基于机器学习计算框架的骨关节炎中巨噬细胞极化相关特征的多组学表征
Multi-omics characterization of macrophage polarization-related features in osteoarthritis based on a machine learning computational framework.
作者信息
Hu Ping, Li Beining, Yin Zhenyu, Peng Peng, Cao Jiangang, Xie Wanyu, Liu Liang, Cao Fujiang, Zhang Bin
机构信息
Department of Othopaedics, Tianjin Medical University General Hospital, 154 Anshan Road, Heping District, Tianjin, 300052, China.
International Science and Technology Cooperation Base of Spinal Cord Injury, Tianjin Key Laboratory of Spine and Spinal Cord Injury, Department of Orthopedics, Tianjin Medical University General Hospital, Tianjin, China.
出版信息
Heliyon. 2024 Apr 27;10(9):e30335. doi: 10.1016/j.heliyon.2024.e30335. eCollection 2024 May 15.
BACKGROUND
OA imposes a heavy burden on patients and society in that its mechanism is still unclear, and there is a lack of effective targeted therapy other than surgery.
METHODS
The osteoarthritis dataset GSE55235 was downloaded from the GEO database and analyzed for differential genes by limma package, followed by analysis of immune-related modules by xcell immune infiltration combined with the WGCNA method, and macrophage polarization-related genes were downloaded according to the Genecard database, and VennDiagram was used to determine their intersection. These genes were also subjected to gene ontology (GO), disease ontology (DO), and Kyoto Encyclopedia of Genes and Genomes (KEGG) functional enrichment analyses. Using machine learning, the key osteoarthritis genes were finally screened. Using single gene GSEA and GSVA, we examined the significance of these key gene functions in immune cell and macrophage pathways. Next, we confirmed the correctness of the hub gene expression profile using the GSE55457 dataset and the ROC curve. Finally, we projected TF, miRNA, and possible therapeutic drugs using the miRNet, TargetScanHuman, ENCOR, and NetworkAnalyst databases, as well as Enrichr.
RESULTS
VennDiagram obtained 71 crossover genes for DEGs, WGCNA-immune modules, and Genecards; functional enrichment demonstrated NF-κB, IL-17 signaling pathway play an important role in osteoarthritis-macrophage polarization genes; machine learning finally identified CSF1R, CX3CR1, CEBPB, and TLR7 as hub genes; GSVA analysis showed that CSF1R, CEBPB play essential roles in immune infiltration and macrophage pathway; validation dataset GSE55457 analyzed hub genes were statistically different between osteoarthritis and healthy controls, and the AUC values of ROC for CSF1R, CX3CR1, CEBPB and TLR7 were more outstanding than 0.65.
CONCLUSIONS
CSF1R, CEBPB, CX3CR1, and TLR7 are potential diagnostic biomarkers for osteoarthritis, and CSF1R and CEBPB play an important role in regulating macrophage polarization in osteoarthritis progression and are expected to be new drug targets.
背景
骨关节炎（OA）因其发病机制仍不清楚，且除手术外缺乏有效的靶向治疗方法，给患者和社会带来了沉重负担。
方法
从基因表达综合数据库（GEO）下载骨关节炎数据集GSE55235，使用limma软件包分析差异基因，然后通过xcell免疫浸润结合加权基因共表达网络分析（WGCNA）方法分析免疫相关模块，根据基因卡片数据库下载巨噬细胞极化相关基因，并用VennDiagram确定它们的交集。对这些基因进行基因本体论（GO）、疾病本体论（DO）和京都基因与基因组百科全书（KEGG）功能富集分析。利用机器学习最终筛选出关键的骨关节炎基因。使用单基因基因集富集分析（GSEA）和基因集变异分析（GSVA），研究这些关键基因功能在免疫细胞和巨噬细胞途径中的意义。接下来,我们使用GSE55457数据集和ROC曲线验证了核心基因表达谱的正确性。最后，我们使用miRNet、TargetScanHuman、ENCOR和NetworkAnalyst数据库以及Enrichr预测转录因子（TF）、微小RNA（miRNA）和可能的治疗药物。
结果
VennDiagram得到了差异表达基因（DEGs）、WGCNA免疫模块和基因卡片的71个交叉基因；功能富集表明核因子κB（NF-κB）、白细胞介素-17信号通路在骨关节炎-巨噬细胞极化基因中起重要作用；机器学习最终确定集落刺激因子1受体（CSF1R）、CX3C趋化因子受体1（CX3CR1）、CCAAT增强子结合蛋白β（CEBPB）和Toll样受体7（TLR7）为核心基因；GSVA分析表明CSF1R、CEBPB在免疫浸润和巨噬细胞途径中起重要作用；验证数据集GSE55457分析显示核心基因在骨关节炎患者和健康对照之间存在统计学差异，CSF1R、CX3CR1、CEBPB和TLR7的ROC曲线下面积（AUC）值均大于0.65。
结论
CSF1R、CEBPB、CX3CR1和TLR7是骨关节炎潜在的诊断生物标志物，CSF1R和CEBPB在骨关节炎进展过程中调节巨噬细胞极化方面起重要作用，有望成为新的药物靶点。
Zotero 插件