Abstract

Background: Breast cancer progression has been gradually recognized as a bidirectional interaction between cancer cells and tumor microenvironment including stroma cells, immune cells, and the dynamically altered ECM. However, there still lacks direct experimental evidences about how ECM properties modulate the activities of stroma and immune cells.Method: The transcriptomic data and corresponding clinical information of breast cancer pawere obtained from TCGA. Patients were divided into ECM-high, ECM-median and ECM-low groups based on ssGSEA scores of C-ECM genes. The prognostic value of ECM was confirmed by univariate/multivariate Cox regression and survival analyses. GO and KEGG analyses were performed between ECM-high and -low groups. Then associations between ECM characteristics and clinical stages were verified by Masson’s trichrome and Sirius red/Fast Green staining of clinical breast cancer tissues. To evaluate the effects of ECM on CAF induction and T cell activation, the MRC-5, NIH/3T-3, primary T cells and Jurkat T cells were encapsulated in 3D collagen with different densities and organizations, and the expression levels of CAF biomarkers and secretion levels of IL-2 were assessed.Results: ECM scores showed broad variation across paracancerous and cancer samples as well as breast cancer molecular subtypes, and patients with different ECM groups showed distinct prognosis. Immunological activity and ECM associated biology processes were identified by GO and KEGG analyses across ECM-high and -low groups. According to MCP-counter algorithm, the infiltration of T cells was significantly lower in the ECM-high group, while CAF abundance was significantly higher. It is furtherly confirmed by clinical samples that collagen density and organization were associate with breast cancer progression. Finally, in vitro 3D-cultured fibroblasts and T cells validated that the