Breast cancer metabolic subtypes analysis and risk prognostic models construction based on glycolysis metabolism genes

Abstract Glycolysis metabolism is an essential part of cancer research due to its role in cancer initiation and progression. However, its characteristics and prognostic value in breast cancer (BRCA) have not been systematically evaluated.We collected glycolysis metabolism gene expression profiles and clinical information of BRCA patients from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. After excluding individuals lacking clinical information and the presence of genetic mutations, we performed consistent clustering of the remaining patients based on glycolysis metabolism gene expression profiles and selected stable clustering results to group patients. Differentially expressed genes (DEGs) and gene set enrichment analysis (GSEA) were compared between subgroups, while metabolic signature identification and decoding the tumor microenvironment were performed. In addition, we explored the survival status of patients among different subgroups and identified signature genes affecting survival by Least Absolute Shrinkage and Selection Operator (LASSO) regression. Finally, we selected signature genes to construct risk prognostic models by multivariate Cox regression.By consensus clustering, patients were distinguished into two stable subpopulations, GSEA and metabolic signature identification effectively defined two completely different subtypes of glycolysis metabolism: glycolysis hyperactive subtype and glycolysis hypoglycemia subtype. Among them, patients with the glycolysis hyperactive subtype had a poorer prognosis, with a significantly lower proportion of Macrophages M1 infiltration within the tumor microenvironment than others. Eight key genes, DEP domain containing 1(DEPDC1), Ras related GTP binding D(RRAGD),Phosphoglycerate kinase 1(PGK1),Secreted Frizzled-related Protein 2(STC2),Syndecan-1(SDC1), Lactate Dehydrogenase A(LDHA) ,Calpain 5(CAPN5),calcium channel, voltage-dependent, alpha 1H subunit(CACNA1H), were selected by multivariate Cox regression, which constructed a risk prognostic model.Our study revealed the heterogeneity of glycolysis metabolism in BRCA patients, defined two completely distinct subtypes of glycolysis metabolism, and finally established a novel glycolysis metabolism-related risk prognostic model. The study contributes to the early risk assessment and monitoring of individual prognosis and provides data to support individualized patient treatment.