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.

Background: The intestinal microbiota early in life play a key role in the development and functional maturation of the immune system. Our previous work showed that the introduction of solid food at weaning is accompanied by a vast expansion of gut microbiota which induce a strong but transient immune response in the intestine termed “weaning reaction”. This reaction is necessary to prevent colitis and colorectal cancer susceptibility in adult mice. However, to which extend dietary compounds upon weaning influence the intestinal microbiota and the susceptibility to develop gut inflammation later in life remains unknown.

Aims: This project aims to determine how a fiber-based nutrition early in life affects immune responses and the susceptibility to develop colitis and colorectal cancer later in life.

Methods: Pregnant mice raised in specific pathogen-free (SPF) conditions were fed a fiber-rich or fiber-poor diet until 4 weeks postdelivery and co-housed on normal chow after weaning. Using a model of dextran sulfate sodium-induced colitis and Apc^Min/+ mice developing spontaneous intestinal cancer, the consequences of feeding different amounts of dietary fibers early in life on the severity of colonic inflammation and carcinogenesis in adult mice was assessed by multiple clinical and immunological tests. 16S-ribosomal RNA-sequencing was used to determine the dietary effect on fecal microbiota composition and flow cytometry to study changes of intestinal immune cells in the offspring at weaning. The causal relationships between diet-shaped microbial composition and colitis development later in life was studied by colonization of germ-free (GF) mice by gut microbiota of mice exposed to different amounts of dietary fibers.

Results: Our results show that the dietary fiber amounts and types during weaning determine colitis and colorectal cancer susceptibility later in life. Dietary fibers strongly affected fecal bacterial phyla composition of the weaning mouse, leading to expansion of Firmicutes in response to a high fiber diet, while a low-fiber diet increased Proteobacteria which dictate the type of immune responses at weaning. We also show that B cells are key immune populations for the regulation of the weaning reaction. Transfer of these microbial communities to neonatal GF mice, but not to adult, was sufficient to recapitulate the disease severity observed in SPF mice. This indicates that early life is a time window of opportunity to prevent gut inflammation later in life.

Conclusion: Altogether, our results indicate that the composition of gut microbiota imprints the immune system early in life which influences the long-term susceptibility to develop colitis and colorectal cancer. This work provides fundamental evidence that the intestinal commensal bacteria early in life can prevent the development of inflammatory bowel diseases and colorectal cancer.

Background/Aim: Inactive rhomboid protein 2 (iRhom2) regulates the release of tumor necrosis factor-α (TNF-α) by activating TNF-α converting enzyme. TNF-α is a well-known pro-inflammatory cytokine that also affects gastrointestinal (GI) motility. This study aimed to assess the effect of iRhom2 in regulating GI motility and explore the mechanisms of immune-enteric nervous system (ENS)-microbiome interaction.
Methods: GI motility in wild-type (WT) and iRhom2 knockout (iRhom2^-/-) mice was assessed by fecal output, whole gastrointestinal transit time (WGTT), and a colonic bead expulsion test. GI motility was also evaluated after antibiotic treatment or a 6-week cohousing to elucidate the role of the microbiome. Lamina propria mononuclear cells of the colon were prepared and colon-derived macrophages were polarized to M1 or M2 activation states by fluorescence-activated cell sorting. The ENS components were quantified by three-dimensional imaging using an IMARIS system. The gut microbiota profile was determined by 16S rDNA sequencing. In addition, GI motility evaluation and immune-microbiome analysis were performed on young (3 months old) and old (24 months old) mice.
Results: Compared to WT mice, iRhom2^-/- mice exhibited increased GI motility, as indicated by the frequency of defecation, total fecal weight, WGTT, and a colonic bead expulsion test (Figure 1). As compared to WT mice, iRhom2^-/- mice showed a lower M1/M2 ratio and fewer neuronal nitric oxide synthase-positive neurons, while the total volume of choline acetyltransferase-positive neurons increased significantly. The iRhom2 deletion led to increased microbial diversity and Firmicutes/Bacteroidetes ratio (Figure 2). After antibiotic treatment or 6-week cohousing, there was no significant difference in GI motility, and both WT and iRhom2^-/- mice had similar microbiota composition. Additionally, aging decreased GI motility and increased the M1/M2 ratio of macrophages, which is associated with a rise in pro-inflammatory cytokines. Old mice also showed decreased microbial diversity and a microbiome composition distinct from young mice.
Conclusion: iRhom2 is associated with the regulation of GI motility through immune-ENS-microbiome interaction and could be considered a potential target for the treatment of GI dysmotility disorders.

Background: Lactobacillus acidophilus has been shown to enhance the intestinal epithelial tight junction (TJ) barrier in a strain-specific and Toll-like receptor 2 (TLR2)-dependent manner. However, the intracellular mechanisms involved in LA enhancement are not fully understood. The nucleotide-binding oligomerization domain containing 1 (NOD1) is a cytosolic pathogen recognition receptor (PRR) that has been shown to recognize bacterial products and pathogen-associated molecular patterns (PAMPs). However, the association between LA, TLR2 and NOD1has not been examined. Aim: the major aim of this study was to delineate the intracellular mechanism involved in LA enhancement of intestinal epithelial TJ barrier in an in-vivo mouse model. Methods: In-vivo model system of recycling intestinal perfusion of live mice) was used to assess intestinal TJ permeability. Results: 1) LA oral gavage (1x10⁹ cfu/ml) caused a significant decrease in dextran 10kda flux in-vivo in a strain-specific manner, suggesting an enhancement in mouse intestinal epithelial TJ barrier. 2) LA oral gavage caused an increase in TLR2 protein expression in mouse intestinal tissue, which was associated with an increase in NOD1 expression in a strain-specific manner. 3) LA treatment did not cause enhancement of mouse intestinal barrier in TLR2 deficient mice (TLR2 ^-/-), nor in NOD1 deficient mice (NOD1 ^-/-). 4) LA treatment did not cause an increase in NOD1 expression in TLR2 deficient mice (TLR2 ^-/-). 5) LA treatment did not cause an increase in TLR2 expression in NOD1 deficient mice (NOD1 ^-/-). 6) Immunostaining images showed that LA caused TLR2 localization with NOD1 at the apical membrane in mouse intestinal cells, that was aberrant on TLR2 deficient mice. Conclusions: These findings show that TLR2/NOD1 interaction plays an integral role in LA enhancement of mouse intestinal epithelial TJ barrier. The LA-mediated TLR2/NOD1-dependent enhancement of the intestinal epithelial TJ barrier might be a potential therapeutic tool for treatment or prevention of gastrointestinal diseases associated with a defective intestinal TJ barrier.

Lauren Kaminsky

Appearances