The highly immunosuppressive nature of head-neck squamous cell cancer (HNSCC) is not fully understood. Exosomes play crucial roles in the communication between cancer and non-cancer cells, but the clinical significance of the expression of exosome-related genes (ERGs) remains unclear in HNSCC. This study aimed to establish an HNSCC-ERGs model by using mass spectrometry (MS)-based label-free quantitative proteomics in combination with the TCGA primary HNSCC dataset. The study managed to classify the HNSCC patients into two subtypes based on the expression level of prognostic ERGs, which showed significant differences in prognosis and immune infiltration. LASSO regression algorithm was used to establish a risk prediction model based on seven risky genes (PYGL, ACTN2, TSPAN15, EXT2, PLAU, ITGA5), and the high-risk group was associated with poor survival prognosis and suppressive immune status. HPRT1 and PYGL were found to be independent prognostic factors through univariate and multivariate Cox regression analyses. Immune and ssGSEA analysis revealed that HPRT1 and PYGL were significantly related to immunosuppression, immune response, and critical signaling transduction pathways in HNSCC. Immunohistochemistry results further validated the expression level, clinical value, and immunosuppressive function of HPRT1 and PYGL in HNSCC patients. In conclusion, this study established molecular subtypes and a prediction risk model based on the ERGs. Furthermore, the findings suggested that HPRT1 and PYGL might play critical roles in reshaping the tumor microenvironment.
Keywords: bioinformatics; exosome; head and neck squamous cell cancer; immune infiltration; prognosis.