Profiling the human luminal small intestinal microbiome using a novel ingestible medical device

Abstract

The invasive nature of sample collection for studying the small intestinal (SI) microbiome often results in its poor characterization. This study evaluated a novel ingestible medical device (MD) for SI luminal sample collection. A monocentric interventional trial (NCT05477069) was conducted on 15 healthy subjects. Metagenomics, metabolomics and culturomics assessed the MD's effectiveness in characterizing the healthy SI microbiome and identifying potential biomarkers. The SI microbiota differed significantly from the fecal microbiota, displaying high inter-individual variability, lower species richness, and reduced alpha diversity. A combined untargeted and semi-targeted LC-MS/MS metabolomics approach identified a distinct SI metabolic footprint, with bile acids and amino acids being the most abundant classes of metabolites. Host and host/microbe-derived bile acids were particularly abundant in SI samples. The application of a fast culturomics approach to two SI samples enabled species-level characterization, resulting in the identification of 90 bacterial species, including five potential novel species. The present study demonstrates the efficacy of our novel sampling MD in enabling comprehensive SI microbiome analysis through an integrative multi-omics approach, allowing the identification of distinct microbiome signatures between SI and fecal samples.

Keywords: Healthy subjects; Microbiome; Multi-omics; Sampling medical device; Small intestine.

Figure 1:

Photograph of the sampling MD evaluated in this clinical investigation.

Figure 2:

Culturomics workflow in anaerobic and aerobic conditions performed on small intestinal contents (the figure was created in BioRender).

Figure 3:

Microbial composition of the small intestinal contents (n=14) and fecal samples (n=14). (a) Bacterial richness and (b) bacterial diversity in small intestinal contents and fecal samples. The samples (SI content vs feces) for each subject are connected by dotted lines. (c) Microbial community β-diversity at ASV level, which was demonstrated using principal coordinates analysis (PCoA) of the Bray-Curtis distance matrix. The samples (SI content vs feces) for each subject are connected by dotted lines. (d) Relative abundance of the 20 most predominant bacterial genera in small intestinal contents. Remaining genera are summarized as “Other”. Each column represents one subject. (e) Relative abundance of the 20 most abundant bacterial genera in fecal samples. Other genera are summarized as “Other”. Each column represents one subject.

Figure 4:

Comparison of metabolite abundance between the SI and fecal samples revealed significant differences across a wide range of compounds. (a) Principal Component Analysis (PCA) of the filtered features (n=574 metabolites) obtained in positive ionization mode. To assess statistical differences between the metabolomics profiles, PERMANOVA based on 999 Monte Carlo permutations was performed (F-value: 45.419; R-squared: 0.63595; p-value permutations: 0.001). (b) Volcano plot of the metabolites showing significant differences in the abondance of metabolites between the SI and the feces groups using a fold change (FC) threshold of 2 and a t-test threshold of 0.05. The log transformed FC and P-values are represented on x- and y-axes.

Figure 5:

Relative quantification by semi-targeted metabolomics of the metabolites of interest in small intestinal contents and fecal samples (in unit/mg for fecal and unit/μL for small intestinal content). GCA: Glycocholic acid; GCDCA: Glycochenodeoxycholic acid; TCA: Taurocholic acid; TCDCA: Taurochenodeoxycholic acid; IAA: Indole-3-acetic acid; IPA: Indole-3-propoinic acid. *Indicates significant differences (p<0.05) between groups, **(p-0.001) and ***(p<0.0001).

Figure 6:

Determination of the Bile Acids (BAs) profile in SI contents (n=14) compared to their corresponding fecal samples (n=14). (a) Boxplots of the overall concentration of BAs in SI (n=14) and fecal (=n=14) samples. Concentration of total Microbe-derived BAs and total Host-derived BAs were depicted and compared separately as well as total Host/microbe-derived BAs (nmol/mg for fecal samples and μM for SI contents). (b) Boxplots of the most common BAs quantified in SI content or fecal sample. Microbe-derived BAs colored in green, Host-derived BAs colored in yellow and Host/Microbe-derived BAs colored in blue. DCA: Deoxycholic acid; LCA: Lithocholic acid; GCDCA: Glycochenodeoxycholic acid; GCA: Glycocholic acid; TCDCA: Taurochenodeoxycholic acid; TCA: Taurocholic acid; CDCA: Chenodeoxycholic acid; CA: Cholic acid and UDCA: Ursodeoxycholic acid. (c) Global profile of bile acids in SI contents and fecal samples. Each column represents a given sample from one subject. BAs have been grouped according to their chemical family, Unc: unconjugated BA. (d) Volcano plot demonstrating the log2 fold changes in the peak area abundances of the 14 subjects between their SI contents and fecal samples. *Indicates significant differences (p<0.05) between groups, **(p-0.001), ***(p<0.0001) and ****(p<0.00001).

Figure 7:

Isolated bacterial species by means of culturomics for small intestinal contents sampled from two different healthy volunteers. (a) Total number of microbial species identified for each sample. SI1: SI content 1 and SI2: SI content 2. (b) Percentage of identified species for each bacterial phylum for the SI content 1. “Other” contains two phyla: Thermodesulfobacteriota (1,4%) and Campylobacterota (1,4%). (c) Percentage of identified species for each bacterial phylum for the SI content 2. (d) Total number of distinct microbial species identified during both culturomics and including the number of the discovered new bacterial species. (e) Core bacterial species identified in both SI contents (figure created in BioRender).