Abstract
PURPOSE: Reliable biomarkers to predict the outcome and treatment response of estrogen receptor (ER)-negative breast cancer (BC) are urgently needed. Since immune-related signaling plays an important role in the tumorigenesis of ER-negative BC, we asked whether Notch genes, alone or in combination with other immune genes, can be used to predict the clinical outcome and immune checkpoint blockade (ICB) for this type of cancer. METHODS: We analyzed transcriptome data of 6918 BC samples from five independent cohorts, 81 xenograft triple-negative BC tumors that respond differently to ICB treatment and 754 samples of different cancer types from patients treated with ICB agents. RESULTS: We found that among four Notch genes, the expression levels of NOTCH1 and NOTCH4 were positively associated with recurrence of ER-negative BC, and that combined expression of these two genes (named Notch14) further enhanced this association, which was comparable with that of the Notch pathway signature. Analysis of 1182 immune-related genes revealed that the expression levels of most HLA genes, particularly HLA-DMA and -DRA, were reversely associated with recurrence in ER-negative BC with low, but not high Notch14 expression. A combined expression signature of NOTCH1, NOTCH4, HLA-DMA and HLA-DRA was more prognostic for ER-negative and triple-negative BCs than previously reported immune-related signatures. Furthermore, we found that the expression levels of these four genes were also synergistically associated with T cell exclusion score, infiltration of specific T cells and ICB efficacy in ER-negative BC, thereby providing a potential molecular mechanism for the synergistic effect of these genes on BC. CONCLUSIONS: Our data indicate that a gene signature composed of NOTCH1, NOTCH4, HLA-DMA and HLA-DRA may serve as a potential promising biomarker for predicting ICB therapy efficacy and recurrence in ER-negative/triple-negative BCs.