DNA methylation signature to identify treatment response in triple negative breast cancer.

1079 Background: The triple negative breast cancer (TNBC) subtype is an aggressive phenotype with scarce treatment alternatives. Chemotherapy (CT) is only effective in about 40% of patients. DNA methylation could play a role in this differential response between TNBC. The objective of our work was to define a DNA methylation profile with potential to predict response to CT in TNBC patients. Methods: Tumor samples from patients diagnosed with a TNBC stage I to III and considered candidate for neoadjuvant CT with anthracycline and taxane were identified. DNA was extracted from FFPA samples obtained by pre-surgery biopsy. Patients were classified according to residual cancer burden (RCB) index in responders (RCB = 0, n = 10) vs. non-responders (RCB > 0, n = 14). DNA methylation was performed by the Infinium HumanMethylation450 array (Illumina) that allows interrogating more than 485,000 methylation CpG sites per sample. The selection criteria were: p < 0.05 and mean difference between methylation groups of ≥ 20%. The functions associated to genes were analyzed using the Gene Ontology FDR. Results: We identified 24 samples from TNBC patients. Median age was 58.4 (48.5-69.0) years. Most patients presented T1-2 tumors (72%), no axillar lymph involvement (53%). Proliferative scores were high with 53% of patients with ki67 > 60% and 88% of grade 3 tumors. We have detected 133 differentially methylated CpGs allowing clearly separate the responders from the non-responders. Thirty five were located at CpG islands (CGIs) or CGI shores at promoter regions. We selected 11 CpGs corresponding to 11 genes with the less variation intragroup (standard deviation ≤ 20%). Nine of them showed a consistent DNA methylation profile of consecutive CpGs. Of these, 5 genes (genes 1 to 5) increased and 4 genes (genes 6 to 9) decreased methylation in non-responding patients compared to those who responded to treatment. The increase ratio varies between 0.20 and 0.27, and the decrease ranges between -0.21 and -0.22. Some of these genes were related to Wnt and Hedgehog pathways, epithelial-mesenchymal transition and cell migration. Conclusions: DNA methylation profile is an interesting predictive tool for response to neoadjuvant CT in TNBC.