Evaluation of genetic radiation hazards in man: History

J. Genet., Vol. 65, Nos. 1 &2, August 1986, pp. 79-102. �9 Printed in India.

Eva luat ion o f genet ic radiat ion hazards in man: his tory , present s tatus and perspect ives

K SANKARANARAYANAN Department of Radiation Genetics and Chemical Mutagenesis, Sylvius Laboratories, State University of Leiden, Wassenaarseweg 72, 2300 RA Leiden, The Netherlands

and

The J A Cohen Interuniversity Institute tbr Radiopathology and Radiation Protection, Rijnsburgerweg 86, 2333 AD Leiden, The Netherlands

MS received 18 December 1985

Abstract. The evaluation of genetic radiation hazards in man is an ongoing scientific enterprise from about the mid-1950s. Since estimates of genetic risks are essential for providing a basis for protecting our genetical endowment and since strictly relevant human data are limited, there is no alternative at present but to use the data from mouse and certain non- human primates. This paper reviews the general principles and methods that have thus far been used, appraises the evolution of the conceptual framework, the data base and the assumptions involved, presents current estimates of genetic risks and provides some perspective of the advances that are likely to be made in the near future.

Currently, risk estimates are made using the so-called "direct method" and the "doubling dose method". Both these methods involve a number of assumptions and consequent uncertainties. With the direct method, it is now estimated that following low LET, low dose- rate or low-dose irradiation of males, there will be (i) about 10-20 cases of affected children per million births per rad of exposure, who will suffer from the effects o f induced mutations having dominant effects and (ii) about 1 to 10 cases of congenitally malformed children (again per million births per rad), a consequence of the induction of reciprocal translocations. For irradiation of females under similar conditions, the estimated risks are 0-9 and 0-3 affected children per million births per rad, these being the consequence of induction of dominant mutations and of reciprocal translocations, respectively.

The doubling dose method is used to estimate risks to a population under continuous irradiation. If the population is exposed to low LET irradiation at a rate of 1 rad/generation (1 generation = 30 years), the expected total increments in the frequencies of genetic diseases are about 20 cases per million births in the first generation and about 150 cases per million births at equilibrium. These expected increments are very small fractions of the spontaneous prevalence of genetic and partially genetic disorders, currently estimated to be about 10.6 %.

Keywords. Genetic radiation'hazards; doubling dose methods; genetic risk evaluation.

1, Introduction

Widespread and serious concern over the genetic consequences of exposures of large numbers of people to low levels of ionizing radiation first arose in the aftermath of World War II, some 20 years after the discovery of the mutagenic effects of X-rays in

Although in current usage, the unit of absorbed dose is Gray (symbol, Gy; 1 Gy = lO0rad) and that of dose equivalent, Sievert (symbol, So; 1 Sv = 100 rein), the units used earlier, namely, rb'ntgen (R) for exposure dose, tad for absorbed dose and rein for dose equivalent have been retained to facilitate reference to the literature cited in this paper.

79

80 K Sankaranarayanan

Drosophila (Muller 1927) and in barley and maize (Stadler 1928a, b). This led to the launching of extensive research programmes to examine, among other things, the magnitude and consequences o f genetic damage from radiation exposure. One of these was directed at the survivors of the atomic bombs in Hiroshima and Nagasaki, while several others were focused on animal studies. In the mid-50s, one major international and several national scientific bodies came into existence, among which are: the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR), the Committee on the Biological Effects of Atomic Radiation (the BEAR Committee, renamed in 1972 as the Committee on the Biological Effects of Ionizing Radiation or the BEIR Committee) set up by the us National Academy of Sciences and the Committee of the British Medical Research Council. At least the UNSCEAR and the BEIR Committee have continued their work up to the present, keeping under review the levels of radiation to which the world population is or may be exposed and improving assessments of risks (genetic and somatic) entailed by exposure to radiation. Over the past 30 years, a number of reports have been published on this subject by these Committees, the latest of these being the one published by UNSCEAR in 1982 and by the BEIR Committee in 1980.

In spite of extensive efforts so far, genetic studies on Hiroshima and Nagasaki populations (as well as those on the offspring of individuals who, for one reason or another, have been exposed to large doses of radiation in radiological practice) have provided no unequivocal evidence for the induction of transmissible genetic damage in the progeny of exposed parents. However, the prodigious amount of data that has accumulated from mammalian radiation genetic studies and similar ones in a variety of biological systems amply document the fact that ionizing radiation is capable of causing damage to the genetic material. It is, therefore, unlikely that human germ cells are immune to the induction of genetic damage by ionizing radiation. Since estimates of genetic risks are essential for providing a basis for protecting our genetical endowment, and since strictly relevant human data are limited, there is no alternative to the use of experimental data in this regard. Therefore, the estimation of genetic risks to man from exposure to radiations remains a major exercise in extrapolation. In this, animal data (primarily those collected using the mouse and, in recent years, using some non-human primate species as well) are analyzed and those considered pertinent are used to assess (i) the amount of transmissible genetic damage likely to result from the exposure of the human population to radiation and (ii) their effects in terms of increased levels of genetic and partially genetic diseases.

Extrapolation inevitably involves a number of uncertainties and their nature and magnitude are dependent on the strengths and weaknesses of the data used and the kinds of assumptions that are made. The aim of this paper is three-fold: (i) to provide a general background of the principles and methods that are used in the evaluation of genetic radiation hazards in man; (ii) to trace the development of essential concepts, the data base, the assumptions and extrapolations involved, and (iii) to indicate in which areas progress is likely.

Evaluation of genetic radiation hazards in man 81

2. Germ cell stages and radiation conditions relevant in the context of evaluation of genetic radiation hazards

2.1 Germ cell stages

From the standpoint of hazard evaluation, the effects of radiation on two germ cell stages are particularly important. In the male, these are the stem-cell spermatogonia which constitute a permanent germ cell population in the testes and which continue to multiply throughout the reproductive life-span of the individual. In the female, the corresponding stage is the oocyte, primarily the immature ones. Female mammals are born with a finite number of oocytes already formed during foetal development. These primordial oocytes, as they are called, grow and a sequence of nuclear changes comprising meiosis takes place in them. The latter, however, are arrested at a particular stage until just before ovulation. Because oocytes are not replenished by mitosis during adult life, these are clearly the cell stages in the female whose irradiation has great potential significance for hazard evaluation.

The radiation exposures received by human populations are usually delivered as small doses at high dose rate (e.g., diagnostic radiology) or are greatly protracted (e.g., continuous exposures from natural and man-made sources). In therapeutic radiology, high doses of the order of several hundreds of rads may be delivered (and at high dose rates); however, such exposures, warranted on medical grounds, are given only to Selected individuals for treatment of specific cancers. In estimating genetic hazards to the population therefore, the relevant radiation conditions are: low doses and chronic (or low-dose-rate) exposures. ~

2.2 General assumptions

In using data from mouse (or those from non-human primates) to make quantitative estimates of risks in man, three principal assumptions are made and these are the following: unless there is evidence to the contrary, (i) the amount of genetic damage induced by a given type of radiation under a given set of conditions is the same in human germ cells and in those of the test species which serves as the model; (ii) the

i various biological (e,g., sex, germ cell stage, age, etc.) and physical (quality of radiation, dose rate, etc.) factors affect the magnitude of the damage in similar ways and tO similar extents in the mouse (and in other experimental species from which extrapolations are made) and in humans and (iii) at low doses and at low dose-rates of low LET irradiation (the radiation conditions chosen for hazard evaluations), there is a linear relationship between dose and the frequency of genetic effects studied. Other, more specific assumptions and considerations will be discussed in the appropriate sections below.

2.3 Methods

The methods that have been used thus far in quantitative genetic risk assessments can be broadly grouped under two headings: the "doubling dose method" or the "relative mutation risk method", and the "direct methods". The aim of the doubling dose method is to provide an estimate of risks in terms of the additional number of cases of


genetic disorders due to radiation exposures using the spontaneous prevalence of such disorders in the population as a frame of reference. To use this method, one needs to have (i) an estimate of the doubling dose; (ii) information on the spontaneous prevalence of genetic disorders, and (iii) the amount of radiation sustained by the population. The doubling dose is the amount of radiation necessary to double the spontaneous mutation rate and is obtained by dividing the spontaneous mutation rate by the rate of induction per unit dose of radiation. Thus, for instance, if the average spontaneous rate is ml per locus and the average rate of inductionis mz per locus per unit dose of radiation, then the doubling dose c = ml/m2. The reciprocal of the doubling dose, 1/c, is the relative mutation risk per unit dose. It is easy to see that the lower the doubling dose, the higher the relative mutation risk and vice versa.

This method of calculation of the doubling dose is valid for~ mutational events, whether they are dominants or recessives, as well as for chromosomal aberrations; however, the major application ~ of the doubling dose method is for simple, dominantly inherited traits; this is because of the fact that equilibrium frequencies of the responsible mutant genes can be assumed to be directly proportional to the mutation rate (BEIR 1972; Denniston 1982). The above assumption is almost as good for X-linked traits (BEIR 1972). Thus, if the spontaneous prevalence of these disorders is p ~o, and the population is exposed at a rate of x rads per generation, then the expected increase at equilibrium (under conditions of continuous radiation) can be estimated as a produc~ of p, 1/c and x.

An increase in mutation rates of autosomal recessive genes will not lead to a ~corresponding increase in the frequency of recessive disorders because (i) when recessive mutations first arise, they are present in a heterozygous condition and their fate depends strongly on the way selection acts (BEIR 1972) and (ii) recessive mutations have to become homozygous to manifest the disorders and this may take several generations, depending on a number of factors. The application of the doubling dose method for non-numerical chromosome aberrations and to disorders of complex aetiology is somewhat more involved (BEIR 1972, 1980; Crow and Denniston 1981, 1985; Denniston 1982) and will be considered later.

In addition to the method discussed above, a number of so-called "direct methods" have been used over the years for the estimation of genetic risks. The advantage of these methods is that they aim at expressing absolute risks in terms of effects expected in the progeny for the different kinds of genetic damage on the basis of experimental data. However, as will be discussed later, it has not been always possible to satisfactorily bridge the gap between the estimates of rates of induction and the actual effects expected in terms of genetic disease. Furthermore, when one considers the kinds and number of assumptions that have been (and have to be) used, all of thesemethods are no more direct than the doubling dose method.

3. Brief history of the approaches to genetic risk assessments

3.1 Doubling dose method

The general radiation genetic principles that guided the BEAR Committee (1956), the British Medical Research Council Committee (MRC 1956) and the UNSCEAR (1958) in the preparation of their respective reports in the mid- and late 1950s were those that


emerged from extensive work with Drosophila (primarily mature sperm irradiation), ongoing work with experimental mammals (primarily in the mouse) and the few human data. Among these principles, the following deserve mention: (i) mutations, spontaneous or induced, are usually harmful; (ii) any dose of irradiation, however small, entails some genetic risk; (iii) the number of mutations produced is proportional to the dose, so that linear extrapolation from high-dose data provides a valid estimate of the low-dose effects and (iv) the effect is independent of the dose-rate at which radiation is delivered and of the spacing between exposures.

The BEAR Committee (1956) estimated that the doubling dose was probably between 30R and 80 R and that it would be reasonable to use 40 R in computations. It also assumed that about 2 ~ of all liveborn children are, or will be, seriously affected by defects of "simple genetic origin". Under the further assumption that for this fraction of human genetic defects, the incidence is proportional to the mutation rate, the effect at equilibrium after a continuing exposure to the then recommended limit of 10 R per generation was computed. The conclusion was that there would be about 5,000 new instances of "tangible inherited defects" per million births with about one-tenth of this number in the first generation, after the beginning of radiation exposure. The general philosophy, the range and the best estimates of the doubling doses arrived at by the British Medical Research Council Committee (MRC 1956) and by UNSCEAR (1958) were roughly similar.

With increasing reliance on mouse data, attention was focused on the collection of more extensive information after irradiation of spermatogonia and oocytes, the cell stages most at risk from the standpoint of genetic risks. The results that became available (after the publication of the three reports mentioned earlier) demonstrated that (i)chronic gamma irradiation of spermatogonia was mutationally less effective (by a factor of about 3) than the same total dose of high dose-rate X-irradiation (Russell and Russell t958; Russell et al 1958, 1960); (ii) following acute X-irradiation at high doses, the mutation rate in mature and maturing oocytes was higher than in spermatogonia (Russell et a11959) and (iii) the dose-rate effect in females was even more pronounced than in males (Russell 1963).

On the basis of these results, it was suggested (UNSCEAR 1962)that for chronic irradiation of males, the doubling dose is probably about 3-4 times the value of 30 R used in 1958 (UNSCEAR 1958). The possibility was also noted that, for similar radiation conditions, the doubling dose for females could be higher than for males. It was pointed out that a " . . . . permanent doubling of the mutation rate would ultimately double the prevalence of those serious defects determined by the unconditionally harmful gene s which are estimated to affect about 1.0 ~o of those born alive" (UNSCEAR 1962).

By the time the 1966 UNSCEAR report was written, the earlier mouse results had been amply confirmed and extended and besides, new data had been obtained in females which showed that there was a dramatic effect of the interval between irradiation and conception: in the first seven weeks after irradiation, the mutation frequency was high, but subsequently, 11o mutations at all were recovered (Russell 1965). In view of the wealth of the mouse data which by then had accumulated, the UNSCEAR abandoned the doubling dose approach in favour of a more "direct approach", using primarily the mouse specific locus data as a basis.

In 1972, the UNSCEAR revived its interest in the doubling dose approach, but gave it a low profile. Part of the reason for this renewed interest was that Liining and Searle (1971) summarized a number of estimates of doubling doses for different kinds of


genetic damage in the mouse (semi-sterility, specific locus mutations, dominant visibles, mutations affecting the skeleton and autosomal recessive lethals). They all fell within a range of 16-51 rad, averaging about 30 rad for spermatogonia exposed to high acute X- ray doses. On the basis of the dose-rate studies on the induction of specific locus mutations in male mice, it was inferred that the doubling dose for chronic low LET irradiation conditions could be about three times the above value, i.e,, 100 rad. Lfining and Searle (1971) gave no doubling dose estimates for females, since very little information on spontaneous rates ~was available in the open literature.

On the assumption, based on the Northern Ireland survey (Stevenson 1959), that in man about 3.0 5/o of liveborn are affected by deleterious traits maintained by mutation (now including simple dominants, some traits of uncertain genetic origin and chromosomal anomalies), and that the doubling dose is 100 rad, the UNSCEAR (1972) estimated about 300 extra Cases per million births for each rad of low dose-rate low LET irradiation to the males in the parental generation. It argued that " . . . . the great majority of these will be (the result of) gene muiations with an unknown degree of dominance . . . if, however, the range observed in Drosophila (2-5 5/o) is used as an upper limit to the average dominance in man as expressed by the frequency of deleterious traits among liveborn, then 6 to 15 affected individuals per million liveborn would be expected in the first generation following irradiation, the rest of the damage being expressed in subsequent generations".

In 1972, the BEIR Committee also brought out a report (BEIR 1972). In this report, the doubling dose was calculated by using an assumed range for the rate of spontaneous mutations in man (0.5 x 10- 6 to 0.5 x 10- s per gene) and an induction rate from mouse specific locus data (0'25 x 10 - v per locus per rem; average rate taking into account both sexes and assumed to apply to low dose rate low LET irradiation conditions). The range of 20-200 rem thus derived represents a "hybrid" doubling dose range. On the basis of the Northern Ireland survey (Stevenson 1959; this was also the basis for the UNSCEAR figures), the BEIR Committee assumed that the prevalence of genetic disorders in the human population is of the order of 6'0 5/o (including 1.0 ~ dominant and X-linked disorders, 1"0 ~ chromosomal and recessive disorders and 4.0 5/oo congenital anomalies and other disorders of complex aetiology), and proceeded to estimate the effects of 5 rem per generation on a population of one million livebirths.

Further considerations/assumptions used in this exercise were the following: (i) the incidence of dominant and X-linked traits is essentially proportional to the mutation rate and hence their frequency will be increased by the dose times the relative mutation risk per rem; (ii) the incidence of recessive diseases is only very indirectly related to mutation rate; (iii) diseases caused by chromosomal anomalies are not likely to be very much increased by low-level irradiation; (iv) for disorders of complex aetiology, the mutational component (i.e., the proportion of their incidence which is directly proportioual to the mutation rate) is likely to be in the range of 5 to 50~; (v) for dominant and X-linked diseases, the expected increase in the first generation is likely to be of the order of 20 5/o of that at equilibrium (based on the finding that, in the northern Ireland survey, the population incidence was only about 4/5 of that among newborn and this is roughly equivalent to assuming that the average mutant persists in the population for five generations) and (vi) for disorders of complex aetiology, the first generation incidence will be about one-tenth of that at equilibrium.

The estimated effect of 5 rem per generation on a population of one million livebirths was 300 to 7,500 new cases of genetic diseases at equilibrium and about 60 to 1,000 cases

Evaluation of genetic radiation hazards & man 85

in the first generation (relative to the assumed prevalence of 60,000 case s of genetic and partially genetic diseases per million livebirths).

At the time of the preparation of its report, the UNSCEAR (1977) had access to (i) detailed analyses of mouse data obtained after chronic gamma ray exposures and estimates of doubling doses (Sankaranarayanan 1976; Searle 1976), (ii) the results of the continuing studies of mortality rates among children born to survivors of the atomic bombs in Hiroshima and Nagasaki, published by Neel et al (1974), and (iii) new data on the prevalence of Mendelian and chromosomal disorders as well as that of disorders of complex aetiology, from an extensive survey carried out in the Canadian province of British Columbia and published by Trimble and Doughty (1974).

The estimates of the doubling doses for the mouse for different genetic endpoints fell in the range between 80 and 240 rad. Analysis of the Japanese mortality data by Neel and colleagues showed no significant effects of parental exposure on mortality of children through the first 17 years of life, but suggested that the doubling dose for this kind of damage is of the order of 46 rem for the fathers and about 125 rem for mothers, for acute radiation conditions obtained during the bombings. Neel et al (1974) suggested that, on the basis of mouse data, the gametic doubling dose for humans would be expected to be 3 to 4 times the value of 46 rem for males and as much as 1,000 rem for females.

The UNSCEAR examined both the mouse and the Japanese human data mentioned above and concluded that it would be prudent to continue to use the doubling dose of 100 rad for estimating human radiation hazards. It also appraised the British Columbia figures (Trimble and Doughty 1974) taking into account, among other things, the results of the northern Ireland survey, those from several ad hoc surveys for specific dominant conditions and from newborn surveys for chromosomal anomalies and the uncertainties involved in the aetiology of irregularly inherited disorders. The following figures that were arrived at were used for hazard evaluation: 1.0 ~ dominant and X- linked disorders, 0 ' 1~ recessive disorders (excluding those maintained by heterozygous advantage of the relevant genes), 0.4 ~ chromosomal disorders (including sex- chromosomal aneuploids, autosomal trisomies and unbalanced forms of translocations, but excluding mosaics and balanced structural rearrangements)'and 9.0 irregularly inherited disorders, together a total of 10.5 ~.

Using a doubling dose of 100 rad, and a prevalence figure of 10.5 ~o mentioned above, the UNSCEAR estimated that, if the population is continuously exposed to low LET irradiation at a rate of 1 tad per generation, there would be a total of about 185 cases of genetic (and partially genetic) disease per million livebirths at equilibrium of which about one-third would be expressed in the first generation (0.17~ and 0.06~, respectively, of the assumed prevalence of 10"5 ~). These figures were arrived at using the following assumptions: (i) for dominant and X-linked diseases, the first generation incidence will be about one-fifth of that at equilibrium; (ii) for recessive diseases, there will be no perceptible increase; (iii) for chromosomal diseases, all those due to numerical anomalies and 3/5 of those due to unbalanced structural rearrangements will be expressed in the first generation; (iv) the mutation component of irregularly inherited diseases will be about one-tenth of that at equilibrium.

Subsequent to the publication of the 1977 UNSCEAR report, the BEIR Committee produced another report (BEIR 1980). Its summary of risk assessments is reproduced in table 1. It should be noted that (i) the prevalence figures of genetic and partially genetic diseases are essentially the same as those used in the 1977 UNSCEAR report; (ii) the


Table 1. Estimate of genetic effects of an average population exposure of 1 rem per 30-year generation (BEIR 1980)

Type of genetic disorder a

Current incidence

per million liveborn offspring

Effect of 1 rem per generation per million tiveborn offspring

First generation b Equilibrium c

Autosomal dominant, X-linked 10,000~ 5-65 d 40-200 f

Irregularly inherited 90,000J 20-900 e

Recessive 1,100 Very few; effects Very slow increase in heterozygotes accounted tbr in top row

Chromosomal aberrations r 6,000 Fewer than 10g Increase only slightly

alncludes disorders and traits that cause serious handicap at some time during life-span; bestimated directly from measured phenotypic damage or from observed cytogenetic effects; Cestimates by relative mutation risk method; dno first generation estimate available for X-linked disorders; the expectation is that it would be relatively

small (note that a typographical error in the BEIR report is corrected here); %ome estimates have been rounded off to eliminate the impression of considerable precision; fincludes only aberrations expressed as congenital malformations, resulting from unbalanced segregation

products of transtocations and from numerical aberrations; gthe majority of the subcommittee felt that it is considerably closer to zero, but one member felt that it

could be as much as 20.

calculation of risks is based on an assumed doubling dose range of 50 to 250 rem (instead of the 20 to 200 rein used in 1972); (iii) the risks are expressed for a population exposure of 1 rem per generation (instead of the 5 rem/generation used earlier) and (iv) whereas in 1972, the doubling dose method was the preferred one to express risks at equilibrium and in the first generation, in the 1980 report only the equilibrium values were obtained using the doubling dose method (the first generation values given in table 1 were arrived at by using a direct method to be discussed later).

As far as the new range of 50 to 250 rem used in 1980 is concerned, the BEIR Committee has this to say: " . . . . this is based mainly on our best substantiated estimate of the doubling dose namely, 114 R for mouse spermatogonia; we approximately halve and double this to get our range of 50 to 250 reln". Secondly, although a direct method was used to obtain first generation figures, the report states that such figures can also be arrived at using the equilibrium values (estimated using the doubling dose method) and making assumptions similar to those used in 1972 (namely, for dominant and X-linked diseases, it will be one-fifth of that at equilibrium and for irregularly inherited diseases, about one-tenth of that at equilibrium).

Finally, the UNSCEAR (1982) report can be taken to reflect the current views in this area. (There are no major conceptual advances or numerical changes in the estimates of genetic risks in the UNSCEAR report expected to be published in 1986.) The 1982 estimates are summarized in table 2. Two important changes relative to the 1977 report are the following: (i) for dominant and X-linked diseases, the first generation incidence is now assumed to be 15 ~ of that at equilibrium (instead of the 20 ~ assumed in 1977)


Table 2. Estimated effect of I rad per generation of low dose or low dose-rate, low LET irradiation on a population of 1 million liveborn according to the doubling dose method (based on the 1982 UNSCEAR report) (doubling dose: 100 rad).

Current incidence per one million Effect of 1 tad per generation

liveborn Disease classification a offspring b First generation c Equilibrium

Autosomal dominant and 10,000 d 15 100 X-linked diseases

Recessive diseases 2,500 e Slight Slow increase

Chromosomal diseases Structural 400 f 2.4 4 Numerical 3,000g Probably very Probably very

small small Congenital anomalies, anomalies expressed later and constitutional and degenerative diseases

Total

90,000 h 4.5 45 i

105,900 22 149

aFollows that given in the BEIR report (1972), except that chromosomal diseases are divided into those with a structural and those with a numerical basis;

bbased on the results of the British Columbia survey and other studies; Cthe first generation incidence is assumed to be about 15 ~ of tile equilibrium incidence for autosomal

dominant and X-linked diseases, about 60 ~ of the equilibrium incidence tbr structural anomalies and about 10 % of the equilibrium incidence for diseases of complex inheritance;

dincludes diseases with both early and late onset; ealso includes diseases maintained by heterozygous advantage; rbased on the pooled results of surveys for chromosomal anomalies in newborn children but excluding

euploid structural rearrangements, Robertsonian translocations and "others" (mainly mosaics); gexcluding mosaics; hincludes an unknown proportion of numerical (other than Down syndrome) and structural chromosome

anomalies; ibased on the assumption of a 5 % mutational component.

and this revision is based on the calculations of Childs (1981), a n d (ii) diseases due to chromosomal anomalies are split up into those due to numerical and to structural ones; it is assumed now that the increase due to the induct ion of numerical anomalies will probably be quite small and the first generation incidence of those due to structural anomalies will be about 3/5 of that at equilibrium. Consequent ly , the expected total increases at equi l ibr ium (149 cases per mil l ion livebirths) and in the first generation (22 cases per mil l ion livebirths) are somewhat lower than the corresponding values

estimated in 1977 (185 and 63 cases, respectively). It is worth poin t ing out that the UNSCEAR, during the prepara t ion of its 1982 report

had at its disposal the papers of Neel et al (1982) and of Schull et al (1980, 1982). In these papers, the authors presented an analysis of all the available genetic data obtained in the

cont inuing studies of the Hiroshima and Nagasaki popula t ions (a follow-up of the material presented by Neel et al 1974). These data pertain to un toward pregnancy outcomes (i.e., those resulting in children with major congenital defects, still-births and


deaths in the neonatal period), survival through childhood, incidence of sex- chromosomal anomalies and incidence of biochemical variants (erythrocyte and plasma protein variants analyzed using one-dimensional electrophoresis). In all the calculations, an aBE (relative biological efficiency) of 5 for neutrons has been assumed. As Schull et al (1982) point out, " . . . . in no instance is there a statistically significant effect of parental exposure . . , but for all indicators, the observed effect is in the direction suggested by the hypothesis that genetic damage resulted from the exposure".

Doubling dose estimates were made only for the first three of the indicator traits since the data on biochemical variants were considered too preliminary. The gametic doubling dose estimates presented by Schull et al (1982) were the following: untoward pregnancy outcomes, (60___ 93 rem); survival through childhood (F1 mortality), 135 + 388 rein) and sex-chromosomal aneuploids (535 + 2416 rein). The weighted average of these estimates is (139 _+ 156 rem). The authors considered that this estimate should be multiplied by a factor of 3 to convert it into a relevant estimate for low doses and low dose-rates.

The main message from these papers is 'that the doubling dose for genetic effects in man may be of the order of about 400 rad, i.e., about four times the value used by UNSCEAR, which means that the relative risks estimated by UNSCEAR are higher by a factor of 4. The UNSCEAR examined these data and the analysis presented and concluded that in view of the lack of statistically significant effects and the high standard deviations associated with the doubling dose estimates, it would be premature at present to use these new results for genetic risk assessments. Therefore, the earlier doubling dose estimate (based entirely on mouse data) of 100 rad was retained.

3.2 Doubling dose method in retrospect

The rationale for the continued use of the doubling dose method stems from the fact that it permits one to express risks in tangible terms and that whole classes of genetic effects can be handled as a unit in the absence of information about, for instance, the number of loci involved or their individual mutation rates. The early estimates of doubling doses fell in the range of 10 to 100 rad (then for acute, high dose-rate irradiation) and the possible representative values, between 30 and 40 rad. With the discovery of the dose-rate effects in the mouse, in the early 1960s, the UNSCEAR adopted 100 rad as the best estimate and this value is still being used, although (i) individual estimates for different kinds of genetic damage vary over a range from about 80 to 240 rad in the mouse, and (ii) analysis of the Hiroshima and Nagasaki data continue to suggest that the value may be higher than 100 rad. The principal reasons for adhering to the 100rad estimate are that the human evidence is still too far from conclusive to warran(revision and that caution and prudence constitute the main guiding principle in this endeavour. Furthermore, even with a 100 rad doubling dose, the risk estimates are quite low, relative to the naturally-occurring burden of genetic and partially genetic disorders. The changes in the estimates of relative risks, from the mid-1950s to the present, thus reflect greatly the increasing knowledge on the prevalence of genetic and partially genetic disorders in the human population and the evolving assumptions concerning their possible responses to an increase in mutation rate and are not due to any real breakthrough in our understanding of the genetic sensitivity of human germ cells to ionizing radiation.

Evaluation of 9enetic radiation hazards in man 89

3.3 The direct methods

Some of the basic radiation genetic principles that guided the BEAR Committee during the preparation of its 1956 report were briefly alluded to earlier. In addition, advances in radiation genetics, theoretical and experimental population genetics had already documented the thesis that the changes due to mutated genes are seldom fully expressed in the first generation progeny of irradiated individuals and that these mutant genes would persist in the population for shorter or longer periods of time, depending on their deleterious effects, until they are eventually eliminated from the population. The concept of "genetic death" enunciated by Muller (1950) was current and influenced much of the thinking of geneticists. It seemed logical therefore to apply the concept to the estimation of genetic risks. This viewpoint was succinctly stated by BEAR Committee (1956): " . . . One way of thinking about this problem of genetic damage is to assume that all kinds of mutations on the average produced equivalent damage, whether as a drastic effect on one individual who leaves no descendants because of this damage, or a wider effect on many. Under this view, the total damage is measured by the number of mutations induced by a given increase in radiation, this number is to be multiplied in one's mind by the average damage from a typical mutation".

It was therefore thought that the "total risk" due to induced mutations could be obtained as the product of three factors (i.e., number of genes at which mutations can occur x radiation dose x rate per gene per unit dose) and the expression of this "risk" in the first and succeeding generations estimated by population genetic methods. Some limited mouse data were available on mutation rates in the mouse, but the main difficulty centred around the problem of having reliable gene numbers in man. To circumvent this problem, the BEAR Committee used Drosophila data, dividing the total mutation rate (for recessive lethals per gamete) by that for individual genes, multiplied this ratio by 2 to 3 to allow for mutations with less than lethal effects and arrived at a figure of about 104; this figure was used along with the mutation rate inferred from mouse studies to estimate the "total number of mutant genes that would enter the population in the next generation if everyone in the United States received a dose of 10 R to the reproductive glands" (= 5 x 106 mutant genes) (BEAR 1956). It is thus clear that this estimate was for a "hypothetical organism whose mutation rate per gene is that for the mouse and whose gene number is that of Drosophila" (see BEIR 1972 for a discussion). The Committee, however, concluded that " . . . this kind of estimate is not a meaningful one to certain geneticists. . , their principal reservation is doubtless a feeling that, hard as it is to estimate the number of mutants, it is much harder still, at the present state of knowledge, to translate this over into a recognizable statement of harm to individual persons" (BEAR 1956). This form of"gene number method" was therefore subsequently abandoned (BEIR 1972, 1980).

In its 1966 report, the UNSCEAR used a variation of the "gene number method". It was argued that (i) if it is assumed that the average rate of induction of recessive mutations in the mouse (estimated as 1 x 10-7 per locus per R from data obtained for specific locus mutations, following high dose, high dose-rate X-irradiation of spermatogonia) is applicable to man, and (ii) if, for the purpose of computations, the total number of gene loci in man can be assumed to be 20,000 (range: 7,000 to 70,000), then, the total "risk" from the induction of point mutations will be 2 x 10- 3 mutations per gamete per R). This estimate was higher than the One arrived at (0.5 x 10- 3 mutations per gamete per R) using the limited data on the induction of autosomal recessive lethals in mouse


spermatogonia as a basis, although not significantly so. The "expression" of this risk in the first and succeeding generations was then estimated using the average degree of semi-dominance of recessive lethal mutations in Drosophila.

For estimating more directly the "risk" from the induction of dominant mutations, the Committee assumed that the rate is likely to be in the range from 10 - 9 to 10 - 7 per locus per R and the number of loci determining dominant disorders in man in the range of 50 to 500. The total "risk" was thus estimated to be in the range from 5 x 10 -8 to 5 x 10- s mutations per gamete per R. Similar estimates of risk from the induction of chromosomal aberrations (translocations and deletions) were arrived at using the limited data on the induction of heritable semi-sterility in the mouse and those on the induction of dicentrics and deletions in human peripheral blood lymphocytes. It was pointed out that all the risks would be lower for chronic irradiation conditions.

In its 1972 report, the UNSCEAR went about the problem of estimating gene numbers in the following way. In the mouse, there are about 20 functional units per cross-over unit (Russell 1971) and there are some.l,250 map units in the entire genome (Green et al 1970). This gives about 25,000 functional units in the mouse genome. The DNA contents of the mouse and human diploid genomes were estimated to be 4"7 x 10 9 and 5.6 x 10 9

nucleotide pairs, respectively. Thus, the estimate of the total number of functional units in the human genome was 25,000 x (5.6/4.7) = 30,000. By assuming that the rate of induction of specific locus mutations in the mouse spermatogonia under conditions of chronic gamma ray exposures (0.5 x 10-7 per locus per rad; average estimate based on 12 loci studied) was applicable to man, the total "risk" from the induction of point mutations in man was estimated as 1,500 mutations per rad per million gametes (i.e., 0'5 x 10-7 x 30,000). This was considered an overestimate because specific locus mutations may involve more than one functional unit.

The estimate of total risk based on the new mouse data on the induction of autosomal recessive lethals (spermatogonial irradiation) after similar correction for DNA content, gave a figure of 36 mutations per rad per million gametes, and this was considered a possible underestimate of the risk. The first generation expression of these risks was computed, using, as before, the degree of semi-dominance of recessive lethals (2-5 ~) inferred from Drosophila studies in conjunction with the total rate of 1,500 or 36 mutations per rad per million gametes.

A modified version of the gene number approach was also used in the 1972 UNSCEAR report for estimating the rate of induction of dominant mutations in man. For doing this, the rate of induction of dominant visible mutations in mouse spermatogonia (4"96 x 10 -7 per rad per gamete) was used. Since this rate was for acute high dose-rate X-

irradiation, this was divided by 3 to correct for effects at low dose-rates, and divided by 75 (the then presumed number ofloci in the mouse which mutated to dominant visibles) to get a rate of 2.2 x 10- 9 per locus per gamete per tad. The Compendium of McKusick (1971) listed 415 autosomal dominant loci in man plus another 528 for which the evidence was inconclusive. Assuming that there was "good reason to predict that the number will not be less than 1,000 based on progress of research in this area", the Committee multiplied 2.2 x 10 - 9 by 1,000 and obtained an estimate of about 2 dominant mutations per rad per million gametes. (Parenthetically, it may be mentioned that if one were to attempt such a calculation now, the "risk" will be only slightly different: in the mouse, autosomal dominant and codominant mutations have now been recovered at 194 loci (data o f T H Roderick, cited in McKusick 1983) and a total of 934 confirmed (+ 893 presumed) loci are known in man at which autosomal dominant


mutations have been identified (McKusick 1985); the estimate of risk of induction of dominant mutations in man would therefore be 9-17 per rad per ten million gametes).

In its 1972 report, the UNSCEAR also provided an estimate of risk from the induction of reciprocal translocations in man, using a "direct method". This was done by assuming that (i) the rate of induction of balanced reciprocal translocations in man would be twice that in the mouse germ cells [i.e., 2 x 0'3 x 10 . 4 per gamete per rad; based on the arm number hypothesis of Brewen et al (1973) which was then (but not now) considered valid]; (ii) at low doses or dose-rates, the rate in males will be reduced by factors of 4 or 7, respectively; (iii) the ratio of balanced to unbalanced translocation products will be 1:2 and (iv) only about 6.0 ~ of the unbalanced products will result in liveborn children with multiple congenital anomalies. Taking all these into account, it was estimated that the risk was 1-2 congenitally malformed progeny per rad per million liveborn, following irradiation of males. It is worth mentioning that in its report (BEIR

1972), the BEIR Committee also provided an estimate of risk from the induction of reciprocal translocations; using the same basic data and a different set of assurnptions, they arrived at an estimate which is about an order of magnitude higher than that of UNSCEAR.

3.4 "Direct methods" used until 1972, in retrospect

Despite gallant efforts to weld the principles of population genetics, the concept of "genetic death" and radiation genetics to provide "risk" estimates for the induction of mutations, the vision projected by the gene number methods until 1972 ran into considerable problems. Apart from difficulties of reliably estimating gene numbers, the conceptual difficulties involved in bridging the gap betweeen the dynamics of mutant genes in the population, on the one hand, and genetic disease, on the other, were so many that the method fell short of expectation and slowly faded into oblivion. Furthermore, the degree of semi-dominance of recessive mutant genes used to extract "first generation effects" from the "total risks" refers in reality to the effects of these genes on biological (reproductive) fitness and not to genetic diseases as such.

As far as the "direct methods" used for predicting risks from the induction of reciprocal translocations in human germ cells are concerned, they still retain their usefulness; the nature of the data base has improved, although a number of assumptions have still to be made, as will be discussed in the next section.

3.5 "Direct methods" used in the 1977 and 1982 UNSCEAR reports and in the 1980 BEIR report for predicting first generation effects

In the UNSCEAR (1977) report, a major conceptual change was introduced. At the time of writing the above report, the Committee had as its disposal the earlier data collected by Ehling (1965, 1966) on the induction of dominant mutations affecting the skeleton of the progeny of irradiated male mice and new data on these collected by Selby and Selby (1977). The data of the latter authors established that these "skeletal mutations" are in fact transmissible and also showed that these mutations had incomplete penetrance and variable expressivity for many or all of the phenotypic effects they caused and besides, some of them behaved as recessive lethals when made homozygous. Doubtless, some of


the properties of these mutations are similar to those of rare dominants and rare irregularly inherited dominants in man.

However, to convert the rate of induction of mutations causing skeletal abnormalities in the mouse into an overall rate for all mutations with dominant phenotypic effects in man, one needs information on (i) the proportion of dominant conditions in man whose main effect is in the skeleton, and (ii) the proportion of skeletal abnormalities studied in the mouse which, at the human level, is likely to cause a serious handicap. After careful deliberations, the UNSCEAR arrived at values of 10 ~o for item (i) [taking into account the proportion of clinically relevant autosomal dominants in man whose main effect is in the skeleton, which was assessed at 20 ~, and the ease of diagnosis of skeletal defects (possibly higher by a factor of 2) relative to those affecting other bodily systems]. The proportion Of skeletal anomalies in the mouse that might cause serious handicaps, should they occur in man, was assessed as one-half.

In risk estimation, the frequency of skeletal mutations observed after high fractionated (or high acute) X-ray doses was corrected by suitable factors to arrive at a rate that would be applicable for low dose rate, low LET irradiation conditions. The resultant figure of 4 x 10 .6 was multiplied by 10 and divided by 2. The figure thus arrived at, namely 20 x 10- 6, is one of induction of mutations causing dominant effects on any of the bodily systems in man. In other words, tbllowing 1 rad of paternal (spermatogonial) irradiation, 20 per million progeny would carry mutations which cause one or another kind of dominant genetic disease, in the first generation.

In its 1980 report, the BEIR Committee, however, gave a different estimate (5-65 cases per million per rein; see table 1) starting from the same data base as that used by UNSCEAR. The reasons for this difference are the following: (i) to convert the rate of induction of skeletal mutations in mice to an overall rate involving all bodily systems in humans, the UNSCEAR used a multiplication factor of 10 as mentioned in the preceding paragraph; this was divided by a factor of 2 to exclude mutations whose effects are slight; the BEIR

Committee, however, used multiplication factors of 5-15 to make the first conversion and a range of 0.25 to 0'75 to make the second conversion. These operations gave the range of 5-45 cases per million births. (ii) To take into account the effects of irradiation of females, the BEIR Committee multiplied the upper limit by 1.44 (the 1977 UNSCEAR report assumed that the risk for irradiated human females is negligible and did not give any quantitative estimate). The rationale tbr the multiplication by 1.44 was set forth as follows: " . . . The mutational response of the resting oocytes in mice is negligible~ compared with that of spermatogonia, and mature and maturing oocytes in mice have a mutation rate no greater than 0.44 times that found in spermatogonia. We do not know which of the two classes of oocytes would have a mutational response more similar to that of the arrested oocytes in women. To incorporate this range of uncertainty into our risk estimate for the combined effects of irradiation of both sexes, we have simply kept the lower limit of our estimate the same as it was (assuming a negligible mutation frequency i n resting oocytes) and multiplied the upper limit by 1.44 (assuming the maximal estimate of the mutation frequency in mature and maturing oocytes). This gives an estimate of 5-65 induced serious disorders per million liveborn as the first generation express ion . . . ".

For estimating the risk from the induction of reciprocal translocations, the UNSCEAR (1977) used the marmoset and human cytogenetic data (Brewen et a11975) as the basis and arrived at an estimate of between 2 and 10 congenitally malformed children per million births per tad of low LET irradiation of males. Briefly, the calculations were the

Evaluation o f genetic radiation hazards in man 93

fol lowing: 7 x 10 . 4 (rate o f t r ans loca t ion induc t ion ; spe rma tocy te data) x 0-25 (mult i -

p l i ca t ion factor to get the rate for her i tab le t r ans loca t ions in the p rogeny) x 0"1 or 0"5 (mul t ip l ica t ion factors to a cco u n t for dose-ra te effect af ter ch ron ic g a m m a a n d low-

dose X-rays, respectively) x 2 (mul t ip l i ca t ion factor to es t imate the rate o f p r o d u c t i o n o f u n b a l a n c e d p roduc t s o f rec iproca ! t r ans loca t ions ) x 6 ~o (the p r o p o r t i o n o f un -

ba l anced p roduc t s tha t was a s s u me d to give rise to viable, b u t congen i ta l ly m a l f o r m e d progeny). The risk for the i r r ad ia t ion o f h u m a n females was cons ide red to be small, b u t no quan t i t a t ive es t imate was given.

F o r the BEIR C o m m i t t e e ' s ca lcu la t ions o f risks for i r r ad ia t ion o f males, the s t a r t ing po in t was the same cy togene t ic da t a a s those used by UNSCEAR. The es t imate o f risks (see table 1) was also near ly the same, b u t the m e t h o d o f ca lcu la t ion a n d the a s s u m p t i o n s used were s o m e w h a t different. F u r t h e r m o r e , it a s sum ed tha t the r isk for i r rad ia ted h u m a n females will be the same, as it d id in its 1972 report .

The mos t recent pub l i shed es t imates o f risk are those arr ived a t in the UNSCEAR (!982) r epo r t ( table 3). These are basical ly the same as those arr ived at in the 1977 report , except for the fo l lowing add i t ions /changes : (i) an add i t i ona l and i n d e p e n d e n t es t imate o f r isk f rom the i n d u c t i o n o f m u t a t i o n s hav ing d o m i n a n t effects in the F1 p r o g e n y has been presented , us ing the new da ta o n the i n d u c t i o n o f d o m i n a n t ca tarac t m u t a t i o n s in male mice fo l lowing s p e r m a t o g o n i a l i r rad ia t ion ; (ii) the p r o b a b l e m a g n i t u d e o f risk for i r r ad ia t ion o f females has been arr ived at very indirect ly, by assuming , o n the basis o f specific locus data, tha t the rate in females is l ikely to be close

Table 3. Risk of induction of genetic damage in humans per rad of low dose-rate low LET irradiation according to the direct method (based on the 1982 UNSCEAR report).

Risk from the induction of

Expected frequency (per 106) of genetically abnormal children in the first generation

after irradiation of

Males Females

Mutations having dominant efl'ects a 10-20 b

Unbalanced products of reciprocal translocations 0.3-10 d

0 - - 9 c

0 - 3 e

alncludes the risk from the induction of mutations, as well as recessive mutations, deletions and balanced reciprocal transtocations with dominant effects;

bThe lower limit of about 10 is derived from the data on cataract mutations and the upper limit of 20 from data on skeletal mutations and is the same as that arrived at in UNSCEAR (t977);

Cthe lower limit of zero is based on the assumption that the mutational sensitivity of human immature oocytes i~ similar to that of mouse immature oocytes. The upper limit is based on the assumption that the sensitivity of human oocytes is similar to that of mature and maturing mouse oocytes and that the latter is at best, 0-44 times that of spermatogonia;

dthe lower limit of 0-3 is based on rhesus monkey cytogenetic datal the upper limit is based on combined marmoset and human cytogenetic data;

ethe lower limit of zero is based on the assumption that the sensitivity of the human immature oocytes to the induction of heritable translocations will be similar to that of the mouse with respect to the induction of other types of chromosome aberrations; the upper limit of 3 is based on the assumptions that the sensitivity of the human immature oocytes to the induction of translocations will be one-half of that of the human and marmoset spermatogonia (based on results with mice on heritable translocations), that the frequency of unbalanced products will be six times that of recoverable balanced reciprocal translocations and that 6 % of the unbalanced products will result in congenitally malformed children. ' |


to zero (o11 the assumption that the mutational sensitivity of the human immature oocytes may be similar to those of the mouse immature oocytes) or, not more than 44 of the spermatogonial rate [on the assumption derived from the finding that the higher ratio of the oocyte (mature and maturing) to spermatogonial rate is no more than 0.44 to 13 (see Russell 1977); and (iii) for estimating risks from the induction of reciprocal transtocations, all the available primate data (i.e., the marmoset,the rhesus monkey and the limited human data) have been used.

The mouse cataract mutation data were obtained in experiments involving high acute or fractionated gamma-ray exposures (Ehling 1980; Ehling et al 1982; Kratochvilova and Ehling 1979). From these results, a rate applicable to chronic gamma-ray exposures was derived by using empirical correction factors derived from specific locus studies. This was about 2.6 x 10- 7 per gamete per rad: In man, about 2.7 ~o of all known and proven dominant mutations are associated with one or another form of cataract. The reciprocal of this (i.e., 100/2.7 = 36.8) was used to multiply the rate of 2.6 x 10 .7 to arrive at approximately 10 x 10 .6 per rad. In other words, about 10 individuals per million born will be affected by one or another kind of serious, clinically important genetic disease per rad of low dose rate, low LET irradiation. It is worth pointing out that in these calculations, the multiplication factor of 2 and the division factor of 2 used in similar computations with the skeletal mutations to take into account ease of diagnosis and severity of effects, respectively, have not been employed. (In retrospect though, it would appear that the multiplication factor of 2 could have been used, as after all, the ease of detection of a cataract in the eye, just like the ease of detection of a skeletal anomaly, is certainly higher than that of defects involving many other bodily systems!).

The extensive cytogenetic data on translocation induction in mate rhesus monkeys show that the rate of induction, at least for high dose-rate X-irradiation conditions, is much lower than that in marmosets (0.86 x 10 .4 per rad per gamete versus 7.7 x 10 -4 per rad; see Van Buul 1980). Since it is not known whether the human spermatogonial sensitivity is similar to that of the marmoset or to that of rhesus monkey, the rates for both these species were used, one to define a probable upper limit and the other the lower limit of risk. The correction factors used to convert the rhesus monkey acute rate into the one applicable to human radiation conditions are the same as those used in the 1977 report. Other correction factors and the underlying assumptions are explained in the footnotes to table 3.

4. Current status, uncertainties and perspectives

4.1 Current status

As of now, there does not appear to be any compelling reason to believe that major revisions of the UNSCEAR (1982) genetic risk estimates are warranted. However, this view is not meant to imply that there have been no advances in our knowledge in areas relevant for genetic risk evaluation in recent years; rather, it is an admission of the fact that the impact of these advances has not been, as yet, fully assessed as will be illustrated below.


4.2 77w prevalence of Mendelian, chromosomal and other disorders

4.2a Mendelian disorders: There is no need to belabour the point that a sound knowledge of the contribution of gene mutations and of chromosomal aberrations tO genetic diseases is of paramount importance not only for gaining a perspective of these diseases to which man is literally heir, but also to provide a frame of reference to appraise the increases expected as a result of radiation (or other mutagenic) exposures. Our current estimates of the prevalence of Mendelian disorders (autosomal dominants and X-linked recessives, together 1.0~, and autosomal recessives, excluding those maintained by heterozygous advantage of the relevant genes, about 0.1 ~) are based on a total of less than about 50 disorders, the individual birth frequencies of which range from about 1/10 s to about 1/104 with some upward adjustment (in the total frequency) for disorders yet to be discovered. For obvious reasons, these estimates pertain to a very small proportion of those listed in the recent update of McKusick's (1983) Compendium (i.e., a total of 1739 conditions which are well-documented as being inherited in a Mendelian manner and a further 1886 for which such evidence is incomplete; McKusick 1985). The disparity is even greater when one considers the fact that the total amount of nuclear DNA in the human haploid genome is about 3 x 106 kb, which in principle can accommodate well over a million genes (on the assumption that an average gene is about 1 to 2 kb long). However, one should hasten to add that this method of estimating gene numbers is probably invalid, since it neglects introns, pseudogenes, highly repetitive sequences such as satellite DNA and other interspersed and non-transcribed DNA sequences (Britten and Kohne 1968; Jelinek and Schmid 1982; Evans 1984; Emery 1984; Dutta 1984; Vogel 1984). Judging from the phenomenal progress in the understanding of the human genome that has been made possible in recent years by the application of recombinant DNA methods, one can optimistically look tbrward to (i) the bridging of the gaps between genes, mutations and their effects on health and disease, and (ii) the application of the knowledge so derived in the context of genetic risk evaluation as well.

4.2b Chromosomaldisorders: At the chromosomal front, the application of banding techniques to the study of human chromosomes has led to the identification of a variety of chromosomal defects, particularly deletions and duplications involving almost every chromosome of the human complement (e.g., Lewandowski and Yunis 1977; Sanchez and Yunis 1977). Most of the currently available information on these partial monosomies and trisomies, however, is in the nature of case reports and, consequently, their prevalences and their contribution to disease states cannot be readily determined.

One exciting development in human cytogenetics in recent years is that pertaining to the discovery of fragile sites on chromosomes which can be made visible by appropriate culture conditions (reviewed by Opitz and Sutherland 1984; Sutherland 1983). In particular, the fragile site on the X-chromosome (at position Xq27) has elicited considerable attention because it is associated with one form of X-linked mental retardation. The prevalence of fragile-X-associated mental retardation has been estimated to be of the order of about 4 x 10- 4 both in males and females (Sherman et al 1984); this would make the fragile-X the most common chromosomal abnormality associated with mental retardation, next only to the Down syndrome. It is currently not known whether this abnormality can be induced by radiation and therefore genetic risk assessments are not possible.


4.2c Congenital anomalies and other irregularly inherited disorders: Our current estimates of the prevalence of' disorders of complbx aetiology-congenital anomalies (4.3 ~) and other irregularly inherited disorders (4.7 ~ ) - a r e based on the results of the British Columbia survey (Trimble and Doughty 1974) in which the follow-up period was from birth to age 21. Recent studies of Czeizel and Sankaranarayanan (1984, and unpublished) lend credence to the view that the above prevalence figures are underestimates. Their data show that in Hungary (i) the congenital anomalies, have a birth prevalence of about 6.0 70 and (ii) the population prevalence (i.e., including all age groups) of the other irregularly inherited disorders b is at least 60.0 70. The inheritance patterns of most of these do not neatly fit the Mendelian mould and are complex (Carter 1976; Czeizel and Tusnady 1984; Smith 1975). Twin and family studies indicate that their aetiology is muttifactorial, depending on polygenic genetic predisposition and environmental factors that may also be multiple. For most of these, not much is known of the mechanisms by which .the genetic predisposition acts or of the environmental factors involved. Nonetheless, it is clear that most of the genetic ill- health in human populations is due to these, and not due to Mendelian and chromosomal disorders. From the standpoint of genetic risk of radiation, the question of whether and to what extent their prevatences will increase as a result of radiation exposures, remains largely 'a matter of conjecture. Furthermore, the usefulness of the method suggested by Crow and Denniston (1981) to estimate the mutational component from heritability estimates for these disorders has not yet been explored.

4.3 Mutational mechanisms

A major growing point of contemporary genetic research is that associated with studies on movable genetic elements, the conceptual framework for which was laid by McClintock over three decades ago with her work on "controlling elements" in maize (McClintock 1948, 1950, 1952, 1955). Now, movable genetic elements have been discovered in a number ofprokaryotes and eukaryotes and some of these elements have been characterized at the molecular level. The rapidly accumulating corpus of evidence documents the fact that they play an important role in the genesis of spontaneous mutations and of chromosomal aberrations and this in turn, alters our concepts about mutational mechanisms and the stability of the genome (see Shapiro 1983). The view that certain repetitive sequences in the human genome such as the "Alu" sequences may

"Congenital anomalies listed in Chapter XIV, entries 740-759 of the International Classification of Diseases (WHO 1977), include central nervous system anomalies, eye anomalies, anomalies of ear, lhce and neck, cardiovascular anomalies, anomalies of the respiratory system, cleft lip d: cleft palate, anomalies of the alimentary system, anomalies of the genital and urinary system, anomalies of the musculoskeletal system, anomalies of the skeletal system, chromosomal anomalies and multiple congenital anomalies.

bIncludes the so-called common familial disorders (about 25 entities) such as thyrotoxicosis, diabetes, gout, schizophrenia, affective psychoses, multiple sclerosis, epilepsy, glaucoma, essential hypertension, acute and sub-acute myocardial infarction, varicose veins of lower extremities, allergic rhinitis, bronchial asthma, ulcers, psoriasis and related disorders, systemic lt, pus erythematosus, rheumatoid arthritis, ankylosing spondylitis, etc.), each of which has a population prevalence of at least 1/104; excludes presenile and senile dementias, refractory errors of the eye, mental retardation, blindness and deaf-mutism, tuberculosis, all cancers, etc. These disorders are not manifest at birth but have variable age of onset, many manifesting late in life.


behave as movable genetic elements has been entertained and discussed in the literature (e.g., Jelinek and Schmid 1982; Schmid and Paulson 1984), although there is no evidence as yet that these sequences move from one genomic location to another or that they are associated with detectable phenotypic effects.

I f a majority of spontaneous mutations in man arise as a consequence of the mobility of the Alu and similar repetitive sequences, and if their mobility is not altered significantly by exposure to radiation or other mutagens (as the limited evidence from other experimental systems would lead one to believe), then the use of the doubling dose method of genetic risk evaluation becomes questionable. I have recently addressed this problem (Sankaranarayanan 1986) and have concluded that (i) it is unlikely that a major fraction of spontaneous mutations in man that lead to disease states is due to mobile genetic elements, and (ii) the use of the doubling dose method for the evaluation of genetic radiation hazards in man would appear to be valid at the present time.

4.4 Nature of spontaneous and radiation-induced mutations

The methods of recombinant DNA technology have enabled the direct detection of the molecular heterogeneity of spontaneously arising mutational events that lead to disease states in man, as will be evident from the spectacular progress in studies of globin (Forget et a11983; Orkin and Kazazian 1984), HPRT (Wilson et a11983; Yang et a11984; Albertini et al 1985; Turner et al 1985), haemophilia B (Rees et al 1985) and gene loci among others (reviewed in Caskey and White 1983). Similar studies are being carried out to analyse the nature of radiation-induced mutations in mammalian somatic cells (e.g., Vrieling et al i985; Thacker 1986). It is to be expected that these advances will add more precision both in the formulation of questions and the search for answers that are relevant for genetic risk assessments.

4.5 Fragile sites and spontaneous chromosome breakage

There are suggestive indications for the thesis that certain autosomal fragile sites may . predispose to chromosome breakage. Hecht and Hecht (1984a, b) have adduced some

evidence in this regard through analysis of the information on the location of hreakpoints in chromosomal anomalies (deletions, duplications, inversions and non- Robertsonian translocations) found in amniocentesis studies and in those on spontaneous abortions, stillbirths and newborns. They, however, caution that in view of the heterogeneity of the database from which the pertinent information was extracted, more evidence is required to provide convincing proof that fragile sites are regions in the chromosome predisposed to breakage in meiosis. The question of whether those fragile sites may predispose the chromosomes to radiation-induced breakage events is as yet unknown.

4.6 Are cancer induction studies the province of geneticists or of cancer biologists?

The remarkable insights into oncogenic transformation that have emerged in recent years through the convergence of thought and techniques from viral oncology, cell


genetics and molecular biology amply testify to the fact that cancer studies have become an exciting area for both geneticists as well as cancer biologists. There isnow extensive evidence that mammalian genomes harbour nncleotide sequences related to the retroviral oncogenes as part of their normal genetic make-up and that their activation (either spontaneously or through mutagenic exposures) can initiate cancers (Bishop 1982, 1983; Cooper 1982, 1984; Sukumar et a11983; Batmain and Pragnel11983; Guerra

et a11984; Croce and Klein 1985). There is thus no a priori reasonto assume that germ cells will be immune in this respect, especially after exposure to mutagens. In fact, Nomura (1982, 1983; 1985) has demonstrated high tumour yields in the progeny of irradiated mice and thus has highlighted the possibility that genetic changes induced in germ cells of parents can cause tumours in the progeny. It would therefore appear worthwhile to independently confirm these results in other experimental systems and subsequently extend them to man, by looking for cancers in the progeny of those who have survived after radio- and/or chemotherapy. Whether this component of "genetic risks" would be negligible, small or significant, is difficult to state at present.

4.7 Other work in progress

Genetic studies on the Hiroshima and Nagasaki populations are continuing and will remain a major source of direct human data. An international collaborative programme on genetic effects in the offspring of cancer patients exposed to physical and chemical agents for therapeutic purposes is in the process of being initiated (Oftedal and Mulvihill 1985). This is a highly commendable enterprise and it is hoped that this study will provide information on hazards attributable to germ celt mutations and will undoubtedly have practical importance for clinicians who must advise cancer survivors or current cancer patients who have initiated or are considering pregnancy.

5. Conclusions

The foregoing examples illustrate.the fact that a number of advances have taken place during the last several years, although as stated earlier, their impact on genetic risk assessments'has not been worked out. It is also becoming increasingly clear that radiation geneticists cannot effectively work in isolation from molecular biologists, tumourologists and epidemiologists. I am certain that when it becomes possible to assess the impact of the advances on genetic risk assessments, the latter field will enter a new phase of development marked by greater precision than has hitherto been possible.

Acknowledgments

The author is grateful to Prof. F H Sobels for his warm encouragement. The writing of this paper was supported by Euratom contract No. BIO-E-406-81-NL with the University of Leiden.

E v a l u a t i o n o f gene t i c rad ia t ion h a z a r d s in m a n 99

R e f e r e n c e s

Albertini R J, O'Neill J P, Nicklas J N, Heintz N H and Kelleher P C 1985 Alteration of the hprt gene in human in vivo derived 6-thioguanine resistant T lymphocytes. Nature (London) 316: 369-37!

Balmain A and Pragnell I B 1983 Mouse skin carcinomas induced in vivo by chemical carcinogens have a transforming Harvey-rat oncogene. Nature (London) 303:72-74

BEAR Report 1956 The biological effects of atomic radiation. Natl. Acad Sci., Natl. Res. Council, Washington, D.C.

BEIR Report 1972 The effects on populations of exposure to low levels of ionizing radiation. Natl. Acad. Sci., Natl. Res. Council, Washington, D.C.

BEIR Report 1980 The effects on populations of exposure to low levels of ionizing radiation. Natl. Acad. Sci., Natl. Res. Council, Washington, D.C.

Bishop M J 1982 Oncogenes. Sci. Am. 246:68-78 Bishop M J 1983 Cellular oncogenes and retroviruses. Annu. Rev. Biochem. 52:301-354 Brewen J G, Preston R J and Gengozian N 1975 Analysis of X-ray-induced chromosomal translocations in

human and marmoset spermatogonial cells. Nature (London) 253:468-470 Brewen J G, Preston R J, Jones K P and Gosslee D G 1973 Genetic hazards of ionizing radiation: cytogenetic

extrapolation from mouse to man. Mutat. Res. 17:245-254 Britten R J and Kohne D E 1968 Repeated sequence in DNA. Science 161:529-540

Carter C O 1976 Genetics of common single malformations. Br.. Med. Bull. 32:21-216 Caskee C T and White R L (eds) 1983 Recombinant DNA applications to human disease. Banbury Report 14,

Cold Spring Harbor Laboratory Childs J D 1981 The effect of a change in mutation rate on the incidence of dominant and X-linked disorders

in man. Murat. Res. 83:145-158 Cooper G M 1982 Cellular transforming genes. Science 218:801-806 Cooper G M 1984 Activation of transforming genes in neoplasms. Br. J. Cancer 50:137-142 Croce G M and Klein G 1985 Chromosome translocation and human cancer. Sci. Am. 252:44-50 Crow J F and Denniston C 1981 The mutation component of genetic damage. Science 212:888-893 Crow J F and Denniston C 1985 Mutation in human populations. In Advances in human genetics (eds)

H Harris and K Hirschhorn (New York: Academic Press) vol. 14, pp. 59-123 Czeizel A and Sankaranarayanan K 1984 The load of genetic and partially geneti c disorders in man.

I. Congenital anomahes. Estimates of detriment in terms of years of life lost and years of impaired life. Mutat. Res. 128: 73-103

Czeizel A and Tusn~dy G 1984 Aetiological studies of isolated common congenital abnormalities in Hungary. Akad6miai Kiad6, Budapest

Denniston C 1982 Low level radiation and genetic risk estimation in man. Annu. Rev. Genet. 16:329-355 Dutta S K 1984 Chromosomal split genes. In Genetics: New frontiers Proc. X V Int. Cong. Genet. (eds) V L

Chopra, B C Joshi, R P Sharma and H C Bansal (New Delhi: Oxford and IBH) pp. 175-186 Ehling U H 1965 The frequency of X-ray-induced dominant mutations affecting the skeleton of mice.

Genetics 51:723-732 Ehling U H 1966 Dominant mutations affecting the skeleton in offspring of X-irradiated male mice. Genetics

54:1381-1389 Ehling U H 1980 Strahlengenetisches Risiko des Menschen. Umsch. Wiss. Tech. 80:754-759 Ehling U H, Favor J, Kratochvilova J and Neuh/iuser-Klaus A 1982 Dominant cataract and specific locus

mutations in mice induced by radiation or ethylnitrosourea. Mutat. Res. 92:181-192 Emery A E H 1984 An introduction to recombinant DNA (New York: John Wiley) Evans H J 1984 Structure and organization of the human genome. In Mutation in man (ed.) G Obe (Berlin:

Springer Verlag) pp. 58-100 Forget B G, Benz E J and Weissman S M 1983 Normal human globin gene structure and mutations causing

the fl-th;dassemia syndromes. In Recombinant DNA applications to human disease (eds) C T Caskey and R L White, Banbury Report 14, Cold Spring Harbor Laboratory, New York, pp. 3-17

Green M C, Lane P C and Southard 1970 In 41st Annual Report of the Jackson Laboratory, Bar Harbor, Maine (1969-1970); cited in UNSCEAR 1972

Guerro I, Caizada P, Mayer A and Pellicar A 1984 A molecular approach to leukaemogenesis: mouse lymphomas contain an activated c-ras oncogene. Proc. Natl. Acad. Sci. USA 81:202-205

Hecht F and Hecht B K 1984a Fragile sites and chromosome breakpoints in constitutional rearrangements. I. Amniocentesis. Clin. Genet. 26:169-173

100 K S a n k a r a n a r a y a n a n

Hecht F and Hecht B K 1984b Fragile sites and chromosome breakpoints in constitutional rearrangements. II. Spontaneous abortions, stillbirths and newborns. Clin. Genet. 26:174-177

Jelinek W E and Schmid C W 1982 Repetitive sequences in eukaryotic DNA and their expression. Annu. Rev. Biochem. 51:813-844

Kratochilova J and Ehling U H 1979 Dominant cataract mutations induced by gamma irradiation of male mice. Mutat. Res. 63:221-223

Lewandowski R G and Yunis J J 1977 Phenotypic mapping in man. In New chromosomal syndromes (ed.) J J Yunis (New York: Academic Press) pp. 369-394

Lfining K G and Searle A G 1971 Estimates of genetic risks from ionizing radiation. Murat. Res. 12:291-304 McClintock B 1948 Mutable loci in maize. Carnegie lnst. Washington, Yearb. 47:155-163 MeClintock B 1950 The origin and behaviour of mutable loci in maize. Proc. Natl. Acad. Sci. USA 36:344-355 McClintock B 1952 Chromosome organization and genie expression. Cold Spring Harbor Symp. Quant. Biol.

16:13-47 McClintock B 1955 Controlled mutations in maize. Carnegie Inst. Washington, Yearb. 54:245-258 McKusick V A 1971 Mendelian inheritance in man. Catalooues ofautosomal dominant, autosomal recessive

and X-linked phenotypes, 3rd edn (Baltimore, Md: Johns Hopkins University Press) McKusick V A 1983 Mendelian inheritance in man. Catalogues ofautosomal dominant, autosomal recessive

and X-linked phenotypes, 6th edn (Baltimore, Md: Johns Hopkins University Press) McKusick V A 1985 The human gene map 1 December 1984. Clin. Genet 27:207-239 MRC 1956 The hazards to man of nuclear and allied radiations, emd 9780, Medical Research Council, H M S

Office, London Muller H J 1927 Artificial transmutation of the gene. Science 66:84-87 Muller H J 1950 Our load of mutations. Am. 3. Hum. Genet. 2:111-176 Noel J V, Kato H and Sehull W J 1974 Mortality in the children of atomic bomb survivors and controls.

Genetics 76:311-326 Neel J V, Schull W J and Otake M 1982 Current status of genetic follow-up studies in Hiroshima and

Nagasaki. In Prooress in mutation research (eds) K C Bora, G R Douglas and ER Nestmann (Amsterdam: Elsevier Biomedical Press) pp. 39-51

Nomura T 1982 Parental exposure to X-rays and chemicals induces heritable turnouts and anomalies in mice. Nature (London) 296:575-577

Nomura T 1983 X-ray-induced germ-line mutations leading to tumours: its manifestation by post-natally given urethane in mice. Mutat. Res. 121:59-65

Nomura T 1985 Further studies on X-ray and chemically-induced germ-line alterations causing tumours and malformations in mice. In Proc. I V Int. Conf Environmental Mutagens, Stockholm (in press)

Oftedal P and Mulvihill J J 1985 Possible genetic effects of cancer therapy seen in surviving cancer patients' offspring. In Proc. I V Int. Conf Environmental Mutagens, Stockholm (in press)

Opitz J M and'Sutherland G R 1984 Conference report. Int. Workshop on the fragile-X and X-linked mental retardation. Am. J. Med. Genet. 17:5-94

Orkin S H 1983 A review of fl-thalassemias. In Recombinant DNA applications to human disease (eds) C T Caskey and R L White, Banbury Report 14, Cold Spring Harbor Laboratory, New York, pp. 19--28

Orkin S H and Kazazian H H 1984 The mutation and polymorphism of the human fl-globin gene and its surrounding DNA. Annu. Rev. Genet. 18:131-171

Rees D J G, Rizza C R and Brownlee G G 1985 Haemophilia B caused by a point mutation in a donor splice junction of the human factor IX gene. Nature (London) 316:634-645

Russell L B 1971 Definition of functional units in a small chromosome segment of the mouse and its use in interpreting the nature of radiation-induced mutations. Mutat. Res 11:107-123

Russell W L 1963 The effects of radiation dose rate and fraetionation on mutation in mice. In Repair from genetic radiation damage (ed.) F H Sobels (Oxford: Pergamon Press) pp. 313-352

Russell W L 1965 Effect of intelwal between irradiation and conception on mutation frequency in female mice. Proc. Natl. Acad. Sci. USA 54:1552-1557

Russell W L 1977 Mutation frequencies in female mice and the estimation of radiation hazards in women. Proc. Natl. Acad. Sci. USA 74:3523-3527

Russell W L and Russell L B 1958 Radiation-induced genetic damage in mice. In Proe. II Int. Cotr Peaceful Uses of Atomic Energy. Biological Effects of Radiation (Geneva: United Nations) vol. 22, pp. 360-365

Russell W L, Russell L B and Cupp M B 1959 Dependence of mutation frequency on radiation dose-rate in female mice. Proe. Natl. Acad. Sci. USA 45:18-23

Russell W L, Russell L Band Kelly E M 1958 Radiation dose rate and mutation frequency. Science 128: 1546-1550

E v a l u a t i o n o f gene t ic rad ia t ion hazards in m a n 101

Russell W L, Russell L B and Kelly E M 1960.Dependence of mutation rate on radiation intensity. In Intermediate and low level effects of ionizin9 radiation (ed.) A A Buzzati-Traverso (London: Taylor and Francis) pp, 311-320

Sanchez O and Yunis J J 1977 New chromosomal syndromes (ed.) J J Yunis (New York: Academic Press) pp. 1-54

Sankaranarayanan K 1976 Evaluation and re-evaluation of genetic radiation hazards in man. III. Other relevant data and risk assessment. Murat. Res. 35:387-414

Sankaranarayanan K 1986 Transposable genetic elements, spontaneous mutations and the doubling-dose method of radiation genetic risk evaluation in man. Mutat. Res. 160:73-86

Schmid C W and Paulson K W 1984 Interspersed repeats in mammalian DNA: a status report. In Genetics: New frontiers, Proc. X V Int. Cong. Genetics (eds) V L Chopra, B C Joshi, R P Sharma and H C Bansal (New Delhi: Oxford and IBH) pp. 255-267

Schull W J, Otake M and Neel J V 1980 Genetic effects of the atomic bombs: a re-appraisal. Science 213: 1220-1227

Schull W J, Neel J V, Otake M, Awa A A, Satoh C and Hamilton H B 1982 Hiroshima and Nagasaki: three and a half decades of genetic screening. In Environmental mutagens and carcinogens (eds) T Sugimura, S Kondo and H Takebe (Tokyo: University Press and New York: Alma R. Liss) pp. 687-700

Searle A G 1976 Paper presented at the Seminar on Radiation Biology and Radiation Protection, Euratom Meeting, Orsay, France (May 1976)

Selby P B and Selby P R 1977 Gamma-ray-induced dominant mutations that cause skeletal abnormalities in mice. I. Plan, summary of results and discussion. Mutat. Res. 43:357-375

Shapiro J A (ed.) 1983 Mobile genetic elements (New York: Academic Press) Sherman S L, Morton N E, Jacobs P A and Turner G 1984 The marker-X syndrome: a cytogenetic and

genetic analysis. Ann. Hum. Genet. (London) 48:21-37 Smith C 1975 Quantitative inheritance. In Text book of human .genetics (eds) G R Fraser and O Mayo

(Oxford: Blackwell Scientific Publications) ch. 9, pp. 382-411 Stadler L J 1928a Genetic effects of X-rays in maize. Proe. Natl. Aead. Sci. USA 14:69-75 Stadler L J 1928b Mutations in barley induced by X-rays and radium. Science 68:186 Stevenson A (3 1959 The load of hereditary defects in human populations. Radiat. Res. suppl. 1:306-325 Sukumar S, Notario V, Martin-Zanca D andBarbaeid M 1983 Induction of mammary carcinomas in rats by

nitrosomethylurea involves malignant activation of N-rasl locus by single point mutations. Nature (London) 306:658-661

Sutherland G R 1983 The fragile-X chromosome, lnt. Re,x Cytol. 81:107-143 Thacker J 1986 The nature of mutants induced by ionizing radiation in cultured hamster cells. III. Molecular

characterization of HPRT-deficient mutants induced by gamma-rays or alpha particles showing that the majority have deletions of all or part of the hprt gene. Mutat. Res. 160:267-275

Trimble B K and Doughty J H 1984 The amount of hereditary disease in human populations. Ann. Hum. Genet. (London) 38:199-229

Turner D R, Morley A A, Halindros M, Kutlaca R and Sanderson B J 1985 In vivo somatic mutations in human lymphocytes frequently result from major alterations. Nature (London) 315:343-345

UNSCEAR 1958 Report of the United Nations Scientific Committee on the Effects of Atomic Radiation. General Assembly, Official Records, 13th session, Suppl. 17 (A/3838), United Nations, New York

UNSCEAR 1962 Report of the United Nations Scientific Committee on the Effects of Atomic Radiation. General Assembly, Official Records, 17th session, Suppl. 16 (A/5216), United Nations, New York

UNSCEAR 1966 Report of the United Nations Scientific Committee on the Effects of Atomic Radiation. General Assembly, Official Records, 21st session, Suppl. 14 (A/6314), United Nations, New York

UNSCEAR 1972 Ionizing radiation: Levels and effects. A Report of the United Nations Scientific Committee on the Effects of Atomic Radiation to the General Assembly with Annexes, vol. 2. Effects. United Nations, New York

UNSCEAR 1977 Sources and effects of ionizing radiation. Report of the United Nations Scientific Committee on the Effects of Atomic Radiation, United Nations, New York.

UNSCEAR 1982 Ionizing radiation: Sources and biological effects. Report of the United Nations Scientific Committee on tile Effects of Atomic Radiation to the General Assembly with Annexes, United Nations, New York

Van Buul P P W 1980 Dose-response relationship for X-ray-induced reciprocal translocations in stem cell spermatogonia of the rhesus monkey (Macaca mulatta). Murat. Res. 73:363-375

Vogel F 1984 Gene or point mutations. In Mutations in man (ed.) G Obe (Berlin: Springer Verlag) pp. 101-127


Vrieling H, Simons J W I M, Arwert F, Natarajan A T and van Zeeland A A 1985 Mutations induced by X- rays at the HPRT locus in cultured Chinese hamster cells are mostly large deletions. Murat. Res. 144: 281-286

WHO 1977 International class~cation of diseases, injuries and causes of death (Geneva: World Health Organization) vols. 1 and 2

Wilson J M, Young A B and Kelly W N 1983 Hypoxanthine guanine-phosphoribosyl-transferase deficiency: the molecular basis of clinical syndromes. New Engl. J. Med. 309:90(0-910

Yang T P, Patel P I and Chinault A C 1984 Molecular evidence for new mutations at the HPRT locus in Lesch- Nyhan patients. Nature (London) 310:412-414

Documents

Evaluation of genetic radiation hazards in man: History