Gut Microbiome Signatures Are Biomarkers for Cognitive Impairment in Patients With Ischemic Stroke

Abstract

Post-stroke cognitive impairment (PSCI) is a common neuropsychiatric complication of stroke. Mounting evidence has demonstrated a connection between gut microbiota (GM) and neuropsychiatric disease. Our previous study revealed the changes in the GM in a mouse model of vascular dementia. However, the characteristic GM of PSCI remains unclear. This study aimed to characterize the GM of PSCI and explored the potential of GM as PSCI biomarkers. A total of 93 patients with ischemic stroke were enrolled in this study. The patients were divided into two groups according to their MoCA scores 3 months after stroke onset. Clinical data and biological variables were recorded. GM composition was analyzed using 16S ribosomal RNA sequencing, and the characteristic GM was identified by linear discriminant analysis Effect Size (Lefse). Our results showed that <i>Proteobacteria</i> was highly increased in the PSCI group compared with the post-stroke non-cognitive impairment (PSNCI) group, the similar alterations were also observed at the class, order, family, and genus levels of <i>Proteobacteria</i>. After age adjustments, the abundance of <i>Firmicutes</i>, and its members, including <i>Clostridia</i>, <i>Clostridiales</i>, <i>Lachnospiraceae</i>, and <i>Lachnospiraceae_other</i>, were significantly decreased in the age-matched PSCI group compared with the PSNCI group. Besides, the GM was closely associated with MoCA scores and the risk factors for PSCI, including higher baseline National Institute of Health Stroke Scale score, higher homocysteine (Hcy) level, higher prevalence of stroke recurrence, leukoaraiosis, and brain atrophy. The KEGG results showed the enriched module for folding, sorting and degradation (chaperones and folding catalysts) and the decreased modules related to metabolisms of cofactors and vitamins, amino acid, and lipid in PSCI patients. A significant correlation was observed between PSCI and the abundance of <i>Enterobacteriaceae</i> after adjustments (<i>P</i> = 0.035). Moreover, the receiver operating characteristic (ROC) models based on the characteristic GM and <i>Enterobacteriaceae</i> could distinguish PSCI patients from PSNCI patients [area under the curve (AUC) = 0.840, 0.629, respectively]. Our findings demonstrated that the characteristic GM, especially <i>Enterobacteriaceae</i>, might have the ability to predict PSCI in post-stroke patients, which are expected to be used as clinical biomarkers of PSCI.