NEW YORK (GenomeWeb) – An international team led by investigators at Vanderbilt University and the QIMR Berghofer Medical Research Institute in Australia has used a transcriptome-wide association approach to track down four dozen genes with apparent ties to breast cancer.
Using genetically predicted gene expression profiles for nearly 123,000 individuals with breast cancer and almost 106,000 without, the researchers searched for loci and genes contributing to breast cancer — a disease that has a great deal of unexplained heritability despite being the subject of genome-wide association studies going back more than a decade.
Based on available genotype-expression data from the Genotype-Tissue Expression (GTEx) project and the Cancer Genome Atlas (TCGA), the team got a glimpse at expression profiles for thousands of genes in the cases and controls, uncovering breast cancer associations for 34 genes previously implicated in breast cancer and 14 new candidate risk genes. The results appeared online today in Nature Genetics.
"[O]ur study has identified multiple gene candidates that can be further functionally characterized," corresponding and co-senior author Georgia Chenevix-Trench, a cancer researcher at QIMR Berghofer, and co-authors wrote.
Following from a TWAS published in PLOS Genetics last year that turned up five genes with significant ties to breast cancer from tens of thousands of cases and controls, the researchers performed an even larger TWAS based on 122,977 breast cancer cases and 105,874 unaffected individuals of European ancestry.
The team used insights gleaned from transcriptome and genotype profiles for 67 GTEx participants, together with expression insights for tumor and normal samples assessed for TCGA, to predict individuals' gene expression from their genotypes, namely SNP summary statistics.
With the help of a MetaXcan association analysis approach described previously, the researchers searched 8,597 genes for potential ties to breast cancer, uncovering 48 genes with predicted expression profiles that coincided with breast cancer risk. The set encompassed 14 genes at 11 loci not implicated in the disease in past studies.
Then team went on to incorporate expression quantitative trait loci, GWAS target prediction insights, gene pathway clues, and in vitro functional assays in an effort to get a better look at some of the biological processes that may go awry during breast cancer development. This included processes involving genes flagged by variants identified in GWAS.
The pathway analysis highlighted the potential importance of apoptosis- and immune system-related pathways, for example, while systematic knockdown of candidate genes in breast cells led to declines in cell proliferation and colony formation that was not detected after lowering levels of unrelated control genes.
"The silencing experiments we performed suggest that many of the genes identified are likely to mediate risk of breast cancer by affecting proliferation or [colony-forming efficiency], two hallmarks of cancer," the authors concluded, noting that "[f]urther investigation of genes identified in our study will provide additional insight into the biology and genetics of breast cancer."