NEW YORK – An international team led by investigators at Cedars-Sinai Medical Center and the University of Cambridge has tracked down a handful of new genetic loci linked to specific ovarian cancer subtypes, while highlighting other candidate risk genes, causal variants, and apparent regulatory roles for these variants.
"[W]e describe a comprehensive genetic association study for ovarian cancer," co-senior and co-corresponding authors Paul Pharoah, a cancer genetic epidemiology researcher at Cambridge, and Simon Gayther, molecular epidemiologist at Cedars-Sinai, and their colleagues wrote in the American Journal of Human Genetics on Wednesday.
The work "expands upon previous findings indicating that different genetic alleles confer risks to different ovarian cancer histotypes, but also that several risk regions confer risk to two or more histotypes," they explained, "suggesting there are similarities in the underlying biology driving the disease spectrum of ovarian cancer."
For their study, the researchers began by performing a genome-wide association study on 25,981 individuals with epithelial ovarian cancer and more than 105,700 unaffected controls, searching for genetic loci with ties to specific histological subtypes of the disease. They then performed a series of fine-mapping, transcriptome-wide association study, colocalization, and interactome analyses to tease out "credible causal variants" (CCVs) and candidate risk genes, along with their apparent functional consequences.
The study participants included individuals with the five main invasive epithelial ovarian cancer histotypes, including 13,609 high-grade serous ovarian cancer (HGSOC) cases; 2,749 low-grade serous ovarian cancer (LGSOC) cases; 2,587 individuals with mucinous ovarian cancer (MOC); 2,877 endometroid ovarian cancer (ENOC) cases; and 1,427 individuals with clear-cell carcinoma ovarian cancer (CCOC). Another 753 individuals had less common histotypes of the disease.
Based on SNP profiles spanning 470,825 directly genotyped variants and nearly 10.2 million imputed variants, the investigators flagged five loci with previously unappreciated ties to specific ovarian cancer histotypes, while shoring up associations at more than two dozen loci previously implicated in ovarian cancer. Half a dozen of the ovarian cancer-related regions contained additional association signals that turned up in their conditional analyses.
"Our findings confirm previous reports that genetic risk factors reflect the underlying disease heterogeneity," the authors reported.
In all, the researchers uncovered 16 new risk associations — the five from their primary analysis and 11 from their subsequent conditional analyses — and noted that seven were strongly associated with HGSOC, two with LGSOC, and three with MOC. Four additional loci were linked to all of the histological subtypes, except MOC.
The team subsequently highlighted more than 4,000 CCVs and 86 suspected susceptibility genes at the ovarian cancer-associated loci by bringing together chromatin interaction, epigenomic, TWAS, and other data to explore the consequences of the ovarian cancer-related alterations.
Because more than 99 percent of the CCVs identified turned up in noncoding portions of the genome, for example, the researchers speculated that regulatory elements such as promoters, enhancers, and insulators that have functions in tissues affected by ovarian cancer may play an outsized role in the disease.
"By combining genetic, TWAS, and epigenomic profiling with HiChIP-seq interactome analysis, we … show the likely functional mechanisms underlying several [epithelial ovarian cancer] risk loci, representing hitherto unrecognized aspects of biology in the development of different [epithelial ovarian cancer] histotypes," the authors reported, noting that the broader analytical strategy used "represents a functional framework for understanding the biological mechanisms underlying common low penetrance risk regions for ovarian and other cancers in the future."