Society: AGA

This session will cover the pathogenic and non-pathogenic constituents of the gut microbiome that induce local and systemic inflammatory response and play an important role in disease development.

Background: The intestinal epithelium is an essential interface between the microbial ecosystem and the underlying mucosa, orchestrating both immune-mediated and metabolic functions. Mitochondrial signaling and metabolism control the phenotype and function of intestinal epithelial cells (IECs) and mitochondrial perturbation is associated with chronic inflammation. We hypothesize that disruption of microbial-metabolic circuits and impaired metabolic flexibility of IECs contributes to the initiation and progression of tissue injury and chronic inflammation. Methods: To address the role of mitochondrial impairment in epithelial homeostasis, conditional deletion of Heat shock protein 60 (Hsp60) was induced in IECs of both specific-pathogen free and germ-free (GF) Hsp60^Δ/ΔIEC mice as well as in a colitis mouse model (Hsp60^Δ/ΔIEC;Il10^-/-) and in AhR-deficient mice (Hsp60^Δ/ΔIEC;AhR^-/-). Histological analysis, 16S rRNA and shallow shotgun sequencing were performed. Different molecular mechanisms were analyzed in Hsp60-deficient organoids. Results: Hsp60 deletion in IECs induces metabolic injury, defined as a disrupted mitochondrial metabolism accompanied by an irregular crypt architecture causing transient tissue injury. Mitochondrial perturbation is associated with dysbiotic changes in the microbiota including a rapid drop in species richness and changes in the bacterial community profile of Hsp60^Δ/ΔIECmice with an increased abundance of Bacteroides spp.. Tissue lesions are completely absent in the distal colon of GF mice, indicating bacterial contribution to metabolic injury. In Hsp60^Δ/ΔIEC;Il10^-/- mice, mitochondrial impairment of the epithelium accelerates inflammation, paralleled by a persistent Bacteroides spp. expansion. Clinical relevance of mitochondrial stress signaling of the epithelium and increased abundance of Bacteroides spp. are confirmed in a combined analysis of three IBD cohorts (N=560 patients). Upon colonization of GF Hsp60^Δ/ΔIECwith the synthetic minimal consortium OligoMM¹² inducing epithelial lesions, B. caecimuris expands during metabolic injury, strengthening our hypothesis. Indeed, B. caecimuris triggers tissue aberration when mono-colonizing Hsp60^Δ/ΔIECmice, confirming its deleterious role in tissue homeostasis. While lack of Aryl hydrocarbon Receptor (AhR) signaling upon metabolic injury is lethal in Hsp60^Δ/ΔIEC;AhR^-/-mice, antibiotic exposure abrogates the severity of the phenotype, highlighting that AhR ligands protect from metabolic injury. Conclusion: Bacteroides spp. were identified as key players regulating mitochondrial metabolism and host-sensing pathways critically affected cellular plasticity upon metabolic injury. Mitochondria emerge as intriguing interceptors of milieu signals in the intestine, and unresolved metabolic injury is a novel concept in the pathogenesis of intestinal inflammation.

Background: Primary sclerosing cholangitis (PSC) is an immune-mediated, chronic cholestatic liver disease often associated with a unique phenotype of ulcerative colitis (UC), suggesting a bidirectional interplay of the gut-liver axis. Several studies have provided evidence that the microbiota is distinct in PSC, as compared to UC and controls. However, the duodenal mucosa-associated microbiome (MAM) as the key site for mucosal inflammation and dysbiosis in PSC remains unexplored.
Aim: To characterize the duodenal, ileal, and colonic MAM in PSC patients with/without UC, UC patients without PSC, and controls.
Methods: Biopsies from the duodenum, terminal ileum, ascending colon, and rectum from 69 controls, 46 UC patients without PSC and 20 PSC patients with/without UC were collected. Biopsies were cryopreserved at -80C prior to total DNA extraction and production of 16S rRNA gene amplicon (V6-V8) libraries. Data processing and filtering and were performed with QIIME2 and taxonomic assignments made using SILVA v138 rRNA database. Samples producing >500 reads were retained; microbiota diversity and statistical analyses were performed using various packages within the R environment.
Results: The MAM profiles of 122 subjects (Controls n=65, UC n=38, PSC n=19) were retained after quality filtering and exceeding the read threshold. Grouping data from all sites, Kruskal-Wallis tests suggests the Shannon diversity metrics measured for the controls were significantly greater than those of UC patients with/ without PSC, and those with PSC alone (p=0.01), which was confirmed via pairwise comparisons using Dunn’s test (p_adj=0.02). Although the PSC patients without UC had the lowest Shannon diversity, there was no significant difference with the UC groups, Fig. 1. Some site-specific differences in microbiota composition were observed between the groups, with the relative abundances of Haemophilus and Alloprevotella spp. at the duodenum being less in the controls. At the terminal ileum, the relative abundance of Anaerostipes spp. was greater and that of Butyricoccus less in the controls as compared to other two groups. In contrast the relative abundances of Lachnoclostridium spp. at the terminal ileum and Collinsella spp. at the ascending colon were discriminatory of the PSC subjects from the other two groups, while the relative abundance of Anaerostipes spp. at the rectum was reduced in the PSC and UC groups, Fig. 2.
Conclusions: MAM profiles were different between controls, UC patients with/without PSC, and patients only with PSC. These variations were detectable in the MAM communities of the proximal, distal small intestine, and the colon. Such findings suggest the gut-liver-microbiome axis is not restricted to the colon in UC and/or PSC and warrants further investigation to advance our understanding of the pathophysiology and treatment options for both conditions.

<b>Figure 1. </b>Box-violin plot for overall microbiome alpha diversity (Shannon) with Kruskal-Wallis and Dunn’s test.

Figure 1. Box-violin plot for overall microbiome alpha diversity (Shannon) with Kruskal-Wallis and Dunn’s test.

<b>Figure 2. </b>Relative differential analysis on top 10 genera in each site using ANOVA test. DU, Duodenum; TI, Terminal ileum; AC, ascending colon; R, rectum.

Figure 2. Relative differential analysis on top 10 genera in each site using ANOVA test. DU, Duodenum; TI, Terminal ileum; AC, ascending colon; R, rectum.

Background: Preterm neonates are at risk for late-onset sepsis (LOS) which is the leading cause of morbidity and mortality at neonatal intensive care units (NICUs). The pathophysiology of LOS is complex and remains poorly understood. Given the widespread use of antibiotics at NICUs, we hypothesize that intestinal fungi play an important role in preterm intestinal barrier integrity and immunity, and that this may contribute to LOS development caused by Escherichia coli (E.coli).

Methods: Preterm neonates from two Dutch neonatal intensive care units who were diagnosed with E. coli caused LOS (n=10) were matched to healthy preterm neonates (n=10) based on birth center, birth weight and antibiotic use. Fecal samples were collected longitudinally prior to (5 and 4 days), during and after (4 and 5 days) diagnosis of LOS. Mycobiome (fungal microbiome) profiles were determined with Internal Transcribed Spacer 1 (ITS1) sequencing on isolated DNA. A gradient boosting classification model was used to identify amplicon sequence variants (ASVs) most predictive of LOS caused by E.coli. Subsequently, the possible role of the fungus Candida parapsilosis (Cp) was studied in the preterm intestine in vitro. Primary human fetal mucosal immune cells and intestinal organoids were derived from fetal intestinal tissues (18-21 weeks of gestation). Cp-induced loss of barrier integrity was assessed by trans epithelial resistance (TEER). E.coli translocation was determined by basolateral E.coli colony forming unit concentrations after a 6 hour infection period. Inflammatory responses were measured by cytokine levels using cytokine bead array (CBA).

Results: Gradient boosting modeling identified Candida parapsilosis ASVs as most predictive of LOS caused by E.coli. Moreover, Cp abundance was increased in feces of preterm neonates with LOS already 5 days prior to diagnosis. When human fetal colon organoid monolayers (n=5 donors) were exposed to Cp derived conditioned media (CDM) for 5 days, relative TEER levels were significantly decreased by 22% (p=0.016). In addition, a trend of increased relative E.coli translocation was observed after 5 days Cp CDM exposure followed by 6 hours of E.coli infection (n=4, p=0.1). Inflammatory IL8 responses of colon organoid monolayers (n=3) towards Cp CDM were increased basolaterally compared to control conditions (1240 pg/ml and 671 pg/ml respectively, p=0.1). When exposed to living Cp for 24 hours, the inflammatory response of mucosal intestinal immune cells was most prominently characterized by increased IL1β cytokine secretion compared to control conditions (587.08 pg/ml and 33.27 pg/ml respectively; n=4, p=0.008).

Conclusions: Our results suggest that increased abundance of intestinal Candida parapsilosis in preterm neonates could precede E.coli LOS by potentially mediating intestinal barrier dysfunction and pro-inflammatory responses.

Tim De Meij

Appearances