The results of Dubrova and colleagues from the three post-Chernobyl studies (two in Belarus and one in Ukraine) and from a study conducted on the population in the vicinity of the nuclear test site in Semipalatinsk (Kazakhstan) provide evidence that mutations at minisatellite loci can be induced by radiation in human germ cells (Dubrova and others 1996, 1997, 2002b). The dose-response relationships, however, remain uncertain because of considerable difficulties in the estimation of parental gonadal doses. For example, in the first Belarus study (Dubrova and others 1996) the level of surface contamination by 137Cs was used as a broad dose measure, and the children of parents inhabiting heavily contaminated areas (>250 kBq m−2) were found to have twice the frequency of mutations compared to those of parents from less contaminated areas (<250 kBq m−2). In the second Belarus study (with more exposed families and more loci sampled), based on estimates of individual doses, two groups were defined: <20 mSv and >20 mSv (Dubrova and others 1997). The mutation frequency in children from the latter group was 1.35 times that in the former and the frequency in both groups was about twofold higher than in unexposed UK controls.
In the Ukraine study (Dubrova and others 2002b), a 1.6-fold increase in mutation rate in the exposed fathers but not exposed mothers (both relative to unexposed controls) was found, but again the dose-response relationship is uncertain. The authors noted that the doses from external chronic irradiation and internal exposures together were of the order of ~100 mSv (excluding short-lived isotopes). In the Semipalatinsk study (Dubrova and others 2002a), again there was a 1.8-fold increase in the first-generation progeny of parents receiving relatively high doses of radiation (cited as >1.0 Sv, but could have been higher or lower). In this study, through the use of three-generation families, the authors obtained evidence for a decline in mutation frequency as population doses decreased. Although this is what one normally would expect, it becomes a puzzling observation in view of the earlier evidence from the authors on ESTR loci on transgenerational mutagenesis in mice (i.e., the persistence of high mutation rate for at least two generations after the initial radiation exposure).
It is intriguing that in all studies discussed above, there is roughly a twofold increase in mutation rate (often less) despite the fact that the estimates of doses range from about 20 mSv to 1 Sv. Also noteworthy is that studies of the children of Chernobyl cleanup workers (estimated dose: <0.25 Sv; Livshits and others 2001) and of children of A-bomb survivors (estimated dose: 1.9 Sv; Kodaira and others 1995; Satoh and Kodaira 1996) do not show any increase in minisatellite mutation frequency. The same is true also of studies of cancer patients who had sustained chemo- and/or radiotherapy (Armour and others 1999; May and others 2000; Zheng and others 2000). The question of whether the induced mutation frequencies reach a plateau at low doses (unlike in the case of ESTR loci in mice) remains open. In a more recent study of the children born to Estonian Chernobyl cleanup workers, Kiuru and colleagues (2003) found that the minisatellite mutation rate was slightly but not significantly increased among children born after the accident relative to that in their siblings born before the accident; the recorded dose levels at which such an effect was seen were 200 mSv. At lower doses, there was no effect. It is obvious that much work is needed to validate the potential applications of minisatellite loci for monitoring mutation rate in human populations.
As discussed in Annex 4F, ESTR loci in mice and minisatellite loci in humans differ in a number of ways: the composition and size of the arrays, their distribution (apparently random in the case of ESTRs and subtelomeric in the case of minisatellites), the manifestation of instability (in both somatic cells and germline in the case of ESTRs, but almost completely restricted to the germline in the case of minisatellites, although the end result is the change in the number of repeat cores with both ESTRs and minisatellites), and mechanisms (ESTR instability appears to be a replication- or repair-based process involving polymerase slippage during replication, whereas minisatellite instability is due to gene conversion-like events involving recombinational exchanges). To what extent these differences may help explain the differences in response between mouse ESTR loci and human minisatellite loci remains to be determined. As pointed out by Yauk and others (2002), “… the use of mouse ESTR loci as models for human minisatellite instability should be treated with considerable caution.” Apart from the evidence that the mutational events represent nontargeted ones, no real insights have emerged thus far on the mechanisms of instability or radiation mutagenesis at these loci. In view of this and the fact that “mutational events” at the mouse ESTR and human minisatellite loci do not pertain to protein-coding genes, these data are not used in risk estimation.
In this section, advances in knowledge reviewed in earlier sections are recapitulated briefly and used to revise the estimates of genetic risks presented in BEIR V (NRC 1990). Additionally the consistency of the main finding of the genetic studies carried out on atomic bomb survivors in Japan (i.e., lack of demonstrable adverse genetic effects of radiation) with the present estimates and the strengths and weaknesses of the latter are discussed. Risks are estimated using the doubling dose method for Mendelian and chronic multifactorial diseases. For congenital abnormalities, mouse data on developmental abnormalities are used without recourse to the doubling dose method. No separate risks are estimated