中国组织工程研究 ›› 2025, Vol. 29 ›› Issue (32): 6893-6904.doi: 10.12307/2025.912

• 脊柱组织构建 spinal tissue construction • 上一篇    下一篇

脊髓损伤后氧化应激相关基因及分子机制:基于GEO数据库的数据分析及验证

王子恒,吴  霜   

  1. 贵州医科大学附属医院康复医学科,贵州省贵阳市  550000
  • 收稿日期:2024-07-13 接受日期:2024-10-31 出版日期:2025-11-18 发布日期:2025-04-26
  • 通讯作者: 吴霜,博士生导师,主任医师,主任,贵州医科大学附属医院康复医学科,贵州省贵阳市 550000
  • 作者简介:王子恒,1998年生,贵州省遵义市人,硕士,主要从事神经康复方面的研究。
  • 基金资助:
    国家自然科学基金地区项目(82060419,82260452):项目负责人:吴霜;贵州省科技计划项目[黔科合基础-ZK(2022)重点045]:项目负责人:吴霜

Oxidative stress-related genes and molecular mechanisms after spinal cord injury: data analysis and verification based on GEO database

Wang Ziheng, Wu Shuang   

  1. Department of Rehabilitation Medicine, Affiliated Hospital of Guizhou Medical University, Guiyang 550000, Guizhou Province, China 
  • Received:2024-07-13 Accepted:2024-10-31 Online:2025-11-18 Published:2025-04-26
  • Contact: Wu Shuang, Doctoral supervisor, Chief physician, Department of Rehabilitation Medicine, Affiliated Hospital of Guizhou Medical University, Guiyang 550000, Guizhou Province, China
  • About author:Wang Ziheng, MS, Department of Rehabilitation Medicine, Affiliated Hospital of Guizhou Medical University, Guiyang 550000, Guizhou Province, China
  • Supported by:
    National Natural Science Foundation of China (Regional Project), Nos. 82060419 and 82260452 (both to WS); Guizhou Provincial Science and Technology Plan Project, No. ZK(2022)045 (to WS) 

摘要:


文题释义:
生物信息分析:通过信息科学的方法和技术,探究生物体系和过程中的信息存储、内涵及传递,并分析生物分子数据,为理解生命复杂性提供新视角。此次研究利用了两种机器算法(最小绝对收缩和选择算法和支持向量机递归特征消除)鉴定出与氧化应激相关的生物标志物。最小绝对收缩和选择算法在变量选择、正则化、降低过拟合及高维数据处理上表现出色,而支持向量机递归特征消除则适用于多种数据类型并提升模型可解释性,结合两种算法,可提高生物标志物鉴定的准确性和可靠性。
脊髓损伤:可源自创伤、疾病或先天性因素,此损伤会阻碍损伤部位以下的感觉与运动功能,极大地削弱患者的活动能力、自理能力和工作能力。当前,氧化应激相关基因在脊髓损伤进展中的具体作用仍不明确。故而深入探究氧化应激相关基因在脊髓损伤演变过程中的角色,并明晰其背后的分子机制,对于促进脊髓损伤治疗手段的进步具有至关重要的意义。

背景:在脊髓损伤中,氧化应激的生物学意义尚未得到系统研究。
目的:探讨氧化应激相关基因在脊髓损伤发展过程中的作用并阐明相关的分子机制。
方法:从基因表达综合数据库(GEO)获取了GSE151371数据集,并从GeneCards数据库中筛选出了与氧化应激高度相关的基因集(包含相关得分≥7的899个基因)。首先,对脊髓损伤患者和健康对照组的样本进行了差异表达基因分析,之后将这899个氧化应激基因与差异表达基因进行交叉比对,识别出与脊髓损伤紧密关联的氧化应激相关基因。采用最小绝对值收缩和选择算子以及支持向量机递归特征消除算法,基于与脊髓损伤紧密关联的氧化应激相关基因推导出潜在的生物标志物,对这些生物标志物进行了基因本体论和京都基因与基因组百科全书富集分析,以探究其在生物学过程中的作用及参与的信号通路。此外,构建了一个竞争性内源性RNA网络,对免疫微环境进行了深入分析,并预测了与这些生物标志物相关的转录因子,探索了可能的潜在治疗药物。从贵州医科大学附属医院收集了8例急性期脊髓损伤患者和7例对照组的15份血液样本进行RT-qPCR分析,评估生物标志物在脊髓损伤和对照样本之间的表达情况。
结果与结论:①脊髓损伤和对照样本之间有2 511个差异表达基因,通过交叉分析氧化应激基因和差异表达基因,得到了151个与脊髓损伤紧密关联的氧化应激相关基因;通过最小绝对值收缩和选择算子及支持向量机递归特征消除算法鉴定出6个生物标志物,分别为S100A9、PLAU、CASP4、GAPDH、CYP1B1和HSPA1B,它们共同参与了对烟酰胺腺嘌呤二核苷酸磷酸的氧化还原酶活性、氧化磷酸化等;②6个信使RNA (mRNA)、8个微小RNA(miRNA)和50个长链非编码RNA (lncRNA)构成了mRNA-miRNA-lncRNA网络,包括S100A9-hsa-mir-16-5p-HCG18、CYP1B1-hsa-mir-429-KCNQ1OT1等关系对;③在脊髓损伤和对照样本之间,9种免疫细胞的浸润水平明显不同,如浆细胞、静息的自然杀伤细胞、M0巨噬细胞等;④预测的前5个转录因子是FOSB、ZNF581、SP110、NFIL3和BATF2,其中ZNF581和NFIL3在脊髓损伤和对照样本之间的表达差异显著;⑤基于6个生物标志物预测了多种药物,分子对接结果显示 β-萘黄酮与CYP1B1的靶蛋白具有良好的结合活性,它们形成的相互作用对包括DB01373-S100A9、DB01065-CYP1B1等;⑥在临床RT-qPCR验证中,S100A9、GAPDH、CYP1B1和HSPA1B在脊髓损伤组和对照组之间显示出显著差异,而PLAU和CASP4在脊髓损伤组中呈上升趋势;⑦此次研究识别了6个关键生物标志物,包括S100A9、PLAU、CASP4、GAPDH、CYP1B1和HSPA1B,然后基于6个生物标志物预测了100种小分子药物,对国人的健康管理和疾病防治具有重要的借鉴意义,为国人在脊髓损伤治疗方案选择上提供了新的思路。
https://orcid.org/0009-0005-6672-5033(王子恒)

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: 脊髓损伤, 氧化应激, 差异表达基因, 生物信息学, 生物标志物, 工程化组织构建

Abstract: BACKGROUND: The biological significance of oxidative stress in spinal cord injury has not been systematically investigated. 
OBJECTIVE: To investigate the role of oxidative stress-related genes in the development of spinal cord injury and to elucidate the related molecular mechanisms. 
METHODS: The GSE151371 dataset was obtained from the Gene Expression Omnibus (GEO) database and selected a gene set highly relevant to oxidative stress from the GeneCards database (including 899 genes with relevance scores of no less than 7). Firstly, we performed differential expression gene analysis on samples from spinal cord injury patients and healthy controls. Subsequently, we cross-referenced these 899 oxidative stress genes with the differential expression genes to identify oxidative stress-related genes closely associated with spinal cord injury (spinal cord injury-reactive oxygen species-related genes). Then, the Least Absolute Shrinkage and Selection Operator and Support Vector Machine-Recursive Feature Elimination algorithms were used to derive potential biomarkers based on the spinal cord injury-reactive oxygen species-related genes. Following this, Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analyses on these biomarkers were conducted to explore their roles in biological processes and the signaling pathways they participate in. Additionally, we constructed a competing endogenous RNA network and performed an in-depth analysis of the immune microenvironment. We predicted transcription factors related to these biomarkers and explored potential therapeutic drugs. Finally, fifteen blood samples from eight patients with acute spinal cord injury and seven controls were collected from Affiliated Hospital of Guizhou Medical University for reverse transcription quantitative polymerase chain reaction analysis to evaluate the expression of biomarkers between spinal cord injury and control samples. 
RESULTS AND CONCLUSION: (1) There were 2 511 differential expression genes between spinal cord injury and control samples. Cross-analysis of oxidative stress genes and differential expression gene resulted in 151 spinal cord injury-reactive oxygen species-related genes. Least Absolute Shrinkage and Selection Operator and Support Vector Machine-Recursive Feature Elimination identified six biomarkers-S100A9, PLAU, CASP4, GAPDH, CYP1B1, and HSPA1B-involved in activities such as nicotinamide adenine dinucleotide phosphate oxidoreductase activity and oxidative phosphorylation. (2) Six messenger RNAs (mRNAs), eight microRNAs (miRNAs), and 50 long noncoding RNAs (lncRNAs) formed the mRNA-miRNA-lncRNA network, including relationships such as S100A9-hsa-mir-16-5p-HCG18 and CYP1B1-hsa-mir-429-KCNQ1OT1. (3) Infiltration levels of nine immune cells, including plasma cells, resting natural killer cells, and M0 macrophages, showed significant differences between spinal cord injury and control samples. (4) The top five predicted transcription factors were FOSB, ZNF581, SP110, NFIL3, and BATF2, with ZNF581 and NFIL3 showing significant expression differences between spinal cord injury and control samples. (5) The study predicted a variety of drugs based on the six biomarkers, and the molecular docking results showed that β-naphthoflavone had good binding activity to the target protein of CYP1B1. The interaction pairs they formed included DB01373-S100A9 and DB01065-CYP1B1. (6) In reverse transcription quantitative polymerase chain reaction verification, S100A9, GAPDH, CYP1B1, and HSPA1B showed significant differences between spinal cord injury and control groups, while PLAU and CASP4 showed an increasing trend in spinal cord injury group. (7) This study identified six key biomarkers, including S100A9, PLAU, CASP4, GAPDH, CYP1B1, and HSPA1B, and subsequently predicted 100 small-molecule drugs based on these biomarkers. These findings provide valuable insights for health management and disease prevention among the Chinese population, offering new perspectives for spinal cord injury treatment options.

Key words: spinal cord injury, oxidative stress, differentially expressed gene, bioinformatics, biomarker, engineered tissue construction

中图分类号: