Two recent publications have examined the role of heritable genetic factors in the cancer process and how genetic heterogeneity in cancer susceptibility might impact upon radiation protection practices. The International Commission on Radiological Protection (ICRP) report Genetic Susceptibility to Canceris the more extensive and begins with a comprehensive appraisal of radiation-induced DNA damage and repair processes. The misrepair of DNA double strand lesions leading to chromosome exchanges and deletions is considered characteristic of radiation exposure and the role of the XRCCand radfamilies of genes in this process is described. Ataxia telangiectasia is used as a model for elucidating the mechanisms involved in cancer proneness, with misrepair of DNA double strand breaks, disruption of signal transduction and impaired cell cycle checkpoint functions being implicated. Evidence is also presented for the role of recombinational processes in double strand break repair and the induction of deletion mutants as the prime mechanism for producing gene mutations by radiation. Experimental studies on the induction of genomic instability by radiation are reviewed, although it is concluded that the mechanistic basis for this phenomenon and its relationship with neoplastic development remains to be established. Although increased radiation-induced cellular lethality has been claimed for a range of genetic disorders, it is judged that only a minority show unambiguous evidence of cellular radiosensitivity. However, developments in chromosomal assays are thought to hold more promise for the discrimination of variations in radiosensitivity, with such studies leading to an understanding of how the low penetrance genes involved in this response may be related to cancer predisposition.

A section devoted to solid tumours provides a comprehensive discussion of the role of oncogenes and tumour suppressor genes and how mutations in these, together with those in DNA repair and cell cycle control genes, contribute to tumourigenesis. The functions associated with oncogenes and tumour suppressor genes are dealt with in some detail, particularly in relation to regulation of the cell cycle, DNA damage checkpoint control and the roles of the RB1and p53genes in these processes. The predominant event in the genesis of solid tumours is the somatic homozygous loss of function of tumour suppressor genes and a number of these genes have also been identified as predisposing germ line mutations expressing as familial cancer. Mutations in DNA mismatch repair genes are a critical component of both sporadic and some familial solid tumours. Defects in DNA repair, DNA replication and chromosomal segregation are suggested as early events which lead to an unstable genotype and thus an increase in probability of damage to critical oncogenes and tumour suppressor genes. Gene involvement and function in heritable susceptibility to solid tumours is outlined with reference to diseases in specific organ systems. It is noted that factors that affect the penetrance and expressivity of germ line tumour gene mutations are, at present, not well understood but are probably influenced by genotypic profile and, more importantly, environmental factors such as diet and genotoxic exposure. Highly penetrant germ line mutations are considered to account for around 5% of all solid cancers but, as yet, the contribution from mutations of lower penetrance which do not manifest as familial cancer is not clear, and a larger overall contribution to risk cannot be ruled out.

An introduction on the structure, development and function of the lymphohaemopoietic system leads onto a section on leukaemias and lymphomas. Leukaemogenesis is viewed as an uncoupling of the process of proliferation and maturation in relatively immature cells and this can occur at various stages in the development of the different haemopoietic cell lineages. The most consistent feature of human leukaemias and lymphomas is the presence of disease-specific chromosome rearrangements involving changes in the activity of proto-oncogenes. The presence of non-random chromosome deletions is also well established, and, although these are predicted to involve tumour suppressor genes, few have been identified to date; this is recognised as a crucial area for the future study of the leukaemogenic process. Further proto-oncogene, tumour suppressor and gene amplification mutations are thought to occur as secondary events following the lineage-specific chromosome changes which initiate the neoplastic process. Familial susceptibility to these disorders is judged to be extremely rare. In support of this view, studies examining incidence patterns in families, and particularly in twins, are reviewed. It is observed that the well characterised germ line mutations predisposing to solid cancers in a wide range of cancer family syndromes, in the main, do not impact on leukaemia or lymphoma incidence, the exceptions being Li Fraumeni syndrome and neurofibromatosis. Genetic predisposition is, however, evident in a number of rare specific recessive disorders associated with chromosome instability and/or immunodeficiency, many of which are characterised by defects in DNA damage processing, e.g. ataxia telangiectasia, Fanconi anaemia and Bloom syndrome. The specific somatic chromosome exchanges that characterise many types of leukaemia and lymphoma have not been reported in the germ line of patients and are thus thought to play little or no role in the germ line predisposition to haemopoietic malignancy. However, inherited chromosome imbalance, most notably that resulting in Down syndrome, is associated with greater incidence of disease. It is concluded that the inherited genetic contribution to leukaemogenesis, although still poorly defined, is less than that for solid tumours and is probably greatest in childhood disease where it may account for 3-4% of disease. Nevertheless, in view of the current uncertainty in respect of leukaemogenic mechanisms, a value of 5% is judged to be reasonable for the genetic contribution for application to a population of all ages.

Having laid the groundwork for an examination of the relationship between cancer radiosensitivity and heritable predisposition to cancer, it is suggested that, in most cases, tumour cancer predisposition associated with DNA repair or tumour suppressor genes will be associated with increased radiation cancer risk, this being most certain for inherited deficiencies in tumour suppressor genes. Support for this is drawn from animal models, although it is advised that extrapolation of animal data to humans should be treated with caution since it has been observed that differences in genetic background may have profound effects on gene expression. Further support comes from studies of cancer patients receiving radiotherapy, with such studies indicating increased radiosensitivity to further malignant disease in those patients with genetically predisposing cancer disorders, e.g. retinoblastoma and Li Fraumeni syndrome. Epidemiological studies of breast cancer in Japanese A-bomb survivors, which found a 6-fold increase in excess relative risk per Sv for early onset cancer in women exposed below the age of 20, are suggestive of the presence of a genetic subgroup of radiosensitive cases, but these and other such studies are unlikely to provide unequivocal evidence of genetic influences on radiation cancer risks. The committee emphasises the paucity of data currently available for developing judgements applicable to radiation protection practices. However, an interim judgement, based on animal and human data, suggests that while the risk of cancer following irradiation may be elevated up to 100-fold in some heritable cancer disorders, a single best estimate of a 10-fold increase in risk is appropriate for the purposes of modelling radiological impact.

The report emphasises the difficulty in making estimates of excess cancer attributable to a genetically predisposed subgroup in an irradiated human population when the current knowledge of two critical parameters is sparse, i.e. the number of individuals in the population that carry the mutant genes that predispose to cancer and the degree of tumourigenic radiosensitivity conferred to mutant gene carriers. A model is described into which estimates of disease prevalence, strength of predisposition and tumourigenic radiosensitivity can be introduced as data accumulates, and this is used to explore the impact of two diseases with a genetic component for which some data is currently available, namely breast and colon cancer. As a result of this exercise it is concluded that irradiation of a heterogeneous population containing gene carriers of predisposing cancer genes will result in higher cancer risks than if such individuals are not present, but unless the mutant gene frequencies, the proportion of cancers attributable to the gene being studied, the radiosensitivity differentials and the strength of the predisposition are sufficiently large, the enhancement in risks is small. For the two diseases examined, breast and colon cancer, the mutant gene frequencies based on current knowledge are too low to make a significant impact on radiological risk in typical human populations.

In view of the current state of knowledge only interim judgements can be made on the impact of genetic factors on radiation-induced carcinogenic risk and hence radiological protection. With respect to highly penetrant genetic factors responsible for familial disorders, the gene frequencies are considered too low and the increase in radiogenic cancer risk not sufficient to provide a significant distortion of risk in the whole population. In terms of individual risk, the excess risk needs to be considered in the context of the overall increased spontaneous risk conferred by the mutant gene. For disorders associated with cancer at a specific site, the impact of low dose protracted exposures of the order of 100 mSv on overall cancer risk will be small, although for disorders associated with excess cancer of multiple sites the absolute risk will be less modest. For high doses experienced in radiotherapy, risk will be dependent on a number of factors, in particular age of exposure. The problem of ascertainment of cancer predisposing mutations of low penetrance means that no firm judgements are possible on their implications for radiological protection. Their impact on the pattern of population risk will be dependent on the number of different genes and the extent of their individual contribution to radiation-induced cancer. In general the prevalence of carriers would need to be high in order to affect current judgements on risk, although if specific gene combinations conferred particular sensitivity this could affect population risk distribution. For individual risk assessments, low dose effects are expected to be minor although high dose exposure may be cause for concern. However, the possibility still remains that there are rare disorders where the spontaneous cancer risk is not elevated but where risk is substantially increased by radiation exposure, and in such cases the individual risk of excess cancer even after low doses may be high relative to baseline risk.

Whilst recognising that extensive research will continue in the field of cancer genetics, the ICRP is of the view that radiological protection research should seek to contribute to a better understanding of cancer predisposing genes of low penetrance and the gene-gene and gene-environment interactions that influence phenotypic expression. Epidemiological studies of irradiated populations and cancer case series are part of this approach, but these will need to incorporate appropriate molecular analyses if they are to lead to advances in knowledge of genetic effects on radiation risk. Experimental animal studies on gene deficient mice and molecular analyses of interstrain differences in radiosensitivity should be pursued. There is also great scope for further work on chromosomal radiosensitivity in relation to novel genes controlling radiation response, and associated with this is the need to establish specific banks of cells with known or suspected altered radiosensitivities.

On the basis of the data reviewed and the judgements derived, ICRP examine the practical implications for radiological protection practices. The Commission recognise that existing measures of risk, which are based on large heterogeneous populations, will include a genetic contribution and thus risk will not be uniformly distributed. However, since only around 1% of the population fall into the familial cancer category with potentially increased radiosensitivity, known genetic factors should not impose an appreciable degree of distortion, and thus existing recommendations on cancer risk in irradiated populations are not subject to unacceptable genetic uncertainty. For specific individuals with familial cancer there is little value in recommending specific restrictions for low dose exposure. Although the enhanced risk to familial cancer cases associated with high dose is difficult to estimate with real confidence, ICRP recommend that the clinical benefits of radiotherapy should be viewed against a possible 10-fold or more increased chance of a second cancer arising in irradiated normal tissue.

Since genetic testing is currently the subject of wide ranging ethical debate and is likely to become a legislative issue, the Commission feel it is inappropriate to make specific recommendations on the employment of genetic testing in the context of radiological protection. It is, however, observed that such testing may in future play a role in the development of therapeutic strategies involving high dose radiotherapy regimes, but for low dose exposure the cost involved and the small elevated risk for which avoidance is being sought suggest that genetic testing will not play a significant role in the context of occupational exposure. In conclusion, it is stressed again that the data currently available are limited and the judgements given should be regarded as preliminary. Emphasis is placed on the need for knowledge on the identity, prevalence and impact of weakly expressing mutations which may influence response but do not manifest as familial cancer.

In providing the background to their deliberations on the current assessment of the implications of genetic variability in cancer susceptibility to radiological protection, ICRP have produced an important review and source of reference. Whilst perhaps rather daunting for those without a scientific background in radiobiology and/or cancer genetics, it illustrates how the increasing depth of scientific knowledge amassing in this area will inevitably lead to ongoing reassessments of risks and hence radiological protection practices.

An easier read is a publication from the National Radiological Protection Board (NRPB) entitled Genetic Heterogeneity in the Population and its Implications for Radiation Risk . This report summarises current knowledge on the heterogeneity in response to ionising radiation arising from individual genetic variation, and essentially comes to the same conclusions and highlights the same uncertainties as expressed by ICRP. A review of the known cancer-prone disorders concludes that high penetrating cancer-predisposing genes are carried by less than 1% of western populations, but also recognises that identifying and judging the impact of weakly expressing mutations and those subject to significant gene-gene and gene-environment interactions is more difficult. Current understanding of DNA damage and repair mechanisms suggests that many cancer prone disorders will show an enhanced radiation-induced cancer risk. In common with the ICRP report, evidence of heterogeneity in inherited susceptibility to radiation-induced cancer is drawn from chromosomal radiosensitivity studies in cultured cells, animal models of cancer-predisposing conditions and epidemiological studies of radiation-exposed populations. Thus a similar interim judgement of an average 10-fold increase in risk of radiation-induced cancer for known genetic disorders associated with a specific tumour is derived with the same provisos that the data underpinning this judgement are limited. Thus, lack of information on the genetic component of all tumour types, the penetrance and prevalence of the genes involved, and the degree of radiosensitivity conferred by such genes, leads NRPB to conclude that computational modelling of the impact of genetic factors on overall radiation-induced population cancer risk is not currently appropriate. However, modelling of specific disorders associated with high prevalence genes for which some data are available can be informative. Whilst agreeing with the ICRP conclusions from this approach that the prevalence of highly penetrant genes is too low to significantly distort population radiation risk, the NRPB express caution in the interpretation of the impact of low dose exposure on individual risk. Whilst the increase in risk in susceptible individuals in relative terms may be small, to regard this as acceptable involves an ethical decision which needs to be addressed since the same absolute risk in non-predisposed individuals might not be acceptable. Nevertheless, doubt is expressed as to the economic viability of a genetic screening strategy in the context of occupational exposure and medical diagnostic procedures. Individual genetic testing, however, is viewed as potentially beneficial for establishing appropriate cancer therapy regimes. Further work is recommended and echoes that outlined by ICRP. It is concluded that the present state of knowledge offers no reasons to alter current radiological standards and practices. However NRPB envisage a time when the identification of radiosensitive individuals will be possible and this will raise ethical issues that may have implications for radiological protection in the future.

E Janet Tawn Westlakes Research Institute Moor Row Cumbria CA24 3JY UK