Prognosis and immunotherapy response in head and neck squamous cell carcinoma could be predicted by a signature based on three genes, as revealed by an integrated analysis of single-cell and bulk RNA-sequencing data

Abstract Background: The microenvironment of head and neck squamous cell carcinoma (HNSC) is made up of cancer and non-cancerous cells, and their interactions have profound effects on anti-tumor immunity. However, a thorough understanding of the genetic and cellular-level intercellular communication networks involved in tumor progression remains a significant obstacle. Material/Methods: 460 HNSC patients from various cohorts were included. To identify the marker genes, we analyzed single-cell RNA-sequencing (scRNA-seq) data from GEO database. An analysis of immunological infiltrating cell density was carried out using cell-type identification by calculating relative subsets of RNA transcripts (CIBERSORT). The bulk RNA-seq dataset from TCGA database was used to construct signature, and the GSE 65858 were used for validation. And the expression of related proteins were verified using HPA database and western blotting. Results: A three-gene signature (CES1, ELF3 and SERPINE1) was developed for prognostic prediction in the TCGA dataset, which divided patients into high-risk and low-risk categories based on overall survival. The prognostic potential of the signature was confirmed by GSE 65858. The signature protein expression was validated by HPA database and western blotting. Furthermore, the riskScore was identified as a significant prognostic factor in the multivariate analysis, indicating that the signature had high predictive ability. In addition, patients with high-risk scores obtained fewer benefits from immunotherapy. Conclusions: Our study identified a distinctive predictive signature for HNSC patients based on CES1, ELF3, and SERPINE1. The signature may be used as a predictor for immunotherapy and as an indicator of survival in patients with HNSC.