CELL INTELLIGENCE · 肿瘤细胞治疗研究

The bulk RNA-seq data for patients with BRCA were collected from the TCGA-BRCA cohort, GSE20685, and GSE96058 cohorts. STING pathway-related genes (SRGs) were obtained from the Reactome database. Differentially expressed genes were analyzed using the limma package. Immune cell infiltration was analyzed using the IOBR package. Gene Ontology biological processes, Kyoto Encyclopedia of Genes and Genomes pathways, and cancer hallmark pathways were analyzed using the MSigDB database. Prognostic models were prepared using the least absolute shrinkage and selection operator and multiple-factor Cox regression analysis. Single-cell analysis was performed using the Seurat and SCP pipeline.

The expression patterns and clinical relevance of SRGs were analyzed in patients with BRCA. Transcriptional differences in the SRGs were observed between normal and tumorous tissues, with global down-regulated STING1 and up-regulated TBK1 in BRCA tissue. Tumor tissues were classified through consensus clustering analysis into two distinct groups, with differences in clinical characteristics and immune infiltration. A prognostic model related to the differences in STING pathway activity-high prognostic stratification potency-was developed and validated. Correlation analysis revealed suppressed overall immune activation in patients with BRCA having higher risk scores. Gemcitabine had a more favorable outcome in the low-risk group. The activity of the prognostic model at the single-cell level was confirmed through single-cell analysis, particularly in CD8 T cells and intratumor natural killer cells.

A STING pathway-related prognostic model developed and validated and the model could accurately predict BRCA patient outcomes. These findings have important implications for the personalized treatment and management of patients with BRCA.