Abstract
The incidence of head and neck squamous cell carcinoma (HNSCC) remains high, accompanied by low 5-year survival rates. Identifying prognostic factors is essential for advancing personalized treatment approaches. Increasing evidence implicates aberrant alternative splicing (AS) plays a key role in tumor progression. Utilizing data from TCGA and TCGA SpliceSeq, prognosis-associated AS events were identified through Cox regression analysis. A prognostic risk model was developed via multivariate Cox and LASSO regression, with validation conducted using Kaplan-Meier survival analysis and ROC curve analysis. The correlation between splicing factors (SFs) and prognosis-associated AS events was analyzed using Pearson's method, followed by the construction of an SF-AS regulatory network. Key splicing factors (KSFs) were identified using Cytoscape software. Expression of KSFs in HNSCC was confirmed by quantitative PCR and Western blotting. SiRNA-mediated knockdown in HNSCC cell lines (HONE1, HN4, SAS) demonstrated effects on cell proliferation, invasion, and migration, as assessed by CCK8, colony formation, Transwell, and wound healing assays. Tumor growth was further evaluated in a subcutaneous tumor model in vivo. A total of 2347 survival-related AS events were identified, of which eleven were used to construct the prognostic model. Patients in the low-risk group exhibited significantly improved outcomes (P = 0e + 00), underscoring the model's predictive accuracy. Notably, DDX39B and PRPF39 emerged as key splicing factors, exhibiting high expression in HNSCC and correlating with poor prognosis, positioning them as potential biomarkers and therapeutic targets.