The metastatic potential of breast cancer cells is believed to correspond with the preferential enrichment of a tumor population known as cancer stem cells (CSCs). Thus, evaluating the unique molecular characteristics of breast CSCs is of great interest from a prognostic and therapeutic standpoint. In order to provide a more detailed profile of breast cancer stem cells, we compared the 450K DNA methylation landscape of the EpCAM-/CD49f- cancer stem cell subpopulation from the isogenic MCF10A p53-/PTEN- breast cell line against the corresponding EpCAM+/CD49f+ and EpCAM-/CD49f+ subpopulations to determine how DNA methylation varies within the different genomic regions of CSCs. In addition, we also overlapped the 450K DNA methylation profiles from 16 established breast cancer cell lines of varying subtypes and aggressiveness to determine how these cell lines relate epigenetically with the isolated CSCs. Based on unsupervised PCA and matrix dissimilarity clustering, we were able to identify 3 distinct groups that appeared to independently cluster based on EpCAM-/CD49f- enrichment status. It is also interesting to note that the aggressive cell lines SUM149 and MDA-MB-231 would cluster away from the EpCAM-/CD49f- subset when utilizing strictly promoter probes for clustering, whereas SUM149 and MDA-MB-231 would cluster more closely with the EpCAM-/CD49f- subset when utilizing gene body probes, potentially indicating a more sensitive correlation between gene body methylation and CSC-associated aggressiveness. To further investigate the differing promoter and gene body DNA methylation patterns in CSCs, we performed differential methylation analysis between the 3 previously defined groups. Based on our results, we discovered 1432 differentially methylated promoter probes and 7243 differentially methylated gene body probes (ANOVA FDR p-value <0.001), with the majority of the promoter probes being hypermethylated in the CSC group and, inversely, the majority of the gene body probes being hypomethylated in the CSC group. Examples of genes that demonstrated significant hypomethylation throughout the gene body within CSCs included the guanine nucleotide exchange factor MCF2L, the synaptic-associated protein SHANK2, and the protein kinase C isoform PRKCZ. For the promoter regions, genes that were significantly hypermethylated in CSCs included the metabolic regulator GPD2, the e-cadherin protein CDH1, and the transcriptional regulator IRF6, all of which were transcriptionally suppressed in the CSC populations. Based on these findings, our work helps provide clarity to the stochastic nature of DNA methylation changes throughout the genome in EpCAM-/CD49f- breast CSCs and provide further motivation for determining the prognostic potential of these identified CSC-associated epigenetic events.

Citation Format: Austin Y. Shull, Caroline E. Dyar, Maria Ouzounova, Nicole L. Hudson, Max S. Wicha, Hasan Korkaya. Deciphering the DNA methylation signature of EpCAM-/CD49f- breast cancer stem cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5321.