Li Sijun, Sun Lanfeng, Huang Hongmi, Wei Xing, Lu Yuling, Qian Kai, Wu Yuan
Department of Neurology, the First Affiliated Hospital of Guangxi Medical University, Guangxi Medical University, Nanning, Guangxi, China.
Department of Neurology, the First Affiliated Hospital of Guangxi Medical University, Guangxi Medical University, Nanning, Guangxi, China.
Neurobiol Dis. 2025 Feb;205:106789. doi: 10.1016/j.nbd.2025.106789. Epub 2025 Jan 11.
One of the underlying mechanisms of epilepsy (EP), a brain disease characterized by recurrent seizures, is considered to be cell death. Disulfidptosis, a proposed novel cell death mechanism, is thought to play a part in the pathogenesis of epilepsy, but the exact role is unclear. The gene expression omnibus series (GSE) 33000 and GSE63808 datasets were used to search for differentially expressed disulfidptosis-related molecules (DE-DRMs). A correlation between the DE-DRMs was discovered. Individuals with epilepsy were then used to investigate molecular clusters based on the expression of DE-DRMs. Following that, the best machine learning model which is validated by GSE143272 dataset and predictor molecules were identified. The correlation between predictive molecules and clinical traits was determined. Based on the in vitro and in vivo seizures models, experimental analyses were applied to verify the DE-DRMs expressions and the correlation between them. Nine molecules were identified as DE-DRMs: glycogen synthase 1 (GYS1), solute carrier family 3 member 2 (SLC3A2), solute carrier family 7 member 11 (SLC7A11), NADH:ubiquinone oxidoreductase core subunit S1 (NDUFS1), 3-oxoacyl-ACP synthase, mitochondrial (OXSM), leucine rich pentatricopeptide repeat containing (LRPPRC), NADH:ubiquinone oxidoreductase subunit A11 (NDUFA11), NUBP iron‑sulfur cluster assembly factor, mitochondrial (NUBPL), and NCK associated protein 1 (NCKAP1). NDUFS1 interacted with NDUFA11, NUBPL, and LRPPRC, while SLC3A2 interacted with SLC7A11. The optimal machine learning model was revealed to be the random forest (RF) model. G protein guanine nucleotide-binding protein alpha subunit q (GNAQ) was linked to sodium valproate resistance. The experimental analyses suggested an upregulated SLC7A11 expression, an increased number of formed SLC3A2 and SLC7A11 complexes, and a decreased number of formed NDUFS1 and NDUFA11 complexes. This study provides previously undocumented evidence of the relationship between disulfidptosis and EP. In addition to suggesting that SLC7A11 may be a specific DRM for EP, this research demonstrates the alterations in two disulfidptosis-related protein complexes: SLC7A11-SLC3A2 and NDUFS1-NDUFA11.
癫痫（EP）是一种以反复发作性癫痫发作为特征的脑部疾病，其潜在机制之一被认为是细胞死亡。二硫化物诱导的细胞死亡（Disulfidptosis）是一种新提出的细胞死亡机制，被认为在癫痫发病机制中起作用，但确切作用尚不清楚。利用基因表达综合数据库系列（GSE）33000和GSE63808数据集搜索差异表达的二硫化物诱导的细胞死亡相关分子（DE-DRMs）。发现了DE-DRMs之间的相关性。然后利用癫痫患者基于DE-DRMs的表达研究分子簇。随后，确定了经GSE143272数据集验证的最佳机器学习模型和预测分子。确定了预测分子与临床特征之间的相关性。基于体外和体内癫痫模型，进行实验分析以验证DE-DRMs的表达及其之间的相关性。九个分子被鉴定为DE-DRMs：糖原合酶1（GYS1）、溶质载体家族3成员2（SLC3A2）、溶质载体家族7成员11（SLC7A11）、NADH：泛醌氧化还原酶核心亚基S1（NDUFS1）、线粒体3-氧代酰基-ACP合酶（OXSM）、富含亮氨酸的五肽重复序列（LRPPRC）、NADH：泛醌氧化还原酶亚基A11（NDUFA11）、线粒体NUBP铁硫簇组装因子（NUBPL）和NCK相关蛋白1（NCKAP1）。NDUFS1与NDUFA11、NUBPL和LRPPRC相互作用，而SLC3A2与SLC7A11相互作用。最佳机器学习模型显示为随机森林（RF）模型。G蛋白鸟嘌呤核苷酸结合蛋白α亚基q（GNAQ）与丙戊酸钠耐药相关。实验分析表明SLC7A11表达上调，SLC3A2和SLC7A11复合物形成数量增加，NDUFS1和NDUFA11复合物形成数量减少。本研究提供了二硫化物诱导的细胞死亡与癫痫之间关系的前所未有的证据。除了表明SLC7A11可能是癫痫的特异性DRM外，本研究还证明了两种二硫化物诱导的细胞死亡相关蛋白复合物的改变：SLC7A11-SLC3A2和NDUFS1-NDUFA11。
