CANCER AND SUPPRESSOR GENES
The notion that there might be such things as tumour suppressor genes came originally from fusing a tumour cell with a normal cell in vitro and showing that the resulting hybrids were generally non-tumourigenic. That was a slightly surprising result at the time but we now have a ready explanation: the normal cell provides a gene or genes that have been lost from the tumour cell (‘complementation’). Two such genes that can switch a cell from normality into cancer mode simply by ceasing to function are the retinoblastoma gene (RB1) and P53, both key cell cycle regulators.
The suggestion that some familial cancers may be caused by loss of function of both copies of a gene was first made by Robert DeMars in 1969, his idea is that an individual might be born with an inherited mutation in one copy of the gene (i.e. one allele) but be perfectly normal and that cancer would only appear if a subsequent (somatic) mutation knocked out the other copy (so the individual then became homozygous for the mutant cancer-causing gene). Alfred Knudson (1971) put his idea on a firm basis mainly by thinking about retinoblastoma, a rare, inherited childhood disease (incidence 1 in 20,000) in which tumours develop in the eye – specifically in photoreceptor cells in the retina. Cancer comes in two forms, sporadic and familial. In sporadic retinoblastoma, there is no family history and just one tumour develops in one eye. In the inherited form tumours arise in both eyes (Fig. 1).
1. Retinoblastoma development.
Knudson suggested that the incidence and growth of retinoblastomas might indeed be explained if both copies of a gene had to be inactivated and that the two forms might arise if those inheriting the disease were born with one defective gene and went on to lose the other (by somatic mutation), whereas in the sporadic form individuals were born genetically normal but acquired somatic mutations in both genes within one cell.
The requirement for two genetic events gave rise to the term ‘two-hit model’ and it was a remarkably perceptive bit of thinking, based as it was on no molecular evidence whatsoever. It has turned out to be absolutely correct for the retinoblastoma gene although it was a very long journey from Knudson’s hypothesis to the identification and cloning of RB1. It emerged that both copies of the gene are indeed defective in all retinoblastomas but the real importance of RB1 in cancer is that it is also knocked out in a substantial proportion of many other tumours: it is lost, for example, in 20 to 30% of lung, breast and pancreatic carcinomas. RB1 is, therefore, the classical model for a tumour suppressor gene in that both paternal and maternal copies of the gene must be inactivated for a tumour to develop (this behaviour is called ‘recessive’). Oncogenes, by contrast, are ‘dominant’, because mutation of just one allele is sufficient for an effect to be seen. It’s perhaps worth noting a possible confusion of terms herein that, although the RB1 gene behaves in a recessive manner, within families’ germinal mutations in RB1 are inherited in an autosomal dominant manner, that is, one copy of the altered gene increases the risk of cancer.