Document yrnyZRwrRqMVJ21rQDL65kgKn

Int Arcft Occup Environ Health (1996) 68: 394-398 p; 11 v+bf' Springer-Verljg im)fl M.-J. Marion I. DeVivo S. Smith J.-C. Luo P. W. Brandt-Rauf The molecular epidemiology of occupational carcinogenesis in vinyl chloride exposed workers 9 <"0 ( Introduction The production of specific mutations in cellular onco genes and tumor suppressor genes is thought to be in volved in the development of certain cancers related to exposures to chemical carcinogens. The occurrence of certain point mutations in oncogenes and tumor suppres sor genes results in the expression of mutant forms of their encoded protein products, which is believed to con tribute to the process of cellular transformation and thus cancer. The detection of the expression of such mutant protein products in viva could therefore serve as potential biomarkers for the molecular epidemiologic study of chemical carcinogenesis in human populations exposed to carcinogens [1]. A model system for such study is pro vided by workers who have been occupationally exposed to high levels of vinyl chloride (VC) and are at risk for the development of a sentinel neoplasm, angiosarcoma of the liver (ASL) [2], VC is a known animal and human carcinogen which is rapidly absorbed following respiratory exposure and is primarily metabolized in the liver by the cytochrome P450 2E1 system to the electrophilic metabolites, chtoroethylene oxide (CEO) and chloroacetaldehyde (CAA). CEO and CAA react with DNA bases to form adducts that Work presented at the `23rd Congress on Occupational and Envi ronmental Health in the Chemical Industry" (Medichem 1995) "The Chemical Industry as a Global Citizen - Balancing Risks and Benefits". 19-22 September 1995. Massachusetts Institute of Technology. Cambridge. Massachusetts M.-J. Marion Unit for Research on Hepatitis and Human Retroviruses, INSERM. Lyon. France I. DeVivo S. Smith J.-C. Luo P. W. Brandt-Rauf (E3) Division of Environmental Health Sciences. Columbia University. School of Public Health. 60 Haven Avenue. B-l, New York. NY. 10032. USA FAX: 212-305-4012 are mutagenic in bacterial systems and mammalian cells, including: 7-(2'-oxoethyl) guanine; 1, N6-ethenoadenine: 3. NJ-ethenocytosine: and N:-3-ethenoguanine 13]. In an imal experiments, the oxoethyl adduct is the major liver DNA adduct formed, representing 98% of all adducts, but it is the least persistent with a half-life of about 62 h [4] On the other hand, the less common etheno adducts are highly persistent with half-lives of more than 30 days, suggesting that they are poorly recognized by the liver DNA repair system (5]. Thus, the production of such etheno adducts could account for the occurrence of spe cific point mutations (including G.C--A.T transitions and A. T--T.A transversions) identified in the cellular onco genes and tumor suppressor genes of VC-associated ASLs (Fig. 1). For example, the cellular ras oncogenes are frequently found to contain specific point mutations in many human malignancies [6]. In a study of ras oncogene mutations in tumors of VC-exposed workers. 15 of 18 (83%) ASLs ex amined to date have been found to contain a G-->A transi tion at the second base of codon 13 of the c-Ki-/us-2 gene [7,8]. This ras mutation at codon 13 (GGC-^GAC) 6 known to be capable of converting the gene into an ac tively tranforming oncogene in model systems, and. thus, could be causally related to the carcinogenic process (9), This ras mutation results in the substitution of an aspartic acid for the normal glycine at amino acid residue 13 in the encoded p21 protein product. This amino acid substitution is believed to produce a conformational change in p2l which may be responsible for altering its intrinsic GTPase activity, affecting signal transduction within the cell con tributing to cellular transformation [10]. This altered Asp 13 p21 protein can be distinguished from normal p21 ami other mutant p21s immunologieally with a mouse mono clonal antibody that is specific for this mutant protein HI). In addition, the p53 tumor suppressor gene is also fre quently found to contain point mutations in a significant proportion of human malignancies [12]. In a study ot P-'" gene mutations in tumors of VC-exposed workers. 2 ot 4 (50%) ASLs examined to date have been found to contain AS1000013963 vinyl chloride Fig. 1 Proposed carcinogenic mechanism ot sinyl chloride in the production of' angiosarcomas of the liver via ihe induction of mu tations in p21 and p53 an A--tranxversion at the first base of codon 249 or 255 of p53 L131. These mutations result in the substitution of a tryptophan for the normal arginine at amino acid residue 249 or of a phenylalanine for the normal leucine at amino acid residue 255 in the encoded p53 protein product; these amino acid substitutions occur in a domain of the p53 pro tein which has been highly conserved throughout evolu tion. suggesting that this domain is critical for determina tion of the structure and hence function of the protein [12]. Other amino acid substitutions in this conserved re gion of p53 have in fact been demonstrated to alter the protein's normal structure and function and to result in the stabilization of the mutant protein, significantly prolong ing its half-life in comparison to the wild-type protein and thus causing its accumulation in transformed cells [12|. The resultant increased amounts of mutant p53 can be de tected by a mouse monoclonal antibody that is specific for mutant conformations of the protein [14]. Thus, similar VC-induced amino acid substitutions would also be ex pected to alter p53's normal ability to act as a negative regulator of cell division by causing a conformational change in the protein that would result in its accumulation which could be detected immunologically in the same fashion. For cells in culture that contain a mutant ras gene and express mutant p21 or contain a mutant p53 gene and ex press mutant p53 protein, it is possible to use monoclonal antibodies to detect the mutant p21 or mutant p53. respec tively. in the extracellular supernatant [15,16]. In an anal ogous situation in vivo where individuals may contain cells that have mutations in ras or p53 due to exposure to chemical mutagens, one might expect to be able to detect the presence of mutant p21 or mutant p53 in their extra cellular fluids, including serum. We have therefore used immunological techniques, including immunoblotting for mutant p21 and an enzyme linked immunosorbent assay (ELISA) for mutant p53, to examine the expression of these proteins in the serum of individuals with VC expo sure. Methods From a cohort of more than 400 workers employed in VC-polymerization plants in France since 1950. a sample of 27 of the most heavily exposed individuals have been studied for their serum ex pression of both mutant p21 and mutant p53 proteins [15.161. For these individuals, information was available on age, gender, race, smoking status, alcohol consumption, occupational exposure his tory. and current medical status, including the presence of hepatic or other neoplasia. Estimates of VC exposure were based on veais worked with VC as well as on estimated ppm-years of VC expo sure based on the exposure matrix of Heldaas et al. (for production, autoclave cleaning, maintenance and packing/do ing job cate gories exposures averaged 2000 ppm trom 1950 through 1954. ASI 000013964 1000 ppm from 1955 through 1959, 500 ppm from 1960 through 1967, 100 ppm from 1968 through 1974, and I ppm thereafter) fl7]. The 21 individuals studied were white males with an average age ot 58 (range = 57-741 and with an average exposure to VC ot 21 years (range = 2-57) or 7.107 ppm-searx iranse - 9-26,708). These individuals included five cases of ASL (four of which were known to contain codon 15 c-Ki-ms gene mutations and two of which were known to contain codon 249 or 255 p,55 gene muta tions), one case of hepatocellular carcinoma (HCC) without either the ras or p55 gene mutations, and 21 VC-exposed workers with out hepatic malignancies. For all workers, serum samples had been collected between 1987 and 1992 by routine venipuncture tech niques and stored frozen at -20C until the time of analysis. A sample of 18 control patients - group-matched for age, gen der and race - was selected from a previously described cohort of surgical patients hospitalized with non-cancer diagnoses [15. 16|. These controls were white males with an average age ot 65 (range = 54-81) and with no known exposure to VC. and they were also comparable to the exposed cohort in terms of cigarette smoking and alcohol consumption. For all controls, serum samples had been collected between 1987 and 1989 by routine venipuncture tech niques and stored frozen at -80C until the time of analysis. Serum analysis for mutant p21 was performed by immunoblotting as described previously using the primary mouse monoclonal antibody. D146. raised against a synthetic peptide corresponding to ammo acid residues 5-16 of the p21 protein with aspartic acid at position 13 [15], Briefly. 2 pi of serum was diluted with 4 ml of sample buffer (0.125 M tnzma base. pH 6 8, 4% SDS, 20% glyc erol. 10% 2-mercaptoethanol). placed in a boiling water bath tor 2 min, and then loaded into individual wells of a 5%~17% gradient polyacrylamide gel with a 5% stacking gel and electrophoretically separated at constant current at room temperature The samples were then electrophoretically transferred onto nitrocellulose mem branes at constant current at 4C. The nitrocellulose blots were blocked with 3% BSA in PELS for 2 h at room temperature and in cubated with the primary antibody DI46 (t 250) overnight at 4C. After washing, the blots were developed with a secondary anti body (horse antt-mousej-avidin-biotin-peroxidase detection sys tem. Serum 21 kDa bands were scored in comparison with the 21 kDa bands of cell lysate from the positive control human cancer cell line. HCT 116. known to contain the c-Ki-ras codon 15 aspar tic acid mutation, by laser densttometric analysts of the volumetric integration of the optical density of the bands. Positive mutant p21 serum bands were confirmed as c-Ki-ru.t p21 by immunoblotting with another primary mouse monoclonal antibody. 147-67C6. raised against a synthetic peptide corresponding to ammo acid residues 157-180 of the c-Ki-ras p21 protein. Serum analysis for mutant p53 was performed as described previously using an ELISA based on the mouse monoclonal anti body PAb 240 which identifies an epitope between amino acids 210 and 214 that is normally concealed in the wild-type p53 pro tein but which is revealed m many mutant p53 proteins due to the aforementioned conformational changes produced by the muta tions [18], Briefly, microtiter wells were precoated with PAb 240 and then 100 |il of serum was added to each well and incubated overnight at 4C, After washing, 100 |4l of a rabbit polyclonal re porter antibody for p53 was added to each well and allowed to in cubate at room temperature for 2 h. After washing again, the re maining reporter antibody was bound to a horseradish peroxidaseconjugated goat antt-rabbit IgO. and color was developed by incu bation with the chromogenic substrate 2, 2'-azmo-dt-[3-ethylbenzthiazoline sulfate]. Absorbance of each well was read on a spectrophotometric plate reader at 405 ntn, and the concentration of mutant p53 was determined by comparison to the absorbance of standard solutions of purified, recombinant, human mutant p53. Results and Discussion The four of five cases of ASL (80%) among the VC-ex posed workers that were known to contain c-Ki-am gene mutations in their tumor DNA were found to have de tectable amounts of the corresponding mutant p21 protein in their serum, whereas the one case of ASL and the one case of HCC that were known not to contain such a muut tion in their tumor DNA were found not to have mutant p21 in their serum. Similarly, the two of five cases of ASL (40%) that were known to contain p53 gene mutations m their tumor DNA were found to have detectable amounts of mutant p53 protein in their serum, whereas the three cases of ASL and one case of HCC that were known not to contain p53 mutations in their tumor DNA were found not to have mutant p53 in their serum. These results sug gest that detection of serum mutant p21 and serum muum p53 can be valid surrogates for mutant ras and mutant py) gene expression at the tissue level in vivo. The tesults among the other VC-exposed workers without malignancies and the unexposed controls showed considerable differences between these two groups in terms of their serum expression of these mutant proteins None of the 18 unexposed controls (0%) were found to have detectable levels of the mutant p21 in their serum, whereas 9 of 21 VC-exposed workers (43%) were serum positive for mutant p21. Similarly, only 1 of 18 unexposai controls (6%) were serum-positive for mutant p53 com pared to 3 of the 21 VC-exposed workers (14%). These results suggest that mutational activation of the ras and p53 genes may be an early event in VC-induced carcino genesis and that identification of the mutant p21 and mu tant p53 proteins may be potential biomarkers of cancer risk in these individuals. For example, one can make an estimate of the predictive value of the biomarkers based on the current data. At this time, each biomarker by itself has a relatively low estimated positive predictive value (0,29 for mutant p21 and 0.40 for mutant p53). However, with the results of the two tests combined, the positive predictive value for both tests positive at the same time in creases to 0.67 (with a negative predictive value it boih tests are negative or only one test is positive of 0.88), Fur thermore. over time one would anticipate an increase m the predictive value of the tests if they accurately repre sent critical events in the causal pathway for carcinogene sis. as presumed. Follow-up of the serum-positive and serum-negative workers in this cohort should provide confirmation of this. For example, to date, all serum-neg ative individuals remain healthy, but one of the serum positive individuals has developed a lesion consistent with ASL. In addition, the current results with the higher percentage of p21 positives among the exposed compared to p53 positives may indicate that ras gene mutation is a very early event in VC-induced carcinogenesis but that p53 gene mutation may occur at a somewhat later stage Such a step-wise model of VC-induced carcinogenesis would thus be similar to other models of carcinogenesis such as that proposed for colon cancer in which ras gene mutation is an early event and p53 gene mutation occurs later [19]. The serum results for all workers and controls were also stratified by degree of VC exposure ;n terms of year' worked and ppm-years exposed (Table 1 A and B). 9mli ASI 000013965 I 397 Table 1 Dose-response relationship between serum biomarkers, mutant p21 and mutant p53. and VC exposure in years and PPMYears. Maximum likelihood logistic analysis statistically signifi cant for increasing biomarker positivity with increasing exposure (tor A. P = 0 0021: for B, P = 0.0019) Both One Both biomarkers biomarker biomarkers Negative Positive Positive A. Years of exposure 0 17 1 0 < 15 340 15-25 45 1 > 25 44 B. PPM-Years of exposure 0 17 l 0 A b O 1.80CM 8.000 > 18.000 440 6 7 "> 121 increasing exposure by years or ppm-years. there was a statistically significant increasing likelihood of positivity for one or both serum biomarkers (maximum likelihood logistic analysis, P < 0.01). For example, among the indi viduals without either serum biomarker, the percentage of individuals in the strata defined by years fell from 94% (17/18) among the unexposed to 40% (4/10) among those with more than 25 years of exposure, whereas among the individuals with both serum biomarkers, the percentage rose from 0% (0/18) among the unexposed to 20% (2/10) among those with more than 25 years of exposure. Simi larly, among the individuals without either serum bio marker, the percentage of individuals in the strata defined by ppm-years fell from 94% (17/18) among the unex posed to 25% (1/4) among those with more than 18,000 ppm-years of exposure, whereas among the individuals with both serum biomarkers, the percentage rose from 0% (0/18) among the unexposed to 25% (1/4) among those with more than 18,000 ppm-years of exposure. Likewise, the average exposure can be seen to be increasing with the presence of the biomarkers from 7.8 years (2,421 ppmyears) among those negative for both biomarkers to 19.3 years (6,571 ppm-years) among those positive for one biomarker to 26.3 years (10,708 ppm-years) among those positive for both biomarkers. Once again assuming that the presertce of the biomarkers represent the occurrence of critical events in the causal pathway for carcinogenesis, these results would be consistent with epidemiologic stud ies that have suggested an increased risk of ASL with in creasing VC exposure f 17], An important goal of molecular epidemiology studies such as this is to contribute to improved disease preven tion and control. For example, as noted, biomarkers of presumed critical events in the disease pathway may be better able to identify individuals at risk before clinical onset and thus allow more effective intervention. At the present, there is no specific intervention that could be of fered to individuals with the presence of these biomarkers that is known to mitigate the effects of VC-induced can cer-related events. However, progress is being made in developing such interventions. For example, we have al ready demonstrated that the compound L-p-(5-hydroxy-2pyridyl)-alanine (azatyrosine) is capable of causing per manent reversion of the human cancer cell line HCT116 (which is known to contain the Asp 13 c-Ki-tus VC-in duced cancer-related mutation) to a normal cellular phe notype that will no longer grow in soft agar and will no longer produce tumors in nude mice ft]. It is hoped that this type of compound will serve as a prototype chemoprophylactic agent that can eventually be employed in hu man studies for the prevention of ASLs in VC-exposed cohorts. Furthermore, it is anticipated that studies of biomarkers such as these may have broader implications for car cinogen risk assessment and regulation. For example, the presence or absence of biomarkers among subgroups of workers with varying levels of exposure could provide in termediary evidence for a potentially protective exposure level. Thus, in most Western countries, the workplace ex posure limit for VC has been 1 ppm since 1974. However, if some individuals with such low exposure levels (e.g., below 40 ppm-years or 1 ppm for 40 working years) tested positive for these biomarkers, this could suggest that current permissible exposure limits are not ade quately protective. Conversely, if it is only individuals with much higher exposures who test positive for these biomarkers, this would support the level of protection provided by current exposure limits. As noted above, most of the individuals in the present cohort had ex tremely high levels of exposure. However, the range of exposure was quite broad and included a few individuals with low levels of exposure. Thus, it is interesting to note that one of these worl.-rs with an estimated exposure of only 9 ppm-years tested positive for mutant p21 (although negative for mutant p53). Expansion of studies such as these to include greater numbers of workers with low lev els of exposure will be necessary to adequately address these regulatory concerns, as well as related issues of can cer prevention and control. Acknowledgements We thank Dr. C. Trepo. D. Tang. Y Zhang and S. Kahn for their assistance with this research. This work was supported in part by grants from the U.S. Environmental Pro tection Agency (R818624). the National Institutes of Health (CA69243), the Association pour la Recherche sur le Cancer and le Groupment des Enterprises Francatses pour la Lutte contre le Cancer. References 1. Brandt-Rauf PW. et al. (1995) The molecular epidemiology of growth signal transduction proteins. J Occup Environ Med 37, 77-83 2. Brandt-Rauf PW. et al. (1995) Mutant p21 protein as a btomarker ot chemical carcinogenesis in humans. In: Mendelsohn ML. Peelers JP. Normandy MJ (eds) Biomarkers and Occupa tional Health: Progress and Perspectives, Washington. DC. Joseph Henry Press, pp 163-173 a AS I 000013966 398 3 Agency for Toxic Substances and Disease Registry (1993) Toxicological Profile for Vinvl Chloride. TP-92/20. Atlanta. USDHHS/PHS 4. Barbin A. Bartsch M (1986) Mutagenic and promutagemc pro perties of DNA adducts formed by vinyl chloride metabolites. In: Singer B, Bartsch H (eds) The Role of Cyclic Nucleic Acid Adducts in Carcinoaenesis and Mutagenesis, Leon, IARC, pp 345-358 5.Swenberg JA, et al. (1992) Etheno adducts formed in DNA of vinyl chloride-exposed rats are highly persistent in liver, Car cinogenesis 13:727-729 6 Barbacid M (19871 Ras genes. Ann Rev Biochem 56'779-827 7 Marion M-i. et al, (1993) Ras gene as a molecular epidemiologic marker m human cancers induced by chemicals. Proceedings of the International Congress on Occupational Health 24 175 8. Marion M-J. et al. (1991) Activation of Kt-ras gene by point mutations in human liver angiosarcoma associated with \myl chloride exposure. Molec Carcinogen 4,450--354 9 Bos JL. et al (I985i Ammo acid substitutions at codon 13 of the N-ras oncogene in human acute imeloid leukemia Nature 315.726-730 ' lO.Ptncus MR. et al. (1985) Correlation of structure with trans forming activity of the p21 proteins with substitutions of Lummo acids for Oly at position 13. ) Protein Chem 4 345-352 ll.LaVecchio JA, et al. (1990) Characterization of monoclonal antibodies specific to the activated ras p21 with aspartic acut position 13. Oncogene 5: 1173-1 178 I2.Soussi T. et al. (1994) Multifactonal analysts ot p53 alteration in human cancer: a review. Int J Cancer 57 1-9 13. Hollstein M. et al. (1994) P53 mutations at A.T hase pairs m angiosarcomas of vinyl chloride-exposed factory workers. Cjictnoeenesis 15:1-3 14. Gannon JV. et al. (1990) Activating mutations in p53 produce a common conformational effect-a monoclonal antibody spe cific tor the mutant form, EMBO Journal 9:1595-1602 15. DeVivo I. et al. (1994) Mutant c-Kt-ras p21 protein in chemi cal carcinogenesis in humans exposed to vinyl chloride. Cancer Causes Control 5 :273-278 16. Smith S. et al. (1996) Mutant p53 protein as a biomarker ui chemical carcinogenesis in humans J Occup Environ Med tin press) 17. Heldaas SS. et al. (19X4) Incidence of cancer among vinyl chloride workers Br J lndustr Med 41 :25-30 IX Luo J-C. et al. ( 19V5) Lc'vels of p53 antigen in the plasma ol patients with adenomas and carcinomas of the colon. Cancel Lett 91 : 235-240 19. Fearon ER, Vogelstein B (1990) A genetic model for colorec tal tumorigenesis. Cell 61 :759-761 ASI 000013967