http://www.cell.com/cell-reports/pdf/S2 ... 0387-8.pdf
Abstract
APOBEC3B cytosine deaminase activity has recently emerged as a significant mutagenic factor in human cancer. APOBEC activity is induced in virally infected cells, and APOBEC signature mutations occur at high frequency in cervical cancers (CESC), over 99% of which are caused by human papillomavirus (HPV). We tested whether APOBEC-mediated mutagenesis is particularly important in HPV-associated tumors by comparing the exomes of HPV+ and HPV− head and neck squamous cell carcinomas (HNSCCs) sequenced by The Cancer Genome Atlas project. As expected, HPV− HNSCC displays a smoking-associated mutational signature, whereas our data suggest that reduced exposure to exogenous carcinogens in HPV+ HNSCC creates a selective pressure that favors emergence of tumors with APOBEC-mediated driver mutations. Finally, we provide evidence that APOBEC activity is responsible for the generation of helical domain hot spot mutations in the PIK3CA gene across multiple cancers. Our findings implicate APOBEC activity as a key driver of PIK3CA mutagenesis and HPV-induced transformation.
We (me + other authors) worked out that in HPV-driven cancers, which are mostly from the head and neck and the uterine cervix the mutations you see are characteristic of proteins from the APOBEC family of proteins. These proteins go round mutating viral DNA, changing cytosines to uracils which then get repaired to form thymine or guanine but by accident end up mutating cellular DNA, generating the necessary genetic changes to convert a cell that has acquired integrated copies of the HPV genome but hasn't turned cancerous yet to fully cancerous cells.
Even cooler, one of the genes that is most often mutated across cancer types, called PIK3CA encodes a component of a complex called PI3 kinase and there are two regions in the protein that are mutated recurrently - one in a "helical" domain and the other in the kinase domain. We found that in the HPV+ tumours there was a strong preference for the former instead of the latter, even if it has been shown both mutant versions are equally capable of making cells cancerous. The secret, it turns out, is that the helical domain mutations have a DNA sequence that is recognised and edited by the candidate APOBEC enzymes we were looking at while the kinase ones do not. We then looked across tumours from different tissues and showed that strong skews towards helical hotspot mutations were associated with greater activity of APOBEC proteins, which we could measure by looking for characteristic mutational patterns in sequences of all protein coding genes from The Cancer Genome Atlas (TCGA).
We therefore ended up showing that mutational processes can actively determine what genes are mutated and how, and this adds to our fundamental understanding of mutation and selection in the evolution of cancers.
Additional background and descriptions of how we got to our findings is here on my blog http://exploreable.wordpress.com/2014/0 ... mutations/ and the journal did a little Q&A with the two lead authors (me and the other informatician) and the senior author http://news.cell.com/cellreports/cell-r ... colleagues . Oversimplified press release is here http://www.ucl.ac.uk/news/news-articles ... -infection
