Xu Jingwei, Wang Yuan, Han Yating, Liu Ningfeng, Liu Zhenming, Guo Huailian, Zou Xiajuan, Zhang Jun
Department of Neurology, People's Hospital, Peking University, Beijing, China.
Department of Neurology, The First Medical Center of PLA General Hospital, Beijing, China.
Front Neurol. 2022 May 23;13:860555. doi: 10.3389/fneur.2022.860555. eCollection 2022.
Migraine is a common neurological disease, but its pathogenesis is still unclear. Previous studies suggested that migraine was related to immunoglobulin G (IgG). We intended to analyze the immune characteristics of migraine from the perspective of IgG glycosylation and provide theoretical assistance for exploring its pathogenesis.
The differences in the serum level of IgG glycosylation and glycopeptides between patients with episodic migraine and healthy controls were analyzed by applying the poly(glycerol methacrylate)@chitosan (PGMA@CS) nanomaterial in combination with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS). We constructed a binary classification model with a feedforward neural network using PyTorch 1.6.0 in Python 3.8.3 to classify the episodic migraine and healthy control groups.
Twenty patients with migraine and 20 healthy controls were enrolled and the blood samples and clinical information were collected. Forty-nine IgG N-glycopeptides were detected in the serum of the subjects. The serum level of N-glycopeptide IgG1 G0-NF ( = 0.012) was increased in patients with migraine. The serum level of N-glycopeptide IgG3/4 G2FS ( = 0.041) was decreased in patients with migraine with family history of headache. It was found that the serum level of the IgG1 G1 ( = 0.004) and IgG2 G0 ( = 0.045) was increased in patients with migraine with aura, while the serum level of IgG2 G0N ( = 0.043) in patients with migraine with aura was significantly lower than that in patients with migraine without aura. In addition, a linear feedforward neural network (FFNN) was used to construct a binary classification model by detected IgG N-glycopeptides. The area under the curve (AUC) value of the binary classification model, which was constructed with 7 IgG N-glycopeptides, was 0.857, suggesting a good prediction performance. Among these IgG N-glycopeptides that were constructed the model, IgG1 G0-NF was overlapped with the differential IgG N-glycopeptide between patients with migraine and healthy controls detected with MALDI-TOF-MS.
Our results indicated that the serum level of N-glycopeptides IgG1 G0-NF might be one of the important biomarkers for the diagnosis of migraine. To the best of our knowledge, this is the first study about the changes of IgG N-glycosylation in patients with migraine by the method of MALDI-TOF-MS. The results indicated a relationship between the migraine and immune response.
偏头痛是一种常见的神经系统疾病，但其发病机制仍不清楚。以往研究提示偏头痛与免疫球蛋白G（IgG）有关。我们旨在从IgG糖基化角度分析偏头痛的免疫特征，为探索其发病机制提供理论支持。
应用聚（甲基丙烯酸甘油酯）@壳聚糖（PGMA@CS）纳米材料结合基质辅助激光解吸/电离飞行时间质谱（MALDI-TOF-MS）分析发作性偏头痛患者与健康对照者血清IgG糖基化水平及糖肽的差异。我们使用Python 3.8.3中的PyTorch 1.6.0构建了一个具有前馈神经网络的二元分类模型，以对发作性偏头痛组和健康对照组进行分类。
纳入20例偏头痛患者和20例健康对照者，采集血样及临床资料。在受试者血清中检测到49种IgG N-糖肽。偏头痛患者血清中N-糖肽IgG1 G0-NF水平升高（P = 0.012）。有头痛家族史的偏头痛患者血清中N-糖肽IgG3/4 G2FS水平降低（P = 0.041）。发现有先兆偏头痛患者血清中IgG1 G1水平升高（P = 0.004），IgG2 G0水平升高（P = 0.045），而有先兆偏头痛患者血清中IgG2 G0N水平显著低于无先兆偏头痛患者（P = 0.043）。此外，利用检测到的IgG N-糖肽，采用线性前馈神经网络（FFNN）构建二元分类模型。由7种IgG N-糖肽构建的二元分类模型的曲线下面积（AUC）值为0.857，提示具有良好的预测性能。在构建模型的这些IgG N-糖肽中，IgG1 G0-NF与MALDI-TOF-MS检测到的偏头痛患者与健康对照者之间的差异IgG N-糖肽重叠。
我们的结果表明，N-糖肽IgG1 G0-NF血清水平可能是偏头痛诊断的重要生物标志物之一。据我们所知，这是首次采用MALDI-TOF-MS方法研究偏头痛患者IgG N-糖基化变化的研究。结果提示偏头痛与免疫反应之间存在关联。
