NEW YORK (GenomeWeb) – Researchers at the Albert Einstein College of Medicine in New York have developed a functional genomic assay to predict breast cancer risk.
A portion of families at high risk for breast cancer harbor mutations in BRCA1 and BRCA2 as well as in other genes that have a role in repairing double-stranded DNA breaks. But as panel-based sequencing tests often find variants of unknown significance in such genes, Einstein's Harry Ostrer and his colleagues sought to develop an alternative way to gauge dysfunction in that pathway and predict breast cancer risk.
As they reported today in Genetics in Medicine, the researchers used flow-variant approaches (FVAs) to assess the extent of damage to the double-stranded repair pathways by monitoring whether the BRCA1 and BRCA2 proteins localize to the nucleus after damage occurs and whether there's a change in p53 phosphorylation status. A classification score based on these three FVAs was 95 percent accurate, the researchers reported.
"A direct functional test at the protein level queries whether key biological functions — BRCA1 and BRCA2 nuclear localization and p53 phosphorylation — are being altered, " Ostrer and his colleagues wrote in their paper. "As shown in this study, tests based on multiple, direct, functional analyses appear to be more sensitive and specific for identifying genetic risks and identify those at high risk, even when a mutation cannot be identified by sequencing."
Ostrer and his colleagues reported in Human Molecular Genetics in 2015 that they'd developed flow cytometry-based FVAs that could determine the effect of heterozygous mutations within the double-strand break repair pathway. In particular, they measured the response of cultured or circulating cells harboring those mutations after being exposed to radiomimetic agents. They reported that BRCA1, BRCA2, FANCC, and NBS1 mutation carriers all had reduced BRCA1 nuclear localization and a reduced phosphorylated p53 to total p53 ratio after radiomimetic exposure.
In their new paper, the researchers relied on the same approach to assess various mutations within BRCA1, BRCA2, and ATM, as well as single mutations within FANCC, FANCF, FANCD2, and NBS1 in lymphoblastoid cell lines. Again, they found that BRCA1 mutant cells had reduced BRCA1 and BRCA1 nuclear localization and added that cell lines with other mutations exhibited similar effects.
Based on these assays, the researchers developed a multivariate classification score. Individually, they reported that the BRCA1 nuclear localization assay, the BRCA2 nuclear localization assay, and the phosphorylated p53 to total p53 ratio assay had sensitivities ranging between 83 percent and 91 percent and specificities between 86 percent and 89 percent. But by combining the three assays into one score, the sensitivity, specificity, and accuracy could be boosted to 91 percent, 100 percent, and 95 percent, respectively.
The researchers tested their classification score on B cells collected from a cohort of 29 women at high risk of breast cancer. Of these, 20 had breast or ovarian cancer and affected relatives — the BOC-positive group — and nine did not have cancer but had affected relatives — the BOC-negative group. All of the women had previously tested negative for BRCA1 or BRCA2 mutations.
The researchers found that their classifier could distinguish the BOC-positive and BOC-negative groups.
Within the BOC-negative group, the researchers noted two distinct clusters. They suggested that these clusters might represent women with varying degrees of disease risk.
Ostrer and his colleagues also conducted whole-genome sequencing for about a dozen of the members of the BOC-positive group. Through this, they uncovered 59 variants, but the only one expected to affect splice sites or lead to a truncated protein was a benign variant.
Fifteen nonsynonymous missense variants were also classified as variants of unknown significance. This suggested to the researchers that their test could uncover breast cancer risk even when sequencing could not.