NEW YORK (GenomeWeb) – An Institute of Cancer Research-led team has identified potential target and survival-related genes in breast cancer using chromatin interaction profiles to map parts of the genome previously implicated in risk of the disease.
As they reported online today in Nature Communications, the researchers turned to a long-range interaction profiling technique called Capture Hi-C (CHi-C) to investigate dozens of regions previously implicated in breast cancer risk. Based on CHi-C data for four breast cancer cell lines with or without enhanced estrogen receptor (ER) activity, a cancer-free breast epithelial cell line, and a cell line from another tissue type, they focused in on 110 potential target genes at 33 of the risk loci.
By adding in related risk SNP data, RNA sequence profiles, and somatic mutation data reported previously, the team looked at overlap between these proposed targets, expression quantitative trait loci (eQTL), and genes prone to mutation in breast cancer tumors. With the help of gene expression and outcome data from the Metabric breast cancer cohort, meanwhile, the authors highlighted 32 genes with ties to breast cancer survival time.
"Identifying these new genes will help us to understand in much greater detail the genetics of breast cancer risk," senior author Olivia Fletcher, a genetic epidemiology and molecular biology team leader at ICR, said in a statement. "Ultimately, our study could pave the way for new genetic tests to predict a woman's risk, or new types of targeted treatment."
Around 100 risk loci have been described through prior genome-wide association studies, mapping analyses, and the like, the team noted. Even so, the causal genes and variants behind risky genetic loci are not always known, and the distinct factors contributing to various breast cancer subtypes are largely unknown.
To get a better look at 63 of the previously reported breast cancer risk loci, the researchers relied on CHi-C to identify related interactions in two ER-positive breast cancer cell lines, two ER-negative breast cancer lines, a normal breast epithelial cell line, and a lymphoblastoid cell line, used as a non-breast tissue control.
"Our study took the high-level maps of breast cancer risk regions and used them to pull out specific genes that seem to be associated with the disease," Fletcher explained. "We studied how DNA forms loops to allow physical interactions between a DNA sequence in one part of the genome and a risk gene in another."
After using Illumina HiSeq2000 instruments to sequence target-enriched Hi-C libraries, the team tallied the CHi-C interaction peaks at each risk locus. A dozen risk loci lacked discernable interaction peaks, leaving interaction peaks at 51 loci for further consideration within and across the six cell lines.
The researchers uncovered potential gene targets at 33 of the breast cancer-associated sites, spanning 94 protein-coding and 16 non-coding RNA gene targets neighboring the risk loci or falling more far afield. They subsequently considered the target genes alongside eQTLs (informed by RNA sequence data from the Cancer Genome Atlas) and somatic mutation profiles identified in hundreds of breast cancer genome sequences, which supported the notion that CHi-C can help unearth authentic target genes.
The team cautioned that far more data is needed to evaluate the list of target genes proposed from the current analyses, but reasoned that their general long-range interaction-based approach may serve as a strategy for "short-listing candidates for follow-up studies.
"[W]e would argue that a high-throughput CHi-C analysis can contribute to on-going efforts to functionally annotate GWAS risk loci and that CHi-C target genes that are supported by additional data sources are strong candidates for in-depth functional follow-up studies," the authors concluded.