Proceedings of the National Academy of SciencesVol. 67, No. 2, pp. 908-915, October 1970

Evaluation of the Effects of Chemical Mutagens on Man:The Long Road Ahead*

James V. Neel

DEPARTMENT OF HUMAN GENETICS, UNIVERSITY OF MICHIGAN MEDICAL SCHOOL,

ANN ARBOR, MICH. 48104

Contributed to the Symposium on Aids and Threats to Society from Technology, April 29, 1970

Abstract. By analogy with the problem of evaluating the genetic risks of radia-tion, it appears that it will be extremely difficult to assess the mutagenicity forman of the wide range of chemicals to which human populations are currentlyexposed. Nevertheless, the potential significance of this problem calls for a majoreffort at such an evaluation.

One of the recurring themes in any consideration of environmental pollution,whether by chemicals or radiation, has been the question of the genetic dangersinvolved. Although by invitation I am to speak on the problem of chemicalmutagenesis in man, in fact most of these remarks will, for three reasons, beconcerned with radiation mutagenesis. Firstly, the two subjects are conceptu-ally almost inextricably interrelated. Secondly, the very same problems thathave arisen in trying to understand the genetic threat of increased radiationexposures will arise with respect to trace chemicals-not to mention a few newones. There is much to be said for viewing the potential threat of chemicalmutagenesis from the perspective of 40 years of radiation genetics. Thirdly,such is our ignorance of chemical mutagenesis in man or any other mammal thatit is difficult to find hard data to sustain even as brief a presentation as this.The thesis of this presentation will be that while there seems no immediate

danger of massive genetic damage from the trace chemicals, i.e., no geneticcatastrophe, it does seem quite possible that current exposures to trace chemicalshave increased human mutation rates. Since mutations are thought to havedeleterious effects rather more often than beneficial effects, this increase is to thedetriment of man. Unfortunately, it seems impossible to maintain the levels ofenergy production and food consumption on which our culture is based withoutsome increased exposure to mutagenic agents. Accordingly, unless we are pre-pared to forego many aspects of our present culture and at the same time dras-tically reduce population numbers, we shall probably have to live in the foresee-able future with some increase in mutation rates. We are thus confronted withthe problem of striking the best balance possible between genetic damage andthe benefits our culture has thus far shown no signs of abandoning, and to do thisintelligently we must collect the necessary information. It is not now at hand.The quantitation of genetic damage: It is impossible in the brief time avail-

able to begin to do justice to the vast literature on experimental mutagenesis.908

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1

VOL. 67, 1970 SYMPOSIUM ON SOCIETY AND TECHNOLOGY 909

The bulk of this literature deals with bacteria, fungi, protozoa, a variety of in-sects, and many plants, but there is an accumulating body of knowledge onmammals, to which these remarks will be restricted. There are in general twoapproaches to measuring induced genetic damage. One, to be termed the"specific locus approach," attempts to measure mutation at specific genetic loci.This approach, thus far limited among mammals to the house mouse, permitsvery precise measurements of mutation at certain loci, but then depends onsomewhat dubious assumptions concerning the number of genes and the mannerin which the average mutation expresses itself, for estimates of the actual harmsuffered by populations. The second approach, to be termed the "populationcharacteristics approach," bases the case on the effects of a mutagen on suchcomplex but important aspects of the study population as the frequency of con-genital malformations, survival rates from fertilization onwards, sex ratio,physical growth and development, and fertility. Each of these latter indicatorsis undoubtedly influenced by mutation at many different loci, as well as by en-vironmental variables, so that careful consideration of environmental factors isessential, and simple genetic explanations of positive findings unlikely. How-ever, there is little room for debate about the significance for a population of adefinite finding with respect to one of these indicators.The two approaches should be thought of as complementary rather than com-

peting, with, ultimately, each one illuminating the results of the other. Witheither of these approaches, it has been convenient to use the "doubling dose" asa reference point. This is the amount of radiation or chemical exposure whichwould result in a mutation rate twice the "normal" spontaneous rate, or whichwould double the genetic contribution to congenital malformations or earlydeath rate. Inasmuch as we have some knowledge of the contribution of spon-taneous mutation to human morbidity and mortality, some geneticists, includingmyself, feel that the calculation of a doubling dose provides an important frameof reference.The complexity of assessing the effects of a mutagen, as illustrated by de-

velopments in mammalian radiation genetics: The extensive work on radiationmutagenesis in small animals, especially the mouse, now provides many evidencesof the potential complexity of the chemical mutagenesis problem. With respectto the specific locus approach, the early studies on the mouse, concerned withthe genetic yield from spermatogonia receiving large doses of radiation over ashort period of time, resulted in estimates of the probability of mutation perlocus per R unit of approximately 0.6 X 10-7 (Harwell data) or 2.2 X 10-7 (OakRidge data). Estimates of the yield from irradiated oocytes fertilized shortlyfollowing the radiation treatment were slightly higher. The spontaneous rateof mutation per locus per generation in male mice is about 0.8 X 10-5. Theamount of radiation of this type necessary to increase the mutation rate to twiceits spontaneous value (i.e., the doubling dose) would thus be about 40-80 R.'12However, short of nuclear catastrophe, our species will not receive most of itsradiation in this manner. Rather, the exposures (above background) of bothsexes will usually be small, intermittent, and/or at low dose rates, and progenywill often be conceived at long intervals following the radiation. There have

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1

910 N. A. S. SYMPOSIUM: J. V. NEEL PROC. N. A. S.

been two important developments that indicate that under these circumstances,the doubling dose will be considerably higher. Firstly, there are dose rate anddose fractionation effects. With respect to spermatogonia, a dose rate of 0.009R/min produced only about one quarter as many specific locus mutations as wereinduced by a dose rate of 90 R/min.' '3 There is a similar dose rate effect infemales. Indeed, in oocytes receiving a dose of 400 R at the rate of 0.009 R/min,the mutation rate was not significantly elevated.4'5 Fractionation of a givendose of radiation also results in a lower yield of mutations than administration ofthe total dose in one exposure.6 Secondly, although in the case of spermatogoniathe yield appears to be essentially the same no matter whether progeny tests areconducted shortly after radiation or a year later, in the female it is quite different.Thus, whereas an increased mutation rate was readily demonstrated in the prog-eny of female mice receiving 50 R at the rate of 90 R/min when those progenywere conceived within 6 weeks of radiation, progeny conceived later demon-strated no increase in the mutation rate. These findings in my opinion clearlysuggest that given the human reproductive pattern, the doubling dose of radia-tion as received from industry and medicine is not likely to be less than 200 R.With respect to the "population characteristics" approach, in addition to the

mouse, rather extensive data are now available for the rat, pig and man. Asregards man, our own studies on children born to the survivors of the atomicbombings of Hiroshima and Nagasaki reveal no effect of the exposure on still-birth rate, frequency of congenital malformation, birthweight, death duringinfancy and childhood, or rate of physical growth and development.78 An ap-parent effect on sex ratio in the early years of the study was not borne out in thelater years.9 Incidentally, the other investigations of a possible sex-ratio effectin man have with impressive consistency yielded results similar to those of theearly years in Hiroshima and Nagasaki (reviewed in refs. 10 and 11). On ad-mittedly tenuous grounds we have suggested that for man the "doubling dose"of radiation of the Hiroshima-Nagasaki acute, high dose-rate type is probably notless than 50 R,8 an estimate in satisfactory agreement with the estimate of 40-80R derived from mice for the same type of radiation.There are now much better, controlled observations on experimental mammals

than those on man, involving large cumulative doses over several generations.In both the mouse and rat, autosomal recessive lethal effects from acute sperma-togonial irradiation occur with a frequency of approximately 1 X 10-4 per gameteper R. Dominant lethal effects, resulting in the death of the animal in the firstpost-radiation generation and so quickly eliminated, have about the same fre-quency (reviewed in ref. 12). Surprisingly, net dominant heritable damage hasnot been observed, i.e., there is no evidence for induced dominant detrimentalmutations, thought to be relatively frequent in Drosophila (reviewed in ref. 13).Taylor and Chapman'4 find the over-all rates of mutation to recessive lethaleffects and visible mutants in rats to be lower than anticipated from estimatesbased on the specific-locus rate studies in the mouse. The now-extensive swinedata reveal that 300 R of spermatogonial radiation has a significant positive effecton litter size in one of the two breeds studied, and no effect on post-natal mor-tality.'5 Thus, these studies tend to suggest that the doubling dose for the en-

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1

VOL. 67, 1970 SYMPOSIUM ON SOCIETY AND TECHNOLOGY 911

tire genome is greater rather than less than the above quoted figures based onspecific-locus studies.

Simultaneously with the accumulation of the foregoing data, there have beentwo important conceptual developments. During much of the past 20 years,a prevailing viewpoint, most readily associated with the writings of Muller, hasbeen that the mutations produced by radiation were except in very rare instancesharmful, and that each harmful mutation eventually resulted in one geneticdeath. However, as we emphasized some years ago,'6 selection is directed to-wards the total phenotype rather than a specific locus, so that the death of asingle particularly unfortunate individual might remove from the population anumber of deleterious mutants. Recently this concept, of threshold effects inselection, has been greatly refined mathematically;'7-'9 clearly under this as-sumption the cost of increased mutation rates in individual mortality and mor-bidity is not as great as under earlier formulations. Furthermore, serious doubtsare being raised as to whether the proportion of mutations which are deleteriousis as high as previously assumed.20I-22There have been two recent challenges to the foregoing rather conservative

view of the genetic risks of radiation. Sternglass23-2' has received wide publicityfrom his statement that in certain areas of the United States, fallout radiationfrom nuclear testing produced about a 60% increase in the fetal death rate. Thecomplete failure to demonstrate a causal relationship in this situation has been welldiscussed by Sagan26 and Tamplin.27 Secondly, Mbeyer, Diamond, and Merz28,29have recently reported that among offspring arising from oocytes exposed before30 weeks of fetal life to quite low levels of radiation (diagnostic pelvimetry)there is a relative preponderance of males (37 males, 18 females; ratio 67:33).The control sample contains 234 male and 254 female births (ratio 48:52).The probability of this or a greater difference is 0.0066 (not <0.001 as given intheir paper). The control sex ratio is unusual; were it the more usual 51.5:48.5,the probability of this or a greater difference would be 0.026. This deviation isopposite in direction to the (questionable) effects observed in studies on radiationof adult females (reviews in refs. 10 and 11), and opposite to the ratio in childrenin their own series resulting from oocytes radiated between 30 weeks of pregnancyand termination. In neither this nor the preceding instance do the authorsconcerned survey the totality of the data on the subject and point out just howmuch greater a sensitivity is implied than in the other studies on man by myrough calculation well over 100-fold, a difference difficult to accept. They alsooverlook pertinent contradictory data from the radiation of fetal mice.30 In-vestigators whose results and/or conclusions in so sensitive an area are so atvariance with a large body of knowledge, or those who quote them, have in myopinion a moral obligation to present their findings in the context of all the re-sults to date. Failure to do so can only further increase lay doubts of scientificresponsibility.

In summary, despite 40 years of active, imaginative, and extensive work onthe mutagenic effects of radiation-an agent whose tissue dose can be preciselyquantitated-we are still many steps removed from the ability to predict pre-

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1

912 N. A. S. SYMPOSIUM: J. V. NEEL PROC. N. A. S.

cisely the biologically significant results of human exposure, with public concernover any exposure still running high.The study of chemical mutagenesis in mammals, including man: There is

of course a very extensive literature on chemical mutagenesis in a variety oforganisms ranging from viruses to Drosophila, and it is out of the finding thatliterally dozens of chemicals demonstrably mutagenic in these organisms are nowpresent in the human environment that the present concern for the risks to mansprings. I will not stop to enumerate the usual litany of potential offenders,ranging from well-known pesticides and herbicides to antibiotics and othermedications.31-33 Mammalian experimentation with chemical mutagens islargely limited to the past decade; there is a paucity of data. Most of the ob-servations involve the induction of dominant lethals in the mouse or rat, thesedefined as dominant genetic changes incompatible with survival of the conceptus(whose "dominance," paradoxically, thus cannot be analyzed in the usual geneticsense). Some of the problems in mutagenicity tests of this type have been re-viewed by Rohrborn.34 Much of the early work involved a variety of alkylatingagents, especially the alkanesulphonic esters, but such other compounds asaflatoxin, trimethylphosphate, and benzo(a)-pyrene have also been found to bemutagenic in mice (reviews in refs. 35-40). These are potent mutagens in lowerorganisms. Caffeine, a weaker mutagen, has not yet been shown to increase thefrequency of dominant lethals in mice (review in ref. 41), and at the momentstands as an example of the difficulties in extrapolating from lower organisms tomammals.Most geneticists experience great difficulty in projecting from dominant

lethals to effects on populations. The more laborious demonstration of the pro-duction of specific locus mutations is just getting under way, Cattanach42'43 havingshown that triethylene melamine increases the frequency of point mutations in themouse. While the choice of this particular chemical was probably dictated byits known effectiveness in other experimental material, humans receiving thisor other cytostatic agents therapeutically seldom reproduce subsequently44;from the standpoint of human implications, an urgent problem is to identify justwhich mutagens should be given priority in investigation. Strain and sex varia-tion has already been demonstrated in the sensitivity of mice to dominant lethalinduction with triethylene melamine, 2,3,5-tri(ethylene-imino)-p-benzoquinone(Trenimon), and ethyl methanesulfonate,4547 indicating the complexity of whatis to come. Thus far the routes of administration have usually been intraperito-neal, and there is to my knowledge no evidence for the mutagenic effects ofchemicals administered through physiological portals, either gastrointestinal orrespiratory. The problem of estimating gonad doses of both a drug and itspossibly mutagenic metabolites has been discussed by Goldstein.32

Since the structures of the various chemical mutagens differ widely, it wouldappear unwise to extrapolate from one to another. Ideally, the effect of eachagent in physiological doses should be studied as carefully as radiation. Butwe have just seen how many uncertainties still persist for radiation itself. Thenecessary evaluations demand a major effort. Under the circumstances, thereis an urgent need for reliable shortcuts and screening procedures. Currently

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1

VOL. 67, 1970 SYMPOSIUM ON SOCIETY AND TECHNOLOGY 913

there is a very lively interest in the possibility that one can screen the knownchemical mutagens for their mutagenicity in the intact animal by their effective-ness in causing chromosomal damage or inducing somatic mutation in tissueculture material (discussion in refs. 40 and 49). However, while I count myselfan enthusiastic supporter of these various screening methods, I do believe thatas a basis for formulating policy concerning human exposures there is absolutelyno substitute for data on the ability of potential chemical mutagens to induceheritable changes in proteins and physiologically important traits in intact or-ganisms, preferably man.

I would like to close this presentation with what must be labeled a "typicalgenetic utterance." Man's most precious possession is his genetic endowment.Each generation holds it in trust for subsequent generations. As we struggle tomove towards those new levels of social and technological organization which willenable us to meet the kinds of problems we have been discussing in this sympo-sium, surely an important element in the decision-making process must be knowl-edge of the genetic cost. At the outset, I voiced optimism that the chemi-cal mutagens posed no threat of massive genetic damage. This viewpoint wasbased on rough extrapolations from experimental material, but I would be hard-pressed for a rigorous documentation. For instance, with the exception ofcaffeine,32 I have been unable to find any data concerning the accumulation inthe human gonad of potential mutagens following exposure. Furthermore, thequestions of possible synergism of different chemicals and of cumulative effects inlong-lived animals is largely unexplored.There have been several recent discussions of the various approaches to moni-

toring human populations for increased mutation rates from chemical expo-sures.33,40,U,49 In my opinion, recent technical developments in our ability todetect variant proteins by the relatively cheap techniques of electrophoresis setthe stage for the much more effective monitoring of human populations than inthe past. A single blood specimen permits the inspection of at least 20 and pos-sibly 30 different proteins for evidences of mutational damage. If 100,000 bloodspecimens from the umbilical cords of newborn infants were monitored for 20proteins each year, and if the spontaneous rate of mutation is 1 X 10-5, then anincrease from an observed 20 mutants in 2,000,000 determinations to 35 in asimilar number (or decrease to 9) would be significant at the 5% level. If oneassumed the lower figure provided a "baseline" estimate, then one could detectunder these circumstances a 72.5% increase in mutation rates. More sensitivedetection systems (and most would feel the foregoing to be minimal) can beachieved simply by increasing sample size. Such studies could be combinedwith a search for an increase in chromosomal abnormalities. This approach hasthe advantage that the investigator can select important target proteins, ratherthan (as with the specific locus approach in the mouse) being forced to utilizegenetic loci known to us by virtue of mutations with clear phenotypic effects-which may be nonrepresentative for that very reason.The practical difficulties-and there are some-do not appear insuperable.

Should a significant increase in genetically abnormal proteins be detected, weare of course confronted with the problems of translating this kind of information

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1

914 N. A. S. SYMPOSIUM: J. V. NEEL PROC. N. A. S.

into relevance to morbidity and mortality, as mentioned earlier. In the case ofthe chemical mutagens, we also face the issue of identifying which of many pos-sible offenders is at fault. But if there is no evidence of increased damage, wewill have substantially lessened one of the major concerns of a large sector of thebiomedical community. As for the cost to the United States-it is nothing be-side the cost of a year's logistical support for Viet Nam or even one supersonictransport. Finally, let me reiterate the opening statement: in order to supportour technology we may have to compromise with the desire of the geneticist forno increase in mutation rates-but we owe it to our offspring to see that the com-promise is based on knowledge rather than a guess we may later regret.

* The original investigations mentioned herein were supported in part by the U.S. AtomicEnergy Commission.

1 Russell, W. L., Jap. J. Genet., 40, 128 (1965).2 Searle, A. G., in International Congress of Radiation Research, ed. G. Silini (Amsterdam:

North-Holland Publishing Co., 1967), p. 469.3 Russell, W. L., L. B. Russell, and E. M. Kelly, Science, 128, 1546 (1958).4Russell, W. L., An. Acad. Brasil. Cienc., 39, 65 (1967).5 Searle, A. G., and R. J. S. Phillips, in Symposium on Effects of Radiation on Meiotic Sys-

tems (Vienna: International Atomic Agency, 1968), p. 17.6 Russell, W. L., and E. M. Kelly, Science, 154, 427 (1966).7 Neel, J. V., and W. J. Schull, The Effect of Exposure to the Atomic Bombs on Pregnancy

Termination in Hiroshima and Nagasaki (Washington, D.C.: National Academy of Sciences-National Research Council, 1956).

8 Kato, H., W. J. Schull, and J. V. Neel, Amer. J. Hum. Genet., 18, 339 (1966).9 Schull, W. J., J. V. Neel, and A. Hashizume, Amer. J. Hum. Genet., 18, 328 (1966).

10 Neel, J. V., Changing Perspectives on the Genetic Effects of Radiation (Springfield: CharlesC Thomas, 1963).

l Schull, W. J., Nucleonics, 21, 54 (1963).12 Taylor, B. A., and A. B. Chapman, Genetics, 63, 455 (1969).13 King, J. L., Genetics, 58, 625 (1968).14 Taylor, B. A., and A. B. Chapman, Genetics, 63, 441 (1969).15 Mullaney, P. D., and D. F. Cox, Mutat. Res., 9, 337 (1970).16 Neel, J. V., and H. F. Falls, Science, 114, 419 (1951).17 Sved, J. A., T. E. Reed, and W. F. Bodmer, Genetics, 55, 469 (1967).18 King, J. L., Genetics, 55, 483 (1967).19 Milkman, R. D., Genetics, 55, 493 (1967).20 Kimura, M., Nature, 217, 624 (1968).21 Kimura, M., Proc. Nat. Acad. Sci. USA, 63, 1181 (1969).22 King, J. L., and T. H. Jukes, Science, 164, 788 (1969).23 Sternglass, E. J., Bull. At. Sci., 25, 18 (1969).24 Sternglass, E. J., Bull. At. Sci., 25, 26 (1969).25 Sternglass, E. J., Bull. At. Sci., 25, 29 (1969).26 Sagan, L. A., Bull. At. Sci., 25, 26 (1969).27 Tamplin, A. R., Bull. At. Sci., 25, 23 (1969).2 Meyer, M. B., E. L. Diamond, and T. Merz, Johns Hopkins Hosp. Bull., 123, 123 (1968).29 Meyer, M. B., T. Merz, and E. L. Diamond, Amer. J. Epidemiol., 89, 619 (1969).30 Carter, T. C., Genet. Res., 1, 59 (1960).31 Barthelmess, A., Arzneim. Forsch., 6, 157 (1956).32 Goldstein, A., in Mutations, ed. W. J. Schull (Ann Arbor: University of Michigan Press,

1962), p. 167.33 U.S. Department of Health, Education, and Welfare, Report of the Secretary's Commission

on Pesticides and Their Relationship to Environmental Health (Washington: GovernmentPublications Office, 1969).

34 Rohrborn, G., Humangenetik, 6, 345 (1968).35 Loveless, A., Genetic and Allied Effects of Alkylating Agents (London: Butterworth and

Company, 1966).

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1

VOL. 67, 1970 SYMPOSIUM ON SOCIETY AND TECHNOLOGY 915

36 Vogel, F., G. Rohrborn, E. Schleiermacher, and T, M. Schroeder, Dtsch. Med. Wschr., 49,2249 (1967).

37 Vogel, F., J. Kriiger, G. Rohrborn, E. Schleiermacher, and T. M. Schroeder, Dtsch. Med.W8chr., 92, 2382 (1967).

8 Epstein, S. S., and H. Shafner, Nature, 219, 385 (1968).3Epstein, S. S., W. Bass, E. Arnold, and Y. Bishop, Science, 168, 584 (1970).40 Neel, J. V., and A. D. Bloom, Med. Clin. N. Amer., 53, 1243 (1969).41 Adler, I. D., Humangenetik, 7, 137 (1969).42 Cattanach, B. M., Mutation Res., 3, 346 (1966).43 Cattanach, B. M., Mutation Res., 4, 73 (1967).44Vogel, F., and P. Jager, Humangenetik, 7, 287 (1969).0 Cattanach, B. M., Int. J. Radiatn Biol., 3, 288 (1959)." Rohrborn, G., Humangenetik, 2, 81 (1966).47 Generoso, W. M., and W. L. Russell, Mutation Res., 8, 589 (1969).'" Crow, J. F., Scientist and Citizen, June-July, 113 (1968).49 Crow, J. F., in Environmental Chemical Mutagen8, ed. A. Hollaender (New York: Plenum

Publishing Corporation, in press).

Dow

nloa

ded

by g

uest

on

Apr

il 8,

202

1