Changing perspectives on the genetic doubling dose of ionizing radiation for humans, mice, and Drosophila.

Early on in the various studies of humans and experimental organisms that were directed at the genetic effects of radiation, the issue arose as to the most appropriate manner in which to present these effects, both to the scientific community and to a concerned public. One school of thought favored presenting findings simply in terms of risk of mutation per gene per generation per unit of radiation [in the early days, per roentgen (rad) or per roentgen equivalent man (rem), now per Gray (Gy) (100 rad 5 1 Gy) or per Sievert (Sv) (100 rem 5 1 Sv)]. A second school argued that inasmuch as (spontaneous) mutation was a fact of life, presenting ‘‘added’’ risk with no references to ‘‘customary’’ risk (from spontaneous events) failed to provide the necessary perspective, and that a more useful estimate was the ‘‘doubling dose.’’ This is defined as the exposure to ionizing radiation (or other mutagen) that will produce the same frequency of mutation per generation as occurs spontaneously. It may be expressed either as the dose per gamete or the dose per zygote. Because radiation-induced rates may differ in the male and female, the zygotic doubling dose is not simply twice the spermatogonial doubling dose. In the studies in Japan on the children of atomic bomb survivors, because we studied the results of the radiation of a total population, the estimate we have derived is the zygotic doubling dose. Although there have been, and remain, uncertainties concerning the total impact of spontaneous mutation on any animal population, most geneticists have an (intuitive?) feeling for this impact, and what doubling the impact might imply for a population. The ‘‘doubling dose’’ treatment has become standard, especially for across-species comparisons, and is pursued in this paper. As we shall see, simple though the concept is, the derivation of a figure that will include the impact of all aspects of the complex mutational process has proved very difficult. In particular, because it depends on a ratio, of induced to spontaneous rate, each component of which has biological variability as well as an error of estimation, the error in the estimate of the doubling dose is apt to be large but indeterminate. The estimate of the doubling dose can also be influenced by such factors as stage of germ cell at irradiation, type of mutation scored, sex of animal, or population structure (humans) or precise research design (experimental organisms), and in across-species comparisons, one must employ similar indicators as far as possible (Muller, ’59). In the decade following the atomic bombings, the uncertainties were such that no less a presence than Haldane (’55) could suggest that for humans, the doubling dose might be 0.03–0.05 Gy of acutely delivered ionizing radiation, suggesting that background radiation from natural sources might make a substantial contribution to so-called spontaneous mutations. There are two especially troublesome issues in the comparisons to be made. The first is the differing radiation exposures across species at which doubling doses are derived. For instance, most of the Drosophila data were derived at acute exposures of approximately $30 Gy, whereas the mouse data were usually derived at exposures of 3–6 Gy. The average combined (mother and father) gonadal exposure of the atomic bombexposed parents of the children examined in Japan is currently estimated at 0.4to 0.5-Sv equivalents, but with continuing uncertainty about the magnitude of the (highly effective) neutron component in the ionizing radiation. Because of the special interest in the human doubling dose, we shall, wherever possible in comparisons, use the ‘‘low-dose’’ data from experimental organisms. The second issue arises from the well-documented greater sensitivity of mature and maturing sperm than spermatogonia to ionizing radiation. As the human data are predominantly based on the irradiation of gonia-stage cells (or immature oocytes), we will also resort, wherever, possible to similar data from experimental organisms.