Mutations affecting cancer risk exist in the 90% of the genome regarded as “junk DNA,” which typically does not code for proteins. A recent GWAS-based study by Dana-Farber researchers found, for the first time, that such mutations can influence cancer risk through an epigenetic effect. The study also opens the way to therapies that—by disrupting that mechanism—can reduce at-risk people’s likelihood of developing certain cancers.
The new research appeared in Nature Genetics, and investigators found that in the overwhelming number of cases examined, non-coding mutations have an epigenetic effect. That, in turn, affects how open those locations are to binding to other sections of DNA or certain proteins, all of which can influence the activity of genes involved in cancer.
“Studies have identified an enormous number of mutations across the genome that are potentially involved in cancer,” says Alexander Gusev, PhD, of Dana-Farber, the Eli and Edythe L. Broad Institute and Brigham and Women’s Hospital, who co-authored the paper with Dana-Farber’s Dennis Grishin, PhD. “The challenge has been understanding the biology by which these variations increase cancer risk. Our study has uncovered an important part of that biology.”
Of the many cancer-causing germline mutations, such as recent findings in testicular cancer, only a small percentage are in coding portions of the genome. Breast cancer is another example. “More than 300 mutations have been identified that are associated with an increased risk of the disease,” Gusev states. “Less than 10% of them are actually within genes. The rest are in ‘desert’ regions, and it hasn’t been clear how they influence disease risk.”
A New Approach
Typically, researchers approach this by doing “colocalization studies,” looking for overlap between GWAS data, showing mutations in a specific type of cancer, and data on another genomic feature of that cancer type, such as an abnormally high or low level of activity in certain genes. But colocalization studies have so far turned up very few such correspondences.
Instead of beginning with the premise that non-coding mutations might influence gene expression, Gusev and Grishin asked whether these mutations affect the coiling of DNA in their immediate vicinity.
“We hypothesized that if you look at the effect of these mutations on local epigenetics—specifically, whether they caused nearby DNA to be wound more tightly or loosely—we’d be able to detect changes that wouldn’t be evident in expression-based studies,” Gusev relates.
To do that, they took GWAS data on cancer-related mutations and data on epigenetic changes in seven common types of cancer and examined whether—and where—they intersected.
They used 406 cancer ATAC-Seq samples across 23 cancer types to identify 7,262 germline allele-specific “accessibility QTLs (as-aQTLs)”. Cancer as-aQTLs, they reported, had stronger enrichment for cancer risk heritability (up to 145 fold) than any other functional annotation across seven cancer GWAS. To connect as-aQTLs to putative risk mechanisms, they also introduced the “regulome-wide associations study (RWAS).” The RWAS helped them identify genetically associated accessible peaks at >70% of known breast and prostate loci and discovered new risk loci in all examined cancer types.
These results were in stark contrast to those from colocalization studies. “We found that whereas most non-coding mutations don’t have an effect on gene expression, most of them do have an impact on local epigenetic regulation,” Gusev states. “We now have a basic biological explanation of how the vast majority of cancer-risk mutations are potentially linked to cancer, whereas previously no such mechanism was known.”
Integrating as-aQTL discovery, motif analysis and RWAS identified candidate causal regulatory elements and their probable upstream regulators. Our work establishes cancer as-aQTLs and RWAS analysis as powerful tools to study the genetic architecture of cancer risk.