Identification and validation of tumor microenvironment-related prognostic biomarkers in breast cancer

Background: Stromal cells and immune cells in tumor microenvironment (TME) have been reported to have significant value in the diagnosis and prognosis of cancers. We aimed to identify key biomarkers predicting survival in the TME of breast cancer. Methods: Cell type enrichment analysis was performed to estimate cell scores using the xCell method with gene expression data from public database. Least absolute shrinkage and selection operator (LASSO) regression was used to identify key signature from the cell scores. Results: Totally, 50 cells in TME had different scores between 1,078 breast cancer tissues and 112 adjacent normal tissues. We identified a 4-cell signature predicting breast cancer survival, including myocytes, natural killer T cell (NKT), conventional dendritic cell (cDC) and sebocytes, which was validated in the test set. Further analysis showed that cDC score was a key signature predicting prognosis of breast cancer. cDC score was significantly associated with molecular classification and stage of breast cancer, as well as expression level of Ki67. Spearman's correlation analysis found that cDC score was inversely correlated with the expression level of HER2. High cDC score may predicate better pathological complete response rate. Mechanism analysis indicated high cDC score was associated with elevated immune activity; IL-2 was a key gene associated with high cDC score; and Breast cancer patients with high IL-2 expression had a longer survival time. Conclusions: In conclusion, cDC score was a key signature predicting prognosis for breast cancer. cDCs may exert antitumor effects by upregulating IL-2.