Nakahara memorial lecture. Hereditary cancer, oncogenes, and anti-oncogenes.

Dominantly heritable predisposition to cancer is well known, even if rare. Such heritable cases are known for most cancers. Predisposition is usually to one or a few specific cancers, and the kinds of pedigrees that are found suggest the existence of an array of 50 or so such cancer genes. Penetrance is usually high, as is the relative risk for a particular tumor. The inherited event is insufficient to cause cancer; at least one other (somatic) event must occur. For one tumor, retinoblastoma, the total number of necessary events seems to be two for both the hereditary and non-hereditary forms, and those events involve mutation or loss of the two copies of a tumor suppressor gene on chromosome 13. Analyses of tumors with polymorphic syntenic DNA probes and for abnormality of the gene itself have provided a picture of the kinds of first and second events that can occur. Several other tumors may follow the retinoblastoma scenario, including tumors associated with multiple endocrine neoplasia type 1 and with neurofibromatosis type 2, and identify the loci of several other putative anti-oncogenes. Still other hereditary cancer genes have been mapped, but evidence that they are suppressor genes is incomplete. It is possible that the inherited mutation may sometimes occur in an oncogene, although no such cases have been validated. In most, perhaps all, tumors, further genetic changes occur. In some instances these appear to be in oncogenes, and sometimes in anti-oncogenes. Some of these events appear to play a role similar to that of promotion in experimental carcinogenesis, whereas others are clearly important in progression. The process is particularly complicated in some of the common carcinomas, where changes in both oncogenes and anti-oncogenes are common.