Strange Landscape of Cancer Cells

Compared with the genome of a normal cell, that of a cancer cell has sections that are missing, repeated or scrambled. Although some of the mutations are probably driving the cancer, others may have little relevance to the cancer’s progression. To understand which mutations are important in the development and spread of cancer, a multi-national team including researchers from the Broad Institute and the Dana-Farber Cancer Institute have studied over 3000 primary human cancer specimens representing more than 26 cancer types, including lung, prostate, breast, ovarian, colon, oesophageal, liver, brain, and blood cancers.

Classically, cancers have been classified (and treated) by their tissue of origin but, more recently, it has been realised that oncogenes are not restricted to one type of cancer. In the new study, the team found that amplifications and deletions of regions of the genome, known as somatic copy-number alterations or SCNAs, are not evenly distributed across the cancer genome but are concentrated in less gene-rich regions where they may be better tolerated. As with other mutations seen in cancer cells, many SCNAs may not play an important role in cancer growth but some will identify genes that help to promote and maintain cancers.

The most common SCNAs were found to be either relatively long (the length of an entire chromosome or a single chromosome arm) or short (about 1.8 million base pairs). More than 150 regions of short, or focal, SCNA – which because of their compact size are more likely to pinpoint important cancer-linked genes – were found to be altered at significant frequency across several cancer types. Of these, 122 did not coincide with genes already known to be amplified or deleted in cancer but the team also found copy number changes in BCL2 family genes – which are already linked to cancer and are the target of some cancer drugs – in over half of all samples. Most focal SCNAs were not unique to one type of cancer suggesting that common genetic mutations underlie different types of tumour, a finding which may lead to improved diagnosis and more targeted treatments for cancer.

The study is published in the journal Nature and the researchers have also built an online tool, the Tumorscape portal, to allow other researchers to freely access their data.