Identification and validation of disulfidptosis-related gene signature revealed DKK1 as a therapeutic target in head and neck squamous cell carcinoma
摘要
Head and neck squamous cell carcinoma (HNSCC) poses significant prognostic challenges due to recurrence, metastasis, and drug resistance. Disulfidptosis, a newly identified cellular self-destruction pathway, is implicated in various diseases, including cancer. This study aimed to identify biomarkers by analyzing disulfidptosis-related gene expression. Recent studies have evaluated disulfidptosis in HNSCC; we therefore position this work by integrating multi‑cohort transcriptomic analysis with a focused evaluation of DKK1 and functional assays to clarify its potential clinical relevance. Utilizing transcriptomic data from The Cancer Genome Atlas(TCGA) and Gene Expression Omnibus(GEO)datasets, unsupervised clustering grouped HNSCC patients, and Least Absolute Shrinkage and Selection Operator (LASSO) regression pinpointed central genes to develop a prognostic and therapeutic prediction model. The integration of a nomogram enhanced prediction accuracy. Among 527 HNSCC patients, we identified 10 disulfidptosis-related genes with dysregulated expression, contributing to HNSCC progression. A risk score model incorporating clinical factors such as age and stage was developed. Analyses based on gender, stage, and risk score revealed significant differences in tumor mutation burden and immune landscape between risk categories. The model showed predictive utility predicted chemotherapy sensitivity and identified candidates for programmed death-1/programmed death-ligand 1(PD-1/L1)-targeted therapy. Integrating this with clinical features could guide personalized follow-up strategies. Experimental validation confirmed DKK1 upregulation in oral squamous cell carcinoma cell (OSCC) lines (CAL27 and SCC15), and its knockdown inhibited proliferation in vitro, supporting its potential involvement in OSCC development, however, mechanistic assays of β‑catenin activity/localization and cystine‑starvation–induced disulfidptosis (e.g., phalloidin staining, disulfide‑stress readouts) were not performed, so the DKK1–disulfidptosis link is interpreted as associative rather than causal. Our study proposes a 10-gene disulfidptosis signature that predicts HNSCC outcomes and treatment responses, with DKK1 overexpression implicated as a potential contributor to progression and immune modulation.