Individualized Lipid Metabolism-Associated Six-Gene Signature for Prediction of Overall Survival, Immune Infiltration, Immunotherapy Response, and Potential Candidate Drugs in Lung Squamous Cell Carcinoma Patients: Development and Validation

Abstract

AbstractBackgroundDisordered lipid metabolism is a novel hallmark of cancer initiation. However, the role of lipid metabolism-associated genes (LAGs) in the immune microenvironment, the prognosis, and the overall survival of lung squamous cell carcinoma (LUSC) remain unclear.MethodsThis study examined 1064 LAGs in 502 LUSC tumors and 49 normal lung tissues from the Cancer Genome Atlas lung squamous cell carcinoma (TCGA-LUSC) cohort. Using univariate, least absolute shrinkage and selection operator (LASSO), and multivariate Cox regression analysis, a LAG-based gene signature was constructed for predicting the overall survival of patients with LUSC from the TCGA training-set. The predictive power of the gene signature was validated using the TCGA-internal validation cohort and six independent cohorts ( GSE73403, GSE74777, GSE157009, GSE157010, GSE157011, and GSE67061), obtained from the Gene Expression Omnibus (GEO) database.The prognosis was determined using a Kaplan-Meier analysis. The immunological aspects were examined using CIBERSORT, gene set enrichment analysis (GSEA), and tumor immune dysfunction and exclusion (TIDE) analysis. The Human Protein Atlas (HPA) database was utilized to validate the protein expression of the gene signature between LUSC tissues and normal lung tissues via immunohistochemistry staining (IHC) and Immunofluorescent . The expression of the gene signature was also assessed in various lung cancer-associated cell lines using HPA database. In addition, candidate small-molecule therapeutics for treating LUSC, were also identified using the connectivity Map (CMap) database.ResultsA total of 112 lipid metabolism-associated DEGs (LADEGs) were detected in LUSC. A six-LAG-based prognostic signature (ALOX15B, CYP24A1, PPP2R2C, PTGIS, SPTSSB, and UGT2B17)was successfully constructed and validated to predict the prognosis of LUSC patients. Functional analysis suggested that the LAGs were significantly enriched in arachidonic acid metabolism, ether lipid metabolism, metabolism of xenobiotics by cytochromeP450, glycerophospholipid metabolism, steroid hormone biosynthesis, the fatty acid metabolic process, oxidoreductase activity, andPPARsignaling pathways. Furthermore, we determined that the six-gene signature was an independent risk factor. The prognostic model was closely related to immune cell infiltration. The expression ofSPTSSBwas low in high-risk patients, whileALOX15B, CYP24A1, PPP2R2C, PTGIS, andUGT2B17were higher in high-risk patients than in low-risk ones. Identical outcomes were seen in the HPA database. In addition, the model showed that the high-risk score group had a worse overall survival (OS), and the validated cohorts showed the same result. Moreover, the prognostic signature of six LAGs predicted overall survival more accurately than conventional clinical characteristics.Based on the CMap dataset, small-molecule drugs exhibiting anticancer properties could be potential therapeutic therapies for LUSC patients.ConclusionsThis study demonstrated, for the first time, that LAG expression plays a crucial role in LUSC. Furthermore, this high-efficiency six-gene LAG prognostic signature could serve as a predictive model to separate LUSC patients into high- and low-risk groups and potentially facilitate risk-stratified immunotherapy in LUSC patients. In addition, a number of small molecule drugs with significant therapeutic potential for LUSC patients were found.