In utero human intestine contains maternally derived bacterial metabolites

Abstract

Understanding when host-microbiome interactions are first established is crucial for comprehending normal development and identifying disease prevention strategies. Furthermore, bacterially derived metabolites play critical roles in shaping the intestinal immune system. Recent studies have demonstrated that memory T cells infiltrate human intestinal tissue early in the second trimester, suggesting that intestinal immune education begins in utero. Our previous study reported a unique fetal intestinal metabolomic profile with an abundance of several bacterially derived metabolites and aryl hydrocarbon receptor (AHR) ligands implicated in mucosal immune regulation. To follow up on this work, in the current study, we demonstrate that a number of microbial byproducts present in fetal intestines in utero are maternally derived and vertically transmitted to the fetus. Notably, these bacterially derived metabolites, particularly short chain fatty acids and secondary bile acids, are likely biologically active and functional in regulating the fetal immune system and preparing the gastrointestinal tract for postnatal microbial encounters, as the transcripts for their various receptors and carrier proteins are present in second trimester intestinal tissue through single-cell transcriptomic data.

Figure 1.. Sample separation and differential expression of individual metabolites.

(A) t-distributed stochastic neighbor embedding (t-SNE ) plot using all metabolites. (B) Heatmap showing normalized abundance of the top 20 expressed metabolites in the fetal GI samples (SI and LI), as well as their abundance in the meconium (SI Mec and LI Mec), PV, and maternal decidua samples. (C), (D) and (E) Volcano plots of differentially abundant metabolites between the GI and decidua groups, GI and meconium groups, and GI and PV groups respectively. Top ten most abundant metabolites are labeled with the metabolite name.

Figure 2.. Pathway enrichment.

(A), (B) and (C) Integrated pathway analysis for differentially altered pathways between GI and decidua groups, between GI and meconium groups, and between GI and PV groups. The length of the bar is proportional to the q value. Pathways with positive values on the x axis (orange bar) are those enriched for in the decidua, meconium, and PV respectively. Those with negative values (blue bars) are pathways enriched for in the GI samples.

Figure 3.. Correlation between tissue groups of microbial metabolites, xenobiotics, and fetal metabolites.

(A) Schematic of how the correlation between tissue would identify the source of bacterial metabolites. (B) Heatmap showing normalized abundance of the 41 microbial associated metabolites across sample types. (C) Pairwise correlation matrix of the 41 microbial metabolites between paired tissue samples from subject 1146. (D) Boxplots visualizing the correlations between GI and decidua groups, and between GI and meconium groups based on 41 microbial metabolites, 47 xenobiotics, and 13 fetal-derived metabolites respectively from Supplementary Table S4. Red asterisk points represent the pairwise correlations between tissue samples from subject 1146. *P < 0.05, ***P<0.001.

Figure 4.. ANOVA analysis of SCFA, bile acids and aromatic acids across tissues.

(A-C) Boxplots of individual metabolite’s abundance for primary and secondary bile acids, SCFA, and aromatic amino acids and aromatic acids. *P < 0.05, **P<0.01, ***P<0.001. (D) Heatmap showing that microbial metabolites without significant difference across tissue groups (upper panel) and microbial metabolites significantly enriched in decidua samples (lower panel).

Figure 5.. Cell type specific expression of genes associated with bile acid and SCFA transport and signaling.

Uniform Manifold Approximation and Projection (UMAP) plot visualization of bile acid and SCFA associated transport and signaling gene expression in fetal small intestine epithelial (A) and fetal immune cells (B). (C) Boxplots of selected individual gene expression in small intestine across developmental stages (related to Figure 1 (33)). *P < 0.05, **P<0.01, ***P<0.001.