NEW YORK (GenomeWeb) – Researchers from University College London reported in Genome Medicine this week that a DNA methylation signature they've uncovered could predict the risk of a woman developing breast cancer.
Although they developed the signature by examining the blood of women who are BRCA1 carriers, it appeared to also be able to predict risk of non-hereditary breast cancer.
"We identified an epigenetic signature in women with a mutated BRCA1 gene that was linked to increased cancer risk and lower survival rates," Martin Widschwendter, the study's lead author and head of UCL's Department of Women's Cancer, said in a statement. "Surprisingly, we found the same signature in large cohorts of women without the BRCA1 mutation, and it was able to predict breast cancer risk several years before diagnosis."
Breast cancer is the most common type of cancer, excluding some skin cancers, that affects women. While mutations in BRCA1 and BRCA2 affect a person's risk of cancer, those familial mutations cause less than 10 percent of breast cancers, leaving a large percentage that cannot be easily predicted.
Because aberrant DNA methylation is common in cancer, especially at promoters of stem cell differentiation genes, the researchers suspected that an epigenetic signature might be able to predict cancer risk.
To develop their signature, the UCL researchers examined the DNA methylation profiles of white blood cell samples from 72 women with BCRA1 mutations and 72 women without either BRCA1 or BRCA2 mutations using the Illumina Infinium Human Methylation27 Bead Chip.
Through a multivariate analysis, they ranked some 27,500 CpG islands based on the association between their DNA methylation profile and mutation status. Following an elastic net framework approach, the researchers whittled that number down to a signature of 1,829 CpGs — 1,074 hypermethylated sites and 755 hypomethylated sites — that reflected breast cancer risk.
Those hypermethylated sites, the researchers noted, are enriched for stem cell polycomb group target genes, which are involved in stem cell differentiation.
They validated this epigenetic signature in two cohorts of women, 152 women from the MRC National Survey of Health and Development cohort and 241 from the United Kingdom Collaborative Trial of Ovarian Cancer Screening cohort.
For the NSHD cohort, the researchers' signature could predict the development of breast cancer, as well as the development of invasive non-breast cancer. Further, they noted that their signature from peripheral blood could predict disease, while profiles generated from buccal swabs at the same time could not, indicating that the signature is tissue specific.
Within the full 202,600-member UKCTOCS cohort, the researchers also performed a nested case-control study to examine not only whether the signature could predict disease risk in that group, but also whether it could predict mortality.
They found that women with higher-than-average methylation signature scores in this cohort were nearly 8.5-fold more likely to die from breast cancer than women with lower-than-average scores.
Neither epidemiological breast cancer risk factors nor hormones, the researchers added, were associated with their methylation signature.
Remarkably, they reported that their DNA methylation signature could also predict incidence of and death from breast cancer in women in the UKCTOCS cohort who did not have any family history of the disease.
"The data is encouraging since it shows the potential of a blood-based epigenetic test to identify breast cancer risk in women without known predisposing genetic mutations," Widschwendter said.
Though it isn't yet clear how the epigenetic signature reflects cancer risk, the researchers said it is consistent with the notion that epigenetic changes to immune cells may hinder the immune system's ability to prevent the development of cancer.
The UCL researchers are now studying whether this signature merely reflects breast cancer risk or whether it is involved in breast cancer progression.