IL-22 alters gut microbiota composition and function to increase aryl hydrocarbon receptor activity in mice and humans

Abstract

Background: IL-22 is induced by aryl hydrocarbon receptor (AhR) signaling and plays a critical role in gastrointestinal barrier function through effects on antimicrobial protein production, mucus secretion, and epithelial cell differentiation and proliferation, giving it the potential to modulate the microbiome through these direct and indirect effects. Furthermore, the microbiome can in turn influence IL-22 production through the synthesis of L-tryptophan (L-Trp)-derived AhR ligands, creating the prospect of a host-microbiome feedback loop. We evaluated the impact IL-22 may have on the gut microbiome and its ability to activate host AhR signaling by observing changes in gut microbiome composition, function, and AhR ligand production following exogenous IL-22 treatment in both mice and humans.

Results: Microbiome alterations were observed across the gastrointestinal tract of IL-22-treated mice, accompanied by an increased microbial functional capacity for L-Trp metabolism. Bacterially derived indole derivatives were increased in stool from IL-22-treated mice and correlated with increased fecal AhR activity. In humans, reduced fecal concentrations of indole derivatives in ulcerative colitis (UC) patients compared to healthy volunteers were accompanied by a trend towards reduced fecal AhR activity. Following exogenous IL-22 treatment in UC patients, both fecal AhR activity and concentrations of indole derivatives increased over time compared to placebo-treated UC patients.

Conclusions: Overall, our findings indicate IL-22 shapes gut microbiome composition and function, which leads to increased AhR signaling and suggests exogenous IL-22 modulation of the microbiome may have functional significance in a disease setting. Video Abstract.

Keywords: Aryl hydrocarbon receptor; Colitis; Gastrointestinal microbiome; IL-22; Indole; Inflammatory bowel disease; Tryptophan.

Fig. 1

IL-22Fc consistently alters microbiota composition along the GI tract in mice. A Shannon’s diversity following IL-22Fc or aRW treatment. B NMDS ordination plot of Bray-Curtis distances. Dashed ellipses represent the 95% CI for the centroid of each stratification group. r² calculated by PERMANOVA. For A and B, triangle, square, and circle markers indicate ileal, colonic, and fecal samples, respectively. Teal and red markers indicate animals treated with IL-22Fc (n=8) or isotype control (n=8), respectively. C Taxonomic tree depicting all bacterial genera detected within the ileum, colon, or stool of IL-22Fc or aRW-treated mice. The outer heatmap displays the difference in mean, normalized abundance of a given bacterial genera between IL-22Fc and aRW-treated mice (Log₂(∆)) in the ileum, colon, or stool. All non-zero Log₂(∆) values have a BH-adjusted p value <0.05. Highly IL-22Fc responsive genera, defined as the top ten genera based on average Log₂(∆) and significant enrichment or depletion in IL-22Fc-treated animals across multiple tissues, are labeled. Note that only seven IL-22Fc-enriched genera met these criteria

Fig. 2

IL-22Fc alters microbial tryptophan metabolism in mice. A Fecal concentrations of L-Trp, I3C, and 5-HIAA following IL-22Fc or aRW treatment. B Log₂ fold change from baseline in fecal concentrations of L-Trp, I3C, and 5-HIAA. C Predicted metagenomic counts, as calculated by PICRUSt2, mapping to the L-tryptophan biosynthesis pathway in IL-22Fc or aRW-treated animals. D Predicted metagenomic counts, as calculated by PICRUSt2, of microbial enzymes capable of synthesizing tryptophan (tryptophan synthase), indole (tryptophanase), and 5-HIAA (aldehyde dehydrogenase). Statistical significance determined by rank-sum test (*p<0.05, **p<0.01). Triangle, square, and circle markers indicate ileal, colonic, and fecal samples, respectively. Teal and red markers indicate animals treated with IL-22Fc (n=8) or isotype control (n=8), respectively

Fig. 3

Microbially derived AhR ligand availability in the intestine increases following IL-22Fc treatment in mice. A AhR activity following treatment with prepared fecal samples obtained from IL-22Fc or aRW-treated animals, reported as Log₂(fold change over unstimulated wells). B Scatterplot of predicted fecal metagenomic counts of tryptophan synthase, tryptophanase, and aldehyde dehydrogenase vs. fecal AhR activity. Spearman’s rank correlation coefficient (rho) was calculated to determine correlation strength. C AhR activity, reported as Log₂(fold change over unstimulated wells), following treatment with 100uM of either L-Trp, I3C, or 5-HIAA (see the “Methods” section). D AhR activity following treatment with sterile-filtered bacterial culture supernatant from bacterial isolates representing either highly IL-22Fc depleted (n=14) or enriched (n=8) genera, reported as Log2(fold change over matched sterile culture media). E Proportion of bacterial isolates representing highly IL-22Fc-depleted or -enriched genera capable of activating AhR in vitro, as defined by an in vitro AhR activity statistically greater than that of matched sterile culture media (see Fig. S5). For A and B, teal and red markers indicate animals treated with IL-22Fc (n=8) or isotype control (n=8), respectively. For panels A, C, and D, all samples were measured in triplicate and statistical significance was determined by T test. For panel E, the statistical significance was determined by Barnard’s test. For all panels, *p<0.05 and **p<0.01

Fig. 4

Efmarodocokin alfa administration increases fecal AhR activity in humans. A Fecal concentrations of L-Trp, I3C, and IPA at screening in HV (n=69) compared to UC (n=46). P value determined by rank-sum test. B AUC of Log₂(fold change from screening) for fecal AhR activity in enrolled UC patients following administration of efmarodocokin alfa (n=8) or placebo (n=5). C Scatterplot of fecal AhR activity vs. fecal concentrations of I3C and IPA at screening only (n=23) and at all visits (n=104) in enrolled UC patients. Spearman’s rank correlation coefficient (rho) was calculated to determine correlation strength. D Mean AUC of Log₂(fold change from screening) for normalized abundance of IL-22Fc-depleted (n=7) or -enriched (n=5) genera detected in enrolled UC patients receiving efmarodocokin alfa treatment. E Mean ΔAUC (AUC in efmarodocokin alfa-treated UC patients—AUC in placebo-treated UC patients) of IL-22Fc-depleted (n=7) or -enriched (n=5) genera detected in enrolled UC patients. For panels B, D, and E, only enrolled UC patients with a complete sample collection history (screening, day 29, day 43, day 64, and day 85) were considered for AUC calculations. P value determined by Student’s T test