中国神经再生研究(英文版)

中国神经再生研究(英文版) ›› 2026, Vol. 21 ›› Issue (6): 2442-2453..doi: 10.4103/NRR.NRR-D-24-00037

• 原著:脊髓损伤修复保护与再生 • 上一篇    下一篇

机器学习确定脊髓损伤后铁死亡过程中的关键细胞及治疗靶点

  

  • 出版日期:2026-06-15 发布日期:2025-09-19

Machine learning identifies key cells and therapeutic targets during ferroptosis after spinal cord injury

Yigang Lv1, #, Zhen Li1, #, Lusen Shi2, #, Huan Jian1 , Fan Yang3, 4, Jichuan Qiu5 , Chao Li1 , Peng Xiao6 , Wendong Ruan1 , Hao Li2 , Xueying Li4, *, Shiqing Feng1, 2, 4, 7, *, Hengxing Zhou1, 2, 4, 8, *
  

  1. 1 Department of Orthopedics, Tianjin Medical University General Hospital, International Science and Technology Cooperation Base of Spinal Cord Injury, Tianjin Key Laboratory of Spine and Spinal Cord, Tianjin, China;  2 Department of Orthopedics, Qilu Hospital of Shandong University, Cheeloo College of Medicine, Shandong University, Jinan, Shandong Province, China;  3 Key Laboratory Experimental Teratology of the Ministry of Education and Department of Physiology and Pathophysiology, School of Basic Medical Sciences, Shandong University, Jinan, Shandong Province, China;  4 Shandong University Center for Orthopedics, Advanced Medical Research Institute, Cheeloo College of Medicine, Shandong University, Jinan, Shandong Province, China;  5 State Key Laboratory of Crystal Materials, Shandong University, Jinan, Shandong Province, China;  6 Key Laboratory Experimental Teratology of the Ministry of Education, Department of Biochemistry and Molecular Biology, School of Basic Medical Sciences, Cheeloo College of Medicine, Shandong University, Jinan, Shandong Province, China;  7 Department of Orthopedics, The Second Hospital of Shandong University, Cheeloo College of Medicine, Shandong University, Jinan, Shandong Province, China;  8 Center for Reproductive Medicine, Shandong University, Jinan, Shandong Province, China
  • Online:2026-06-15 Published:2025-09-19
  • Contact: Hengxing Zhou, PhD, zhouhengxing@sdu.edu.cn; Shiqing Feng, PhD, sqfeng@tmu.edu.cn or shiqingfeng@sdu.edu.cn; Xueying Li, MS, xueyingli@sdu.edu.cn.
  • Supported by:

    This study was supported by the National Natural Science Foundation of China, No. 81972073 (to HZ); a grant from the Taishan Scholars Program of Shandong Province-Young Taishan Scholars, No. tsqn201909197 (to HZ); a grant from Tianjin Key Medical Discipline (Specialty) Construct Project, No. TJYXZDXK027A (to SF); and a grant from Academic Expert International Innovation Summit, No. 22JRRCRC00010 (to SF).

PDF

5

摘要：

铁死亡是一种主要涉及铁代谢失衡和脂质过氧化的细胞死亡，与脊髓损伤后出血引起的吞噬反应密切相关。因此实验首先从基因表达数据库中获取GSE47681和GSE5296的批量RNA测序数据以及GSE162610的单细胞RNA测序数据进行差异分析和免疫浸润分析，继而通过随机森林和最小绝对收缩和最小绝对收缩与选择算子（LASSO）算法确定Atf3和Piezo1为铁死亡的关键基因。对单细胞RNA测序数据的进一步分析揭示了铁死亡与巨噬细胞/小胶质细胞等细胞的密切关系以及它们的内在状态转换过程。在铁死亡相关细胞中发现了转录因子调节和细胞间通讯网络的差异，证实了关键基因Atf3和Piezo1在这些细胞中存在高表达。进一步分子对接结果证实，这些基因编码的蛋白质可与环己酰亚胺结合。最后以T8脊髓损伤小鼠模型验证，发现低剂量环己酰亚胺可改善小鼠亚急性期损伤后的神经功能，降低促炎细胞因子诱导性一氧化氮合酶的水平，并增加抗炎细胞因子Arg-1水平，同时巨噬细胞/小胶质细胞中与铁死亡相关的基因Gpx4的表达增加，而Acsl4表达减少。该实验通过识别脊髓损伤后铁死亡的关键细胞类型和基因以及验证潜在药物治疗的有效性，揭示了铁死亡在脊髓损伤治疗中的重要作用，为脊髓损伤的治疗指明了新的方向。

https://orcid.org/0000-0003-0053-8970 (Hengxing Zhou); https://orcid.org/0000-0001-9437-7674 (Shiqing Feng); 
https://orcid.org/0009-0008-7589-8499 (Xueying Li)

关键词: 机器学习, 铁死亡, 生物信息学分析, 批量RNA测序, 单细胞RNA测序, RNA速度分析, 转录因子分析, 细胞通信分析, 治疗药物, 神经功能

Abstract: Ferroptosis, a type of cell death that mainly involves iron metabolism imbalance and lipid peroxidation, is strongly correlated with the phagocytic response caused by bleeding after spinal cord injury. Thus, in this study, bulk RNA sequencing data (GSE47681 and GSE5296) and single-cell RNA sequencing data (GSE162610) were acquired from gene expression databases. We then conducted differential analysis and immune infiltration analysis. Atf3 and Piezo1 were identified as key ferroptosis genes through random forest and least absolute shrinkage and selection operator algorithms. Further analysis of single-cell RNA sequencing data revealed a close relationship between ferroptosis and cell types such as macrophages/microglia and their intrinsic state transition processes. Differences in transcription factor regulation and intercellular communication networks were found in ferroptosis-related cells, confirming the high expression of Atf3 and Piezo1 in these cells. Molecular docking analysis confirmed that the proteins encoded by these genes can bind cycloheximide. In a mouse model of T8 spinal cord injury, low-dose cycloheximide treatment was found to improve neurological function, decrease levels of the pro-inflammatory cytokine inducible nitric oxide synthase, and increase levels of the anti-inflammatory cytokine arginase 1. Correspondingly, the expression of the ferroptosis-related gene Gpx4 increased in macrophages/microglia, while the expression of Acsl4 decreased. Our findings reveal the important role of ferroptosis in the treatment of spinal cord injury, identify the key cell types and genes involved in ferroptosis after spinal cord injury, and validate the efficacy of potential drug therapies, pointing to new directions in the treatment of spinal cord injury.

Key words: bioinformatic analyses, bulk-RNA sequencing, cellular communication analysis, ferroptosis, machine learning analysis, neurological function, RNA velocity analysis, single-cell RNA sequencing, therapeutic drugs, transcription factor analysis