Zhong Haihui, Wang Jie, Zhu Yaru, Shen Yefeng
Department of Thoracic Surgery, Meizhou People's Hospital (Huangtang Hospital), Meizhou Hospital Affiliated to Sun Yat-sen University, Meizhou Academy of Medical Sciences, Meizhou, China.
Institute for Pathology, University Hospital of Cologne, Cologne, Germany.
Front Cell Dev Biol. 2021 Sep 1;9:700607. doi: 10.3389/fcell.2021.700607. eCollection 2021.
Lung adenocarcinoma (LUAD) is the most common malignancy, leading to more than 1 million related deaths each year. Due to low long-term survival rates, the exploration of molecular mechanisms underlying LUAD progression and novel prognostic predictors is urgently needed to improve LUAD treatment. In our study, cancer-specific differentially expressed genes (DEGs) were identified using the robust rank aggregation (RRA) method between tumor and normal tissues from six Gene Expression Omnibus databases (GSE43458, GSE62949, GSE68465, GSE115002, GSE116959, and GSE118370), followed by a selection of prognostic modules using weighted gene co-expression network analysis. Univariate Cox regression, least absolute shrinkage and selection operator (LASSO), and multivariate Cox regression analyses were applied to identify nine hub genes () that constructed a prognostic risk model. The RNA expressions of nine hub genes were validated in tumor and normal tissues by RNA-sequencing and single-cell RNA-sequencing, while immunohistochemistry staining from the Human Protein Atlas database showed consistent results in the protein levels. The risk model revealed that high-risk patients were associated with poor prognoses, including advanced stages and low survival rates. Furthermore, a multivariate Cox regression analysis suggested that the prognostic risk model could be an independent prognostic factor for LUAD patients. A nomogram that incorporated the signature and clinical features was additionally built for prognostic prediction. Moreover, the levels of hub genes were related to immune cell infiltration in LUAD microenvironments. A CMap analysis identified 13 small molecule drugs as potential agents based on the risk model for LUAD treatment. Thus, we identified a prognostic risk model including CBFA2T3, CR2, SEL1L3, TM6SF1, TSPAN32, ITGA6, MAPK11, RASA3, and TLR6 as novel biomarkers and validated their prognostic and predicted values for LUAD.
肺腺癌(LUAD)是最常见的恶性肿瘤,每年导致超过100万人死亡。由于长期生存率较低,迫切需要探索LUAD进展的分子机制和新的预后预测指标,以改善LUAD的治疗。在我们的研究中,使用稳健秩聚合(RRA)方法从六个基因表达综合数据库(GSE43458、GSE62949、GSE68465、GSE115002、GSE116959和GSE118370)的肿瘤组织和正常组织中鉴定癌症特异性差异表达基因(DEG),随后使用加权基因共表达网络分析选择预后模块。应用单因素Cox回归、最小绝对收缩和选择算子(LASSO)以及多因素Cox回归分析来鉴定构建预后风险模型的九个枢纽基因()。通过RNA测序和单细胞RNA测序在肿瘤组织和正常组织中验证了九个枢纽基因的RNA表达,而来自人类蛋白质图谱数据库的免疫组织化学染色在蛋白质水平上显示出一致的结果。风险模型显示,高危患者的预后较差,包括晚期和低生存率。此外,多因素Cox回归分析表明,预后风险模型可能是LUAD患者的独立预后因素。还构建了一个纳入该特征和临床特征的列线图用于预后预测。此外,枢纽基因的水平与LUAD微环境中的免疫细胞浸润有关。CMap分析基于LUAD治疗的风险模型鉴定出13种小分子药物作为潜在药物。因此,我们鉴定了一个包括CBFA2T3、CR2、SEL1L3、TM6SF1、TSPAN32、ITGA6、MAPK11、RASA3和TLR6的预后风险模型作为新的生物标志物,并验证了它们对LUAD的预后和预测价值。