Development of an anoikis-related gene signature for lung squamous cell carcinoma to predict prognosis, immune landscape, and immunotherapy response

Abstract

Abstract Background Lung squamous cell carcinoma (LUSC) is a highly invasive malignancy associated with a poor prognosis. Anoikis plays a crucial role in the malignant progression of various cancers. However, studies focusing on the role of anoikis in LUSC are needed. Methods A total of 357 anoikis-related genes (ARGs) were obtained from the GeneCards database and Harmonizome portals. Univariate Cox regression was used to identify prognostic ARGs for patients with LUSC, and the R package ConsensusClusterPlus was used to determine the subtypes. Lasso regression was used to build a prognostic risk model called the ARG score. Functional enrichment analysis was used to investigate the potential biological function of the ARGs. The ESTIMATE, ssGSEA, and CIBERSOT algorithms were used to evaluate the pattern of immune infiltration of patients with different LUSC subtypes and risk groups. The R package timeROC, survival, and survminer were used to test the accuracy of the prognostic model and to construct a nomogram. RT-PCR was used to detect differential expression of the 9 hub genes in LUSC and normal lung epithelial cells. Finally, OncoPredict was used to identify drugs with therapeutic value for patients in the high and low-risk groups. Results A total of 138 differentially expressed ARGs were selected, among which 15 survival-related ARGs were identified, and 3 subtypes (A, B, and C) with significant differences in the immune infiltration and survival outcomes were determined. Nine prognostic ARGs were identified to construct prognostic models, and the predictive capacity was verified in the GSE73403 and GSE30219 datasets. Patients with a high-risk score had worse outcome, and were associated with Tregs, resting memory CD4 + T cells, neutrophils, immune escape phenotypes, inflammatory response, and epithelial–mesenchymal transition. A nomogram with excellent clinical diagnostic accuracy was established for LUSC patients. The 9 hub genes had lower expression in LUSC cell lines, except for CHEK2, FADD, and SPINK1. Finally, we identified several drugs for patients in different risk groups. Conclusions ARGs play a key role in LUSC, and the ARG score devised in this study may be a valuable tool for determining clinical prognosis and the immunotherapy response in patients with LUSC.