Abstract
BackgroundA number of clinico-pathological criteria and molecular profiles have been used to stratify patients into high- and low-risk groups. Currently, there are still no effective methods to determine which patients harbor micrometastatic disease after standard breast cancer therapy and who will eventually develop local or distant recurrence. The purpose of our study was to identify circulating DNA methylation changes that can be used for prediction of metastatic breast cancer (MBC).ResultsDifferential methylation analysis revealed ~5.0 × 106 differentially methylated CpG loci in MBC compared with healthy individuals (H) or disease-free survivors (DFS). In contrast, there was a strong degree of similarity between H and DFS. Overall, MBC demonstrated global hypomethylation and focal CpG island (CPGI) hypermethylation. Data analysis identified 21 novel hotspots, within CpG islands, that differed most dramatically in MBC compared with H or DFS.ConclusionsThis unbiased analysis of cell-free (cf) DNA identified 21 DNA hypermethylation hotspots associated with MBC and demonstrated the ability to distinguish tumor-specific changes from normal-derived signals at the whole-genome level. This signature is a potential blood-based biomarker that could be advantageous at the time of surgery and/or after the completion of chemotherapy to indicate patients with micrometastatic disease who are at a high risk of recurrence and who could benefit from additional therapy.Electronic supplementary materialThe online version of this article (doi:10.1186/s13148-015-0135-8) contains supplementary material, which is available to authorized users.
Highlights
A number of clinico-pathological criteria and molecular profiles have been used to stratify patients into high- and low-risk groups
Clinical characteristics of samples We characterized the plasma methylome of metastatic breast cancer (MBC) by paired-end whole-genome bisulfite sequencing (WGBS) to identify differentially methylated regions that were uniquely found in circulating cell-free DNA (cfDNA) of a pool of 40 MBC when compared with a pool of 40 H and a pool of 40 disease-free survivors (DFS)
WGBS demonstrated global hypomethylation and focal hypermethylation in cfDNA of MBC compared with H and DFS, which had a high degree of similarity To assess the similarity of each sample group to the others, we used methylKit (25) to compute pair-wise Pearson correlation coefficients, hierarchical clustering (Ward’s method, correlation distance metric), and Principal Component Analysis (PCA) on % CpG methylation profiles. These analyses demonstrated that the H cohort closely resembled DFS, evidenced by Pearson correlation coefficient (0.83) and close proximity by hierarchical clustering and PCA (Fig. 1)
Summary
A number of clinico-pathological criteria and molecular profiles have been used to stratify patients into high- and low-risk groups. A number of clinico-pathological criteria have been established as breast cancer prognostic markers to determine risk of recurrence and stratify patients into highand low-risk groups. Despite the huge quantity of information gleaned from these gene signatures, none can precisely predict the clinical course of an individual and rely on the presence of tissue at a single time point. They are not able to monitor a patient’s risk status after completion of therapy due to residual disease. Even with the clinico-pathological features, there are patients deemed high-risk who do very well with standard therapy and never experience a recurrence and patients with low-risk profiles who still die of breast cancer. We report a 21-gene DNA hypermethylation signature, detectable in the circulation of MBC patients, which maybe useful in the pre-macrometastatic setting to indicate patients at a high risk of recurrence
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
