Abstract 3652: Identification of cancer-specific signaling networks: what is “normal” control

Abstract Success of precision medicine depends on definitive identification of cancer-specific alterations in signaling pathways. However, identifying cancer-specific signaling networks is challenging because of lack of proper control tissue for differential gene expression analyses. Most studies in breast cancer utilize tumor-adjacent normal tissue or reduction mammoplasty samples as “normal” controls. We recently reported that breast epithelial cells from healthy donors as well as tumor-adjacent normal are in different differentiation states compared with tumor cells and the differences in differentiation status alone could account for major transcriptome variations between normal and tumor. To overcome these limitations, we propagated breast epithelial cells from three healthy donors (healthy-normal), two high-risk patients, two tumor-adjacent normal (HR/AD-normal) and five tumor samples of different molecular subtypes. Phenotypically defined (CD49f+/EpCAM+) luminal progenitor cells were sorted from these cultures and subjected to RNA-seq analyses. Pathway analysis revealed activation of cell-intrinsic pro-inflammatory signaling in HR/AD-normal cells compared with healthy-normal cells. This signaling network was further amplified in tumor cells. The pro-inflammatory chemokine CCL2, which is overexpressed in highly aggressive breast cancer, and the cytokine TNFRSF11B were elevated in HR/AD-normal luminal progenitor cells. Despite using phenotypically defined cells in the transcriptome analyses, cancer-specific signaling network identification was directly influenced by the type of controls used; healthy-normal or HR/AD-normal. While cancer-enriched PI3K and NF-κB activation was observed when compared to any kind of control, SRC kinase activation was noted only when cells from healthy-normal were used as a control. In general, the number of tumor signaling networks identified using healthy-normal as a control was higher than when compared with HR/AD-normal as a control. These results suggest that considerable attention should be placed on the type of tissues used as control for definitive identification of cancer-specific signaling networks and therapies to target such pathways. Additionally, these data show that non-cancer tissues of breast cancer patients acquire a cell intrinsic pro-inflammatory phenotype, which may be prerequisite for cancer development and potentially an early-detection tool. Citation Format: Manjushree Anjanappa, Yangyang Hao, Howard J. Edenberg, Yunlong Liu, Harikrishna Nakshatri. Identification of cancer-specific signaling networks: what is “normal” control. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 3652.