ENGINEERED HIGH ACETATE PRODUCING SACCHAROMYCES BOULARDII ATTENUATES INFLAMMATION AND SHOWS DOSE-DEPENDENT HISTOLOGIC IMPROVEMENT IN DSS-INDUCED COLITIS

A key feature in intestinal immunity is the dynamic intestinal barrier, which separates the host from resident and pathogenic microbiota through a mucus gel impregnated with antimicrobial peptides. The mechanisms underlying the maintenance and function of this intestinal barrier are not completely understood. Using a mouse forward genetic screen for defects of intestinal homeostasis, we have found a mutation in Tvp23b, which conferred susceptibility to chemically induced and infectious colitis. Golgi apparatus membrane protein TVP23 homolog B (TVP23B) is a transmembrane protein conserved from yeast to humans. The protein is localized to the intestinal epithelium and its deficiency in the hematopoietic extrinsic compartment was essential to protecting against colitis. We found that TVP23B controls the homeostasis of Paneth cells and function of goblet cells in vivo, leading to a decrease in antimicrobial peptides as well as a defective and more penetrable mucus layer. As a result, Tvp23b^-/- mice displayed decreased barrier function and a loss of host-microbe separation. Using unbiased glycomics, TVP23B-deficient colonocytes have a loss of core-3 O-glycosylation of colonic proteins, which is the major O-glycosylation present on gel forming mucins. TVP23B binds with another Golgi protein, YIPF6, which is similarly critical for intestinal homeostasis. The Golgi proteomes of YIPF6 and TVP23B-deficent colonocytes have a common deficiency of several critical glycosylation enzymes, including those necessary for core-3 glycosylation of mucins. TVP23B is necessary for the formation of the sterile mucin layer of the intestine and its absence disturbs the balance of host and microbe in vivo.

Background: Endoplasmic reticulum (ER) stress is critical in the pathogenesis of inflammatory bowel disease (IBD). ER stress is observed under various pathologic situations, including inflammation. Recent studies have revealed that chronic ER stress augments the suppressive phenotypes of immune cells. Existing data suggest that XBP1, a key component of ER stress pathways is a “synergistic” transcription factor for FoxP3 with critical involvement in Treg development. We hypothesize that ER stress during inflammation is required to activate Treg development and function via XBP1.

Methods: We aimed to validate this hypothesis using both in vivo and in vitro methods. In vivo we used a TNF-driven model of Crohn’s disease, the B6.129S-Tnf^tm2GKl/Jarn strain (TNF^ΔARE/+; MGI:3720980) mouse model, characterized by a spontaneous chronic murine ileitis which we crossed with the XBP1^loxpCD4^lckcre to assess XBP1 deletion in an inflamed mouse and confirmed the effects in an adoptive transfer (CD45RB^hi) model of colitis using XBP1^loxpCD4^lckcre CD45RB^hi cells. In vitro, Treg development was assessed in a Treg conversion assay using ER stress activators. Treg function was measured using a 3-day suppression assay and IL-10 output was measured by ELISA and by intracellular cytokine staining.

Results: Isolated CD4⁺CD25⁺FoxP3⁺ Tregs from TNF^ΔARE/+-Foxp3GFP mice expressed high levels of XBP1 (20-fold) compared to WT FoxP3-GFP mice, suggesting that the ER stress pathways are activated in FoxP3⁺ Treg cells with inflammation. This was recapitulated using in vitro conversion assays identifying an increase in the spliced (active) form of XBP1. We next demonstrated that glucose depletion and tunicamycin (ER stress activators) increased Treg development by 32% in vitro. With increased expression of XBP1 with inflammation we next evaluated conditional XBP1 knockout mice for development of Tregs but only identified a mild decrease in Treg development without inflammation. We thus hypothesized that inflammation would trigger a more significant effect, therefore we crossed the XBP1^loxpCD4^lckcre to the TNF^ΔARE/+. These mice had decreased Treg numbers and more severe inflammation compared to the TNF^ΔARE/+ littermates. This was confirmed using the CD45RB^hi (XBP1^loxpCD4^lckcre) adoptive transfer model of colitis.

Conclusion: While ER stress has been shown to be important in intestinal inflammation, models have focused on the deleterious effects. Here we demonstrate that ER stress has beneficial anti-inflammatory effects as well and that XBP1 is required for Treg development in the inflamed intestine. Understanding ER stress in a critical cell type for intestinal homeostasis and disease pathogenesis is indispensable. elucidating these mechanisms will impact our understanding of IBD pathogenesis and will provide novel methods to expand more potent Tregs ex vivo or in vivo.

Background: Crohn’s disease (CD) is a chronic inflammatory disease that arises from defects in epithelial barrier function and mucosal tolerance. Intraepithelial lymphocytes bearing the γδ T cell receptor (γδ IEL) continuously survey the villous epithelium and confer protection against injury and infection; however, their contribution to the etiology of CD remains unclear. Recent reports indicate that γδ IEL number is decreased in ileal CD patients and these T cells exhibit reduced expression of CD39, which helps to convert ATP to adenosine to suppress inflammation. Epithelial butyrophilins (BTN) shape the composition of the γδ IEL compartment, and reduced expression of BTN3/8 is associated with IBD. However, the kinetics by which BTN expression contribute to the regulation of γδ IEL number and composition during the progression of chronic ileitis have yet to be elucidated. Methods: Histological analysis was conducted on ileal sections of 4-16 wk old TNF^+/+ (WT) or TNF^DARE/+ mice, which develop spontaneous CD-like ileitis. Ileal γδ IELs and whole ileal homogenates from WT and TNF^DARE/+ mice (4-10 wks) were analyzed by flow cytometry and qPCR, respectively. Immunostaining or RNAScope were performed on ileal sections. Results: Although the histological onset of ileitis is first detected in TNF^DARE/+ mice at wk 8, we observed a 2.6-fold decrease in the frequency and 3.2-fold reduction in the number of γδ IELs in TNF^DARE/+ mice beginning at wk 5 (p<0.001, p<0.05). Expression of butyrophilin-like (Btnl) 1/6, the murine homologs of BTN3/8, are also decreased at 5 wks of age in TNF^DARE/+ mice compared to WT (p<0.0001, p<0.001). Further, expression of HNF4g, a regulator of BTNL expression in the small intestine, is reduced in TNF^DARE/+ mice at 4 wks of age. As a result, the frequency of Vγ7+ IELs is reduced at 6 wks of age in TNF^DARE/+ mice with slight increases in Vγ1+ and Vγ4+ IEL subpopulations. These Vγ7- IELs were CD5+ CD44- CD8α- CD39- suggesting recent recruitment from the periphery. In support of this, we found that CD103 expression is upregulated on Vγ1+ and Vγ4+ T cells in the mesenteric lymph node (MLN) in 4 wk old TNF^DARE/+ mice with a significant accumulation in the MLN observed at 6 wks of age. Lastly, CD39 expression was reduced across all γδ IEL subsets in 4 wk old TNF^DARE/+ mice compared to WT. Suppression assays demonstrate that CD39+ γδ IELs are capable of inhibiting IFNγ production by effector memory CD8 T cells. Conclusions: Together, our data demonstrate that the γδ IELs compartment is dysregulated early in the pathogenesis of chronic ileitis in TNF^DARE/+ mice with a loss of the epithelial HNF4g/BTNL axis leading to impaired γδ IEL survival and an influx of immature peripheral γδ T cells. We posit that the loss of tissue-resident immunoregulatory γδ T cells contributes to the initiation of chronic ileitis.

BACKGROUND & AIMS: T-helper (Th)1 effector cells are critical in the pathogenesis of autoimmune inflammatory diseases, including inflammatory bowel diseases (IBD), but the intrinsic factors that control the pro-inflammatory programs of Th1 cells remain unclear. The stimulator of interferon genes (STING), an intracellular DNA sensor, has been shown to regulate innate anti-viral infection and the pathogenesis of various cancers. However, whether and how intrinsic STING signaling in Th1 cells regulates intestinal homeostasis is still unknown.
METHODS: Dextran sodium sulfate (DSS)-induced colitis and wild-type/STING-deficient CD4⁺T cell adoptive transfer models were used to analyze the role of STING in regulating colitis. The effect of STING agonists on Th1 cells was determined by flow cytometry, RNA sequencing, metabolic assays, and mitochondrial functions. 16s rRNA sequencing and germ-free mice were used to investigate whether the microbiota mediate STING regulation of T cell function. The in vivo effect of STING agonists in preventing and treating murine colitis was determined. The expression and role of STING in human T cells were also determined by quantitative PCR, immunofluorescent staining, and flow cytometry.
RESULTS: STING^-/- CD4⁺ T cells induced more severe colitis in Rag1^-/- mice, which was characterized by the reduction of IL-10 expression in intestinal Th1 cells, but not Th17 cells or regulatory T cells. Activation of STING transformed pro-inflammatory IFNγ⁺ Th1 cells into self-limited IL-10⁺ IFNγ⁺ Th1 cells, which were dramatically less pathogenic in inducing colitis. STING promoted Th1 IL-10 production by inducing STAT3, but not STAT1, translocation into nuclear and mitochondria, which promoted Blimp1 expression and mitochondrial oxidation, respectively. Blockade of glucose or glutamine-derived oxidation, but not lipid-derived oxidation, suppressed CD4⁺ T cell production of IL-10 induced by STING agonist, which highlighted the importance of glucose and glutamine in this process. Gut microbiota were altered in STING^-/- mice. However, the transfer of fecal bacteria from wild-type and STING^-/- mice into germ-free mice induced similar levels of intestinal IL-10-producing CD4⁺ T cells, indicating that the altered gut microbiota did not mediate STING effects on intestinal CD4⁺ T cell production of IL-10. Translationally, treatment of STING agonists suppressed both DSS-induced acute colitis and CD4⁺CD45RB^hi T cell-induced chronic colitis. Intestinal STING⁺ CD4⁺ T cells were increased in the inflamed colonic tissue of IBD patients, and STING agonists upregulated IL-10 production in human CD4⁺ T cells.
CONCLUSIONS: These findings establish a crucial role of T cell-intrinsic STING in switching off the pathogenesis of Th1 cells in intestinal inflammation, which provides a foundation for developing STING agonists as therapeutics for colitis.

It has been hypothesized that the probiotic potential of the yeast Saccharomyces boulardii (Sb) is associated with its acetate production. This SCFA is of interest in IBD due to its cross-feeding potential with beneficial butyrate-producing bacteria and low toxicity to epithelial cells . Our previous in vitro work supports this hypothesis (Deleu et al, 2022) but still requires in vivo validation. Therefore, we evaluated the effect of different engineered Sb strains producing variable amounts of acetate in DSS-induced colitis in mice.

Nine week old female C57/Bl6 mice (N=120) were allocated to 12 treatment groups receiving drinking water or 2.75% DSS in combination with PBS (control), Baker’s yeast (non-probiotic control), SDH1 (non-acetate producing Sb), ENT (transient acetate producing Enterol strain-probiotic control), SbP (high acetate producing Sb) and ENT3 (extra high acetate producing Sb). Disease activity including weight loss, diarrhoea and the presence of occult blood was scored daily. On day 7, the DSS groups were transferred to regular drinking water and on day 14 mice were sacrificed. Colonic tissue and blood were collected for resp. histologic and cytokine analysis.

Disease activity, determined by the area under the curve (Fig. 1A), in DSS subgroups was lower for SbP compared to PBS and Baker’s yeast (both p<0.05). Remarkably, Sb SDH1 showed even higher disease activity compared to the Sb strains ENT, SbP and ENT3 (all p<0.05). At sacrifice, macroscopic damage score (Fig. 1C) in DSS subgroups was lower for SbP and ENT3 (both p<0.05) compared to Sb SDH1 and the colon weight/length-ratio (Fig. 1B) was decreased for ENT and SbP compared to Sb SDH1 (resp. p=0.06 and p=0.08). Higher histologic inflammation (Fig. 1D) was noted in the non- or only transient-acetate producing strains on DSS compared to healthy PBS control (all p<0.05), whereas this increase was not observed for both high-acetate producing strains SbP and ENT3 on DSS (p=NS). Lower IL1β, IL2 and IL4 concentrations (Fig. 2) for DSS groups on SbP and ENT3 compared to DSS groups on Sb SDH1 and ENT were observed (all p<0.05). In contrast, IL10, TNFα and KC/GRO were lower for the DSS groups on Sb SDH1 and ENT compared to DSS groups on SbP, ENT3, and even PBS for IL2 and 4 (all p<0.05).

Engineered high acetate producing Sb strains show a significant trend towards improved attenuation of DSS-induced colitis compared to the parent Sb strain on disease activity, macroscopical damage score and show production-dependent response on histology. Mixed pro-inflammatory serum profiles were observed potentially pointing in the direction of other effects of acetate accumulation. Together with our previous in vitro work, these data indicate a role for Sb-produced acetate in attenuating inflammation.