Identification of the core genes to endoplasmic reticulum stress in chronic obstructive pulmonary disease based on multi-chip integrated analysis of GEO database and experimental validation.

Objective Bioinformatics and experimental validation reveal the core genes of endoplasmic reticulum stress(ERS) in chronic obstructive pulmonary disease.Methods The microarray datasets GSE5058、GSE8545andGSE19407were downloaded from the Gene Expression Omnibus (GEO) database to identify differentially expressed genes (DEGs) between COPD smokers and non-smokers airway epithelial，and the common differentially expressed genes were obtained after overlapping with endoplasmic reticulum stress-related genes and enriched for analysis.Three machine learning algorithms, LASSO, SVM-RFE, and RF, were used to screen the characteristic genes of endoplasmic reticulum stress, and their diagnostic performance was verified and evaluated in the GSE10006, next, CIBERSORT algorithm was used to analyze immune cell infiltration.Finally the mRNA-miRNA network was constructed and the mRNA expression level of lung tissue in mouse emphysema model were verified.Results A total of 153 common DEGs were acquired. GO analysis showed that they were involved in endoplasmic reticulum stress response、protein folding、ubiquitin-like protease and other processes, KEGG analysis mainly involved protein processing、lipid and atherosclerosis、P53 signaling pathway、PI3K-Akt signaling pathway etc ,and DO analysis was mainly enriched in pulmonary vascular occlusive disease and chronic obstructive pulmonary disease.Immune infiltration analysis showed that COPD samples contained significantly more M0 macrophages、resting dendritic cells and eosinophils, while regulatory T cells and M2 macrophages were significantly less.A total of four key endoplasmic reticulum stress genes (including THBS1, BCL2, USP13 and RNFT2) were identified by machine learning algorithms, and they showed good diagnostic performance in both the training and validation sets. At the same time, co-expressed mRNA and miRNA were selected to construct the mRNA-miRNA interaction network.Compared with the air-exposed mice, the mRNA expression levels of THBS1 and RNFT2 in the lung tissues of emphysema mice induced by cigarette smoke exposure were statistically significant, and the mRNA expression levels of BCL2 and USP13 decreased (P<0.05).Conclusion THBS1、BCL2、USP13 and RNFT2 may be the core genes formed by endoplasmic reticulum stress during the pathogenesis of chronic obstructive pulmonary disease, which are expected to be the targets of immunotherapy for chronic obstructive pulmonary disease.

**研究目标**：本研究通过生物信息学分析与实验验证，揭示慢性阻塞性肺疾病（chronic obstructive pulmonary disease, COPD）中内质网应激（endoplasmic reticulum stress, ERS）的核心基因。**研究方法**：从基因表达综合数据库（Gene Expression Omnibus, GEO）下载微阵列数据集GSE5058、GSE8545与GSE19407，以识别吸烟与非吸烟COPD患者气道上皮细胞中的差异表达基因（differentially expressed genes, DEGs）；将所得差异表达基因与内质网应激相关基因取交集，获得共同差异表达基因并进行富集分析。采用LASSO、SVM-RFE及RF三种机器学习算法筛选内质网应激特征基因，并在数据集GSE10006中验证与评估其诊断效能；随后使用CIBERSORT算法分析免疫细胞浸润情况。最后构建mRNA-miRNA互作网络，并在小鼠肺气肿模型中验证肺组织的mRNA表达水平。**研究结果**：共获得153个共同差异表达基因。GO富集分析显示，这些基因参与内质网应激应答、蛋白质折叠、泛素样蛋白酶等生物学过程；KEGG富集分析主要涉及蛋白质加工、脂质与动脉粥样硬化、P53信号通路、PI3K-Akt信号通路等；DO富集分析主要集中于肺血管闭塞性疾病与慢性阻塞性肺疾病。免疫浸润分析显示，COPD样本中M0巨噬细胞、静息树突状细胞及嗜酸性粒细胞占比显著升高，而调节性T细胞与M2巨噬细胞占比显著降低。通过机器学习算法共筛选得到4个内质网应激关键基因（包括THBS1、BCL2、USP13及RNFT2），它们在训练集与验证集中均表现出良好的诊断效能；同时筛选共表达的mRNA与miRNA，构建mRNA-miRNA互作网络。与空气暴露组小鼠相比，香烟烟雾诱导的肺气肿小鼠肺组织中THBS1与RNFT2的mRNA表达水平显著升高，BCL2与USP13的mRNA表达水平显著降低（P<0.05）。**研究结论**：THBS1、BCL2、USP13及RNFT2可能是慢性阻塞性肺疾病发病过程中内质网应激介导的核心基因，有望成为慢性阻塞性肺疾病免疫治疗的潜在靶点。