It is now forty years since the publication of a seminal analysis of retinoblastoma, which suggested that hereditary cases of this paediatric eye tumour probably arise following a single somatic mutation. This led to the development of what has been called the "two-hit hypothesis" of the role of tumour suppressor genes in hereditary tumours.

In children with hereditary retinoblastoma, there is a germline mutation of one allele of the tumour suppressor gene involved (now known as RB1). Since there is a strong chance that a somatic mutation to the other allele will arise during development, the tumour appears to be inherited in a dominant manner.

In contrast, spontaneous cases of retinoblastoma may arise in rare instances where both alleles are mutated during development. Thus, either an inherited or a somatic mutation of the first allele causes cancer susceptibility; mutation of the second allele causes cancer.

Pier Paolo Pandolfi and Alice Berger from Harvard Medical School, Boston, Massachusetts, USA and Alfred Knudson from Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA have now reviewed how our knowledge of the role of tumour suppressor genes in cancer has developed since 1971.

They survey the full range of inactivating mutations now known to be involved in cancer development, and propose a model of tumour suppressor gene function that could explain differences in their function.

Knudson's analysis of retinoblastoma not only led to the establishment of the two-hit model of tumour suppression, but also provided proof of principle that tumour suppressor genes could be identified from the study of chromosomal deletions and genetic linkage.

This strategy led to the identification of now well-known and widely studied tumour suppressor genes including p53 and the breast cancer susceptibility genes BRCA1/2.

It became clear fairly early on that the two-hit model would only explain a few rare paediatric cancer syndromes. In most cases, where cancer arises later in life, mutations in at least four different genes are necessary for a cancer to develop.

In these cases, the two-step model relates to the inactivation of one particular, possibly rate-determining tumour suppressor gene, but other genetic changes are necessary for tumour initiation.

Furthermore, later investigations into recurrent chromosomal deletions in cancer led to the discovery that an aberration of one allele of a gene would often be found consistently in a particular cancer, but this would not be associated – as would seem necessary by the two-gene hypothesis – with any functional changes to the other allele. This led to the concept of haploinsufficiency; that is, to the loss of one allele and consequent depletion of the associated protein being sufficient for cancer development.

It is now known that the role a gene plays in cancer development may depend on its genetic background and the tissue in which it is expressed, and also that the relationship between the degree to which a gene product is expressed and its ability to function as a tumour suppressor may be more complex that understood by haploinsufficiency alone.

The gene PTEN, which encodes a protein phosphatase, is a good example of this phenomenon.

In tissues with wild type p53, PTEN haploinsufficiency of is actually more tumourigenic than its complete loss, because absence of PTEN expression triggers a p53-dependent fail-safe mechanism. This phenomenon is known as quasi-haploinsufficiency. In contrast, where p53 expression is absent, complete loss of PTEN expression will trigger a fatal cancer.

Pandolfi and his co-workers discuss many other examples of this phenomenon and propose that the simple, classic two-step model for tumour suppression be replaced by a continuum model. In this, tumour suppression activity is related to protein expression level rather than simply to functional gene copy number.

In general terms, the potential for malignancy will rise as the expression level of a tumour suppressor protein drops, although this can be reversed at low expression levels in quasi-haploinsufficiency.

The same paradigm can be applied to oncogenes, with malignancy increasing with expression level. This could have important implications for therapy; as one example, a tumour with a haploinsufficient tumour suppressor gene may respond differently to treatment from a similar tumour in which that gene is fully deleted.

Source: Berger, A.H., Knudson, A.G. and Pandolfi, P.P. (2011). A continuum model for tumour suppression. Nature 476, 163-169. doi:10.1038/nature10275

Knudson, A. G. Jr. (1971) Mutation and cancer: statistical study of retinoblastoma. Proc. Natl Acad. Sci. USA 68, 820–823