. To this end, absorptive stripping voltammetry is being developed for metal speciation, for example, speciation of nickel(II)-hi$tidine as a biomarker for nickel exposure (Wu and Wong, 1991). CMA 121562 20 Biomarkers in the Study of Environmental Liver Disease S25 Effect/Diagnostic Application of Biomarkers of Hepatic Effects Caused by Xenobiotic Exposure Hepatic biomarkers of environmental exposure can identify three major outcomes: acute, chronic, and latent. Chronic and latent diseases (e.g., can cer) are complex multistepped processes. Therefore, any single biomarker is likely to identify only one or a few of the various steps. However, two major steps are essential to all chronic and carcinogenic processes: tissue injury and cellular repair or regeneration. In acute or chronic hepatic exposure, only incomplete cellular repair of toxic injury allows identification of the expo sure. Without detectable injury, there is no clinically meaningful exposure. In the carcinogenic process, malignant transformation cannot occur without both tissue injury and cellular regeneration (i.e., dead cells or cells unable to replicate cannot become malignant). Therefore, the most useful hepatic biomarkers assess injury or identify cellular replication. Presently, the most frequently used markers of hepatic injury are enzy matic or biochemical ones (Table 20.1). These markers are relatively nonspe cific with respect to etiology, but are the clinical standard for the absence or presence of hepatic injury. Depending on the degree (level) of hepatic injury, these markers have relative diagnostic usefulness in the detection of heparotoxic exposure (Table 20.4). Level 1: adaptive response At this level, clinical exposure is followed by metabolic or biologic changes that result in no injury, for example, gamma glutamyl transpeptidase (GGT) enzyme induction after alcohol exposure or P450 induction (via abnormal ABT) after synthetic hydrocarbon exposure. Enzyme induction is a physiologic or structural adaptation. There is no cel lular damage or death. All other biochemical liver tests (BLTs) and tests of synthetic function are normal. Level 2: acute injury, mild This level of clinical exposure causes cellular changes that are nonprogressive, reversible, without disruption of cellular TABLE 20.4 Diagnostic Effect: Biomarkers for Various Outcomes* Adaptive response: Biological change, no injury (P-4JOs, GGT, ABT) Acute injury: Mild (AST/SGOT, ALT/SGPT, ICG) Acute injury: Severe (total bilirubin, albumin, FT, transferrin) Chronic injury (cholyglycine, alkaline phosphatase, procollagen III) Disease Nonmalignant: Cirrhosis (albumin, PT, cholyglycine, alkaline phosphatase) Malignant (a-fetoprotein, lactic dehydrogenase) ' Markers in parentheses represent those that are more characteristic of a condition. CMA121563 526 Carlo H. Tamburro and John L. Wong function, and without evidence of residual injury, for example, alcoholinduced fatty liver with cellular enzyme leakage (shown by increased ala nine aminotransferase (ALT)/aspartate aminotransferase (AST), indocyanine green (ICG) clearance). All other BLTs are normal. At this level, there is structural adaptation without functional impairment or permanent architec tural damage, even though some histologic changes are identifiable. Level 3: acute injury, severe Clinical exposure to this degree causes dis ruption of cellular function and leaves residual evidence of liver injury, for example, carbon tetrachloride exposure causing cellular necrosis (shown by increased ALT/AST), disrupted function (by elevated bilirubin), and synthe sis (by lowered albumin). There are specific histologic changes (by pericentral necrosis) and later fibrosis and scarring (residual injury shown by elevated alkaline phosphatase). True cellular injury has occurred, with repair. Even with repair, there is residual evidence of damage without major architectural changes. Genetic injury may have occurred but is unlikely to be clinically significant or permanent. Level 4: chronic injury Clinical exposure under these circumstances causes cellular disruption and architectural changes that reduce functional hepatic capacity, for example, vinyl chloride-induced fibrosis (shown by procollagen III, IV) and portal hypertension (evidenced by elevated cholylglycine and alkaline phosphatase with decreased ICG clearance). The clinically sig nificant injury is permanent, often with characteristic structural changes. Ge netic injury can occur with risk of cancer development. Level 5: disease At this stage, clinical exposure has caused permanent structural damage, impaired organ function, and reduced capacity. Genetic injury can cause disruption of cellular control and, with active regeneration, may ultimately lead to malignant transformation, for example, chronic viral hepatitis with cirrhosis (shown by HBV-DNA, HBsAg), primary hepatocel lular carcinoma, vinyl chloride fibrosis, peliosis hepatis, and hepatic angio sarcoma. Changes in these molecular and BLT markers correlate with the degree and type of hepatic tissue response to various environmental hepatotoxins (Liss etal., 1985). Susceptibility Tests of susceptibility to hepatic xenobiotic injury are governed by the ability of the liver to metabolize and detoxify reactive metabolites. Therefore, mark ers that identify the oxidative pathway of a xenobiotic or the degree of pu nitive metabolite detoxification are the best ones to use to assess individual susceptibility. Although some markers (e.g., P450 levels) may indicate in- 1I -r CMM2A564 20 Biomarkers in the Study of Environmental Liver Disease S27 creased oxidation and others (e.g., GSH) indicate detoxification, none of the presently available hepatic markers are sufficiently specific or sensitive to identify the metabolic capabilities of an individual. Monoclonal antibodies or adducts with GSH, albumin, or hemoglobin are able to identify exposure or reactive metabolites of specific hepatotoxins. However, tests of future risk must be able to identify specific hepatic changes that will make an adverse outcome more likely. Such hepatic changes include scar formation (fibrosis, i.e., incomplete repair), active regeneration, DNA adduct formation, and on cogene activation. These changes are all related to increased risk of cancer development. In exposed individuals, for example, identification of pS3 and ras gene activity requires concurrent assessment of the putative metabolite (e.g., by adduct occurrence) with the histologic and clinical markers (BLT and sero logic) of hepatotoxic injury. Without this form of combined assessment, the differentiation of an exposed individual with subclinical hepatic injury and competent reparative capability from a susceptible individual with genetic injury and high-risk outcomes cannot be accomplished effectively. The detection of viral confounders (hepatitis B, C, and D), which en hance susceptibility, has improved vastly with the use of PCR. These biomarkers provide proper classification of individuals with chronic liver disease who have exposures to various hepatotoxic chemicals and allow causal dif ferentiation [e.g., Vietnam veterans with viral hepatitis B and dioxin expo sure (Tamburro, 1992) and alcoholics with viral hepatitis C (Mendenhall etai, 1991)]. Case Studies Aflatoxtn (Hepatocellular Carcinoma): Natural Environmental Toxin Human hepatocellular carcinoma (HCC) has been causally associated with chronic active hepatitis (CAH), secondary to hepatitis B virus (HBV) and moldy food grain contaminated by aflatoxins (AF), mycotoxin metabolites of the Aspergillus fungus (Harris, 1990). HCC is prevalent in certain regions of Africa and Asia, where HBV carriers and dietary AF, typically AFB,, are common. AFB, has been shown to be a potent carcinogen; its activity de pends on the balance of AFB, metabolism between oxidation by specific cy tochrome P450 phase I isozymes, that produce either the less toxic hydroxylated AFB, products or the carcinogenic 2,3-epoxide, and conjugation by glutathione. This balance has been shown (Schrager etai, 1990) to shift with nutritional modulation and chemical intervention, both of which may en hance or diminish liver cancer induced by AFB, in rats. The AFB, epoxide covalently binds to DNA at the N7-guanine site, as well as to proteins, for example, via lysine e-amino groups (Figure 20.2). Such chemical reactions I~=~T CMA121565 S28 Carlo H. Tamburro and John L. Wong FIGURE 20.2 AFB, epoxide covalently bonding to DNA, which might lead to hepatocellular transformation. may lead to transformation of hepatocytes whose clones may expand during the regenerative phase of CAH, CAH behaves as a "viral partial hepatectomy," liberating endogenous proliferative factors. In addition to environmental monitoring of food contaminants by AFB, using TLC and HPLC, noninvasive biologic screening of populations to de termine the "internal dose" of AFB, in HOC etiology have been carried out. Immunoassays of the major AFB,-serum albumin adduct, aflatoxin-lysine, have been applied to human populations (Sabbioni et al,, 1990). Quantifi cation of this adduct in human serum is achieved by combined immunoaffinity chromatography and HPLC with fluorescence detection. For this method, serum is digested by pronase and the adducts are purified by monoclonal antibody (MAb). The MAb was obtained from a hybridoma of mouse SP-2 myeloma cells with spleen cells of mice immunized with a synthetic antigen of AFB, epoxide covalently bound to bovine gamma globulin (Sabbioni et al., 1990). One MAb isolated (2B11) was found to be a high IgM antibody with an affinity constant for AFB, and derivatives of about 1 X lO' liter/mol. A significant correlation coefficient of 0.82 was obtained between the afla toxin-lysine adduct levels and AFB, consumption for an epidemiologic study in China. The human data revealed an average aflatoxin-lysine adduct level of 0.38 ng adduct/p.g AFB, from the diet, or a daily albumin adduct burden of 2.9% of the AFB, daily intake. The MAb 2B11 also showed significantly cross-reactivity for the major aflatoxin-DNA adducts, the N7-guanosyl, and the corresponding imidazole ring opened derivative, suggesting that these adducts share a common anti genic determinant. The antibody was applied by Groopman et al. (1985) to quantify AFB,-N7-G in urine. The MAb first was bound covalently to Sepharose 4B, which made a reusable preparative column for isolating aflatoxin derivatives from human urine. As a measure of MAb sensitivity, a com petitive radioimmunoassay (RIA) showed a 50% inhibition value of approxi- CM* 121566 20 Biomarkers in the Study of Environmental Liver Disease S29 mately 300 fmol for AFB,. When this methodology was applied to human urine samples, the aflatoxin metabolites detected were AFBi-N?-G and the hydroxylated aflatoxins M, and Pi in individuals exposed to AFB! through dietary contamination at levels of 10-250 ppb. Although antibody technology facilitates isolation and detection of uri nary metabolites of aflatoxins, some studies of aflatoxin exposure may be subject to criticisms. In a cross-sectional ecological survey in China of possible risk factors for primary liver cancer (PLC; Campbell et al., 1990), multiple regression analyses for various combinations of risk factors were attempted that showed that aflatoxin exposure consistently remained un associated with PLC mortality. In contrast, HBsAg and plasma cholesterol were associated. This unique comprehensive survey included 48 county sites, approximately 600-fold aflatoxin exposure range, a 39-fold range of PLC mortality rates, a 28-fold range of FIBsAg carrier prevalence, and estimation of other life-style features. The aflatoxin exposure was determined from 4-hr urine samples, which were analyzed by isolating oxidative aflatoxin metabo lites such as AFM, (excluding nucleic acid adducts) on an antiaflatoxin MAb affinity column and quantifying them by a competitive JFI-base RIA. This analysis procedure, however, was faulted (Wild and Montesano, 1991) for not being representative of aflatoxin intake; the aflatoxin-albumin adduct was suggested as the proper biomarker for determining recent past exposure to AF. The counterargument (Campbell etal., 1990) is the strong correlation between the intake of AFBi and the urinary excretion of AFM!, as well as the correlation between serum aflatoxin-albumin adduct levels and urinary AFM,. Since the null effect of aflatoxin in this study contrasts sharply with other surveys, the new provocative conclusion makes it imperative to confirm that the aflatoxin exposure measured during the survey period can represent past intakes when PLC was forming. The larger question is how to relate aflatoxin exposure to oncogene ac tivation in the etiology of liver cancers. Evidence of such a relationship has appeared. McMahon et al. (1987) showed that AFB,-N7-G adducts were distributed nonrandomly in tumor-derived DNA of aflatoxin-induced HCC in rats. Such liver tumors also were found to contain activated c-Ki-ras on cogenes as identified in NIH3T3 mouse transformants. A single G-C to A T base mutation in codon 12 was found to activate the ras gene. In view of an accumulating body of evidence concerning single base mutations in codons 12,13, or 61 that arise in cellular ras genes after administration of chemical carcinogens, this AF activation of a ras gene may not serve the purpose of an exposure biomarker for aflatoxins. However, a combination of positive im munoassay of AFB,-N7-G in a dose-response manner, with the presence of multiple c-Ki-rus oncogene alleles, will make a compelling case for carcino genesis induced by aflatoxins. Further, studies have elucidated a significant mutation in the pS3 gene during the development of liver tumors. The pS3 nuclear phosphoprotein appears to function as a cell cycle regulatory mole- CMA 121567 S30 Carlo H. Tamburro and John L. Wong cule, controlling cell proliferation. The wild-type pS3 gene is a tumor sup pressor gene and has been mapped to chromosome 17p, a region often re duced to homozygosity in common cancers. It is the most frequently altered gene in human cancers (Jones et al., 1991). In analysis for mutations of p53 in HCC, in patients from China (Hsu et al., 1991) and from Africa (Bressac et al., 1991), 11 of 13 mutations have resulted in an arginine to serine sub stitution in codon 249 (AGG) of pS3. Additionally, 12 of 13 point mutations found in these patients were G -- T transversions. Aflatoxin-FF-G is the most likely cause of mutation. The specific mutant pS3 acts as a dominant oncogene and may interact further with a hepatitis B protein to provide a growth advantage in hepatomas. Other types of mutations, including frameshift and deletion, also may enhance clonal expansions. It appears that pS3 mutations in colon cancer, leukemias, and sarcomas are not induced by carcinogen-DNA adducts (Jones et at., 1991); therefore, patterns of base changes in p53 induced by aflatoxins may be considered footprints of their activities on DNA. Vinyl Chloride (Angiosarcoma): Synthetic Environmental Toxin The original association of vinyl chloride (VC) with angiosarcoma of the liver (ASL) in humans was made at a Louisville plastics and synthetic rubber plant in 1973. Since that initial discovery, the University of Louisville and B. F. Goodrich Company have been involved in a 17-year cooperative pro spective medical surveillance study involving 600-1200 active and 150-200 retired employees of the Louisville plant. The biologic data include annual historical, physical, radiologic, physiologic, pathologic, and biochemical data obtained on each employee. The environmental data include rankordered exposure estimates to 22 toxic chemicals and yearly individual job and area monitoring for specific vinyl monomers. The prospective human study of VC-associated ASL illustrates the fol lowing points. First, the initial discovery of ASL, a very rare liver tumor, was not linked specifically to VC. Polyvinyl chloride (PVC) and acrylonitrile (AN), as well as other chemicals, were also initially suspect. Medical exami nations did not identify the causal agent(s) and, in only a few cases, the ex istence of liver disease. Basic biochemical screening tests identified one or more abnormalities in 30-35% of the work force. Federally required specific liver tests found abnormalities in 10-20% of the work force. Definitive in vestigation confirmed only 10% of the work force as having persistent or significant liver dysfunction; 0.4% (four) had pre- or malignant disease. Individual rank-ordered retrospective or prospective work histories for 22 major work-related chemicals (Table 20.5) were used to identify which of these chemicals' cumulative exposure ranked months (CERMs) correlated with liver disease and angiosarcoma (Figure 20.3A,B). Only four chemicals were associated with ASL cases: VC, hexane, dimethyl maleate (DMM), and CMA 121568 20 Biomarkers in the Study of Environmental Liver Disease S31 TABLE ZO.S Selected Chemicals for Exposure Indices Chemical code Chemical name 01 Acrylic acid 02 Acrylamides--acrylamide, methyl, n-octyl 03 Acrylonitrile 04 Acetylene 05 Acrylates--ethyl, methyl, methyl-meth, 2-ethyl hexyl, N-butyl OS Bisphenol A 07 Butadiene 08 Caprylyl chloride 09 Chlorinated solvents--carbon tetrachloride, chloroform, trichloroethylene 10 Chloroethyl vinyl ether 11 Diethyl maleate 12 Mercuric chloride 13 Methanol 14 Phenol 15 Toluene 16 Vinyl chloride 17 Vinylidene chloride 18 Vinyl acetate 19 PVC dust 20 Catalysts 21 Styrene 22 Hexane catalysts. All other plastics-related chemicals and all the synthetic rubber chemicals showed no relationship. The catalyst group was used for VC prod ucts only and hexane was the major solvent for the VC catalyst, therefore both were always present when VC was used. DMM was a specific catalyst for a specialized PVC product and was used only periodically. This chemical is used by toxicologists to deplete GSH in animals in order to potentiate the toxicologic effect of the agent under study. Among the A$L cases, individuals with DMM exposure have shorter latency periods (Tamburro etai, 1984). VC also causes characteristic histologic liver injury (Tamburro, 1984). These histologic characteristics correlate very well with total (CERMs) rela tive VC exposure job rank, as shown in Figure 20.4. Study of biochemical and metabolic liver markers in detecting chemical injury, using CERMs and liver histology for specific lesions, revealed that ICG clearance provided the best combination of sensitivity and specificity (Figure 20.5A); GGT provided the highest sensitivity but also had the lowest specificity (highest false posi tivity; Figure 20.5B); and AP had the highest specificity (Figure 20.5C). Fi nally, individual job and area monitoring of VC and AN were shown to be CMA 121569 S32 Carlo H. Tamburro and John L. Wong Index liver angiosarcoma cases FIGURE 20.3 Relative vinyl chloride (A) and acrylonitrile (B) exposure rankings of index cases of hepatic angiosarcoma relative to their controls (individuals who worked the same years and number of years as index cases). .Angiosarcomas; CZ=1, matched controls. CMA 121570 HISTOLOGY FIGURE 20.4 Correlation of hepatic injury and chemical exposure illustrated by the signifi cantly larger percentage of markers whose liver histology showed evidence of chemical liver injury (CLI) that had vinyl chloride (VC) exposure ranking of 4 or greater. LD, Liver disease, nonchemical; NH, normal histology. FIGURE 20.5 (A) Sensitivity, (B) specificity, and (C) sum (sensitivity and specificity) for he patic biochemical biomarkers in chemical (} and nonchemical (O) liver injury. CMA 121571 534 Carlo H. Tamburro and John L. Wong 100 A 100 B 10 I I * 1 fc 0.11----*1 ii* 1 2 34 5 6 Job rank order 0.1 I--------1--------1--------1I-------- 1--------L. 1 2 34 5 6 Job rank order FIGURE 20,6 Correlation of job-specific acrylonitrile environmental exposure with job ex posure ranking, at maximum level (A) and at mean/average level (B). highly correlated to CERMs, to verify the relative exposure estimates of the CERMs, and to provide a ppm value for the CERMs (Figure 20.6). This study identified and verified the cellular toxicity and carcinogenicity level of VC and established its biologic threshold level for humans. These data now can be used to estimate human risk to past and future exposure accurately (Tamburro, 1984). Finally, similar analysis of the other chemicals, via the relational database system, provided strong evidence that no associ ation or relationship existed between VC exposure and other malignancies in the cohort. A 1982 report summarized the initial multidisciplinary re search developments on techniques and methods for the detection and pre vention of carcinogenesis in this cohort of industrial workers (Tamburro et al., 1982). As illustrated in the ASL case, because of the multiple metabolic and synthetic roles of the liver, no single marker, biologic or analytical, is suffi cient for molecular epidemiologic purposes. The essential requirements for effective epidemiologic study in the occupational surveillance of hepatic in juries are listed in Table 20.6. Guidelines for detection of hepatotoxicity due to chemical exposure were outlined by Davidson etal. (1979). A comprehensive review of the VC epidemiologic studies (Doll, 1988) confirms that the association between VC and cancer is confined to ASL. However, concern remains regarding individual human variation and prior VC animal exposure studies showing other cancers (Maltoni et al., 1981). CMA 121572 20 Biomarkers in the Study of Environmental Liver Disease S3S TABLE 20.6 Essential Elements for Prospective Surveillance of Occupational Environments___________________ 1. Medical history 2. Physical examination 3. Basic biochemical screening tests 4. Specific biochemical markers of liver (target organ) 5. Medical protocol for evaluation of positive finding 6. Defined investigation (radiologic, physiologic, pathologic) for hepatic evaluation 7. Biologic storage bank (blood, tissue) 8. Individual work history with rank-ordered cumulative exposure to key chemicals in work environment 9. Individual job and area analytical monitoring of key chemicals (agents) in the work environment 10. Computerized relational database for storage of all data Therefore, a dosimetry method based on molecular markers of VC metabo lites is needed to evaluate the current regulatory exposure limit of 1 ppm. Potential biomarkers for the metabolites shown in the scheme in Figure 20.7 are under current development (Tamburro etal., 1982; IARC, 1986; Joseph etal, 1990). Among the DNA adducts detected after exposure of experimental ani mals to VC, the major product, N7-(2-oxoethyl)guanine (OEG), is derived from guanine N7 alkylation by chloroethylene oxide (CEO). The detection limit was 10 pmol OEG/p.mol unmodified guanine (Fedtke etal., 1990). Rat tissue DNA was depurinated using mild acid hydrolysis. The hydrolysates CHj=CH-CI vc tytPW (0) CHj-CH-CI * Cl--CH,-- 0 CEO 0 CAA K H; a riboM FIGURE 20.7 Potential biomarkers for vinyl chloride metabolites. CMA 121573 TAMBURRO,CARLO M.D. 302 852 8927 P.82 526 Carlo H. Timhurro and John I, 'Woo; were analyzed by HPLC on * reversed-phase jcrong cation exchange column with a fluorescence detector (excitation at 225 nm with a 340-nm emission cut off filter). This method appears to give more reproducible results than reduction of the N'-oxoethyl group with tririared sodium borohydride or derivatizarion of the oxo group with O-methylhydroxylamine for GC-MS analysis. Other DNA adducts are the cyclic erheno derivatives whose formation occurs in the order of F.G > EC > EA (Eior ethenoj G, C, and A are DNA bases). The concentration of EG was determined in mild acid DNA hydroly sates by MS. The EG chromatography fraction was electfopbore-labeied using pentafluorobenzyl bromide. The dipencafluorobenzyl derivative was quantified relative to an internal standard '`Q-EG using GC-MS with nega tive ion chemical ionization by measuring the m/z ion ratio of 354/358. The limit of detection was 60 fmot EG/pmol guanine (Fcdtkc at <4., 1990), an improvement over the previous HPLC method with fluorescence detection. Thus, the ratio between EG and OEG was found to be approximately 1:100 in all tissues of rats immediately after VC exposure (600 ppm by inhalation, 4 hr per day, 5 days per week). This ratio in the liver increased to 1:141 week after exposure. It was calculated that the half-life of OEG was 62 hr, but that of EG was greater than 30 days, showing a greater persistence of the ethenoguanine adduct. Whether any of these two major VC adducts is required for cell transformations is not known. Analysis of the two minor cyclic nucleosides EC and EA can be achieved by MP-post)abeling or by radioimmunoassay. The former procedure reported by Watson and Crane (1989) was preceded by separating the adducts at 3'-monophosphates by ion-pair reversed-phase HPLC. Then the molecules were postlabcled using [y-^PJATP, and the mixture was treated with a nucle ase for 3' dephosphorylarion. The etheno[5'-MP}monophosphates, collected from HPLC, were quantified by liquid scintillation counring, yielding detec tion limits of 3-6 fmol FA and EC/p.g DNA. Monoclonal antibodies that specifically recognize ethenoadenosine or ethcnocytidine at approximately 200 fmol have been developed as an alternative detection method (Young and Santella, 1988). A radioimmunoassay for their presence in exposed rat tissues was reported (Cirousse) tt at, 1990). The concentrations measured were 0.49 pmol EC/pmol deoxycytidine and 0.13 pmol EA/ptnol deoxyadenosine in the liver DNA of rats exposed to 500 ppm VC (7 hr per day for 14 days). These values were an order of magnitude lower than the EG value from a similar exposure experiment described earlier. Analysis of these four VC-DNA adducts from animal tissues can be adapted to human dosimetry. The major DNA adduct formed in livers of rats acutely exposed to VC was OEG, whereas the etheno derivatives (but nor OEG) were found in liver DNA of rats chronically exposed to VC. In addi tion to its role as a biomarker of exposure to VC-type chemicals, each adduct should be considered for its role in genotoxicity. The predominant adduct CMA 121574 20 Biomarkers in the Study of Environmental Liver Disease S3 7 OEG, derived from the putative metabolite CEO, has a short half-life and lacks miscoding properties (IARC, 1986). It probably contributes only indi rectly to the mutagenic effects of VC via depurination and mispairing oppo site the apurinic sites. In contrast, the three minor etheno adducts have been reported to be efficient in causing mispairing during DNA replication (Jacob sen etal,, 1989; Singer et al,, 1987), although conflicting data point to low miscoding efficiency of EA and EC (Bartsch and Singer, 1985). AH three cy clic adducts can be attributed to the other putative metabolite chloroacetaldehyde (CAA), thereby suggesting CAA to be responsible for VC genotoxicity. However, bacterial mutagenesis assays showed CEO to be much more potent than CAA (Perrard, 1985). Also, under comparable conditions when CEO was found to produce skin tumors in mice, CAA produced no increase in benign or malignant tumors (Zajdela et al., 1980). Thus, the detoxifica tion of CEO and CAA must be considered in assessing individual risk to VC exposure. Strengths and Limitations Enzymes Microsomal P450 The induction enzymes, such as cytochrome P450, can be measured di rectly from liver samples obtained by needle or surgical biopsies (McPherson et al., 1982). Such measurements have limitations based on differences in regional distribution of P450 and other enzymes, and on the different forms of the groups of enzymes, the levels of which may be reduced or induced by the xenobiotics themselves. In addition, these methods are limited by over lap (Table 20.2), variable xenobiotic induction, and background induction caused by high natural diet exposure, air pollutants, and life-style factors (Watkins, 1990). Knowledge of the "usual" background level (steady state) of induction is required also to identify any changes attributable to the sus pect xenobiotic(s). Indirect measurement by ABT is the alternative to using tissue for these enzyme determinations. Aminopyrene Breath Test ABT itself has several limitations. It cannot distinguish among the vari ous levels of liver disease (Hepner and Vesell, 1975). The overlap between individuals with adaptive or mild liver dysfunction makes the test less useful for those in most need of such evaluation, that is, individuals with subclinical disease. ABT has been found to be more reliable in predicting short-term changes, clinical improvement, and the histologic severity of chemical liver disease (e.g., alcohol related) than the more conventional liver tests. The pre dictive value of ABT for steatonecrosis, pericentral fibrosis, and cirrhosis (in- ' ' CMA 121575 S38 Carlo H. Tamburro and John L. Wong active) is less than the standard predicted value of a BLT. At the moment, there is no evidence that one breath test has anything to offer over another. Such "surrogate" methods (ABT) with high sensitivity are desirable. How ever, without concurrent high specificity and high disease occurrence, such surrogate markers can be potentially more psychologically or socioeconom ically harmful because of high false-positive and false-negative rates. Glutathione S-Transferase The clinical usefulness of GST as a potential biomarker is uncertain. For example, although glutathione S-transferase is involved in catalyzing the de toxification of diol-epoxide carcinogenic metabolites of polycyclic aromatic hydrocarbons (Vander Jagt et al, 1985), these transferases have distinct but overlapping substrate specificities bordering on the complexities of the cyto chrome P450s. Metabolic and Physiologic Tests ICG and other clearance tests mainly reflect hepatocellular injury or physio logic dysfunction. They provide only indirect evidence of xenobiotic injury. They are effective markers when the agent(s) and its exposure level are known and when other toxic associations can be excluded by epidemiologic or statistical analysis. Their major strength is their selectiveness for the liver. Proteins The major limitation of tests of antigen or antibody induced by xenobiotics is their sensitivity and specificity for the chemical agent. Exposure to the chemical agents acting as antigens may not be followed by antibody induc tion due to inadequate antigen production or structure derangement caused by its hepatic metabolism. Hepatic biomarkers of this type can be enhanced by PCR amplification for better detection. Adducts Monoclonal antibodies for specific chemical agents or their metabolites pro vide the most promising biomarker methods. A major limitation, at present, is that many adduct markers are not hepatically selective. Adducts with GSH, albumin, and hemoglobin reflect highly sensitive methods of identify ing hepatic exposure over various periods of time. Monoclonal antibodies to various hepatic metabolites allow identification of different routes of me tabolism (albumin and hemoglobin adducts) and the effectiveness of detoxi fication (GSH adducts). DNA adducts are best used to identify high-risk ef fects of exposure and provide a potential method of assessing the reparative capability of individual DNA. This methodology can be applied indirectly (circulating tissue: white blood cells) and directly (hepatic tissue: via biopsy). CMA 121576 r 20 Biomarkers in the Study of Environmental Liver Disease 539 The data suggesting strong xenobiotic associations or even causations in "group" data are insufficient for use on an individual basis. Such adduct markers still require confirmation of their specificity and sensitivity in indi viduals whose exposure and hepatic disease has been well characterized. The adduct surrogate (e.g., hemoglobin adduct for a hepatotoxic xenobiotic also must be shown to reflect the target organ (the liver) under surveillance cor rectly (i.e., selectivity). Genetic Markers Polymorphism Use of gene mutation in the clinical setting of hepatic disease may be limited because (1) gene mutation may be present only in the end stages of the carcinogenic processes, (2) gene mutation may require multiple "hits" before becoming established, and (3) gene mutation may be seen only in liver tissue and not in the more accessible body tissues. Oncogene Markers In human cancer, the ras oncogene was found in 90% of adenocarci noma in the pancreas, 50% in the colon, 30% in the lung, 50% in the thyroid, and in 30% of myeloid leukemia (Bos, 1989), The NIH3T3 trans fection-transformation assay may not be sensitive enough to select ras acti vation in all the liver tumor DNA. Until a more sensitive assay is used, inter pretation of the detection percentages of activated ras gene as a biomarker cannot be made with confidence. However, one should note the potential of the ras oncogene as a specific disease marker for the causative agent. Increasing evidence suggests that mutational spectra are highly correlated with each chemical carcinogen and reflect the predicted base substitu tion, that is, G-C --> T-A transversion for benzo[a]pyrene, which forms predominantly the AP-BPDE-deoxyguanosine adduct, and A:T T:A resulting from N6-deoxyadenosine bonded to the diol-epoxide of 7,12dimethylbenzanthracene (Singer and Grunberger, 1983). It is plausible that molecular analysis of mutationally activated ras genes (a feat readily achiev able with PCR) will reflect promutagenic DNA adduct formation and the mutagenic activities elicited by specific environmental carcinogens. With more complete molecular information, such structure-function correla tions between ras DNA adducts and ras activities may be made, even in the presence of confounding factors such as spontaneous mutations producing G-C -- T-A transversions. The p53 tumor suppressor gene appears to be more suited as a hepatic biomarker. The p53 gene is mutated in diverse types of human cancers (Hollstein et al., 1991); germ line mutations in p53 predispose to cancers of the breast, soft tissues, and brain. Mutant p53 has been found in hepatocellular carcinoma in connection with aflatoxin and hepatitis B. However, it is not I --wr ~ CMM21577 S40 Carlo H. Tamburro and John L, Wong certain which of these two agents has caused the pS3 mutations in the China and southern Africa studies. Analysis of liver tumors from regions in which either aflatoxin or the hepatitis B virus is the predominant agent will be in structive. At present, mutant pS3 is associated with driving selective clonal growth, a critical step in the neoplastic process. Its detection in liver and other tissues means a risky prognosis. Clinical Field Application ofBiomarkers Further application of these methods in the early detection of hepatic injury in multiple exposure environments, such as the workplace, is needed. Their application for determining hepatic cancer risk, however, will have strong socioeconomic and ethical impacts. Their field application is vital in showing their ability to: (1) identify high-risk individuals, (2) identify an individual's specific hepatic metabolism for xenobiotics (e.g., degree of oxidation/detoxi fication, DNA adduct/repair), (3) confirm the safety of work environments containing potential carcinogens (e.g., acrylonitrile, TCDD, PCB), (4) show levels of individual exposure not associated with hepatic functional or struc tural changes beyond the adaptive response (Level 1), and (5) differentiate the cause of hepatic injury in multiple agent involvement (e.g., acrylonitrile and vinyl chloride). Due to the multiple and complex functions of the liver, molecular epi demiologic investigations require joint disciplinary research between the basic molecular biologist and the clinical hepatologist. By nature, human investigation must be conducted in well-characterized environments, in a prospective surveillance-type system containing and applying the essential elements set forth in Table 20.6. This control is especially relevant to hepatotoxin exposure and hepatogenetic markers. More than any other hepatic biomarkers, the genetic markers raise serious ethical and social questions. In industrially developed countries, the incidence of hepatic cancer is low, whereas in underdeveloped countries it is high. Having DNA damage or a cancer-susceptibility gene does not necessarily lead to cancer, although it may identify an individual as high risk. Determining whether such individual information outweighs the benefits requires continuous reassessment. In the low HCC-incidence populations, high risk identification is associated with increased anxiety, discrimination, depression, decreased job security, or uninsurability. In high HCC-incidence populations, such information often im pairs personal economic growth or opportunity. Research Needs In the hepatic organ system, no single molecular marker will provide ade quate information about exposure, effect, or susceptibility (risk) nor will it CMA 121578 20 Biomarkers in the Study of Environmental Liver Disease S41 answer all clinically relevant needs. Research in hepatic molecular markers is needed in four distinct but interdependent areas. The first need deals with exposure identification of causative agent(s) when liver disease is found in a new occupational or environmental setting. Because of the variety of chemicals customarily present in such situations, it is often neither practical nor feasible to use specific markers such as MAbs to assess exposures. Under these circumstances, screening approaches are needed to identify the environmental chemicals or their hepatic metabolites. Mature analytical techniques for analysis of body fluids, such as GC-MS, are well developed for stable and volatile compounds such as dioxin (TCDD) and polychlorinated biphenyls (PCB). However, gaseous or gas-like com pounds such as formaldehyde, methyl chloride, vinyl chloride, acrylonitrile, or butadiene escape easily from the aqueous samples to be so determined. Better techniques are needed for detection of their hepatic metabolites or conjugates in body fluids or tissue. In addition, more technological develop ment is needed for nonvolatile agents or by-products. Intensive research is on-going to develop mass spectrometry for trace analysis of highly polar and nonvolatile compounds of complex mixtures; none, however, has been di rected to liver-specific assays. Under development are derivatizations of adducts or adduct hydrolysates to increase volatility followed by GC-MS; liquid chromatography-mass spectrometry (LC-MS); tandem mass spec trometry (TMS or MS-MS) with desorption ionization (DI) via fast atom bombardment (FAB) or laser microprobe (LAM); LC-MS-MS; and so forth. Particularly promising for liver tissue analysis is the TMS technique. Here, the key innovation is the DI technique to produce ions from nonvolatile sur faces for mass analysis. Ions are formed from sputtered molecules after irra diating samples with a high-energy particle beam (FAB) or a focused laser beam (LAM). TMS is a nonchromatographic method for direct-mixture analysis that can yield molecular weight and structural information. A TMS experiment is performed as the name implies: two mass spectrometers (MS-1 and MS-2) are connected together so that MS-1 separates a particular ion Ma+ (molecular weight information), formed by direct ionization of the sample, and the fragment ions formed by dissociation of Ma+ are massanalyzed by MS-2 (structure identification). Thus, TMS performs both the separation and the analysis step with sensitivity of detection reaching to sub program levels. The technique has been applied to biomolecules such as vi tamin B12, chlorophyll, and bradykinin (Burlingame et al., 1984). Burlin game and co-workers reported assignment of specific residues in human hemoglobin modified by styrene oxide using TMS (Kaur etal, 1989). In the second area, for the more defined exposures under which routine screening of biological samples is required, MAbs to metabolites of com modity or known high-risk chemicals such as vinyl chloride, acrylonitrile, and the environmental toxin aflatoxin are very much wanted. However, the problem of false positives must be addressed more keenly, since cross CMA 121579 542 Carlo H. Tamburro and John L. Wong reactivity of MAbs that is not identified during laboratory development may become significant in its field applications. To date, liver histology (biopsy) remains the standard of environmental liver injury. However, specific hepatic tissue biomarker assays are needed for hepatocytes, biliary ductal cells, reticuloendothelial cells, and macrophages to help identify organ tissue target sites. Along this line, liver specific tissue antigen biomarkers could help identify specific liver function impairment (ef fect) without the need for liver tissue. Further, hepatic biomarkers with high specificity are more important socioeconomically than those with high sen sitivity. These markers are needed to avoid the high false-positive findings that often outweigh the true positive benefits. Especially needed for chronic exposure are quantitative markers of he patic reserve, both anatomical and functional. Special attention needs to be directed to methods that can quantitate hepatic collagen content, decreased protein substrate, or synthesis. Quantitative biomarkers of hepatic collagen (effect of injury) or substrate content or synthesis (effect of function impair ment) would have very high clinical diagnostic value. Susceptibility research, the third area of hepatic biomarkers of risk, should be directed at two major categories. Methods that will identify puta tive metabolic pathways of exposure in individuals (individual susceptibil ity), for example, MAbs and MS, are needed. Another area of susceptibility involves tests to detect gene injury. In liver tissue, susceptibility of groups or populations will depend greatly on the history of exposure (dose, duration, and recentness) and the evidence of hepatic injury. Tests directed at gene in jury or activation, for example, analysis of pJ3 and the ras gene using PCR and RLFP techniques, offer promise for hepatic biomarker development. This development should start with verification of the effect of each gene presently implicated in hepatic tumors (e.g., hepatocellular carcinoma and HBV-aflatoxin or angiosarcoma and vinyl chloride). Retrospective and pro spective applications should be in a defined population, for example, firstgeneration Asian immigrants to the United States and vinyl monomer workers. Both are valid populations in which to apply the outcomes of sus ceptibility research to demonstrate its benefits or limitations. Finally, the fourth area, the most time consuming and the most vital, is research in the further development and maintenance of well-characterized prospective surveillance programs of at-risk populations. These studies are absolutely essential to validate the application and interpretation of these evolving techniques. Long-term cooperative nonadversarial efforts between industry, its workforces, and the scientific community already have shown that such programs can be financially, ethically, and scientifically feasible. These programs need to be enhanced by incorporating the standardization of hepatic nomenclature and criteria for hepatic function, injury, and disease (Leevy et al,, 1976) into environmental surveillance. Issues of background marker frequency, determination of relative risk assessment, and disease fre- T IW1" CMA 121580 20 Biomarkers in the Study of Environmental Liver Disease 543 quency (M, P, and R, respectively) in the workplace and in noncommercial exposed environments cannot be addressed properly without such popula tion or group studies. References Baker, A. L., Kotake, A. N., and Schoeller, D. A. (1983). Clinical utility of breath tests for the assessment of hepatic function. Sent. Liver Dis. 3,318-329. Bartsch, H., and Singer, B. (1985). International meeting on the role of cyclic nucleic acid ad ducts in carcinogenesis and mutagenesis. Cancer Ret. 45,5205-5209. Bos.J. L. (1989). Ras oncogenes in human cancer: A review. Cancer Res. 49,4682-4689. Bressac, B., Kew, M., Wands, J., and Ozturk, M. (1991). Selective G to T mutations of p53 gene in hepatocellular carcinoma from southern Africa. Nature (London) 350,429-431. Burke, M. D., Murray, G. I., and Lees, G. M. (1983). Fluorescence-microscopic measurement of intracellular cytochrome P-450 enzyme activity (ethoxyresorufin O-deethylation) in un fixed liver sections. Biochem, ]. 212,15-24. Burlingame, A. L., Whitney, J. O., and Russell, D. H. (1984). Mass spectrometry: Fundamental reviews. Anal. Chent, 56,417R, Campbell, T. C., Chen, J., Liu, C., Li, J., and Parpia, B. (1990). Nonassociation of aflatoxin with primary liver cancer in a cross-sectional ecological survey in the People's Republic of China. China Res, 50,6882-6893. Ciroussel, F., Barbin, A., Eberle, G., and Bartsch, H. (1990). Investigations on the relationship between DNA ethenobase adduct levels in several organs of vinyl chloride-exposed rats and cancer susceptibility. Biochem. Pharmacol 39,1109-1113. Colli, A., Buccino, G., Cocciolo, M., Parravicini, R,, and Scaltrini, G. (1988). Disposition of a1 flow-limited drug (lidocaine) and a metabolic capacity-limited drug (theophylline) in liver cirrhosis. Clin. Pharmacol. Tber. 44,642-649. Davidson, C. S., Leevy, C. M., and Chamberlayne, E. C. (1979). "Guidelines for Detection of Hepatotoxicity Due to Drugs and Chemicals." DHEW (NIH) Publ. No. 79-313. U.S. Gov ernment Printing Office, Washington, D.C. Doll, R. (1988). Effects of exposure to vinyl chloride. Scand. ]. Work Environ. Health 14, 61-78. Fedtke, N., Boucheron, J. A., Walker, V. E., and Swenberg, J. A. (1990). Vinyl chloride-induced DNA adducts. II. Formation and persistence of y-(2'-oxoethyl)guanine and N-2,3-erhenoguanine in rat tissue DNA. Carcinogenesis 11,1287-1292. Glatt, H., Friedberg, T., Grover, P. L, Sims, P., and Oesch, F. (1983). Inactivation of a diolepoxide and a K-region epoxide with high efficiency by glutathione transferase X. Cancer Res. 43,5713-5717. Groopman, J, D., Donahue, P, R., Zhu, J., Chen, ]., and Wogan, G. N. (1985), Aflatoxin me tabolism in humans: Detection of metabolites and nucleic acid adducts in urine by affinity chromatography. Proc. Natl. Acad. Sci. USA. 82,6492-6496. Harris, C- C. (1990). Review on hepatocellular carcinoma: Recent advances and speculations. Cancer Cells 2,146-148. Harris, C. C. (1991). Chemical and physical carcinogenesis: Advances and perspectives for the 1990s. Cancer Res. Suppl. 51,5023s-5044s. Hayasaka, A., Maddrey, W. C., and Hahne, E. G. (1988). Serum concentrations of amino ter minal propeptide of collagen. III. (PIIPJ Study state levels of messenger RNA of alpha-1 chain of procollagen III in carbon tetrachloride-induced fibrotic rat livers. Hepatology 8, 1304-1309. Hepner, G. W., and Vesell, E. S. (1975). Quantitative assessment of hepatic function by breath analysis after oral administration of C-aminopyrine. Ann, Int. Med. 83,632-638. CMA 121581 544 Carlo H. Tamburro and John L. Wong Holistein, M., Sidransky, D., Vogelstein, B,, and Harris, C. C. (1991). p53 mutations in human cancers. Science 253,49-53. Hsu, I. C., Metcalf, R. A., Sun, T., Welsh, J. A., Wang, N. J., and Harris, C. C. (1991). Muta tional hotspots in the p53 gene in human hepatocellular carcinomas. Nature (London) 350,427-428. International Agency for Research on Cancer (1986). "The Role of Cyclic Nucleic Acid Adducts in Carcinogenesis and Mutagensis," (B. Singer, and H- Bartsch, eds.). IARC, Lyon, France. Jacobsen, J. S., Perkins, C. P., Callahan, J. T., Sembimupti, K., and Humayun, M. Z. (1989). Mechanism of mutagenesis by chloroacetaldehyde. Genetics 121,213-222. Jones, P. A., Burckley, J. D., Henderson, B. E., Ross, R. K., and Pike, M. C. (1991), From gene to carcinogen: A rapidly evolving field in molecular epidemiology. Cancer Res. 51, 3617-3620. Joseph, J. T., Elmore, J. D., and Wong, J. L. (1990). Comparative sulfhydryl reaction pathways of chlorooxirane and chloroacetaldehyde. J. Org. Chem. 55,471-474. Kaur, S., Hollander, D., Haas, R., and Burlingame, A. L. (1989). Characterization of structural xenobiotic modifications in proteins by high sensitivity tandem mass spectrometry. /. Biol, Chem.264,16981-16984. Leevy, C. M., Popper, H., and Sherlock, S. (1976). "Disease of the Liver and Biliary Tract: Stan dardization of Nomenclature, Diagnostic Criteria, and Diagnostic Methodology." Fogarty International Center Proceedings No. 22, DHEW (NIH) Publ. No. 76-725. U.S. Govern ment Printing Office, Washington, D.C. Le Provost, E., Cresteil, T., Codumelli, S., and Leroux, J. P. (1983). Immunological and enzy matic comparison of hepatic cytochrome P-450 fractions from phenobarbital-, 3-methylcholanthrene-, beta-naphrhoflavone-, and 2,3,7,8-tetrachlorodibenzo-p-dioxin-treated rats. Biochem. Pharmacol. 32,1673-1682. Liss, G, M., Greenberg, R, A., and Tamburro, C. H. (1985). The use of serum bile acids in the identification of vinyl chloride hepatotoxicity. Am. /. Med. 78,68-76. McMahon, G., Davis, E., and Wogan, G. N. (1987). Characterization of c-Ki-ras oncogene alleles by direct sequencing of enzymatically amplified DNA from carcinogen-induced tu mors. Proc. Nat. Acad. Sci. U.S.A. 84,4974-4978. McMahon, G., Huber, L. J., Moore, M. J., Stegcman, J. J., and Wogan, G. N. (1990), Mutations in c-Ki-ras oncogenes in diseased livers of winter flounder from Boston Harbor. Proc. Nat. Acad. Sci. U.S.A. 87,841-845. McPherson, G. A. D., Benjamin, I. S., Boobis, A. R., Brodie, M. J., Hampden, C., and Blumgart, L. H. (1982). Antipyrine elimination as a dynamic test of hepatic functional integrity in obstructive jaundice. Gut 23,734-738. Maltoni, C., Lcfemine, C., Ciliberti, A., Cotti, G., and Carretti, D. (1981). Carcinogenicity bio assays of vinyl chloride monomer: A model of risk assessment on an experimental basis. Environ. Health Perspect. 41, 3-29. Mendenhall, C, L., Seeff, L., Diehl, A. M., Ghosn, S. J., French, S. W., Gartside, P. S., Rouster, S. D., Buskell-Bales, Z., Grossman, C. J., Roselle, G- A., Weesner, R. E., Garcia-Pont, P., Goldberg, S- J., Kieman, T. W., Tamburro, C. H., Zetterman, R., Chedid, A., Chen, T., Rabin, L., and the VA Cooperative Study Group (No. 119) (1991). Antibodies to hepatitis B virus and hepatitis C virus in alcoholic hepatitis and cirrhosis; Their prevalence and clini cal relevance. Hepatol. 14,581-589. Ozturk, M,, Bressaci, B., Puisieux, A., Kew, M., Volkmann, M., Bozcali, S., Bella Mura, J., de la Monte, S., Carlson, R., Blum, H., Wands, J., Takahashi, H., von Weizsacker, F., Galun, E., Kar, S., Carr, B., Schroder, C., Erken, E., Varinli, S., Rustgi, V., Prat, J., Toda, G., Koch, H., Lian, X., Tang, Z., Shouval, D,, Lee, H., Vyas, G., and Sarosi, I. (1991). p53 hotspot mutation in hepatocellular carcinoma is linked to aflatoxin exposure. Hepatol. 14, 94A. Perera, F,, and Weinstein, I. B. (1982). Molecular epidemiology and carcinogen-DNA adduct detection: New approaches to studies of human cancer causation. ]. Chron. Dis. 35, 581-600. CMA 121582 20 Biomarkers in rhe Study of Environmental Liver Disease S4S Perera, F., Santella, R., and Poirier, M. (1986). Biomonitoring of workers exposed to carcino gens: Immunoassays to benzo(a)pyrene-DNA adducts as a prototype. ]. Occup. Med. 28, 1117-1123. Perrard, M, H. (1985). Mutagenicity and toxicity of chloroethylene oxide and chioroacetaldehyde. Experientw 41,676-677. Sabbioni, G., Ambs, S., Wogan, G. N., and Groopman, J. D- (1990). The aflatoxin-lysine ad duct quantified by high performance liquid chromatography from human serum albumin samples. Carcinogenesis 11,2063-2066. Santella, R. M., Weston, A., Perera, F. P,, Trivers, G. T., Harris, C. C., Young, T. L., Nguyen, D., Lee, B. M., and Poirier, M. C. (1988). Interlaboratory comparison of antisera and immunoassays for benzo[]pyrene-dioI-epoxide-I-modified DNA. Carcinogenesis 9, 1265-1269. Savolainen, E. R., Brocks, D., Ala-Kokko, L., and Kivirikko, K. I. (1988). Serum concentrations of the N-terminal propeptide of type HI procollagen and two type IV collagen fragments and gene expression of the respective collagen types in liver in rats with dimethylnitrosamine-induced hepatic fibrosis. Biochem.J. 249,753-7J7. Schrager, T. F., Nerbeme, P. M., Pikul, A. H., and Groopman, J. D. (1990). Aflatoxin-DNA adduct formation in chronically dosed rats fed a choline-deficient diet. Carcinogenesis 11, 177-180. Singer, B., and Grunberger, D. (1983). "Molecular Biology of Mutagens and Carcinogens." Plenum Press, New York. Singer, B., Spengler, S. J., Chavez, F., and Kusmierek, J. T. (1987). The vinyl chloride-derived nucleoside, !N-3-ethenoguanosine is a highly efficient mutagen in transcription. Carcino genesis 8,745--747. ,Tamburro, C. H. (1984). Relationship of vinyl monomers and liver cancer; angiosarcoma and hepatocellular carcinoma. Sent. Liver Dis. 4 159-169, Tamburro, C. H. (1991). Laboratory evaluation of liver disease. In "Diseases of the Liver and Biliary Tract" (G. Gitnick, D. R. LaBrecque, F. G. Moody, eds.), pp. 158-162. Mosby, St. Louis, Missouri. Tamburro, C. H. (1992). Chronic liver injury in phenoxy herbicide exposed Vietnam veterans. Environ, Res. 59,175-188. Tamburro, C. H., and Liss, G. M. (1986). Tests of hepatotoxicity; Usefulness in screening work ers. J. Occup. Med. 28,1034-1044. Tamburro, C. H., Kupchella, C. E., Wong, J. L., Barrows, G. H., Du, J. T., Espinosa, E., Feldhoff, R. C., Fortwengler, H. P., Schrodt, G. R,, Sonnenfetd, G., Streips, U. N., Tseng, M, T., and Waddell, W. J. (1982). Report on research techniques and methods for the detection and prevention of carcinogenesis in the industrial worker. Rep. Chem. Manufact. Assoc. vc7.0, pp. 1-104. Tamburro, C. H., Makk, L., and Poppet, H. (1984). Early hepatic histologic alterations among chemical (vinyl monomer) workers. Hepatology 4,413-418. Vander Jagt, D. L., Hunsaker, L. A., and Royer, R. E. (1985). Glutathione 5-transferase in the human liver; 13 forms of GST have been identified. Isolation of the multiple glutathione 5transferases from human liver.}. Biol. Chem. 260,1163-1166. Wang, P. P., Beaune, P., Kaminsky, L. S., Dannan, G. A., Kadlubar, F. F., Larrey, D,, and Gengerich, F. P. (1983). Purification and characterization of six cytochrome P450 isoenzymes from human rnicrosomes. Biochem. 22,5375-5383. Watkins, P. B. (1990). Role of cytochromes P450 in drug metabolism and hepatotoxicity. Sem. Liver Dis. 10,235-350. Watson, W. P., and Crane, A. E. (1989). HPLC-MP-postlabelling analysis of l,N`-ethenodeoxyadenosine and 3,N4-ethenodeoxycytidine. Mutagenesis 4,75-77. Wild, C. P., and Montesano, R. (1991). Correspondence re: T. C. Campbell et at. and Reply. Cancer Res. 51,3825-3827. Winer, F. R., Eskreis, D. S., Compton, K. V., Orrego, H., and Zem, M. A. (1988). Haplotype CMA121583 S46 Carlo H. Tamburro and John L. Wong analysis of a type-I collagen gene and its association with alcoholic cirrhosis in man. Mol. Asp. Med. 10,159-168. Wong, J. L., Ma, F, F., and Zhang, Y. (1990). Antibodies to acrylonitrile-glutathione conjugate. Antibody Immunoconj. Radiopharmaceu. 3,194. Wu,T. G-, and Wong, J. L. (1991). Adsorptive stripping voltammetric speciation of nickel(II)histidinc in aqueous ammonia. Anal. Chim. Acta 246,301 -308. Young, T. L., and Santella, R. M. (1988). Development of techniques to monitor for exposure to vinyl chloride: Monoclonal antibodies to ethenoadenosine and ethenocytidine. Card* nogenesis 9,589-592. Zajdela, F., Croisy, A., Barbin, A., Malaveille, C,, Tomatis, L_, and Bartsch, H. (1980). Carci nogenicity of chloroethylcne oxide, an ultimate reactive metabolite of vinyl chloride and bis(chloromethyl)ether after subcutaneous administration and in initiation-promotion ex periments in mice. Cancer Res. 40,3J2-3J6. CMA 121584