The linear no-threshold extrapolation from a dose-response relationship for ionizing radiation derived at higher doses to doses for which regulatory standards are proposed is being challenged by some scientists and defended by others. It appears that the risks associated with exposures to doses of interest are below the risks that can be measured with epidemiological studies. Therefore, many have looked to biology to provide information relevant to risk assessment. The workshop reported here, "The Impact of Biology on Risk Assessment", was planned to address the need for additional information by bringing together scientists who have been working in key fields of biology and others who have been contemplating the issues associated specifically with this question. The goals of the workshop were to summarize and review the status of the relevant biology, to determine how the reported biological data might influence risk assessment, and to identify subjects on which more data are needed.

A linear dose response was observed for radon-induced mutations at the CHO-hprt locus with an induction frequency of $1.4\times 10^{-4}$ mutants per viable cell per gray. Mutants isolated after two levels of radon exposure were evaluated using Southern blot techniques and polymerase chain reaction (PCR) exon amplification. No significant differences in mutational spectra were detected at these two exposure levels. Of 52 radon-induced mutations, 48% sustained a gene deletion, 23% underwent a rearrangement of the banding patterns or loss of one or more exons, and 29% showed no change from the parental line. These mutants were compared with mutants produced after X irradiation (3 Gy) and with spontaneous mutants from untreated cells. The spectra of mutation types in cells treated with radon and X rays were not significantly different. In contrast, 31 spontaneous mutations exhibited a low percentage of gene deletion events (16%); most spontaneous mutants showed no change (74%); the remaining 10% were classified as alterations. In conclusion, the principal lesion seen at the CHO-hprt locus after radiation exposure in gene deletion, while the predominant class of spontaneous mutations is composed of smaller events not detectable by Southern blot or PCR exon analysis.

The cytotoxic and mutagenic effects of radon and its progeny were compared in murine lymphoblast L5178Y-R16 cells after exposure at three institutions. The cells were exposed to 222 Rn at Case Western Reserve University (CWRU) and Pacific Northwest Laboratories (PNL) and to 212 Bi, a decay product of 220 Rn, at the University of Chicago (UC). The dose to the cell nucleus was calculated using a dosimetric model which addressed both the contribution of the dose from the radioactivity in the medium and that associated with the cells. The dose-response curves for cell survival showed $D_{0}\text{'}{\rm s}$ of 0.30 Gy at CWRU, 0.20 Gy at PNL, 0.37 Gy for chelated 212 Bi, and 0.13 Gy for unchelated 212 Bi. Induced mutant frequencies at the thymidine kinase locus at the 37% survival level were $1470\times 10^{-6}$ at CWRU, 1518 at PNL, and $2414\times 10^{-6}$ at UC using combined results for chelated and unchelated 212 Bi. The variation between institutions was greater than obtained in a previous interlaboratory comparison of the effects of radon on CHO cells. Since less radioactivity was associated with CHO cells than L5178Y cells, we have concluded that the variation between institutions in the case of L5178Y cells is caused by the differences in cell-associated radioactivity and errors related to the measurement of this parameter.

We have developed a model to calculate the dose to the cell nucleus in cells exposed in suspension to radon and/or radon progeny. The model addresses the influence of (1) different radiation qualities and energies in the irradiation milieu; (2) the contribution to dose from radioactivity in the medium surrounding the cell after exposure to the radon gas as well as that from excess radon progeny associated with the cell; (3) the geometry of the cell and of the radiosensitive target, the cell nucleus; (4) the intracellular localization of the radionuclides; (5) attenuation of the α particles by the cytoplasm; (6) the radionuclide concentrations in the medium; and (7) the length of exposure. Investigation of the influence of these various parameters was made using an irradiation system in which cells were exposed to 212 Bi, which decays to stability with the emission of an α particle (either 6.05 or 8.78 MeV). The information from these studies was then used to develop the system further for more complex systems in which 222 Rn and its progeny are present. The model takes into account the contribution of dose from different radiation sources using scintillation counts of the medium and the cells, and it is useful for calculations of dose in situations where cells are exposed in suspension culture.

Chinese hamster ovary cells were given two 450-rad doses of X rays separated by 1 or 2 hr, during which the cells were held at room temperature to minimize progression through the cell cycle. The frequency of induced mutations to 6-thioguanine resistance was compared with that of mutations induced by a single dose of 450 or 900 rad. Mutation frequencies induced by the split dose were intermediate between those induced by the two single doses, suggesting that "premutational" lesions were repaired.

Synchronous Chinese hamster ovary cells obtained by mitotic selection were X-irradiated at various points in the cell cycle and were analyzed for the frequency of mutants resistant to 8-azaguanine (AG). The induced frequencies were highest in mitosis ( $40\times 10^{-5}$ for 5 Gy), in early G 1 ( $30\times 10^{-5}$ ), and at the G 1 / S border ( $30\times 10^{-5}$ ); were low in mid- to late G 1 ( $15\times 10^{-5}$ ); and were lowest in S phase ( $6\times 10^{-5}$ ). Qualitatively, the variation in sensitivity during the cell cycle for induction of mutants corresponded with that reported for induction of chromosomal aberrations. However, a quantitative comparison of the two endpoints and plots of the logarithm of survival versus mutant frequencies or aberration frequencies indicates that the mutant frequencies were much higher than expected when the cells were irradiated in mitosis or at the G 1 / S border. A spectrum of mutants was observed and classified either as complete, i.e., failing to survive in azaserine-hypoxanthine medium and incorporating $[{}^{3}{\rm H}]\text{hypoxanthine}$ ( $[{}^{3}{\rm H}]{\rm Hx}$ ) at a rate less than 1% of that for wild-type cells, or as partial, i.e., surviving in azaserine-hypoxanthine medium and incorporating $[{}^{3}{\rm H}]{\rm Hx}$ at a rate greater than 1% of that for wild-type cells. The percentage of mutants classified as partials was 0-22% for selection in 30 μg/ml AG and 15-60% for selection in 5 μg/ml AG. However, the number of partial relative to complete mutants had no or little dependence on radiation dose or phase of the cell cycle irradiated. Seventeen mutants (10 partials and 7 completes) were cultured for 12-33 weeks in the absence of AG, and only one partial reverted to wild type as assayed by AG resistance or uptake of $[{}^{3}{\rm H}]{\rm Hx}$ . Furthermore, the complete mutants and some of the partials were resistant to 6-thioguanine.