The implication of endoplasmic reticulum stress-related genes in lung adenocarcinoma revealed by bioinformatics and single cell RNA sequencing analysis

Abstract

Background: Endoplasmic reticulum stress (ERS) has the potential to treat a variety of malignant tumors, including lung adenocarcinoma (LUAD). In this study, by leveraging bioinformatics, ERS-related genes were screened to construct a prognostic model for lung adenocarcinoma so as to find new therapeutic strategies. Methods: A variety of bioinformatics algorithms (co-expression analysis, univariate Cox analysis, multivariate Cox analysis and cross-validation) were used to screen ERS-related genes and construct a risk model. Patients with LUAD were divided into training group and testing group in a 1:1 ratio. Receiver operating characteristic curve (ROC), nomogram, independent prognostic analysis and principal component analysis were used to compare the risk scores of the high and low risk groups to verify the validity of the prognostic model. In addition, consensus clustering was used to classify different clusters of LUAD patients which were compared in terms of survival time, immune microenvironment and pathway regulation. The deletion-associated genes were analyzed, combined with single cell sequencing (scRNA-seq), to further investigate screened prognostic risk genes. Results: This study demonstrated the feasibility of a model based on six ERS related genes (SLC2A1, ASPH, SERPINH1, TLR4, CAV3, and SLC6A4), as well as the identification of UMI77, YM155, MG132, and lapatinib as potential therapeutic strategies for LUAD. Risk scores based on this model could be used as independent prognostic factors for LUAD (HR > 1; p < 0.001) and had the highest accuracy in predicting survival compared to clinical features. scRNA-seq found that SERPINH1, ASPH, and SLC2A1 were mainly expressed in malignant cells of various cancer. Conclusions: An accurate risk model was constructed based on six ERS-related genes, which can be used for prognostic prediction and therapeutic strategy design in clinical practice.