Acidosis-associated gene signature defines novel subtypes and dual-target therapeutic candidates in breast cancer

BACKGROUND: Breast cancer remains the most common malignancy in women, and substantial heterogeneity in treatment response and prognosis persists despite multimodal therapies. The acidic tumor microenvironment (TME), driven by metabolic reprogramming and lactate accumulation, is recognized as a key driver of tumor adaptation, immune evasion, and therapeutic resistance. However, the genomic and transcriptomic patterns of acidosis tolerance in human breast cancer, and their implications for subtype stratification and targeted therapy, remain poorly understood. METHODS: We integrated GEO datasets, sgRNA-seq data and breast cancer-associated genes to identified breast cancer acidosis tolerance genes (BCATGs). GO and KEGG enrichment analysis were used to define BCATGs function. We divided breast cancer patients into two subtypes based on BCATGs by consensus clustering. GSEA and GSVA were used to characterized the molecular mechanisms of BCATGs subtypes. CIBERSORT, ESTIMATE, IPS, and TIDE, and oncoPredict were applied to depict the immune microenvironment of BCATGs subtypes. A LASSO-Cox prognostic model was developed and validated, with clinical correlations assessed by Cox regression. Virtual screening against key prognostic genes employed AutoDock Vina, followed by molecular dynamics simulations and MM/PBSA binding energy calculations in GROMACS. RESULTS: Seventeen BCATGs were identified, predominantly enriched in mitotic regulation and cell cycle, with low mutation rates, predominant copy number gains and notable co-occurrence patterns. Consensus clustering revealed two subtypes: Subtype I and Subtype II. Subtype II exhibited marked activation of proliferative signatures and suppression of differentiation pathways, coupled with a pro-inflammatory yet immunosuppressive immune profile. Subtype II showed greater sensitivity to cell-cycle inhibitors, apoptosis inducers, and proteasome inhibitors. A five-gene LASSO risk model (AURKA, CCNA2, CDC45, EXO1, KIF4A) demonstrated robust prognostic performance, particularly for disease-specific survival (1-year AUC 0.731), with CCNA2 and CDC45 retaining independent prognostic significance after multivariate adjustment. Virtual screening and molecular dynamics identified four lead compounds (SBC-115337, CDK2-IN-4, Bractoppin, Corylin) with stable binding to CCNA2 and CDC45. CONCLUSIONS: This study establishes a novel framework linking acidosis adaptation to breast cancer heterogeneity, identifying BCATGs-driven subtypes with distinct molecular, immunological, and pharmacological profiles. The prognostic model highlights CCNA2 and CDC45 as key drivers of adverse outcomes. These findings provide a foundation for patient stratification, risk assessment, and targeted therapy in breast cancer.