Document k6jBqEObjoqbmzp2QZobMvp2n

v' - --**' Job-rsated cancer. likely **< w *rv0 7.. to, increasMeS//exEsssi savs J*.T. *t A Xi tu !** al<Vd. . -, M , DY JOHN OKODA Chronicle Staff t ' These phenomena were reportedly caused by Exposure lo cancer*'' 'chemicals, buried during .. causing, substances is ktlt- 1 - tlie 1930's. that reached the ;'Ing tens of thousands of' surface and contaminated ''/workers, and the situation * the surrounding area. ' wilt probably worsen, a re- Eckhardl said the feder . .searcher-who has. spent al government has spent decades studying job-,1 $23 million trying to "clear related caucer said here. .1 up the area." in contrast, Dr. Irving. J. Selikofrof he.said, to an estimated -'Ml Sinai H'osnital in-New < "dost of $4 million to York said Friday that in.1 properly dispose of the tense exposure to carcino- j -chemicals, s-genssticn atasbostos can/TEckhardl said-;.lhe- i.; produce cancer,rates as.:./proper handling of car* ,:<;muchas 1,000timesgreat-*c)-cinogcns- may,be much er thau occur naturally in $1 cheaper than industry* -/.humans. .- r*-.-.--- -- jf opponents argue. He said' /.^Speaking at/a-'.confer->1 that when regulations for ence on job-related can- aLthe carcinogen vinyl chlo- - cer,' Sslikoff said cancer'^ rtide were imposed in 1974. '; rales arc highest among the government cstwpated workers who expose them* that industrial changes -'`selves-, to carcinogens would cost Sl.85 billion, '(cancer-causing sub. stances!'on the-jol) and: <* wh^also smoke. : v. Most of Selikoffs audi*. cnce were workers and but the industry estimate ;was close to $3 billion. Eckhardt said a -recent .study found "thrt the eftange had cost industry A* - inference was Dr: Eula with a panel of area work* gin,"ham, head of the U.S. ers voicing their griev- ' Occupational Safety and - ances against local indus Heaith Administration. try. reporting Incidents in - (OSHA), who said that one . ,which, they said, their em third of all cancer deaths ployers refused to take * * may be job-related. ' steps to protect them from ' '*1 am now convinced industrial carcinogens. that many factories have a . "What scares workers carcinogen-riche nviron- more than anything else te- ment," she said. "And if ' the latent period of can- , there are-carcinogens in* cer," said Rod Morgan of . the work place, there are Texas City. "They may. carcinogens outside of the have something today and'- ! .work place." *v... r-'.not find out for 2D years." ' ... Dr. Bingham said some- Fred Mabry, represent** industries that use car-' ring the United Steelwork . cinogens are moving to. ers of America, criticized. licxico and South America - companies that refuse to- . to escape OSHA restric take action to protect: tions. She said her depart workers from carcinogens- ment communicated with ontii after statistics have its counterparts south of proven a danger exists: i the border to urge the "Ycu'-ve got to count the. adoption of American fair dead-bodies before any-1 labor standards, "which thing gets done. You've' Include the protection cot to fmd the bodies be**. 'from chemicals.'! ' fore anybody cares," he* .. The conference began said. \ Bvenunent officials, lie onlv_&i27 million. about hed mainly aboutasbes* - ose-fifteenih of the govern los. 'atlhough lie referred ment estimate and one* -to other carcinogens such forty-fifth of tha. industry .as benzene and vinyl chio- estimate. .. - *' ride. , \ ;> <.. . ^ Eckhard said the lower -**- Selikoff reported-a *gov-' ; eminent estimate that : about 10 percent ofall can-Xzr deaths'in the United 'States may be related to cost could be attributed to innovations in disposal and control ofthe chemical. Another speaker u the ......... V-*--" -asbestos exposure. Then he noted that a local can cer expert. Dr. R.A. SlaJ- Jone. her concluded that this carcinogen may cause even more deaths, as -many as 14 percent of ell 'cancer fatalities in the na tion.- Sefikoff said one of five asbestos workers dies cf cancer, although the dis ease often appears 1$ cr 2D years after first exposure . to the carcinogen. Congressman Bob Eck- V r Saturday, March 31,1379' Houston Chronicle `bardl spefco to the confer ence about the effects end costs of chemical dimming grounds. He said dumping at Love Canal dump near . Ningra Falls was followed ! by miscaiTiugos m several women and by rates of sickness and cancer far hif.lvcr tliaji the national average. APOOO18430 421 LFTTEH chemicals cause pathological damage to sensitive as his test on rodents: chloro tissues and increase their susceptibility form would have to be more than S times Chemical Carcinogens to cancer. At low doses few tumors as potent (on a milligram per kilogram should appear. However, the evaluation per day basis) in dogs as in rodents to be Thomas H. Maugh It's recent articles ofchloroform bioassay data attributed to positive in his tests. Thus, Roe's experi on chemical carcinogens (Research News, Stokinger in support of this argument is ments on the carcinogenicity of chloro 29 Sept. 1978. p. 1200; 6 Oct. 1978, p. 37) incorrect. Stokinger is eited as saying form are consistent with those from NCI were both provocative and timely, al that rodents in the National Cancer In and support the conclusion that chloro though we were disturbed by several er stitute (NCI) test had "severely dis form is a renal and hepatic carcinogen in rors of fact and interpretation. Our con tended abdomens, markedly shortened rodents. cern stems in part from preliminary re* life-spans, and extensive liver dam Thresholds, or a disproportionately suits ofour 2-year study (/) investigating age. .. .XTumors were observed in only small effect at low doses, may someday quantitatively the carcinogenic potencies the most debilitated animals." In fact, be demonstrated for certain carcinogens, of more than 500 chemicals adequately however, in the bioassay on mice, (i) ab but there is no evidence that they are the tested and reported in the literature. We dominal distension was caused by the general case or that they `nave been dem have also quantified the sensitivity of liver tumors induced by chloroform onstrated for vinyl chloride or chloro negative cancer tests, expressing how treatment: (iQ the life-span of three ofthe form. potent a carcinogen would have to be to four dose/sex groups did not differ from In addition to the possibility of thresh be detected by these tests. the controls, although each group did de olds, Maugh also suggests that predicting A dispute of long standing in regula velop a high incidence of liver tumors: human risk from animal cancer tests is tory circles is whether chemical carcino and (iii) neither extensive liver damage extremely difficult because of large vari gens exhibit "thresholds," or doses be (for example, necrosis), reduced weight ations in carcinogenic response between low which there is either no effect gain, nor any other pathological sign of exes and species. Whether this is a diffi or an effect considerably less than that nonneoplastic debilitation was associat culty, however, can only be decided by a predicted by linear extrapolation from ed with chloroform treatment (7). Fur systematic, quantitative analysis of the responses at higher doses. Two investi thermore, in the NCI rat bioassay, data, including an analysis of the power gators are cited by Maugh in support of chloroform treatment increased kidney of negative cancer tests. Our preliminary the "threshold theory," Gehring on vi and thyroid tumors significantly without results using this type ofanalysis suggest nyl chloride and Stokinger on chloro any increase in nonneoplastic damage to that there is generally good agreement form. Their arguments are either mis these tissues. between species, sexes, and strains. leading or inaccurate and do not support The assertion that Roe "fed chloro Considering the more than millionfold the contention that the risk to society form to mice, rats, and dogs at doses that range in carcinogenic potency observed from these chemicals is negligible be did not produce pathological damage, among chemical carcinogens, even an cause of thresholds. and these animals show no evidence of order-of-magnitude agreement between The first argument is that mechanisms malignant tumors," is incorrect in two sexes and species would be quite useful, detoxifying carcinogens are over respects: Roe's doses in rodents (i) did especially in supporting priorities for whelmed at the high dose levels often produce pathological damage and (il) regulation. used in rodent bioassays, and a dispro did produce malignant tumors,'although With respect to the relationship be portionately large carcinogenic effect oc Roe found no relationship between path tween carcinogenic potency and muta curs at these doses. As an example, Geh ological damage and tumor incidence in a genic potency from short-term assays, ring implies that at dose levels of vinyl tissue (8). Roe's test on dogs was not as Maugh refers to the reservations ex chloride above 50 to 150 parts per mil lion, where detoxification mechanisms may be saturated, a disproportionate in pressed by Ashby and Styles that such a relationship is tenuous, at best. Yet Maugh fails to mention, as we observed crease in carcinogenic effect in rodents would occur. Yet the animal bioassay re elsewhere (9), that their reservations were based on experiments with serious sults of Maltoni (2), Lee et oi. (5), and Viola tt at. (4) on vinyl chloride suggest, if anything, just the opposite: as in conceptual and technical errors. While we would not expect that the correlation would be precise, preliminary evidence halation exposures increase beyond 150 (JO) suggests that a battery of short-term ppm, the increase in carcinogenic re tests may be quite useful in approximate sponse is less than that predicted from a linear extrapolation of the responses in ly estimating carcinogenic potency and thus in helping to. set priorities among the dose range 0 to 150 ppm (Pig. 1). Thus, the risk from exposure to low lev large numbers of new mutagens that have aot been tested in animals. els ofvinyl chloride might be more than a mQlionfoid higher than Gehring's esti Maugh also quotes Lijinsky as saying the Salmonella (Ames) test is not a use mate (5), Based on recent pharmacoki ful qualitative predictor of carcinogenic netic studies on vinyl chloride (5), the ex ity on the basis that it has only a 55 per planation for this proportionately greater activity at low doses may be that the mechanisms that activate vinyl chloride to the proximate carcinogen are saturat ed at high doses. A second argument cited in support of the threshold theory Is that high doses of Pig. I. Tumors from inhaled doses of vinyl chloride (?). The percent of Sprague-Dawley rets or Swiss mice with turnon is plotted against the vinyl chloride concentration in air (4 hours daily) over two dose ranges, 0 to 500 ppm (lifetime observation) and 0 to 5 ppm (87week observation). There is no evidence for a threshold. cent correlation for a group of polycyclic hydrocarbons that he tested for both mutagenicity and carcinogenicity (//). However, the Ames test has been exten sively validated in three independent studies (12,13), all of which agreed that approximately 90 percent of the carcino- M2 SCIENCE. VOL. 203 I t APOOO18431 usa^merjussHsmsneu-Miie } gens from a wide variety ofclasses were mutagens in the test. Furthermore* Lijinsky's study does show that the Salmo nella mutagenicity assay is an excellent test for the detection of carcinogenic polycyclic hydrocarbons: 88 percent of the carcinogenic hydrocarbons analyzed (14 out of 16) were mutagenic in Salmo- MUa. This agrees with a larger study on polycyclics by Coombs et el. {14), in which 38 out of 38 carcinogenic polycy clic* were mutagenic in the test. Mjlnsky obtained the "35 percent cor relation** estimate by combining his re sults on the 16 carcinogens with the 8 "noncarcinogens," of which 8 out of 8 were mutagens. Unfortunately* Lftinsky accepts the results of limited or in adequate cancer tests as definitive evi dence that a chemical is a "noncarcinogen." But the word "noncar- emogen" has meaning only in the con text of an analysis of the power of the test. We have discussed elsewhere US) the problems with using Lflinsky's in sensitive cancer tests for determining whether or not a chemical is a "non- carcinogen." His analysis of carcinoge nicity was based entirely on skin painting studies, which are considerably less sen sitive than the standard NCI feeding studies and do not meet NCI criteria for an adequate test. It remains problem atical whether his skin painting studies would have detected very many of the known human carcinogens at the doses used, although the Salmonella test de tects almost all of these {13). N. Kim Hooper Robert H. Harris Bruce N. Ames Department ofBiochemistry, University of California, Berkeley 94720 Rcbmect and Netet 1. N. K. Hoopar, C. Sawyar, R. peto, R. Harm, A. Friedman, B. N. Amei, in preparation. 2. C. Mallosi. in Conferences on Cell Prolifero Hen, voL 4, Origins of Human Cancer, H. H. Hiatt, J. D. WaUon, 1. A. Wtaran. Edi. (Cold Spring Harbor Laboratory. Cold Sprint Harbor, N.Y., 1977), book A, pp- 119-144. 9. C. C. Lee, ), C. Bhandari, M. Winston, W. B. House, R. L. Diuo, J. S. Woods, I. Toxicol. Environ. Health 4, IS (1971). 4. P. L. Viola, A. BaSMii, A, Caputo, CanerrFes. tusutwi). 5. Cebrinf t estimate, bated at an extrapolation (torn rodent data. It that the ride of Hver angio sarcoma in humant is about t0~* from esposun for over half a lifetime to 1 ppm of vinyl chlo ride. This contrasts with a risk of Ever angiosar coma* in tots of greater than !0~* (and mam mary carcinomas of greater than 10-*} from ex posure for over half e lifetime to I ppm C): in extrapolating to humans we estimate the risk of cancer might be in the range of t0~* to 10"'. 4. F. 0. Wetanabe, J. A. Zempel, 0.0. Pug- F JCehriitg, Toxical. Appl. Pharmacol. 44, 971 11978): P. ). Cehrtng, P. O. Wataaabe, C. N. Part,/Wrf.,p. 981. It should alu> be noted Ihat, jf Maugh'i figure I, which wgiesu a threshold. Is repotted with dote on a linear scale. th data ftt equally well a straight line passing through the origin, suggesting no threshold. 7. "Report on the carcinogenesis bios*say of chloroform" (NationalCancer Institute, Belbea ts FEBRUARY 1979 da, Md., 1 March 1974); "Individual animal pa thology table for chloroform" (National Cancer Institute, Bethesds, Md.. 24 September 1974). I. F. J. C. Roe, unpublished prtNmrairy repori of long-term tests of chloroform in rati, mice, and dogi. In four strains of mice. tumors developed la tissue which showed no tnher pathological damage, and tumors did not develop in tissue which showed pathological damage: ICI-Swisg mica had kidney tumors but no kidney damage; CBA and CRN mice had kidney damage but tto kidney tumors; and ICI-Swisi ntitt had fiver damage but no Ever tumors. 9. E. N. Ames and N. K. Hooper* Mature {Lon don) 774. 19(1978). 19. M. Meselton and K. Russel), ie Conferences on Celt FroBftrailon, vol. 4, Origins of Human Canter, H. H. Htan, J. D. Watson.I A. Wiaatea, Eds. (Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1977), book C. pp. 1473--1481. H. A. W. Andraws, L. H. TWbaolt, W. LEtnsky, Mutat. Res. 91,311 (1978). 12. I. F. H. Purchase. E. Longetaff, J. Ashby, J. A. Styles, D. Anderson, P. A. Lefevrt, F. R. Westwood, Harare (London) 844.624(1976);T. Supmura #< at., m Fundamentals la Cancer Pre vention, P. N. Magee rf at., Eds. (University of Tokyo Pmss-Univenlty Pash Prats, Baltimore, 1974), o. 191. 13. J. McCann, E. Choi, E. Vunasiki, B. N. Ames, Proc, Hail. Acad. Set. U.S.A. 72. 9133 (1979): J. McCann and B. N. Amei, ibid. 73, 950 (1974); in Conferences on Celt Prolif eration,' 'A.t.OrtglnsofHumanCancer, H, H. Hiatt. J. O. Watson, J. A. Winslet). Eds. (Cold Spring Hsrbor Laboratory, Cold Spring Harbor, N.Y., 1977), book C.pp. 1431--1450. 14. M. M. Coombs. C. Dixon. A.-M. RJtsonerghii. Canter Re*. 34, 4525 (1974). 15. B. N. Ames and L. Haroun. in preparation. For example, Lijinsky (//) lists benzo|r]pyrtne as one of the eight-'noneareinogen*,"' The Salmo nella test shows that benzofelpyrene is a muta gen, but it is only 1/360 ms potent as beazofelpynne. Other short-term tests using ani mal ceils also show that bcnzo[e]pyrene is a mutagen. Before Lqinsky can conclude that these tests are unreliable because ben[r)pyren* is not acardnogen. he should show that trie power of the cancer test such that it could detect chemical 1/340 as potent as beetxetdlpyrene. In his review (Research News, 6 Oct. 1978, p. 40) of some of the arguments concerning the existence of threshold dotes in carcinogenesis, Thomas Maugh reproduces a graph {from David B. Clay- son (/, p. 164)] as the basis for some people's argument that high dose-low dose extrapolation of animal data (and hence extrapolation from animal to man) b a nearly impossible task. The graph re produced is incomplete in that the dose scale--log dose--that was in the original article is left out. That graph implies m biological mod el--the logiuithm of the percent of ani mals developing a tumor is linearly re lated to the logarithm of the dose. Not one ofthe extrapolation techniques that I know (Mantet-Bryan, Crump tt al,, Al- beit-Altschuler, one-hit, multi-hit, log- logistic, and others) uses this model. I know no theory of carcinogenesis that is consistent with such a model. Gently put, the model is inappropriate. Finally, the confidence intervals as drawn can come only from an experiment without a low response at some low dose, that is, an experiment without a control group. If the point is that an inappropriate bi ological model combined with a defec tive experiment is likely to give nonuse ful answers, I agree wholeheartedly. If the point is that that graph provides a se rious argument against extrapolation us ing a reasonable biological model from data gathered in a well-conducted exper iment, then serious questions arise in my mind about its pertinence and their logic. By way ofanalogy, I find that a picture of a centipede is not proof (or even evi dence) that a horse is not a quadruped. Marvin A. Schneiderman National CancerInstitute, Bethesda, Maryland20014 RcSh*mm* t. D. B. Ctayaoa, Cnor Butt, 29, SCI (1977). There is a continuing national debate over determining the risks of small amounts of toxic substances by extrapo lations of massive exposure studies. The legend of figure 1 of Maugh's 6 October 1978 article on chemical carcinogens states "Most scientists, however, now think that the actual response is in dicated by the smooth curve passing through zero dose and zero response-" Dose Is plotted on a logarithmic scale. In figure 2 response is plotted on a logarith mic scale. How do "most" scientists plan to extrapolate these logarithmic scales to zero? All models of dose versus response in which the curve increases monotonically must include the points (0.0) and (*, 100 percent), but this does not help to distin guish among them. In probit theory the relationship between log dose and sus ceptibility is a normal curve. Response is the integral ofthis normal curve which is "S shaped" but can give the impression of a straight line over narrow dose ranges. Probit analysis requires at least two real points to independently eval uate potency and the standard deviation of a diverse population to this effect. It is ironic that so much energy is wasted on a debate over what amounts to the proper choice ofcoordinate axes, Frank J. Weioert StratfordApartments, J-10, Shipley and Naamans Roads, Wilmington, Delaware I98W Mexican OU Reserves All of us owe a debt to WiDiam D. Metz for his commentary "Mexico: The premier oil discovery in the Western Hemisphere" (News and Comment, 22 Dec. 1978, p. 1261). Recognition in the United States of the Mexican accom plishment is long overdue, and Metz makes his point in a most effective man- 603 APOOOf 8432