NEW YORK – A UK team led by investigators at the University of Cambridge has focused in on rare coding variants contributing to heritable cancer susceptibility risk. In a large-scale UK Biobank study, they identified protein-truncating germline variants with ties to cancer risk as well as clusters of cancer types with overlapping genetic contributors.
"[W]e looked systematically at all coding genes and could model the contribution of all genes," Douglas Easton, a researcher at the University of Cambridge Centre for Cancer Genetic Epidemiology, said in an email.
For a paper published in the American Journal of Human Genetics on Tuesday, Easton and colleagues used analyzed exome sequence data for about 450,000 UK Biobank participants — coupled with national cancer registration data, diagnostic codes, and some self-reported cancer diagnostic data — to search for rare, cancer risk-related variants and genes linked to 11 cancer types.
"We have known for some years that carriers of rare variants in some genes, like BRCA1 and BRCA2, are at a high risk of cancer," Easton explained. "Many of these genes are involved in DNA repair. However, most genes have not been evaluated systematically, so it was not clear how much contribution was made by coding variants in the rest of the genome."
Using sequence data spanning more than 15,000 genes, the team uncovered several with apparent ties to individuals' germline risk of pancreatic, endometrial, ovarian, and other cancer types.
"We previously described a method for evaluating the contribution of the gene-wise burden of rare coding variants to cancer heritability that fits models to the burden effect sizes using an empirical Bayesian approach," the authors explained. "This approach can be implemented using gene burden summary statistics and is not computationally intensive."
Though the investigators cautioned that "[m]any of these associations were … based on low carrier counts," they highlighted an exome-wide significant, gene-level association between melanoma risk and the MED9 gene. Their analyses also led to an association between breast and ovarian cancer and rare variants in the DNA repair gene NHEJ1.
Protein-truncating variants in the germline appeared to have the most pronounced impact on the overall heritability of ovarian cancer, an estimated 46 percent.
Even so, Easton noted that "for the 11 cancers we looked at, protein-truncating coding variants could only explain a minority of the heritability, and much of these could be explained by the genes previously known," suggesting that 'missing' heritability for cancer susceptibility "must be largely due to variants in the noncoding genomes, presumably involved in gene regulation, rather than in genes themselves."
More broadly, the authors explained, their cancer pair-focused analyses revealed susceptibility gene overlap between several pairs of cancer types.
When they used an extended version of their analytical approach to search for contributors to cancer pairs, for example, they saw signs that rare germline variants in overlapping gene sets — particularly genes involved in DNA repair and tumor-suppressor genes — may contribute to heritable risk for breast, ovarian, prostate, and pancreatic cancer.
Though additional research is needed to better understand why cancers with shared genetic contributors develop in specific tissue types, Easton explained, the results so far may provide insights for improving cancer risk prediction and management tools, while revealing common mechanisms across multiple cancer types.
"We were able to define lists of genes that will be worth following up in still larger sequencing studies, to confirm associations and define the cancer risks," Easton noted. Members of the team are conducting a large follow-up study involving about 100,000 samples as part of the Breast Cancer Association Consortium, he added, in order to confirm some of the new genes and determine precise risk estimates for different breast cancer subtypes.