In summary, while these more recent data on adaptive responses for radiation-induced tumorigenesis may act as a focus for further research, they do not provide coherent evidence of the generality of this mechanism and its importance for judgments on low-dose cancer risk.
It has been known for many years that there are individuals and families within human populations who carry heritable mutations that can increase their lifetime probability of spontaneously developing cancer. Indeed, family pedigrees providing evidence of strongly expressing predisposition, particularly to colon carcinoma, were published in the early part of the 1900s, but it was not until the development of molecular genetic techniques in the 1970s that the whole field of human cancer genetics began its rapid development.
The primary mechanistic association between heritable cancer in humans and exposure to an environmental carcinogen was made in the late 1960s when Cleaver (1968) demonstrated an excess of skin cancer in sun-exposed DNA, repair-deficient xeroderma pigmentosum (XP) patients (i.e., there was likely to be a direct association between heritable DNA repair or damage response capacity and cancer development). Since the 1970s the generality of this crucial association has been much more firmly established by a combination of clinical, epidemiologic, and molecular genetic approaches. These developments have included the elucidation of two rare human genetic disorders of cancer, ataxia-telangiectasia (AT; Easton 1994) and Nijmegen breakage syndrome (NBS), in which the DNA damage response defects concern the form of DNA damage (Brenner and Ward 1995) critical for cellular response to ionizing radiation (Taylor and others 1994a; Savitsky and others 1995). The DNA damage response defects in these human disorders are considered in depth elsewhere in this report. However, as evident from the data outlined in the following sections, genetic susceptibility to radiogenic tumors extends beyond a simple relationship between DNA damage response deficiency, cellular radiosensitivity, and neoplastic development (ICRP 1998; NRPB 1999).
The first objectives of this section are to outline the data that relate to (1) cancer-prone human genetic disorders determined by strongly expressing genes, (2) less strongly expressing cancer-associated genes, and (3) the evidence available on radiosensitivity and predisposition to radiation tumorigenesis. The principal conclusions from these reviews will then be applied in the development of judgments on the identification of human subgroups having potentially increased cancer risk after radiation and the likely magnitude of that increased risk. In developing these judgments, particular attention will be given to the uncertainties involved.
The whole field of cancer genetics has expanded dramatically in the last 15 years, and it is appropriate to provide only a brief overview here. Detailed reviews are given elsewhere (Eeles and others 1996; ICRP 1998).
Published genetic catalogs (McKusick 1998; Mulvihill 1999) show that around 6% of recorded human disorders and mutant genes have some degree of association with neoplastic disease. The number of such disorders for which the association is unambiguously strong remains small (less than 50) and tends to be restricted to rare autosomal recessive and autosomal dominant diseases. Highly expressing autosomal dominant diseases usually manifest as familial cancer, often without other major clinical features. As a genetic grouping, these have received much attention in recent years. Autosomal recessive diseases tend to be more rare, and excess cancer is usually accompanied by other characteristic clinical features. Since their manifestation demands a genetic input from both parents, these disorders do not typically express as familial cancer.
The majority of human genetic diseases associated with DNA damage response and repair fall into this category. Table 3-3 outlines examples within this category including AT and NBS. There are also examples of autosomal recessive and X-linked disorders of the immune system, which manifest as susceptibility to virally associated neoplasia (ICRP 1998); these are not considered here.
In this category are examples of mutations in DNA damage response or repair genes, in proto-oncogenes, and in tumor-suppressor genes. Table 3-4 outlines examples of human disorders that make up this grouping.
In considering the examples given in Tables 3-3 and 3-4, a number of general points can be added to the descriptions. First, there are genetic disorders that might qualify for inclusion in both DNA damage response or repair and tumor-suppressor categories. The prime example is Li-Fraumeni syndrome, which may be ascribed to DNA damage response and tumor suppression activity of the responsible TP53 gene (ICRP 1998). However, on the basis of their autosomal dominant inheritance and gene loss in tumors, DNA mismatch repair defects in hereditary nonpolyposis colon cancer and, possibly, BRCA-type heritable breast cancer might also be included in the tumor-suppressor category.
Second, there are general clinical and medical genetic features of the cancer-prone disorders of Tables 3-3 and 3-4 that are important for the judgments to be developed. For autosomal dominant human mutations of cancer to be detected readily in the population via family studies, the