Table 3_Immune-related diagnostic indicators and targeted therapies for COPD combined with NASH were identified and verified via WGCNA and LASSO.xlsx

IntroductionThe incidence of chronic obstructive pulmonary disease (COPD) and non-alcoholic fatty liver disease (NAFLD) has increased significantly in past decades, posing a significant public health burden. An increasing amount of research points to a connection between COPD and NAFLD. This study aimed to identify the key genes of these two diseases, construct a diagnostic model, and predict potential therapeutic agents based on critical genes.
MethodsNAFLD and COPD datasets were obtained from the GEO database, differential genes were identified by differential analysis and WGCNA, PPI networks were constructed and enriched for differential genes and COPD-associated secreted proteins, small molecule compounds were screened, and immune cell infiltration was assessed. Meanwhile, LASSO and RF further screened the essential genes, and finally, two key genes were obtained. Subsequently, the nomogram diagnostic model and lncRNA-miRNA-mRNA network were constructed based on these two core genes, subjected to drug prediction and GSEA enrichment analysis, and validated in an external cohort using qRT-PCR.
ResultsKEGG enrichment analysis indicated that the NF-kappa B and TNF signaling pathways may be associated with COPD and NASH co-morbidities. Ten small-molecule drugs associated with COPD and NASH were identified through cMAP analysis, including ansoprazole and atovaquone. In addition, we further identified the hub genes S100A9 and MYH2 for NAFLD and COPD by machine learning methods. The immune infiltration indicated that these two core genes might be involved in the immunomodulatory process of NASH by regulating the function or recruitment of specific immune cell types. A nomogram diagnostic model was constructed based on these two core genes. The AUC value for S100A9 was 0.887, for MYH2 was 0.877, and for the nomogram was 0.889, demonstrating excellent diagnostic efficacy. Two hundred fifty-four potential drugs targeting S100A9 and 67 MYH2 were searched in the DGIdb database. Meanwhile, the lncRNA-miRNA-mRNA network was constructed by predicting target miRNAs of biomarkers and further predicting lncRNAs targeting miRNAs. qRT-PCR analysis revealed that S100A9 was upregulated in both COPD and NAFLD, consistent with bioinformatic predictions, while MYH2 showed increased expression in COPD but decreased expression in NAFLD, diverging from the predicted downregulation in both diseases. These findings suggest that S100A9 serves as a common inflammatory marker for both diseases, whereas MYH2 may be regulated by disease-specific mechanisms, highlighting its potential for distinguishing COPD from NAFLD.
ConclusionThe hub genes S100A9 and MYH2 in COPD and NASH were identified by various bioinformatics methods and a diagnostic model was constructed to improve the diagnostic efficiency. We also revealed some potential biological mechanisms of COPD and NASH and potential drugs for COPD-related NASH. Our findings provide potential new diagnostic and therapeutic options for COPD-associated NASH and may help reduce its prevalence.

引言：近数十年来，慢性阻塞性肺疾病（Chronic Obstructive Pulmonary Disease, COPD）与非酒精性脂肪性肝病（Non-Alcoholic Fatty Liver Disease, NAFLD）的发病率显著上升，给公共卫生带来了沉重负担。越来越多的研究表明，COPD与NAFLD之间存在密切关联。本研究旨在明确这两种疾病的关键基因，构建诊断模型，并基于关键基因预测潜在治疗药物。
方法：本研究从基因表达综合数据库（Gene Expression Omnibus, GEO）获取NAFLD与COPD数据集，通过差异表达分析与加权基因共表达网络分析（Weighted Gene Co-expression Network Analysis, WGCNA）筛选差异基因；构建蛋白质相互作用（Protein-Protein Interaction, PPI）网络，并对差异基因及COPD相关分泌蛋白进行富集分析；筛选小分子化合物，评估免疫细胞浸润情况。同时，采用最小绝对收缩和选择算子（Least Absolute Shrinkage and Selection Operator, LASSO）与随机森林（Random Forest, RF）进一步筛选核心基因，最终得到2个关键基因。基于这两个核心基因构建列线图（nomogram）诊断模型与长链非编码RNA-微小RNA-信使RNA（long non-coding RNA-microRNA-messenger RNA, lncRNA-miRNA-mRNA）调控网络，进行药物预测与基因集富集分析（Gene Set Enrichment Analysis, GSEA），并通过实时荧光定量聚合酶链反应（quantitative Real-Time Polymerase Chain Reaction, qRT-PCR）在外部队列中进行验证。
结果：京都基因与基因组百科全书（Kyoto Encyclopedia of Genes and Genomes, KEGG）富集分析显示，核因子κB（Nuclear Factor-kappa B, NF-κB）与肿瘤坏死因子（Tumor Necrosis Factor, TNF）信号通路可能与COPD及非酒精性脂肪性肝炎（Non-Alcoholic Steatohepatitis, NASH）共病相关。通过连通性图谱（Connectivity Map, cMAP）分析，本研究筛选出10种与COPD和NASH相关的小分子药物，包括安索拉唑（ansoprazole）与阿托伐醌（atovaquone）。此外，通过机器学习方法进一步筛选得到NAFLD与COPD的核心枢纽基因S100A9和MYH2。免疫浸润分析表明，这两个核心基因可能通过调控特定免疫细胞类型的功能或募集，参与NASH的免疫调节过程。基于这两个核心基因构建的列线图诊断模型，其受试者工作特征曲线下面积（Area Under the Receiver Operating Characteristic Curve, AUC）分别为：S100A9的AUC为0.887，MYH2的AUC为0.877，列线图整体AUC为0.889，展现出优异的诊断效能。在药物相互作用基因数据库（Drug Gene Interaction Database, DGIdb）中检索得到254种靶向S100A9的潜在药物，以及67种靶向MYH2的潜在药物。同时，通过预测生物标志物的靶向miRNA，并进一步预测靶向这些miRNA的lncRNA，构建了lncRNA-miRNA-mRNA调控网络。qRT-PCR分析显示，S100A9在COPD与NAFLD中均呈上调表达，与生物信息学预测结果一致；而MYH2在COPD中表达升高，但在NAFLD中表达降低，与两种疾病中均下调的预测结果不符。上述研究结果表明，S100A9可作为两种疾病共有的炎症标志物，而MYH2可能受疾病特异性机制调控，其有望用于区分COPD与NAFLD。
结论：本研究通过多种生物信息学方法筛选得到COPD与NASH的核心枢纽基因S100A9和MYH2，并构建了诊断模型以提升诊断效能。本研究同时揭示了COPD与NASH的潜在生物学机制，以及针对COPD相关NASH的潜在治疗药物。本研究结果为COPD相关性NASH提供了潜在的新型诊断与治疗策略，或有助于降低其发病率。