Abstract
Hypoxia exerts a profound influence on the tumor microenvironment and immune response, shaping treatment outcomes and prognosis. Utilizing consistency clustering, we discerned two hypoxia subtypes in OPSCC bulk sequencing data from GEO. Key modules within OPSCC were identified through weighted gene correlation network analysis (WGCNA). Core modules underwent CIBERSORT immune infiltration analysis and GSEA functional enrichment. Univariate Cox and LASSO analyses were employed to construct prognostic models for seven hypoxia-related genes. Further investigation into clinical characteristics, the immune microenvironment, and TIDE algorithm prediction for immunotherapy response was conducted in high- and low-risk groups. scRNA-seq data were visually represented through TSNE clustering, employing the scissors algorithm to map hypoxia phenotypes. Interactions among cellular subpopulations were explored using the Cellchat package, with additional assessments of metabolic and transcriptional activities. Integration with clinical data unveiled a prevalence of HPV-positive patients in the low hypoxia and low-risk groups. Immunohistochemical validation demonstrated low TDO2 expression in HPV-positive (P16-positive) patients. Our prediction suggested that HPV16 E7 promotes HIF-1α inhibition, leading to reduced glycolytic activity, ultimately contributing to better prognosis and treatment sensitivity. The scissors algorithm effectively segregated epithelial cells and fibroblasts into distinct clusters based on hypoxia characteristics. Cellular communication analysis illuminated significant crosstalk among hypoxia-associated epithelial, fibroblast, and endothelial cells, potentially fostering tumor proliferation and metastasis.