中国组织工程研究 ›› 2024, Vol. 28 ›› Issue (27): 4312-4318.doi: 10.12307/2024.571

• 组织工程相关大数据分析 Big data analysis in tissue engineering • 上一篇    下一篇

类风湿性关节炎中免疫相关基因表达及免疫细胞的双样本孟德尔随机化分析

范以东1,秦  刚2,何凯毅2,龚玉芳3,李威材1,吴广涛1   

  1. 1广西中医药大学,广西壮族自治区南宁市  530222;2广西中医药大学第一附属医院,广西壮族自治区南宁市  530022;3贵港市人民医院,广西壮族自治区贵港市  537199
  • 收稿日期:2023-10-08 接受日期:2023-11-25 出版日期:2024-09-28 发布日期:2024-01-26
  • 通讯作者: 秦刚,博士,主任医师,广西中医药大学第一附属医院,广西壮族自治区南宁市 530022
  • 作者简介:范以东,男,1993年生,河南省辉县市人,汉族,广西中医药大学在读硕士,主要从事骨关节退变与缺血性疾病的防治研究。
  • 基金资助:
    国家自然科学基金(81860793),项目名称:基于IFITM1/Wnt/β-catenin信号通路调控骨肉瘤细胞恶性增殖探讨抑瘤汤抗骨肉瘤的机制,项目负责人:秦刚;国家自然科学基金(82360939),项目名称:circ_CDK14激活miR-520a-3p/SUV39H1诱导骨肉瘤细胞铁死亡及抑瘤汤防治机制研究,项目负责人:秦刚;广西自然科学基金(2020GXNSFAA297140),项目名称:抑瘤汤通过LncRNA MALAT1激活MAPK通路磷酸化抑制骨肉瘤的基础研究,项目负责人:秦刚;广西中医药大学2022年研究生教育创新计划项目(YCSY2022028),项目名称:枸杞多糖通过miRNA-20a介导的PTEN/PI3K/AKT信号通路诱导骨肉瘤细胞自噬的机制研究,项目负责人:范以东

Expression of immune-related genes in rheumatoid arthritis and a two-sample Mendelian randomization study of immune cells

Fan Yidong1, Qin Gang2, He Kaiyi2, Gong Yufang3, Li Weicai1, Wu Guangtao1   

  1. 1Guangxi University of Chinese Medicine, Nanning 530222, Guangxi Zhuang Autonomous Region, China; 2The First Affiliated Hospital of Guangxi University of Chinese Medicine, Guangxi Zhuang Autonomous Region, China; 3The People’s Hospital of Guigang, Guigang 537199, Guangxi Zhuang Autonomous Region, China
  • Received:2023-10-08 Accepted:2023-11-25 Online:2024-09-28 Published:2024-01-26
  • Contact: Qin Gang, MD, Chief physician, The First Affiliated Hospital of Guangxi University of Chinese Medicine, Guangxi Zhuang Autonomous Region, China
  • About author:Fan Yidong, Master candidate, Guangxi University of Chinese Medicine, Nanning 530222, Guangxi Zhuang Autonomous Region, China
  • Supported by:
    National Natural Science Foundation of China, Nos. 81860793 and 82360939 (both to QG); Natural Science Foundation of Guangxi Zhuang Autonomous Region, No. 2020GXNSFAA297140 (to QG); Postgraduate Education Innovation Program of Guangxi University of Chinese Medicine in 2022, No. YCSY2022028 (to FYD)

摘要:


文题释义:

孟德尔随机化:是一种遵循孟德尔遗传定律的遗传变量分析,该方法能利用单核苷酸多态性作为工具变量来评估暴露因素与临床结局之间的因果关系。
单核细胞:是人体内重要的先天性免疫细胞,能分泌细胞因子、趋化因子和炎症递质等促进炎症反应的进行。


背景:类风湿性关节炎是一种慢性全身性自身免疫疾病,研究其涉及的免疫学特征改变对诊断和治疗具有重要意义。

目的:通过生物信息学方法筛选类风湿性关节炎中免疫相关生物标志物,分析其免疫细胞浸润变化及与免疫细胞的潜在因果关系。
方法基于GEO与ImmPort数据库,采用差异表达分析筛选类风湿性关节炎中表达上调的免疫相关基因,采用GO与KEGG富集分析探索这些上调基因的潜在作用。采用支持向量机递归特征消除(SVM-RFE)、最小绝对收缩和选择算法(Lasso)筛选类风湿性关节炎中免疫特征基因,独立的数据集进行差异验证,绘制特征基因的受试者工作特征曲线评价诊断性能。采用免疫细胞浸润分析类风湿性关节炎中免疫细胞的差异情况及特征基因与免疫细胞的相关性。最后,进行孟德尔随机化分析,以确定免疫细胞中单核细胞与类风湿性关节炎的因果效应。

结果与结论:①筛选得到了39个在类风湿性关节炎中高表达的差异基因,GO富集分析显示其主要富集在趋化性、细胞因子活性、免疫受体活性等过程,KEGG富集分析显示其在肿瘤坏死因子信号通路、白细胞外周迁移等途径显著富集;②通过Lasso与SVM-RFE两种机器学习算法确定了5个特征基因,分别为CXCL13、SDC1、IGLC1、PLXNC1和SLC29A3,独立的数据集对特征基因验证发现他们同样在类风湿性关节炎样本中高表达,5个特征基因在2个数据集中的AUC值均大于0.8;③免疫细胞浸润发现大多数免疫细胞如单核细胞、自然杀伤细胞等在类风湿性关节炎样本里表现出更高的浸润水平,相关性分析显示5个特征基因与记忆B细胞、不成熟B细胞等呈正相关性,逆方差加权法发现单核细胞与类风湿性关节炎发病风险相关。

https://orcid.org/0009-0000-3044-7311(范以东)

中国组织工程研究杂志出版内容重点:组织构建;骨细胞;软骨细胞;细胞培养;成纤维细胞;血管内皮细胞;骨质疏松;组织工程

关键词: 类风湿性关节炎, 机器学习, 免疫浸润, 全基因组关联研究, 孟德尔随机化

Abstract: BACKGROUND: Rheumatoid arthritis is a chronic systemic autoimmune disease. It is important to study the immunological changes involved in it for diagnosis and treatment.
OBJECTIVE: To identify immune-related biomarkers associated with rheumatoid arthritis utilizing bioinformatics techniques and examine alterations in immune cell infiltration as well as the relationship between immune cells and biomarkers.
METHODS: Differential expression analysis was used to identify the immune-related genes that were up-regulated in rheumatoid arthritis based on the GEO and Immport databases. Kyoto encyclopedia of genes and genomes (KEGG) and gene ontology (GO) enrichment analyses were used to investigate the possible function of these elevated genes. The immunological characteristic genes associated with rheumatoid arthritis were screened using least absolute shrinkage and selection operator (Lasso) and support vector machine recursive feature elimination (SVM-RFE). Independent datasets were used for difference validation, and the diagnostic performance was evaluated by plotting receiver operating characteristic curves for feature genes. Immune cell infiltration was used to analyze the differential profile of immune cells in rheumatoid arthritis and the correlation between the characterized genes and immune cells. In order to ascertain the causal relationship between monocytes and rheumatoid arthritis in immune cells, Mendelian randomization analysis was ultimately employed.
RESULTS AND CONCLUSION: There were 39 upregulated differentially expressed genes in rheumatoid arthritis. The genes were primarily enriched in chemotaxis, cytokine activity, and immune receptor activity, according to GO enrichment analysis, while kEGG enrichment analysis revealed that the genes were considerably enriched in the tumor necrosis factor signaling pathway and peripheral leukocyte migration. Lasso and SVM-RFE identified five feature genes: CXCL13, SDC1, IGLC1, PLXNC1, and SLC29A3. Independent dataset validation of the feature genes found them to be similarly highly expressed in rheumatoid arthritis samples, with area under the curve values greater than 0.8 for all five feature genes in both datasets. Immune cell infiltration indicated that most immune cells, including natural killer cells and monocytes, exhibited increased levels of infiltration in rheumatoid arthritis samples. The correlation analysis revealed a significant positive correlation between memory B cells and immature B cells and these five feature genes. Correlation analysis showed that the five feature genes were positively correlated with memory B cells and immature B cells. The inverse variance weighting method revealed that monocytes were associated with the risk of developing rheumatoid arthritis.

Key words: rheumatoid arthritis, machine learning, immune infiltration, genome-wide association studies, Mendelian randomization

中图分类号: