Identification and validation of key biomarkers for chemoresistance in oral squamous cell carcinoma

This study aimed to elucidate the molecular mechanisms underlying cisplatin resistance in oral squamous cell carcinoma (OSCC) and to identify prognostic biomarkers that can guide personalized treatment strategies. Differentially expressed genes (DEGs) between cisplatin-sensitive and resistant OSCC cells were identified using transcriptomic data from the GSE197561 dataset. Functional enrichment analyses were conducted to explore involved pathways. Weighted Gene Co-expression Network Analysis (WGCNA) and random survival forest algorithms were employed to identify key resistance-related gene modules and hub genes. Clinical relevance was validated using The Cancer Genome Atlas (TCGA) dataset, while immune cell profiling and functional validation in SCC9/CAL27 cisplatin-resistant models were conducted to elucidate biological significance. DEGs were significantly enriched in processes such as oxidative stress response, cell cycle regulation, and immune-inflammatory pathways. WGCNA and survival analysis identified a chemotherapy resistance-associated gene module and 10 core genes, including DUSP3 and SACM1L, whose high expression was correlated with poorer overall survival. A nomogram integrating TNM stage, age, and DUSP3 expression demonstrated superior predictive performance for patient prognosis. In analyses stratified by chemotherapy resistance, upregulated DUSP3 expression appeared to synergistically cooperate with malignant tumor progression. Immune profiling revealed reduced NK cell activity and decreased populations of resting CD4⁺ memory T cells in the resistant phenotype. Functional assays confirmed that cisplatin-resistant SCC9 and CAL27 cells exhibited accelerated proliferation and epithelial-mesenchymal transition (EMT), characterized by downregulated E-cadherin and upregulated N-cadherin and Vimentin. Notably, knockdown of DUSP3 significantly reversed these resistant phenotypes. This study provides a comprehensive multi-omics-based framework for understanding cisplatin resistance in OSCC. DUSP3 was identified as a key prognostic indicator, and the proposed nomogram may enhance precision oncology efforts for cisplatin-resistant OSCC patients. Multi-omics analysis identifies key pathways driving cisplatin resistance in OSCC; Ten hub genes linked to resistance and prognosis were screened via Weighted Gene Co-Expression Network Analysis (WGCNA) and Random Survival Forest (RSF); DUSP3 overexpression predicts poor survival and is integrated into a prognostic nomogram; Immune profiling reveals reduced NK cell activity in cisplatin-resistant OSCC cells; SCC9/CAL27 cisplatin-resistant models exhibit EMT features and upregulated hub gene (DUSP3) expression.