Systematic Analysis of Tumor Stem Cell-related Gene Characteristics to Predict the PD-L1 Immunotherapy and Prognosis of Gastric Cancer

Abstract

Aims:: We aimed to develop a prognostic model with stemness-correlated genes to evaluate prognosis and immunotherapy responsiveness in gastric cancer (GC). Background:: Tumor stemness is related to intratumoral heterogeneity, immunosuppression, and anti-tumor resistance. We developed a prognostic model with stemness-correlated genes to evaluate prognosis and immunotherapy responsiveness in GC. Objective:: We aimed to develop a prognostic model with stemness-correlated genes to evaluate prognosis and immunotherapy responsiveness in GC. Methods:: We downloaded single-cell RNA sequencing (scRNA-seq) data of GC patients from the Gene-Expression Omnibus (GEO) database and screened GC stemness- related genes using CytoTRACE. We characterized the association of tumor stemness with immune checkpoint blockade (ICB) and immunity. Thereafter, a 9-stemness signature-based prognostic model was developed using weighted gene co-expression network analysis (WGCNA), univariate Cox regression analysis, and the least absolute shrinkage and selection operator (LASSO) regression analysis. The model predictive value was evaluated with a nomogram. Results:: Early GC patients had significantly higher levels of stemness. The stemness score showed a negative relationship to tumor immune dysfunction and exclusion (TIDE) score and immune infiltration, especially T cells and B cells. A stemness-based signature based on 9 genes (ERCC6L, IQCC, NKAPD1, BLMH, SLC25A15, MRPL4, VPS35, SUMO3, and CINP) was constructed with good performance in prognosis prediction, and its robustness was validated in GSE26942 cohort. Additionally, nomogram and risk score exhibited the most powerful ability for prognosis prediction. High-risk patients exhibited a tendency to develop immune escape and low response to PD-L1 immunotherapy. Conclusion:: We developed a stemness-based gene signature for prognosis prediction with accuracy and reliability. This signature also helps clinical decision-making of immunotherapy for GC patients.