NEW YORK (GenomeWeb News) – The latest in a series of papers from The Cancer Genome Atlas' Pan-Cancer Initiative is highlighting more than 125 genes that appear to contribute to cancer development or progression.
As they reported online today in Nature, researchers from Washington University's Genome Institute and elsewhere used TCGA data from nearly 3,300 tumors to look for genes recurrently altered by point mutations or small insertions and deletions. The search unearthed 127 genes that appear prone to mutation across the 12 cancer types tested.
That significantly mutated gene set included well-established cancer players such as TP53, as well as genes with cancer roles that are just beginning to be revealed. As in previous pan-cancer studies, the investigators uncovered clusters of genetic glitches characterizing tumors that originated in one to several tissue types.
"Many oncologists and scientists have wondered whether it's possible to come up with a complete list of cancer genes responsible for all human cancers," senior author Li Ding, a researcher affiliated with Washington University's Genome Institute and the Siteman Cancer Center, said in a statement. "I think we're getting closer to that."
Together with information on the contributions that such genes make across cancer types and their potential role in treatment outcomes, such a list of core cancer contributors is expected to prove useful for designing new methods for diagnosing and treating cancer, she and her colleagues explained.
"Ultimately, these data and their associations with different clinical features and sub-types should contribute to the formulation of a reference candidate gene panel for all tumor types that could be helpful for prognosis at various clinical time points," they wrote.
Last month, members of TCGA's Pan-Cancer Initiative began publishing a spate of studies on everything from oncogenic signatures shared across cancers to recurrent copy number changes characterizing multiple tumor types.
For the new analysis, researchers used exome sequence profiles and other data types for 3,281 tumors to define mutation types and frequencies within the 12 cancer types considered. In the process they were able to narrow in on 127 genes that were significantly mutated across the complete tumor set.
Some of those significantly mutated genes fell in pathways with long-documented roles in cancer — such as the MAP kinase and Wnt/beta-catenin signaling pathways — while others point to processes that are recently recognized as potential cancer contributors.
Consistent with past findings from smaller sample sets, the researchers saw that some significantly mutated genes were prone to mutation in the same tumors, while others occurred in a more mutually exclusive manner — patterns that they were able to explore in both cancer type-specific and cross-cancer contexts.
The extent to which members of this gene set were mutated varied depending on the tumor and the cancer type considered, the group noted. Even so, more than 3,000 of the tumors contained mutations involving at least one significantly mutated gene and most harbored alterations affecting between two and six of them.
Along with an examination of the frequency and distribution of such somatic mutations, the team's analysis included a cross-cancer look at mutations that might offer therapeutic and/or prognostic insights.
In particular, some 42 percent of those included in the study harbored point mutations or indels affecting TP53. Mutations in that well-known cancer gene turned up in a large swath of the serous endometrial carcinomas tested and in most of the serous ovarian tumors. They were also quite common in some forms of breast cancer and in some cancers lacking mutations to most other significantly mutated genes, the researchers noted.
For at least some of the cancer types tested, such TP53 alterations signaled poor outcomes. Likewise, glitches in genes such as BAP1, FBXW7, and KDM5C appeared to have poor prognostic implications for individuals with kidney renal cell carcinoma and other cancer types.
Other significantly mutated genes corresponded with somewhat better than usual outcomes, meanwhile. The team generally saw more favorable outcomes in individuals with ovarian cancer whose tumors contained BRCA2 glitches, for example, or glioblastoma patients with IDH1-mutated tumors.
When they considered ways in which significantly mutated genes grouped together across the tumor set, investigators saw clusters characterizing tumors within and across cancers from the same tissue types.
For example, results from the study hint that tumors from many cancer types tend to harbor mutations affecting chromatin remodeling or histone modification genes. In contrast, though, glitches in transcription factors and genes with more tissue-specific roles occurred in a more limited range of cancer types.
Using information on variant allele patterns in the significantly mutated gene set, meanwhile, the group was also able to start exploring tumor heterogeneity and delving into the sequence with which mutations appear in significantly mutated genes.
"Although a common set of driver mutations exists in each cancer type, the combination of drivers within a cancer type and their distribution within the founding clone and sub-clones varies for individual patients," the study authors wrote. "This suggests that knowing the clonal architecture of each patient's tumor will be crucial for optimizing their treatment."