In breast cancer, it is unclear the functional modifications at a transcriptomic level that are associated with the evolution from epithelial cells and ductal carcinoma in situ (DCIS) to basal-like tumors. By applying weighted gene co-expression network analysis (WGCNA), we identified 17 gene co-expression modules in normal, DCIS and basal-like tumor samples. We then correlated the expression pattern of these gene modules with disease progression from normal to basal-like tumours and found eight modules exhibiting a high and statistically significant correlation. M4 included genes mainly related to cell cycle/division and DNA replication like CCNA2 or CDK1. The M7 module included genes linked with the immune response showing top hub genes such as CD86 or PTPRC. M10 was found specifically correlated to DCIS, but not to basal-like tumor samples, and showed enrichment in ubiquitination or ubiquitin-like processes. We observed that genes in some of these modules were associated with clinical outcome and/or represented druggable opportunities, including AURKA, AURKB, PLK1, MCM2, CDK1, YWHAE, HSP90AB1, LCK, or those targeting ubiquitination. In conclusion, we describe relevant gene modules related to biological functions that can influence survival and be targeted pharmacologically.
Keywords: breast cancer; carcinoma in situ; transcriptomic evolution.