Exploration of precise classification and therapeutic targets of breast cancer based on genes related to neuro-cancer crosstalk.

This investigation examines neural-cancer crosstalk in breast cancer, utilizing TCGA/GEO transcriptomic and single-cell data to develop a neural-cancer crosstalk prognostic signature and evaluate its clinical relevance. We identified four core genes-L1CAM, TACR1, GFRA1, and NTRK3-using univariate Cox regression and LASSO-Cox regression to develop a risk scoring model. The prognostic signature showed consistent predictive accuracy across validation cohorts, with high-risk patients demonstrating markedly reduced survival. Multivariate analysis established the risk score as an independent prognostic indicator. Immune profiling identified an immunosuppressive phenotype in high-risk cases, featuring elevated Treg infiltration and impaired antigen presentation. Genomic analysis revealed significantly higher TP53 mutation frequency in high-risk versus low-risk patients. Drug sensitivity predictions indicated that the high-risk group exhibited increased sensitivity to targeted therapies such as Lapatinib, but enhanced resistance to conventional chemotherapy. Single-cell sequencing revealed the heterogeneity of breast cancer cells and elucidated a neuro-immune-tumor interaction network dominated by MIF and MDK signaling pathways. Functional validation studies demonstrated that L1CAM knockdown effectively suppressed malignant phenotypes in breast cancer cells, including proliferation, migration, and invasive capacity in vitro. This study provides new insights into the regulatory mechanisms of the neural microenvironment in breast cancer, and the established prognostic model and identified potential therapeutic targets hold significant clinical implications for personalized treatment.