Liu Po-Chun, Cheng Shao-Ying, Li Chih-Chi, Wang Yu-Ke, Tseng Yufeng Jane, Kuo Ming-Che
Cathay General Hospital, Taipei, Taiwan.
Department of Neurology, National Taiwan University Hospital Bei-Hu Branch, Taipei, Taiwan.
Front Neurosci. 2025 Aug 13;19:1623165. doi: 10.3389/fnins.2025.1623165. eCollection 2025.
The alterations in the gut microbial network in multiple system atrophy (MSA) remain poorly understood. This study aimed to identify key gut microbial interaction networks in MSA through comprehensive multimodal analyses.
Demographic information and frozen fecal specimens were collected from 119 participants [MSA, = 26; Parkinson's disease (PD), = 66; healthy control (HC), = 27]. Raw amplicons of the bacterial 16S rRNA V3-V4 gene region were processed using two methods: DADA2-denoising and clustering into operational taxonomic units. We conducted univariate and multivariable analyses to assess the differential abundance of bacterial genera and predicted metabolic pathways using four statistical methods: ANCOM, ANCOM-BC, ALDEx2, and MaAsLin 2. Interbacterial interactions were assessed using four correlation and two network analyses.
We consistently observed lower levels of in MSA patients and lower levels of in PD patients compared with HCs (q < 0.05), both before and after adjusting for comorbidities, diet, and constipation status. The random forest classifiers effectively differentiated between MSA and PD, achieving high AUCs (0.75-0.78) in 5-fold cross-validation. A significant positive interbacterial interaction between group and was uniquely observed in MSA patients. Additionally, we identified an increase in the ARGORNPROST-PWY pathway (L-arginine degradation, q = 0.003) and a decrease in the PWY-6478 pathway (GDP-D-glycero-α-D-manno-heptose biosynthesis, q = 0.015) in MSA patients compared with HCs.
Future studies are warranted to determine whether fecal microbiome-derived signatures can serve as reliable biomarkers for MSA.
多系统萎缩（MSA）患者肠道微生物网络的变化仍知之甚少。本研究旨在通过全面的多模态分析确定MSA中关键的肠道微生物相互作用网络。
收集了119名参与者的人口统计学信息和冷冻粪便样本[MSA，n = 26；帕金森病（PD），n = 66；健康对照（HC），n = 27]。细菌16S rRNA V3-V4基因区域的原始扩增子使用两种方法进行处理：DADA2去噪和聚类为可操作分类单元。我们使用四种统计方法（ANCOM、ANCOM-BC、ALDEx2和MaAsLin 2）进行单变量和多变量分析，以评估细菌属的差异丰度和预测的代谢途径。使用四种相关性分析和两种网络分析评估细菌间的相互作用。
在调整合并症、饮食和便秘状态前后，我们一致观察到与健康对照相比，MSA患者的[具体细菌名称1]水平较低，PD患者的[具体细菌名称2]水平较低（q < 0.05）。随机森林分类器能够有效区分MSA和PD，在5折交叉验证中获得了较高的AUC（0.75 - 0.78）。在MSA患者中独特地观察到[细菌组名称]组和[另一细菌名称]之间存在显著的正细菌间相互作用。此外，与健康对照相比，我们发现MSA患者中ARGORNPROST-PWY途径（L-精氨酸降解，q = 0.003）增加，PWY-6478途径（GDP-D-甘油-α-D-甘露庚糖生物合成，q = 0.015）减少。
未来有必要进行研究以确定粪便微生物组衍生特征是否可作为MSA的可靠生物标志物。
