Construction and validation of a prognostic signature based on seven endoplasmic reticulum stress-related lncRNAs for patients with head and neck squamous cell carcinoma

Abstract Head and neck squamous cell carcinoma (HNSCC) is the most prevalent form of head and neck cancer, with an escalating incidence rate and dismal survival outcomes. Endoplasmic reticulum stress (ERS) occurs when misfolded or unfolded proteins accumulate in the endoplasmic reticulum (ER), and it is often observed in tumors, including HNSCC. Relevant studies have demonstrated the prognostic significance of ERS-related long non-coding RNAs (lncRNAs) in various cancers. However, the relationship between ERS and lncRNAs in HNSCC has received limited attention in previous studies. Based on this, we aimed to identify the lncRNAs that impact the clinical prognosis of HNSCC patients and develop an ERS-related lncRNA prognostic model in this study. Transcriptomic, gene mutation, and clinical data were acquired from The Cancer Genome Atlas (TCGA) database. Correlation analysis, Cox regression analysis, and least absolute shrinkage, and selection operator (LASSO) regression analysis were used to establish an ERS-related lncRNAs prognostic model. Subsequently, the survival and predictive ability of this model were evaluated using Kaplan-Meier analysis and time-dependent receiver operating characteristics (ROC), while nomograms and calibration curves were constructed. Additionally, Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Set Enrichment Analysis (GSEA) functional enrichment analyses, tumor mutation burden (TMB), tumor infiltration of immune cells, single sample Gene Set Enrichment Analysis (ssGSEA), drug sensitivity analysis were performed. Furthermore, we conducted a consensus cluster analysis to compare differences between subtypes of tumors. Finally, we validated the expression of the seven lncRNAs in HNSCC tissues through qRT-PCR. A total of 423 lncRNAs were identified as ERS-related lncRNAs in HNSCC tissues. The prognostic signature was established based on seven ERS-related lncRNAs, namely ACTN1-AS1, AP003774.2, DOCK8-AS1, DDX11-AS1, MIR924HG, MIR9-3HG, and AL451085.2. The high-risk group had a poorer prognosis in comparison to the low-risk group. As an independent predictor, the model showed better predictive performance than other clinicopathological features. The area under the ROC curve (AUC) predicted by this model for 3-year survival rates of HNSCC patients was 0.805. Enrichment analysis revealed that the differentially expressed genes were primarily enriched in pathways related to immune responses and signal transduction. Low-risk patients had lower TMB, more immune cell infiltrations, and enhanced anti-tumor immunity. Cluster analysis indicated that cluster 3 may have a better prognosis and immunotherapy effect. In addition, the result of qRT-PCR was consistent with our analysis. This prognostic model based on seven ERS-related lncRNAs is a promising tool for risk stratification, survival prediction, and immune cell infiltration status assessment.