Machine learning framework develops neutrophil extracellular traps model for clinical outcome and immunotherapy response in lung adenocarcinoma

Abstract

Neutrophil extracellular traps (NETs) are novel inflammatory cell death in neutrophils. Emerging studies demonstrated NETs contributed to cancer progression and metastases in multiple ways. This study intends to provide a prognostic NETs signature and therapeutic target for lung adenocarcinoma (LUAD) patients. Consensus cluster analysis performed by 38 reported NET-related genes in TCGA-LUAD cohorts. Then, WGCNA network was conducted to investigate characteristics genes in clusters. Seven machine learning algorithms were assessed for training of the model, the optimal model was picked by C-index and 1-, 3-, 5-year ROC value. Then, we constructed a NETs signature to predict the overall survival of LUAD patients. Moreover, multi-omics validation was performed based on NETs signature. Finally, we constructed stable knockdown critical gene LUAD cell lines to verify biological functions of Phospholipid Scramblase 1 (PLSCR1) in vitro and in vivo. Two NETs-related clusters were identified in LUAD patients. Among them, C2 cluster was provided as "hot" tumor phenotype and exhibited a better prognosis. Then, WGCNA network identified 643 characteristic genes in C2 cluster. Then, Coxboost algorithm proved its optimal performance and provided a prognostic NETs signature. Multi-omics revealed that NETs signature was involved in an immunosuppressive microenvironment and predicted immunotherapy efficacy. In vitro and in vivo experiments demonstrated that knockdown of PLSCR1 inhibited tumor growth and EMT ability. Besides, cocultural assay indicated that the knockdown of PLSCR1 impaired the ability of neutrophils to generate NETs. Finally, tissue microarray (TMA) for LUAD patients verified the prognostic value of PLSCR1 expression. In this study, we focus on emerging hot topic NETs in LUAD. We provide a prognostic NETs signature and identify PLSCR1 with multiple roles in LUAD. This work can contribute to risk stratification and screen novel therapeutic targets for LUAD patients.

Keywords: Cluster immunophenotype; Lung adenocarcinoma; Machine learning; Neutrophil extracellular traps; PLSCR1; Prognostic model.