Gut microbiota is causally associated with poststroke cognitive impairment through lipopolysaccharide and butyrate

Abstract

Background: Poststroke cognitive impairment (PSCI) is prevalent in stroke patients. The etiology of PSCI remains largely unknown. We previously found that stroke induces gut microbiota dysbiosis which affects brain injury. Hereby, we aimed to investigate whether the gut microbiota contributes to the pathogenesis of PSCI.

Methods: 83 stroke patients were recruited and their cognitive function were measured by Montreal Cognitive Assessment (MoCA) scores 3 months after stroke onset. The peripheral inflammatory factor levels and gut microbiota compositions of the patients were analyzed. Fecal microbiota transplantation from patients to stroke mice was performed to examine the causal relationship between the gut microbiota and PSCI. The cognitive function of mice was evaluated by Morris water maze test.

Results: 34 and 49 stroke patients were classified as PSCI and non-PSCI, respectively. Compared with non-PSCI patients, PSCI patients showed significantly higher levels of gut Enterobacteriaceae, lipopolysaccharide (LPS) and peripheral inflammation markers. Consistently, stroke mice that received microbiota from PSCI patients (PSCI mice) presented a higher level of Enterobacteriaceae, intestinal Toll-like receptor-4 (TLR4) expression, circulating LPS, LPS-binding protein (LBP) and inflammatory cytokines, and a lower level of fecal butyrate, severer intestine destruction and cognitive impairment than mice that received microbiota from nPSCI patients (nPSCI mice). In addition, we observed exacerbations in blood-brain barrier (BBB) integrity, microglial activation, neuronal apoptosis in the CA1 region of the hippocampus, and Aβ deposition in the thalamus of PSCI mice in comparison with nPSCI mice. Intraperitoneal injection of LPS after stroke caused similar pathology to those seen in PSCI mice. Supplementation with sodium butyrate (NaB) via drinking water rescued these detrimental changes in PSCI mice.

Conclusions: Our data indicate a cause-effect relationship between gut microbiota and PSCI for the first time, which is likely mediated by inflammation-regulating metabolites including LPS and butyrate.

Keywords: Fecal microbiota transplantation; Hippocampal apoptosis; Lipopolysaccharide; Post-stroke cognitive impairment; β-Amyloid.

Fig. 1

PSCI patients exhibit excessive peripheral inflammation and increased abundance of Enterobacteriaceae. a Mini-Mental State Examination (MMSE) and the Montreal Cognitive Assessment scale (MoCA) score between poststroke cognitive impairment (PSCI) patients (n = 34) and non-PSCI patients (n = 49) 3 months after stroke. Lower scores indicate severer cognitive impairment. b Levels of interleukin-6 (IL-6), IL-1β, lipopolysaccharide (LPS), LPS-binding protein (LBP), d lactate (DLA) and erythrocyte sedimentation rate (ESR) in peripheral blood. c Average relative abundances of prevalent microbiota at the family levels in the two groups. d Linear discriminant analysis effect size (LEfSe) shows bacterial taxa with significantly different abundances between the two groups. e Heatmap shows the association between gut microbiota and markers inflammation by Spearman’s rank correlation. f Changes in the relative abundance of Enterobacteriaceae from stroke onset to 3 months after stroke. Data are expressed as mean ± SEM. Student’s t test comparing MoCA, MMSE, inflammatory factors and abundances of Enterobacteriaceae at the same timepoint between PSCI and nPSCI patients; Wilcoxon matched-pairs signed rank test comparing abundances of Enterobacteriaceae between two timepoints of PSCI or nPSCI patients, *P < .05, **P < .01, ***P < .001, ****P < .0001

Fig. 2

PSCI mice receiving present higher Enterobacteriaceae abundance and lower fecal butyrate level than nPSCI mice. a Experimental design. After acclimatization for 1 week, mice received antibiotics in the drinking water for 2 weeks, followed by middle cerebral artery occlusion (MCAO). 3 days after MCAO, gut microbiota from PSCI patients or non-PSCI patients was transferred to stroke mice by FMT. Sodium butyrate (NaB) was provided via drinking water. b Principal-coordinate analysis (PCoA) plot of unweighted UniFrac distances between the three groups. c Average relative abundances of prevalent microbiota at the family levels in the three groups. d LEfSe shows bacterial taxa with significantly different abundances between the three groups. e Relative abundances of Enterobacteriaceae and fecal butyrate levels in the three groups. f Correlations between Enterobacteriaceae and butyrate by Spearman’s rank correlation. Data are expressed as mean ± SEM, n = 15–17 mice per group, Nonparametric Kruskal–Wallis test comparing the abundance of Enterobacteriaceae and fecal butyrate level. *P < .05, **P < .01

Fig. 3

Stroke mice receiving FMT from PSCI patients display cognitive decline. a Modified Neurological Severity Score (mNSS) 3 days after stroke. b Escape latency of the three groups from days 1 to 5 by Morris water maze. c Duration of stay in quadrant III of the three groups. d Platform crossing frequency of the three groups. e Representative images of motion traces of the three groups. Data are expressed as mean ± SEM, n = 10–12 mice per group, one-way ANOVA comparing mNSS, duration of stay and platform crossing frequency. Escape latency was compared by repeated-measure ANOVA, ns, not significant, *P < .05, **P < .01

Fig. 4

PSCI-associated gut microbiota promotes hippocampal apoptosis and thalamic Aβ deposition. a Morphology of the ileum was assessed using H&E staining (scale bar = 100 μm) and the average villus height and crypt depth was analyzed. b Expression of intestinal TLR-4 protein in the three groups. c Levels of LPS, LBP, IL-6, IL-1β and TNF-α in the peripheral blood. d Expression of cerebral tight junction proteins ZO-1, Occludin and Claudin-4 in the three groups. e Double immunostaining for Iba-1 (microglial marker) and f Nissl staining, and g TUNEL staining were performed in the hippocampal CA1 region to detect apoptotic neurons, and h Aβ staining was performed in the thalamus (scale bar = 50 μm). The cells were counted per 40× field of view (FOV). Data are expressed as mean ± SEM, n = 10–12 mice per group, one-way ANOVA, *P < .05, **P < .01, ***P < .001, ****P < .0001

Fig. 5

Intraperitoneal injection of LPS after stroke causes similar pathology to those seen in mice receiving PSCI-associated gut microbiota. a Experimental design. After acclimatization for 1 week, mice were subjected to MCAO. Three days after MCAO, mNSS score was assessed, followed by intraperitoneal injection of PBS or LPS daily. After 1 month, Morris water maze was performed after which mice were sacrificed. b Results of the mNSS score and c–f Morris water maze between the two groups. g Morphology of the ileum was assessed using H&E staining (scale bar = 100 μm) and the average villus height and crypt depth was analyzed. h Expression of intestinal TLR-4 protein in the two groups. i Levels of LPS, LBP, IL-6, IL-1β and TNF-α in the peripheral blood. j Expression of cerebral tight junction proteins Occludin and Claudin-4 in the two groups. k Nissl staining in the hippocampal CA1 region of the two group. l Double immunostaining for Iba-1 and m TUNEL staining in the hippocampal CA1 region, and n Aβ staining in the thalamus (scale bar = 50 μm). The cells were counted per 40× FOV. Data are expressed as mean ± SEM, n = 9–10 mice per group, Student’s t test, escape latency was compared by repeated-measure ANOVA, ns, not significant, *P < .05, **P < .01, ***P < .001, ****P < .0001

Fig. 6

PSCI-associated gut microbiota promotes hippocampal apoptosis and thalamic Aβ deposition. The gut microbiota of PSCI patients is hallmarked by an increased abundance of Enterobacteriaceae and a decreased level of butyrate, which result in a disrupted gut barrier. The LPS constantly travels through the leaky gut into circulation, causing a chronic peripheral inflammation. The continuous inflammation destroys the integrity of BBB, leading to constant infiltration of peripheral LPS and inflammatory cytokines, which promote the neuronal apoptosis in the CA1 region of hippocampus, a brain region that is critical for the cognition. In addition, the low-grade chronic inflammation in the brain promotes the deposition of Aβ plaque in the thalamus, a toxic chemical that is associated with cognitive impairment