α-synuclein overexpression and the microbiome shape the gut and brain metabolome in mice

Abstract

Pathological forms of α-synuclein contribute to synucleinopathies, including Parkinson's disease (PD). Most cases of PD arise from gene-environment interactions. Microbiome composition is altered in PD, and gut bacteria are causal to symptoms in animal models. We quantitatively profiled nearly 630 metabolites in the gut, plasma, and brain of α-synuclein-overexpressing (ASO) mice, compared to wild-type (WT) animals, and comparing germ-free (GF) to specific pathogen-free (SPF) animals (n = 5 WT-SPF; n = 6 ASO-SPF; n = 6 WT-GF; n = 6 ASO-GF). Many differentially expressed metabolites in ASO mice are also dysregulated in human PD patients, including amine oxides, bile acids and indoles. The microbial metabolite trimethylamine N-oxide (TMAO) strongly correlates from the gut to the plasma to the brain in mice, notable since TMAO is elevated in the blood and cerebrospinal fluid of PD patients. These findings uncover broad metabolomic changes that are influenced by the intersection of host genetics and microbiome in a mouse model of PD.

Fig. 1. αSyn overexpression and microbiome presence alter global metabolomic profiles in mice.

a tSNE plot of all metabolomic samples in this study, colored by tissue. b UpSet plot of unique and shared metabolite sets across all samples. Intersection size describes the number of metabolites with a significant genotype×microbiome interaction effect (p < 0.05). The dots below the bar chart indicate the sample source of the metabolites. Singular dots with no vertical lines connecting to other tissues indicate a set of metabolites which are uniquely altered in a particular tissue. c Stacked barplots depicting average effect sizes for biochemical classes containing molecules significantly associated (p < 0.05) with either a genotype, microbiome, or interaction effect in a linear regression model.

Fig. 2. The gut microbiome shapes metabolism similarly across the GI tract.

a Lollipop plot of relative enrichment/depletion of significantly altered (p < 0.05) metabolites in gut samples, organized by molecular class. Data points are colored by sample source and sized by number of metabolites. Volcano plots showing metabolites significantly enriched by genotype or microbiome status in the colon (b) and cecal contents (d). Scatterplots of metabolites in the colon (c) and cecal contents (e) affected by the microbiome and genotype in a linear model. Colored points indicate metabolites significantly (p ≤ 0.05) altered by genotype and/or microbiome. Labeled points indicate the top 10 metabolites with the most significant genotype×microbiome interaction effect.

Fig. 3. The genotype and microbiome alter metabolite levels differentially across the brain.

a Lollipop plot of relative enrichment/depletion of metabolites whose levels are significantly altered (p < 0.05) by genotype, microbiome, or their interaction in the brain. Data points are colored by tissue and sized by number of metabolites. b Volcano plots showing metabolites significantly enriched by genotype or microbiome status in the striatum. c Scatterplot of metabolites in the striatum affected by the microbiome and genotype in a linear model. Colored points indicate metabolites significantly (p < 0.05) altered by genotype and/or microbiome. Labeled points indicate the top 10 metabolites with the most significant genotype×microbiome interaction effect.

Fig. 4. αSyn overexpression and microbiome influence circulating metabolites.

a Lollipop plot of relative enrichment/depletion of metabolites whose levels are significantly altered (p < 0.05) by genotype, microbiome, or their interaction in plasma. Data points are colored by metabolite class and sized by number of metabolites. b Volcano plots showing metabolites significantly enriched by genotype or microbiome status in the plasma. c Scatterplot of metabolites in the plasma affected by the microbiome and genotype in a linear model. Colored points indicate metabolites significantly (p < 0.05) altered by genotype and/or microbiome. Labeled points indicate the top 10 metabolites with the most significant genotype×microbiome interaction effect.

Fig. 5. A microbially-produced metabolite links the gut-brain axis in ASO mice.

a Lollipop plot showing metabolites whose levels in each indicated sample correlate strongly (Spearman’s Rho > 0.75) with the level of the same metabolite in plasma. b Network visualization of metabolites shown in a. A metabolite is connected to a tissue node if its abundance there is strongly correlated with its abundance in plasma. TMAO is the most highly connected metabolite, with strong correlations to plasma levels in six different tissue types.