Unique epigenetic gene profiles define human breast cancers with poor prognosis.

Epigenetic enzymes are at the nexus of cellular regulatory cascades and can drive cancer-specific deregulation at all stages of the oncogenic process, yet little is known about their prognostic value in human patients. Here, we used qRT-PCR to profile at high resolution the expression of fifty-five epigenetic genes in over one hundred human breast cancer samples and patient-matched benign tissues. We correlated expression patterns with clinical and histological parameters and validated our findings in two independent large patient cohorts (TCGA and METABRIC). We found that human breast malignancies have unique epigenetic profiles and cluster into epigenetic subgroups. A subset of epigenetic genes defined an Epigenetic Signature as an independent predictor of patient survival that outperforms triple negative status and other clinical variables. Our results also suggest that breast cancer grade, but not stage, is driven by transcriptional alterations of epigenetic modifiers. Overall, this study uncovers the presence of epigenetic subtypes within human mammary malignancies and identifies tumor subgroups with specific pharmacologically targetable epigenetic susceptibilities not yet therapeutically exploited.