Multi-level analysis of gut microbiome extracellular vesicles-host interaction reveals a connection to gut-brain axis signaling

Abstract

Microbiota-released extracellular vesicles (MEVs) have emerged as a key player in intercellular signaling. However, their involvement in the gut-brain axis has been poorly investigated. We hypothesize that MEVs cross host cellular barriers and deliver their cargoes of bioactive compounds to the brain. In this study, we aimed to investigate the cargo capacity of MEVs for bioactive metabolites and their interactions with the host cellular barriers. First, we conducted a multi-omics profiling of MEVs' contents from ex vivo and stool samples. Metabolomics analysis identified various neuro-related compounds encapsulated within MEVs, such as arachidonyl-dopamine, gabapentin, glutamate, and N-acylethanolamines. Metaproteomics unveiled an enrichment of enzymes involved in neuronal metabolism, primarily in the glutamine/glutamate/gamma-aminobutyric acid (GABA) pathway. These neuro-related proteins and metabolites were correlated with Bacteroides spp. We isolated 18 Bacteroides strains and assessed their GABA production capacity in extracellular vesicles (EVs) and culture supernatant. A GABA-producing Bacteroides finegoldii, released EVs with a high GABA content (4 µM) compared to Phocaeicola massiliensis. Upon testing the capacity of MEVs to cross host barriers, MEVs exhibited a dose-dependent paracellular transport and were endocytosed by Caco-2 and hCMEC/D3 cells. Exposure of Caco-2 cells to MEVs did not alter expression of genes related to intestinal barrier integrity, while affected immune pathways and cell apoptosis process as revealed by RNA-seq analyses. In vivo, MEVs biodistributed across mice organs, including the brain, liver, stomach, and spleen. Our results highlight the ability of MEVs to cross the intestinal and blood-brain barriers to deliver their cargoes to distant organs, with potential implication for the gut-brain axis.

Importance: Microbiota-released extracellular vesicles (MEVs) have emerged as a key player in intercellular signaling. In this study, a multi-level analysis revealed presence of a diverse array of biologically active molecules encapsulated within MEVs, including neuroactive metabolites, such as arachidonyl-dopamine, gabapentin, glutamate, and N-acylethanolamines, and gamma-aminobutyric acid (GABA). Metaproteomics also unveiled an enrichment of neural-related proteins, mainly the glutamine/glutamate/GABA pathway. MEVs were able to cross epithelial and blood-brain barriers in vitro. RNA-seq analyses showed that MEVs stimulate several immune pathways while suppressing cell apoptosis process. Furthermore, MEVs were able to traverse the intestinal barriers and reach distal organs, including the brain, thereby potentially influencing brain functionality and contributing to mental and behavior.

Keywords: Bacteroides; extracellular vesicles; gut microbiome; gut-brain axis; neuroactive metabolites; γ-aminobutyric acid (GABA).

Fig 1

Physical and metabolic characteristics of MEVs. (A) TEM image depicting isolated vesicles. TEM pictures presented are for MEVs isolated from D2S, D2F, and D3S, respectively. (B) Nanoparticle size analysis of the isolated MEVs using tunable resistive pulse sensing (TRPS, IZON) showing the concentration as a function of the size range, the graph presents the particle size range of MEVs of D3F. (C) Lipid profiles of MEVs determined by nLC-ESI-MS/MS analysis in the positive mode. (D ) Identification of neuro-related metabolites within MEVs using nLC-ESI-MS/MS analysis in the positive mode.

Fig 2

Correlation between predominant microbiota bacterial species and neuroactive metabolites. The relative abundances of microbiota bacterial species, detected in >50% of the tested samples, were correlated with the identified neuroactive metabolites from the same samples. (A) Correlation analysis of stool-derived MEVs. (B) Correlation analysis of fermented MEVs. The color scale represents the spearman correlation coefficient (R), and asterisks indicate the significance levels (*P < 0.05, **P < 0.01); blue arrows highlight taxa with positive correlations and red arrows highlight taxa with negative correlations.

Fig 3

Metaproteome of MEVs is enriched in proteins with neuroactive potential. (A) Number of proteins identified in both stool and ex vivo-generated MEVs. (B and C) Predominance of the identified proteins at the level of COG categories (B) and their distribution in each sample (C). (D) Relative abundance of bacterial families identified as the source of MEVs proteome using MetaLab (25). (E) Glutamine/glutamate/GABA pathway is enriched in MEVs; red arrows highlight the proteins identified by the current study. (F) Dominant proteins with neuroactive potential and their corresponding microbiota taxa source.

Fig 4

EVs released by Bacteroides spp. exhibit distinct neurometabolic and protein contents. (A) Quantification of GABA production in 18 representative Bacteroides spp. isolates by competitive ELISA. (B) Measurement of GABA and glutamic acid intensity in cell-free supernatants and EVs of B. finegoldii (high GABA producer) and P. massiliensis (low GABA producer) by nLC-nESI-MS/MS metabolomics analysis. (C) Identification of different enzyme categories in the proteomes of B. finegoldii and P. massiliensis using nano-LC-MS/MS proteomics analysis.

Fig 5

Transport of MEVs across the intestinal and blood-brain barriers via paracellular transport and endocytosis. (A–F) Paracellular transport of FITC-labeled MEVs at various concentrations across Caco-2 (A) and hCMEC/D3 (D) cells, presented as A.U. TEER values of Caco-2 cells (B and C) and hCMEC/D3 (E and F) before and after MEV addition. Asterisks indicate statistical significance (*P < 0.05, **P < 0.0, ***P < 0.001). (G) Endocytosis of Cyanine 7 (Cy7)-labeled MEVs by different cell types, including Caco-2, hCMEC/D3, and RIN-14B.

Fig 6

RNA-seq reveals that interaction of MEVs with Caco-2 cells modulates host immune system. (A) Volcano plot illustrating differentially expressed genes in MEVs-treated Caco-2 cells compared to control. The x-axis represents the log2 fold change, while the y-axis represents −log10 (P-value). Significance threshold is set at adjusted P-value ≤0.05. (B) Dot plot presenting enrichment results of significantly enriched biological processes and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways for upregulated and downregulated genes, with a P_adj threshold cut-off of 0.05. Enrichment significance is indicated by bubble color, while bubble size corresponds to gene count in the term.

Fig 7

Biodistribution of MEVs in C57BL/6 mice. (A and B) Average radiant efficiency of various organs following injection (A) or oral gavage (B) of Cy7-labeled MEVs compared to control. Asterisks indicate statistical significance (**P < 0.01, ***P < 0.001). Fluorescence imaging (IVIS Lumina XR) of different organs following intravenous injection of MEVs.

Fig 8

Proposed pathway mechanism of microbiota-released extracellular vesicles in mediating host-microbiome communication and transporting bioactive metabolites.