New perspectives on cancer evolution from genome sequencing

Researchers at the MRC Human Genetics Unit led by Professor Colin Semple have just published one of the largest studies of tumour WGS so far: August 2016

Cancer is a disease of the genome, involving widespread disruptions (mutations) to human DNA accumulating as the cells within a tumour evolve. These mutations have been studied for many years to provide insights into how tumours occur and progress. However up until recently technological and financial limitations meant that it was only possible to examine the DNA within genes, encompassing only around 3% of the total genome sequence. The other 97% of the genome is not well studied, but recent whole genome sequencing (WGS) datasets have emerged worldwide, based upon sequencing hundreds or thousands of tumours from cancer patients.

Researchers at the MRC Human Genetics Unit led by Professor Colin Semple have just published one of the largest studies of tumour WGS so far, including 11 different cancer types(Kaiser et al, 2016, PLOS Genet 12:(8):e1006207). They studied the patterns of mutations accumulating across cancer genomes using carefully controlled comparisons in a novel computational analysis, paying close attention to functional regulatory sites. These sites are short DNA segments that act as molecular switches, and are known to turn the expression of genes on or off in a variety of human cell types.

They discovered strikingly high rates of mutation at functional regulatory sites across different cancers, relative to matched control sequences. This excess of mutations is predicted to disrupts the binding sites of most known transcription factors, which bind to these sites and activate genes. Particular factors, such as CTCF, suffer unusually high loads of mutation suggesting widespread impacts on the physical organisation of chromosomes. These unusual patterns could conceivably be generated by selection during tumour evolution, as the tumour cells adapt to invade normal tissue more effectively.

However, Kaiser et al show that the patterns are likely to simply reflect mutational bias, with functional regulatory sites being inherently prone to accumulating large numbers of mutations. Why this bias should exist remains mysterious, and is the focus of ongoing work.

The study suggests that tumours accidentally but unavoidably alter the regulation of many, and perhaps most, human genes as they develop over time. Understanding this progression better may allow us to develop new approaches to detecting earlier and later stages of cancers. The work was made possible by core MRC funding to Semple and his colleagues.