179P Integrative analysis revealed the signature of cancer stem cells and its immunosuppressive role in lung adenocarcinoma

Cancer stem cells (CSCs) could induce immunosuppression to promote tumor progression and therapy resistance. Due to the absence of specific markers, CSCs and their phenotypes remain unexplored in lung adenocarcinoma (LUAD). In the single-cell level, CytoTRACE package was employed to calculate CytoTRACE score representing the level of stemness. Malignant cells with top 25% CytoTRACE score were defined as CSCs. Then, we developed a strict procedure involving weighted gene co-expression network analysis (WGCNA) and metacell algorithm to identify tumor-specific CSC-related genes. Our in-house serum proteomics data was used for screening CSC-related genes for biomarkers of immunotherapy. Based on over 20 000 single cells from five datasets, we identified a group of CSCs with highest CytoTRACE score. CSCs exhibited enhanced proliferating activities, such as cell cycle, G2M checkpoint and MYC targets. In the WGCNA analysis, a module (correlation = 0.9, p < 0.0001) was identified as having a significant association with CytoTRACE Score. A total of 91 genes overlapping between this module and tumor specific genes were exploited to construct the Stemness Score. In bulk transcriptomics level, Stemness Score was positively associated with immunosuppressive cells (correlation >0.3), including the T cells regulatory (Treg) and myeloid-derived suppressor cells (MDSC). Checkpoints PDCD1, CD274, CTLA4 and TIM3 were also observed to have positive association with Stemness Score (correlation >0.3). Stemness Score also linked to undesirable prognosis (HR = 1.68, p = 0.0004) in the cohort TCGA LUAD. In our in-house cohort of 57 samples from 17 non-small cell lung cancer (NSCLC) patients treated with immunotherapy, the serum proteomics data revealed that ENO1 was significantly higher in non-responders (p = 0.0023) and its level increased as disease progressed (p = 0.005). This study characterized and thoroughly elucidated CSCs, through integrative analyses of multi-omics data, identifying ENO1 as a novel serum biomarker for predicting immunotherapeutic outcome in NSCLC.