Characterisation of LPS+ bacterial extracellular vesicles along the gut-hepatic portal vein-liver axis

Abstract

Gut microbiome dysbiosis is a major contributing factor to several pathological conditions. However, the mechanistic understanding of the communication between gut microbiota and extra-intestinal organs remains largely elusive. Extracellular vesicles (EVs), secreted by almost every form of life, including bacteria, could play a critical role in this inter-kingdom crosstalk and are the focus of present study. Here, we present a novel approach for isolating lipopolysaccharide (LPS)+ bacterial extracellular vesicles (bEV^LPS) from complex biological samples, including faeces, plasma and the liver from lean and diet-induced obese (DIO) mice. bEV^LPS were extensively characterised using nanoparticle tracking analyses, immunogold labelling coupled with transmission electron microscopy, flow cytometry, super-resolution microscopy and 16S sequencing. In liver tissues, the protein expressions of TLR4 and a few macrophage-specific biomarkers were assessed by immunohistochemistry, and the gene expressions of inflammation-related cytokines and their receptors (n = 89 genes) were measured using a PCR array. Faecal samples from DIO mice revealed a remarkably lower concentration of total EVs but a significantly higher percentage of LPS+ EVs. Interestingly, DIO faecal bEV^LPS showed a higher abundance of Proteobacteria by 16S sequencing. Importantly, in DIO mice, a higher number of total EVs and bEV^LPS consistently entered the hepatic portal vein and subsequently reached the liver, associated with increased expression of TLR4, macrophage markers (F4/80, CD86 and CD206), cytokines and receptors (Il1rn, Ccr1, Cxcl10, Il2rg and Ccr2). Furthermore, a portion of bEV^LPS escaped liver and entered the peripheral circulation. In conclusion, bEV could be the key mediator orchestrating various well-established biological effects induced by gut bacteria on distant organs.

Keywords: bacterial extracellular vesicles; dysbiosis; gut microbiome; inflammation; lipopolysaccharide.

FIGURE 1

Characterisation of bacterial extracellular vesicles (bEV) isolated from faeces. (a) Bar graph representing the average body weight (mean ± SEM) of lean and diet induced obese (DIO) mice (n = 5/group) at 28 weeks of age. ****p ≤ 0.0001 (b) Bar graph representing the concentration of faecal EV total (fEV^total) from lean and DIO mice as estimated by nanoparticle tracking analysis (NTA) (n = 5/group). Concentration (particles/mL) was normalised with the corresponding weight of faecal matter. ****p ≤ 0.0001 (C) Bar graph representing the size (nm) of fEV^total from lean and DIO mice as estimated by NTA (n = 5/group). *p = 0.03 (d) Representative micrograph captured at 98,000X magnification showing the size and surface expression of LPS on lean and DIO fEV^total by immunogold labelling coupled with transmission electron microscopy (n = 3/group) (scale bar: 100 nm). (e) Representative micrograph showing the size and surface expression of LPS on lean and DIO fEV^total by super‐resolution microscopy (n = 3/group) (scale bar: 100 nm). (f) Bar graph showing expression of LPS‐positive EV in lean and DIO fEV^total as estimated by flow cytometry (n = 3/group). ****p ≤ 0.0001 (g) Bar graph showing expression of OmpC‐positive EV in lean and DIO fEV^total as estimated by flow cytometry (n = 3/group). ***p = 0.001 (h and i) Bar graphs representing the concentration (p = 0.058) and size (nm) of bEV^LPS in lean and DIO mice faeces as estimated by NTA (n = 4‐5/group). (j) Bar graph showing 16S rRNA sequencing data representing the percentage of total phyla in lean and DIO bEV^LPS (n = 3/group, *p = 0.0497).

FIGURE 2

TLR4 expression in the intestine. TLR4 expression in the small (a) and large (b) intestinal tissue in lean and DIO mice (n = 5/group) was measured by immunohistochemistry. Images were captured using Olympus VS120 Slide Scanner and quantified by subjectively scoring them. The bar graph in the left panel represents TLR4 positivity (%), whilst the right panel shows representative images (20X magnification, scale bar: 100 μm). **p = 0.003.

FIGURE 3

Characterisation of total EVs isolated from hepatic portal vein. (a) Bar graph representing the concentration of EV^total from lean and DIO mice in hepatic portal vein plasma as estimated by NTA (n = 5/group). **p = 0.002 (b) Representative micrograph captured at 98,000X magnification showing the size and surface expression of LPS on lean and DIO hepatic portal vein plasma‐EV^total by immunogold labelling coupled with transmission electron microscopy (n = 3/group) (scale bar: 100 nm). (c) Bar graph showing the number of LPS+ EV, normalised by volume, in lean and DIO EV^total isolated from hepatic portal vein plasma by flow cytometry (n = 5/group). *p = 0.03 (d) Bar graph showing 16S rRNA sequencing data representing the percentage of total phyla in lean (n = 5) and DIO (n = 3) bEV^LPS from hepatic portal vein plasma (*p = 0.040).

FIGURE 4

Isolation and characterisation of total EVs from liver tissue. (a) Representative micrograph captured at 98,000X magnification showing the size and surface expression of LPS in lean and DIO liver EV^total by immunogold labelling coupled with transmission electron microscopy (n = 3/group) (scale bar: 100 nm). (b) Bar graph showing expression of LPS‐positive EVs in lean and DIO liver‐EV^total as estimated by flow cytometry (n = 5/group). p = 0.081 (c) Bar graph showing expression of OmpC‐positive EV in lean and DIO liver EV^total as estimated by flow cytometry (n = 3/group). (d) TLR4 expression in the liver tissues of lean and DIO mice (n = 5/group) was measured by immunohistochemistry. Images were captured using Olympus VS120 Slide Scanner and quantified using VisioPharm digital pathology analysis software. The bar graph in the right panel represents TLR4 positivity (%), whilst the left panel shows representative images (at 20X magnification, scale bar: 100 μm). Non‐parametric Kolmogorov‐Smirnov test was utilised for analysis of significance. **p = 0.0079 (e) Immunofluorescence staining for macrophage markers F4/80, CD86 and CD206 in lean and DIO mice liver tissue (n = 3/group). The left side shows representative images (10X magnification) of expression of macrophage markers (CD86, F4/80 and CD206) and the nuclei stain (DAPI) whilst the bar graph on the right side represents the mean intensity of expression of F4/80, CD86 and CD206 respectively. ****p < 0.0001 (f) Left panel: All genes analysed by PCR array are tabulated. The housekeeping genes are highlighted in green. Right panel: Volcano plot representing the differential changes in the genes related to mouse inflammatory cytokines and receptors in the liver tissues of lean and DIO mice (n = 3/group). The dotted line on the x‐axis represents the cutoff at 1.5‐fold. The dotted line on the y‐axis represents the p‐value of 0.05. Genes showing a statistically significant change in lean and DIO mice liver tissue are shown in red.

FIGURE 5

Characterisation of bEV in cardiac plasma from lean and DIO mice. (a and b) Bar graphs representing the size (nm) and concentration (particles/mL) of EV^total in cardiac plasma from lean and DIO mice as estimated by NTA (n = 5/group). (c) Bar graph showing expression of LPS‐positive EVs in lean and DIO mice cardiac plasma EV^total as estimated by flow cytometry (n = 5/group). *p = 0.03 (d and e) Bar graphs representing the size (nm) and concentration (particles/mL) of bEV^LPS in lean and DIO mice cardiac plasma as estimated by NTA (n = 5/group). (f) Representative micrograph captured at 98,000X magnification showing the size and surface expression of LPS on lean and DIO mice plasma EV^total by IG/TEM (n = 3/group) (scale bar: 100 nm). (g) Representative micrograph showing the size and surface expression of LPS on lean and DIO mice plasma EV ^total by super‐resolution microscopy (n = 3/group) (scale bar: 100 nm). *p ≤ 0.05.