The long latent periods and the complexity of the neoplastic process have been formidable obstacles in identifying specific radiation effects that initiate the sequence of events in cancer development at the cellular and molecular level. However, some generalizations can be made. Both in vitro and in vivo studies have amply demonstrated that radiation acts principally at the level of initiation of the carcinogenic process and is considerably less effective in promoting already-initiated cells or in influencing the progression of neoplasia (Han and Elkind 1982; Hill and others 1987, 1989; Bowden and others 1990). Mechanisms by which radiation initiates carcinogenesis are still poorly understood. It is generally accepted that the carcinogenic effects of radiation are related to its clastogenic and mutagenic effects, but no causal relationship between changes in specific genes and the development of radiation-induced cancer has been established. In fact, initiation frequencies derived from recent studies that used in vivo/in vitro models for radiation-induced cancer (with initiation frequencies around 10-2 initiated cells per Gy) are not compatible with a target whose size is limited to a specific gene or even a family of several genes (Kennedy 1985; Gould and others 1987; Selvanayagam and others 1995).
Rather, those frequencies indicate that the cellular target for the initiation of carcinogenesis after irradiation constitutes a substantial fraction of the entire genome. Such results have led to new approaches in the exploration of possible mechanisms of radiation carcinogenesis. A major focus of current research is on the role of radiation-induced genetic instability in carcinogenesis.
Over the next few years, two closely linked approaches using animal models of carcinogenesis are likely to contribute to the understanding of the mechanisms of radiation-induced cancer. Researchers conducting this new generation of animal studies are taking advantage of the current rapid development of molecular genetics. A number of laboratories have begun to use genetically engineered mice with alterations in specific genes to determine the influence of these genes (such as ATM, BRCA1, and BRCA2) on susceptibility to radiation-induced cancer. At the same time, other laboratories are focusing on the inherent differences in susceptibility to radiation-reduced cancer among different mouse strains and beginning to dissect genes involved in controlling susceptibility. Both approaches should yield useful information on susceptible subpopulations and might into the underlying lesions and the processing of these lesions, which initiate carcinogenesis after exposure to ionizing radiation. Progress on both fronts should be substantial over the next 4-5 years and results of relevance to risk estimates are expected to be available for an important BEIR VII phase-2 study.