Point Mutations, Deletions, and Radiation Carcinogenesis

Abstract

A recent symposium summary and a letter in Radiation research prompt several comments. Fry, in his summary (1), asks: "Are deletions and perhaps changes in suppressor genes more important in radiation carcinogenesis than are point mutations?" Sentiments along similar lines have been expressed by Weischselbaum and his co-workers (2, 3). Rossi in his letter (4) based on microdosimetric considerations suggests that the production of point mutations is likely to be due to single ionization events, while large deletions are likely to be the result of "two-hit" processes, which can occur in much larger target volumes. He concludes that the event frequency (per gray) for large deletions is much higher than for point mutations and hence that suppressor genes frequently play the predominant role in radiation carcinogenesis. I should like to make the following points: (i) Carcinogenesis is generally regarded as a multistep process, involving various genetic changes, such as the activation of proto-oncogenes, the inactivation of suppressor genes, and the functional inactivation of genes which control cellular senescence [see for example Ref. (5); for review see Ref. (6)]. Each step is a sine qua non for full tumorigenicity; no event can be regarded as more significant than another. Indeed, studies of the clinical progression of colorectal tumors and of glioblastomas (7, 8) have identified genetic aberrations corresponding either to mutations or to deletions and even to amplifications, all of which appear to be equally important. (ii) DNA deletions and rearrangements can materially affect the expression (at the transcriptional and translational levels) of proto-oncogenes, thus converting them to full-blown oncogenes; such DNA deletions are thus not restricted to suppressor genes alone. A good case in point is the induction of skin tumors in rats (9, 10) by low-LET radiation (0.8-MeV electrons). DNA from 10 of the 12 tumors showed amplification of the third exon of the c-myc oncogene, and deletion of the first exon, which is known to regulate expression of the gene. (iii) Rossi (4) argues (correctly, I believe) that the production of point mutations is quite unlikely to be due to multiple ionization events. He then concludes that such production is therefore dose-rate independent and cannot be a major cause of radiation carcinogenesis when it is reduced by dose protraction. This point is open to question. It should be noted that point mutations are the result of the misrepair of a DNA lesion. It is conceivable that cellular repair processes (or a component thereof) are dose-rate dependent; in particular, at high dose rate error-prone repair pathways may predominate over high-fidelity repair mechanisms. A number of reports have demonstrated that mutation induction is in fact decreased when the radiation dose is fractionated or delivered at a low dose rate (11, 12). These studies involve traditional mutation assays, namely induction of resistance to selective agents, and it is not clear if the proliferating mutant cells are the result of gene deletion or of point mutations. Clarity on this issue awaits the results of DNA sequencing experiments. In the interim, a useful pointer can be obtained from Southern blot experiments. For example, Hill and Zhu (13) have shown that, of 21 independent mutations (in the HPRT gene) induced by 7 rays, fully one-third showed no observable change in the banding pattern of the gene. This implies that these mutants are due to extremely small deletions or to point mutatio s. In addition, a recent analysis' of the y-ray-induced mutations at the human HPRT locus showed that up to 50% were due to point mutations or small (2 bp) deletions. (iv) In a series of experiments, Pellicer and colleagues (14, 15) have identified activated K-ras oncogenes in four of seven radiation-induced mouse lymphomas. These genes were all activated by point mutations in the 12th codon of the gene. The work of Sawey et al. (9) is consistent with these findings; 6 out of 12 tumors exhibited point mutati ns in the K-ras gene. It is thus clear that point mutations are frequently found in tumors initiated by radiation. However, it is improbable that the incident radiation could have directly produced alterations in such a small DNA target with such a high (observed) frequency. This remains one of the basic conundrums of radiobiology. Perhaps the answer